Oncogene (2003) 22, 1557–1567
& 2003 Nature Publishing Group All rights reserved 0950-9232/03 $25.00
www.nature.com/onc
Generation and functional characterization of intracellular antibodies
interacting with the kinase domain of human EGF receptor
Stephen Hyland1, Roger R Beerli2,3,4, Carlos F Barbas III2, Nancy E Hynes3 and Winfried Wels*,1
1
Chemotherapeutisches Forschungsinstitut Georg-Speyer-Haus, D-60596 Frankfurt am Main, Germany; 2The Skaggs Institute for
Chemical Biology and the Department of Molecular Biology, The Scripps Research Institute, La Jolla, CA 92037, USA;
3
Friedrich Miescher Institute, CH-4002 Basel, Switzerland
Intracellular expression of single-chain antibodies (scFvs)
represents a promising approach for selective interference
with cellular proto-oncogenes such as the epidermal
growth factor receptor (EGFR). Previously, we have
shown that intrabodies targeted to the lumen of the
endoplasmic reticulum prevent the transit of EGFR or the
related ErbB2 molecule to the cell surface, thereby
inactivating their transforming potential. While intramolecular disulfide bridges important for antibody stability
are correctly formed during expression in the secretory
pathway, scFvs expressed in the reducing environment of
the cytosol are often inactive. To overcome this problem
and to generate antibody fragments that interact with the
intracellular domain of human EGFR in the cytoplasm,
here we have chosen a two-step approach combining
classical selection of scFvs by phage display with
subsequent expression in yeast. After enrichment of
EGFR-specific antibody fragments from a combinatorial
library by biopanning, a yeast two-hybrid screen was
performed using the intracellular domain of EGFR as
bait. Screening of 1.5 105 preselected scFv plasmids
under highly stringent conditions yielded 223 colonies that
represented at least five independent scFv clones functional in the intracellular milieu of eukaryotic cells.
Interaction of selected antibody fragments with the
intracellular domain of EGFR was confirmed in GST
pull-down and coimmunoprecipitation experiments. Upon
cytoplasmic expression in human tumor cells, scFvs
colocalized with EGFR at the plasma membrane demonstrating their functionality in vivo.
Oncogene (2003) 22, 1557–1567. doi:10.1038/sj.onc.
1206299
Keywords: epidermal growth factor receptor; receptor
tyrosine kinase; single-chain Fv antibody; intrabody;
yeast two-hybrid
*Correspondence: Winfried Wels, Chemotherapeutisches Forschungsinstitut Georg-Speyer-Haus, Paul-Ehrlich-Strae 42-44, D-60596
Frankfurt am Main, Germany;
E-mail: wels@em.uni-frankfurt.de
4
Current address: Cytos Biotechnology AG, CH-8952 Zürich-Schlieren, Switzerland
Received 18 September 2002; revised 3 December 2002; accepted 5
December 2002
Introduction
Epidermal growth factor receptor (EGFR, ErbB) is a
member of the ErbB family of receptor tyrosine kinases
that also includes the closely related ErbB2 (HER2/
Neu), ErbB3 (HER3), and ErbB4 (HER4) proteins
(Olayioye et al., 2000; Yarden and Sliwkowski, 2001).
They share a common molecular architecture and
consist of a glycosylated extracellular domain to
which peptide ligands bind, a single transmembrane
region, and a cytoplasmic domain with tyrosine kinase
activity. EGFR was originally identified as a cellular
proto-oncogene because of its homology with the
retrovirally transduced chicken v-erbB (Downward
et al., 1984). In humans, EGFR gene amplification
and protein overexpression have been found in glioblastoma and many cancers of epithelial origin, and
have been linked with cancer development and progression (Gullick, 1991; Klapper et al., 2000). This has
prompted the development of therapeutic strategies
aimed at targeted interference with EGFR function
(Mendelsohn and Baselga, 2000). One of these approaches is based on the antagonistic, partially humanized antibody C225 that blocks ligand binding to the
extracellular domain of EGFR and has shown promising effects in clinical studies (Baselga et al., 2000;
Shin et al., 2001).
Experimentally, ligand-induced EGFR activation
could be prevented in an autocrine fashion, and
anchorage-independent growth could be inhibited upon
expression of a single-chain (sc) Fv fragment of
monoclonal antibody 225 directly in EGFR-transformed fibroblasts (Beerli et al., 1994a, 1996). Also,
retention of EGFR or the related ErbB2 molecule (King
et al., 1985) in the endoplasmic reticulum (ER) by ERresident scFvs directed to the receptors’ extracellular
domains severely affected receptor signaling and growth
of tumor cells (Beerli et al., 1994b, 1996; Deshane et al.,
1994; Jannot et al., 1996). Cell lines rendered physiologically ErbB2 negative in this way have helped to
identify ErbB2 as the preferred coreceptor in the ErbB
receptor signaling network (Beerli et al., 1995; Karunagaran et al., 1996; Graus-Porta et al., 1997). Thus,
antibodies expressed intracellularly (intrabodies) might
be generally applicable for selective interference with
signal transducing proteins or other cellular targets
Intracellular antibodies specific for human EGF receptor
S Hyland et al
1558
(Marasco, 1995, 1997; Steinberger et al., 2000; Goncalves et al., 2002).
Upon ligand-induced activation, EGFR connects
extracellular signals to multiple intracellular pathways
depending on adapter proteins and substrates associated
with its cytoplasmic domain (Gschwind et al., 2001;
Yarden and Sliwkowski, 2001). While scFv antibodies
targeted to distinct intracellular regions of EGFR might
allow to interfere with these signaling pathways
individually, the application of intrabodies for the
analysis of cytoplasmic target proteins or protein
domains is complicated by the limited functionality of
scFvs in the cytosol. For stability and functional
activity, antibodies are usually dependent on the
formation of intradomain disulfide bonds within the
variable regions of antibody heavy and light chains.
These disulfide bridges are correctly assembled upon
expression in the secretory pathway of cells including the
ER, but they do not form in the reducing cytoplasmic
environment (Biocca et al., 1995; Proba et al., 1997;
Martineau et al., 1998). This requires alternative
stabilizing features of intracellular antibodies to be
potent enough to promote correct folding and prevent
aggregation under these conditions (Biocca et al., 1995;
Proba et al., 1997; Wörn and Plückthun, 1998; Cattaneo
and Biocca, 1999; Zhu et al., 1999). Only in some cases,
individual antibody fragments not specifically selected
for intracellular expression could be produced as
functional cytoplasmic proteins (Biocca et al., 1994;
Duan et al., 1994; Mhashilkar et al., 1995; Gargano and
Cattaneo, 1997). Therefore, different strategies have
been suggested to more reliably generate antibody
fragments that are active in the cytosol (Cattaneo and
Biocca, 1999; Ohage et al., 1999; Wörn and Plückthun,
2001). Among these, expression of scFvs in yeast is
attractive since it is possible to select from antibody
libraries and directly test for binding activity in a
eucaryotic system (Visintin et al., 1999; Wörn et al.,
2000; De Jaeger et al., 2000; Tse et al., 2002).
Here we have applied a two-step strategy to isolate
intracellular scFv fragments that interact with the
cytoplasmic domain of human EGFR. After enrichment
of phage-displayed, EGFR-specific antibody fragments
from a combinatorial library, a yeast two-hybrid screen
was performed to identify scFvs that are functional in
the intracellular milieu of eukaryotic cells. Subsequently,
interaction of antibody fragments with the intracellular
domain of EGFR in vitro and in mammalian cells in vivo
was confirmed in GST pull-down and coimmunoprecipitation experiments. Confocal laser scanning micro-
scopy proved that upon cytoplasmic expression in
human tumor cells, scFvs colocalize with EGFR at the
plasma membrane.
Results
Enrichment of EGFR-specific scFvs from a phage
displayed mouse antibody library
A scFv antibody library was generated from cDNA
derived from the spleens of mice immunized with
recombinant human EGFR kinase domain following
procedures established previously (Barbas et al., 1991;
Andris-Widhopf et al., 2001; Rader and Barbas, 1997).
ScFvs in this library consist of an N-terminal k lightchain variable domain and a C-terminal heavy-chain
variable domain, linked by a seven or an 18 amino-acid
flexible linker. The initial library was highly diverse,
consisting of more than 2 108 independent clones.
Antibodies binding the EGFR kinase domain were
enriched by several rounds of panning against immobilized EGFR antigen. In the third panning round, the
ratio of output to input phage started to increase and
enhanced binding to EGFR was detected by ELISA,
indicating a successful enrichment of EGFR-specific
scFvs (Table 1). This preselected library was subcloned
for yeast two-hybrid screening.
Construction of a yeast library for expression of EGFRspecific scFv fragments
ScFv fragments were isolated from phagemids and
inserted into modified pGADT7 yeast two-hybrid prey
vector. The resulting plasmids encode fusion proteins
consisting of the GAL4 transactivation domain (AD), a
sequence tag derived from influenza virus hemagglutinin
(HA), and a C-terminal scFv antibody (Figure 1a, upper
panel). A yeast two-hybrid bait vector was constructed
by inserting an EGFR cDNA fragment encoding aminoacid residues 645–1186 30 of the GAL4 DNA-binding
domain (BD) into plasmid pGBKT7 (Figure 1a, lower
panel). Saccharomyces cerevisiae AH109 cells were
transformed with the resulting pGBKT7–EGFR vector
and expression of the GAL4 BD–EGFR fusion protein
was verified by immunoblot analysis of yeast lysates
using GAL4 BD- and EGFR-specific antibodies
(Figure 1b). Subsequently, yeast cells harboring the bait
construct were also transformed with the pGADT7–
scFv library.
Table 1 Enrichment of EGFR-specific scFv antibody phage by biopanning
Panning round
Unselected
1
2
3
4
Input
Output
Output/input
EGFR ELISAa (A 405 nm)
–
6.0 1012
2.0 1012
2.1 1012
1.1 1012
–
1.8 107
8.4 105
2.4 107
4.3 108
–
3.0 106
4.2 107
1.1 105
3.9 104
0.149
0.093
0.126
0.663
0.734
a
Binding of phage pools to immobilized EGFR antigen was determined in enzyme-linked immunosorbent assays using M13 phage-specific primary
and horseradish-peroxidase coupled secondary antibodies
Oncogene
Intracellular antibodies specific for human EGF receptor
S Hyland et al
1559
Selection of EGFR-specific scFv fragments by yeast twohybrid screening
Upon intracellular coexpression of GAL4 BD–EGFR
with GAL4 AD–scFv molecules strongly interacting
with EGFR, functional GAL4 is reconstituted and
drives expression of the GAL4 regulated ADE2, HIS3,
and MEL1/lacZ reporter genes integrated in the genome
of S. cerevisiae AH 109 (Figure 2a, upper panel). In
contrast, expression of inactive antibody fragments or
scFvs not binding to EGFR cannot result in reporter
gene activation and survival on selection media
(Figure 2a, lower panel). Yeast cells containing EGFR
bait and scFv prey vectors, and expressing ADE2 and
HIS3 reporter genes were selected on high-stringency
selection media lacking tryptophan, leucine, histidine,
and adenine. Subsequently, the resulting yeast colonies
were replated on the same selection media containing Xa-Gal to test for expression of a-galactosidase
(Figure 2b). A total of 223 blue yeast colonies were
obtained. Control cells transformed with pGBKT7–
EGFR plasmid and empty pGADT7 vector, or empty
pGBKT7 and pGADT7-scFv did not grow on highstringency selection media demonstrating that interaction between bait and prey in the selected yeast clones
was specific and dependent on the simultaneous
presence of scFv and EGFR protein domains (data
not shown).
Molecular analysis of EGFR-specific scFv fragments
Figure 1 Vectors for intracellular expression of the anti-EGFR
scFv antibody library and EGFR cytoplasmic domain in yeast. (a)
ScFv antibody fragments from a preselected anti-EGFR phagemid
library were subcloned into modified pGADT7 yeast two-hybrid
prey vector for fusion to the GAL4-activation domain (AD) and
the influenza virus hemagglutinin tag (HA). A cDNA fragment
encoding the intracellular domain of human EGFR was fused to
the GAL4 DNA-binding domain (BD) in bait vector pGBKT7. (b)
Expression of GAL4 BD-EGFR fusion protein in yeast was
confirmed by immunoblot analysis of yeast lysates with EGFRand GAL4 BD-specific antibodies followed by HRP-coupled
species-specific antibodies and chemiluminescent detection
Plasmid DNA of 24 randomly chosen, a-galactosidaseexpressing colonies was isolated and upon PCR
amplification of scFv inserts, the diversity of selected
antibody fragments was analysed by digestion of PCR
products with restriction enzyme Mval (Figure 3a).
Based on the resulting restriction patterns, scFvs were
subgrouped into five different groups represented by
scFv fragments # 2 (group of 13 clones), # 8 (group of
seven clones), and # 29 (group of two clones). ScFv
clones # 22 and # 30 displayed unique Mval restriction
Figure 2 Selection of intracellularly active, EGFR-specific scFv fragments by yeast two-hybrid screening. (a) Upon expression of
correctly folded, EGFR-binding scFv fragments, functional GAL4 is reconstituted and drives expression of the GAL4-regulated
ADE2, HIS3, and MEL1/lacZ reporter genes integrated in the genome of S. cerevisiae AH109 (upper panel). In yeast, clones
expressing misfolded antibody fragments or scFvs not binding to EGFR, GAL4 BD and AD remain separated and transcription of
reporter genes is not induced (lower panel). Yeast clones containing bait and prey vectors and expressing ADE2 and HIS3 reporter
genes were selected on media lacking tryptophan, leucine, histidine, and adenine. Then positive clones were tested for expression of agalactosidase on media containing X-a-Gal. Representative yeast clones expressing anti-EGFR scFv fragments # 2, 8, 22, 29, and 30
are shown (b). Yeast cells expressing GAL4 BD-p53 and GAL4 AD-SV40 T antigen, and cells expressing GAL4 BD-EGFR and a
GAL4 AD fusion protein with Myc-tag-specific scFv(9E10) were included as positive and negative controls, respectively
Oncogene
Intracellular antibodies specific for human EGF receptor
S Hyland et al
1560
represented by scFvs 2, 8, and 30 contained the short
seven amino-acid linker.
Binding of in vitro translated scFv fragments to
recombinant EGFR
Interaction of yeast-selected scFv antibody fragments
with the cytoplasmic portion of human EGFR was
tested in glutathione S transferase (GST) pull-down
experiments as a cell-free system. Bacterially expressed
GST–EGFR fusion protein containing EGFR aminoacid residues 645–1186, or GST control protein were
immobilized on glutathione beads and incubated with in
vitro translated, 35S-labeled scFvs 2, 8, 22, 29, and 30.
Upon precipitation of beads, bound proteins were
analysed by SDS–PAGE and autoradiography of the
gels. As shown in Figure 4a, all scFvs tested displayed
specific binding to recombinant GST-EGFR protein but
not to GST alone. To further confirm specificity for
EGFR, a similar GST pull-down experiment was
conducted with 35S-labeled scFv 2 using a recombinant
GST fusion protein containing the intracellular portion
of human ErbB2, a tyrosine kinase closely related to
EGFR. Whereas binding to GST-EGFR could be
verified, no binding of scFv to GST-ErbB2 could be
detected (Figure 4b).
Figure 3 Molecular analysis of yeast-selected scFv clones. After
yeast two-hybrid selection, individual pGADT7-scFv plasmids
were isolated, scFv inserts were amplified by PCR and analysed by
digestion with Mval restriction enzyme. (a) Based on their DNA
fingerprinting patterns, isolated clones were subgrouped into five
different groups represented by scFv fragments # 2, 8, 22, 29, and
30. (b) Sequencing of scFv cDNA of these clones revealed high
homology in the deduced amino-acid sequences of the complementarity determining regions (CDRs) of antibody light (VL)- and
heavy-chain variable domains (VH)
patterns. Reconstitution of functional GAL4 upon
coexpression of the individual GAL4 AD–scFv fragments and GAL4 BD–EGFR protein was confirmed for
all five clones upon amplification of plasmid DNA in
Escherichia coli and retransformation of yeast cells. In
contrast, transformation of yeast with pGBKT7–EGFR
and a control pGADT7–scFv vector facilitating intracellular expression of Myc-tag-specific scFv(9E10) fused
to GAL4 AD did not result in colonies surviving
selection (Figure 2b).
Subsequent analysis of DNA and deduced amino-acid
sequences revealed high homology in the framework and
complementarity determining regions (CDRs) (Kabat
et al., 1991) of the scFvs’ heavy and light chains
(Figure 3b and data not shown). The CDRs in the VH
region of scFvs 22 and 30 were identical, but VH
framework sequences of these clones differed in several
amino-acid positions. Interestingly, while the unselected
yeast library consisted primarily of scFv plasmids with
an 18 amino-acid linker, the majority of selected clones
Oncogene
Interaction of scFv fragments with EGFR in the
cytoplasm of mammalian cells
To investigate whether scFv fragments selected in yeast
are also functional upon cytoplasmic expression in
mammalian cells, we performed coimmunoprecipitation
experiments. COS-7 cells were cotransfected with
plasmids encoding full-length human EGFR and HAtagged scFv fragments. After transfection, cell lysates
were incubated with humanized C225 antibody binding
to EGFR extracellular domain (Mendelsohn, 1997) and
protein G sepharose beads. Upon precipitation, bound
proteins were analysed by SDS–PAGE and immunoblotting with anti-HA antibody as shown in Figure 5.
Protein bands migrating at the size expected for scFv
fragments were detected in EGFR-immunoprecipitates
of cells expressing scFvs 2, 8, 29, and 30, indicating that
these antibody fragments are functional and bind to
EGFR upon cytoplasmic expression. In contrast,
intracellularly expressed Myc-tag-specific control
scFv(9E10) did not coprecipitate with EGFR (Figure 5,
lower panel). Also, in EGFR immunoprecipitates from
cells expressing scFv 22 no signal was detected. While
this might indicate that this antibody is not functional in
mammalian cells, it could also be because of the low
level of scFv expression found in these cells (data
not shown). In most samples, a second protein
band migrating below the scFv fragment was detected.
This protein most likely bound nonspecifically to
protein G beads since it was also present when cell
lysates were precipitated in the absence of anti-EGFR
antibody.
Intracellular antibodies specific for human EGF receptor
S Hyland et al
1561
Figure 4 Binding of in vitro translated scFv fragments to EGFR. (a) Binding of yeast-selected scFv fragments to the cytoplasmic part
of EGFR was investigated in glutathione S transferase (GST) pull-down assays using bacterially expressed GST–EGFR fusion protein
and in vitro translated, 35S-labeled antibody fragments. Binding of scFvs to glutathione-coupled beads either carrying GST alone (G) or
GST–EGFR (E) was analysed by SDS–PAGE of proteins present in precipitated complexes, followed by autoradiography. scFv input
(10%) were included as a control (I). (b) Specificity of scFv interaction with EGFR was confirmed further by conducting GST pulldown assays using GST, GST–EGFR and GST–ErbB2 fusion proteins as indicated. The presence of equivalent amounts of GST
fusion proteins on precipitated glutathione-coupled beads was controlled by Coomassie staining of SDS–PAGE gels (lower panel)
Intracellular localization of scFv fragments in human
tumor cells
To investigate activity and localization of scFv fragments upon cytoplasmic expression in human tumor
cells overexpressing EGFR, A431 epidermoid carcinoma cells were transiently transfected with plasmids
encoding HA-tagged, EGFR-specific antibody fragments scFv 2 and scFv 30. Cells were fixed and
permeabilized, and presence and localization of scFvs
and EGFR target antigen were analysed by detection
with HA-tag-specific and anti-EGFR C225 antibodies,
followed by FITC- and Cy5-labeled secondary antibodies and confocal laser scanning microscopy. As
expected, strong EGFR staining was detected at the
plasma membrane of A431 cells (Figure 6a, d, g).
Whereas intracellularly expressed Myc-tag-specific
scFv(9E10) was diffusely distributed throughout the
cytosol, and to a lesser extent the nucleus of transfected
cells (Figure 6h), anti-EGFR scFvs 2 and 30 were
exclusively localized to the plasma membrane and
vesicular structures also detected with anti-EGFR antibody (Figure 6b, e). Overlays of EGFR and scFv
specific signals confirmed colocalization of scFvs 2 and
30, but not control scFv(9E10) with EGFR (Figure 6c, f,
i). These results demonstrate that the anti-EGFR scFvs
are functional upon cytoplasmic expression in human
tumor cells and interact with the intracellular domain of
the target receptor in vivo.
Discussion
Single chain antibody fragments specific for growth
factor receptors and expressed directly in target cells
represent a promising concept to interfere selectively
Oncogene
Intracellular antibodies specific for human EGF receptor
S Hyland et al
1562
Figure 5 Binding of intracellular scFv fragments to EGFR in vivo.
COS-7 cells were transiently cotransfected with plasmids encoding
full-length human EGFR and HA-tagged, EGFR-specific antibody
fragments. Interaction of cytoplasmic scFvs with EGFR was tested
in coimmunoprecipitation experiments. Cleared cell lysates were
incubated with humanized C225 antibody binding EGFR extracellular domain and protein G sepharose beads (+), or protein G
sepharose beads in the absence of C225 (). Upon precipitation of
beads, bound proteins were analysed by SDS–PAGE and
immunoblotting with rat anti-HA IgG followed by HRP-coupled
secondary antibody and chemiluminescent detection. In vitro
translated anti-EGFR scFv 30 was included as a control for
detection of scFv. Lanes loaded either with C225 antibody, or
immunoprecipitate from cells transfected with EGFR plasmid and
cDNA encoding Myc-tag-specific scFv(9E10) served as negative
controls. An additional protein band not representing scFv but
crossreacting with the antibodies used for detection migrates below
the EGFR-specific antibody fragments
with the multiple signaling processes associated with
these proteins (Gschwind et al., 2001; Yarden and
Sliwkowski, 2001). Previously we could show that scFvs
targeted to the lumen of the ER prevent the transit of
EGFR or the related ErbB2 molecule to the cell surface,
thereby inactivating the transforming potential of these
receptor tyrosine kinases (Beerli et al., 1994b, 1996;
Jannot et al., 1996). While intramolecular disulfide
bridges important for antibody stability are correctly
formed during expression in the secretory pathway of
cells (Steinberger et al., 2000), single chain antibodies
expressed in the reducing environment of the cytosol are
often inactive and have a tendency to aggregate
(Cattaneo and Biocca, 1999). This complicates their
application for interference with intracellular target
proteins. To overcome this problem and to generate
antibody fragments that interact with the intracellular
domain of EGFR in the cytosol of mammalian cells,
here we have chosen a two-step approach combining
classical selection of EGFR-binding scFvs by phage
display and biopanning with subsequent screening in a
yeast two-hybrid system.
Different strategies have been developed to more
reliably create or identify intrinsically stable antibody
fragments (Wörn and Plückthun, 2001). Mutagenesis of
scFvs and in vitro selection under reducing conditions
resulted in molecules correctly folded in the absence of
Oncogene
disulfide bridges (Jermutus et al., 2001), a prerequisite
for functional cytoplasmic intrabodies (Biocca et al.,
1995; Cattaneo and Biocca, 1999). Grafting of the
antibody combining site into stabilized framework
sequences (Ohage et al., 1999) and fusion of scFvs to
another protein serving as a chaperone (Bach et al.,
2001) have led to the expression of antigen-binding
intrabodies in E. coli and might also be applicable in
eukaryotic cells. Expression of scFvs in yeast is an
attractive alternative since established methodologies
are available that allow screening of diverse libraries.
Furthermore, functionality of antibody fragments in the
cytoplasm and nucleus of lower eukaryotic cells appears
to be predictive for their behavior upon expression in
the intracellular milieu of mammalian cells (Visintin
et al., 1999; De Jaeger et al., 2000; Wörn et al., 2000; Tse
et al., 2002). After enrichment of phage-displayed,
EGFR-specific antibody fragments from a combinatorial library and transfer into a yeast two-hybrid screen,
we could isolate a panel of scFvs that are active in the
cytosol and bind strongly to the intracellular domain of
human EGFR.
In unselected combinatorial libraries antibody fragments that fold correctly in the absence of disulfide
bonds and interact with their target antigen in the
reducing intracellular environment appear to be very
rare. When a mixture of 5 105 plasmids from a random
scFv library was screened in a yeast two-hybrid system
using the coat protein of a plant virus as a model
antigen, no functional intrabody could be selected,
whereas a scFv fragment previously shown to bind to
this target antigen under cytoplasmic conditions and
added to the library at a frequency of 1 in 5 105 was
successfully recovered (Visintin et al., 1999). Indeed,
direct yeast two-hybrid screening of a portion of the
unselected anti-EGFR scFv library (approximately 105
independent plasmids) did not result in the identification
of functional EGFR-specific intrabodies (data not
shown). Owing to limitations in transformation efficiency, quantitative transfer of combinatorial scFv
libraries of more than approximately 106 independent
clones into a yeast screening system is not practical (Tse
et al., 2002). Therefore, EGFR-binding antibody fragments in our initial phage display library were enriched
by three rounds of biopanning, reducing the complexity
of the library to a value below 106, which proved
suitable for subsequent expression in yeast. Screening of
1.5 105 pGADT7-scFv plasmids yielded 223 a-galactosidase-expressing yeast colonies that represented at
least five independent scFv clones.
Interaction of the isolated antibody fragments with
the intracellular domain of EGFR in vitro and in
mammalian cells in vivo was confirmed in GST pulldown and co-immunoprecipitation experiments. Thereby no crossreactivity of scFvs with the tyrosine kinase
domain of the closely related ErbB2 protein was found.
Two of the scFv molecules were also tested in
colocalization studies in human tumor cells. Upon
cytoplasmic expression, the EGFR-specific antibody
fragments were found together with EGFR at the
plasma membrane, whereas a control scFv was diffusely
Intracellular antibodies specific for human EGF receptor
S Hyland et al
1563
Figure 6 Intracellular localization of scFv fragments in A431 cells. EGFR overexpressing A431 tumor cells transiently transfected
with plasmids encoding HA-tagged, EGFR-specific antibody fragments scFv 2 (a, b, c) and scFv 30 (d, e, f) were fixed and
permeabilized. Cells were stained with EGFR-specific C225 antibody followed by Cy5-labeled secondary antibody (a, d, g), and HAtag-specific antibody followed by FITC-labeled secondary antibody (b, e, h). Localization was analysed by confocal laser scanning
microscopy. Cy5 and FITC channels were merged to demonstrate colocalization (c, f, i). Myc-tag-specific scFv(9E10) was included as a
control (g, h, i)
distributed throughout the cytosol and nucleus. We have
also investigated possible effects of intrabody expression
on EGFR activation. For anti-EGFR scFv(225) that
blocks ligand binding to the extracellular domain of the
receptor, it was shown that autocrine production of the
scFv inhibits EGF-dependent activation and growth of
transformed cells (Beerli et al., 1994a, 1996). In contrast,
in COS-7 cells transiently transfected with constructs
encoding human EGFR and cytoplasmic intrabodies
specific for the intracellular part of EGFR, no apparent
effects on EGFR tyrosine phosphorylation were noted
in the presence or absence of EGF (data not shown).
This suggests that the scFvs do not directly block
sequences crucial for EGFR kinase activity such as the
ATP-binding site (Russo et al., 1985), while it does not
rule out the possibility of subtle effects on the
interaction of the receptor with individual downstream
substrates or adapter proteins.
In the case of p21Ras or phospholipase C-g1,
cytoplasmic intrabodies inhibited activity of their target
proteins by sequestering them to antibody aggregates
with a different subcellular localization and increased
proteasomal degradation of the resulting antibody–
antigen complexes (Lener et al., 2000; Cardinale et al.,
2001; Yi et al., 2001). Thereby for p21Ras-specific
intrabodies to be effective, direct interference with the
intrinsic GTPase activity was not required (Lener et al.,
2000). Here confocal laser scanning microscopy showed
that in the presence of cytoplasmic scFvs, EGFR
remained at its normal location at the plasma membrane. Intrabodies were not sequestered to aggregates
but colocalized with the target receptor, making scFvdependent redistribution of EGFR unlikely to occur.
However, modification of scFv sequences by introducing intracellular routing and ubiquitination signals
might allow to downmodulate EGFR signaling by
redirecting the receptor to alternative cellular compartments or enhance its proteasomal degradation (Persic
et al., 1997; Cattaneo and Biocca, 1999; Parisien et al.,
2001; Kile et al., 2002).
The scFv plasmid library used for selection in yeast
consisted of a mixture of antibody fragments either
containing a seven or an 18 amino-acid linker sequence
to connect light- and heavy-chain variable domains.
Oncogene
Intracellular antibodies specific for human EGF receptor
S Hyland et al
1564
Before yeast selection, the majority of scFvs carried the
long linker. However, 21 of 24 anti-EGFR intrabodies
individually analysed after selection harbored the short
linker, suggesting an enrichment of these clones during
yeast two-hybrid screening. Short linker sequences
usually prevent proper intramolecular association of a
scFv’s heavy- and light-chain sequences, but favor
noncovalent interaction of VH and VL from two
neighboring scFv molecules resulting in a bivalent
antibody fragment termed diabody (Holliger et al.,
1993). During screening of scFvs under highly stringent
conditions in yeast, bivalent intrabodies interacting
more strongly with the target antigen in the cytosol
might have had a selective advantage over similar
monovalent molecules.
Despite different linker length, the amino-acid
sequences of heavy- and light-chain variable domains
of the isolated anti-EGFR scFvs were highly homologous. Previously it has been described that functional BCR-ABL-specific intrabodies selected from a
library in yeast carried only a limited set of different
framework regions, possibly because of their contribution to enhanced intracellular stability and solubility
(Tse et al., 2002). However, while in that case a high
degree of diversity was still found in the CDR3 regions
that are most important for antigen specificity (Xu
and Davis, 2000), here CDR3 sequences of different
independent scFv clones were either identical or differed only in a few amino-acid positions. Therefore,
most of the scFvs isolated from yeast might be
directed against the same EGFR epitope. This could
be because of the preferential enrichment of scFvs
recognizing an immunodominant epitope during biopanning of the phage display library, or selection of
antibody fragments recognizing a sequence best accessible upon expression of the cytoplasmic domain of
EGFR in yeast.
We showed that antibody phage display methodology
combined with subsequent selection in a yeast twohybrid system is a suitable strategy to identify EGFRspecific scFv antibodies that are functional upon
cytoplasmic expression in tumor cells. Since an appropriate screening system has now been established,
starting out from a more diverse, nonimmune phage
display library (Tse et al., 2002), and using for selection
a variety of smaller EGFR fragments might allow to
isolate scFv intrabodies recognizing additional EGFR
epitopes. Our results suggest that similar approaches
could also be applicable for the generation of intrabodies targeted to the cytoplasmic domains of other ErbB
receptor tyrosine kinases and a wide range of more
diverse growth factor receptors.
bad, CA, USA). First strand cDNA was prepared from 20 mg
of total RNA using AMV Reverse Transcriptase (Invitrogen).
Heavy- and light-chain variable domain coding regions were
amplified by PCR and linked by overlap extension as described
previously (Andris-Widhopf et al., 2001). The final PCR
product encoding a library of scFv antibodies, where heavyand light-chain variable domains are linked either by a seven
(GGSSRSS) or an 18 (SSGGGGSGGGGGGSSRSS) aminoacid linker, was digested with the restriction endonuclease Sfil,
gel-purified and cloned into Sfil-digested pComb3H phagemid
DNA (Barbas et al., 1991; Rader and Barbas, 1997). The
primary library consisted of 2.4 108 independent transformants. Phage displaying scFv was produced as described
(Barbas et al., 1991, 2001).
Selection of EGFR-specific scFv antibodies by phage display
Panning was carried out against immobilized antigen in wells
of a 96-well plate, using 1 mg of human EGFR kinase domain
in 25 ml of TBS for coating. Coated wells were blocked with
TBS/3% BSA for 1 h at 371C. In total, 2–3 1012 PFU of scFvdisplaying phage was then applied per antigen-coated well in
50 ml TBS/1% BSA and incubated for 2 h at 371C. In the first
selection round, two wells were used, whereas in the
consecutive rounds, only one well was used. Wells were
washed five times (first round) or 10 times (second and third
round) with TBS/0.1% Tween-20. Bound phage were then
eluted with 50 ml Trypsin (10 mg/ml) for 30 min at 371C and
used to infect exponentially growing E. coli ER2537 (New
England Biolabs, Beverly, MA, USA). Phage was grown and
purified from the supernatants of overnight cultures as
described (Barbas et al., 1991).
Construction of an anti-EGFR scFv library for yeast two-hybrid
screening
A derivative of pGADT7 yeast two-hybrid prey vector
(Clontech, Heidelberg, Germany) suitable for subcloning of
scFv antibody fragments was constructed by amplifying a
cDNA fragment encoding Myc-tag-specific scFv(9E10) antibody via PCR using plasmid pWW152-9E10 (Altenschmidt
et al., 1997) as template and oligonucleotides scFv 50 Ndel/Sfil
50 -aaaaaacatatggcccaggcggccCAGGT(G/C)(A/C)A(A/G)CTGCAG(G/C)AGTC(A/T)GG-30 and scFv 30 Sfil/Xbal/Sacl
50 -aaaaaagagctcactctagacggccggcctggccAA(G/T)CTCGAG(C/
T)TT-(G/T)GT(G/C)C-30 (scFv sequence in upper case)
introducing Ndel and Sfil at the 50 end, and Sfil, Xbal and
Sacl restriction sites at the 30 end of the PCR fragment. The
amplified scFv(9E10) fragment was digested with Ndel and
Sacl and ligated in frame 30 of GAL4 activation domain in the
prey vector, resulting in plasmid pGADT7-9E10. pComb3H
phagemid DNA containing the preselected anti-EGFR scFv
library after the third panning round was digested with Sfil,
scFv DNA fragments were isolated and used to replace
scFv(9E10) in Sfil-digested pGADT7-9E10. The complexity of
the resulting pGADT7-scFv library was 1.5 105.
Yeast two-hybrid screening
Materials and methods
Construction of a mouse scFv phage display library
Total RNA was isolated from the spleens of Balb/c mice
immunized with purified recombinant human EGFR kinase
domain (kindly provided by Dr N Lydon, now Amgen Inc.,
Thousand Oaks, CA, USA) using Trizol (Invitrogen, CarlsOncogene
For construction of the yeast two-hybrid bait vector, a cDNA
fragment encoding the cytoplasmic part of EGFR (amino-acid
residues 645–1186) was derived by PCR using human EGFR
cDNA as a template and oligonucleotides EGFR 50 Smal 50 ggcctcccggggCGAAGGC-GCCAC-30 and EGFR 30 Sall 50 atactagtcgacgtggTCATGCTCC-30 that introduced Smal and
Sall restriction sites at the 50 and 30 ends of the PCR product
(EGFR sequences in upper case). The EGFR cDNA fragment
Intracellular antibodies specific for human EGF receptor
S Hyland et al
1565
was digested with Smal and Sall and inserted into plasmid
pGBKT7 (Clontech). The resulting vector pGBKT7–EGFR
encodes GAL4 DNA-binding domain fused in frame to the
cytoplasmic fragment of human EGFR.
Yeast transformation and two-hybrid screening procedures
were carried out according to the manufacturer’s recommendations (Matchmaker GAL4 two-hybrid system 3; Clontech).
Briefly, competent S. cerevisiae AH109 cells were prepared by
the PEG/lithium acetate method and sequentially transformed
by heat shock with pGBKT7–EGFR and the anti-EGFR
pGADT7–scFv library. Transformants were selected for the
presence of both vectors on media lacking leucin and
tryptophan. Expression of the bait protein was confirmed by
immunoblot analysis of yeast extracts using anti-GAL4 BD
IgG (Clontech) or 15E polyclonal anti-EGFR serum (Gullick
et al, 1985). Yeast clones expressing all three reporter genes
ADE2, HIS3, and MEL1/lacZ were selected on high-stringency selection media lacking adenine and histidine, and
containing X-a-Gal for detection of a-galactosidase activity.
pGADT7–scFv plasmids were isolated from single yeast
colonies and scFv inserts were amplified by PCR using 30
AD and T7 primers (Clontech). ScFv PCR products were
digested with Mval and fragments were separated by gel
electrophoresis. One prototype pGADT7–scFv plasmid for
each group of identical Mval restriction patterns was chosen,
and scFv interaction with EGFR was confirmed by retransformation and selection of yeast cells. Subsequently,
pGADT7-scFv plasmids containing functionally active antibody fragments were digested with Sfil and scFv cDNAs were
inserted into a pcDNA3.1/Zeo (+) vector (Invitrogen,
Karlsruhe, Germany) previously modified by including Sfil
restriction sites for subcloning and a C-terminal HA-tag for
protein detection.
GST pull-down assays
For bacterial expression of a GST–EGFR fusion protein, a
cDNA fragment encoding amino-acid residues 645–1186 of
human EGFR was amplified by PCR using oligonucleotides
EGFR 50 Sall/Xhol 50 -aaaaaagtcgactcgagCGAAGGCGCCACATCGTTC-30 and EGFR 30 Xbal/Notl 50 -aaaaaagcggccgctctagaGCTCCAATAAATTCACTGC-30 (EGFR sequences in
upper case), and was cloned in frame with GST in the vector
pGEX-4T-1 (Amersham Biosciences, Freiburg, Germany). E.
coli BL21-CodonPlus (DE3)-RP (Stratagene, Heidelberg,
Germany) harboring the resulting pGEX-4T-1-EGFR vector
were induced with 1 mm IPTG at an OD600 of 0.8 and grown
overnight at 201C. Bacterial cell lysates were prepared by
sonication in PBS containing 0.5 mm PMSF, 1 mm EDTA, 1%
Triton X-100, and were cleared by centrifugation. Cleared
lysates were incubated with glutathione sepharose beads
(Amersham Biosciences) for 1 h at 41C, and beads were
subsequently washed four times with PBS. Binding reactions
typically contained 1 mg of recombinant GST–EGFR fusion
protein or 10 mg GST protein as a negative control immobilized on glutathione sepharose beads. In vitro translation of
scFv antibody fragments was carried out with the T7 promoter
containing pGADT7-scFv plasmids using the TNT T7 Quick
Coupled Transcription/Translation System (Promega, Mannheim, Germany) in the presence of 35S-labeled methionine
according to the manufacturer’s recommendations, in vitro
translated scFv fragments were added to GST–EGFR or GST
bound to glutathione sepharose beads in 1.3 ml binding buffer
containing 50 mm Tris-HCl pH 8.0, 100 mm NaCl, 10 mm
MgCl2, 10% glycerol, 0.1% NP-40, 0.3 mm DTT, 1 mm EDTA
and complete EDTA-free protease inhibitor mix according to
the manufacturer’s recommendation (Roche Diagnostics,
Mannheim, Germany). Samples were incubated with gentle
inversion for 2 h at 41C, followed by three washing steps with
binding buffer. Washed beads were resuspended in 25 ml of
5 SDS loading buffer, boiled, and liberated proteins were
analysed by SDS–PAGE.
Maintenance of cells and transfection procedures
COS-7 SV40-transformed African green monkey kidney cells
and A431 human epidermoid tumor cells were maintained in
Dulbecco’s modified Eagle’s medium (DMEM) containing
10% fetal bovine serum (FBS), 2 mm l-glutamine, 100 U/ml
penicillin, 100 mg/ml streptomycin. For analysis of interaction
of scFv antibody fragments with human EGFR in vivo, COS-7
cells were transiently cotransfected with full-length EGFR
encoding plasmid pLTR-EGFR (Hills et al., 1995) (kindly
provided by Dr WJ Gullick, University of Kent at Canterbury,
Canterbury, UK), and pcDNA3.1/Zeo (+) vectors encoding
HA-tagged anti-EGFR scFvs or Myc-tag-specific scFv(9E10)
as a control. As a transfection reagent in all experiments
Fugene 6 (Roche Diagnostics) was used according to the
manufacturer’s recommendations.
Coimmunoprecipitation and immunofluorescence staining
At 48 h after transfection, COS-7 cells were lysed in 2 ml of icecold lysis buffer containing 1% Triton X-100, 50 mm Tris pH
7.5, 150 mm NaCl, 5 mm EGTA, and complete EDTA-free
protease inhibitor mix. Following 30 min incubation on ice,
lysates were cleared by centrifugation and proteins binding
unspecifically to protein G sepharose were removed by
incubation of 900 ml cleared lysates for 3 h at 41C with 50 ml
of a 50% slurry of protein G sepharose beads (Amersham
Biosciences). After separation from the beads by centrifugation, supernatant was collected and incubated for 1 h at 41C
with 2 mg of monoclonal C225 humanized anti-EGFR IgG
(kindly provided by Dr J Mendelsohn, MD Anderson Cancer
Center, Houston, TX, USA) followed by addition of 50 ml of
protein G sepharose beads and gentle rotation for 3 h at 41C.
Beads were collected by centrifugation and washed three times
using lysis buffer. Bound proteins were eluted in 50 ml of SDS
loading buffer and analysed by SDS–PAGE and immunoblotting with rat anti-HA tag antibody followed by horseradish
peroxidase-conjugated goat anti-rat IgG and chemiluminescent detection.
Intracellular localization of scFv antibody fragments was
analysed in A431 cells transiently transfected with pcDNA3.1/
Zeo-scFv constructs. Transfected cells were grown on coverslips and 48 h after transfection were fixed with 3% paraformaldehyde for 10 min at room temperature, followed by
permeabilization with 0.1% Triton X-100 in PBS for 5 min.
Cells were washed with PBS and incubated for 1 h with
primary antibodies C225 anti-EGFR and rat anti-HA tag
(Roche Diagnostics) in PBS containing 3% BSA, after
intensive washing followed by incubation with Cy-5 and
FITC-labeled Alexa FluorTM 633 goat anti-human IgG and
Alexa FluorTM 488 goat anti-rat IgG (Molecular Probes,
Eugene, OR, USA). Then samples were analysed using a Leica
TCS SL laser scanning microscope (Leica Mikrosysteme,
Bensheim, Germany).
Acknowledgements
We thank Gesa Brochmann-Santos for excellent technical
assistance during the initial phase of the project. This work
was supported in part by grants to WW from the National
Genome Research Network (NGFN) program of the German
Oncogene
Intracellular antibodies specific for human EGF receptor
S Hyland et al
1566
‘Bundesministerium für Bildung, Wissenschaft, Forschung und
Technologie’ (BMBF), and the German–Israeli Foundation
for Scientific Research and Development. The laboratory of
NEH is supported by the Novartis Forschungsstiftung
Zweigniederlassung Friedrich Miescher Institute for Biomedical Research.
References
Altenschmidt U, Schmidt M, Groner B and Wels W. (1997).
Int. J. Cancer, 73, 117–124.
Andris-Widhopf J, Steinberger P, Fuller R, Rader C and
Barbas III CF. (2001). Phage Display: A Laboratory Manual.
Barbas III CF, Burton DR, Scott JK and Silverman GJ
(eds). Cold Spring Harbor Laboratory Press: Plainview, NY,
pp. 9.1–9.113.
Bach H, Mazor Y, Shaky S, Shoham-Lev A, Berdichevsky Y,
Gutnick DL and Benhar I. (2001). J. Mol. Biol., 312,
79–93.
Barbas III CF, Burton DR, Scott JK and Silverman GJ (2001).
Phage Display: A Laboratory Manual. Cold Spring Harbor
Laboratory Press: Plainview, NY.
Barbas III CF, Kang AS, Lerner RA and Benkovic SJ. (1991).
Proc. Natl. Acad. Sci. USA, 88, 7978–7982.
Baselga J, Pfister D, Cooper MR, Cohen R, Burtness B, Bos
M, D’Andrea G, Seidman A, Norton L, Gunnett K, Falcey
J, Anderson V, Waksal H and Mendelsohn J. (2000). J. Clin.
Oncol., 18, 904–914.
Beerli RR, Graus-Porta D, Woods-Cook K, Chen X,
Yarden Y and Hynes NE (1995). Mol. Cell. Biol., 15,
6496–6505.
Beerli RR, Wels W and Hynes NE (1994a). Biochem. Biophys.
Res. Commun., 204, 666–672.
Beerli RR, Wels W and Hynes NE (1994b). J. Biol. Chem.,
269, 23931–23936.
Beerli RR, Wels W and Hynes NE (1996). Breast Cancer Res.
Treat., 38, 11–17.
Biocca S, Pierandrei-Amaldi P, Campioni N and Cattaneo A.
(1994). Biotechnology (NY), 12, 396–399.
Biocca S, Ruberti F, Tafani M, Pierandrei-Amaldi P
and Cattaneo A. (1995). Biotechnology (NY), 13,
1110–1115.
Cardinale A, Filesi I and Biocca S. (2001). Eur. J. Biochem.,
268, 268–277.
Cattaneo A and Biocca S. (1999). Trends Biotechnol., 17, 115–
121.
De Jaeger G, Fiers E, Eeckhout D and Depicker A. (2000).
FEBS Lett., 467, 316–320.
Deshane J, Loechel F, Conry RM, Siegal GP, King CR and
Curiel DT. (1994). Gene Ther., 1, 332–337.
Downward J, Yarden Y, Mayes E, Scrace G, Totty N,
Stockwell P, Ullrich A, Schlessinger J and Waterfield MD.
(1984). Nature, 307, 521–527.
Duan L, Bagasra O, Laughlin MA, Oakes JW and
Pomerantz RJ. (1994). Proc. Natl. Acad. Sci. USA, 91,
5075–5079.
Gargano N and Cattaneo A. (1997). J. Gen. Virol., 78, 2591–
2599.
Goncalves J, Silva F, Freitas-Vieira A, Santa-Marta M, Malho
R, Yang X, Gabuzda D and Barbas III C (2002). J. Biol.
Chem., 277, 32036–32045.
Graus-Porta D, Beerli RR, Daly JM and Hynes NE. (1997).
EMBO J., 16, 1647–1655.
Gschwind A, Zwick E, Prenzel N, Leserer M and Ullrich A.
(2001). Oncogene, 20, 1594–1600.
Gullick WJ. (1991). Br. Med. Bull., 47, 87–98.
Gullick WJ, Downward J and Waterfield MD (1985). EMBO
J., 4, 2869–2877.
Oncogene
Hills D, Rowlinson-Busza G and Gullick WJ. (1995). Int. J.
Cancer, 63, 537–543.
Holliger P, Prospero T and Winter G. (1993). Proc. Natl.
Acad. Sci. USA, 90, 6444–6448.
Jannot CB, Beerli RR, Mason S, Gullick WJ and Hynes NE.
(1996). Oncogene, 13, 275–282.
Jermutus L, Honegger A, Schwesinger F, Hanes J and
Plückthun A. (2001). Proc. Natl. Acad. Sci. USA, 98,
75–80.
Kabat EA, Wu TT, Perry HM, Gottesman KS and Foeller C.
(1991). Sequences of Proteins of Immunological Interest, 5
edn. National Institutes of Health: Bethesda.
Karunagaran D, Tzahar E, Beerli RR, Chen X, Graus-Porta
D, Ratzkin BJ, Seger R, Hynes NE and Yarden Y. (1996).
EMBO J., 15, 254–264.
Kile BT, Schulman BA, Alexander WS, Nicola NA, Martin
HM and Hilton DJ. (2002). Trends Biochem. Sci., 27, 235–
241.
King CR, Kraus MH and Aaronson SA. (1985). Science, 229,
974–976.
Klapper LN, Kirschbaum MH, Sela M and Yarden Y. (2000).
Adv. Cancer Res., 77, 25–79.
Lener M, Horn IR, Cardinale A, Messina S, Nielsen UB,
Rybak SM, Hoogenboom HR, Cattaneo A and Biocca S.
(2000). Eur. J. Biochem., 267, 1196–1205.
Marasco WA. (1995). Immunotechnology, 1, 1–19.
Marasco WA. (1997). Gene Ther., 4, 11–15.
Martineau P, Jones P and Winter G. (1998). J. Mol. Biol., 280,
117–127.
Mendelsohn J. (1997). Clin. Cancer Res., 3, 2703–2707.
Mendelsohn J and Baselga J. (2000). Oncogene, 19, 6550–6565.
Mhashilkar AM, Bagley J, Chen SY, Szilvay AM, Helland DG
and Marasco WA. (1995). EMBO J., 14, 1542–1551.
Ohage EC, Wirtz P, Barnikow J and Steipe B. (1999). J. Mol.
Biol., 291, 1129–1134.
Olayioye MA, Neve RM, Lane HA and Hynes NE. (2000).
EMBO J., 19, 3159–3167.
Parisien JP, Lau JF, Rodriguez JJ, Sullivan BM, Moscona A,
Parks GD, Lamb RA and Horvath CM. (2001). Virology,
283, 230–239.
Persic L, Righi M, Roberts A, Hoogenboom HR, Cattaneo A
and Bradbury A. (1997). Gene, 187, 1–8.
Proba K, Honegger A and Plückthun A. (1997). J. Mol. Biol.,
265, 161–172.
Rader C and Barbas III CF (1997). Curr. Opin. Biotechnol., 8,
503–508.
Russo MW, Lukas TJ, Cohen S and Staros JV. (1985). J. Biol.
Chem., 260, 5205–5208.
Shin DM, Donato NJ, Perez-Soler R, Shin HJ, Wu JY, Zhang
P, Lawhorn K, Khuri FR, Glisson BS, Myers J, Clayman G,
Pfister D, Falcey J, Waksal H, Mendelsohn J and Hong WK.
(2001). Clin. Cancer Res., 7, 1204–1213.
Steinberger P, Andris-Widhopf J, Buhler B, Torbett BE and
Barbas III CF (2000). Proc. Natl. Acad. Sci. USA, 97,
805–810.
Tse E, Lobato MN, Forster A, Tanaka T, Chung GT and
Rabbitts TH. (2002). J. Mol. Biol., 317, 85–94.
Visintin M, Tse E, Axelson H, Rabbitts TH and Cattaneo A.
(1999). Proc. Natl. Acad. Sci. USA, 96, 11723–11728.
Intracellular antibodies specific for human EGF receptor
S Hyland et al
1567
Wörn A, Auf der Maur A, Escher D, Honegger A, Barberis A
and Plückthun A. (2000). J. Biol. Chem., 275, 2795–2803.
Wörn A and Plückthun A. (1998). FEBS Lett., 427, 357–361.
Wörn A and Plückthun A. (2001). J. Mol. Biol., 305, 989–1010.
Xu JL and Davis MM. (2000). Immunity, 13, 37–45.
Yarden Y and Sliwkowski MX. (2001). Nat. Rev. Mol. Cell
Biol., 2, 127–137.
Yi KS, Chung JH, Lee YH, Chung HG, Kim IJ, Suh BC, Kim
E, Cocco L, Ryu SH and Suh PG. (2001). Oncogene, 20,
7954–7964.
Zhu Q, Zeng C, Huhalov A, Yao J, Turi TG, Danley D, Hynes
T, Cong Y, DiMattia D, Kennedy S, Daumy G, Schaeffer E,
Marasco WA and Huston JS. (1999). J. Immunol. Methods,
231, 207–222.
Oncogene