EGF-independent activation of cell-surface EGF receptors harboring mutations found in Gefitinib-sensitive lung cancer

Abstract

Several somatic mutations within the tyrosine kinase domain of epidermal growth factor receptor (EGFR) have been identified that predict clinical response of non-small-cell lung carcinoma (NSCLC) patients to gefitinib. To test the hypothesis that these mutations cause constitutive EGF receptor signaling, and to investigate its mechanistic basis, we expressed representative examples in a null background and analysed their biochemical properties. Each mutation caused significant EGF-independent tyrosine phosphorylation of EGFR, and allowed the receptor to promote Ba/F3 cell mitogenesis in the absence of EGF, arguing that these are oncogenic mutations. Active mutated receptors are present at the cell surface and are fully competent to bind EGF. Recent structural studies show that the inactive EGFR tyrosine kinase domain is autoinhibited by intramolecular interactions between its activation loop and alphaC helix. We find that mutations predicted to disrupt this autoinhibitory interaction (including several that have not been described in NSCLC) elevate EGF-independent tyrosine kinase activity, thus providing new insight into how somatic mutations activate EGFR and other ErbB family members.

Full text

ORIGINAL ARTICLE
EGF-independent activation of cell-surface EGF receptors harboring
mutations found in gefitinib-sensitive lung cancer
SH Choi
1
, JM Mendrola
1
and MA Lemmon
Department of Biochemistry and Biophysics, University of Pennsylvania School of Medicine, Philadelphia, PA, USA
Several somatic mutations within the tyrosine kinase
domain of epidermal growth factor receptor (EGFR) have
been identified that predict clinical response of non-small-
cell lung carcinoma (NSCLC) patients to gefitinib. To test
the hypothesis that these mutations cause constitutive
EGF receptor signaling, and to investigate its mechanistic
basis, we expressed representative examples in a null
background and analysed their biochemical properties.
Each mutation caused significant EGF-independent tyr-
osine phosphorylation of EGFR, and allowed the receptor
to promote Ba/F3 cell mitogenesis in the absence of EGF,
arguing that these are oncogenic mutations. Active
mutated receptors are present at the cell surface and are
fully competent to bind EGF. Recent structural studies
show that the inactive EGFR tyrosine kinase domain is
autoinhibited by intramolecular interactions between its
activation loop and aC helix. We find that mutations
predicted to disrupt this autoinhibitory interaction (in-
cluding several that have not been described in NSCLC)
elevate EGF-independent tyrosine kinase activity, thus
providing new insight into how somatic mutations activate
EGFR and other ErbB family members.
Oncogene (2007) 26, 1567–1576. doi:10.1038/sj.onc.1209957;
published online 4 September 2006
Keywords: EGFR; ErbB; somatic mutations; constitu-
tive activation; internalization; autoinhibition
Introduction
Aberrant signaling by receptors from the ErbB or
epidermal growth factor receptor (EGFR) family is
implicated in many human cancers (de Larco and
Todaro, 1978; Holbro et al., 2003). ErbB2 overexpres-
sion clearly correlates with poor prognosis in breast
cancer (Slamon et al., 1987; Slamon et al., 1989), and
with response to ErbB2-targeted therapeutics (Slamon
et al., 2001; Dent and Clemons, 2006). By contrast, the
prognostic significance of EGFR overexpression varies
across studies, and responses to EGFR-targeted ther-
apeutic agents do not correlate with levels of EGFR
(Arteaga, 2002). Structural studies of ErbB receptors
(Burgess et al., 2003) provide one explanation for this
difference. Whereas EGFR dimerization is ‘autoinhib-
ited’ in the absence of EGF, ErbB2 adopts an active-like
conformation even without bound ligand (Burgess et al.,
2003). Accordingly, simply overexpressing ErbB2 trans-
forms NIH3T3 cells (Hudziak et al., 1987; Di Fiore
et al., 1987b), whereas EGFR can only transform
NIH3T3 cells in the presence of EGF (Di Fiore et al.,
1987a).
This distinction between EGFR and ErbB2 was
further underscored in clinical studies of the anilinoqui-
nazoline EGFR tyrosine kinase inhibitor, gefitinib
(Iressa). In phase II clinical trials (Fukuoka et al.,
2003; Kris et al., 2003), responses to gefitinib were seen
in just 10–28% of non-small-cell lung carcinoma
(NSCLC) patients (depending on trial location). When
seen, responses were dramatic. Gefitinib response did
not correlate with EGFR levels, but was instead
associated with somatic point mutations and deletions
in the EGFR tyrosine kinase domain (Lynch et al., 2004;
Paez et al., 2004; Pao et al., 2004). Initial studies of how
these gefitinib-sensitizing mutations affect the biochem-
ical properties of EGFR yielded conflicting results.
Lynch et al. (2004) reported enhanced EGF-induced
receptor activation, whereas Pao et al. (2004) suggested
autophosphorylation levels that were unchanged or
reduced compared with wild-type (WT) EGFR. Another
study suggested that autophosphorylation site utiliza-
tion is changed in EGFR mutants – altering signaling
specificity and enhancing activation of antiapoptotic
pathways (Sordella et al., 2004).
We hypothesized that gefitinib sensitivity of lung
tumors might not reflect enhanced gefitinib sensitivity of
the mutated EGFR per se, but might instead arise
because these tumors are uniquely dependent on
constitutive signaling by the EGFR mutants. To test
this hypothesis, we compared the autophosphorylation,
mitogenic, EGF-binding and localization properties of
mutated receptors expressed at matched levels in a null
Ba/F3 cell background. All mutations associated with
gefitinib sensitivity in NSCLC caused elevated EGF-
independent signaling by EGFR, suggesting that these
are indeed oncogenic mutations. We show that the
active mutated receptors are present (and active) at the
Received 17 March 2006; revised 19 July 2006; accepted 19 July 2006;
published online 4 September 2006
Correspondence: Professor MA Lemmon, Department of Biochem-
istry and Biophysics, University of Pennsylvania School of Medicine,
809C Stellar-Chance Laboratories, 422 Curie Boulevard, Philadelphia,
PA 19104-6059, USA.
E-mail: mlemmon@mail.med.upenn.edu
1
These authors contributed equally to this work.
Oncogene (2007) 26, 1567–1576
&
2007 Nature Publishing Group
All rights reserved 0950-9232/07 $30.00
www.nature.com/onc

cell surface. In investigations of the mechanistic basis
for EGFR activation by NSCLC mutations, we also
found that simply disrupting autoinhibitory interactions
observed crystallographically in the tyrosine kinase
domain (Wood et al., 2004; Zhang et al., 2006) is suffi-
cient to generate similar constitutively active variants
of EGFR.
Results and discussion
We focused on five of the reported gefitinib-sensitizing
NSCLC mutations: three missense mutations, and two
deletion mutations. We refer in this paper to residue
numbers in mature EGFR (where numbering reflects
loss of the 24aa signal sequence). The G695S mutation
(G719S in pro-EGFR) is found at the first glycine in the
GXGXXG motif of the kinase domain P-loop (Fig-
ure 1); the L834R and L837Q mutations (L858R and
L861Q in pro-EGFR) are both in the activation loop
(A-loop) of the kinase domain (Figure 1). Deletion
mutations associated with gefitinib-sensitivity cluster in
the N-lobe of the kinase domain, around the loop that
precedes the aC helix (Figure 1). We analysed two
deletion mutants. In one, residues L723 to P729 of
mature EGFR (L747-P753 in pro-EGFR) are replaced
with a single serine to give the DL723-P729insS mutant
(Lynch et al., 2004; Paez et al., 2004). In the second,
residues 728–735 (mature numbering) were deleted (Paez
et al., 2004), yielding DS728-I735 EGFR. The positions
of the two deletions in the active EGFR kinase domain
(Stamos et al., 2002) are colored red and cyan
respectively in Figure 1.
We expressed full-length mutated (or WT) EGFR in
the murine hematopoietic Ba/F3 cell line. Ba/F3 cells
lack endogenous EGFR, ErbB2 and ErbB4 [although
ErbB3 is detectable by reverse transcription–polymerase
chain reaction (RT–PCR) (Riese et al., 1995)], allowing
us to focus on the signaling properties of EGFR itself,
without complications from heterodimerization with
other ErbB family members. Selected Ba/F3 cell
transfectants were sorted using fluorescence-activated
cell sorting (FACS) into pools that express matched
(within B2-fold) levels of the EGFR variants at the cell-
surface (Figure 2).
Gefitinib-sensitizing mutations enhance basal EGFR
autophosphorylation
We first assessed tyrosine phosphorylation of each
mutated receptor by Western blotting both with
antiphosphotyrosine antibodies (to detect total pEGFR)
and with commercial antibodies against four of the
Figure 1 Location of gefitinib-sensitizing mutations in the EGFR
tyrosine kinase domain. Ribbon representation of the active EGFR
tyrosine kinase domain (PDB ID 1M17) bound to erlotinib
(Stamos et al., 2002). Erlotinib (in the ATP-binding site) is
represented with black sticks. The P-loop and catalytic loop are
colored blue, the activation loop (A-loop) is colored green and
helix aC is cyan. Side chains of G695, L834 and L837 (mutated in
NSCLC) are shown as yellow spheres. Regions deleted in the
DL723-P729insS and DS728-I735 mutations are colored red
(residues 723–728) and cyan (residues 729–735), respectively.
Figure 2 Comparable cell-surface expression levels of WT and
mutated EGFR. Parental Ba/F3 cells or sorted pools expressing
WT or mutated EGFR were incubated with R-phycoerythrin
conjugated anti-EGFR antibody for 30 min on ice, washed,
resuspended in 2% FBS/PBS at 10
6
cells/ml, and analysed by flow
cytometry. 10 000 cells were analysed for each FACS analysis.
Constitutively active EGFR mutants
SH Choi et al
1568
Oncogene

major EGFR tyrosine phosphorylation sites (Figure 3a).
All EGFR mutants showed significantly higher basal
(EGF-independent) tyrosine phosphorylation levels
than the WT receptor, with evidence for constitutive
autophosphorylation at Y992, Y1045 and Y1068, and
(presumably Src-mediated (Biscardi et al., 1999)) Y845
phosphorylation. The G695S and L834R mutants
appeared least heavily phosphorylated (but nonetheless
elevated over WT), while basal activity appeared similar
for the L837Q, DL723-P729insS and DS728-I735 mu-
tants. These findings concur with other studies published
while this paper was in preparation (Greulich et al.,
2005; Jiang et al., 2005; Chen et al., 2006), and contrast
with initial reports that these mutations do not increase
ligand-independent receptor activity (Lynch et al., 2004;
Pao et al., 2004).
Although elevated basal EGFR phosphorylation
was seen in all cases, the mutants seem to fall into two
classes with respect to their EGF response. Single
amino-acid substitutions in the P-loop (G695S) or
A-loop (L834R and L837Q) do not appear to impair
maximal EGF response, whereas the two deletion
mutants (DL723-P729insS and DS728-I735) showed
substantially reduced maximal levels of EGF-induced
autophosphorylation. Reduced maximal activity has also
been observed in other recent studies of deletion mutants
(Pao et al., 2004; Amann et al., 2005; Chen et al., 2006).
Autophosphorylation specificity is not altered in
gefitinib-sensitizing mutants
By contrast with two other recent reports (Sordella
et al., 2004; Chen et al., 2006), we could not detect
differences in phosphorylation site utilization in mu-
tated forms of EGFR. As shown in Figure 3a, the
relative degree of phosphorylation at tyrosines 845, 992,
1045 and 1068 was indistinguishable in WT and mutated
EGFR, with- and without EGF. We saw no evidence for
the selective phosphorylation of tyrosine 845 (primarily
a Src phosphorylation site (Biscardi et al., 1999; Tice
et al., 1999)) reported for the L834R EGFR in mouse
mammary epithelial cells (Sordella et al., 2004), or for
enhanced autophosphorylation of Y992 or Y1068
reported for L834R and DL723-P729insS in the same
study (Sordella et al., 2004). Similarly, we could not
reproduce the increased Y1045 phosphorylation re-
ported by Chen et al. (2006) for EGFR mutants
expressed in the H1299 NSCLC cell line. We note that
these previous reports differ starkly from one another in
which phosphorylation sites appeared to be used in
gefitinib-sensitizing EGFR mutants, and our studies
disagree with both – showing no altered selectivity at all.
The three studies were performed in quite different cell
lines that differ in their complements of ErbB receptors
(and other kinases), suggesting that this observation
could be cell-type dependent. Our findings do argue,
however, that there is no clear case to be made for an
alteration in the autophosphorylation substrate specifi-
city of the mutated EGFR itself.
The two previous studies of mutated EGFR phos-
phorylation used the same set of phosphospecific anti-
EGFR antibodies that we employed here (see Materials
and methods). As a control, we assessed apparent EGF-
induced phosphorylation at tyrosine 845 of a Y845F
EGFR mutant expressed in CHO-K1 cells (which lack
endogenous EGFR). As shown in Figure 3b, EGF
treatment promotes recognition of Y845F EGFR by the
phospho-Y845-specific antibody even though it lacks a
tyrosine at position 845. Thus, even when alterations in
recognition by phosphospecific antibodies are observed
in particular cell lines for EGFR mutants, these changes
should only be ascribed to increased phosphorylation
at specific sites when appropriate controls with phenyl-
alanine substitution mutants have been performed.
EGFR mutants promote Ba/F3 cell proliferation
independent of EGF or Interleukin-3
As gefitinib-sensitizing mutations enhance constitutive
EGFR autophosphorylation, we reasoned that they
might also promote ligand-independent cell prolifera-
tion. Ba/F3 cells normally require interleukin-3
(IL-3) for survival and proliferation. However, in
Ba/F3 cells expressing WT EGFR, EGF can promote
α-pY845
α-pY992
α-pY1045
α-pY1068
total pEGFR
total EGFR
actin
WT G695S L834R L837Q
∆L723
P729iS
∆S728
I735
- +
- +
- +
- + - +
- +
EGF
total EGFR
pY845
- +
- +
- +
EGF
WT Y845FMock
a
b
Figure 3 Constitutive tyrosine phosphorylation of EGFR mu-
tants associated with gefitinib sensitivity. (a) Cells expressing WT
or mutated EGFR were serum-starved overnight, and treated on
ice with ( þ ) or without () 100 ng/ml EGF. Cells were lysed, and
samples of whole-cell lysate matched for protein concentration
were subjected to SDS–PAGE and Western blotting with the noted
antibodies (see Materials and methods). The bottom panels show
blots with anti-EGFR antibody Ab15 (total EGFR) for receptor-
level normalization and anti-actin antibodies as a loading control.
This experiment is representative of more than three repeats. (b)
WT or Y845F mutated EGFR was transiently expressed in CHO-
K1 cells. Serum-starved cells were stimulated on ice with- or
without 100 ng/ml EGF, and lysed. EGFR was immunoprecipi-
tated with anti-EGFR antibody 528, and subjected to immuno-
blotting with anti-pY845 (Cell Signaling Technology) and Ab15.
The anti-pY845 antibody from Biosource gave identical results.
Constitutively active EGFR mutants
SH Choi et al
1569
Oncogene

IL-3-independent Ba/F3 cell survival (Riese et al., 1996)
or mitogenesis and proliferation (Collins et al., 1988;
Walker et al., 2004). As shown in Figure 4a, expression
of WT EGFR in Ba/F3 cells does not increase DNA
synthesis above levels seen in parental Ba/F3 cells (in the
absence of IL-3 or EGF). However, expression of any of
the five gefitinib-sensitizing EGFR mutants in Ba/F3
cells led to significant growth factor-independent
[
3
H]thymidine incorporation. EGF addition increased
DNA synthesis further (by >15-fold in the WT case,
2.5- to 5-fold for the mutants). Thus, each EGFR
mutant studied here (but not WT EGFR) can promote
constitutive growth signaling in Ba/F3 cells. Jiang et al.
(2005) recently reported a similar observation for just
the G695S and L834R EGFR mutants in Ba/F3 cells.
Moreover, Greulich et al. (2005) found that several of
the EGFR mutants studied here can transform NIH 3T3
cells. Thus, from our own studies and those published
recently by other groups (Greulich et al., 2005; Jiang
et al., 2005; Chen et al., 2006; Sakai et al., 2006), it
appears that the EGFR mutations identified in gefitinib-
responsive lung cancer patients are oncogenic. The
response to gefitinib and other EGFR inhibitors of
patients expressing these mutants is therefore likely to
reflect the dependence of their tumors on constitutive
EGFR signaling per se, rather than any alteration in the
response of mutated EGFR to the inhibitors. Consistent
with this, recent studies of inhibitor binding (Fabian
et al., 2005) have shown that none of the mutations
analysed in our study alter gefitinib or erlotinib binding
by more than threefold.
EGFR mutants are active at the cell surface
The molecular basis for EGFR activation by gefitinib-
sensitizing mutations in NSCLC is currently not clear. It
is possible that the mutations elevate kinase activity by
promoting conformational changes that mimic the
normal ligand-induced activation mechanism. However,
it is also known that many ErbB receptor mutants
misfold and are retained in the endoplasmic reticulum,
where they can be potently transforming (Hudziak and
Ullrich, 1991; Ekstrand et al., 1995). Although cell-
permeable tyrosine kinase inhibitors potently inhibit all
the EGFR mutants studied here, the cetuximab anti-
body fails to inhibit their EGF-independent EGFR
signaling in non-small-cell lung cancer cell lines (Amann
et al., 2005; Mukohara et al., 2005) or Ba/F3 cells (data
not shown). This raises the critical question as to
whether the active, mutated, receptors are accessible to
inhibitory antibodies at the cell surface, or are instead
retained (and active) in the endoplasmic reticulum.
To assess subcellular location of the constitutively
activated EGFR mutants, we used immunofluorescence
microscopy. Staining of nonpermeabilized cells with an
antibody against the EGFR extracellular region (Ab5,
see Materials and methods) suggested that active
mutated receptors are present at the cell surface. Ab5-
stained WT EGFR is distributed evenly around the
periphery of nonpermeabilized Ba/F3 cells without EGF
(Figure 5). When EGF is added, the receptor forms
patches or clusters at the cell surface, a well-described
phenomenon that is always associated with EGFR
activation (Schechter et al., 1979; Schreiber et al., 1983).
The constitutively active EGFR mutants found in
NSCLC formed similar cell-surface patches even in the
absence of EGF (Figure 5), the most dramatic examples
being the L837Q and DL723-P729insS mutants. Only
the G695S mutant required EGF addition to form such
patches. We also monitored the ability of mutated
EGFR to internalize bound Ab5 in the absence of EGF
stimulation (Figure 6). WT EGFR showed little
evidence of antibody patching or internalization at 0
or 371C, consistent with a low level of activation in the
absence of EGF. Ab5 that was bound to the mutated
receptors remained primarily at the cell surface when
cells were kept on ice (although did show clustering), but
showed a punctate, largely intracellular, localization in
cells that had been placed at 371C (Figure 6, right-hand
panels). These data suggest that each EGFR mutant is
sufficiently active in the absence of EGF to promote
their constitutive internalization from the cell surface.
Cell-surface patching and receptor internalization was
seen for all DL723-P729insS-expressing cells examined,
for 91% of DS728-I735-expressing cells, 94% of L837Q-
expressing cells, 87% of L834R-expressing cells and
65% of G695S-expressing cells.
0
5
10
15
20
25
30
35
40
45
Control WT G695S L834R L837Q
∆L723
P729iS
∆S728
I735
-EGF
+EGF
EGFR
Actin
Thymidine Incorporated ( x 10
3
cpm)
a
b
Figure 4 EGF-independent stimulation of DNA synthesis by
gefitinib-sensitizing EGFR mutants. (a)[
3
H]thymidine incorpora-
tion by Ba/F3 cells expressing WT or mutated EGFR was analysed
with (black bars) and without (gray bars) 100 ng/ml EGF, as
described in Materials and methods. Data represent the average of
at least three independent assays, and error bars represent the
standard deviation. (b) Ba/F3 cells expressing WT or mutated
EGFR were harvested at the same time as taken for the
[
3
H]thymidine incorporation assay. Cell extracts were prepared
and subjected to Western blot analysis using anti-EGFR (Ab15)
and anti-actin antibodies, for receptor-level and loading controls.
Constitutively active EGFR mutants
SH Choi et al
1570
Oncogene

The fact that EGFR variants harboring gefitinib-
sensitizing mutations are present at the cell surface
argues strongly that the constitutive activity evident in
Figures 3 and 4 is not simply a result of EGFR
aggregation in the endoplasmic reticulum. Our qualita-
tive experiments also suggest that the majority of the
cell-surface mutated EGFR is being internalized without
EGF, yet EGF addition leads to a strong increase in
autophosphorylation of the G695S, L834R and L837Q
mutants. One hypothesis suggested by these observa-
tions is that the gefitinib-sensitizing mutations cause a
low level of EGF receptor activation (and autophos-
phorylation) that is sufficient to promote receptor
internalization and Ba/F3 cell mitogenesis, but does
not entirely recapitulate the activated state induced by
EGF (and so is insensitive to cetuximab inhibition).
Influence of gefitinib-sensitizing mutations on
ligand-binding affinity of EGFR
We hypothesized that mutations causing elevated basal
EGFR autophosphorylation might also promote ligand-
independent dimerization of the receptor, which might
in turn be reflected in the affinity of the cell-surface
receptor for EGF. We therefore analysed binding of
125
I-
labeled EGF to Ba/F3 cells expressing WT or mutated
EGFR (Figure 7). EGF binding saturated at a level
equivalent to between 5000 and 15 000 receptors per cell
for each mutant. WT EGFR gave typical Scatchard
plots indicating the expected two classes of binding site
(Ullrich and Schlessinger, 1990), with B3% of sites (the
high-affinity sites) described by a K
D
value of 25 pM, and
the remaining 97% (low-affinity sites) by an apparent
K
D
of 0.45 nM (Figure 7a). The G695S mutant appeared
almost identical to WT EGFR, with an apparent K
D
value of B20 pM for high-affinity sites (2–3%) and
0.4 n
M for low-affinity sites (Figure 7b). The L834R
mutant was again very similar, but suggested an
apparent K
D
for the low-affinity sites of approximately
0.25 n
M, representing an affinity increase of B2-fold
over WT (Figure 7c). Finally, the DL723-P729insS
mutant gave more linear Scatchard plots (Figure 7d),
which were fit best by assuming a single class of (low-
affinity) sites with an apparent K
D
for EGF binding of
0.8 n
M (B2-fold lower affinity than WT).
These findings confirm that the G695S, L834R and
DL723-P729insS mutants are all present at the cell
surface at normal levels, and that they are competent to
bind EGF. It is difficult to interpret the small (B2-fold)
differences in apparent K
D
for EGF-binding seen in
L834R and DL723-P729insS EGFR. However, it is
interesting to note that abolishing the autoinhibitory
tether in the extracellular region of EGFR leads to a
mere 2- to 3-fold increase in EGF-binding affinity
(Ferguson et al., 2003; Ozcan et al., 2006). Thus,
although the effect is small, the influence of the L834R
mutation on cell-surface EGF binding is consistent with
the loss of an autoinhibitory interaction in the receptor
(or with a slightly increased tendency to dimerize).
Activation of EGFR by mutations that destabilize
autoinhibitory interactions in the inactive state
Recent structural studies provide potential mechanistic
explanations for how gefitinib-sensitizing mutations
elevate EGF-independent EGFR kinase activity (Wood
et al., 2004; Zhang et al., 2006). In the inactive
conformation of the EGFR tyrosine kinase domain
with bound lapatinib (Wood et al., 2004) or AMP-PNP
(Zhang et al., 2006), the aC helix and the activation loop
interact directly with one another. Both of these
catalytically essential elements are prevented by this
autoinhibitory interaction from adopting the active
configuration. As depicted in Figure 8a, a short a-helix
in the A-loop (green) that includes residues L834 and
L837 (colored yellow) packs against the displaced aC
helix (cyan). The L834 and L837 side chains contribute
to a hydrophobic core formed by residues from the
activation loop, helix aC, the P-loop, and elsewhere
(Wood et al., 2004; Zhang et al., 2006). An L834R or
L837Q mutation (as seen in gefitinib-sensitive NSCLC)
could disrupt the autoinhibitory interaction between the
Figure 5 Cell-surface clustering of mutated EGFR suggests constitutive activity. Ba/F3 cells expressing WT or mutated EGFR were
treated with- or without EGF as noted, fixed in formaldehyde, and incubated with an antibody against the EGFR extracellular region
(Ab5) for one hour. Cells were then labeled with FITC-conjugated anti-mouse antibody (green) for 45 min, and nuclei were visualized
with Hoecht 33342 dye (blue).
Constitutively active EGFR mutants
SH Choi et al
1571
Oncogene

A-loop and aC helix of EGFR simply by destabilizing
this set of hydrophobic interactions. This would free the
regulatory elements of the kinase domain to adopt the
active configuration shown in Figure 1.
The side chains of I735 and M742 from helix aC (red
in Figure 8a) also appear to make important contribu-
tions to the hydrophobic core between the A-loop and
aC helix, and we hypothesized that their mutation to
0ⴗC37ⴗC
G695S
wild-type
L834R
L837Q
∆L723-P729
insS
∆S728-I735
Figure 6 EGF-independent internalization of mutated, but not
WT, EGFR. Ba/F3 cells expressing WT or mutated receptors were
serum starved overnight, prechilled on ice, and incubated for 1 h on
ice with a monoclonal antibody (Ab5) against the EGFR
extracellular region. All manipulations were performed in the
absence of EGF. Half of the cells were then incubated at 371C for
20 min to permit internalization (right panels), and half remained
on ice (left panels). Cells were then fixed, permeabilized and
internalized (plus cell-surface-bound) anti-EGFR antibody was
detected with an anti-mouse IgG Texas red conjugate (red).
a
WT
b
G695S
c
L834R
d
∆L723-P729insS
0.02 0.03 0.040.00 0.01
0.00
0.05
0.10
0.15
0.20
Specific Binding (nM)
Bound/Free
0 1 2 3 4
0.00
0.01
0.02
0.03
0.04
[EGF] (nM)
Specific Binding (nM)
0.01 0.02 0.030.00
0.00
0.05
0.10
0.15
0.20
Specific Binding (nM)
Bound/Free
0 1 2 3 4
0.00
0.01
0.02
0.03
0.03 0.04 0.05 0.06 0.070.00 0.01 0.02
0.00
0.05
0.10
0.15
0.20
0.25
Specific Binding (nM)
Bound/Free
0 1 2 3 4
0.00
0.01
0.02
0.03
0.04
0.05
0.06
[EGF] (nM)
[EGF] (nM)
Specific Binding (nM)Specific Binding (nM)
0.00 0.05 0.10 0.15 0.20
0.00
0.05
0.10
0.15
0.20
0.25
0.30
Specific Binding (nM)
Bound/Free
0.00
0 1 2 3 4
0.05
0.10
0.15
[EGF] (nM)
Specific Binding (nM)
Figure 7 Effects of EGFR tyrosine kinase domain mutations on
cell-surface EGF binding affinity. Binding of
125
I-labeled EGF to
Ba/F3 cells expressing WT or mutated EGFR was analysed as
described in Materials and methods. Results averaged for at least
four independent experiments are plotted both as Scatchard plots
and as saturation binding curves (insets), with best fits shown as
solid lines. Data for WT (a), G695S (b) and L834R (c) EGFR were
all best fit to a two binding site model in which the high-affinity
EGF-binding site (K
D
B25 pM) account for approximately 3% of
the sites. The DL723-P729insS data (d) could not be fit
satisfactorily to a two-site model.
Constitutively active EGFR mutants
SH Choi et al
1572
Oncogene

alanine should also disrupt autoinhibitory interactions
in the kinase domain. As shown in Figure 8b, both
I735A and M742A EGFR mutants show robust EGF-
independent basal autophosphorylation when expressed
in CHO-K1 cells. Mutation of the more peripherally
associated L764 side-chain had little effect. These results
suggest that many EGFR alterations that destabilize
kinase domain autoinhibition are likely to be oncogenic.
The regions deleted from the N-terminus of helix aCin
DL723-P729insS and DS728-I735 EGFR were seen to
interact with the A-loop in inactive EGFR (Zhang et al.,
2006), arguing that their removal may activate EGFR
by disrupting the autoinhibited kinase domain config-
uration. However, these deletion mutants might
also destabilize the receptor sufficiently to promote
aggregation, which could contribute to constitutive
activity. Indeed, DL723-P729insS EGFR is degraded
very rapidly following treatment of cells with inhibitors
of Hsp90, suggesting that it is particularly reliant
on chaperone proteins for its stability (Shimamura
et al., 2005).
Conclusion
It now seems clear that EGFR mutations found in
gefitinib-responsive NSCLC patients cause EGF-inde-
pendent activation of the receptor, which is important
for tumor initiation and/or progression. We show that
these mutations do not significantly affect the cell-
surface localization or EGF binding properties of the
EGF receptor. Rather, in the light of recently published
structural studies (Wood et al., 2004; Zhang et al., 2006),
it appears that NSCLC mutations disrupt autoinhibi-
tory interactions that ordinarily suppress the basal
kinase activity of EGFR. These mutations are likely to
mimic in part the activating conformational changes
induced in the EGFR intracellular domain upon
receptor dimerization (Zhang et al., 2006). In testing
this hypothesis, we identified two EGFR mutations
(I735A and M742A), predicted to mimic the conforma-
tional effects of NSCLC mutations, that activate EGFR
but have not been described in cancer patients. There
are likely to be many similar activating mutations, and
the increasingly sophisticated picture of EGF receptor
activation and autoinhibition should allow a rapid
assessment of their consequences when they are ob-
served clinically.
Materials and methods
Reagents
Anti-EGFR antibodies, Ab15 and Ab5 were purchased from
LabVision (Fremont, CA, USA), and Ab528 was from Santa
Cruz Biotechnology (Santa Cruz, CA, USA). Phosphotyrosine
antibody (PY20) was from Zymed Laboratories Inc. (San
Francisco, CA, USA). Anti-actin antibody was from Sigma (St
Louis, MO, USA). R-phycoerythrin conjugated anti-EGFR
antibody was purchased from BD Biosciences (San Diego, CA,
USA). [
3
H]thymidine and
125
I-labeled recombinant human
EGF were purchased from Amersham Biosciences (Piscat-
away, NJ, USA). ITS was purchased from Collaborative
Biomedical Products (Bedford, MA, USA). IL-3 was pur-
chased from was purchased from Pepro Tech. (Rocky Hill, NJ,
USA), and EGF from Chemicon (Temecula, CA, USA).
Phospho-specific antibodies against EGFR phosphorylation
sites (anti-pY845, anti-pY992, anti-pY1045 and anti-pY1068)
were purchased from Cell Signaling Technologies (Beverly,
MA, USA). An additional anti-pY845 antibody was pur-
chased from Biosource (Camarillo, CA, USA).
Generation of mutated EGFR constructs
WT EGFR was cloned into pcDNA3.1( þ ) (Invitrogen,
Carlsbad, CA, USA) using KpnI and XhoI restriction sites.
Missense mutations (G695S, L834R and L837Q) were
generated by site-directed mutagenesis using the Quikchange
method (Stratagene, La Jolla, CA, USA), and deletion
mutations (DL723-P729insS and DS728-I735) were generated
by 4-primer PCR, with ligation of the resulting products
between internal BspEI (introduced) and ClaI (natural) sites in
Figure 8 EGF-independent EGFR activation by mutations that
disturb interactions between helix aC and the activation loop in the
inactive kinase domain. (a) Close-up view of the packing between
the displaced aC helix (cyan) and a helical segment of the activation
loop observed by Wood et al. (2004) in a crystal structure of the
complex between lapatinib and inactive EGFR tyrosine kinase
domain (PDB code 1XKK). The P-loop, catalytic loop and
activation loop (A-loop) are colored and labeled as in Figure 1.
G695, L834 and L837 are colored yellow. The residues colored red
(I735, M742 and L764) all have aliphatic side chains that appear to
contribute to van der Waals interactions between helix aC and the
activation loop in this inactive conformation, and could stabilize it.
Each was individually mutated to alanine for the experiment shown
in (b). (b) Full-length EGFR bearing the I735A, M742A or L764A
mutation (or WT) was transiently expressed in CHO-K1 cells.
Serum-starved cells were stimulated on ice with- or without EGF as
marked, and lysed. EGFR was immunoprecipitated with antibody
528, and subjected to immunoblotting with PY20 and Ab15.
Constitutively active EGFR mutants
SH Choi et al
1573
Oncogene

the EGFR sequence. All constructs were fully sequenced to
confirm introduction of only the desired mutations.
Generation of stable cell lines
Ba/F3 cells were maintained in Roswell Park Memorial
Institute media (RPMI)-1640 media supplemented with 10%
fetal bovine serum (FBS), 10 m
M 2-[4-(2-hydroxyethyl)-1-
piperazinyl]ethanesulfonic acid (HEPES), 2 m
M sodium pyr-
uvate and IL-3 (1 ng/ml). Plasmids directing expression of WT
or mutated EGFR were electroporated into Ba/F3 cells using
an Electrocell manipulator (ECM600, BTX Genetronics, San
Diego, CA, USA). Transfected cells were selected in medium
containing 1 mg/ml G418 for 2 weeks, and viable cells were
sorted by fluorescence-activated cell sorting on a FACSCalibur
machine (BD Biosciences, San Jose, CA, USA) for EGFR
expression level using R-phycoerythrin conjugated anti-EGFR
antibody. EGFR expression levels were also confirmed by
Western blot analysis using anti-EGFR (Ab15) antibody,
normalizing with anti-actin antibodies. Cell pools with similar
expression levels (within twofold of the WT stable line) were
used for all subsequent experiments.
Receptor activation assays
Cells were serum starved for 24 h in supplemented RPMI-1640
media (described above) lacking FBS. Starved cells were
prechilled on ice, and incubated for 10 min in ice-cold
starvation medium containing 100 ng/ml EGF. Cells were
pelleted, washed extensively in phosphate-buffered saline
(PBS), lysed in radioimmunoprecipitation assay (RIPA) buffer
(1% Na deoxycholate, 1% NP-40, 0.1% sodium dodecyl
sulfate (SDS), 150 m
M NaCl, 10 mM Na phosphate buffer pH
7.2, 1 m
M phenylmethylsulphonyl fluoride (PMSF), 1 mg/ml
leupeptin, 1 mg/ml aprotinin, 25 m
M NaF, 5 mM Na
2
MoO
4
,
0.2 m
M Na
3
VO
4
), and subjected to Western blot analysis using
the indicated antibodies.
[
3
H]thymidine incorporation assay
Cells growing in log phase were harvested, washed twice with
RPMI-1640 to remove residual IL-3, resuspended in RPMI-
1640 supplemented with 0.5% ITS (ITS plus BSA and linoleic
acid) and seeded into 96-well plates at 1  10
5
cells/150 ml. EGF
(100 ng/ml), IL-3 (10 ng/ml), or no growth factor/cytokine was
added to each well as indicated, and cells were incubated for
15 h at 371Cin5%CO
2
. Tritiated thymidine (1 mCi) was added
to each well, and cells were incubated for another 7 h at 371C.
Cells were then harvested onto a glass fiber filter using a
Harvester 96 MACH III M (TomTec, Hamden, CT, USA),
washed to remove unincorporated [
3
H]thymidine, and
3
H-
incorporation was measured by liquid scintillation counting of
cells harvested from the plate using a Wallac 1450 Microbeta
counter (Perkin-Elmer, Boston, MA, USA).
125
I-EGF binding assay
Cells (10
6
) were washed with binding buffer (cold RPMI-1640
supplemented with 0.1% bovine serum assay (BSA), incubated
on ice for 2 h with 50 p
M
125
I-EGF (specific activity B8–
12  10
5
cpm/pmol) supplemented with various concentrations
of cold EGF (for total (EGF) of 0.003 n
M–6 nM). Cells were
then washed once with 0.5 ml of binding buffer, and were
centrifuged through a cushion of 0.7 ml bovine serum to
remove unbound ligand. Cells were collected, and bound
radioactivity was measured by scintillation counting. Each
experiment was performed in triplicate, with a 100-fold excess
of unlabeled ligand added to the third tube at each
concentration to measure nonspecific binding. After correcting
for nonspecific binding, results were analysed using GraphPad
Prism 4.0.
Immunofluorescence
Ba/F3 cells were fixed in 3.7% formaldehyde for 20 min and
then washed thoroughly with PBS. Cells for surface staining
were blocked in 1% BSA diluted in PBS (BSA/PBS) for
20 min. Cells were incubated with anti-EGFR mAb5 for 1 h
and then with fluorescein isothiocyanate (FITC)-conjugated
anti-mouse for 45 min at room temperature. To visualize
nuclei, 1.5 mg/ml Hoecht 33342 dye was included with
secondary antibodies. Cells were washed thoroughly in PBS,
mounted with Fluoromount G, and images were collected with
a Hamamatsu Orca CCD camera using a  100 objective on a
Leica DM IRBE microscope.
For analysis of EGFR internalization, cells were serum
starved overnight, prechilled on ice and incubated with the
anti-EGFR mAb5 (diluted 1/50 in serum starvation media) for
1 h on ice to allow antibody binding to cell-surface receptors.
Cells were then incubated for 20 min either on ice, or at 371C
to allow internalization of active receptors. Internalization was
stopped by incubation on ice followed by extensive washing to
remove unbound antibody. Cells were fixed in 4% parafor-
maldehyde (30 min), washed and blocked in 1% BSA/PBS
supplemented with 0.2% saponin. Anti-EGFR antibody was
then visualized with an anti-mouse IgG-Texas red conjugate
(Santa Cruz Biotechnology, Santa Cruz, CA, USA). Cells were
washed thoroughly with PBS, mounted with Fluoromount G,
and serial Z section images were taken and deconvoluted using
the Volocity software package (Improvision, Lexington, MA,
USA).
Acknowledgements
We thank Kate Ferguson and members of the Lemmon
laboratory for valuable discussions and comments on the
paper. This work was supported by NIH grant R01-CA096768
to MAL.
References
Amann J, Kalyankrishna S, Massion PP, Ohm JE, Girard L,
Shigematsu H et al. (2005). Aberrant epidermal growth
factor receptor signaling and enhanced sensitivity to EGFR
inhibitors in lung cancer. Cancer Res 65: 226–235.
Arteaga CL. (2002). Epidermal growth factor receptor
dependence in human tumors: more than just expression?
Oncologist 7(Suppl 4): 31–39.
Biscardi JS, Maa MC, Tice DA, Cox ME, Leu TH, Parsons
SJ. (1999). c-Src-mediated phosphorylation of the epidermal
growth factor receptor on Tyr845 and Tyr1101 is associated
with modulation of receptor function. J Biol Chem 274:
8335–8343.
Burgess AW, Cho HS, Eigenbrot C, Ferguson KM, Garrett
TP, Leahy DJ et al. (2003). An open-and-shut case? Recent
insights into the activation of EGF/ErbB receptors. Mol Cell
12: 541–552.
Chen YR, Fu YN, Lin CH, Yang ST, Hu SF, Chen YT et al.
(2006). Distinctive activation patterns in constitutively active
and gefitinib-sensitive EGFR mutants. Oncogene 25:
1205–1215.
Collins MK, Downward J, Miyajima A, Maruyama K, Arai
K, Mulligan RC. (1988). Transfer of functional EGF
receptors to an IL3-dependent cell line. J Cell Physiol 137:
293–298.
Constitutively active EGFR mutants
SH Choi et al
1574
Oncogene

de Larco JE, Todaro GJ. (1978). Epithelioid and fibroblastic
rat kidney cell clones: epidermal growth factor (EGF)
receptors and the effect of mouse sarcoma virus transforma-
tion. J Cell Physiol 94: 335–342.
Dent R, Clemons M. (2006). Trastuzumab after primary
treatment for early stage HER2-positive breast cancer
reduces recurrence. Cancer Treat Rev 32: 144–148.
Di Fiore PP, Pierce JH, Fleming TP, Hazan R, Ullrich A, King
CR et al. (1987a). Overexpression of the human EGF
receptor confers an EGF-dependent transformed phenotype
to NIH 3T3 cells. Cell 51: 1063–1070.
Di Fiore PP, Pierce JH, Kraus MH, Segatto O, King CR,
Aaronson SA. (1987b). ErbB-2 is a potent oncogene when
overexpressed in NIH/3T3 cells. Science 237: 178–182.
Ekstrand AJ, Liu L, He J, Hamid ML, Longo N, Collins VP
et al. (1995). Altered subcellular location of an activated
and tumour-associated epidermal growth factor receptor.
Oncogene 10: 1455–1460.
Fabian MA, Biggs 3rd WH, Treiber DK, Atteridge CE,
Azimioara MD, Benedetti MG et al. (2005). A small
molecule-kinase interaction map for clinical kinase inhibi-
tors. Nat Biotechnol 23: 329–336.
Ferguson KM, Berger MB, Mendrola JM, Cho HS, Leahy DJ,
Lemmon MA. (2003). EGF activates its receptor by
removing interactions that autoinhibit ectodomain dimer-
ization. Mol Cell 11: 507–517.
Fukuoka M, Yano S, Giaccone G, Tamura T, Nakagawa K,
Douillard JY et al. (2003). Multi-institutional randomized
phase II trial of gefitinib for previously treated patients with
advanced non-small-cell lung cancer (The IDEAL 1 Trial).
J Clin Oncol 21: 2237–2246.
Greulich H, Chen TH, Feng W, Janne PA, Alvarez JV,
Zappaterra M et al. (2005). Oncogenic transformation by
inhibitor-sensitive and -resistant EGFR mutants. PLoS Med
2: e313.
Holbro T, Civenni G, Hynes NE. (2003). The ErbB receptors
and their role in cancer progression. Exp Cell Res 284:
99–110.
Hudziak RM, Schlessinger J, Ullrich A. (1987). Increased
expression of the putative growth factor receptor p185HER2
causes transformation and tumorigenesis of NIH 3T3 cells.
Proc Natl Acad Sci USA 84: 7159–7163.
Hudziak RM, Ullrich A. (1991). Cell transformation potential
of a HER2 transmembrane domain deletion mutant retained
in the endoplasmic reticulum. J Biol Chem 266:
24109–24115.
Jiang J, Greulich H, Janne PA, Sellers WR, Meyerson M,
Griffin JD. (2005). Epidermal growth factor-independent
transformation of Ba/F3 cells with cancer-derived epidermal
growth factor receptor mutants induces gefitinib-sensitive
cell cycle progression. Cancer Res 65: 8968–8974.
Kris MG, Natale RB, Herbst RS, Lynch Jr TJ, Prager D,
Belani CP et al. (2003). Efficacy of gefitinib, an inhibitor of
the epidermal growth factor receptor tyrosine kinase, in
symptomatic patients with non-small cell lung cancer: a
randomized trial. JAMA 290: 2149–2158.
Lynch TJ, Bell DW, Sordella R, Gurubhagavatula S, Okimoto
RA, Brannigan BW et al. (2004). Activating mutations in the
epidermal growth factor receptor underlying responsiveness
of non-small-cell lung cancer to gefitinib. N Engl J Med 350:
2129–2139.
Mukohara T, Engelman JA, Hanna NH, Yeap BY, Kobayashi
S, Lindeman N et al. (2005). Differential effects of gefitinib
and cetuximab on non-small-cell lung cancers bearing
epidermal growth factor receptor mutations. J Natl Cancer
Inst 97: 1185–1194.
Ozcan F, Klein P, Lemmon MA, Lax I, Schlessinger J.
(2006). On the nature of low- and high-affinity EGF
receptors on living cells. Proc Natl Acad Sci USA 103:
5735–5740.
Paez JG, Janne PA, Lee JC, Tracy S, Greulich H, Gabriel S
et al. (2004). EGFR mutations in lung cancer: correlation
with clinical response to gefitinib therapy. Science 304:
1497–1500.
Pao W, Miller V, Zakowski M, Doherty J, Politi K, Sarkaria I
et al. (2004). EGF receptor gene mutations are common in
lung cancers from ‘never smokers’ and are associated with
sensitivity of tumors to gefitinib and erlotinib. Proc Natl
Acad Sci USA 101: 13306–13311.
Riese DJ, Kim ED, Elenius K, Buckley S, Klagsbrun M,
Plowman GD et al. (1996). The epidermal growth factor
receptor couples transforming growth factor-alpha, heparin-
binding epidermal growth factor-like factor, and amphi-
regulin to Neu, ErbB-3, and ErbB-4. J Biol Chem 271:
20047–20052.
Riese 2nd DJ, van Raaij TM, Plowman GD, Andrews GC,
Stern DF. (1995). The cellular response to neuregulins is
governed by complex interactions of the erbB receptor
family. Mol Cell Biol 15: 5770–5776.
Sakai K, Arao T, Shimoyama T, Murofushi K, Sekijima M,
Kaji N et al. (2006). Dimerization and the signal transduc-
tion pathway of a small in-frame deletion in the epidermal
growth factor receptor. FASEB J 20: 311–313.
Schechter Y, Hernaez L, Schlessinger J, Cuatrecasas P. (1979).
Local aggregation of hormone-receptor complexes is re-
quired for activation by epidermal growth factor. Nature
278: 835–838.
Schreiber AB, Libermann TA, Lax I, Yarden Y, Schlessinger
J. (1983). Biological role of epidermal growth factor-receptor
clustering. Investigation with monoclonal anti-receptor
antibodies. J Biol Chem 258: 846–853.
Shimamura T, Lowell AM, Engelman JA, Shapiro GI. (2005).
Epidermal growth factor receptors harboring kinase domain
mutations associate with the heat shock protein 90
chaperone and are destabilized following exposure to
geldanamycins. Cancer Res 65: 6401–6408.
Slamon DJ, Clark GM, Wong SG, Levin WJ, Ullrich A,
McGuire WL. (1987). Human breast cancer: correlation of
relapse and survival with amplification of the HER-2/neu
oncogene. Science 235: 177–182.
Slamon DJ, Godolphin W, Jones LA, Holt JA, Wong SG,
Keith DE et al. (1989). Studies of the HER-2/neu proto-
oncogene in human breast and ovarian cancer. Science 244:
707–712.
Slamon DJ, Leyland-Jones B, Shak S, Fuchs H, Paton V,
Bajamonde A et al. (2001). Use of chemotherapy plus a
monoclonal antibody against HER2 for metastatic breast
cancer that overexpresses HER2. N Engl J Med 344:
783–792.
Sordella R, Bell DW, Haber DA, Settleman J. (2004).
Gefitinib-sensitizing EGFR mutations in lung cancer acti-
vate anti-apoptotic pathways. Science 305: 1163–1167.
Stamos J, Sliwkowski MX, Eigenbrot C. (2002). Structure of
the epidermal growth factor receptor kinase domain alone
and in complex with a 4-anilinoquinazoline inhibitor. J Biol
Chem 277: 46265–46272.
Tice DA, Biscardi JS, Nickles AL, Parsons SJ. (1999).
Mechanism of biological synergy between cellular Src and
epidermal growth factor receptor. Proc Natl Acad Sci USA
96: 1415–1420.
Ullrich A, Schlessinger J. (1990). Signal transduction by
receptors with tyrosine kinase activity. Cell 61: 203–212.
Constitutively active EGFR mutants
SH Choi et al
1575
Oncogene

Walker F, Orchard SG, Jorissen RN, Hall NE, Zhang HH,
Hoyne PA et al. (2004). CR1/CR2 interactions modulate the
functions of the cell surface epidermal growth factor
receptor. J Biol Chem 279: 22387–22398.
Wood ER, Truesdale AT, McDonald OB, Yuan D, Hassell A,
Dickerson SH et al. (2004). A unique structure for epidermal
growth factor receptor bound to GW572016 (Lapatinib):
relationships among protein conformation, inhibitor off-
rate, and receptor activity in tumor cells. Cancer Res 64:
6652–6659.
Zhang X, Gureasko J, Shen K, Cole PA, Kuriyan J. (2006).
An allosteric mechanism for activation of the kinase
domain of epidermal growth factor receptor. Cell 125:
1137–1149.
Constitutively active EGFR mutants
SH Choi et al
1576
Oncogene