Peptide-based activity-based probes (ABPs) for target-specific profiling of protein tyrosine phosphatases (PTPs)

Article abstract of DOI:10.1039/b919744c

Synthesis of a novel unnatural amino acid (2-FMPT) for the solid-phase synthesis of peptide-based probes suitable for target-specific activity-based profiling of protein tyrosine phosphatases from crude proteomes is reported.

Full text of DOI:10.1039/b919744c

View Article Online / Journal Homepage / Table of Contents for this issue
COMMUNICATION
www.rsc.org/chemcomm | ChemComm
Peptide-based activity-based probes (ABPs) for target-specific profiling
of protein tyrosine phosphatases (PTPs)w
Karunakaran A. Kalesh,a Lay Pheng Tan,b Kai Lu,b Liqian Gao,c Jigang Wangb and
Shao Q. Yao*abc
Received (in Cambridge, UK) 22nd September 2009, Accepted 2nd November 2009
First published as an Advance Article on the web 17th November 2009
DOI: 10.1039/b919744c
Synthesis of a novel unnatural amino acid (2-FMPT) for the
solid-phase synthesis of peptide-based probes suitable for target-
specific activity-based profiling of protein tyrosine phosphatases
from crude proteomes is reported.
developed a PTP probe that uses a-bromobenzylphosphonate
as a non-hydrolyzable phosphotyrosine (pTyr) mimic.5c,d The
probe was shown to be specific towards PTPs in a proteomic
experiment, but the highly reactive/unstable nature of the
probe renders it impractical for widespread use. The same
group developed a newer version of PTP probes by using
phenyl vinyl sulfone/sulfonate as the enzyme-targeting
warhead.5e Vinyl sulfones/sulfonates are, however, well-
established irreversible inhibitors of other cysteine-utilizing
enzymes (i.e. cysteine proteases), and are likely to cross-label
them as well. Notwithstanding, all four existing classes of PTP
probes have so far failed to address one of the most important
issues in PTP biology—how can one monitor/profile the active
state of individual PTPs in a crude proteome, where these
enzymes’ native cellular environment is closely emulated?
Herein, we report the design and chemical synthesis of a
novel unnatural amino acid, 2-FMPT, and its successful
incorporation into peptide-based probes for potential target-
specific, activity-based profiling of individual PTPs (Fig. 1(a)).
Apart from an additional 2-fluoromethyl group located in the
aromatic ring, 2-FMPT is structurally identical to pTyr and
should cause minimal disruption in PTP recognition. The most
significant advantage of 2-FMPT over other existing PTP
probes is that, with its N- and C-terminus (as in the case of
naturally occurring amino acids), essential peptide recognition
elements which occupy the proximal positions of pTyr in a
naturally occurring PTP substrate could be introduced. With
this, the corresponding peptide-based probes could achieve
target-specific binding to/profiling of different PTPs. As shown
in Fig. 1(b), the peptide-based probe works by first binding to
the active site of the target PTP in a sequence-specific manner,
Reversible protein dephosphorylation at the tyrosine residues
plays a pivotal role in signal transduction.1 In vivo, it is regulated
by the opposite action of protein tyrosine phosphatases (PTPs)
and protein tyrosine kinases (PTKs).2 In humans, there are
well over 100 known PTPs, which, together with several
hundred protein kinases, control the phosphorylation of as
many as 1–2% of all human proteins. In order to maintain an
intricate balance of this highly complex phosphoproteome
network, PTPs carry out dephosphorylation with a high
degree of specificity inside the cell, although in vitro they have
shown only moderate specificity towards synthetic peptide
substrates.3 Consequently, chemical and biological methods
that report not only the global PTP activities, but more
importantly the precise enzymatic activity of individual PTPs,
may offer unprecedented views on how these enzymes carry
out their biological functions under native physiological
settings. Activity-based protein profiling (ABPP) is a chemical
proteomics strategy that makes use of mechanism-based small
molecule probes to detect active enzymes present in a crude
proteome, and has been successfully applied to different
classes of enzymes.4 In order to develop activity-based probes
(ABPs) that detect global PTP activities, Lo et al. first
explored small molecules containing a 4-fluoromethyl phenyl
phosphate (FMPP) moiety which serves as a phosphotyrosine
mimic, and upon dephosphorylation by a PTP, generates a
highly reactive quinone methide intermediate that sub-
sequently alkylates nucleophiles present in the PTP active site.5a
We made similar PTP probes using 2-difluoromethyl phenyl
phosphate (DFPP) which presumably works via a similar
mechanism as FMPP.5b Unfortunately, due to the diffusible
nature of the quinone methide generated, both classes of
probes possess poor PTP specificity and cross-react with other
nearby proteins in the crude proteome. Zhang et al. recently
a Department of Chemistry, National University of Singapore,
3 Science Drive 3, Singapore 117543. E-mail: chmyaosq@nus.edu.sg;
Fax: (65) 67791691; Tel: (65) 65162925
b Department of Biological Sciences, National University of Singapore,
3 Science Drive 3, Singapore 117543
Fig.
1 (a) Structures of the unnatural amino acid, 2-FMPT,
c NUS MedChem Program of the Life Sciences Institute, National
University of Singapore, 3 Science Drive 3, Singapore 117543
w Electronic supplementary information (ESI) available: Detailed
experimental procedures and results. See DOI: 10.1039/b919744c
and its corresponding peptide-based ABPs. (b) Proposed mechanism
of the sequence-specific, activity-based labeling of PTPs using the
peptide-based probes.
ꢀc
This journal is The Royal Society of Chemistry 2010
Chem. Commun., 2010, 46, 589–591 | 589

View Article Online
and is subsequently dephosphorylated (I; as in the case of a
synthetic PTP substrate). The ensuing charge delocalization
followed by spontaneous elimination of –F generates the
highly reactive quinone methide intermediate which, upon
attack by a nearby nucleophilic residue in the enzyme, would
give rise to the covalent enzyme-probe complex (II). We
anticipated that, in our newly developed probes, introduction
of the additional peptide fragments specific towards individual
PTPs should improve probe binding affinity as well as
specificity, thus overcoming the diffusion problem observed
in the previous ABPs (i.e.FMPP/DFPP).
The synthesis of the Fmoc-protected 2-FMPT (10 in
Scheme 1) was carried out from Boc-L-tyrosine. Briefly,
formylation of 2 at the ortho position of the phenolic group
by Reimer–Tiemann reaction gave 3 in moderate yield (39%).6
Subsequently, 3 was protected to give the corresponding benzyl
ester 4 (80%), followed by phosphorylation with chlorodiethyl-
phosphate to give 5 (70%). Reduction of the CHO group in 5
with NaBH4 gave 6 (69%) followed by fluorination with DAST,
giving 7 (74%). Finally, deprotection of the benzyl ester in 7,
followed by Boc-to-Fmoc conversion gave 10. To show that 10
could be conveniently incorporated into peptides using
standard SPPS protocols, eleven peptide probes, designed
against five different PTPs (PTP1B, TCPTP, SHP1, SHP2
and LMWPTP), were synthesized using Rink amide resin
(Scheme 1, Table 1). For comparison, ten corresponding
phosphopeptides were also synthesized, in which the naturally
occurring pTyr was used in place of 2-FMPT (see ESIw).
Under standard labeling conditions, the probes were found
to label only PTPs in acivity-dependant manner (see ESIw).
Next, activity-based labeling of the ten rhodamine-containing
probes, P1–P10, with five different PTPs was tested to generate
the so-called activity-based fingerprints.7 As shown in Fig. 2,
clearly distinctive labeling profiles of individual PTPs were
observed, and to a large extent corroborated well with their
known substrate preferences (Table 1 in Scheme 1). For
example, the top-5 probes (P2, P3, P7, P9 and P10) which
generated the strongest labeling against PTP1B, were derived
expectedly from known PTP1B substrates. TCPTP, known to
have similar substrate preferences as PTP1B, gave a similar
activity-based fingerprint. For SHP1, two of the most active
probes are P1 and P6. Of the two SHP2-specific probes
(P8 and P10), one was derived from its known substrate.
Similarly, LMWPTP also gave a distinctive labeling profile,
which is in good agreement with its substrate preference.8
Further characterizations of the labeling profiles were carried
out by kinetic measurements of PTP1B inactivation in the
presence of the ten probes, as well as measurements of PTP1B-
mediated dephosphorylation with ten corresponding phospho-
peptides (see ESIw). Analysis of pseudo-first-order inactivation
rate constants, kobs, as a function of probe concentration showed
time- and concentration-dependant inactivation with saturation
kinetics (typical of irreversible inhibition) and followed the
Kitts–Wilson relationship. The maximal inactivation rate ki
and the dissociation constant Ki for the ten probes were
determined. The dissociation constants of the probes reflect
their substrate preferences, with highest 1/Ki values correlated
well with the most active probes (in their fluorescence labeling),
as well as the most preferred phosphopeptide substrates (in the
Scheme 1 Synthesis of 10 and solid-phase peptide synthesis (SPPS) of
the eleven probes. Reagents and conditions: (a) CHCl3, NaOH, H2O,
reflux, 39%; (b) BnOH, EDC, DMAP, DCM, 0 1C to RT, 80% to give
4, then chlorodiethylphosphate, DBU, DCM, 0 1C to RT, 70% to give
5; (c) NaBH4, EtOH–THF (4 : 1), 0 1C, 69%; (d) DAST, DCM, 0 1C to
RT, 74%; (e) H2, Pd–C, MeOH, 96% to give 8; (f) TFA–DCM (1 : 1),
90% to give 9; (g) FmocOSu, NaHCO3, THF–water (1 : 1), 76%;
(h) standard Fmoc-based SPPS conditions; (i) TMSI, DCM, RT;
(j) TFA–H2O–TIS (95 : 2.5 : 2.5), RT.
2À
amount of PO3 release). All these lines of evidence suggest
our peptide-based activity probes can be used to report not
only PTP enzymatic activities, but more importantly the
substrate specificities of individual PTPs.
Finally, the target-specific, activity-based PTP profiling of
these peptide-based probes was evaluated in crude bacterial
and mammalian proteomes using the P3 probe. First, the high
specificity of this probe towards PTP1B was unambiguously
confirmed by a spike experiment with a bacterial proteome
ꢀc
This journal is The Royal Society of Chemistry 2010
590 | Chem. Commun., 2010, 46, 589–591

View Article Online
biotinylated version of P3. As shown in Fig. 3(c), upon labeling
with P11, the HEK293T cell lysate was subjected to affinity
enrichment and subsequently analyzed by Western blotting;
endogenous PTP1B was successfully detected (lane 2). This
experiment thus demonstrates the feasibility of our peptide-
based activity-based probes for target-specific profiling of
endogenous PTP activities in cellular lysates. We noticed that
in Fig. 2(b), in addition to the endogenous PTP1B, the probe
also labeled a number of other proteins in the HEK293T lysate.
We speculate they might have originated from other endogen-
ous PTPs which were present in the cell lysate and accepted P3
as a substrate. We are currently carrying out large-scale
proteomic/mass spectrometric experiments to identify these
labeled proteins and will report our findings in due course.
In conclusion, we have successfully synthesized a novel
unnatural amino acid (2-FMPT) which is a close mimic of
phosphotyrosine. With the peptide-based probe design, we
were able to demonstrate target-specific profiling of PTPs
present in mammalian cell lysates. Further improvement of
our probes may be achieved by incorporation of longer
peptide recognition sequences or by genetically or semi-
synthetically incorporating the unnatural amino acid into
suitable protein-based substrates of PTPs.10 Ultimately, it
might be possible to extend our strategy to in vivo profiling
of PTP/substrate interactions in living cells.
Fig. 2 (a) Fluorescence labeling profiles (in gray scale) of five different
PTPs with the ten peptide probes (left to right). The labeling profiles of P3
were highlighted (boxed in red). (b) The fluorescence intensity of each band
is quantified using ImageQuant software and the relative fluorescence
intensity (y-axis) of labeling of each PTP were plotted against the panel
of 10 probes (x-axis, P1 to P10 from left to right in each graph).
(free of endogenous PTP activities); as shown in Fig. 3(a), as
low as 5 ng of PTP1B (0.05% total proteome) could be readily
detected with the probe. Next, total cell lysates from two
different mammalian cells (HEK293T and NIH3T3) were
prepared and treated with P3 (Fig. 3(b)). The resulting labeled
lysates were visualized by in-gel fluorescence scanning
(left panel) and Western blotting using anti-PTP1B antibody
(right panel); Western blotting indicated the presence of
endogenously expressed PTP1B in HEK293T (B52 kDa)
but not in NIH3T3 cells. A corresponding fluorescent band
was observed in the fluorescence gel from the HEK293T
lysate, but not from the NIH3T3 lysate (lanes 1 and 2,
respectively, highlighted in red box). This is in agreement with
previous reports that extremely low or undetectable expression
levels of endogenous PTP1B in NIH3T3 cell lines were
observed.9 To further validate the labeling of endogenous
PTP1B, we performed pull-down experiments using P11—the
Funding support was provided by the tier-2 grant (R-143-
000-394-112) from the Ministry of Education of Singapore.
Notes and references
1 (a) L. N. Johnson and R. J. Lewis, Chem. Rev., 2001, 101,
2209–2242; (b) T. Hunter, Cell, 2000, 100, 113–127.
2 A. Alonso, J. Sasin, N. Bottini, I. Friedberg, I. Friedberg,
A. Osterman, A. Godzik, T. Hunter, J. Dixon and T. Mustelin,
Cell, 2004, 117, 699–711.
3 (a) Z.-Y. Zhang, A. M. Thieme-Sefler, D. Maclean, D. J. McNamara,
E. M. Dobrusin, T. K. Sawyer and J. E. Dixon, Proc. Natl. Acad. Sci.
U. S. A., 1993, 90, 4446–4450; (b) S. W. Vetter, Y.-F. Keng,
D. S. Lawrence and Z.-Y. Zhang, J. Biol. Chem., 2000, 275,
2265–2268; (c) C. Blanchetot, M. Chagnon, N. Dube, M. Halle and
´ ´
M. L. Tremblay, Methods, 2005, 35, 44–53; (d) M. Kohn, M. Gutierres-
¨
Rodriguez, P. Jonkheijm, S. Wetzel, R. Wacker, H. Schroeder,
H. Prinz, C. M. Niemeyer, R. Breinbauer, S. E. Szedlacsek and
H. Waldmann, Angew. Chem., Int. Ed., 2007, 46, 7700–7703.
4 For recent reviews, see: (a) M. J. Evans and B. F. Cravatt, Chem.
Rev., 2006, 106, 3279–3301; (b) M. Uttamchandani, J. Li, H. Sun
and S. Q. Yao, ChemBioChem, 2008, 9, 667–675.
5 (a) L. C. Lo, T. L. Pang, C. H. Kuo, Y. L. Chiang, H. Y. Wang and
J. J. Lin, J. Proteome Res., 2002, 1, 35–40; (b) Q. Zhu, X. Huang,
G. Y. J. Chen and S. Q. Yao, Tetrahedron Lett., 2003, 44, 2669–2672;
(c) S. Kumar, B. Zhou, F. Liang, W.-Q. Wang, Z. Huang and
Z.-Y. Zhang, Proc. Natl. Acad. Sci. U. S. A., 2004, 101, 7943–7948;
(d) S. Kumar, B. Zhou, F. Liang, H. Yang, W.-Q. Wang and
Z.-H. Zhang, J. Proteome Res., 2006, 5, 1898–1905; (e) S. Liu,
B. Zhou, H. Yang, Y. He, Z.-X. Jiang, S. Kumar, L. Wu and
Z.-Y. Zhang, J. Am. Chem. Soc., 2008, 130, 8251–8260.
6 M. E. Jung and T. I. Lazarova, J. Org. Chem., 1997, 62, 1553–1555.
7 (a) H. Sun, S. Chattopadhaya, J. Wang and S. Q. Yao, Anal.
Bioanal. Chem., 2006, 386, 416–426; (b) R. Srinivasan, X. Huang,
S. L. Ng and S. Q. Yao, ChemBioChem, 2006, 7, 32–36.
8 (a) L. Bialy and H. Waldmann, Angew. Chem., Int. Ed., 2005, 44,
3814–3839; (b) S. Zhang and Z.-Y. Zhang, Drug Discovery Today,
2007, 12, 373–381.
Fig. 3 (a) Labeling of spiked PTP1B in the presence of bacterial
proteome. Different amounts (left to right: 1000 ng, 400 ng, 200 ng,
100 ng, 20 ng, 5 ng and 0.5 ng) of recombinant PTP1B was spiked
to 10 mg BL21 bacterial proteome and the labeling was performed with
10 mM probe P3 for 1 h. (b) Labeling of PTPs in mammalian cell lysates.
Left panel, in-gel fluorescence analysis of global PTP activity profiles
obtained from total cell lysates of HEK293T and NIH3T3 cells
(20 mg total proteins/lane) with probe P3. Right panel, Anti PTP1B
blot of the two labeled lysates detecting endogenous PTP1B at B52 kDa
from HEK293T cells (lane 1) but not from NIH3T3 cells (lane 2). (c)
Pull-down results using the biotinylated probe P11 with HEK293T
lysates followed by enrichment with NeutrAvidin beads. Anti-PTP1B
WB of the probe treated HEK293T cell lysate after pull-down experi-
ments detected the endogenous PTP1B at B52 kDa (lane 2) while no
PTP1B was detected from the untreated pull-down fraction (lane 1).
9 L. V. Ravichandran, H. Chen, Y. Li and M. J. Quon, Mol.
Endocrinol., 2001, 15, 1768–1780.
10 (a) J. M. Xie and P. G. Schultz, Nat. Rev. Mol. Cell Biol., 2006, 7,
775–782; (b) T. W. Muir, Annu. Rev. Biochem., 2003, 72, 249–289.
ꢀc
This journal is The Royal Society of Chemistry 2010
Chem. Commun., 2010, 46, 589–591 | 591