Fragment-based substrate activity screening method for the identification of potent inhibitors of the Mycobacterium tuberculosis phosphatase PtpB

Article abstract of DOI:10.1021/ja0727520

A new substrate-based fragment approach for the identification of novel PTP inhibitors is presented. This method was applied to Mycobacterium tuberculosis PtpB, a promising new target for the treatment of tuberculosis. This resulted in the development of the most potent PtpB inhibitor reported to date (0.22 μM) with low molecular weight and good selectivity against a panel of other protein tyrosine phosphatases. Copyright

Full text of DOI:10.1021/ja0727520

Published on Web 07/18/2007
Fragment-Based Substrate Activity Screening Method for the Identification of
Potent Inhibitors of the Mycobacterium tuberculosis Phosphatase PtpB
Matthew B. Soellner,† Katherine A. Rawls,† Christoph Grundner,‡ Tom Alber,‡ and
Jonathan A. Ellman*,†
Departments of Chemistry and Molecular and Cell Biology, UniVersity of California, Berkeley, California 94720
Received April 26, 2007; E-mail: jellman@berkeley.edu
Scheme 1. Identification of Phosphatase Inhibitors by Substrate
Activity Screening
The pharmaceutical industry has invested considerable resources
in the production and high throughput screening (HTS) of large
compound libraries to identify drug-like molecules that interact with
clinically relevant biomolecular targets. However, for intractable
targets, such as protein tyrosine phosphatases (PTPs), this approach
has had only limited success.1 Alternative fragment-based ap-
proaches have therefore been developed wherein collections of low-
molecular-weight molecules are screened to identify modest affinity
binders that are subsequently optimized.2
The PtpB protein tyrosine phosphatase encoded by Mycobacte-
rium tuberculosis (M. tb.) is a promising target for new tuberculosis
(TB) drugs.3 M. tb. is one of the three deadliest international
pathogens, yet the last new TB drug was introduced in the 1960s.
Inhibitors against new targets are needed to shorten the 6 month
course of the current standard treatment, cure multi-drug-resistant
infections, and target persistent bacteria.4 PtpB is a secreted
virulence factor that functions within human macrophages.5 Because
a waxy coat impenetrable to many antibiotics surrounds the M. tb.
bacilli, localization outside the bacterial cell makes PtpB an
especially accessible therapeutic target. Here, we have used a new
fragment-based approach termed substrate activity screening (SAS)
to identify selective and low-molecular-weight PtpB inhibitors with
submicromolar inhibitory activity.6
Scheme 2. Spectrophotometric Method for Detection of
Phosphatase Substrates
Our SAS method for identification of phosphatase inhibitors
consists of three steps (Scheme 1): (1) a library of O-aryl
phosphates with diverse, low molecular weight O-aryl groups is
screened to identify phosphatase substrates using a simple spec-
trophotometric-based assay; (2) the identified O-aryl phosphate
substrates are optimized by rapid analogue synthesis and evaluation;
and (3) the optimized substrates are converted to inhibitors by direct
replacement of the phosphate with known phosphate isosteres.
Step 1: A library of 140 O-aryl phosphates was prepared by
solution-phase parallel synthesis from a diverse set of 140 phenol
starting materials. The phenols were selected using hierarchical 2D
extended connectivity analysis from thousands of commercially
available phenols with molecular weights below 300 Da.7 Each of
the phenols was converted into the corresponding O-aryl phosphates
using phosphorus oxychloride followed by aqueous workup.8 All
O-aryl phosphate library members were purified by preparative-
scale reversed-phase HPLC and assayed for purity using LCMS
and NMR spectroscopy.
bance from 330 nm for substrate 1 to 360 nm for product 2, and
thus this assay can be used to continuously monitor the kinetics of
Pi released by phosphatase-catalyzed hydrolysis of O-aryl phosphate
substrates. This coupled assay was adapted for screening in 96-
well plates and spectrophotometric plate readers to enable high
throughput screening of the O-aryl phosphate library.
Importantly, our substrate-based assay eliminates all false
positives commonly found in HTS assays, including false positives
due to irreversible inactivation of the protein target, protein
precipitation, and aggregation of putative inhibitors. Because
inhibition of PNP by a putative phosphatase substrate would lead
to false negatives in our assay, as a control, a subset of substrates
was evaluated for inhibition of PNP and none were found to inhibit
PNP at all concentrations examined.
The 140-member O-aryl phosphate library was then screened
against PtpB using a fast and sensitive continuous spectrophoto-
metric coupled assay method based upon the quantitation of
inorganic phosphate (Pi) released into solution (Scheme 2).9 Purine
nucleoside phosphorylase (PNP) catalyzes the conversion of Pi and
nucleoside 1 to ribose 1-phosphate and 2 (Scheme 2). Substrate
turnover results in a spectrophotometric shift in maximum absor-
Step 2: Screening the initial library of 140 O-aryl phosphates
yielded several promising fragment substrates with greatly improved
KM values relative to the standard assay substrate p-nitrophenyl
phosphate 9 (Figure 1 contains a subset of the initial fragment
results). Of note, the biphenyl scaffold 5 is regarded as a “privileged
† Department of Chemistry.
‡ Department of Molecular and Cell Biology.
9
10.1021/ja0727520 CCC: $37.00 © 2007 American Chemical Society
J. AM. CHEM. SOC. 2007, 129, 9613-9615
9613

C O M M U N I C A T I O N S
Table 1. Inhibitors of M. tb. PtpB
Figure 1. Selected initial substrate hits obtained against PtpB.
a Ki values represent the average of at least six independent measurements
and at least two enzyme concentrations.
Table 2. Selectivity of Inhibitor 19 against a Panel of PTPs
Figure 2. Selected hits from focused library of biphenyl phosphates.
PtpB
PtpA
VHR
TC-Ptp
CD45
LAR
Ki, µM 0.22 ( 0.03 >50 >50 >50 7.70 ( 0.31 21.4 ( 0.78
selectivity >225 >225 >225 35 98
forward to prepare than corresponding inhibitors containing phos-
phate mimetics.
A variety of phosphate isosteres have been reported and could
be introduced in place of the phosphate,11 but many contain at least
two acidic sites and lead to inhibitors with poor cellular perme-
ability.12 We therefore selected two promising monoacidic phos-
phate isosteres, isothiazolidinone13 and isoxazole carboxylic acid,14
which have effectively been used in PTP1B inhibitors with good
cell permeability and activity.
A series of inhibitors was prepared by modified literature
procedures (see Supporting Information), and Ki values against PtpB
were determined using the standard inhibition assay for phos-
phatases with para-nitrophenyl phosphate (pNPP) as the chro-
mogenic substrate.15 The assay was performed with 0.025%
Triton-X 100 detergent, and inhibition was found to be independent
of both enzyme and detergent concentrations. Additionally, the Ki
curves did not display any unusual steepness16 (see Figure 3 for an
initial rate plot for the most potent inhibitor). Together, these
controls demonstrate the observed inhibition is specific and not due
to compound aggregation or micelle formation, which are common
assay artifacts.16
Figure 3. Plot of averaged initial rate versus log [19] used to determine
the averaged Ki for inhibitor 19.
structure” with drug-like properties, and therefore it was the focus
of further optimization.10
To explore the substrate activity relationships (SAR) of the
biphenyl fragment 5, a small, focused library (45 members) of
biphenyl phosphates was synthesized and screened against PtpB.
Determination of the KM values provided clear SAR as represented
by the select substrates shown in Figure 2. The m-trifluoromethyl-
substituted biphenyl phosphate 12 showed a >4-fold decrease in
KM relative to substrate 5. Likewise, the p-fluoro-substituted
derivative 11 conferred a ∼2-fold decrease in KM. Importantly,
merging these two substituents provided substrate 13 with >6-fold
decreased Michaelis constant. Finally, the 2-isopropyl substituent
from the initial fragment library (7, Figure 1) was merged with the
disubstituted biphenyl hit 13 to provide substrate 14 with >13-
fold improved apparent affinity.
Step 3: The final step of this method involves direct conversion
of the optimal substrates to inhibitors by replacement of the
phosphate functionality with a phosphate isostere. Introducing the
phosphate isostere at this later stage greatly facilitates optimization
efforts because phosphate substrates are inherently more straight-
Compounds 15 and 16 allow direct comparison of the selected
pharmacophores. Isothiazolidinone 15 has a Ki of 8.8 µM as a
racemic mixture, while isoxazole carboxylic acid 16 has a Ki of
2.5 µM (Table 1). Due to the modest increase in potency, the
9
9614 J. AM. CHEM. SOC. VOL. 129, NO. 31, 2007

C O M M U N I C A T I O N S
Z. K.; Wilson, D. P.; Follows, B. C.; Binnun, E.; Joseph-McCarthy, D.;
Foreman, K.; Erbe, D. V.; Zhang, Y. L.; Tam, S. K.; Tam, S. Y.; Lee, J.
Bioorg. Med. Chem. 2006, 14, 2162-2177.
isoxazole carboxylic acid isostere was chosen for the synthesis of
all additional inhibitor analogues.17
Addition of the isopropyl group at R2, which was observed to
be a favorable element for binding in substrate 7, provided inhibitor
17 with a Ki of 850 nM. Inhibitor 18 with the fluoro group at R3
was prepared based upon SAR obtained from the small, focused
biphenyl substrate library (Figure 2) and showed further increased
potency, with a Ki of 500 nM. Finally, replacement of the isopropyl
group of 18 with a cyclohexyl group decreased the Ki to 220 nM,
as predicted from substrate 4 (Figure 3).
(2) For reviews on fragment-based screening, see: (a) Erlanson, D. A.;
McDowell, R. S.; O’Brien, T. J. Med. Chem. 2004, 47, 3463-3482. (b)
Carr, R. A. E.; Congreve, M.; Murray, C. W.; Rees, D. C. Drug DiscoVery
Today 2005, 10, 987-992. (c) Hajduk, P. J.; Greer, J. Nat. ReV. Drug
DiscoVery 2007, 6, 211-219.
(3) (a) Greenstein, A. E.; Grundner, C.; Echols, N.; Gay, L. M.; Lombana,
T. N.; Miecskowski, C. A.; Pullen, K. E.; Sung, P. Y.; Alber, T. J. Mol.
Microbiol. Biotechnol. 2005, 9, 167-181. (b) Singh, R.; Singh, A.; Tyagi,
A. K. Tuberculosis (Edinb.) 2005, 85, 325-335.
(4) For an example of a promising tuberculosis drug in development, see:
Andries, K.; et al. Science 2005, 307, 223-227.
It has been shown previously that, for reversible difluorophos-
phonic acid inhibitors of PTP1, inhibitor Ki values correlate to
substrate KM values but not to kcat/KM values.18 Indeed, our inhibitors
exhibit the same trend, with good correlation between substrate
(5) (a) Koul, A.; Choidas, A.; Treder, M.; Tyagi, A. K.; Drlica, K.; Singh,
Y.; Ullrich, A. J. Bacteriol. 2000, 182, 5425-5432. (b) Singh, R.; Rao,
V.; Shakila, H.; Gupta, R.; Khera, A.; Dhar, N.; Singh, A.; Koul, A.;
Singh, Y.; Naseema, M.; Narayanan, P. R.; Paramasivan, C. N.; Ra-
manathan, V. D.; Tyagi, A. K. Mol. Microbiol. 2003, 50, 751-762.
(6) For SAS methodology applied to the development of protease inhibitors,
see: (a) Wood, W. J. L.; Patterson, A. W.; Tsuruoka, H.; Jain, R. K.;
Ellman, J. A. J. Am. Chem. Soc. 2005, 127, 15521-15527. (b) Patterson,
A. W.; Wood, W. J. L.; Hornsby, M.; Lesley, S.; Spraggon, G.; Ellman,
J. A. J. Med. Chem. 2006, 49, 6298-6307. (c) Salisbury, C. M.; Ellman,
J. A. ChemBioChem 2006, 7, 1034-1037. (d) Inagaki, H.; Tsuruoka, H.;
Hornsby, M.; Lesley, S. A.; Spraggon, G.; Ellman, J. A. J. Med. Chem.
2007, 50, 2693-2699.
(7) Phenols were clustered using: Pipeline Pilot (Server version 4.1.0.200).
Developed and distributed by SciTegic, C. D., Suite 401, San Diego, CA
92123-1365; http://www.scitegic.com.
(8) King, E. J.; Nicholson, T. F. Biochem. J. 1939, 33, 1182-1184.
(9) Webb, M. R. Proc. Natl. Acad. Sci. U.S.A. 1992, 89, 4884-4887.
(10) (a) Hajduk, P. J.; Bures, M.; Praestgaard, J.; Fesik, S. W. J. Med. Chem.
2000, 43, 3443-3447. (b) Horton, D. A.; Bourne, G. T.; Smythe, M. L.
Chem. ReV. 2003, 103, 893-930.
(11) For general references and reviews, see: (a) Zhang, Z.-Y. Annu. ReV.
Pharmacol. Toxicol. 2002, 42, 209-234. (b) Rye, C. S.; Baell, J. B. Curr.
Med. Chem. 2005, 12, 3127-3141. (c) Kotoris, C. C.; Chen, M.-J.; Taylor,
S. D. Bioorg. Med. Chem. Lett. 1998, 8, 3275-3280.
(12) Hu, X. Bioorg. Med. Chem. Lett. 2006, 16, 6321-6327.
(13) Combs, A. P.; et al. J. Med. Chem. 2005, 48, 6544-6548.
(14) Liu, G.; Xin, Z.; Pei, Z.; Hajduk, P. J.; Abad-Zapatero, C.; Hutchins, C.
W.; Zhao, H.; Lubben, T. H.; Ballaron, S. J.; Haasch, D. L.; Kaszubska,
W.; Rondinone, C. M.; Trevillyan, J. M.; Jirousek, M. R. J. Med. Chem.
2003, 46, 4232-4235.
KM values and inhibitor Ki values for the isoxazole carboxylic acid
series of inhibitors. Moreover, no correlation was observed between
kcat/KM values and inhibitor Ki values (Figures 1 and 2). On the
basis of transition state theory, these results are consistent with the
selected phosphate mimetics binding as substrate rather than
transition state analogues.19
Isoxazole inhibitor 19 with a Ki of 220 nM represents the most
potent inhibitor of PtpB known in the literature.20 In addition,
inhibitor 19 showed good selectivity against a panel of mycobac-
terial (PtpA) and human PTPs (Table 2). Structural characterization
and biological evaluation of PtpB inhibitor 19 are currently in
progress. The small size of this compound (MW ) 433 Da) leaves
room for additional elaborations to improve its properties.
In conclusion, we have developed a new substrate-based fragment
approach termed SAS for the identification of novel PTP inhibitors.
Application of this method to M. tb. PtpB resulted in the
development of inhibitor 19, the most potent PtpB inhibitor reported
to date.20 Additionally, the method should be generally applicable
to many different phosphatases.21
(15) Montalibet, J.; Skorey, K. I.; Kennedy, B. P. Methods 2005, 35, 2-8.
(16) (a) McGovern, S. L.; Caselli, E.; Grigorieff, N.; Shoichet, B. K. J. Med.
Chem. 2002, 45, 1712-1722. (b) Feng, B. Y.; Shoichet, B. K. Nat.
Protocols 2006, 1, 550-553.
(17) To ensure that binding was not due to pharmacophore alone, the isoxazole
inhibitor corresponding to substrate 9 was synthesized and evaluated for
inhibition against PtpB. This isoxazole inhibitor has a Ki of >150 µM,
clearly indicating that the pharmacophore by itself is not sufficient for
tight binding.
(18) (a) Montserat, J.; Chen, L.; Lawrence, D. S.; Zhang, Z. Y. J. Biol. Chem.
1996, 271, 7868-7872. (b) Taylor, S. D.; Kotoris, C. C.; Dinaut, A. N.;
Wang, Q. P.; Ramachandran, C.; Huang, Z. Bioorg. Med. Chem. 1998, 6,
1457-1468.
Acknowledgment. We thank Karen Dehnert for help with
preparation of the focused biphenyl library, Dr. Jerome Hert for
assistance with 2D clustering of commercially available phenols,
and Dr. Markus Seeliger for helpful discussions. This work was
supported by the NIH grants GM54051 to J.A.E. and AI068135 to
T.A. M.B.S. was supported by NIH postdoctoral fellowship
GM080025, C.G. was supported by a postdoctoral fellowship from
the American Lung Association of California, and K.A.R. was
supported by NIH training grant GM066698.
(19) Mader, M. M.; Bartlett, P. A. Chem. ReV. 1997, 97, 1281-1301.
(20) (a) Noren-Muller, A.; Reis-Correa, I.; Prinz, H.; Rosenbaum, C.; Saxena,
K.; Schwalbe, H. J.; Vestweber, D.; Cagna, G.; Schunk, S.; Schwarz, O.;
Schiewe, H.; Waldmann, H. Proc. Natl. Acad. Sci. U.S.A. 2006, 103,
10606-10611. (b) Grundner, C.; Perrin, D.; Hooft van Huijsduijnen, R.;
Swinnen, D.; Gonzalez, J.; Gee, C. L.; Wells, T. N.; Alber, T. Structure
2007, 15, 499-509.
Supporting Information Available: Complete experimental details,
spectral data for all compounds described, and full author list for refs
4 and 13. This material is available free of charge via the Internet at
http://pubs.acs.org.
(21) Ser/Thr phosphatases should also be accessible targets using an initial
library of diverse O-alkyl phosphates.
References
(1) Workers at Wyeth have definitively reported on the seriousness of this
problem: Moretto, A. F.; Kirincich, S. J.; Xu, W. X.; Smith, M. J.; Wan,
JA0727520
9
J. AM. CHEM. SOC. VOL. 129, NO. 31, 2007 9615