Targeting inactive enzyme conformation: Aryl diketoacid derivatives as a new class of PTP1B inhibitors

Article abstract of DOI:10.1021/ja8068177

There has been considerable interest in protein tyrosine phosphatase 1B (PTP1B) as a therapeutic target for diabetes, obesity, as well as cancer. Identifying inhibitory compounds with good bioavailability is a major challenge of drug discovery programs targeted toward PTPs. Most current PTP active site-directed pharmacophores are negatively charged pTyr mimetics which cannot readily enter the cell. This lack of cell permeability limits the utility of such compounds in signaling studies and further therapeutic development. We identify aryl diketoacids as novel pTyr surrogates and show that neutral amide-linked aryl diketoacid dimers also exhibit excellent PTP inhibitory activity. Kinetic studies establish that these aryl diketoacid derivatives act as noncompetitive inhibitors of PTP1B. Crystal structures of ligand-bound PTP1B reveal that both the aryl diketoacid and its dimeric derivative bind PTP1B at the active site, albeit with distinct modes of interaction, in the catalytically inactive, WPD loop open conformation. Furthermore, dimeric aryl diketoacids are cell permeable and enhance insulin signaling in hepatoma cells, suggesting that targeting the inactive conformation may provide a unique opportunity for creating active site-directed PTP1B inhibitors with improved pharmacological properties.

Full text of DOI:10.1021/ja8068177

Published on Web 11/14/2008
Targeting Inactive Enzyme Conformation: Aryl Diketoacid
Derivatives as a New Class of PTP1B Inhibitors
Sijiu Liu,† Li-Fan Zeng,§ Li Wu,† Xiao Yu,† Ting Xue,§ Andrea M. Gunawan,‡
Ya-Qiu Long,*,§ and Zhong-Yin Zhang*,†,‡
Department of Biochemistry and Molecular Biology and Chemical Genomics Core Facility,
Indiana UniVersity School of Medicine, 635 Barnhill DriVe, Indianapolis, Indiana 46202, and
State Key Laboratory of Drug Research, Shanghai Institute of Materia Medica, Chinese
Academy of Sciences, 555 Zuchongzhi Road, Shanghai 201203, China
Received August 28, 2008; E-mail: zyzhang@iupui.edu; yqlong@mail.shcnc.ac.cn
Abstract: There has been considerable interest in protein tyrosine phosphatase 1B (PTP1B) as a therapeutic
target for diabetes, obesity, as well as cancer. Identifying inhibitory compounds with good bioavailability is
a major challenge of drug discovery programs targeted toward PTPs. Most current PTP active site-directed
pharmacophores are negatively charged pTyr mimetics which cannot readily enter the cell. This lack of
cell permeability limits the utility of such compounds in signaling studies and further therapeutic development.
We identify aryl diketoacids as novel pTyr surrogates and show that neutral amide-linked aryl diketoacid
dimers also exhibit excellent PTP inhibitory activity. Kinetic studies establish that these aryl diketoacid
derivatives act as noncompetitive inhibitors of PTP1B. Crystal structures of ligand-bound PTP1B reveal
that both the aryl diketoacid and its dimeric derivative bind PTP1B at the active site, albeit with distinct
modes of interaction, in the catalytically inactive, WPD loop open conformation. Furthermore, dimeric aryl
diketoacids are cell permeable and enhance insulin signaling in hepatoma cells, suggesting that targeting
the inactive conformation may provide a unique opportunity for creating active site-directed PTP1B inhibitors
with improved pharmacological properties.
Introduction
animal models demonstrate that a reduction in PTP1B leads to
decreases in adipose tissue mass, plasma insulin, and blood
Protein tyrosine phosphatases (PTPs) constitute a large family
of enzymes, which are crucial modulators of tyrosine phospho-
rylation-dependent cellular events.1 Malfunction in PTP activity
is associated with many human diseases, including cancer,
diabetes/obesity, and autoimmune disorders.2 Among members
of the PTP superfamily, PTP1B is considered one of the best-
validated targets for therapeutic development. Biochemical and
genetic evidence indicate that PTP1B plays a key role in
regulating body weight, glucose homeostasis, and energy
expenditure by acting as a key negative regulator of insulin and
leptin receptor mediated signaling pathways. PTP1B-deficient
mice display increased insulin sensitivity and improved glycemic
control, and are resistant to diet-induced obesity.3,4 Moreover,
studies with PTP1B antisense oligonucleotides in diabetes
glucose levels.5 These findings suggest that inhibition of PTP1B
represents an effective strategy to combat metabolic syndromes
such as type 2 diabetes and obesity.
Besides having a role in dampening insulin- and leptin-
mediated processes, PTP1B also augments signaling downstream
of growth factor receptors and integrins. To that end, PTP1B
can remove the inhibitory phosphate from the C-terminus of
Src, thereby promoting Src kinase activation.6-8 In addition,
PTP1B dephosphorylates the scaffolding adapter protein p62DOK
,
leading to activation of the Ras-ERK pathway.9 Given that
PTP1B is capable of promoting both the Src kinase and Ras/
Erk pathways, which are major components in HER2/Neu
signaling, PTP1B may function as an oncogene in the context
of breast cancer. Interestingly, PTP1B is up-regulated in HER2/
† Department of Biochemistry and Molecular Biology, Indiana University
School of Medicine.
‡ Chemical Genomics Core Facility, Indiana University School of
Medicine.
(5) Zinker, B. A.; Rondinone, C. M.; Trevillyan, J. M.; Gum, R. J.;
Clampit, J. E.; Waring, J. F.; Xie, N.; Wilcox, D.; Jacobson, P.; Frost,
L.; Kroeger, P. E.; Reilly, R. M.; Koterski, S.; Opgenorth, T. J.; Ulrich,
R. G.; Crosby, S.; Butler, M.; Murray, S. F.; McKay, R. A.; Bhanot,
S.; Monia, B. P.; Jirousek, M. R. Proc. Natl. Acad. Sci. U.S.A. 2002,
99, 11357–11362.
§ State Key Laboratory of Drug Research, Shanghai Institute of Materia
Medica.
(1) Tonks, N. K. Nat. ReV. Mol. Cell Biol. 2006, 7, 833–846.
(2) Zhang, Z.-Y. Curr. Opin. Chem. Biol. 2001, 5, 416–423.
(3) Elchelby, M.; Payette, P.; Michaliszyn, E.; Cromlish, W.; Collins, S.;
Loy, A. L.; Normandin, D.; Cheng, A.; Himms-Hagen, J.; Chan, C. C.;
Ramachandran, C.; Gresser, M. J.; Tremblay, M. L.; Kennedy, B. P.
Science 1999, 283, 1544–1548.
(6) Bjorge, J. D.; Pang, A.; Fujita, D. J. J. Biol. Chem. 2000, 275, 41439–
41446.
(7) Cheng, A.; Bal, G. S.; Kennedy, B. P.; Tremblay, M. L. J. Biol. Chem.
2001, 276, 25848–25855.
(4) Klaman, L. D.; Boss, O.; Peroni, O. D.; Kim, J. K.; Martino, J. L.;
Zabolotny, J. M.; Moghal, N.; Lubkin, M.; Kim, Y. B.; Sharpe, A. H.;
Stricker-Krongrad, A.; Shulman, G. I.; Neel, B. G.; Kahn, B. B. Mol.
Cell. Biol. 2000, 20, 5479–5489.
(8) Liang, F.; Lee, S.-Y.; Liang, J.; Lawrence, D. S.; Zhang, Z.-Y. J. Biol.
Chem. 2005, 280, 24857–24863.
(9) Dube, N.; Cheng, A.; Tremblay, M. L. Proc. Natl. Acad. Sci. USA
2004, 101, 1834–1839.
9
10.1021/ja8068177 CCC: $40.75
2008 American Chemical Society
J. AM. CHEM. SOC. 2008, 130, 17075–17084 17075

A R T I C L E S
Liu et al.
Neu-transformed cells10 and 90% of all breast tumors overex-
press both HER2/Neu and PTP1B.11 More recent studies reveal
that crossbreeding transgenic mice expressing activated HER2/
Neu with PTP1B-/- mice caused delayed HER2/Neu-induced
mammary tumor development and reduced the incidence of lung
metastases.12,13 These data suggest that PTP1B-specific inhibi-
tors may qualify as an effective treatment for breast cancer.
Given the compelling evidence linking PTP1B to multiple
human diseases, major efforts have been initiated by both the
pharmaceutical industry and academia to target PTP1B for
therapeutic development. Contrary to the widely held belief that
PTPs are challenging targets for small-molecule inhibitor
development, numerous potent and selective PTP1B inhibitors
have been described.14 However, most of the reported com-
pounds contain negatively charged nonhydrolyzable phospho-
tyrosine (pTyr) mimetics, due primarily to the highly positively
charged nature of the active site evolved to bind pTyr. Con-
sequently, poor membrane permeability has limited further
advancement of such compounds as drug candidates. Thus, one
challenge which has previously delayed realization of PTP1B-
based small-molecule therapeutics has been the identification
of novel chemical entities with improved physicochemical
properties and bioavailability.
in anhydrous THF at 0 °C was added dropwise aryl methyl ketone
(1.0 equiv) in DME. The resulting orange-yellow mixture was
stirred at room temperature for a maximum of 1.5 h. The reaction
was quenched with a 1.0 M HCl aqueous solution and the
compound was extracted with CH2Cl2. The combined organic layers
were washed with saturated aqueous NaHCO3 and brine, respec-
tively, dried over anhydrous Na2SO4 and concentrated under
vacuum. Purification by silica gel flash chromatography provided
the desired m-/o-/p-substituted phenyl-4-oxo-2-hydroxy-2-butenoic
acid methyl ester. Subsequent hydrolysis of the resulting aryl
diketoacid methyl ester readily furnished the corresponding aryl
diketoacid by 1 N NaOH in MeOH/THF (1:1) at room temperature.
Synthesis and physicochemical data of the compounds LZP3-4,
LZP6-9 and LZP37-38 were reported previously.17,18
4-(3-(Dibenzylamino)phenyl)-2,4-dioxobutanoic acid (LZP25):
yellow powder, yield 89%. 1H NMR (300 MHz, CDCl3): δ
7.27-7.24 (m, 14H), 4.68 (s, 4H). EI-MS (m/z): 315 (M - 72)+.
Analytical RP-HPLC tR ) 11.1 min (gradient 30-90% of solvent
C over 20 min, purity 100%), tR ) 9.2 min (gradient 30-90% of
solvent B over 20 min, purity 100%).
2,4-Dioxo-4-phenylbutanoic acid (LZP-36): yellow solid, yield
90%. 1H NMR (400 MHz, CDCl3): δ 8.03-7.99 (m, 2H),
7.67-7.61 (m, 1H), 7.55-7.50 (m, 2H), 7.18 (s, 1H); EI-MS (m/
z): 192 (M+).
2,4-Dioxo-4-(4-(trifluoromethyl)phenyl)butanoic acid (LZP39):
1
white needle solid, yield 87.0%. H NMR (400 MHz, CDCl3): δ
We report the characterization of aryl diketoacids as novel
pTyr surrogates for PTP inhibitor development. Surprisingly,
we discovered that amide-linked aryl diketoacid dimers, which
lack any formal charge, also exhibit PTP inhibitory activity.
X-ray crystallographic analyses of ligand-bound PTP1B struc-
tures revealed that although monomeric and dimeric aryl
diketoacid occupy and interact with the enzyme active site in
distinct manner, both compounds stabilize PTP1B in its inactive,
WPD loop open conformation. Importantly, dimeric aryl dike-
toacids are cell permeable and enhance insulin signaling in
hepatoma cells. These properties suggest that it is possible to
develop uncharged PTP1B active site-directed inhibitors with
improved in ViVo efficacy, thereby facilitating the development
of PTP1B-based therapeutics for the treatment of diabetes,
obesity, and breast cancer. Furthermore, targeting the inactive
conformation may constitute a general strategy for PTP inhibitor
design.
8.11 (d, 2H, J ) 8.0 Hz); 7.82 (d, 2H, J ) 8.0 Hz); 7.20 (s, 1H).
EI-MS (m/z, %): 260 (M+, 3.0); 241 (4.0); 215 (100.0); 173 (45.0);
145 (37.0). HR-EIMS calcd for C11H7F3O4: 260.0296, found:
260.0301. Anal. Calcd for C11H7F3O4 · 0.2CH2Cl2 · 0.8CH3OH: C
47.60, H 3.53; Found C 47.33, H 3.34.
1,1′-(Piperazine-1,4-diyl)bis(4-(3-(dibenzylamino)phenyl)bu-
tane-1,2,4-trione) (LZP40). A drop of DMF was added to a turbid
solution of 1.0 eq of LZP25 dissolved in 10 mL of dichloromethane
at 0 °C. Next, 1 mL of SOCl2 was added to the solution. The
reaction mixture was stirred for 2 h at room temperature, after which
the solution was condensed in vaccum. The residue was dissolved
in 10 mL of dichloromethane followed by addition of 0.5 eq of
piperazine and a drop of pyridine. This solution was stirred for
another 2 h. Work up and dried by anhydrous sodium sulfate.
Purification from chromatography (CH2Cl2/CH3OH )20:1) gives
LZP40 as light yellow solid in yield of 65%. 1H NMR (300 MHz,
CDCl3): δ 7.36-7.31 (m, 10H), 7.28-7.24 (m, 10 H), 7.22 (m,
6H), 6.91 (d, 2H, J ) 6.9 Hz), 6.47 (s, 2H), 4.69 (s, 8H), 3.70 (m,
8H). ESI-MS m/z: calc. 824.9, found 823.2 (M - 1)-, 863.3
(M+K)+. m.p. 105-106 °C. tR ) 28.7 min. (30-90% of solvent
B over 30 min, purity 99%). tR )24.1 min (30-90% of solvent C
over 28 min, purity 98%).
tert-Butyl 4-(4-(3-(benzyloxy)phenyl)-2,4-dioxobutanoyl) pip-
erazine-1-carboxylate (LZP70). A solution consisting of com-
pound LZP4 (75 mg, 0.25 mmol), EDCI (73 mg, 0.38 mmol),
DIPEA (66 µL, 0.38 mmol) and HOAt (51 mg, 0.38 mmol) in 10
mL of dry dichloromethane (1.5 mL) was stirred at room temper-
ature for half an hour. To this solution was added tert-butyl
piperazine-1-carboxylate (50 mg, 0.27 mmol) in dry THF. The
reaction mixture was stirred for 12 h at room temperature and then
poured into ice-water and extracted with dichloromethane. The
combined organic layers were washed with water and brine and
dried over anhydrous Na2SO4. The concentration provided the
residue which was purified by chromatography using PE/EtOAc
(1:1) as an eluent, which gives compound LZP70 as a brown
powder with a yield of 47%. 1H NMR (300 MHz, CDCl3): δ 7.53
Materials and Methods
Materials. Polyethylene glycol (PEG3350) and buffers for
crystallization were purchased from Hampton Research Co. p-
Nitrophenyl phosphate (pNPP) was purchased from Fluke Co.
Dithiothreitol (DTT) was provided by Fisher (Fair Lawn, NJ).
Dulbecco’s modified Eagle’s medium (DMEM) and fetal bovine
serum were purchased from ATCC. All of other chemicals and
reagents were of the highest commercially available grade. Poly-
clonal insulin receptor ꢀ-subunit dual-phospho-specific (pY1162/
pY1163) antibody was purchased from BioSource International
(Camarillo, CA). Human insulin, sodium orthovanadate, iodoacetic
acid (IAA) and Nonidet P-40 were from Sigma (St. Louis, MO).
Synthesis of Aryl Diketoacid Derivatives.15,16 To a stirred
mixture of tert-BuONa (2.5 equiv) and dimethyl oxalate (2.0 equiv)
(10) Zhai, Y. F.; Beittenmiller, H.; Wang, B.; Gould, M. N.; Oakley, C.;
Esselman, W. J.; Welsch, C. W. Cancer Res. 1993, 53, 2272–2278.
(11) Wiener, J. R.; Kerns, B. J.; Harvey, E. L.; Conaway, M. R.; Iglehart,
J. D.; Berchuck, A.; Bast, R. C., Jr. J. Natl. Cancer Inst. 1994, 86,
372–378.
(15) Jiang, X.-H.; Song, L.-D.; Long, Y.-Q. J. Org. Chem. 2003, 68, 7555–
7558.
(16) Jiang, X.-H.; Long, Y.-Q. Chin. J. Chem. 2004, 22, 978–983.
(17) Long, Y.-Q.; Jiang, X.-H.; Dayam, R.; Sanchez, T.; Shoemaker, R.;
Sei, S.; Neamati, N. J. Med. Chem. 2004, 47, 2561–2573.
(18) Zeng, L.-F.; Jiang, X.-H.; Sanchez, T.; Zhang, H.-S.; Dayam, R.;
Neamati, N.; Long, Y.-Q. Bioorg. Med. Chem. 2008, 16, 7777–7787.
(12) Julien, S. G.; Dube´, N.; Read, M.; Penney, J.; Paquet, M.; Han, Y.;
Kennedy, B. P.; Muller, W. J.; Tremblay, M. L. Nat. Genet. 2007,
39, 338–346.
(13) Bentires-Alj, M.; Neel, B. G. Cancer Res. 2007, 67, 2420–2424.
(14) Zhang, S.; Zhang, Z.-Y. Drug DiscoVery Today 2007, 12, 373–381.
9
17076 J. AM. CHEM. SOC. VOL. 130, NO. 50, 2008

Aryl Diketo Acid-Based PTP1B Inhibitors
A R T I C L E S
Table 1. PTP1B Inhibition by Aryl Diketoacids
-7.37 (m, 8H), 7.17 (m, 1H), 6.58 (s, 1H), 5.11 (s, 2H), 3.66 (br,
4H), 3.51 (br, 4H), 1.47 (s, 9H); EI-MS (m/z): 466 (M+).
tert-Butyl 4-(4-(3-(dibenzylamino)phenyl)-2,4-dioxobutanoyl)
piperazine-1-carboxylate (LZP71). LZP71 was prepared in a
similar fashion as described for LZP70, resulting in a yellow solid
1
with a yield of 50%. H NMR (400 MHz, CDCl3): δ 7.36-7.21
(m, 13H); 6.93 (m, 1H); 6.44 (s, 1H); 4.69 (s, 4H), 3.62 (m, 4H),
3.48 (m, 4H); 1.47(s, 9H). EI-MS (m/z): 555 (M+).
General Analytical Methods. Unless otherwise stated, the H
1
NMR spectra were recorded on a Varian 300-MHz or 400-MHz
spectrometer. Data are reported in parts per million relative to TMS
and are referenced to the solvent. Elemental analyses were obtained
using a Vario EL spectrometer. Melting points (uncorrected) were
determined on a Buchi-510 capillary apparatus. IR spectra were
recorded on Bio-Rad FTS-185 spectrometers. Mass spectrometry
and HRMS (ESI) spectra were obtained on an APEXIII 7.0 T FTMS
mass spectrometer. The degree of rotation was measured by P-1030
(A012360639) automatic polarimeter. Flash column chromatogra-
phy was performed on silica gel H (10-40 µm). Anhydrous solvents
were obtained using a standard procedure.
Analytical RP-HPLC was performed for selected compounds on
two systems, using a Vydac C18 column (10 mm × 250 mm),
monitored by a UV detector at 254 nm. Solvent gradient system 1,
solvent A, 0.05% TFA in water; solvent B, 0.05% TFA in 90%
acetonitrile in water; solvent gradient system 2, solvent A, 0.05%
TFA in water; solvent C, 0.05% TFA in 90% methanol in water
with a gradient over 20 min at a flow rate of 2.5 mL/min.
PTP1B. The catalytic domain of PTP1B (residues 1-321) and
its mutants were expressed in Escherichia coli and purified to
homogeneity as described previously.19 All mutant forms of PTP1B
were generated using the QuickChange kit from Stratagene.
Kinetic Analysis of PTP1B Inhibition by Aryl Diketoacids
and Their Dimers. Compounds were dissolved in DMSO to a
concentration of 10 mM. The effect of each compound on the
PTP1B-catalyzed pNPP hydrolysis was determined at 25 °C and
pH 7.0 in a 200 µL reaction system in a 96-well plate. Each reaction
contained 10 µL of 2 mM compound in DMSO (final concentration
100 µM) and 190 µL assay buffer (50 mM 3,3-dimethylglutarate,
1 mM EDTA, 1 mM DTT, pH 7.0 with an ionic strength of 0.15
M adjusted by addition of NaCl) containing 2 mM pNPP and 10
nM PTP1B. The PTP-catalyzed reaction was started by addition
of the PTP1B. Ten µL of DMSO (i.e., no aryl diketoacid) was used
as a control. The PTP1B-catalyzed hydrolysis of pNPP was mea-
sured by monitoring the absorbance at 405 nm of the product
p-nitrophenol continuously, with a SpectraMAX Plus 384 micro-
plate spectrophotometer (Molecular Devices). The initial rate was
obtained by calculating the slope of the product versus the time
curve. The IC50 value was determined by plotting relative PTP
activity toward pNPP activity versus inhibitor concentration and
fitting to eq 1 using SigmaPlot 10.0.
of 18,000 M-1 cm-1. The inhibition constant and inhibition pattern
were evaluated using the program SigmaPlot 10.0.
Crystallization of PTP1B with LZP6 and LZP25 and X-ray
Data Collection. All crystallization experiments were carried out
at 4 °C using the hanging drop vapor diffusion methods. For
cocrystallization, 20 µL of PTP1B stock (∼7.0 mg/mL) in 100 mM
MES (pH 6.5), 50 mM NaCl, 0.1 mM EDTA, and 3.0 mM DTT
was mixed with 1 µL of LZP6 or LZP25 stock solutions (20 mM
in DMSO). Crystals of PTP1B ·LZP25 and PTP1B·LZP6 were
obtained at 4 °C by vapor diffusion in hanging drops. The protein
drops were equilibrated against a reservoir solution containing 20%
w/v polyethylene glycol 3350, 200 mM magnesium acetate tet-
rahydrate, and 100 mM HEPES buffer (pH 7.7). The crystals of
PTP1B·LZP6 and PTP1B ·LZP25 both belonged to space group
P3121 (Table 3).
For X-ray data collection, the crystals were transferred into 5
µL of cryoprotectant buffer containing 30% w/v polyethylene glycol
3350, 100 mM NaCl, and 100 mM HEPES (pH 7.7), and were
allowed to soak for 30 min. The crystals were then flash-cooled by
liquid nitrogen. X-ray data were collected at 100 K on a Rigaku
RU-300 rotating anode generator (Rigaku/MSC) equipped with
focusing mirrors (MSC/Yale) and an R-AXIS IV++ image plate
detector. Data were processed using the program HKL2000,20 and
the statistics are provided in Table 3.
Vi ⁄V0)IC50 ⁄(IC50 + [I])
(1)
Structural Determination and Refinement. The structure of
PTP1B·LZP25 was solved by molecular replacement using the
program AMoRe.21 The structure of PTP1B (PDB entry code
1PXH),22 without the solvent molecules and the bidentate inhibitor,
was used as a search model. The resulting difference Fourier map
indicated some alternative tracing, which was incorporated into the
model. The map revealed the density for the bound LZP25 in the
active site of PTP1B. The structure was refined to 2.4 Å resolution
with the program CNS,23 first using simulated annealing at 3,000
K, and then alternating positional and individual temperature factor
refinement cycles. The progress of the refinement was evaluated
by the improvement in the quality of the electron density maps,
and the reduced values of the conventional R factor (R ) Σh||Fo| -
In this case, Vi is the reaction velocity when the inhibitor
concentration is [I], V0 is the reaction velocity with no inhibitor,
and IC50 ) Ki + Ki[S]/Km.
Ki Measurement. The PTP1B-catalyzed hydrolysis of pNPP in
the presence of aryl diketoacid was assayed at 25 °C and in the
assay buffer described above. The reaction was initiated by addition
of PTP1B to a reaction mixture (200 µL) containing various
concentrations of pNPP (ranging from 0.2 to 5 Km) and various
fixed concentrations of aryl diketoacid, and quenched by addition
of 50 µL of 5 M NaOH. Nonenzymatic hydrolysis of the substrate
was corrected by measuring the control without addition of enzyme.
The amount of product p-nitrophenol was determined from the
absorbance at 405 nm detected by a SpectraMAX Plus 384
microplate spectrophotometer using a molar extinction coefficient
(20) Otwinowski, Z.; Minow, W. Methods Enzymol. 1997, 276, 307–326.
(21) Navaza, J. Acta Crystallogr. 1994, A50, 157–163.
(22) Sun, J.-P.; Fedorov, A. A.; Lee, S.-Y.; Guo, X.-L.; Shen, K.; Lawrence,
D. S.; Almo, S. C.; Zhang, Z.-Y. J. Biol. Chem. 2003, 278, 12406–
12414.
(19) Puius, Y. A.; Zhao, Y.; Sullivan, M.; Lawrence, D. S.; Almo, S. C.;
Zhang, Z.-Y. Proc. Natl. Acad. Sci. U.S.A. 1997, 94, 13420–13425.
9
J. AM. CHEM. SOC. VOL. 130, NO. 50, 2008 17077

A R T I C L E S
Liu et al.
|Fc||/Σh|Fo|), and the free R factor (3.5% of the reflections omitted
from the refinement).24 Electron density maps were inspected and
the model was modified on an interactive graphics workstation with
the program O.25 Finally, water molecules were added gradually
as the refinement progressed. They were assigned in the |Fo| - |Fc|
difference Fourier maps with a 3δ cutoff level for inclusion in the
model.
The structure of PTP1B ·LZP6 was solved by molecular replace-
ment using the structure of PTP1B ·LZP25 as the search model
(except that LZP25 and solvent molecules were omitted). The map
revealed very good density for the bound LZP6 in the active site
of PTP1B. The structure was refined at 1.9 Å with the program
CNS (Table 3), following a protocol similar to that described above.
Cell Lines and Tissue Culture. Human hepatoma cells (HepG2)
were obtained from ATCC (HB-8065). HepG2 cells were grown
and maintained in Minimum essential Medium (MEM, Eagle) with
2 mM L-glutamine, 1 mM sodium pyruvate and 10% fetal bovine
serum in a 5% CO2 environment.
Figure 1. Structures of the aryl diketoacid derivatives.
Table 2. Inhibition of PTP1B by Amide-Linked Aryl Diketoacid
Dimers
Immunoblotting. Sixty to eighty percent confluent HepG2 cells
were starved for 5 h in MEM without serum. Next, cells were
treated for 20 min with a range of concentrations of LZP25 or LZP6,
followed by stimulation with or without 100 nM insulin for 5 min.
Incubation was terminated by removing the fluid and cells were
washed with cold PBS. Cells were scraped and lysed with lysis
buffer (50 mM Tris-HCl, pH 7.5, 150 mM NaCl, 1% NP-40, 10
mM sodium phosphate, 10 mM NaF, 1 mM EDTA, 2 mM
orthovanadate, 5 mM iodoacetic acid, 5 mg/mL leupeptin, 5 mg/
mL aprotinin, 0.1 mg/mL PMSF, and 1 mM benzamidine). After
30 min lysing on ice, the cell lysate was centrifuged at 13,200 rpm
for 15 min. Total cellular proteins were separated by SDS-PAGE
and transferred electrophoretically to a nitrocellulose membrane,
which was immunoblotted by antiphospho-IRꢀ antibodies. To
confirm the equal sample loading, the membrane was also probed
with anti-ꢀ-Actin antibody. Blots were developed using a Western
blot detection kit (upstate) according to the manufacturer’s instruc-
tions. Western blotting images were analyzed with Image J (NIH).
Results and Discussion
Aryl Diketoacid Derivatives as PTP1B Inhibitors. One of the
most efficient methods for the acquisition of active site-directed,
reversible, potent and selective PTP1B inhibitors is to tether a
nonhydrolyzable pTyr mimetic to an appropriately functional-
ized chemical moiety, which makes it possible to target both
the active site and a unique nearby subpocket.14 Although
numerous pTyr surrogates have been reported,26 they are
generally not drug-like, and usually have limited cell membrane
permeability. Thus, there is continued interest in developing
novel pTyr mimetics with more acceptable pharmacological
properties.27 The aryl diketoacid scaffold has been used suc-
cessfully for the design of HIV-1 integrase inhibitors.28 Given
the structural similarity between aryl diketoacids and pTyr, we
were intrigued to find out whether aryl diketoacids possess any
inhibitory activity against PTP1B. A number of aryl diketoacid
derivatives were synthesized (Figure 1, and Materials and
Methods). The ability of these compounds to inhibit PTP1B-
catalyzed hydrolysis of p-nitrophenyl phosphate (pNPP) was
assessed at pH 7 and 25 °C (for details see Materials and
Methods). As shown in Table 1, simple aryl diketoacids (LZP36,
LZP3, and LZP39) exhibit no inhibitory activity at 100 µM
against PTP1B, while substitutions at the para or meta position
of the phenyl ring with benzyl moieties bring the IC50 values
of compounds LZP4, LZP38 and LZP25 to the low micromolar
range. Clearly, bulky hydrophobic substitutions on the phenyl
ring enhance the affinity of aryl diketoacids for PTP1B. We
then determined whether dimerization of the aryl diketoacids
would improve inhibition of PTP1B activity. Aryl diketoacid
dimer formation was achieved through either a piperazino or
cyclohexane-1,4-trans-diamine linker (Figure 1). To our delight,
the piperazine- or cyclohexane diamine-linked aryl diketoacid
dimers LZP6, LZP7, LZP37, and LZP40 showed improved
PTP1B inhibitory activity over the corresponding monomers
(LZP4, LZP25, and LZP38), with IC50 values in the low
micromolar range (Table 2).
(23) Bru¨nger, A. T.; Adams, P. D.; Clore, G. M.; DeLano, W. L.; Gros,
P.; Grosse-Kunstleve, R. W.; Jiang, J. S.; Kuszewski, J.; Nilges, M.;
Pannu, N. S.; Read, R. J.; Rice, L. M.; Simonson, T.; Warren, G. L.
Acta Crystallogr., Sect. D 1998, 54, 905–921.
(24) Bru¨nger, A. T. Nature 1992, 355, 472–475.
(25) Jones, T. A.; Zou, J. Y.; Cowan, S. W.; Kjeldgaard, G. J. Acta
Crystallogr., Sect. A 1991, 47, 110–119.
(26) Zhang, Z.-Y. Handbook of Cell Signaling; Bradshaw, R.; Dennis, E.,
Eds.; Academic Press: New York, 2003; Vol. 1, pp 677-684.
(27) Combs, A. P.; et al. J. Med. Chem. 2005, 48, 6544–6548.
(28) Pais, G. C.; Burke, T. R. Drugs of the Future 2002, 27, 1101–1111.
To determine the inhibition constant and mode of inhibition
by LZP25 and LZP6, the effect of the inhibitors on PTP1B-
catalyzed reaction was studied at 8 different pNPP and 4 dif-
ferent compound concentrations. As shown by the Line-
9
17078 J. AM. CHEM. SOC. VOL. 130, NO. 50, 2008

Aryl Diketo Acid-Based PTP1B Inhibitors
A R T I C L E S
Table 3. Crystallographic Data and Refinement Statistics
by LZP25 and LZP6, we decided to solve the three-dimensional
structure of the complexes.
PTP1B·LZP25
PTP1B·LZP6
Structures of the PTP1B ·LZP25 Complex. Using conditions
similar to those described previously,19,22 we produced crystals
of the PTP1B ·LZP25 complex that diffracted to 2.4 Å resolution
(Table 3). The structures were solved by molecular replacement.
The final model for the PTP1B ·LZP25 complex included
PTP1B residues 2-284 and all atoms of the inhibitor. Binding
of pTyr or nonhydrolyzable pTyr mimetics are usually ac-
companied by the closure of the flexible WPD loop (residues
179-187).19,22,30 Thus, it was striking to find that the WPD
loop adopts an open conformation; the overall structure of
PTP1B·LZP25 is similar to the ligand-free PTP1B structure in
the WPD loop open form,31 with root mean square derivation
(rmsd) for all R-carbon positions between the two being 0.499
Å. The major difference between the two structures is electron
density in the PTP1B active site corresponding to LZP25, which
was confirmed by analyzing the 2Fo - Fc and Fo - Fc difference
Fourier maps with contour levels of 1.0 and 2.5 δ, respectively
(Figure 3A).
crystal parameters
space group
cell dimensions (Å)
a ) b
P3121
P3121
88.5
104.7
88.6
104.3
c
data collection
resolution (Å)
50-2.4
17,995
94.7
4.2
8.6
50-1.6
54,300
87.4
3.1
12.4
no. of unique reflections
completeness (%)
redundancy
a
Rmerge
refinement
resolution limit (Å)
no. of reflections used (F g 2δ (F))
no. of protein atoms
no. of inhibitors
50-2.4
17,214
2297
1
157
50-1.9
36,242
2297
1
85
21.3/24.4
no. of waters
c
Rworkb/ Rfree
20.7/25.3
rms deviations from ideal geometry
bond length (Å)
0.0066
1.25
38.0
0.0058
1.17
41.7
bond angle (deg)
average B-factors (Å2)
As shown in Figure 3B, LZP25 occupies the PTP1B active-
site pocket whose base corresponds to the phosphate-binding
loop or P-loop (residues 214-221). The diketoacid moiety
makes two hydrogen bonds with the main chain amides of the
P-loop, one polar interaction with the side chain of Arg221,
one hydrogen bond with the hydroxyl side chain of Ser216,
and two additional polar interactions with the side chains of
Gln266 and Tyr46, respectively. Besides these polar interactions,
there are also ample van der Walls contacts between PTP1B
and LZP25. For example, the phenyl ring immediately attached
to the diketoacid has extensive hydrophobic interactions with
the side chains of residues lining up the active-site cavity,
including Tyr46, Val49, Ala217, and Ile219. There are also van
der Walls contacts between the two carbon atoms of the
terminal ketoacid moiety and the side-chain of Gln262.
Additionally, one of the distal phenyl rings in LZP25 forms
hydrophobic interactions with residues Asp48 and Gln262,
while the other makes van der Walls contacts with residues
Tyr46, Arg47, and Asp48. These additional nonpolar interac-
tions with the distal phenyl ring(s) may be responsible for
the higher affinity of LZP4, LZP38, and LZP25, in compari-
son with those (LZP36, LZP3, and LZP39) that lack the distal
aromatic substitutions (Table 1).
b
a Rmerge
F(h)obsd|/ΣhF(h)obsd
)
ΣhΣi|I(h)i
-
I(h) |/ΣhΣiI(h)i Rwork
)
Σh|F(h)calcd -
,
where F(h)calcd and F(h)obsd were the refined
c
calculated and observed structure factors, respectively. Rfree was
calculated for
a
randomly selected 3.5% (PTP1B ·LZP25) and 4.9%
(PTP1B·LZP6) of the reflections that were omitted from refinement.
Figure 2. The effect of LZP25 (A) and LZP6 (B) on PTP1B-catalyzed
pNPP hydrolysis. The Lineweaver-Burk plot displayed the characteristic
intersecting line pattern, consistent with noncompetitive inhibition. The
experiment was performed at 25 °C and pH 7.0. (A) LZP25 concentrations
were 0 (b), 2.5 (O), 5 (1), and 7.5 µM (3), respectively. (B) LZP6
concentrations were 0 (b), 1.25 (O), 2.5 (1), and 3.75 µM (3), respectively.
weaver-Burk plot in Figure 2, PTP1B-catalyzed pNPP hy-
drolysis in the presence of LZP25 displayed the characteristic
intersecting line pattern for noncompetitive inhibition, with a
Ki value 16.3 ( 1.1 µM. The observation that LZP25, a putative
pTyr surrogate, was not competitive with substrate is rather
intriguing. This is especially so because noncompetitive PTP1B
inhibitors have previously been found to bind an allosteric site
located ∼20 Å from the active site.29 Kinetic analysis indicated
that dimeric LZP6 is also a noncompetitive inhibitor of PTP1B
with a Ki of 12 ( 0.6 µM. Noncompetitive inhibition of PTP1B-
catalyzed reaction by LZP25 and LZP6 was also apparent when
the data were analyzed by Eadie-Hofstee plot (Figure S1). Given
the lack of a formal charge in the aryl diketoacid dimer, it is
not at all clear where the binding site for LZP6 is in PTP1B.
To provide insight into the molecular basis for PTP1B inhibition
Why Is the WPD Loop in the Open Conformation in the
PTP1B·LZP25 Complex? Enzymes require conformational
flexibility to control solvent accessibility, recruit functional
residues for catalysis, and/or facilitate the binding and release
of substrate or product. For the PTPs, the major conformational
change essential to substrate binding and catalysis is restricted
to the movement of the WPD loop, which harbors the general
acid/base (Asp181 in PTP1B). Crystallographic studies have
captured the WPD loop either in the open or closed confor-
mations.30,32,33 Spectroscopic measurements indicate that in
solution ligand-free PTP exists as an equilibrium mixture of
(30) Jia, Z.; Barford, D.; Flint, A. J.; Tonks, N. K. Science 1995, 268,
1754–1758.
(31) Barford, D.; Flint, A. J.; Tonks, N. K. Science 1994, 263, 1397–404.
(32) Stuckey, J. A.; Schubert, H. L.; Fauman, E.; Zhang, Z.-Y.; Dixon,
J. E.; Saper, M. A. Nature 1994, 370, 571–575.
(29) Wiesmann, C.; Barr, K. J.; Kung, J.; Zhu, J.; Erlanson, D. A.; Shen,
W.; Fahr, B. J.; Zhong, M.; Taylor, L.; Randal, M.; McDowell, R. S.;
Hansen, S. K. Nat. Struct. Mol. Biol. 2004, 11, 730–737.
(33) Schubert, H. L.; Fauman, E. B.; Stuckey, J. A.; Dixon, J. E.; Saper,
M. A. Protein Sci. 1995, 4, 1904–1913.
9
J. AM. CHEM. SOC. VOL. 130, NO. 50, 2008 17079

A R T I C L E S
Liu et al.
Figure 3. Crystal structure of PTP1B in complex with LZP25. (A) Electron density map at the active site contoured at 1.0 σ level. The P-loop and WPD
loop are shown in red and blue, respectively. (B) Binding mode of LZP25 in the active site of PTP1B. Atomic colors are as follows: oxygen - red, carbon
- cyan, sulfur - orange, nitrogen - blue. Ligand’s carbon atoms are colored yellow. Key interactions are highlighted with dash lines as follows: electrostatic
interactions with the ligand are shown in red, and hydrophobic interactions with ligands are shown in black. P-loop of PTP1B is shown with transparent
green cartoon. (C) Superposition of PTP1B ·LZP25 and PTP1B ·pTyr, the superposition was calculated with active-site residues without the ligands. Residues
marked with * are in the WPD loop closed conformation. Atomic colors are as follows: oxygen - red, carbon - white, sulfur - orange, nitrogen - blue.
LZP25’s carbon atoms are colored yellow, while carbon atoms from PTP1B ·pY (PDB entry code 1EEN) are colored green.
both the open and closed forms, and that ligand binding
stabilizes the closed conformation.34-36
including Arg221 are less optimal and insufficient to trigger
the WPD loop closure. Moreover, the position of the diketoacid
proximal phenyl ring is nearly perpendicular to that of pTyr,
which would cause a steric clash with Phe182 if the WPD loop
were in the closed conformation. Thus, the WPD loop adopts
an open conformation in the PTP1B ·LZP25 complex. It appears
that, depending on the specific interactions with the PTP, an
active site-directed ligand can either bind the active-site cleft
in the WPD loop open or closed conformation.
Although LZP25 occupies the active site, the manner by
which it interacts with PTP1B is significantly different from
those observed between pTyr and PTP1B (Figure 3C). A key
difference is how Arg221 in the P-loop is engaged by the ligand.
When pTyr or an optimal nonhydrolyzable pTyr mimetic is bound
at the active site, a bidentate hydrogen bond is formed between
the oxyanion and the guanidinium group of Arg221.19,22,30 This
event triggers a series of interactions between Arg221 and the side
chains of the invariant Glu115 as well as Trp179 in the WPD loop,
which are likely responsible for the stabilization of the closed
conformation.37-39 The phenyl ring of pTyr is then effectively
buried within the active-site cavity, created by the nonpolar side-
chains of Ala217 and Ile219 of the P-loop, Phe182 of the WPD
loop, Tyr46, Val49 and Gln262. In stark contrast, the interactions
between the diketoacid moiety of LZP25 and the P-loop
Structure of the PTP1B ·LZP6 Complex. Dimerization of aryl
diketoacids through a piperazino or cyclohexane-1,4-trans-
diamine linkage generated several PTP1B inhibitors with
improved potencies over their monomeric counterparts (Tables
1 and 2). This is a novel finding because, unlike other PTP1B
inhibitors, these dimeric compounds have no negative charge(s),
which are usually required for binding to the highly positively
charged PTP active site. To better understand the mode of
interaction between PTP1B and the aryl diketoacid dimers, we
determined the crystal structure of PTP1B in complex with
LZP6, which was refined to 1.9 Å resolution (Table 3). The
final model for PTP1B ·LZP6 includes residues 2-283 and all
atoms of LZP6. The overall structure of PTP1B in the complex
is similar to that of PTP1B ·LZP25, with rmsd between all
R-carbon positions of the two structures at 0.211 Å. Similar to
LZP25, LZP6 was unambiguously identified by well-defined
electron densities in the active site, and the WPD loop in the
PTP1B·LZP6 structure is also in the open conformation (Figure
(34) Juszczak, L. J.; Zhang, Z.-Y.; Wu, L.; Gottfried, D. S.; Eads, D. D.
Biochemistry 1997, 36, 2227–2236.
(35) Wang, F.; Li, W.; Emmett, M. R.; Hendrickson, C. L.; Marshall, A. G.;
Zhang, Y.-L.; Wu, L.; Zhang, Z.-Y. Biochemistry 1998, 37, 15289–
15299.
(36) Khajehpour, M.; Wu, L.; Liu, S.; Zhadin, N.; Zhang, Z.-Y.; Callender,
R. Biochemistry 2007, 46, 4370–4378.
(37) Keng, Y.-F.; Wu, L.; Zhang, Z.-Y. Eur. J. Biochem. 1999, 259, 809–
814.
(38) Hoff, R. H.; Hengge, A. C.; Wu, L.; Keng, Y.-F.; Zhang, Z.-Y.
Biochemistry 2000, 39, 46–54.
(39) Zhang, Z.-Y. Prog. Nucleic Acid Res. Mol. Biol. 2003, 73, 171–220.
9
17080 J. AM. CHEM. SOC. VOL. 130, NO. 50, 2008

Aryl Diketo Acid-Based PTP1B Inhibitors
A R T I C L E S
Figure 4. Crystal structure of PTP1B in complex with LZP6. (A). Electron density map of LZP6 was contoured at 1.2 δ level in blue. Atoms were colored
according to atom type (carbon in yellow, oxygen in red, and nitrogen in blue). The P-loop and WPD loop are shown in red and blue, respectively. (B).
Binding modes of LZP6 in the active site of PTP1B: Atomic colors are as follows: oxygen - red, carbon - cyan, sulfur - orange, nitrogen - blue. Ligand’s
carbon atoms are colored yellow. Key interactions are highlighted with dash lines as follows: electrostatic interactions with the ligand are shown in red, and
hydrophobic interactions with ligands are shown in black. (C) Superposition of PTP1B ·LZP6 and PTP1B ·pTyr, the superposition was calculated with
active-site residues without the ligands. Active site of PTP1B in the open conformation is shown by transparent surface representation, and key residues are
shown in stick model. Phenyl rings PB1 and PB2 of LZP6 are not shown. Atomic colors are as follows: oxygen - red, carbon - white, sulfur - orange, nitrogen
- blue. Carbon atoms of LZP6 are colored yellow, and pTyr’s carbon atoms are colored green. (D) A symmetry-related interaction between LZP6 monomer
B and PTP1B. LZP6 is shown as a stick model, PTP1B is shown by transparent surface representation. Color scheme for surface: carbon - white; oxygen
- red; nitrogen - blue.
4A). However, the mode of LZP6 binding is completely different
from that of LZP25. Only one of the aryl diketoacid (monomer
A), including the piperazino linker, is actively engaged with
the PTP1B active site, while the other half (monomer B) of
LZP6 appears to be solvent exposed (Figure 4B). Unlike LZP25,
which has numerous polar interactions with PTP1B, only one
electrostatic contact is evident between Lys118 and OB3 of
LZP6, which is mediated by a water molecule. Instead, binding
between PTP1B and LZP6 is dominated by hydrophobic
interactions (Figure 4B). Specifically, the phenyl ring directly
adjacent to the keto group in monomer A makes numerous van
der Waals contacts with the side chains of Tyr46, Ser216,
Ala217, Ile219, and Gln262. Interestingly, this phenyl ring sits
at approximately the same location occupied by the aromatic
moiety of pTyr seen in the PTP1B ·pTyr complex (Figure 4C
and ref.19). In addition, the piperazine linker also makes
hydrophobic contacts with Tyr46. Without any formal charge,
the terminal phenyl ring in monomer A does not point to the
bottom of the active site and interact with the P-loop, as is the
case for the phosphoryl group in pTyr. Rather, the ring extends
across the active-site cleft and makes multiple nonpolar interac-
tions with the side chains of Trp179, Phe182, Arg221, and
Thr263 (Figure 4B). Specific interactions between LZP6
terminal phenyl ring and residues Trp179 and Phe182 also
effectively block the WPD loop closure, thereby providing a
structural basis for the open conformation observed in the
PTP1B·LZP6 complex. Collectively, the structure reveals that
LZP6 binds PTP1B in the WPD loop open conformation and
covers the active-site pocket like a lid.
Although half of the LZP6 molecule appears to be exposed
in the solvent, symmetry operation suggests that the two phenyl
rings in monomer B may interact with the side chains of Phe196
and Phe280 from another PTP1B molecule in the crystal (Figure
4D). This observation is interesting because residues Phe196
and Phe280 have been previously shown to define a hydrophobic
pocket, located ∼20 Å from the active site, that was occupied
by an allosteric inhibitor of PTP1B.29 To ascertain the signifi-
cance of these symmetry related interactions, we replaced either
Phe196 or Phe280 with an Ala and evaluated the effect on LZP6
binding affinity. As shown in Table 4, IC50 values of LZP6 for
F196A and F280A were not significantly different from that of
the wild-type enzyme, suggesting that LZP6 is unlikely to
interact with these residues in solution. Similarly, Arg47 does
9
J. AM. CHEM. SOC. VOL. 130, NO. 50, 2008 17081

A R T I C L E S
Liu et al.
Table 4. IC50 values of LZP6 for wild-type and mutant PTP1B
PTP1B
IC50, µMa
WT
16 ( 2
R47V
F182A
F196A
F280A
Q262A
T263A
16.4 ( 1.1
58.7 ( 2.5
12.6 ( 1.5
11.8 ( 0.8
55.6 ( 2.9
37.6 ( 2.0
a All measurements were made using pNPP as a substrate at pH 7.0,
25 °C, and I ) 0.15 M.
Figure 5. Modeling of pNPP binding to PTP1B in the presence of LZP25
(A) or LZP6 (B). The structures of PTP1B ·LZP25 and PTP1B ·LZP6 were
utilized as the starting points for molecular modeling. pNPP was manually
docked into active site of PTP1B in the presence of LZP25 or LZP6, and
energy minimized using program CNS.23
Figure 6. Cellular activity of aryl diketoacid derivatives. (A) Effect of
LZP25 and LZP6 on insulin-mediated IRꢀ phosphorylation in HepG2 cells.
(B) Effect of LZP70 and LZP71 on insulin-mediated IRꢀ phosphorylation
in HepG2 cells. (C) Concentration dependence of LZP6 on insulin-mediated
IRꢀ phosphorylation in HepG2 cells.. Subconfluent HepG2 cells were
preincubated with either LZP25 or LZP6 20 min, followed by a 5 min
treatment with or without 100 nM insulin. Cell lysates were subjected to
SDS-PAGE and the resolved proteins transferred to nitrocellulose mem-
branes, which were probed with antiphospho-IRꢀ antibodies. ꢀ-Actin was
used as a loading control.
Table 5. Selectivity of LZP25 and LZP6 against a Panel of PTPsa
PTPs
IC50 for LZP25 (µM)
IC50 for LZP6 (µM)
PTP1B
HePTP
SHP2
Lyp
VHR
LAR
20 ( 5
16 ( 2
87 ( 6
72 ( 4
85 ( 8
58 ( 4
73 ( 20
70 ( 24
of LZP70 and LZP71 are slightly higher (2-3 fold) than those
of LZP6 and LZP40, supporting the conclusion that LZP6 is
involved in a novel mode of interaction with PTP1B active site
and only half of the aryl diketoacid dimer is required for binding.
Structural and Mechanistic Basis for Noncompetitive PTP1B
Inhibition by LZP25 and LZP6. Two major mechanisms of
reversible enzyme inhibition are commonly described in text-
books. In competitive inhibition, the inhibitor binds to the active
site in place of the substrate, thus preventing the enzyme-
catalyzed reaction. Competitive inhibition increases the apparent
Km, which can be overcome by high substrate concentration due
to mass action. In noncompetitive inhibition, the inhibitor binds
to the enzyme at an allosteric site that is distinct from the active
site. This type of inhibition reduces the maximum rate of an
enzymatic reaction without affecting the apparent binding
affinity of the enzyme for the substrate. The decrease in Vmax
occurs as a result of conformational changes in the active site
induced by the presence of the inhibitor in the allosteric site.
Our kinetic analyses indicate that LZP25 and LZP6 act as
noncompetitive inhibitors of PTP1B. Unlike classical noncom-
petitive inhibitors, however, both LZP25 and LZP6 occupy the
active site, rather than an allosteric site away from the active
site. Although the specific binding modes differ between the
two compounds, one common feature from the structures of
PTP1B·LZP25 and PTP1B ·LZP6 is that the WPD loop is in
the open conformation. Molecular modeling suggests that
substrate pNPP can still be accommodated by PTP1B’s active
site in the presence of either LZP25 or LZP6 (Figure 5). Because
PTP catalysis requires the WPD loop in the closed conformation,
which positions the general acid Asp181 close to the scissile
113 ( 6
108 ( 6
no inh. at 100 µM
no inh. at 100 µM
no inh. at 100 µM
no inh. at 100 µM
PTPR
a All measurements were made using pNPP as a substrate at pH 7.0,
25 °C, and I ) 0.15 M.
not participate in LZP6 binding and the Arg47 to Val substitu-
tion does not alter the IC50 value for LZP6. On the contrary,
substitutions at Phe182, Gln262, and Thr263 reduced PTP1B’s
affinity for LZP6 by 3-4 fold, which is consistent with the
structural observations that these residues contribute to binding
LZP6 (Figure 4C). Overall mutational data support the conclu-
sion that symmetry-related interactions between LZP6 monomer
B and Phe196 and Phe280 are likely the result of crystal pac-
king.
To further substantiate the observed binding mode between
PTP1B and LZP6, we analyzed several compounds structurally
related to LZP6 (Table 2). LZP7, LZP37, and LZP40 possess
chemical entities similar to those present in LZP6 that are
known to be important for interactions with PTP1B active site.
Accordingly, these compounds possess either comparable
(LZP7) or enhanced (LZP37 and LZP40) inhibitory activities
to that of LZP6. In contrast, LZP8 and LZP9 lack the terminal
phenyl ring predicted to make a number of key contacts with
PTP1B. As expected, they are less potent than LZP6 and exhibit
no inhibitory activity against PTP1B at 100 µM. To directly
test whether one aryl diketoacid moiety is sufficient for binding,
we synthesized monomeric versions of LZP6 and LZP40, in
which one of the aryl diketoacid is replaced by a tert-butyl
dicarbonate (BOC) group. As shown in Table 2, the IC50 values
9
17082 J. AM. CHEM. SOC. VOL. 130, NO. 50, 2008

Aryl Diketo Acid-Based PTP1B Inhibitors
A R T I C L E S
Figure 7. Surface representations of PTP1B in the WPD loop closed (A) and open (B) conformation. The figure was prepared using GRASP.42
oxygen of the substrate for efficient proton transfer,37-39 the
LZP25- or LZP6-bound PTP1B is catalytically incompetent.
Consequently, substrate turnover cannot occur when LZP25 or
LZP6 is present in the PTP1B active site, because the catalyti-
cally essential residues are not optimally aligned to promote
pNPP hydrolysis. Thus, consistent with the mode of noncom-
petitive inhibition, there is no competition between the aryl
diketoacid-based inhibitors and the substrate. The mechanism
for LZP25 and LZP6-mediated PTP1B inhibition derives from
the fact that they stabilize the open, inactive conformation of
the WPD loop, thereby reducing the concentration of the
antibodies. As shown in Figure 6A, incubation of cells with
LZP25 had no significant effect on IRꢀ phosphorylation, while
LZP6 treatment increased insulin stimulated IRꢀ phosphoryla-
tion by 230%. Given that LZP25 and LZP6 have similar affinity
for PTP1B, the results suggest that LZP25 is not cell permeable,
whereas LZP6 is capable of penetrating the cellular membrane.
If the carboxylic acid in LZP25 obstructs cell membrane
penetration, then masking the acid with the piperazino linker
should improve its cellular activity. Indeed, when cells were
treated with LZP70 or LZP71 at ∼4× of their respective IC50
values, both compounds were able to augment the insulin
stimulated IRꢀ phosphorylation to the same extent (∼2-fold)
as effected by LZP6 at equivalent concentration (Figure 6B).
We further demonstrated that LZP6 can potentiate insulin
signaling in a dose dependent manner, exhibiting cellular
activity at both 15 and 30 µM (1 and 2× IC50) concentrations
(Figure 6C).
Implications for PTP1B Inhibitor Development. As men-
tioned above, there is compelling evidence that define PTP1B
as an outstanding target for the treatment of a number of
human diseases. A major challenge in drug discovery
programs targeted toward the PTPs is to identify inhibitory
compounds with pharmacologically acceptable bioavailability.
Most PTP active site-directed pharmacophores identified at
present are negatively charged pTyr mimetics and therefore
cannot readily enter the cell. This lack of cell permeability
limits the utility of such compounds in signaling studies and
further therapeutic development. We show in this study that
LZP25 and LZP6 function as noncompetitive PTP1B inhibi-
tors and bind the active site in distinct fashions that
effectively block the WPD loop closure, thereby preventing
formation of the catalytically competent form of the enzyme.
This is in stark contrast to other pTyr mimetics, which bind
the WPD loop closed conformation. As depicted in Figure
7, the electrostatic potential and topology of the catalytic
cleft are considerably different between the WPD loop open
and closed forms. Moreover, with the WPD loop in the open
conformation the active-site pocket is substantially larger with
greater hydrophobic binding surface. Thus, in comparison
to the WPD closed form, the open conformation presents a
unique opportunity for developing active site-directed PTP
inhibitors with improved pharmacological properties.
catalytically competent enzyme, leading to a decrease in Vmax
.
Cellular Activity of LZP25 and LZP6. Given the observed
potency of LZP25 and LZP6 for PTP1B, we proceeded to
evaluate their ability to inhibit PTP1B inside the cell. Previous
studies have demonstrated that PTP1B serves as a negative
regulator of the insulin activated signaling pathways by cata-
lyzing the dephosphorylation of the insulin receptor ꢀ (IRꢀ)
subunit. Thus, inhibition of PTP1B activity should enhance
insulin signaling. Indeed, IRꢀ phosphorylation in the liver of
the PTP1B knockout mice is prolonged by 2-fold.3,4 PTP1B
knockdown by antisense in rat hepatoma cells yields a 2.3-fold
increase in the level of phosphorylated IRꢀ.40 In addition,
significant enhancements in IRꢀ phosphotyrosine levels are also
observed with cell-permeable analogues of a potent and selective
small-molecule PTP1B inhibitor.41
To examine the specificity of LZP25 and LZP6 for PTP1B,
their inhibitory activities toward a panel of PTPs including
cytosolic PTPs, HePTP, SHP2, and Lyp, the receptor-like PTPs,
LAR and PTPR, and the dual specificity phosphatase, VHR,
were determined. As shown in Table 5, LZP25 and LZP6
display at least several-fold selectivity for PTP1B over all PTPs
examined.
To determine the cellular efficacy of LZP25 and LZP6, we
assessed their effect on insulin signaling in human hepatoma
HepG2 cells. The cells were incubated with either 150 µM (7.5
of the IC50 value) LZP25 or 60 µM (3.75× of the IC50 value)
LZP6 for 20 min and subsequently treated either with or without
insulin (100 nM) for 5 min. Cell lysates were resolved by SDS-
PAGE, then the proteins were transferred to nitrocellulose
membranes and probed with antiphospho-IRꢀ (pY1162/pY1163)
It has proven difficult to develop effective uncharged PTP
active site-directed small-molecule inhibitors, due primarily
(40) Clampit, J. E.; Meuth, J. L.; Smith, H. T.; Reilly, R. M.; Jirousek,
M. R.; Trevillyan, J. M.; Rondinone, C. M. Biochem. Biophys. Res.
Commun. 2003, 300, 261–267.
(41) Xie, L.; Lee, S. Y.; Andersen, J. N.; Waters, S.; Shen, K.; Guo, X. L.;
Moller, N. P.; Olefsky, J. M.; Lawrence, D. S.; Zhang, Z. Y.
Biochemistry 2003, 42, 12792–12804.
(42) Nicholls, A.; Sharp, K. A.; Honig, B. Proteins, Struct., Funct. Genet.
1991, 11, 281–296.
9
J. AM. CHEM. SOC. VOL. 130, NO. 50, 2008 17083

A R T I C L E S
Liu et al.
Data Deposition
to the electrostatic properties of the PTP active site. We
provide the first example in which a neutral compound LZP6
can engage PTP1B active site in such a way as to maintain
the open, inactive conformation of the WPD loop, without
resorting to electrostatic interactions with the P-loop. Notably,
LZP6 exhibits excellent cell permeability and enhances
insulin signaling in HepG2 cells. Further increase in potency
could be achieved by appending an appropriate chemical
group to LZP6 to fill the space taken by pNPP (Figure 5B).
Moreover, binding specificity can be improved by tethering
LZP6 to appropriate functional groups in order to engage
additional unique surface areas that border the PTP1B active
site. Given the conserved structural and catalytic properties
of the PTP active site, it is expected that targeting the
inactive, WPD loop open conformation could represent a
general and effective strategy applicable to all PTPs.
The coordinates for the structures of the PTP1B ·LZP25
complex (accession number 3EB1), and the PTP1B ·LZP6
complex (accession number 3EAX) have been deposited in the
Protein Data Bank.
Acknowledgment. This work was supported by National Institutes
of Health Grant CA126937. The work in Y.-Q.L’s laboratory was in
part supported by National Natural Science Foundation of China (No.
30721005 and No. 20872154).
Supporting Information Available: Eadie-Hofstee plot analy-
sis of PTP1B inhibition by LZP25 and LZP6; complete ref 27.
This material is available free of charge via the Internet at http://
pubs.acs.org.
JA8068177
9
17084 J. AM. CHEM. SOC. VOL. 130, NO. 50, 2008