Modulation of SHP-1 phosphatase activity by monovalent and bivalent SH2 phosphopeptide ligands

Article abstract of DOI:10.1002/bip.21307

A sequence derived from the epithelial receptor tyrosine kinase Ros (pY2267) represents a high-affinity binding partner for protein tyrosine phosphatase SHP-1 and was recently used as lead structure to analyze the recognition requirements for the enzyme's N-SH2 domain. Here, we focused on a set of peptides comprising C-terminally extended linear and conformationally constrained side chain-bridged cyclic N-SH2 ligands based on the consensus sequence LxpYhxh(h/b)(h/b) (x = any amino acid, h = hydrophobic, and b = basic residue). Furthermore, the bivalent peptides described were designed to modulate the activity of SHP-1 through binding to both, the N-SH2 domain as well as an independent binding site on the surface of the catalytic domain (PTP domain). Consistent with previous experimental findings, surface plasmon resonance experiments revealed dissociation constants of most compounds in the low micromolar range. One peptide, EGLNpYc[KVD]MFPAPEEE-NH₂, displayed favorable binding affinity, but reduced ability to stimulate SHP-1. Docking experiments revealed that the binding of this ligand occurs in binding mode I, recently described to lead to an inhibited activation of SHP-1. In summary, results presented in this study suggest that inhibitory N-SH2 ligands of SHP-1 may be obtained by designing bivalent compounds that associate with the N-SH2 domain and simultaneously occupy a specific binding site on the PTP domain.

Full text of DOI:10.1002/bip.21307

Modulation of SHP-1 Phosphatase Activity by Monovalent and
Modulation of SHP-1 Phosphatase Activity by Monovalent and
Bivalent SH2 Phosphopeptide Ligands
Bivalent SH2 Phosphopeptide Ligands
1
1
1
2
3
¨
¨
Kathleen Teichmann, Toni Kuhl, Ina Konnig, * Karin Wieligmann, Martin Zacharias,
Diana Imhof1
1
Center for Molecular Biomedicine, Department of Biochemistry, Peptide Chemistry Group,
Friedrich Schiller University, Hans-Knoell-Strasse 2, D-07745 Jena, Germany
2
Max Planck Research Unit for Enzymology of Protein Folding, Weinbergweg 22, D-06120 Halle, Germany
3
Department of Physics, Technical University Munich, James-Franck-Str. 1, D-85748 Garching, Germany
Received 1 July 2009; revised 26 August 2009; accepted 30 August 2009
Published online 18 September 2009 in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/bip.21307
binding affinity, but reduced ability to stimulate SHP-1.
ABSTRACT:
Docking experiments revealed that the binding of this
A sequence derived from the epithelial receptor tyrosine
ligand occurs in binding mode I, recently described to
kinase Ros (pY2267) represents a high-affinity binding
lead to an inhibited activation of SHP-1. In summary,
partner for protein tyrosine phosphatase SHP-1 and was
results presented in this study suggest that inhibitory
recently used as lead structure to analyze the recognition
N-SH2 ligands of SHP-1 may be obtained by designing
requirements for the enzyme’s N-SH2 domain. Here, we
bivalent compounds that associate with the N-SH2
focused on a set of peptides comprising C-terminally
domain and simultaneously occupy a specific binding site
extended linear and conformationally constrained side
#
on the PTP domain. 2009 Wiley Periodicals, Inc.
chain-bridged cyclic N-SH2 ligands based on the
Biopolymers 93: 102–112, 2010.
consensus sequence LxpYhxh(h/b)(h/b) (x 5 any amino
Keywords: SHP-1; tyrosine phosphatases; SH2 domain
acid, h 5 hydrophobic, and b 5 basic residue).
ligands; surface plasmon resonance; phosphotyrosyl
Furthermore, the bivalent peptides described were
peptides
designed to modulate the activity of SHP-1 through
binding to both, the N-SH2 domain as well as an
independent binding site on the surface of the catalytic
domain (PTP domain). Consistent with previous
experimental findings, surface plasmon resonance
experiments revealed dissociation constants of most
compounds in the low micromolar range. One peptide,
EGLNpYc[KVD]MFPAPEEEÀÀNH2, displayed favorable
This article was originally published online as an accepted
preprint. The ‘‘Published Online’’ date corresponds to the
preprint version. You can request a copy of the preprint by
emailing the Biopolymers editorial office at biopolymers@wiley.
com
INTRODUCTION
he reversible phosphorylation is probably one of the
most important regulatory modifications in proteins
catalyzed by protein tyrosine kinases and protein
tyrosine phosphatases.1 A phosphorylated tyrosine
*Present address: Leibniz Institute for Age Research, Fritz Lipmann Institute,
Beutenbergstrasse 11, D-07745 Jena, Germany.
Correspondence to: Diana Imhof; e-mail: diana.imhof@uni-jena.de
Contract grant sponsor: Deutsche Forschungsgemeinschaft (DFG)
Contract grant numbers: IM 97/2-1, 2-2
residue may thereby associate with recognition sites
T
Contract grant sponsor: Friedrich Schiller University, Jena, Germany
provided by small protein modules, such as phosphotyro-
sine-binding or src homology 2 (SH2) domains.2,3 The latter
C
VV
2009 Wiley Periodicals, Inc.
102
Biopolymers Volume 93 / Number 1

Modulation of SHP-1 Activity by SH2 Phosphopeptide Ligands
103
form independently folding structural units of ꢀ100 amino was generated to elucidate the origin of the high binding affinity
acids and are found in a large number and wide variety of and the inhibitory effect of the cyclic ligands. Considering the
proteins.4–6 SHP-1, a cytosolic protein tyrosine phosphatase, experimental and theoretical findings we suggested that the
consists of two SH2 domains (N-SH2, C-SH2) N-terminal to binding mode and the nature of the residue at pY13 determine
the degree of phosphatase activation.21,23 The lead structure Ros
the catalytic domain (PTP domain) and a short C-terminal
tail. This phosphatase is predominantly expressed in hemato-
poietic cells and at lower levels in epithelial cells. SHP-1 has
been shown to negatively regulate signaling downstream of
transmembrane receptors such as EpoR, c-Kit, CSFR, and
BCR/TCR, as well as receptor tyrosine kinases, such as Ros
and EGFR.4,5,7–10 In this way, SHP-1 is involved in a variety
of cellular activities, including proliferation, differentiation,
and cell development.11 The function of SHP-1 at the differ-
ent stages of embryonic stem cell differentiation as well as
the negative effects on hematopoietic and immune cell func-
tion have been extensively investigated and described.12
pY2267 binds in the class II binding mode in an extended con-
formation where the amino acid side chain at pY13 disturbs
the interaction between the N-SH2 domain and the PTP do-
main. As a result, this mode of association leads to an activation
of the enzyme. In contrast, ligands belonging to consensus
sequence class I binding mode associate in a bent conformation.
In this case, the localization of the pY13 residue is sterically in-
compatible with the N-SH2 conformation that keeps the
enzyme inactive.21 In addition to the studies performed with
peptides of the minimal consensus sequence there was evidence
that residues beyond pY13 may also significantly affect the
ligand-binding affinity and specificity of SHP-1 SH2 domains.
In a recent report it was demonstrated that the positions pY14
and pY15 indeed are part of the specificity determinants of the
N-SH2 domain, because binding of a phosphopeptide ligand is
greatly enhanced by either large hydrophobic (Trp, Tyr, Phe,
Nle) or positively charged (Arg, Lys, or His) residues at these
positions.15 Therefore, one could argue that a peptide sequence
with a considerable high binding affinity should be comprised
of residues pY22 to pY15.
SHP-1 exists in an inactive form in the cytosol with the
N-SH2 domain directly blocking the catalytic domain of the
enzyme. The phosphatase activity of SHP-1 is stimulated
through the binding of a phosphotyrosine (pY)-containing
ligand to the N-SH2 domain (Figures 1A and 1B).16,17 For-
mer studies demonstrated that the high-affinity binding fea-
tures of SHP-1 phosphopeptide ligands are specified by resi-
dues at the pY22, pY11, and pY13 positions (pY is defined
as position 0).18 A combinatorial library screening revealed
that the SHP-1 N-SH2 domain selects two classes of consen-
sus sequences with a preference for LXpY(M/F)X(F/M) (class
I, X 5 any amino acid, M 5 Nle, isostere for Met) (Figure
1C).14,19 Our previous investigations of the interaction of
In either case, these reports are all very useful for the
design and development of compounds specific for SHP-1
N-SH2 to modulate phosphatase activity, particularly inhibi-
tory compounds. Such inhibitors are valuable tools to study
the physiological and pathophysiological role of SHP-1.
However, in contrast to other SH2 domains the knowledge
for SHP-1 N-SH2 inhibitor design is still insufficient. Rea-
sons for this may be the different strategies that are consid-
ered for PTP effectors and the rather large interface between
the ligand and the SH2 domain. However, aside from PTP
active site inhibitors we are interested in N-SH2 inhibitors
that are capable of reducing or even preventing the dissocia-
tion process of the N-SH2/PTP complex. Furthermore, struc-
tural analysis of an inactive form of the enzyme in complex
with an inhibitory ligand may serve as a fundamental basis to
clarify the mechanism of SHP-1 phosphatase activation,
thereby facilitating identification of potent lead compounds
for the design of N-SH2 peptide mimetics.
SHP-1 with
a
sequence from RTK Ros (EGLN-
pY2267MVLÀÀNH2, M 5 Met) showed that this octapeptide
binds to SHP-1 N-SH2 in the low micromolar range.7,20
Based on these studies we suggested that the residues at the
positions pY11 and pY13 and the binding geometry at
these positions play an important role for the mechanism of
SHP-1 activation.21,22 The most potent linear peptides pre-
ferred bulky residues at pY11 (Phe, Nle, or Abu(bPh)) and
also at pY13 (Hfe, Abu(bPh), or Ser(bPh)).22 We later
hypothesized that an optimal occupation of the binding cav-
ities for pY11 and pY13 triggers the allosteric conforma-
tional change within the N-SH2 domain, which in turn leads
to dissociation of the phosphatase domain. The binding of
such a ligand thus mediates a strong SHP-1 activation.
In contrast, cyclic compounds EGLc[K(COCH2NH)-
pYMX]LÀÀNH2 (X 5 Glu or Asp) were found to bind to the
N-SH2 domain with a considerable high affinity, but displayed
a strongly reduced capability to stimulate SHP-1. According to
our model, this is a result of an imperfect fit of the amino acid
side chains occupying the corresponding pockets on the SH2
domain as described earlier.22 The following series of peptides
In this study, we synthesized and evaluated a new series of
linear and cyclic peptides that contain two different motifs:
one reflecting the N-SH2 recognition features and the other
motif representing a sequence selected for binding to a sup-
posed binding site on the surface of the PTP domain. This
strategy of so-called bivalent ligands has also been success-
fully applied in the inhibitor design of tandem SH2 domain
Biopolymers

104
Teichmann et al.
FIGURE 1 (A) Primary structure of the SHP-1 N-SH2 domain (human). All residues that are in
contact with the C-terminus of the ligand are bold underlined. (B) Ribbon representation of the three-
dimensional structure of SHP-1 N-SH2 in complex with lead peptide Ros pY2267.13 (C) Consensus
sequences for the N-SH2 domain determined by combinatorial peptide library screening.14,15
proteins and protein tyrosine kinases.24 We also performed
docking studies for selected ligands on the basis of the terti-
ary structure of SHP-1 and SHP-2 determined by crystal
structure analysis and of SHP-2 N-SH2 in complex with a
peptide.15,17,21,22,24,25 The resulting compounds exhibited an
Novabiochem (Merck Bioscience AG, Schwalbach, Germany), Iris
Biotech GmbH (Marktredwitz, Germany), Orpegen Pharma GmbH
(Heidelberg, Germany), and Neo MPS (Strasbourg, France), respec-
tively. Solvents for chromatography were of analytical grade from
VWR International GmbH (Dresden, Germany). Biomol green was
obtained from Biomol GmbH (Hamburg, Germany). The HiPrep
activity comparable with or lower than the lead peptide Ros 16/60 Sephacryl HR gel filtration column and GSTrap FF column
were purchased from GE Healthcare (Freiburg, Germany). The
pY2267.
One
compound,
EGLNpYc[KVD]NleFPA-
streptavidin-coated (SA) sensor chips for surface plasmon resonance
(SPR) detection were from GE Healthcare.
PEEEÀÀNH2, showed a markedly reduced ability to stimulate
SHP-1 activity, while displaying high binding affinity. This
may serve as a good candidate for further studies and for the
search for suitable and potent lead compounds associated
with both the N-SH2 and the PTP domains.
Solid-Phase Peptide Synthesis
Peptide synthesis was carried out manually on Rink amide MBHA
resin (0.64 mmol/g) using syringes from Intavis AG (Koeln,
Germany). The side chains of the trifunctional amino acids were
protected as follows: Glu(OtBu), Asn(Trt), Tyr(PO3H2) (for linear
peptides) and Tyr[PO(OBzl)OH], Asp(OAll)-OH, and Lys(Alloc)-
OH (for cyclic peptides). A general protocol for the solid phase
MATERIALS AND METHODS
Materials
Rink amide MBHA resin, and Na-Fmoc–protected amino acids and synthesis of the phosphopeptides is described in more detail
coupling reagents (HBTU, HOBt, PyBOP) were purchased from elsewhere.21 Briefly, coupling reactions were performed in DMF
Biopolymers

Modulation of SHP-1 Activity by SH2 Phosphopeptide Ligands
105
using Fmoc amino acids (4 equiv) activated with HBTU (4 equiv)
in the presence of DIEA (8 equiv) for 0.5–1 h (double couplings).
Fmoc-phospho amino acids were coupled in twofold excess with
HBTU (2 equiv) and DIEA (6 equiv). Fmoc removal was effected by
treating the resin twice with 20% piperidine in DMF for 5 and
15 min, respectively. The cyclization procedure was performed using
PyBOP (6 equiv) and DIEA (12 equiv) in DMF for 3 h twice as
described earlier.13,21 All peptides were synthesized in the N-termi-
nal unprotected form for the phosphatase assay and in the biotinyl-
ated version for SPR binding studies.21 Cleavage of the peptides
from the resin with concomitant side-chain deprotection was
achieved by treating the resins with TFA/water/triisopropylsilane
(95:2.5:2.5) for 3 h (peptides containing Tyr(PO3H2)-OH) or 5–6 h
(peptides containing Tyr[PO(OBzl)OH]). The crude peptides were
precipitated in diethyl ether, centrifuged, and washed three times
with diethyl ether. Finally, the peptides were purified on a semipre-
parative reversed-phase HPLC using a Shimadzu LC-8A system
(Shimadzu, Duisburg, Germany) equipped with a C18 column
(Eurospher 100; Knauer, Berlin, Germany). The identity and purity
of the final products were established by analytical reversed-phase
HPLC (LC-10AT; Shimadzu) using a Vydac 218TP column (5 lm
Surface Plasmon Resonance Analysis
The binding affinity of pY peptides to the N-SH2 domain of SHP-1
was assessed by SPR analysis on a BIAcore 2000 instrument (Phar-
macia Biosensor AB, Uppsala, Sweden). Biotinylated pY peptides
were immobilized on SA sensor chips. Prior to immobilization, a
SA sensor chip was conditioned with 1M NaCl/50 mM NaOH
according to the manufacturer’s instructions. The binding assays
were conducted at room temperature in HBS buffer (10 mM
HEPES, 150 mM NaCl, 3.4 mM EDTA, 0.005% Tween 20, pH 7.4).
The phosphopeptides were immobilized on the chip (flow cells 2–4)
at a concentration of 0.1 lM (flow rate: 10 lL/min) to reach an
increase in response units of 50–75 compared with the starting level.
Flow cell 1 was used as the control for correction of nonspecific
binding interactions. Varying concentrations of GST-tagged N-SH2
protein (3.41–1000 nM) were passed over the immobilized SA chip
for 2 min at a flow rate of 15 lL/min. Dissociation (dissociation
time: 20 s) was monitored during subsequent washing of the chip
using HBS buffer. To remove the protein from the peptide, HBS
buffer containing 0.1% SDS (10 s at a flow rate of 30 lL/min) was
applied. The data were analyzed with BIAevaluation version 2.0. To
determine the dissociation constant (KD), the equilibrium response
units (RUeq) were plotted against protein concentration and fit to
the following equation: RUeq 5 RUmax [GST-N-SH2]/(KD 1 [GST-
N-SH2]), where RUmax is maximum response units.
˚
particle size, 300 A pore size, 4.6 3 25 mm), MALDI mass spec-
trometry (Perseptive Biosystems, Weiterstadt, Germany), and amino
acid analysis (Eppendorf-Biotronik, Hamburg, Germany).
Expression and Purification of GST-Fusion Proteins
Full-length SHP-1 and the N-SH2 domain were expressed as GST
(glutathione-S-transferase)-fusion proteins in BL21(DE3)pLys cells
as previously reported.15,19 For both proteins the pGEX-5X1 expres-
sion vector system was used. Purification of GST-SHP-1 was per-
formed as described earlier.15 The GST-N-SH2 protein purification
was carried out using the Akta Prime FPLC Protein Purification Sys-
tem from GE Healthcare (GSTrap FF column). In case of the SPR
measurements, the GST-N-SH2 domain was additionally purified
by gel filtration (HiPrep 16/60 Sephacryl HR gel filtration column)
and freshly prepared prior to use.
Molecular Modeling
Molecular modeling of all peptide-N-SH2 complexes was based on
the crystal structure pdb1aya (in complex to peptide VLpYTAV)
representing mode II peptide binding and on the pdb1ayc (in com-
plex with GGpYMAMG) representing mode I binding.25,27 The
crystal structures contain the N-SH2 domain of SHP-2. Sequence
alignment of N-SH2 of SHP-1 and SHP-2 indicates that the two
proteins differ only in residues outside of the peptide-binding
region. It is therefore assumed that the peptide-binding groove of
N-SH2 of SHP-2 can serve also as a model for the binding pocket of
the N-SH2 domain of SHP-1. All residue substitutions or additions
at the ends of the bound peptides were performed using the SPDBV
molecular modeling program based on the experimental peptide
backbone coordinates.28 Peptide-N-SH2 complexes were energy
minimized using the Amber program keeping the N-SH2 structure
restraint to the crystallographic structure.29
Phosphatase Assay
The peptide concentrations were determined by quantitative amino
acid analysis after complete hydrolysis (6N HCl, 1108C, 24 h) of an
aliquot of the peptide stem solutions and by the malachite green
assay for released inorganic phosphate.21,24,26 Stimulation of SHP-1
activity was measured using p-nitrophenyl phosphate (pNPP) as the
substrate. Peptides were diluted in assay buffer (100 mM HEPES,
150 mM NaCl, 1 mM EDTA, 10 mM DTT) to final assay concentra-
tions of 25, 50, 100, 200, and 500 lM. The reaction mixture (total
volume, 50 lL) contained 0–500 lM peptide, 100 mM pNPP, and
0.4 lg full-length SHP-1 (80 lg/mL). The reaction was initiated by
the addition of SHP-1 and allowed to proceed at room temperature
for 30 min. The reaction was quenched by adding 1M NaOH
(100 lL). The absorbance at 405 nm was measured on a VERSA
max (Molecular Devices GmbH, Ismaning, Germany). The SHP-1
activity reported is given relative to the activation level stimulated
by Ros pY2267 (500 lM, 1.0). Absorption measured for the 500 lM
concentration is set 1.0 for the lead peptide Ros pY2267. All meas-
urements were carried out in duplicate, and the results are given as
the average of two independent experiments.
RESULTS AND DISCUSSION
Phosphopeptide Selection and Synthesis
The natural interaction partner, Ros pY2267 (1), has been
repeatedly used as lead peptide for the investigation of the
recognition features of SHP-1 SH2 domains.13,21,22 On the
basis of these results, new SHP-1 N-SH2 effectors were
derived by applying the concept of a ‘‘two-site binder’’
(Figure 2A).24 For the known binding pockets (pY11, pY13
to pY15) we could already use the knowledge obtained from
previous studies leading to the selection of distinct amino
acids at pY22 to pY15 (motif I, Figure 2A). In addition to
Biopolymers

106
Teichmann et al.
PTP domain are predominantly Arg and Lys residues (R362,
K364, K366) which in turn are predicted to bind to the acidic
residues of the respective ligands (Table I).
The linear phosphopeptides were prepared either with an
N-terminal sequence identical to Ros pY2267 (1), e.g., result-
ing in peptide 4, or in N-terminally truncated versions (5, 6).
In these peptides, positions pY11 to 13 were adopted from
the lead structure (1), while the sequence Nle-Phe-Pro at
pY14 to 16 represents a high-affinity motif beyond pY13
for SHP-1 N-SH2 binding described earlier.15 The latter tri-
peptide sequence was incorporated in peptides 2–10 as it is
or in a slightly modified form with the aim to increase the
binding affinity compared with the lead peptide. The selec-
tion of residues beyond pY16 in peptides 4–10, however,
was according to molecular modeling studies that identified
a cluster of basic amino acids on the surface of the PTP do-
main (Figure 2B). This cluster is supposed to be accessed by
negatively charged residues, e.g., a series of glutamic acid res-
idues (4–10, Table I).
FIGURE 2 (A) Schematic representation of a SHP-1 N-SH2 biva-
lent ligand based on the concept of a ‘‘two-site binder.’’24 (B) Dock-
ing of a bivalent ligand to SHP-1 (space filled model) with the N-
SH2 domain colored red and the PTP domain colored yellow. The
peptide is represented by sticks and colored green and blue. The
cluster of basic amino acids on the surface of the PTP domain is
presented in blue.
Based on the earlier observation that conformationally re-
stricted compounds may have a partially inhibitory effect on
SHP-1 activity, we hypothesized that restriction as well as a
distinct side-chain conformation are responsible for the loss
of phosphatase activity while sustaining high binding affinity
to the protein domain.15,21,22 Thus, it seemed plausible to
combine the inhibitory properties of cyclic compounds with
the motif predicted to occupy a binding pocket on the sur- the high-affinity binding motif at positions C-terminal to
face of the PTP domain (motif II, Figure 2B), there was the pY13 resulting in peptides 2 and 3, which were cyclized
need for a linker sequence to combine both motifs (I ? II). between positions pY12 and 14 and pY12 and 16, respec-
˚
The distance between motif I and II of ꢀ4 A gave rise to
tively. With these compounds the impact of ring size on the
bridge this region by one or two amino acids (Pro, Ala). Key
residues forming the pocket for motif II on the surface of the
stimulation of phosphatase activity and binding affinity can
be analyzed. Furthermore, a C-terminal extension of com-
Table I Physicochemical Characterization of Synthetic Phosphopeptides 1–10
Peptide
Sequence
Ring Atoms
Ring Position
MW (g/mol)
tR (min)a
Yield (%)
1
2
EGLNpYMVLÀÀNH2
—
14
20
—
—
—
17
14
14
20
—
1016.80b
1388.66b
1426.99c
1953.27c
1767.09c
1539.50c
1981.17d
1912.66b
1982.99d
1982.57c
20.62
25.88
32.05
30.63
29.61
27.37
31.58
27.38
26.83
31.29
68
28
32
58
64
60
16
17
27
25
EGLNpYMc[KLD]FPÀÀNH2
12 ? 14
12 ? 16
—
3
EGLNpYMc[KLNleFD]ÀÀNH2
EGLNpYMVLNleFPAPEEEÀÀNH2
__LNpYMVLNleFPAPEEEÀÀNH2
____pYMVLNleFPAPEEEÀÀNH2
EGLc[KpYMD]LNleFPAPEEEÀÀNH2
EGLNpYc[KVD]NleFPAPEEEÀÀNH2
EGLNpYMc[KLD]FPAPEEEÀÀNH2
EGLNpYMc[KLNleFD]APEEEÀÀNH2
4
5
—
—
6
7
21 ? 12
11 ? 13
12 ? 14
12 ? 16
8
9
10
a
Peptides have been used after purification by semipreparative HPLC ([95% pure). HPLC analysis was carried out according to the following conditions:
10–50% eluent B for 40 min (eluent A: 0.1% TFA/water, eluent B: 0.1% TFA/acetonitrile), flow rate: 1.0 mL/min, detection: k 5 220 nm.
b
Molecular weight value was detected as [M 1 H]1.
c
Molecular weight value was detected as [M 1 Na]1.
d
Molecular weight value was detected as [M 1 K]1.
Biopolymers

Modulation of SHP-1 Activity by SH2 Phosphopeptide Ligands
107
pounds 2 and 3 led to peptides 9 and 10 as two representa-
tives of the bivalent ligands according to the model shown in
Figure 2. In addition to these peptides, cyclic ligands were
generated in which the position of the amino acids partici-
pating in the ring closure have also been varied. In contrast
to the previously reported N-SH2 effectors which have been
cyclized between positions pY21 and pY12,21,22 we herein
changed both the position and size of the ring ranging from
14- to 20-membered rings and involving residues at pY21 to
pY16 (7–10, Table I). The amino acids Lys and Asp were
again used for lactam bridge formation as previously
reported.21,22
The synthetic strategies used for the generation of the
linear and cyclic phosphopeptides were according to previ-
ously described investigations using preformed phosphoryl-
ated tyrosine (building block approach).13,21 In brief, for
the linear compounds (1, 4–6) the phosphate-unprotected
tyrosine derivative was used,30 whereas the monobenzyl-
protected Fmoc-phosphotyrosine building block was
applied in case of the cyclic ligands (2, 3, 7–10). The syn-
thesis of the linear phosphopeptides could easily be per-
formed. However, it was interesting to note that a variety of
side products was observed during the preparation of the
cyclic peptides, predominantly of compounds 7 and 8. In
general, side products occurring during the synthesis of
such cyclic phosphopeptides include incompletely depro-
tected linear precursor compounds, e.g., monoallyl-
protected peptides that prevent the formation of cyclic
product. Usually, this can be resolved by repetition of the
Alloc/OAll-cleavage procedure leading to a higher amount
of the unprotected linear precursor and in turn a higher
yield of the cyclic peptide. In addition to that, however, the
most frequently found side products in some peptides are
the monobenzyl-protected linear precursor and cyclic pepti-
FIGURE 3 (A–C) Concentration-dependent stimulation of SHP-
1 activity by the synthetic peptides 1–10. The results are given rela-
tive to the activity of SHP-1 stimulated by Ros pY2267 (1).
des (up to 10% each). Despite a prolonged cleavage time Stimulation of SHP-1 Activity
([6 h) during removal of these peptides from the polymer Stimulation of SHP-1 activity by the phosphopeptides 1–10
support, the phosphotyrosine-protected versions were still was determined at neutral pH using the full-length enzyme
detected by HPLC and mass spectrometric methods. Thus, and pNPP as the substrate. Indeed, differences in the activa-
Alloc/OAll-deprotection, cyclization, as well as Bzl-cleavage tion profile evoked by the individual peptides were observed
were hampered in some of our peptides, probably because as shown in Figure 3.
of an accumulation of hydrophobic, bulky amino acids and
In general, peptides can be classified into three groups: (1)
side chain protecting groups. The content on monobenzyl- peptides that, depending on the concentration, stimulated
protected side products was less in cyclic peptides 2 and 3 SHP-1 activity more efficiently than the lead peptide, (2)
and not detectable in peptides 9 and 10. This fact led to the peptides that showed a similar behavior, and (3) peptides
conclusion that the close proximity of the phosphotyrosine lacking the ability to stimulate the phosphatase (Table II).
residue and the lactam-bridged amino acids might be the Peptides 2 and 3, belonging to the first class, need to be dis-
cause for incomplete deprotection of the benzyl group from tinguished with respect to the concentration that leads to a
the phosphotyrosine side chain. The characterization data more efficient activation compared with Ros pY2267 (1).
of the synthesized peptides reported in this study are sum- Whereas in case of cyclic peptide 3 the activation was delayed
marized in Table I.
relative to 1 resulting in a twofold higher EC50 value, peptide
Biopolymers

108
Teichmann et al.
Table II Half-Maximal Activation, Dissociation Constants, and
Binding Mode of Phosphopeptides 1–10
higher EC50 values than peptide 1 were determined for cyclic
peptides 8 (647 lM) and 9 (144 lM). These results were
interesting because of the fact that peptides 2 and 9 share the
same primary structure at positions pY24 to pY16 as well
as the same ring position and ring size. Therefore, differences
in the ability to stimulate SHP-1 and in the binding to the N-
SH2 domain may only arise from the C-terminal pentapep-
tide sequence in peptide 9 which represents one member of
the group of the bivalent ligands. Slightly improved EC50 val-
ues but lower activation levels at 500 lM were observed for
cyclic peptides 7 and 10 in comparison to 1. However, with
the exception of peptide 8, which contains a 14-membered
ring closed between positions 11 and 13, cyclic peptides 7,
9, and 10 showed a similar behavior with respect to stimula-
tion of SHP-1 activity. Thus, in these cases it seems that nei-
ther ring size nor position strongly influences the ability to
activate the phosphatase.
Peptide
EC50 (lM)
KD (lM)
Binding Modea
1
2
140
12
1.10 6 0.07
0.90 6 0.04
[70
0.07 6 0.01
0.74 6 0.06
NBc
Mode II
Mode II
Mode I
3
272
43
4
Mode I/II
Mode I/II
Mode I/II
Mode I/II
Mode I
5
35
6
NAb
70
7
0.12 6 0.01
0.50 6 0.09
0.34 6 0.03
0.15 6 0.01
8
[500
144
67
9
10
Mode I/II
Mode I/II
a
Predicted binding mode according to models described earlier.21 For
several peptides both modes of binding are theoretically possible.
b
NA, not active.
c
NB, no binding.
2 efficiently stimulated SHP-1 activity up to 100 lM with
half-maximal activation already reached at 12 lM. However, Determination of Binding Affinities to
at concentrations higher than 100 lM this peptide inhibits SHP-1 N-SH2 Domain
phosphatase activity. The same effect was observed earlier SPR has been shown to be exceptionally useful for the study
with the peptides EGLNpYFVh and EGLNpYFVF (h 5 of a wide range of biological interactions, particularly pro-
homophenylalanine) at concentrations up to 100 and tein–protein interactions. In the past, we have established a
200 lM, respectively.21,22 Obviously, dissociation of the SPR-based assay using the GST-SHP-1 N-SH2 domain,21,22
PTP-N-SH2 complex triggered by these peptides at lower whereas the pHis- or MBP-tagged proteins have been
concentrations occurs more efficient compared with the employed by others.14,15 All experiments described herein
other peptides of each series. In addition, this seems to be were carried out under the same conditions using the freshly
combined with an increased ability to compete with pNPP prepared and purified SH2 protein domain. Binding interac-
for the active site at higher concentrations, which is not the tions of GST-N-SH2 at various concentrations with biotin–
case for the other peptides investigated.
peptides immobilized on a SA sensor chip were measured
The linear peptides 4 and 5 synthesized with the aim to using a Biacore 2000 instrument and displayed as a senso-
increase binding affinity showed an efficient stimulation of gram exemplified in Figure 4. The KD values, summarized in
SHP-1 with comparable EC50 values of 43 and 35 lM, respec- Table II, were determined to evaluate whether the increased
tively. This is in contrast to lead peptide 1 which reached half phosphatase activity evoked by the respective ligands is
maximal activation only at 140 lM. However, this result is in caused by a higher binding affinity for the N-SH2 domain.
agreement with what had been expected for these C-termi- In addition, SPR measurements were useful to examine the
nally extended peptides due to the optimal occupation of the inhibitory peptides for their association to the protein do-
binding pockets for residues at pY13 to 15. Peptide 6, main. Indeed, the current results merit comparison with the
which lacks one recognition determinant, namely the residue linear and cyclic pY-peptides reported earlier, displaying KD
at pY22, has been introduced as a negative control and to values in the range of 0.07–1.10 lM.21,22 The binding affin-
find out whether C-terminal prolongation compensates for ities of the linear peptides 4, 5, and 6 directly reflected their
this loss to a certain degree. The latter is not the case because structural differences in the region N-terminal to pY with the
no SHP-1 activity has been detected for this compound. lowest KD value found for compound 4, whereas peptide 5
With respect to half-maximal activation, peptides 4 and 5 are binds with comparable affinity as lead peptide 1, and peptide
among the most efficient activators together with the cyclic 6 did not associate with the protein domain. Thus, additional
peptide 2 (see above). The latter was cyclized between posi- interactions as in peptide 4 increase binding affinity, whereas
tions 12 and 14, while ring closure in peptide 3 showed the recognition determinant pY-2 is essential for association and
delayed stimulation was performed between 12 and 16, is not compensated by the C-terminal extension-carrying
thus representing a larger ring. Comparable or significantly preferred residues at pY14 and pY15. KD values for the
Biopolymers

Modulation of SHP-1 Activity by SH2 Phosphopeptide Ligands
109
FIGURE 4 (A) Sensogram showing the interaction between the GST-N-SH2 domain of SHP-1
and immobilized biotin-phosphopeptide 1. (B) Sensogram for the interaction between the protein
domain and biotin-phosphopeptide 9. The concentrations of the phosphopeptides on the sensor
chip ranged from 50 to 75 RU; protein concentration ranged from 3.4 to 1000 nM. Insets: Data
obtained indicate that the binding is dose dependent.
cyclic peptide 2 is in the range of the native Ros pY2267 pep- 11 and 13, which are most important for recognition of
tide (1), though stimulation of phosphatase activity was C-terminal to pY, retained a high binding affinity (KD
5
enhanced at lower concentrations. The association of peptide 0.5 lM), but exhibited inhibitory properties with respect to
3 to the protein, however, was rather unusual because bind- stimulation of SHP-1 activity.
ing affinity was strongly decreased compared with other pep-
tides. This is somehow consistent with the delayed effect on
phosphatase activity with respect to the half-maximal activa- Molecular Modeling of SHP-1 N-SH2 Domain in
tion, but not with the potency at concentrations higher than Complex with Phosphopeptides
200 lM. However, the reason for this discrepancy has not yet Sterically possible structures of the complexes formed by
been clarified. Beside the linear peptide 4, the best binding various peptides and a model structure of the N-SH2 do-
affinities were found for cyclic peptides 7, 9, and 10 showing main of SHP1 were generated for a qualitative interpreta-
KD values of 0.12, 0.34, and 0.15 lM, respectively. This is in tion of the observed binding and activation data. As already
good agreement with their capacity to stimulate SHP-1 activ- discussed in a previous study,21 it is likely that the Ros pep-
ity. Interestingly, cyclic peptide 8 cyclized between positions tide binds to the N-SH2 domain in the class II mode
Biopolymers

110
Teichmann et al.
FIGURE 5 Molecular modeling of linear and cyclic peptide binding to the SHP1/2-N-SH2 do-
main. (A) Kinked class I mode peptide binding to N-SH2 domain (molecular surface) based on the
crystal structure (pdb1ayc) in complex with the peptide GGpYMDMS (shown as atom-color-coded
stick model). The close proximity of the side chains pY11 (Met) and pY13 (Met, indicated by
arrows) is sterically compatible with the side chain cyclization of peptide 8. (B) Linear class II
mode peptide binding (pdb1aya; same coloring scheme as in A). Arrows indicate residues pY11
and pY13 in mode II peptide binding. (C) Model of peptide 3 (sticks) in mode I binding to the N-
SH2 domain (based on the crystal structure shown in A). The connection between side chains
pY12 (Lys) and pY16 (Asp) is indicated by an arrow. (D) Model of peptide 2 in mode II binding
to the N-SH2 domain (chemical bond between positions 12 (Lys) and 14 (Asp) indicated by an
arrow). The dotted arrows show the location of residue pY16 that is too far from residue 12 to
allow peptide cyclization without disruption of the binding mode.
(Figure 5A) because it is close to the class II consensus pep- sterically only possible for class I peptide binding mode
tide sequence and results in strong activation of the phos- (Figure 5C) which involves a kink in the peptide between
phatase specific for mode II binding. This mode of binding residue pY12 and pY13. This shortens the distance
is also sterically possible for cyclic peptide 2 (Figure 5D) between residues pY12 and pY16 (to bring side chains at
which shows strong binding and efficient activation. Simi- 12 and 16 close enough for cyclization without disrupting
larly, cyclization that involves residues pY21 and pY12 key interactions between peptide and N-SH2 binding
(peptide 7) is compatible with the linear peptide class II region). However, the cyclization is incompatible with any
mode (strong binding and activation). However, peptide 3 close contacts of the peptide’s C-terminal tail and the SH2
with a cyclization that involves a longer peptide segment is domain (Figure 5C), resulting in weaker binding compared
Biopolymers

Modulation of SHP-1 Activity by SH2 Phosphopeptide Ligands
111
with peptide 2. For the linear class II binding mode the dis-
This study demonstrates that bivalent ligands consisting
tance is too large for cyclization without perturbing the of individual motifs can recognize SH2 and PTP domains of
peptide structure at the binding interface (Figure 5B). Con- SHP-1. Future expansions of this approach could include
sistent with our model that only class II peptide binding investigations on how conformationally restricted peptide
allows efficient activation, peptide 3 binds but shows only sequences can be transformed into peptidomimetic struc-
weak activation at lower concentrations.
tures while retaining both high binding affinity and reduced
Similar cyclization that involves residues pY11 and pY13 capability to stimulate SHP-1. This kind of SHP-1 inhibitors
(peptide 8) is only compatible with class I binding because would represent useful tools in order to explore the allosteric
only in this binding mode the side chains of the two residues mechanism of SHP-1 activation.
have a sufficiently short distance (are in contact) to allow
The authors thank Inge Agricola and Dr. Kornelia Hampel for tech-
chemical bonding in a cyclic structure. The restriction to
nical assistance and useful discussions, respectively, and also Professor
mode I binding results again in only weak phosphatase
Frank Grosse (FLI Jena) for the provision of the BIAcore 2000 instru-
ment.
activation. In general, the additional C-terminal acidic resi-
dues result in increased binding (and activation if cyclization
is compatible with class II binding mode as outlined earlier)
possibly due to the proposed additional interactions illus-
trated in Figure 2.
REFERENCES
1. Neel, B. G.; Tonks, N. K. Curr Opin Cell Biol 1997, 9, 193–204.
2. Sawyer, T. K. Biopolymers 1998, 47, 243–261.
3. Sawyer, T. K.; Shakespeare, W. C.; Wang, Y.; Sundaramoorthi,
R.; Huang, W. S.; Metcalf, C. A., III; Thomas, M.; Lawrence,
B. M.; Rozamus, L.; Noehre, J.; Zhu, X.; Narula, S.; Bohacek, R.
S.; Weigele, M.; Dalgarno, D. C. Med Chem 2005, 1, 293–319.
4. Mayer, B. J. Methods Mol Biol 2006, 332, 79–99.
5. Neel, B. G.; Gu, H.; Pao, L. Trends Biochem Sci 2003, 28, 284–
293.
CONCLUSIONS
We have synthesized a set of phosphorylated linear and cyclic
N-SH2 ligands derived from the LNpYMVL sequence of
motif pY2267 of the high-affinity receptor Ros for SHP-1.
Most compounds were designed to bind in a bivalent mode
predicted to interact with different SHP-1 protein domains,
the N-terminal SH2 domain, and the catalytic phosphatase
(PTP) domain. While the majority of the tested peptides
activate the enzyme comparable to the lead compound, one
bivalent ligand (8) displayed a strongly reduced capability to
stimulate SHP-1 and at the same time was identified as a
high-affinity binding partner. A considerable influence of the
ring closure within this ligand was observed. According to
our model of two different binding modes for SHP-1 N-SH2
association, the cyclization in peptide 8 involving residues
pY11 and pY13 is only compatible with class I binding,
indicating that ring position shifting and reduction of the
number of ring atoms is possible with retention of binding
6. Sadowski, I.; Stone, J. C.; Pawson, T. Mol Cell Biol 1986, 6,
4396–4408.
7. Sonnenberg-Riethmacher, E.; Walter, B.; Riethmacher, D.; God-
ecke, S.; Birchmeier, C. Genes Dev 1996, 10, 1184–1193.
8. Tenev, T.; Bohmer, S. A.; Kaufmann, R.; Frese, S.; Bittorf, T.;
Beckers, T.; Bohmer, F. D. Eur J Cell Biol 2000, 79, 261–271.
9. Zhang, J.; Somani, A. K.; Siminovitch, K. A. Semin Immunol
2000, 12, 361–378.
10. Keilhack, H.; Tenev, T.; Nyakatura, E.; Godovac-Zimmermann,
J.; Nielsen, L.; Seedorf, K.; Bohmer, F. D. J Biol Chem 1998, 273,
24839–24846.
11. Chong, Z. Z.; Maiese, K. Histol Histopathol 2007, 22, 1251–
1267.
12. Paling, N. R.; Welham, M. J. Blood 2005, 105, 4290–4297.
13. Imhof, D.; Nothmann, D.; Zoda, M. S.; Hampel, K.; Wegert, J.;
Bohmer, F. D.; Reissmann, S. J Pept Sci 2005, 11, 390–400.
capacity, but simultaneously leads to a shift from mode II to 14. Beebe, K. D.; Wang, P.; Arabaci, G.; Pei, D. Biochemistry 2000,
39, 13251–13260.
15. Imhof, D.; Wavreille, A. S.; May, A.; Zacharias, M.; Tridanda-
mode I binding. The latter is due to the short distance
between the two residues involved in cyclization, resulting in
pani, S.; Pei, D. J Biol Chem 2006, 281, 20271–20282.
a strongly restricted conformation allowing only for weak
16. Plutzky, J.; Neel, B. G.; Rosenberg, R. D. Proc Natl Acad Sci
USA 1992, 89, 1123–1127.
phosphatase activation. The importance of hydrogen-bond-
ing interactions between the protein domains and the pep-
tide backbone for high-affinity binding is in agreement with
our findings, because generally the additional C-terminal
acidic residues result in increased binding compared with the
parent peptide with the exception of ligand 6 in which one
specificity determinant for recognition by SHP-1 N-SH2 is
missing.
17. Yang, J.; Liu, L.; He, D.; Song, X.; Liang, X.; Zhao, Z. J.; Zhou,
G. W. J Biol Chem 2003, 278, 6516–6520.
18. Burshtyn, D. N.; Yang, W.; Yi, T.; Long, E. O. J Biol Chem 1997,
272, 13066–13072.
19. Sweeney, M. C.; Wavreille, A. S.; Park, J.; Butchar, J. P.; Tridan-
dapani, S.; Pei, D. Biochemistry 2005, 44, 14932–14947.
20. Keilhack, H.; Muller, M.; Bohmer, S. A.; Frank, C.; Weidner, K.
M.; Birchmeier, W.; Ligensa, T.; Berndt, A.; Kosmehl, H.;
Biopolymers

112
Teichmann et al.
Gunther, B.; Muller, T.; Birchmeier, C.; Bohmer, F. D. J Cell Biol 27. Lee, C. H.; Kominos, D.; Jacques, S.; Margolis, B.; Schles-
2001, 152, 325–334.
singer, J.; Shoelson, S. E.; Kuriyan, J. Structure 1994, 2, 423–
438.
28. Guex, N.; Peitsch, M. C. Electrophoresis 1997, 18, 2714–2723.
21. Hampel, K.; Kaufhold, I.; Zacharias, M.; Bohmer, F. D.; Imhof,
D. ChemMedChem 2006, 1, 869–877.
22. Imhof, D.; Wieligmann, K.; Hampel, K.; Nothmann, D.; Zoda, 29. Case, D. A.; Pearlman, D. A.; Caldwell, J. W.; Cheatham,
M. S.; Schmidt-Arras, D.; Zacharias, M.; Bohmer, F. D.; Reiss-
mann, S. J Med Chem 2005, 48, 1528–1539.
23. Wieligmann, K.; Pineda De Castro, L. F.; Zacharias, M. In Silico
Biol 2002, 2, 305–311.
T. E., III; Wang, J.; Ross, W. S.; Simmerling, C. L.; Darden,
T. A.; Merz, K. M.; Stanton, R. V.; Cheng, A. L.; Vincent, J.
J.; Crowley, M.; Tsui, V.; Gohlke, H.; Radmer, R. J.; Duan,
Y.; Pitera, J.; Massova, I.; Seibel, G. L.; Singh, U. C.; Weiner,
P. K.; Kollman, P. A. Amber 8; University of California: San
Francisco, 2003.
24. Levitzki, A.; Mishani, E. Annu Rev Biochem 2006, 75, 93–
109.
25. Hof, P.; Pluskey, S.; Dhe-Paganon, S.; Eck, M. J.; Shoelson, S. E. 30. Bonewald, L. F.; Bibbs, L.; Kates, S. A.; Khatri, A.; McMurray,
Cell 1998, 92, 441–450.
J. S.; Medzihradszky, K. F.; Weintraub, S. T. J Pept Res 1999, 53,
161–169.
26. Harder, K. W.; Owen, P.; Wong, L. K.; Aebersold, R.; Clark-
Lewis, I.; Jirik, F. R. Biochem J 1994, 298 (Part 2), 395–
401.
Reviewing Editor: David A. Case
Biopolymers