Rapid assembly and in situ screening of bidentate inhibitors of protein tyrosine phosphatases

Article abstract of DOI:10.1021/ol052895w

We have successfully designed and synthesized a small library of protein tyrosine phosphatase (PTP) inhibitors, in which the so-called "click chemistry" or Cu(I)-catalyzed 1,3-dipolar alkyne-azide coupling reaction was carried out for rapid assembly of 66 different bidentate compounds. Subsequent in situ enzymatic screening revealed a potential PTP1B inhibitor (IC₅₀ = 4.7 μM) which is 10-100 fold more potent than other PTPs.

Full text of DOI:10.1021/ol052895w

ORGANIC
LETTERS
2006
Vol. 8, No. 4
713-716
Rapid Assembly and in Situ Screening
of Bidentate Inhibitors of Protein
Tyrosine Phosphatases
Rajavel Srinivasan,† Mahesh Uttamchandani,‡ and Shao Q. Yao*,†,‡
Departments of Chemistry and Biological Sciences, Medicinal Chemistry Program of
the Office of Life Sciences, National UniVersity of Singapore, 3 Science DriVe 3,
Singapore 117543, Republic of Singapore
chmyaosq@nus.edu.sg
Received November 30, 2005
ABSTRACT
We have successfully designed and synthesized a small library of protein tyrosine phosphatase (PTP) inhibitors, in which the so-called “click
chemistry” or Cu(I)-catalyzed 1,3-dipolar alkyne azide coupling reaction was carried out for rapid assembly of 66 different bidentate compounds.
−
Subsequent in situ enzymatic screening revealed a potential PTP1B inhibitor (IC50
PTPs.
) 4.7 µM) which is 10−100 fold more potent than other
Fragment-based assembly is a recently developed drug
discovery approach which enables high-throughput identi-
fication of small molecule inhibitors using a minimal number
of compounds as building blocks.1 The approach is powerful
especially against protein targets which possess multiple
binding pockets in their active sites. Currently, there are
mainly three methods used to assist the assembly of
compounds: (1) the NMR-based SAR strategy;1b (2) the
tethering method developed by Wells et al.;1c and (3) the in
situ screening method developed by Wong and co-workers,1d
which was largely based on the “click chemistry” pioneered
by Sharpless et al.2 Among them, the click chemistry
approach is highly versatile in that it requires neither
specialized equipment nor mutations in the target proteins,
making it easily adaptable by most research laboratories. To
this end, it has been used successfully in the discovery of
inhibitors against HIV protease, SARS 3CL protease, R-fu-
cosidase, sulfotransferase, and R-1,3-fucosyltransferase.1d,3
Protein tyrosine phosphatases (PTPs) are a large and
structurally diverse class of signaling enzymes. Defective
regulation of these enzymes has significant implications in
various human diseases.4 For example, PTP1B, one of the
best known PTPs, has long been identified as the key enzyme
which causes obesity and diabetes.4b In the past few years,
much effort has been made in attempts to develop potential
drugs, mostly small molecule-based drugs, that target PTP1B
in vivo with high efficacy and minimum side effects.5 The
biggest challenge, however, arises from the high sequence
homology and highly conserved active site structures among
† Department of Chemistry.
‡ Department of Biological Sciences.
(3) Wu, C.-Y. et al. Proc. Natl. Acad. Sci. U.S.A. 2004, 101, 10012-
10017. (b) Wu, C.-Y.; Chang, C.-F.; Chen, J. S.-Y.; Wong, C.-H. Angew.
Chem., Int. Ed. 2003, 42, 4661-4664. (c) Best, M. D.; Birk, A.; Chapman,
E.; Lee, L. V.; Cheng, W.-C.; Wong, C.-H. ChemBioChem 2004, 5, 811-
819. (d) Lee, L. V.; Mitchell, M. L.; Huang, S.-J.; Fokin, V. V.; Sharpless,
K. B.; Wong, C.-H. J. Am. Chem. Soc. 2003, 125, 9588-9589.
(4) (a) Zhang, Z.-Y. Curr. Opin. Chem. Biol. 2001, 5, 416-423. (b)
Johnson, T. O.; Ermolieff, J.; Jirousek, M. R. Nat. ReV. Drug DiscoVery
2002, 1, 696-709. (c) Hunter, T. Cell 2000, 100, 113-127.
(1) For a review, see: (a) Arkin, M. R.; Wells, J. A. Nat. ReV. Drug
DiscoVery 2004, 3, 301-317. (b) Szczepankiewicz, B. G. et al. J. Am. Chem.
Soc. 2003, 125, 4087-4096. (c) Erlanson, D. A.; Braisted, A. C.; Raphael,
D. R.; Randal, M.; Stroud, R. M.; Gordon, E. M.; Wells, J. Proc. Natl.
Acad. Sci. U.S.A. 2000, 97, 9367-9372. (d) Birk, A.; Lin, Y.-C.; Elder, J.;
Wong, C.-H. Chem. Biol. 2002, 9, 891-896.
(2) For a review, see: Kolb, H. C.; Sharpless, K. B. Drug DiscoVery
Today 2003, 8, 1128-1137.
10.1021/ol052895w CCC: $33.50
© 2006 American Chemical Society
Published on Web 01/21/2006

different PTPs,4 making it extremely difficult to identify
small molecules which discriminatively bind to PTP1B over
other PTPs. Recently, Zhang et al. discovered the presence
of a secondary aryl-phosphate binding site near the active
site of PTP1B,6 thus shedding light on the development of
bidentate inhibitors which might impart both potency and
specificity against the enzyme.7 On the basis of the same
model, researchers at Abbott laboratories recently discovered
cell-permeable, micromolar bidentate inhibitors containing
a core N-phenyloxamic acid mimic (Figure 1),8 which, upon
on the above criteria was constructed (Scheme 1): 5 alkyne-
containing core groups and 14 azide-containing peripheral
groups were used as building blocks which were subse-
quently assembled using click chemistry or the Cu(I)-
catalyzed 1,3-dipolar alkyne-azide coupling reaction, pre-
viously shown to be highly modular, efficient, and compatible
to most functional groups.2 More importantly, the assembly
reaction can be conducted in aqueous environments without
the need of any deleterious reagents, thus allowing direct
enzymatic screening and rapid identification of potential
“hits” from the library. It should be noted that, although click
chemistry had previously been used for assembly and
identification of inhibitors against other enzymes,1d,3 our
report herein is, to the best of our knowledge, the first
example for the synthesis and discovery of PTP inhibitors.
We chose cell-permeable analogues of the N-phenyloxamic
acid in Abbott’s inhibitor (shown in Figure 1) as the core
groups (boxed in Scheme 1), as they were shown to bind
tightly to the primary site of PTP1B where the phosphoty-
rosine in a PTP1B substrate normally occupies.8 In addition,
these pharmacophores have been shown to possess “druglike”
properties. Except for the introduction of an alkyne handle,
minimal changes were made to these structures to ensure
they maintained reasonable binding affinity to PTP1B and
possibly other PTPs. Other more potent core groups were
not considered because of their poor pharmacological
properties.7,8a As for the peripheral group of the inhibitor,
14 different azide-containing building blocks were synthe-
sized (Scheme 2), each of which bears an aromatic unit with
different polarity and an alkyl linker of different length (2,
3, or 5 Cs). Aromatic compounds were chosen in the current
study as they were previously shown to preferentially bind
PTP1B at its secondary site.8 As most PTPs are known to
have a highly conserved active site (i.e., primary binding
site), the key advantage of our strategy is that, in the future,
without changes in the core group, a variety of other azide-
containing molecules may be easily introduced in the
bidentate library to generate inhibitors conferring high
specificity against other PTPs while retaining good affinity.
Figure 1. Chemical structure of Abbott’s cell-permeable, bidentate
PTP1B inhibitor. The core and peripheral groups bind to the active
site and the secondary site of PTP1B, respectively.
testing in PTP1B-expressing COS-7 cells, showed much
greater cellular activities (in terms of inhibition and selectiv-
ity) and pharmacokinetic properties than other inhibitors
possessing potent inhibitory activity in vitro but suffering
from low cell permeability and selectivity in vivo.7,8a The
key to their success is the use of the NMR-based fragment
assembly approach which greatly facilitated the rational
improvement of lead compounds. However, the technical
demand, the need for specialized equipment as well as a large
quantity of proteins, and the intrinsic low throughput
invariably limit this method in routine high-throughput
screening (HTS). We aim to develop novel approaches which
enable both high-throughput synthesis and screening of small
molecule inhibitors against PTP1B, as well as other multi-
valent proteins. Herein, we report the use of click chemistry
for rapid assembly, followed by in situ screening, of bidentate
inhibitors against PTP1B.
To take advantage of both the primary and secondary
binding sites within PTP1B and potentially other PTPs, our
inhibitor design entails the following key criteria: (1) each
member is made of a modular, bidentate structure containing
both a core and a peripheral group for potential binding to
the enzyme; (2) whenever possible, each of the two (core
and peripheral) groups should possess optimized pharma-
cological properties; and (3) they may be efficiently as-
sembled in situ for direct screening against potential PTPs.
As a proof-of-concept experiment, a 66-member library based
Of the five different building blocks for the core group,
compound A was synthesized from the commercially avail-
able 4-hydroxyacetophenone, 15, which upon refluxing with
propargyl tosylate in the presence of K2CO3 and benzo-18-
crown-6 afforded 16 with excellent yield (93%). Subse-
quently, condensation between 16 and dimethyloxalate in
the presence of NaOMe, followed by cyclization of the
resulting product, gave the isoxazole carboxylic methylester,
17, in modest yield (two steps) with published procedures.8a
Next, base-catalyzed hydrolysis of 17 gave the free acid A.
Compounds B, C, and E were similarly synthesized starting
from either 5′-chloro- or 5′-fluoro-2′-hydroxyacetophenone
(i.e., 18 or 21). In the case of E, condensation of 19 with
dimethyloxalate followed by direct treatment of the resulting
product with hydrazine sulfate in the presence of p-TsOH
generated the pyrazole carboxylic methylester, 29, which
upon hydrolytic cleavage with NaOH afforded compound
E in 88% yield. For the synthesis of compound D, com-
mercially available 4-nitroacetophenone, 24, was condensed
(5) (a) van Huijsduijnen, R. H.; Bombrun, A.; Swinnen, D. Drug
DiscoVery Today 2002, 7, 1013-1019. (b) Bialy, L.; Waldmann, H. Angew.
Chem., Int. Ed. 2005, 44, 3814-3839.
(6) Puius, Y. A.; Sullivan, M.; Lawrence, D. S.; Almo, S. C.; Zhang,
Z.-Y. Proc. Natl. Acad. Sci. U.S.A. 1997, 94, 13420-13425.
(7) Guo, X.-L.; Shen, K.; Wang, F.; Lawrence, D. S.; Zhang, Z.-Y. J.
Biol. Chem. 2002, 277, 41014-41022.
(8) (a) Liu, G. et al. J. Med. Chem. 2003, 46, 4232-4235. (b) Liu, G. et
al. J. Med. Chem. 2003, 46, 3437-3440.
714
Org. Lett., Vol. 8, No. 4, 2006

Scheme 1. Synthesis of the Alkyne Building Blocks
with dimethyloxalate followed by cyclization to afford 25
Compound 14 was synthesized from the corresponding amino
(36% overall yield). Subsequent reduction of the nitro group
in 25 with 10% Pd-C in the presence of H2 gave 26 (91%),
which upon coupling with mono-Boc protected succinic acid,
36 (Supporting Information), generated 27. Next, deprotec-
tion of the Boc group followed by coupling with propargyl-
amine afforded 28. The methylester in 28 was subsequently
hydrolyzed with NaOH to afford compound D in 88% yield.
The 14 different aromatic azides were synthesized using
highly efficient, two/three-step procedures (Scheme 2).
2-Bromoacetyl chloride and 3-bromopropionyl chloride were
purchased commercially, whereas 5-bromovaleryl chloride,
30, was synthesized from 5-bromovaleric acid, 29, and
thionyl chloride. Subsequently, nine different aromatic
amines, representing different polarities and substituents,
were acylated with one of the above three bromoalkylacyl
chlorides, followed by an SN2 substitution reaction with
sodium azide in DMF to generate the corresponding azides.
acid (i.e., lysine).
The Cu(I)-catalyzed ligation of alkyne blocks and azide
blocks was next carried out. It was found that a mixed solvent
system containing t-BuOH, water, and DMSO (1:1:0.1)
enabled both the alkynes and the azides to be completely
dissolved, and at the same time, the 1,3-dipolar coupling was
carried out with extremely high efficiency (Supporting
Information). More importantly, the products could be taken
directly without further purification and screened for inhibi-
tion against PTPs in subsequent microplate-based enzyme
assays (vide infra). The click chemistry was initiated by
catalytic amounts of CuSO4 and sodium ascorbate. LC-MS
analysis of all 66 coupling reactions indicated the complete
consumption of the alkynes and quantitative fomation of the
triazole products in most cases. A few of the triazole products
precipitated and were separated (by centrifugation) and
redissolved in DMSO.
Org. Lett., Vol. 8, No. 4, 2006
715

Scheme 2. Structure and Synthesis of 14 Azide-Containing
Table 1. IC50 (in µM) of Six Selected Inhibitors
inhibitor PTP1B TC TP YOP LAR λPPase
Blocks
PP1
A13
C10
B1
B11
B4
4.7
58.1
17.2
29.6
19.8
34.5
23.3
17.5
25.5
21.7
49.4
64.5
120.1 >450 >500
24.1 >450 >500
23.9 >450 >500
82.2 >450 >500
176.8 >450 >500
74.6 >450 >500
>750
>750
>750
>750
>750
>750
A5
other hand, this may also underscore the feasibility of our
strategy for potential discovery of inhibitors against specific
PTPs with high potency and specificity. Six hits (A13, C10,
B1, B11, B4, and A5) were chosen, one of which, e.g., A13,10
was a potential potent and specific PTP1B inhibitor, and
further characterized to obtain their IC50 values against each
of the six enzymes. As shown in Table 1, A13 indeed
inhibited PTP1B with an IC50 of 4.7 µM, which is compa-
rable to that of the original Abbott’s inhibitor. More
importantly, it is approximately 5 and 25 times more selective
toward PTP1B than the other two PTPs, TCPTP and YOP,
respectively, and is >100 times more potent than against
LAR, λPPase, and PP1. Another noticeable hit is C10, which
showed good inhibition toward TCPTP (IC50 ) 17.5 µM)
with some degree of selectivity over other phosphatases.
In conclusion, we have developed the first “click chem-
istry” approach for rapid assembly of a bidentate, small
molecule library against protein tyrosine phosphatases (PTPs).
Compounds synthesized from this method are of high quality,
enabling direct in situ screening and high-throughput iden-
tification of potential PTP inhibitors. The strategy is char-
acterized by its modular nature, thus allowing easy adaption
for screening of other PTPs and multivalent proteins, e.g.,
kinases. We identified a specific PTB1B inhibitor with
moderate inhibition (e.g., A13; IC50 ) 4.7 µM) comparable
to that of Abbott’s inhibitor, reported to be one of the most
promising drug candidates against PTP1B to date.8a We
expect that our lead compound should possess favorable
pharmacokinetic properties similar to those for Abbott’s
inhibitor. In vivo experiments are currently being conducted
to confirm this and will be reported in due course.
The 66-member library was next screened directly against
six different phosphatases, including four PTPs (i.e, PTB1B,
TCPTP, YOP, and LAR), a dual-specific phosphatase (i.e.,
λPPase), and a serine/threonine phosphatase (i.e., PP1). A
previously reported fluorescence-based phosphatase assay
was used throughout.9 Full details of the inhibition assays
and profiles of the 66 inhibitors (plus controls with A-E and
1-14) against all six enzymes were shown in the Supporting
Information. Selected results were summarized in Figure 2
Acknowledgment. Funding support was provided by
National University of Singapore (NUS) and the Agency for
Science, Technology, and Research (A*STAR) of Singapore.
Figure 2. Potent inhibitors identified from the screening.
Supporting Information Available: Experimental details
and characterizations of compounds. This material is free
of charge via the Internet at http://pubs.acs.org.
and Table 1. A majority of the library members showed some
degree of inhibition against three of the four PTPs, namely,
PTB1B, TCPTP, and YOP. This is expected when one
considers that the library was purposefully designed to
contain a core group which structurally resembles the original
Abbott’s inhibitor (as shown in Figure 1). On the other hand,
none of the library members showed any inhibition against
LAR, λPPase, and PP1 until they were at very high
concentrations (>450 µM of inhibitor). For LAR, a protein
tyrosine phosphatase, this is somewhat surprising. On the
OL052895W
(9) Zhu, Q.; Huang, X.; Chen, G. Y. J.; Yao, S. Q. Tetrahedron Lett.
2004, 45, 707-710.
(10) Of the 14 azides used (Scheme 2), two have chiral centers (i.e., 13
and 14). 14 was synthesized from the corresponding chiral amino acid
(Supporting Information). 13 was made from its racemic starting material.
No attempt was made to isolate the two enantiomers in A13. It is possible
that one enantiomer may be more potent than the other.
716
Org. Lett., Vol. 8, No. 4, 2006