Structural basis of binding by cyclic nonphosphorylated peptide antagonists of Grb7 implicated in breast cancer progression

Article abstract of DOI:10.1016/j.jmb.2011.07.030

Growth-receptor-bound protein (Grb)7 is an adapter protein aberrantly overexpressed, along with the erbB-2 receptor in breast cancer and in other cancers. Normally recruited to focal adhesions with a role in cell migration, it is associated with erbB-2 in cancer cells and is found to exacerbate cancer progression via stimulation of cell migration and proliferation. The G7-18NATE peptide (sequence: WFEGYDNTFPC cyclized via a thioether bond) is a nonphosphorylated peptide that was developed for the specific inhibition of Grb7 by blocking its SH2 domain. Cell-permeable versions of G7-18NATE are effective in the reduction of migration and proliferation in Grb7-overexpressing cells. It thus represents a promising starting point for the development of a therapeutic against Grb7. Here, we report the crystal structure of the G7-18NATE peptide in complex with the Grb7-SH2 domain, revealing the structural basis for its interaction. We also report further rounds of phage display that have identified G7-18NATE analogues with micromolar affinity for Grb7-SH2. These peptides retained amino acids F2, G4, and F9, as well as the YDN motif that the structural biology study showed to be the main residues in contact with the Grb7-SH2 domain. Isothermal titration calorimetry measurements reveal similar and better binding affinity of these peptides compared with G7-18NATE. Together, this study facilitates the optimization of second-generation inhibitors of Grb7.

Full text of DOI:10.1016/j.jmb.2011.07.030

doi:10.1016/j.jmb.2011.07.030
J. Mol. Biol. (2011) 412, 397–411
Contents lists available at www.sciencedirect.com
Journal of Molecular Biology
journal homepage: http://ees.elsevier.com.jmb
Structural Basis of Binding by Cyclic
Nonphosphorylated Peptide Antagonists of Grb7
Implicated in Breast Cancer Progression
Nigus D. Ambaye1†, Stephanie C. Pero2†, Menachem J. Gunzburg1,
MinYin Yap1, Daniel J. Clayton1, Mark P. Del Borgo1,
Patrick Perlmutter3, Marie-Isabel Aguilar1, Girja S. Shukla2,
Elena Peletskaya2, Michelle M. Cookson2, David N. Krag2,
Matthew C. J. Wilce1 ‡ and Jacqueline A. Wilce1 ‡
⁎
⁎
1Department of Biochemistry and Molecular Biology, Monash University, Wellington Road, VIC 3800, Australia
2Department of Surgery and Vermont Cancer Center, University of Vermont, Burlington, VT 05405, USA
3School of Chemistry, Monash University, Wellington Road, VIC 3800, Australia
Received 8 June 2011;
received in revised form
13 July 2011;
accepted 14 July 2011
Available online
23 July 2011
Growth-receptor-bound protein (Grb)7 is an adapter protein aberrantly
overexpressed, along with the erbB-2 receptor in breast cancer and in other
cancers. Normally recruited to focal adhesions with a role in cell migration, it
is associated with erbB-2 in cancer cells and is found to exacerbate cancer
progression via stimulation of cell migration and proliferation. The G7-
18NATE peptide (sequence: WFEGYDNTFPC cyclized via a thioether bond)
is a nonphosphorylated peptide that was developed for the specific
inhibition of Grb7 by blocking its SH2 domain. Cell-permeable versions of
G7-18NATE are effective in the reduction of migration and proliferation in
Grb7-overexpressing cells. It thus represents a promising starting point for
the development of a therapeutic against Grb7. Here, we report the crystal
structure of the G7-18NATE peptide in complex with the Grb7-SH2 domain,
revealing the structural basis for its interaction. We also report further
rounds of phage display that have identified G7-18NATE analogues with
micromolar affinity for Grb7-SH2. These peptides retained amino acids F2,
G4, and F9, as well as the YDN motif that the structural biology study
showed to be the main residues in contact with the Grb7-SH2 domain.
Isothermal titration calorimetry measurements reveal similar and better
binding affinity of these peptides compared with G7-18NATE. Together, this
study facilitates the optimization of second-generation inhibitors of Grb7.
© 2011 Elsevier Ltd. All rights reserved.
Edited by I. Wilson
Keywords:
Grb7 inhibitor;
SH2 domain;
cyclic peptide;
ITC;
phage display
Introduction
*Corresponding authors. E-mail addresses:
matthew.wilce@monash.edu; jackie.wilce@monash.edu.
† N.D.A. and S.C.P. are joint first authors.
‡ M.C.J.W. and J.A.W. are joint senior authors.
Abbreviations used: Grb, growth-receptor-bound
protein; FAK, focal adhesion kinase; PDB, Protein Data
Bank; ITC, isothermal titration calorimetry; HRP,
horseradish peroxidase.
Growth-receptor-bound protein (Grb)7 belongs to
a family of adapter proteins that interact with
upstream phosphorylated tyrosine kinases in the
propagation of signal transduction pathways.
Under normal cellular conditions, Grb7 plays a
role in integrin-mediated cell migration via direct
interaction with focal adhesion kinase (FAK) at focal
0022-2836/$ - see front matter © 2011 Elsevier Ltd. All rights reserved.

398
Grb7-SH2 Binding by Peptide Antagonists
contacts (reviewed in Refs. 1 and 2). This is
consistent with its evolutionary relationship to the
Mig10 cell migration protein of Caenorhabditis
elegans.3 Grb7 has become of particular interest,
however, in cancer cells where it is massively
overexpressed and interacts with the well-known
propagator of cancerous cell growth, the erbB-2
(also known as Neu/HER-2/EGFR-2) receptor.4
Grb7 is tightly coamplified with erbB-2 in approx-
imately 20–25% of breast cancer cell lines, and there
is also a strong correlation between erbB-2 and Grb7
overexpression in esophageal, gastric, and ovarian
cancers.5–7 Grb7 overexpression has been noted in
32% of Barrett's adenocarcinomas8 and has been
correlated with poor prognosis for hepatocellular
carcinoma.9 Most recently, a study of high Grb7
expression in 638 primary breast cancer specimens
was found to correlate with decreased survival of
the patients.10
There is now compelling evidence that the over-
expression of Grb7 directly impacts the proliferative
and metastatic potential of these cancers (and is not
just a benign result of amplification that occurs due
to the close proximity of the Grb7 gene to the erbB-2
gene on the 17q12 amplicon). In recent studies
correlating Grb7 overexpression with clinical out-
come, Grb7 protein overexpression has been shown
to be an adverse prognostic factor in human breast
cancer independent of erbB-2.10,11 In studies
designed to directly determine the role of Grb7, it
has been shown that the removal of Grb7 from
breast cancer cells (using RNAi) reduces breast
cancer cell viability and increases the activity of the
anti-erbB-2 cancer therapeutic lapatinib.12,13 Grb7
overexpression in breast cancer cells has been shown
to facilitate the phosphorylation of erbB-2 and AKT
and, in mouse xenograft studies, promote the
xenograft growth.14 It thus appears that Grb7
upregulation acts in concert with erbB-2 to promote
cancer cell proliferation. With respect to its role in
metastasis, overexpression of Grb7 has been shown
to enhance cell migration, whereas inhibition of
Grb7 (by overexpression of the Grb7 SH2 domain
acting as a dominant-negative inhibitor) results in
the inhibition of cell migration.15,16 More recently,
knockdown of Grb7 expression in hepatocellular
carcinoma cells resulted in their lower invasive
potential in vitro and the delayed onset of tumors in
nude mice.9 In another study, the knockdown of
Grb7 using RNAi showed a strong antimigratory
effect in a pancreatic cell line.17 Grb7 is thus now
recognized to play a significant role in cancer
progression and to be an important new therapeutic
target.10,18
domain.2,19 It interacts with the phosphorylated
tyrosine of activated upstream partners via its SH2
domain and is phosphorylated at specific tyrosine
residues to propagate downstream events.20,21
While a large number of binding partners function-
ing upstream of Grb7 have been identified,22 the
precise downstream events leading to cell prolifer-
ation and migration are not fully elucidated. The
most recent data suggest that the proliferative
activity of Grb7-overexpressing cells is mediated
via the recruitment of RasGTPases and subsequent
phosphorylation of extracellular signal-regulated
kinase 1/extracellular signal-regulated kinase 2
leading to cell proliferation.23 The way in which
the Grb7/FAK interaction impacts migration in
human cancer cells is not yet determined, but there
is evidence to suggest that the phosphorylation of
Grb7 by FAK impacts at the level of translational
regulation.21,24
The Grb7-SH2 domain has thus been targeted in
the development of a specific peptide inhibitor of
Grb7. The G7-18NATE peptide (sequence: WFE-
GYDNTFPC cyclized via a thioether bond) was
discovered using phage display and was found to be
a specific inhibitor of Grb7, binding to the Grb7-SH2
domain with micromolar affinity, but it does not
bind to the SH2 domains of Grb14 and Grb10, which
are closely related family members, or to the Grb2-
SH2 domain, which is the downstream partner of
the erbB-2 receptor in normal cells.25–27 A cell-
permeable version of G7-18NATE (synthesized with
a Penetratin© or Tat-based C-terminal peptide
extension) was shown to enhance the antiprolifera-
tive effect of Herceptin in breast cancer cells,
successfully validating the principle of Grb7 inhibi-
tion through the blocking of its SH2 domain.28 In
another study, cell-permeable G7-18NATE was
shown to prevent the Grb7 interaction with FAK
and dramatically reduce cell migration in a pancre-
atic cell line.17 Together, these investigations dem-
onstrate that G7-18NATE not only is a valuable tool
for the inhibition of Grb7 in cells but also represents
proof of concept that Grb7 may be blocked via its
SH2 domain and that this may be to therapeutic
advantage.
While G7-18NATE is effective as a Grb7 inhibitor,
a pressing goal is to develop analogues with
enhanced affinity. This would allow cellular exper-
iments to be conducted at lower inhibitor concen-
trations and be an essential criterion for therapeutic
use. What is not yet known is why the G7-18NATE
peptide sequence has affinity and specificity for the
Grb7-SH2 domain. Here, we report a breakthrough
in the understanding of the structural basis for the
G7-18NATE/Grb7-SH2 domain interaction
through structural studies and through further
peptide library screening. We report the crystal
structure of G7-18NATE bound to the Grb7-SH2
domain, revealing the conformation adopted by
Grb7 is a 532-residue protein composed of an N-
terminal proline-rich domain, followed by an RA
(Ras-associating) domain, a central PH (pleckstrin
homology) domain, a BPS (between PH domain and
SH2 domain) domain, and a C-terminal SH2

Grb7-SH2 Binding by Peptide Antagonists
399
G7-18NATE and the critical contacts that underlie
the basis for the interaction. We also report the
outcome of further rounds of phage display that
have provided second-generation peptides based
upon G7-18NATE. These peptides are characterized
and analyzed in light of the structural information
for the further development of potent and specific
Grb7 inhibitors.
Grb7-SH2 domain structure previously reported
[Protein Data Bank (PDB) ID: 2QMS] (across 99
atoms) (Fig. 1d).
G7-18NATE binding interactions with the
Grb7-SH2 domain
The main objective of this study was to determine
the structural basis for the G7-18NATE interaction
with the Grb7-SH2 domain. For clarity, amino acid
residues of G7-18NATE will be referred to using
single-letter code, while those of the Grb7-SH2
domain will be referred to using three-letter code.
The whole of the G7-18NATE peptide was clearly
visible in the electron density map (Fig. 2a). The
structure reveals the G7-18NATE arranged in a cycle
with a pseudo-β-turn formed between the carbonyl
of the N-terminal acetamide (to which the thioether
is connected) and the amine of F9 (Fig. 2b). This turn
is also stabilized by the trans conformation of P10
and the formation of a hydrogen bond between the
backbone amine of C11 and the backbone carbonyl
of F9. A β-turn is formed between F2 and Y5, with a
hydrogen bond between the F2 carbonyl and the Y5
amine. No side-chain-specific intramolecular inter-
actions appear to influence the conformation of the
G7-18NATE. Rather, the side chains are arranged to
either form contacts with the surface of the Grb7-
SH2 domain, including hydrogen-bond, electrostatic,
and van der Waals interactions, or extend into
solvent.
Results
Structure of the G7-18NATE/Grb7-SH2 complex
In order to investigate the basis for the molecular
interaction between the G7-18NATE peptide and the
Grb7-SH2 domain, we solved the structure of the
complex using X-ray crystallography to 2.4 Å
resolution. The final G7-18NATE/Grb7-SH2 com-
plex structure consists of four chains of the Grb7-
SH2 domain and two G7-18NATE molecules in the
asymmetric unit. This unexpected stoichiometry is
most likely due to the crystal packing arrangement.
The G7-18NATE binds at the expected site on the
surface of the Grb7-SH2 molecule, but its presence
occludes the binding site on an adjacent Grb7-SH2
domain. The Grb7-SH2 also formed dimers as
previously observed.26 The crystal is thus composed
of Grb7-SH2 dimers in which one monomer of each
dimer is peptide bound (chains A and B) and the
other is apo-Grb7-SH2 (chains C and D).
The YDN residues of G7-18NATE that comprise
the recognition motif of the Grb7-SH2 domain are
positioned analogously to the pYXN mode of
binding to the Grb2-SH2 domain,30 which also
possesses specificity for a YXN motif. Interestingly,
this occurs despite the lack of a phosphate group
(Fig. 2c). The tyrosine (Y5) side chain projects into
the phosphotyrosine binding pocket with its aro-
matic ring positioned planar to the surface of the
side chain of the critical Grb7 βD6 residue Leu481.
The aspartic acid (D6) residue extends in the
opposite direction and forms hydrogen bonds via
its side-chain carbonyl and backbone amine with the
Grb7 His479 backbone amine and carbonyl, respec-
tively. Asparagine (N7) is extended back towards
Leu481 where it forms bipartite hydrogen bonds
with the backbone amide and carbonyl groups of
Leu481. The G7-18NATE, as mentioned above, is
nonphosphorylated. Grb7 residues that would be
predicted to form electrostatic interactions with a
phosphate include Arg438 and Arg458. In the
current structure, the side chain of Arg438 extends
to hydrogen bond with the carbonyl of glycine (G4).
Strikingly, the aromatic groups of the phenylala-
nines (F2 and F9) are positioned planar to the surface
of the Grb7-SH2 domain contributing to binding via
van der Waals contacts (Fig. 2d). F9 is positioned
over the surface of the EF loop. In particular, this
The Grb7-SH2 domain adopts an SH2 domain
fold, with an αβαββββα configuration as seen also
for its apo-form crystallized under different
conditions26 (Fig. 1a and b). Note that the standard
nomenclature for SH2 domain secondary structures
devised by Eck et al. is adopted throughout this
report.29 A three-stranded antiparallel β-sheet com-
posed of βB, βC, and βD is flanked by the two
helices αA and αB. The βD strand is extended and
also forms a short region of β-sheet with βE. The
intervening loops are named after their adjacent
secondary structural elements. Short regions of β-
sheet (βA, βF, and βG) observed in other SH2
domains are not present in the Grb7-SH2 domain,
and an extra N-terminal helix is observed in this
Grb7-SH2 construct, termed H1.
The G7-18NATE peptide is bound across the βD
strand and makes contact with the BC, EF, and BG
loops (Fig. 1c). The specific Grb7 amino acids that
are involved in interactions are indicated with dots
above the Grb7 sequence shown in Fig. 1a and
further discussed below. The presence of G7-
18NATE does not make significant impact on the
structure of the protein, with the Cα RMSD between
G7-18NATE-bound and apo-forms of the Grb7-SH2
domain being 0.54 Å for the cocrystallized Grb7-
SH2 domain (across 96 atoms) and 0.71 Å for the

(a)
CD loop
AB loop
(b)
(c)
(d)
βC
BG loop
αB
βD
αA
βB
EF
loop
βA
AA
loop
βD’
H1
βE
DE loop
BC loop
Fig. 1 (legend on next page)

Grb7-SH2 Binding by Peptide Antagonists
401
surface is formed by the planar arrangement of the
backbone carbons and Cβ of Asp496. F2 is posi-
tioned over the surface of the Leu481 side chain
adjacent to the surface that Y5 is positioned over.
These residues contribute to a total surface area of
interaction of 458 Å2.
(where X is any amino acid), referred to as the YDN
and YDN-NA biased libraries, respectively. After
three rounds of panning, we observed strong
enrichment of 163-fold in the YDN-NA library and
12-fold in the YDN library. One hundred unique
clones from the eluate collected from the YDN-NA
library were screened for their binding to the SH2
domain of Grb7. There were 24 unique clones
identified from 189 clones sequenced from the
YDN and YDN-NA libraries. From this pool of
unique clones, four clones (5073, 5200, 5242, and
5367) displayed improved binding to the SH2
domain of Grb7. Interestingly, the four strongest
binding clones share a similar amino acid sequence
with each other and with our original G7-18 clone
(Table 1). The motif CX1FX2GYDNX3(X4 =F/Y/W)
(X5 =L/V)C was conserved in these peptides. This
set of data showed that F at position 3 and G at
position 5 are important. There was no preferential
amino acid for the X1 position; however, at the
following positions, preference was shown for the
following amino acids: X2 =W; X3 =E; X4 =F, Y, or
W; and X5 =L or V. The homology in all the top
binding clones occurred in the cyclic ring and not in
the four amino acids flanking the cysteines.
The remaining amino acid residues of G7-
18NATE appear to play no role in forming in-
teractions with the Grb7-SH2 domains at all. W1
projects away from the surface of interaction into
solvent. Likewise, E3 projects away from the
binding site, although in this case, it is not projecting
into solvent. Rather, the E3 side chain forms
hydrogen bonds with Arg458 in the phosphotyr-
osine binding site of an adjacent holo-Grb7-SH2
molecule. T8 also projects outward from the binding
site, not forming any contacts with the Grb7-SH2
protein. The proline (P10) does not make any
specific contribution to binding, although it may
act to stabilize the turn conformation of the G7-
18NATE. Lastly, cysteine appears to play no role in
binding to the Grb7-SH2 domain except to provide
the thiol for the cyclization of the peptide that is
essential for binding.25
Phage display evolution of second-generation
peptides
One more round of maturation was carried out to
further confirm all positions. The oligonucleotides
for this library were designed as follows: CX1FX2-
GYDNX3(X4 =F/Y/W) (X5 =L/V)C, where posi-
tions X1 and X3 were random; position X2
expressed 50% Trp and 50% random amino acids;
position X4 expressed equimolar amounts of F, Y,
and W; and position X5 expressed equimolar
amounts of L and V. This library was created with
excellent diversity, having greater than 10 billion
clones. Following three rounds of panning of this
biased library on Grb7 protein, 96 clones were
randomly picked for clone-binding analysis using
phage ELISA.25 In addition, the remaining eluate
was subjected to further screening using a colony lift
assay31 to the biotinylated Grb7. Colony lift assays
are a way to screen large numbers of clones to
identify rare high-affinity binding phage present in
an excess of low-affinity clones. After the initial
screening, the clones with the strongest signal were
chosen for secondary screening in a colony lift assay,
followed by sequencing and individual clone-
binding analysis using phage ELISA. The binding
In parallel with establishing the determinants of
binding of G7-18NATE to the Grb7-SH2 domain
using a structural method, we also sought to
investigate the sequence requirements of the peptide
using combinatorial peptide library screening. Af-
finity maturation of G7-18NATE was performed
using a biased phage library approach with two
additional rounds of library construction. The first-
round libraries were designed with conserved YDN
amino acids and the cyclic structure of Grb7 peptide
retained using cysteines for disulfide bond cycliza-
tion. The YDN positions were identified from the
original biased library screening where G7-18 was
discovered.25 Two libraries containing the YDN
motif were created, one maintaining the four amino
acids flanking each cysteine and the other with the
flanking amino acids removed since it was not clear
whether these amino acids were important for the
binding. The cysteine-constrained libraries were
designed with X4CX3YDNX4CX4 and CX3YDNX4C
Fig. 1. The Grb7-SH2 structure in complex with G7-18NATE. (a) The sequence of the Grb7-SH2 domain with the
secondary structural elements, as determined in the current structure, highlighted in rainbow colors. The archetypical
SH2 domain secondary structural elements are shown schematically above with the naming system according to Eck et al.29
Black dots above individual amino acids indicate residues that are involved in hydrogen-bond interactions with the
G7-18NATE, and blue dots indicate residues that form part of the buried surface area. The aligned sequences and
percentage identity of Grb14-SH2, Grb10-SH2, and Grb2-SH2 domains are shown for comparison. (b) A cartoon
representation of the Grb7-SH2 domain with coloring correlated with that shown in (a). (c) The structure of the G7-
18NATE peptide is shown in stick representation bound to the Grb7-SH2 domain. (d) Superposed cartoon structures
of Grb7-SH2 bound to G7-18NATE (gray), the cocrystallized apo-Grb7-SH2 domain (orange), and the apo-Grb7-SH2
domain previously structurally determined (purple; PDB ID: 2QMS).

(a)
(b)
T8
W1
F2
D6
N7
C11
P10
Y5
F9
E3
G4
P10
F9
T8
G4
E3
D6
Y5
N7
F2
C11
W1
E3
(c)
D6
(d)
W1
T8
N7
His479
Met495
Leu481
G4
Y5
Y5
Leu481
N7
F9
EF loop
F2
Arg438
Fig. 2. The G7-18NATE structure in complex with the Grb7-SH2 domain. (a) Stereo diagram showing the 2Fo −Fc electron density map contoured at 1.2 σ (blue
mesh) that defines the structure of G7-18NATE at the surface of the Grb7-SH2 domain (gray). (b) G7-18NATE oriented to show intramolecular hydrogen bonds. (c) G7-
18NATE amino acid residues from G4 to N7 (truncated ends of this segment are indicated with black circles) bound to the Grb7-SH2 domain. (d) The G7-18NATE
molecule showing the positioning of aromatic amino acid residues F2 and F9 that form interactions with the Grb7-SH2 domain. In all images, G7-18NATE is shown in
green stick representation, the Grb7-SH2 molecule is shown in gray cartoon representation with a transparent surface, and specific amino acids involved in interactions
with the bound peptide are displayed in gray stick representation.

Grb7-SH2 Binding by Peptide Antagonists
403
Table 1. The newly identified Grb7 binding clones identified
from the YDN and YDN-NA biased libraries have sequence
homology with each other and to our lead G7-18 peptide
only one amino acid change, that is, from L to V at
the X5 position. 08-5992 repeatedly had better
binding than 08-6110, indicating the value of V at
the X5 position. Clone 08-6028 had similar intensity
of binding with clone 08-5992 and also shares the
conserved motif with the exception at position X3.
The fourth clone of interest, 08-6046, which was
narrowly the third highest binder in the ELISA
assay, differed from the consensus sequence at
positions X2 and X3. The notable feature of these
peptides is the complete conservation of F2, G4, and
F9 residues that, with the YDN motif, are precisely
the amino acid residues that were found to form the
basis for the binding interaction in the structural
studies of the G7-18NATE/Grb7-SH2 complex.
These four peptides (08-5992, 08-6110, 08-6046,
and 08-6028) were synthesized as thioether-cyclized
peptides and analyzed for their binding affinity to
the Grb7-SH2 domain as described below. In
addition, peptides 08-5992 and 08-6028 were syn-
thesized as free peptides conjugated to the Tat
sequence and analyzed further to compare their
antiproliferative properties to G7-18NATE-Tat.
These newly identified peptides had the same level
of antiproliferative activity as G7-18NATE-Tat in a
BrdU chemiluminescent cell proliferation ELISA
(Roche Applied Science, Indianapolis, IN) of SK-
BR-3 breast cancer cells (data not shown).
Identical and chemically similar amino acids (L, V and F, Y) are
in boldface and underlined
of the individual clones were compared to G7-18
phage clone. Seventy-two of the strongest binding
clones identified from random picking and colony
lift analysis were grown and normalized using a
plate-based assay developed in our laboratory for
rapid quantification of phage for semi-high-
throughput analysis.32 Figure 3 represents the
sequences of the best binding clones as compared
to G7-18 identified from this panning series. This
affinity maturation identified a series of clones with
strong binding. Multiple clones with identical
sequences were also identified and are presented
in Table 2. The conserved motif of these clones is
CPFWGYDNEFLC, which is the exact sequence of
clone 08-6110. However, clone 08-5992 differs by
Binding affinity of second-generation peptides
Using isothermal titration calorimetry (ITC), we
and others have previously shown that G7-18NATE
binds to the Grb7-SH2 domain with micromolar
affinity.26,27,33,34 Using the same experimental con-
ditions as previously conducted in our laboratory to
arrive at a Kd =35 μM for G7-18NATE, we compared
the binding of second-generation peptides 08-5992,
08-6110, 08-6046, and 08-6028. Similar to G7-
18NATE, all peptides showed an exothermic heat
of reaction (Fig. 4), although these varied between
peptides as reflected in the different magnitudes of
heat loss measured. Analysis of the curves allowed
equilibrium dissociation constants (Kd) to be esti-
mated (Table 3). Kd values were all found to be in the
micromolar range (experiments were conducted in
duplicate). Of these four peptides, 08-6046 displayed
a slightly higher affinity than G7-18NATE with a
Kd =18 μM. The others were of similar or slightly
lower affinity. Since none of these peptides bound
with an affinity remarkably higher than that of the
parent G7-18NATE, this demonstrates that the
amino acids that vary between these peptides do
not significantly have an impact on the binding
interactions with the Grb7-SH2 domain. Conversely,
these data show that the sum total of interactions
underlying binding affinity is made by F2, G4, F9,
and the YDN motif, in agreement with the crystal
structure studies (Fig. 5).
5.E+05
Grb7
5.E+05
4.E+05
4.E+05
3.E+05
3.E+05
2.E+05
2.E+05
1.E+05
5.E+04
0.E+00
Blocker
Normalized Phage Clones
Fig. 3. Phage ELISA demonstrating the binding of the
newly identified phage clones to the purified Grb7-SH2
domain. Equivalent amounts of phage were evaluated to
determine their relative binding affinity to the SH2
domain of Grb7 as compared to our original G7-18
clone.25 Phage were detected with anti-M13-HRP. The
HRP was monitored with chemiluminescent substrate.

404
Grb7-SH2 Binding by Peptide Antagonists
Table 2. Sequence of clones identified from the CX1FX2GYDNX3X4X5 library
C
X1
F
X2
G
Y
D
N
X3
X4
X5
Library
design
No. of
clones
C
Random
F
50% W
G
Y
D
N
Random
33% FWY
50% L/V
C
08-5992a
08-6116
08-6117
08-6028a
08-6127
08-6124
08-6097
08-6086
08-5987
08-6046a
08-6112
08-6089
08-6110a
08-5990
08-5981
08-6126
08-6064
Consensus
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
P
M
Q
P
F
F
F
F
F
F
F
F
F
F
F
F
F
F
F
F
F
F
W
W
W
W
W
F
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
G
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
Y
D
D
D
D
D
D
D
D
D
D
D
D
D
D
D
D
D
D
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
N
E
Q
V
A
S
D
S
Q
A
Q
A
Q
E
F
F
F
F
F
F
F
W
F
F
F
F
F
F
F
F
F
F
V
L
L
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
9
7
L
L
L
V
L
L
L
L
L
L
L
L
L
L
L
L
P
L
K
A
P
L
V
P
Q
L
P
G
G
W
S
Y
W
W
W
W
A
5
6
D
E
W
S
E
R
P
W
W
Column 1 depicts the clone number, and column 2 depicts the number of clones with identical sequence. The amino acids that are
italicized are those that were constant in this library, while those that are not italicized are those enriched for in this round of maturation.
a
Peptides further tested using ITC.
reduce the loss of conformational entropy upon
binding. We have previously reported that the G7-
Discussion
The upregulation and overexpression of Grb7 in
many cancers have a direct impact on their prolifer-
ative and metastatic potential. There is thus signif-
icant current interest in the development of Grb7
inhibitors that could be used in combating these
cancers. The G7-18NATE peptide is the first peptide
that has been discovered that binds with specificity
to the Grb7-SH2 domain. Furthermore, cell-perme-
able forms of G7-18NATE have been shown to
inhibit Grb7 interactions with binding partners in
vivo and to reduce cancer cell migration and
proliferation. The G7-18NATE thus represents an
important starting point in the development of
potent inhibitors that could be used in medical
research and therapeutically. The current work has,
therefore, been carried out in order to understand the
structural basis for the G7-18NATE interaction with
the Grb7-SH2 domain and to potentially identify
peptides of enhanced affinity. While small-molecule
inhibitor development may also represent a path to
targeting Grb7, no small-molecule inhibitor with
specificity or high affinity has been discovered to
date, although we are also pursuing this approach.35
The crystal structure of the G7-18NATE/Grb7-
SH2 domain complex reported here reveals the
structural basis for their interaction. It is remarkable
that the G7-18NATE peptide binds to the Grb7-SH2
domain with micromolar affinity despite being
nonphosphorylated. The binding interaction is, in
part, assisted by the peptide being cyclized to
constrain the peptide. It has been shown that the
linear form of the peptide does not bind with
detectable affinity.25 Cyclization is presumed to
18NATE peptide has a propensity to adopt a turn
about the YDN motif,36 and the current Grb7-SH2/
G7-18NATE structure confirms that, indeed, the G7-
18NATE peptide adopts the anticipated turn when
bound to its target. The YDN motif is positioned
precisely where analogous YXN motifs are posi-
tioned, bound to both Grb2-SH2 and Grb7-SH2
domains,30,37 forming van der Waals and hydrogen-
bond interactions about the βD6 residue Leu481, as
was predicted in our modeling studies.26 At the
phosphotyrosine binding site of the Grb7-SH2
domain, arginine residues (Arg438 and Arg458),
which would be expected to form hydrogen-bond
interactions with the phosphate of the phosphotyr-
osine of a physiological substrate, are engaged in
interactions with the G4 backbone carbonyl and, in
this crystal form, with the E3 glutamic acid side
chain from an adjacent G7-18NATE/Grb7-SH2
domain complex. This explains the selection of a
glycine, since it uniquely possesses the conforma-
tional flexibility required for the interaction. It raises
the question, however, of whether the E3 side chain
would be positioned differently in solution, possibly
even playing a role in G7-18NATE binding. This
answer is provided through the phage display work,
discussed further below, in which a glutamic acid
was not selected from the peptide libraries. It is
therefore unlikely to play a role in the primary G7-
18NATE interaction with the Grb7-SH2 domain but
may have assisted with the crystal formation. An
important additional contribution to G7-18NATE
binding arises from the positioning of two phenyl-
alanine residues (F2 and F9) that contribute to van

Fig. 4. Binding of second-generation G7-18NATE peptides to the Grb7-SH2 domain. ITC thermograms obtained from the isothermal titration of G7-18NATE
analogues (08-5992, 08-6110, 08-6046, and 08-6028) against the Grb7-SH2 domain. The upper panels show raw data obtained from 10-μl injections of peptides at 25 °C.
The lower panels display plots of integrated total energy exchanged (as kilocalories per mole of injected peptides) as a function of molar ratio of the peptides to the
Grb7-SH2 domain.

406
Grb7-SH2 Binding by Peptide Antagonists
Table 3. Thermodynamic binding parameters of G7-18NATE analogues
Binding parameter
08-5992
08-6110
08-6046
08-6028
N
1.2 0.18
1.73 0.23
−1.92 0.37
−3.83
−5.75 0.37
57.8
1.13 0.09
2.26 0.47
−2.89 0.33
−3.05
−5.94 0.33
44.25
1.31 0.08
5.53 0.14
−1.64 0.06
−5.83
−6.48 0.06
18.1
1.05 0.162
1.91 0.40
−1.60 0.33
−4.23
−5.84 0.33
52.36
K (104 M−1
)
ΔH (kcal mol−1
)
−TΔS (kcal mol−1
)
ΔG (kcal mol−1
Kd (μM)
)
der Waals interactions with the Grb7-SH2 domain,
extending the total surface area of interaction to
458 Å2.
It is of interest that, in this crystal form, the G7-
18NATE is bound to the Grb7-SH2 domain in a 1:2
stoichiometric ratio, with only one SH2 domain of
every Grb7-SH2 domain dimer bound by peptide.
This is most likely due to the packing arrangement
of this crystal form in which every second G7-
18NATE binding site is occluded. Previous solu-
tion studies using analytical ultracentrifugation
and ITC have shown that a Grb7-SH2 domain
dimer forms in solution and interacts with G7-
18NATE in a 1:1 stoichiometric ratio—that is, with
a G7-18NATE peptide bound to every available
SH2 domain.26,33,34
The interaction formed between the G7-18NATE
peptide and the Grb7-SH2 domain bears some
resemblance to other ligand SH2 domain interac-
tions and possesses some unique features. Besides
the YXN motif interactions, the positioning of an
aromatic residue against the βD6 side chain (as seen
for F2 in the current structure) has similarly been
observed for ligands bound to the Grb2-SH2
domain.38,39 The G7-18NATE interaction with the
Grb7-SH2 domain is, however, unlike that reported
for the one other ligand-bound structure of the
Grb7-SH2 domain. The structure of a 10-residue
phosphorylated peptide, representing the pY1339
target site of the erbB-2 receptor, bound to the Grb7-
SH2 domain has been solved using NMR
spectroscopy37 (PDB IDs: 1MW4 and 2L4K). In this
structure, although the YXN motif is positioned
similarly, the rest of the peptide is positioned at a
different surface interface from that adopted by G7-
18NATE. No residues are positioned similarly to the
G7-18NATE F2 (against the βD6 side chain) or F9
(over the EF loop). In fact, the structures of the Grb7-
SH2 domain DE and EF loops are dramatically
repositioned by the pY1339 ligand interaction for
reasons that are, as yet, unclear.
The structure provides insight into the basis of
specificity of G7-18NATE for the Grb7-SH2 domain
over SH2 domains of Grb14, Grb10, and Grb2. The
interaction with Leu481 is understood to form the
basis for specificity of peptide binding to the Grb7-
SH2 domain, with a mutation made to Leu481 being
enough to completely abrogate binding or the
addition of Leu481 to the Grb14-SH2 domain
being enough to confer binding.40 While the
equivalent experiment has not been carried out for
G7-18NATE, the central importance of Leu481 is
clearly seen for G7-18NATE binding to the Grb7-
SH2 domain. Not only is the Y5 side chain
positioned over the Leu481 side chain and the N7
side chain hydrogen bonded to the Leu481 backbone
groups, but the F2 side chain is also positioned
against the Leu481 side chain. Any changes to this
residue would be expected to disrupt these opti-
mized interactions. Other interactions made by G7-
18NATE are with amino acid residues that are
conserved across Grb14, Grb10, and Grb2 or with
amino acid backbone functionalities, and it is thus
unlikely that these interactions underlie specificity.
Two phage display libraries were employed to
generate second-generation peptides with affinity
for the Grb7-SH2 domain. These maintained the
YDN motif and established its importance for G7-
18NATE binding. Enriched peptides were screened
for Grb7-SH2 domain binding using an ELISA-
based assay, from which four new peptides that
bind to the Grb7-SH2 domain were discovered. This
established the importance of F5 and G4 before the
YDN motif and an aromatic and hydrophobic
residue after the YDN motif. In a further maturation
step that maintained this consensus pattern, a
HN CO CH2
S
Fig. 5. Sequences of second-generation G7-18NATE
peptides. Shown are the sequences of four second-
generation peptides that were developed using phage
display and further tested as thioether-cyclized peptides
using ITC. Highlighted in red are the amino acid residues
that were conserved among the Grb7-SH2 domain
binding sequences. These residues correspond exactly to
the residues shown to be structurally important for
binding.

Grb7-SH2 Binding by Peptide Antagonists
407
library of over 10 billion clones was screened for an
optimized binder. Multiple clones were identified as
strong binders relative to the G7-18NATE sequence
in the ELISA-based assay. A notable feature of these
peptides is the almost complete conservation of F2,
G4, and F9 residues that, with the YDN motif, are
precisely the amino acid residues that the G7-
18NATE/Grb7-SH2 structure revealed to form the
basis for the binding interaction.
Four of the second-generation peptides were
synthesized as thioether-cyclized peptides by anal-
ogy to the G7-18NATE peptide. These all displayed
similar micromolar binding for the Grb7-SH2
domain. In all cases, their equilibrium dissociation
constants (Kd values) were in the micromolar range
(18–52 μM), similar to that found for G7-18NATE
(35 μM). No second-generation peptide displayed a
significantly enhanced affinity for the Grb7-SH2
domain despite the apparent higher affinity of the
ELISA-screened phage-displayed peptides. In view
of the structural data, this is not unexpected. The
differences between the second-generation peptides
and G7-18NATE occur exclusively at the amino acid
residue positions that, as the structural data show,
do not interact with the Grb7-SH2 domain. The
peptides are completely conserved at the amino acid
residues that form the basis for the interaction.
Together, the data suggest that this class of cyclic
peptide, with the consensus motif XFXGYDNXFXX,
has been optimized for interaction with the Grb7-
SH2 domain.
Materials and Methods
Grb7-SH2 domain preparation
Expression and purification of the SH2 domain of Grb7
(residues 415–532) were conducted using established
protocols.26,41 In brief, the SH2 domain was overexpressed
as a glutathione S-transferase fusion protein in the
Escherichia coli strain BL21(DE3)pLysS. After affinity
purification using glutathione resin, the glutathione S-
transferase was cleaved using thrombin and the protein
was further purified using cation- and size-exclusion
chromatography. Purity was determined using SDS-
PAGE, and the protein was concentrated to 10 mg/ml as
determined spectrophotometrically at 280 nm using an
extinction coefficient of 8480 M−1. The final purified
construct comprised residues 415–532 of Grb7 preceded
by the amino acid residues GS.
Peptide synthesis
The G7-18NATE and analogues were synthesized as
peptide amides using the Fmoc-based solid-phase peptide
synthesis strategy on a rink amide resin (1.0 mmol/g).
Details are as reported previously for the synthesis of G7-
18NATE.26,34 The following peptide sequences were
synthesized: WFEGYDNTFPC (G7-18NATE), PFWGYD-
NEFVC (08-5992), PFWGYDNEFVC (08-6028),
PFSGYDNQFLC (08-6110), and PFWGYDNEFLC (08-
6046). Following incorporation of the residues using the
microwave synthesizer (CEM Liberty System; CEM,
Matthews, NC) and removal of the final Fmoc group, N-
terminal chloroacetylation was carried out using 171 mg of
chloroacetic anhydride, 2 ml of dimethylformamide, and
100 μl of diisopropylethylamine. The reaction was carried
out manually at room temperature for 30 min. The dried
peptide was then cleaved and fully deprotected using a
cocktail of 94.5% trifluoroacetic acid/2.5% triisopropylsi-
lane/2.5% H2O/0.5% ethane dithiothreitol in a 20-fold
excess volume over dry weight of resin peptide for 3 to 4 h
at room temperature. After workup, thioether formation
was effected by dissolving the lyophilized peptide at
2 mg/ml in 50 mM NH4HCO3 in 50% acetonitrile/H2O,
pH 8.0, for 1.5 h at room temperature. The success of
cyclization was followed by mass spectrometry. The
cyclized peptides were purified using preparative re-
versed-phase HPLC, and their successful preparation
was confirmed using electrospray mass spectrometry in
negative ion mode (average molecular masses are listed)
(G7-18NATE: Mexpected=1417.6, Mobserved =1417.1; 08-5992:
Conclusion
This study has revealed the basis for interaction of
the G7-18NATE peptide with its Grb7-SH2 domain
target, which will greatly assist in the development
of inhibitors against Grb7 in cancer. The crystal
structure of the Grb7-SH2/G71-18NATE complex
has revealed the mode of interaction and the identity
of the key functionalities that interact with the
target. Peptide sequences derived from phage
display methods contained these key residues.
They displayed similar and better binding affinity
to the target but did not represent a significantly
improved peptide for Grb7-SH2 inhibition. Togeth-
er, this suggests that the key amino acids that form
the basis for the interaction have been optimized for
this class of cyclic-peptide binding to the Grb7-SH2
domain. It can be concluded that further improve-
ments to the affinity of interaction will not be made
by substitution of the common amino acids at any
position but will require other strategies, such as the
incorporation of nonnatural amino acids or the
preparation of constrained G7-18NATE analogues
that retain the key interactions. Such investigations
are the focus of our current research.
Mexpected=1415.0, Mobserved =1415.4; 08-6028: Mexpected
=
1370.6, Mobserved =1370.8; 08-6110: Mexpected =1428.6,
Mo b se r v e d = 1429.0; 08-6046: Me x p e c t e d = 1328.5,
Mobserved=1328.8).
Crystallization of the Grb7-SH2/G7-18NATE complex
and data collection
The crystallization and diffraction data collection for the
Grb7-SH2/G7-18NATE complex have been previously
reported.41 In brief, the Grb7-SH2 protein, concentrated to
10 mg/ml and in 4-morpholineethanesulfonic acid buffer,

408
Grb7-SH2 Binding by Peptide Antagonists
Table 4. Data refinement parameters
Initial phases were obtained using MOLREP44 and the
coordinates of the apo-Grb7-SH2 structure previously
determined in our laboratory (PDB ID: 2QMS) as the
search model. The correct solution had a correlation
coefficient of 67.7 and an R-factor of 46.8, whereas the next
unrelated peak had a correlation coefficient of 55.1 and an
R-factor of 54.8. Manual model rebuilding was carried out
using Coot,45 and maximum likelihood refinement was
carried out using PHENIX.46 Noncrystallographic sym-
metry restraints were applied between pairs of chains that
occupied the asymmetric unit but were removed for the
final round of refinement. Successive rounds of refinement
and manual building were performed. The refinement
statistics are reported in Table 4. The program
MolProbity43 was used to assess the quality of the final
structures (99% of the amino acids possess backbone
torsional angles within the allowed regions of the
Ramachandran plot).
The final G7-18NATE/Grb7-SH2 complex structure
solved to 2.4 Å resolution consists of four chains of the
Grb7-SH2 domain and two G7-18NATE molecules in the
asymmetric unit. The crystal is composed of Grb7-SH2
dimers in which one monomer of each dimer is peptide
bound (chains A and B) and the other is apo-Grb7-SH2
(chains C and D). The final model also contains 116 water
molecules and has an R-factor of 23.7% (Rfree =27.7%)
(Table 4). Interpretable electron density was present for
residues 423–529 but was missing for the 10 N-terminal
residues of the construct (GSPASGTSLS) and the last 3
residues (VAL). Some disorder was also observed at
residues 460–464 and 487–489 that occur at loop regions.
The four chains did not significantly differ from each other
with Cα RMSDs of 0.32 Å (A/B holo-chains across 106
atoms) and 0.21 Å (C/D apo-chains) between like chains
and Cα RMSDs of 0.54 Å (A/C) and 0.54 Å (A/D) between
unlike chains (across 96 atoms). The whole of the G7-
18NATE peptide was clearly visible in the electron density
map. The coordinates have been deposited at the Research
Collaboratory for Structural Bioinformatics database (PDB
ID: 3PQZ).
Grb7-SH2/G7-18NATE
Data collectiona
Space group
Cell dimensions
a, b, c (Å)
P21
52.7, 79.1, 54.7
90.0, 104.4, 90.0
37.42–2.41 (2.48–2.41)
9.6 (43.2)
α, β, γ (°)
Resolution (Å)
Rmergeb (%)
I/σI
7.3 (2.0)
Reflections measured
Completeness (%)
Redundancy
16,204 (unique)
96.1 (86.3)
4.1 (3.6)
Refinement
Resolution (Å)
No. of reflections used
Rwork/Rfree
No. of atoms
Protein
2.41
16,204 (unique)
23.7/27.7
3227
200
0
Peptide
Ions
Water
116
B-factors (Å2)
Protein
34.0
26.7
—
Peptide
Ions
Water
32.1
RMSDs
Bond lengths (Å)
Bond angles (°)
Ramachandran plotc (%)
Favored regions
Allowed regions
0.010
1.2
95.7
99.0
a
Values in parentheses refer to the highest-resolution shell.
b
Rmerge =∑|I−〈I〉|/∑〈I〉, where I is the intensity of individual
reflections.
c
Determined using the program MolProbity.43
was added to lyophilized peptide to achieve a 1:2 protein-
to-peptide ratio. The mixture formed crystals in a 2-μl
hanging drop using 17.5% (w/v) polyethylene glycol 3350,
0.2 M magnesium chloride hexahydrate, 0.1 M glycine,
10% (w/v) glycerol, and 0.1 M 2-[bis(2-hydroxyethyl)
amino]-2-(hydroxymethyl)propane-1,3-diol, pH 6.0, as the
precipitant solution. The crystals formed as small thin
plates (approximately 0.01 mm×0.06 mm×0.1 mm in size)
over 2 days and, after cryo-freezing, diffracted to 2.4 Å
resolution. X-ray diffraction data were collected on the
high-throughput protein (MX1) crystallography beamline
at the Australian Synchrotron. X-ray diffraction data were
integrated and scaled with XIA2.42 The crystallographic
data and refinement statistics are reported in Table 4. The
diffraction data were consistent with space group P21 with
unit cell dimensions a=52.7 Å, b=79.1 Å, c=54.7 Å,
α=90.0, β=104.4, and γ=90.0 and collected with a
completeness of 96.1% from 37.42 to 2.41 Å resolution
(86.3% in the highest-resolution bin from 2.48 to 2.41 Å
resolution).
Phage library construction
The peptide biased libraries used in these studies
were constructed in the fUSE5 gene III phage-display
system47 according to the methods previously
reported.25 The fUSE5 vector (K91 Kan) and E. coli
host (MC1061F′) strains were generous gifts of Dr.
George Smith at the University of Missouri. There were
two rounds of library construction. The half-site cloning
method used by Cwirla et al. was employed in the
construction of these libraries.48 The two “half-site”
oligos 5′tccgccagccccgt3′ (GS5′) and 5′cggccccgcctcc3′
(GS3′) are complementary to the 5′ and 3′ ends of the
peptide-insert-encoding oligonucleotide, respectively.
The random peptide-encoding oligonucleotides for the
first round of libraries have the following sequences for
the YDN-NA biased library and the YDN biased library,
respectively: 5′gggccggaggatga(NNK)4tatgacaat(NNK)3-
tgcggggccgctg3′ and 5′gggccggagga (NNK)4tga(NNK)4-
tatgacaat(NNK)3tgc(NNK)4ggggccgctg3′ (N = equal
amounts of G, A, T, or C; K=equal amounts of G or
T). The random peptide-encoding oligonucleotides for
the second round of library construction have the
Structure determination of the Grb7-SH2/G7-18NATE
complex
The structure of the Grb7-SH2/G7-18NATE complex
was solved using the molecular replacement method.

Grb7-SH2 Binding by Peptide Antagonists
409
following sequence: 5′tattctcactcggccgacggggccggaggatgc
NNKttc123ggttacgacaacNNK456789tgcggggccgctggggcc-
gaaactgttc3′ (where N and K are the same as noted above).
The oligonucleotide denoted “123” is designed (1=12.5%
G, 12.5% A, 62.5% T, 12.5% C; 2=62.5% G, 12.5% A, 12.5%
T, 12.5% C; 3=66.6% G, 33.3% T) to obtain 50% of the
clones expressing random amino acids and the rest
expressing tryptophan at this position. The oligonucleotide
denoted “456” is designed (4=100% T; 5=33.3% G, 33.3%
A, 33.3% T; 6=33.3% G, 66.6% C) to randomize Phe, Tyr, or
Trp at this position. The oligonucleotide denoted “789” is
designed (7=50% G, 50% C; 8=100% T; 9=50% G, 50% T)
to express Leu or Val at this position. All oligonucleotides
were synthesized by Biosynthesis Inc. (Lewisville, TX).
binding isotherm was generated by plotting the heat
change evolved per injection against the molar ratio of
peptides to the Grb7-SH2 domain receptor. A blank
determination in which the peptides were titrated against
the buffer was carried out to account for heats of dilution
and mixing, which was then subtracted from the binding
data. The corrected data were then employed to fit to a
single binding site model using the nonlinear least-squares
fitter from Origin (MicroCal Software, Northampton,
MA). All of the ITC fitting parameters were kept floating.
Accession numbers
Coordinates and structure factors have been deposited
in the PDB with accession number 3PQZ.
Phage library screening
The methods for panning the newly created biased
libraries against the target and clone-binding analysis
using phage ELISA have been previously reported.25 The
only modification made here was the use of SuperSignal
ELISA Femto chemiluminescent substrate (Pierce, Rock-
ford, IL) instead of ABTS for measuring the horseradish
peroxidase (HRP)-conjugated anti-M13 antibody (Amer-
sham) using a chemiluminescent plate reader (GloRunner,
Turner BioSystems, Sunnyvale, CA). In order to achieve
an equivalent phage input in the binding studies, we
carried out normalization of the phage preparations using
the Chemiluminescent ELISA methods developed in our
laboratory.32 The colony lift assays were performed as
previously described.31 The vector DNA of each selected
clone was purified using QIAprep Spin Miniprep Kit
(Qiagen, Inc., Chatsworth, CA) according to the manufac-
turer's instructions. The pIII sequencing primer, 5-CCC
TCA TAG TTA GCG TAA CG-3, designed by Dr. George
Smith, was used for sequencing the DNA inserts. The
sequence reactions were carried out using BigDye Ver 3.1
Dye Terminator kit (Life Technologies, Foster City, CA) by
the Vermont Cancer Center DNA Analysis Facility.
Acknowledgements
We wish to thank the staff at the Protein
Crystallography Beamline (MX1) at the Australian
Synchrotron, Victoria, Australia, where the diffrac-
tion data were collected. This work was funded by
an Australian Research Council Project Grant
awarded to J.A.W. and D.N.K., a Monash University
Graduate Scholarship awarded to N.D.A., and a
National Health and Medical Research Senior
Research Fellowship awarded to M.C.J.W. The
automated DNA sequencing was performed at the
Vermont Cancer Center DNA Analysis Facility and
was supported by Vermont Cancer Center, Lake
Champlain Cancer Research Organization, and the
University of Vermont College of Medicine. The
phage display work was funded by the National
Cancer Institute (R01 CA80790) and in part by the
SD Ireland Cancer Research Foundation.
ITC determination of peptide-protein affinities
For ITC experiments, the Grb7-SH2 domain protein was
dialyzed against sodium acetate buffer (50 mM
NaOOCCH3, 100 mM NaCl, and 1 mM NaN3, pH 6.6).
The filtered dialysate buffer was used to dilute the
required amount of the Grb7-SH2 domain to a volume
of 2.5 ml and a concentration of 40 to 85 μM. Lyophilized
peptide of accurately determined mass was dissolved in
500 μl of the filtered dialysate buffer. The peptide
concentrations employed were 10 to 30 times higher
than the protein concentration. Following this, both the
peptide and the Grb7-SH2 domain solutions were
degassed for 5 min. ITC experiments were conducted
using a VP-ITC MicroCalorimeter (MicroCal, Northamp-
ton, MA). All experiments were performed in duplicate at
25 °C cell temperature, and the total number of injections
was set to 30 with volume injections of 10 μl, an injection
duration of 20 s, and an injection interval of 250 to 400 s.
The reference power was set at 20 μcal/s, and the initial
injection delay was 60–70 s. The stirring speed was
307 rpm, and the feedback mode gain was set to high
with the fast and auto equilibration options applied. A
References
1. Han, D. C., Shen, T. L. & Guan, J. L. (2001). The Grb7
family proteins: structure, interactions with other
signaling molecules and potential cellular functions.
Oncogene, 20, 6315–6321.
2. Shen, T. L. & Guan, J. L. (2004). Grb7 in intracellular
signaling and its role in cell regulation. Front. Biosci. 9,
192–200.
3. Manser, J., Roonprapunt, C. & Margolis, B. (1997). C.
elegans cell migration gene mig-10 shares similarities
with a family of SH2 domain proteins and acts cell
nonautonomously in excretory canal development.
Dev. Biol. 184, 150–164.
4. Stein, D., Wu, J., Fuqua, S. A., Roonprapunt, C.,
Yajnik, V., D'Eustachio, P. et al. (1994). The SH2
domain protein GRB-7 is co-amplified, overexpressed
and in a tight complex with HER2 in breast cancer.
EMBO J. 13, 1331–1340.
5. Tanaka, S., Mori, M., Akiyoshi, T., Tanaka, Y.,
Mafune, K., Wands, J. R. & Sugimachi, K. (1997).

410
Grb7-SH2 Binding by Peptide Antagonists
Coexpression of Grb7 with epidermal growth factor
receptor or Her2/erbB2 in human advanced esopha-
geal carcinoma. Cancer Res. 57, 28–31.
of cell migration, proliferation, and tumorigenesis.
J. Biol. Chem. 284, 20215–20226.
21. Tsai, N., Bi, J. & Wei, L. (2007). The adaptor Grb7 links
netrin-1 signaling to regulation of mRNA translation.
EMBO J. 26, 1522–1531.
22. Holt, L. J. & Daly, R. J. (2005). Adapter protein
connections: the MRL and Grb7 protein families.
Growth Factors, 23, 193–201.
6. Kishi, T., Sasaki, H., Akiyama, N., Ishizuka, T.,
Sakamoto, H., Aizawa, S. et al. (1997). Molecular
cloning of human GRB-7 co-amplified with CAB1 and
c-ERBB-2 in primary gastric cancer. Biochem. Biophys.
Res. Commun. 232, 5–9.
7. Wang, Y., Chan, D. W., Liu, V. W. S., Chiu, P. & Ngan,
H. Y. S. (2010). Differential functions of growth factor
receptor-bound protein 7 (GRB7) and its variant
GRB7v in ovarian carcinogenesis. Clin. Cancer Res.
16, 2529–2539.
8. Walch, A., Specht, K., Braselmann, H., Stein, H.,
Siewert, J., Hopt, U. et al. (2004). Coamplification and
coexpression of GRB7 and ERBB2 is found in high
grade intraepithelial neoplasia and in invasive Bar-
rett's carcinoma. Int. J. Cancer, 112, 747–753.
9. Itoh, S., Taketomi, A., Tanaka, S., Harimoto, N.,
Yamashita, Y. I., Aishima, S. I. et al. (2007). Role of
growth factor receptor bound protein 7 in hepatocel-
lular carcinoma. Mol. Cancer Res. 5, 667–673.
10. Nadler, Y., González, A. M., Camp, R. L., Rimm, D. L.,
Kluger, H. M. & Kluger, Y. (2010). Growth factor
receptor-bound protein-7 (Grb7) as a prognostic
marker and therapeutic target in breast cancer. Ann.
Oncol. 21, 466–473.
11. Ramsey, B., Bai, T., Hanlon Newell, A., Troxell, M.,
Park, B., Olson, S. et al. (2011). GRB7 protein over-
expression and clinical outcome in breast cancer.
Breast Cancer Res. Treat. 127, 659–669.
12. Kao, J. & Pollack, J. (2006). RNA interference-based
functional dissection of the 17q12 amplicon in breast
cancer reveals contribution of coamplified genes.
Genes Chromosomes Cancer, 45, 761–769.
13. Nencioni, A., Cea, M., Garuti, A., Passalacqua, M.,
Raffaghello, L., Soncini, D. et al. (2010). Grb7
upregulation is a molecular adaptation to HER2
signaling inhibition due to removal of Akt-mediated
gene repression. PLoS One, 5, e9024.
14. Bai, T. & Luoh, S. W. (2008). GRB-7 facilitates HER-2/
Neu-mediated signal transduction and tumor forma-
tion. Carcinogenesis, 29, 473–479.
15. Han, D. C. & Guan, J. L. (1999). Association of focal
adhesion kinase with Grb7 and its role in cell
migration. J. Biol. Chem. 274, 24425–24430.
16. Han, D. C., Shen, T. L. & Guan, J. L. (2000). Role of
Grb7 targeting to focal contacts and its phosphoryla-
tion by focal adhesion kinase in regulation of cell
migration. J. Biol. Chem. 275, 28911–28917.
17. Tanaka, S., Pero, S. C., Taguchi, K., Shimada, M., Mori,
M., Krag, D. N. & Arii, S. (2006). Specific peptide
ligand for Grb7 signal transduction protein and
pancreatic cancer metastasis. J. Natl. Cancer Inst. 98,
491–498.
18. Pero, S. C., Daly, R. J. & Krag, D. N. (2003). Grb7-
based molecular therapeutics in cancer. Expert Rev.
Mol. Med. 5, 1–11.
23. Chu, P. Y., Li, T. K., Ding, S. T., Lai, I. R. & Shen, T. L.
(2010). EGF-induced Grb7 recruits and promotes Ras
activity essential for the tumorigenicity of Sk-Br3
breast cancer cells. J. Biol. Chem. 285, 29279–29285.
24. Tsai, N. P., Ho, P. C. & Wei, L. N. (2008). Regulation of
stress granule dynamics by Grb7 and FAK signalling
pathway. EMBO J. 27, 715–726.
25. Pero, S. C., Oligino, L., Daly, R. J., Soden, A. L., Liu, C.,
Roller, P. P. et al. (2002). Identification of novel non-
phosphorylated ligands, which bind selectively to the
SH2 domain of Grb7. J. Biol. Chem. 277, 11918–11926.
26. Porter, C. J., Matthews, J. M., Mackay, J. P., Pursglove,
S. E., Schmidberger, J. W., Leedman, P. J. et al. (2007).
Grb7 SH2 domain structure and interactions with a
cyclic peptide inhibitor of cancer cell migration and
proliferation. BMC Struct. Biol. 7, 58.
27. Gunzburg, M., Ambaye, N., Del Borgo, M., Hertzog,
J., Pero, S., Krag, D. et al. (2010). Use of SPR to study
the interaction of G7-18NATE peptide with the Grb7-
SH2 domain. Int. J. Pept. Res. Ther. 16, 177–184.
28. Pero, S. C., Shukla, G. S., Cookson, M. M., Flemer, S. &
Krag, D. N. (2007). Combination treatment with Grb7
peptide and Doxorubicin or Trastuzumab (Herceptin)
results in cooperative cell growth inhibition in breast
cancer cells. Br. J. Cancer, 96, 1520–1525.
29. Eck, M., Shoelson, S. & Harrison, S. (1993). Recogni-
tion of a high-affinity phosphotyrosyl peptide by the
Src homology-2 domain of p56lck. Nature, 362, 87–91.
30. Rahuel, J., Gay, B., Erdmann, D., Strauss, A.,
Garcia-Echeverría, C., Furet, P. et al. (1996). Structural
basis for specificity of Grb2-SH2 revealed by a novel
ligand binding mode. Nat. Struct. Biol. 3, 586–589.
31. Sparks, A., Adey, N., Cwirla, S. & Kay, B. (1996).
Screening phage-displayed random peptide libraries.
In Phage Display of Peptides and Proteins, a Laboratory
Manual (Kay, B. K., Winter, J. & McCafferty, J., eds),
pp. 227–253, Academic Press, Inc., San Diego, CA.
32. Shukla, G. & Krag, D. (2005). A sensitive and rapid
chemiluminescence ELISA for filamentous bacterio-
phages. J. Immunoassay Immunochem. 26, 89–95.
33. Spuches, A. M., Argiros, H. J., Lee, K. H., Haas, L. L.,
Pero, S. C., Krag, D. N. et al. (2007). Calorimetric
investigation of phosphorylated and non-phosphory-
lated peptide ligand binding to the human Grb7-SH2
domain. J. Mol. Recognit. 20, 245–252.
34. Ambaye, N. D., Lim, R. C. C., Clayton, D. J.,
Gunzburg, M. J., Price, J. T., Pero, S. C. et al. (2011).
Uptake of a cell permeable G7-18NATE construct into
cells and binding with the Grb7-SH2 domain. Bio-
polymers, 96, 181–188.
19. Daly, R. J. (1998). The Grb7 family of signalling
proteins. Cell. Signalling, 10, 613–618.
20. Chu, P. Y., Huang, L. Y., Hsu, C. H., Liang, C. C.,
Guan, J. L., Hung, T. H. & Shen, T. L. (2009). Tyrosine
phosphorylation of growth factor receptor-bound
protein-7 by focal adhesion kinase in the regulation
35. Ambaye, N., Gunzburg, M., Lim, R., Price, J., Wilce,
M. & Wilce, J. (2011). Benzopyrazine derivatives: a
novel class of growth factor receptor bound protein 7
antagonists. Bioorg. Med. Chem. 19, 693–701.
36. Porter, C. & Wilce, J. (2007). NMR analysis of G7-
18NATE, a nonphosphorylated cyclic peptide

Grb7-SH2 Binding by Peptide Antagonists
411
inhibitor of the Grb7 adapter protein. Biopolymers, 88,
174–181.
37. Ivancic, M., Daly, R. & Lyons, B. (2003). Solution
structure of the human Grb7-SH2 domain/erbB2
peptide complex and structural basis for Grb7 binding
to ErbB2. J. Biomol. NMR, 27, 205–219.
38. Phan, J., Shi, Z., Burke, T. & Waugh, D. (2005). Crystal
structures of a high-affinity macrocyclic peptide
mimetic in complex with the Grb2 SH2 domain.
J. Mol. Biol. 353, 104–115.
39. Ogura, K., Shigam, T., Yokochi, M., Yuzawa, S.,
Burke, T. & Inagaki, F. (2008). Solution structure of the
Grb2 SH2 domain complexed with a high-affinity
inhibitor. J. Biomol. NMR, 42, 197–207.
42. Winter, G. (2010). xia2: an expert system for macro-
molecular crystallography data reduction. J. Appl.
Crystallogr. 43, 186–190.
43. Chen, V., Arendall, W. R., Headd, J., Keedy, D.,
Immormino, R., Kapral, G. et al. (2010). MolProbity:
all-atom structure validation for macromolecular
crystallography. Acta Crystallogr., Sect. D: Biol. Crystal-
logr. 66 , 12–21.
44. Vagin, A. & Teplyakov, A. (1997). MOLREP: an
Automated Program for Molecular Replacement.
J. Appl. Crystallogr. 30, 1022–1025.
45. Emsley, P. & Cowtan, K. (2004). Coot: model-building
tools for molecular graphics. Acta Crystallogr., Sect. D:
Biol. Crystallogr. 60, 2126–2132.
40. Janes, P. W., Lackmann, M., Church, W. B., Sanderson,
G. M., Sutherland, R. L. & Daly, R. J. (1997). Structural
determinants of the interaction between the erbB2
receptor and the Src homology 2 domain of Grb7. J.
Biol. Chem. 272, 8490–8497.
41. Yap, M., Wilce, M., Clayton, D., Perlmutter, P.,
Aguilar, M. I. & Wilce, J. A. (2010). Preparation and
crystallization of the Grb7-SH2 domain in complex
with the G7-18NATE non-phosphorylated cyclic
inhibitor peptide. Acta Crystallogr., Sect. F: Struct.
Biol. Cryst. Commun. 66, 1640–1643.
46. Adams, P., Afonine, P., Bunkóczi, G., Chen, V., Davis,
I., Echols, N. et al. (2010). PHENIX: a comprehensive
Python-based system for macromolecular structure
solution. Acta Crystallogr., Sect. D: Biol. Crystallogr. 66 ,
213–221.
47. Scott, J. & Smith, G. (1990). Searching for peptide
ligands with an epitope library. Science, 249, 386–390.
48. Cwirla, S., Peters, E., Barrett, R. & Dower, W. (1990).
Peptides on phage: a vast library of peptides for
identifying ligands. Proc. Natl. Acad. Sci. USA, 87,
6378–6382.