Parallel synthesis of a library of bidentate protein tyrosine phosphatase inhibitors based on the α-ketoacid motif

Article abstract of DOI:10.1016/j.bmc.2004.03.058

Protein tyrosine phosphatases (PTPases) regulate intracellular signal transduction pathways by controlling the level of tyrosine phosphorylation in cells. These enzymes play an important role in a variety of diseases including type II diabetes and infecti

Full text of DOI:10.1016/j.bmc.2004.03.058

Bioorganic & Medicinal Chemistry 12 (2004) 3289–3298
Parallel synthesis of a library of bidentate protein tyrosine
phosphatase inhibitors based on the a-ketoacid motifq
Yen Ting Chen and Christopher T. Seto*
Department of Chemistry, Brown University, 324 Brook St. Box H, Providence, RI 02912, USA
Received 2 February 2004; revised 25 March 2004; accepted 26 March 2004
Available online 10 May 2004
Abstract—Protein tyrosine phosphatases (PTPases) regulate intracellular signal transduction pathways by controlling the level of
tyrosine phosphorylation in cells. These enzymes play an important role in a variety of diseases including type II diabetes and
infection by the bacterium Yersinia pestis, which is the causative agent of bubonic plague. This report describes the synthesis, using
parallel solution-phase methods, of a library of 104 potential inhibitors of PTPases. The library members are based on the bis(aryl
a-ketocarboxylic acid) motif that incorporates a carboxylic acid on the central benzene linker. This carboxylic acid was coupled with
a variety of different aromatic amines through an amide linkage. The aromatic component of the resulting amides is designed to
make contacts with residues that surround the active site of the PTPase. The library was screened against the Yersinia PTPase and
PTP1B. Based upon the screening results, four members of the library were selected for further study. These four compounds were
evaluated against the Yersinia PTPase, PTP1B, TCPTP, CD45, and LAR. Compound 14 has an IC50 value of 590 nM against
PTP1B and is a reversible competitive inhibitor. This affinity represents a greater than 120-fold increase in potency over compound
2, the parent structure upon which the library was based. A second inhibitor, compound 12, has an IC50 value of 240 nM against the
Yersinia PTPase. In general, the selectivity of the inhibitors for PTP1B was good compared to LAR, but modest when compared to
TCPTP and CD45.
ꢀ 2004 Elsevier Ltd. All rights reserved.
1. Introduction
vation of B and T cell receptor signaling.7 Finally,
bacterial infection by Yersinia pestis, the agent that is
responsible for the bubonic plague, relies on the activity
of a PTPase (YopH). This pathogenic bacterium secretes
the phosphatase into host cells, where it disrupts host
signal transduction processes and thus interferes with
the immune response.8 Since PTPases are important in
such a wide variety of biological processes, there is
currently significant interest in these enzymes as targets
for therapeutic intervention and a great deal of effort is
being invested toward the development of potent and
specific PTPase inhibitors that also show good bio-
availability.
Protein tyrosine phosphatases (PTPases) and kinases
regulate the level of tyrosine phosphorylation of pro-
teins, and as a result they are one of the key mechanisms
for controlling intracellular signal transduction path-
ways. The unregulated activity of phosphatases is
associated with a variety of diseases.1;2 For example,
PTP1B dephosphorylates the insulin receptor and a
number of downstream signaling proteins, and is thus a
potential target for the treatment of type II diabetes.3–5
Several PTPases including PTP-a and the Cdc25 phos-
phatases are implicated in the development of cancer.
PTP-a activates the Src-family of kinases while Cdc25B
activates the cyclin-dependent kinases.6 CD45 is a po-
tential target for the treatment of autoimmune diseases
and inflammation because it is involved with the acti-
Because all PTPases share a common catalytic mecha-
nism, and many of them have a highly conserved active
site,9 one of the most challenging aspects of designing
inhibitors for these enzymes is the ability to achieve
selectivity for one particular PTPase over others. The
active site of PTPases is selective for binding phospho-
tyrosine.10 However, phosphotyrosine by itself has a low
affinity for the enzyme. Thus, it is apparent that regions
of the active site cleft beyond the catalytic residues
are crucial for efficient recognition and binding of
Keywords: Phosphatase; Inhibitor; Library; a-Ketocarboxylic acid.
q Supplementary data associated with this article can be found, in the
online version, at doi:10.1016/j.bmc.2004.03.058
* Corresponding author. Tel.: +1-401-863-3587; fax: +1-401-863-9368;
e-mail: christopher_seto@brown.edu
0968-0896/$ - see front matter ꢀ 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmc.2004.03.058

3290
Y. T. Chen, C. T. Seto / Bioorg. Med. Chem. 12 (2004) 3289–3298
substrates.11 Zhang and co-workers have identified res-
idues in PTP1B (Arg-24, Arg-254, Gln-262) that com-
prise a secondary binding site that is near to the catalytic
site.12 These investigators reasoned that, by the principle
of additivity of the free energy of binding, inhibitors that
bind in both the catalytic site and the secondary site
should have significantly increased affinity for the
enzyme. This notion is supported by studies involving
the biphosphorylated insulin receptor kinase (IRK)
activation loop that is a substrate for PTP1B. The
crystal structure of the enzyme–substrate complex shows
that the two phosphotyrosine residues in the substrate
bind simultaneously to both sites. In addition, the
biphosphorylated substrate has a much greater affinity
for PTP1B than its monophosphorylated analog.13 Since
the residues that comprise the secondary site are not
conserved among all PTPases, it is reasonable to expect
that dual-binding inhibitors will exhibit increased
selectivity, as well as increased affinity.
Figure 1. Computer generated model of compound 2 bound in the
active site of PTP1B.
of PTP1B (Fig. 1), rather than the secondary site. An
Arg residue is present in an analogous position in the
Yersinia PTPase, but not in LAR. The model shown in
Figure 1 illustrates that, while the inhibitor can bind
with one a-ketoacid unit anchored in the active site and
the other extended toward Arg-47 and Asp-48, in this
conformation it does not take advantage of binding
interactions with the secondary site. As a result, we may
be able to improve the potency and selectivity of the
inhibitor by building on additional functionality to the
central benzene core that could extend into the sec-
ondary site. Here, we report on the synthesis and
screening of a 104-membered library of inhibitors with
the general structure 3. These compounds are based on
the structure of inhibitor 2, with the addition of a car-
boxylic acid group that is appended to the central benz-
ene ring. Coupling reactions between this carboxylic
acid and a variety of aromatic amines were used to
generate the library.
Several groups have taken advantage of this concept and
reported inhibitors that contain two phosphotyrosine
mimics, such as aryl difluoromethylenephosphonates,
which are tethered together by a linker.14–16 These com-
pounds display significantly improved inhibition and
selectivity when compared to inhibitors that bear only a
single phosphotyrosine mimic. They generally have
inhibition constants against PTP1B that range from low
micromolar to as low as 2 nM. However, structural
studies have revealed that many of these bidentate
inhibitors make contacts with the active site and other
less conserved, neighboring residues such as Lys-41, Arg-
47, and Asp-48, rather than the secondary binding
site.17–19 Recently, a group from Abbott Laboratories
used a linked-fragment strategy to generate several po-
tent oxalylamide-based inhibitors. These investigators
used X-ray crystallography to confirm that the inhibitors
occupy both the catalytic site and the secondary binding
site.20 Taken together, these results suggest that targeting
a combination of the active site and either the secondary
binding site or Arg-47 and Asp-48 are both effective
strategies for developing PTPase inhibitors.
Over the past several years, we have been working to
develop aryl a-ketocarboxylic acids as inhibitors of
PTPases.21;22 By linking two or three aryl a-ketoacids to
a central aromatic core, we demonstrated that the
potency of the inhibitors against the Yersinia PTPase
can be increased up to 280-fold when compared to
monomeric a-ketoacids such as 4-hydroxyphenylgly-
oxylic acid (1). Of the multivalent compounds that we
have studied, compound 2 showed the highest activity
against the Yersinia PTPase and PTP1B, and it had
good selectivity for these enzymes over LAR.23 This
inhibitor has an IC50 value of 2.7 lM against PTP1B at
pH 5.5. However, the binding affinity decreases signifi-
cantly (IC50 ¼ 73 lM) when the assays are performed at
pH 7.0. Since LAR does not possess the residues that
form the secondary site,12 one reasonable interpretation
of these inhibition data is that compound 2 may be
binding to both the active site and the secondary site.
However, modeling studies suggest that compound 2
may be more likely to bind to the active site and Arg-47
2. Results and discussion
2.1. Chemistry
In our previous studies, we employed a transamination
procedure using copper sulfate and glyoxylic acid to
convert esters of a-amino acids into the corresponding
a-ketoesters. These were subsequently hydrolyzed to

Y. T. Chen, C. T. Seto / Bioorg. Med. Chem. 12 (2004) 3289–3298
3291
yield a-ketocarboxylic acids.23 Although the transami-
nation reaction reliably provided the desired products,
the yields for this transformation were generally poor.
In order to avoid this problem, we have adopted a new
strategy for the synthesis of a-ketoacids in which
a-amino acids are first treated with trifluoroacetic
anhydride to give the corresponding trifluoro-
methyloxazolones following the procedure of Barnish
et al.24 The trifluoromethyloxazolone, which serves as a
protecting group, can be hydrolyzed at the end of the
synthesis to yield the corresponding a-ketoacid.
Reaction of compound 9 with thionyl chloride gave the
corresponding acid chloride. The library of inhibitors
was generated by reacting this acid chloride in parallel
with 103 different commercially available aromatic
amines (Fig. 2) to give amides with the general structure
10. After the coupling reactions were judged to be
complete by TLC analysis, the solvents were removed
and the resulting amides (10) were treated with dilute
sodium hydroxide to hydrolyze the two tri-
fluoromethyloxazolone groups and unmask the final a-
ketocarboxylic acids. Compound 9 was hydrolyzed in
the same manner to give the corresponding a-ketoacid
inhibitor with a carboxylic acid group on the central
benzene ring, rather than an amide. The final products
from these reactions were analyzed by LCMS to verify
that the syntheses were successful. The inhibitors were
screened in crude form, without purification, against the
Yersinia PTPase and PTP1B. Following this initial
screen, we selected four of the inhibitors (compounds
11–14) for further evaluation. These inhibitors were
resynthesized using the same chemistry, except that the
trifluoromethyloxazolone intermediates 10a–d were
purified prior to the final hydrolysis reaction.
This strategy is illustrated in Scheme 1, which begins
with a Williamson ether synthesis using 1 equiv of
compound 4 and 2 equiv of 5 to give 6. This reaction was
performed using sodium ethoxide in ethanol as the base.
These conditions also promoted a transesterification
reaction that converted the methyl ester group in 5 into
an ethyl ester. The synthesis of compounds 4 and 5 have
been reported previously.23;25 The Boc protecting groups
in compound 6 were removed with trifluoroacetic acid to
yield amine 7. Next, the three ester groups in 7 were
hydrolyzed using aqueous sodium hydroxide to give
amino acid 8. Refluxing compound 8 in trifluoroacetic
anhydride gave bis-oxazolone 9. Compared to the
transamination protocol, the oxazolone procedure
requires an additional step, but gives an overall yield
that is significantly higher.
Several details concerning the synthesis of this library
are worth noting. First, the trifluoromethyloxazolone
group is unstable under basic conditions. As a result, we
have limited the library to include only aromatic amines
that generally are less basic than aliphatic amines.
Reaction of the acid chloride derived from compound 9
with aliphatic amines such as benzylamine or diethyl-
amine led to decomposition of the oxazolone groups as
determined by LCMS and 19F NMR analysis. No such
decomposition was observed in coupling reactions with
aromatic amines. Second, coupling reactions using
compounds such as sulfanilic acid, which are insoluble
in the reaction solvent (CH2Cl2 or THF), were not
successful. Third, while coupling reactions with tetra-
and pentafluoroaniline derivatives were successful, the
resulting amides were hydrolyzed back the carboxylic
acid during the hydrolysis of the trifluoromethyloxazo-
lone groups. Finally, for library members that contained
other base-labile functional groups such as an ethyl
ester, benzothiazole, or phthalimide, these groups were
also hydrolyzed during cleavage of the tri-
fluoromethyloxazolones.
2.2. PTPase inhibition
We initially screened the library of inhibitors by deter-
mining the percent inhibition of the compounds against
the Yersinia PTPase and PTP1B at 1 and 5 lM inhibitor
concentrations, respectively (Figs. 3 and 4). These
measurements were performed with p-nitrophenyl
phosphate (p-NPP) as the substrate at 25 ꢁC in 100 mM
acetate buffer at pH 5.5. For reference, compound A104,
which has a carboxylic acid rather than an amide on the
central benzene ring of the inhibitor, gives 12% inhibi-
tion of the Yersinia PTPase at 1 lM concentration.
This compound has a IC50 value of 13 lM at pH 5.5
against this enzyme. The 10 best inhibitors of the Yer-
sinia PTPase are presented in Table 1, using the same
Scheme 1. Reagents and conditions: (a) Na, EtOH, 90%; (b) TFA,
CH2Cl2, 91%; (c) NaOH, H2O; (d) trifluoroacetic anhydride, 76% (two
steps); (e) SOCl2, benzene; (f) RNH2, CH2Cl2 or THF; (g) NaOH,
H2O.

3292
Y. T. Chen, C. T. Seto / Bioorg. Med. Chem. 12 (2004) 3289–3298
Figure 2. Aromatic amines used to synthesize the library. The designator under each aromatic amine refers to the library member with general
structure 3 where the R component is derived from the above amines.
designators provided in Figure 2. In order to improve
the dynamic range of the PTP1B assays, 16 of the best
inhibitors of this enzyme were re-assayed at pH 5.5 at a
concentration of 2 lM. The 10 best PTP1B inhibitors
that emerged from this second round of assays, along
with their % inhibition values, are shown in Table 2. The
library members in Tables 1 and 2 share several com-
mon features. First, inhibitors that contain polycyclic
aromatic groups such as naphthalene (A32), anthracene
(A51 and A98), and pyrene (A53) have good activity.
Second, inhibitors derived from both meta- and para-
phenoxyaniline (A70 and A71) are prominent in both
tables. Finally, the compound derived from 3-amino-
benzoic acid (A10) is a good inhibitor of both the Yer-
sinia PTPase and PTP1B. It is possible that the benzoic
acid moiety is favorable because it can make electro-
static contacts with Arg-24 or Arg-254 in the secondary
binding site of PTP1B and with homologous Arg resi-
dues in the secondary site of the Yersinia PTPase.
the inhibitors under more physiologically relevant con-
ditions and under conditions that provide a direct
comparison with most of the recent studies published in
the literature, we chose to perform the IC50 measure-
ments of these compounds using 50 mM 3,3-dimethyl-
glutarate buffer at pH 7.0 (Table 3). The four inhibitors
were assayed against the Yersinia PTPase, PTP1B,
LAR, CD45, and TCPTP.
Compounds 11–14 are all potent inhibitors of the Yer-
sinia PTPase, with IC50 values that range from 240 to
670 nM. Like many of the previous a-ketoacid-based
inhibitors that we have studied,23 these compounds
show greater potency against the Yersinia PTPase than
PTP1B. Nevertheless, compound 14 in particular is a
good inhibitor of PTP1B with an IC50 value of 590 nM.
For comparison, compound 2 that is the parent struc-
ture upon which this library was constructed, has an
IC50 value of 73 lM under the same conditions. Thus,
our combinatorial strategy that was designed to fill the
secondary binding site of PTP1B with an appropriate
functional group, has lead to a more than 120-fold
increase in the potency of the inhibitor. Lineweaver–
Burk analysis of compound 14 confirms that it is a
reversible, competitive inhibitor of PTP1B (Fig. 5).
Based upon these initial screening results, we chose four
inhibitors (11–14, corresponding to library members
A32, A70, A59, and A53, respectively) to investigate in
more detail. These compounds were resynthesized and
purified using standard techniques. In order to evaluate

Y. T. Chen, C. T. Seto / Bioorg. Med. Chem. 12 (2004) 3289–3298
3293
Figure 3. Assay of library members at 1 lM concentration against the
Yersinia PTPase. Designators on the x-axis refer to the library mem-
bers shown in Figure 2.
Figure 4. Assay of library members at 5 lM concentration against
PTP1B. Designators on the x-axis refer to the library members shown
in Figure 2.
Compounds 11–14 show less than a factor of two
selectivity for PTP1B over TCPTP. TCPTP is a non-
transmembrane PTPase like PTP1B and the Yersinia
PTPase, and it is also the closest structural homologue
to PTP1B.26 These three phosphatases share many res-
idues in common including those that make up the
secondary binding site, and many other residues near
the active site that are important for binding including
Arg-47 and Asp-48. As a result, it is not surprising that
the inhibitors show good affinity for both enzymes. The
lack of selectivity between PTP1B and TCPTP is caused
by the structural similarity of these two enzymes, since
all the factors involved in binding of the inhibitors to
PTP1B are also present in TCPTP. Among the numer-
ous PTPase inhibitors that have been reported by other
groups, only a few exhibit significant selectivity for
PTP1B over TCPTP.15;27;28
Table 1. Percent inhibition of the Yersinia PTPase by selected crude
library members at 1 lM concentrationa
O
CO2H
O
O
HO2C
O
R
O
Entryb
% Inhibition
A32
A59
A70
A51
A93
A10
A55
A8
95
94
94
91
90
90
90
89
89
87
A58
A44
LAR and CD45 are receptor type phosphatases that
have a different architecture when compared to the
nontransmembrane PTPases. Like PTP1B, LAR is
involved in the control of insulin receptor signaling,29
while CD45 is important in T cell regulation.30 Inhibi-
tors 11–14 have greater than 150-fold selectivity for
PTP1B over LAR. LAR does not have a residue that
corresponds to Arg-47 of PTP1B, and also lacks the
residues that make up the secondary binding site.
Because these residues play a key role in governing the
a Assay solutions contained 100 mM acetate buffer at pH 5.5, 1 mM
EDTA, 100 mM NaCl, and 10% DMSO.
b See Figure 2 for specific structures of the inhibitors.
binding interactions between enzyme and inhibitor and
they are missing in LAR, the inhibitors should have
weak affinity for LAR as is observed. In contrast com-
pounds 11–14 are reasonable inhibitors of CD45. The

3294
Y. T. Chen, C. T. Seto / Bioorg. Med. Chem. 12 (2004) 3289–3298
Table 2. Percent inhibition of PTP1B by selected crude library mem-
bers at 2 lM concentrationa
O
CO2H
O
O
HO2C
O
R
O
Entryb
% Inhibiton
A53
A56
A32
A36
A70
A98
A71
A51
A10
A103
74
63
59
55
51
47
45
39
38
33
Figure 5. Inhibition of PTP1B by compound 14. The activity of
PTP1B was measured at pH 7.0 as described under the Experimental
Section in the presence of the following concentrations of 14: (
)
ꢀ
0 lM; (D) 1.27 lM; (h) 2.54 lM; (d) 3.80 lM. Substrate concentra-
tions used in the assays were 0.94, 1.88, 3.76, and 7.53 mM.
a Assay solutions contained 100 mM acetate buffer at pH 5.5, 1 mM
EDTA, 100 mM NaCl, and 10% DMSO.
b See Figure 2 for specific structures of the inhibitors.
shows excellent selectivity for PTP1B over LAR, 10-fold
selectivity over CD45, and less than 2-fold selectivity
over TCPTP. Several of the compounds (12 and 14) are
also potent inhibitors of the Yersinia PTPase. We are
currently investigating the bioavailability of these and
related inhibitors.
selectivity for PTP1B over CD45 ranges from less than a
factor of two for compound 11, to greater than 10-fold
for the best PTP1B inhibitor, compound 14. CD45 also
lacks a residue that corresponds to Arg-47 and the res-
idues that constitute the secondary binding site. How-
ever, CD45 does contain an Asp residue analogous to
Asp-48, while LAR has Asn in its place. If this Asp
makes important contacts with the inhibitors, this
binding interaction may help to explain the difference in
affinities that the inhibitors display for CD45 and LAR.
Alternately, the inhibitors may simply bind in a different
conformation in the two enzymes.
4. Experimental
4.1. General methods
NMR spectra were recorded on Bruker Avance-300 or
Avance-400 instruments. Spectra were calibrated using
TMS (d ¼ 0:00 ppm) for 1H NMR, CDCl3
(d ¼ 77:0 ppm), acetone-d6 (d ¼ 29:5 ppm), or DMSO-d6
3. Conclusion
(d ¼ 39:5 ppm)
for
13C
NMR,
and
CFCl3
(d ¼ 0:00 ppm) for 19F NMR. Mass spectra were
recorded on either a Shimadzu LCMS-QP8000 or an
Applied Biosystems QSTAR electrospray mass spec-
trometer. HPLC analyses were performed on a Rainin
HPLC system with Varian C8 and C18 columns, and
UV detection. Semi-preparative HPLC was performed
on the same system using a semi-preparative column
(21.4 · 250 mm). Reactions were conducted under an
atmosphere of dry nitrogen in oven dried glassware.
THF was distilled from sodium and benzophenone.
In summary, we have described the parallel synthesis of
104 divalent a-ketocarboxylic acid inhibitors for PTP-
ases. These inhibitors are designed to make contacts
with the active site, the secondary binding site, and Arg-
47 and Asp-48 of PTP1B. Using this approach, we have
discovered an inhibitor that is more than 120 times more
potent against PTP1B than compound 2, the parent
structure upon which the library was based. The most
potent PTP1B inhibitor in this series, compound 14,
Table 3. Inhibition of phosphatases by compounds 11–14a
Compd
IC50 (lM)
Yersinia PTPase
PTP1B
LARb
CD45
TCPTP
2
73 12
11
12
13
14
0.51 0.08
0.24 0.05
0.67 0.15
0.27 0.05
4.3 1.0
1.5 0.3
4.5 1.0
0.59 0.07
0% at 650 lM
0% at 650 lM
0% at 650 lM
0% at 200 lMc
7.2 0.8
6.6 0.9
14.2 2.2
6.3 2.0
5.3 1.2
2.4 0.5
5.5 0.9
0.96 0.13
a Average of two measurements. Assays were performed in 50 mM 3,3-dimethylglutarate buffer at pH 7.0, 1 mM EDTA, 50 mM NaCl, and 10%
DMSO.
b % Inhibition at the given concentration.
c Insoluble beyond this concentration.

Y. T. Chen, C. T. Seto / Bioorg. Med. Chem. 12 (2004) 3289–3298
3295
ꢀ
Solvents were of reagent grade and were stored over 4 A
molecular sieves. All reagents were used as received.
Solvent removal was performed by rotary evaporation
at water aspirator pressure.
3 M HCl and the precipitate (compound 8) was collected
by centrifugation. Without further purification, trifluo-
roacetic anhydride (20 mL) was added in two portions to
compound 8 and the mixture was heated at reflux for
4 h, then concentrated to dryness. The residue was
treated with ethyl acetate (30 mL) and the insoluble
material was removed by filtration. The filtrate was
evaporated and the remaining material was purified by
flash chromatography (1:1:2 chloroform–ethyl acetate–
hexanes) to give compound 9 (0.248 g, 0.390 mmol, 76%)
4.1.1. Compound 6. To a solution of sodium (0.320 g,
13.9 mmol) dissolved in ethanol (40 mL) was added N-
Boc-D
-4-hydroxyphenylglycine methyl ester (5,23 3.93 g,
13.9 mmol) and then compound 425 (2.33 g, 6.65 mmol).
The mixture was stirred at room temperature for 14 h
under a N2 atmosphere, the solvent was removed, and
the residue was treated with H2O (30 mL) and ethyl
acetate (30 mL). The layers were separated, and the
aqueous layer was extracted with two portions of ethyl
acetate (15 mL). The organic layers were combined and
washed twice with a saturated solution of aqueous
Na2CO3 (30 mL), and brine (40 mL), dried over Na2SO4,
and concentrated to dryness. Purification by flash col-
umn chromatography (1:2:6 methylene chloride–ethyl
1
as a colorless solid: H NMR (300 MHz, DMSO-d6) d
5.33 (s, 2H), 5.59 (s, 2H), 7.05 (q, J ¼ 4:5 Hz, 2H), 7.22
(d, J ¼ 8:9 Hz, 2H), 7.26 (d, J ¼ 8:9 Hz, 2H), 7.70 (m,
2H), 8.06 (s, 1H), 8.30 (d, J ¼ 8:7 Hz, 4H), 13.25 (br s,
1H); 13C NMR (75 MHz, DMSO-d6) d 68.7, 69.6, 92.3
(q, JCF ¼ 33:2 Hz), 116.2, 116.4, 121.0 (two overlapping
resonances), 121.8 (q, JCF ¼ 280 Hz), 129.3, 130.6, 130.7,
131.5, 131.6, 132.2, 137.2, 138.1, 159.8, 163.2, 163.4,
164.2, 168.7; 19F NMR (376 MHz, DMSO-d6) d )78.5 (t,
J ¼ 4:1 Hz); HRMS-ESI (M)Hþ, negative ion mode)
calcd for C29H17F6N2O8 635.0889, found 635.0902.
acetate–hexanes) yielded compound
6
(4.60 g,
6.01 mmol, 90%) as colorless foam: 1H NMR (300 MHz,
CDCl3) d 1.23 (t, J ¼ 7:1 Hz, 6H), 1.39 (t, J ¼ 7:1 Hz,
3H), 1.45 (s, 18H), 4.19 (m, 4H), 4.38 (q, J ¼ 7:1 Hz,
2H), 5.10 (s, 2H), 5.25 (d, J ¼ 7:2 Hz, 2H), 5.52 (m, 4H),
6.96 (dd, J ¼ 8:7, 1.8 Hz, 4H), 7.30 (d, J ¼ 8:7 Hz, 2H),
7.31 (d, J ¼ 8:7 Hz, 2H), 7.62 (dd, J ¼ 8:0, 1.7 Hz, 1H),
7.76 (d, J ¼ 8:0 Hz, 1H), 8.10 (d, J ¼ 1:5 Hz, 1H); 13C
NMR (75 MHz, CDCl3) d 14.5, 14.7, 28.7, 57.5, 61.7,
62.1 (two overlapping resonances), 68.5, 69.6, 80.5,
115.5 (two overlapping resonances), 128.3, 128.7, 128.8
(two overlapping resonances), 129.9, 130.1, 130.2, 131.9,
136.5, 139.5, 155.2, 158.9, 159.0, 161.7, 171.7, 171.8;
HRMS-ESI (M+Hþ) calcd for C41H53N2O12 765.3599,
found 765.3615.
4.2. General experimental for the synthesis of the library
To a suspension of compound 9 (40 mg, 62.8 lmol) in
benzene (3 mL) was added SOCl2 (0.5 mL). The mixture
was heated at reflux for 3 h under a N2 atmosphere, then
the solvents were removed under vacuum. The residue
was dissolved in methylene chloride and split into eight
equal portions. Each portion was added to a solution of
a different aromatic amine (2.2 equiv) dissolved in either
CH2Cl2 or THF. See Figure 2 for a list of aromatic amines
that were used in this reaction. The reaction was stirred
overnight at 4 ꢁC and the solvent was removed. Without
further purification, a solution of 0.25 M NaOH
(0.5 mL) was added and the mixture was stirred for 2 h.
The reaction was quenched with 0.5 M HCl (1.5 mL)
and the precipitate was collected by centrifugation.
After washing the solid twice with 0.1 M HCl, the
solid was resuspended in water (0.5 mL) and extracted
twice with EtOAc (0.5 mL). The organic layers were
combined, dried over Na2SO4, and concentrated to
dryness.
4.1.2. Compound 7. Compound 6 (1.85 g, 2.67 mmol) was
dissolved in a solution of 25% trifluoroacetic acid in
methylene chloride (10 mL) and the solution was stirred
in an ice bath for 3 h. The solvent was removed, and the
residue was taken up in ethyl acetate (30 mL) and
washed twice with a saturated solution of aqueous
Na2CO3 (15 mL) and brine (15 mL). The organic layer
was dried over Na2SO4 and the solvent was removed by
rotary evaporation to yield compound 7 (1.19 g,
2.42 mmol, 91%) as a colorless solid: 1H NMR
(300 MHz, DMSO-d6) d 1.12 (t, J ¼ 7:1 Hz, 6H), 1.25 (t,
J ¼ 7:1 Hz, 3H), 2.21 (br s, 4H), 4.04 (m, 4H), 4.26 (q,
J ¼ 7:1 Hz, 2H), 4.43 (s, 2H), 5.16 (s, 2H), 5.38 (s, 2H),
6.92 (d, J ¼ 8:5 Hz, 2H), 6.97 (d, J ¼ 8:5 Hz, 2H), 7.29
(d, J ¼ 8:5 Hz, 4H), 7.66 (s, 2H), 7.96 (s, 1H); 13C NMR
(75 MHz, DMSO-d6) d 14.8, 14.9, 58.4, 61.1, 61.8, 68.3,
69.1, 115.2, 115.4, 128.9, 129.5, 130.0, 130.1, 132.2,
134.4 (two overlapping resonances), 137.9, 138.4, 158.2,
158.4, 167.3, 175.1; HRMS-ESI (M+Hþ) calcd for
C31H37N2O8 565.2550, found 565.2553.
4.2.1. Compound 10a. To a suspension of compound 9
(70.7 mg, 0.111 mmol) in benzene (4 mL) was added
SOCl2 (1 mL). The mixture was heated at reflux for 3 h
under a N2 atmosphere, then the solvents were removed
under vacuum. The residue was dissolved in methylene
chloride (1.5 mL) and a solution 1-aminonaphthalene
(39.8 mg, 0.278 mmol) in methylene chloride (0.25 mL)
was added. The solution was stirred at 4 ꢁC overnight,
diluted with methylene chloride (25 mL), and washed
twice with 1 N HCl (15 mL). The organic layer was dried
over Na2SO4, concentrated to dryness, and the residue
was washed twice with benzene to yield compound 10a
1
(50.9 mg, 66.9 lmol, 60%) as a colorless solid: H NMR
4.1.3. Compound 9. Compound 7 (0.291 g, 0.515 mmol)
was suspended in a solution of 2.5 M NaOH (10 mL)
and the mixture was heated at reflux for 4 h. After
cooling, the resulting solution was acidified to pH 7 with
(300 MHz, acetone-d6) d 5.42 (s, 2H), 5.62 (s, 2H), 6.82
(m, 2H), 7.27 (d, J ¼ 9:0 Hz, 2H), 7.33 (d, J ¼ 8:5 Hz,
2H), 7.51 (m, 3H), 7.79 (m, 3H), 7.91 (d, J ¼ 7:4 Hz,
1H), 7.96 (dd, J ¼ 7:4, 2.0 Hz, 1H), 8.13 (s, 1H), 8.22

3296
Y. T. Chen, C. T. Seto / Bioorg. Med. Chem. 12 (2004) 3289–3298
(d, J ¼ 8:1 Hz, 1H), 8.39 (d, J ¼ 9:0 Hz, 2H), 8.44 (d,
J ¼ 9:0 Hz, 2H), 9.78 (s, 1H); 13C NMR (75 MHz, ace-
tone-d6) d 68.2, 69.7, 92.4 (q, JCF ¼ 34:0 Hz), 115.8,
121.0, 121.1, 121.5 (q, JCF ¼ 280 Hz), 122.9, 123.0,
125.9, 126.4, 126.5, 127.7, 128.7, 129.0, 129.7, 129.9,
131.3 (two overlapping resonances), 133.9, 134.7, 135.6,
136.8, 137.3, 159.9, 163.5, 163.6, 163.7, 168.0; HRMS-
FAB (M+Naþ) calcd for C39H25F6NaN3O7 784.1494,
found 784.1477.
analogous to the preparation of 10a. 10d (49.4 mg,
59.1 lmol, 82%): 1H NMR (300 MHz, DMSO-d6) d 5.42
(s, 2H), 5.56 (s, 2H), 7.06 (m, 2H), 7.28 (d, J ¼ 8:9 Hz,
2H), 7.35 (d, J ¼ 9:0 Hz, 2H), 7.75 (m, 2H), 8.10 (m,
3H), 8.20 (m, 3H), 8.31 (m, 8 H), 10.99 (s, 1H); 13C
NMR (75 MHz, DMSO-d6) d 68.5, 69.9, 92.3 (q,
JCF ¼ 33:3 Hz), 116.3, 116.4, 121.0, 121.1, 121.8 (q,
JCF ¼ 280 Hz), 124.6, 125.2, 125.5, 125.7, 125.9, 126.0,
126.2, 127.3, 127.8, 128.0, 128.1, 128.6, 129.2, 129.7,
130.4, 130.5, 131.3, 131.6 (two overlapping resonances),
132.3, 132.5, 135.3, 137.2, 137.5, 158.8, 163.4, 163.5,
164.2, 168.7; HRMS-FAB (M+Naþ) calcd for
C45H27F6NaN3O7 858.1651, found 858.1636.
4.2.2. Compound 10b. To a suspension of compound 9
(43.8 mg, 68.8 lmol) in benzene (3 mL) was added SOCl2
(0.5 mL). The mixture was heated at reflux for 3 h under
a N2 atmosphere, then the solvents were removed under
vacuum. The residue was dissolved in methylene chlor-
ide (1 mL) and a solution of 4-phenoxyaniline (31.5 mg,
170 lmol) in methylene chloride (0.25 mL) was added.
The solution was stirred at 4 ꢁC overnight, diluted with
methylene chloride (15 mL), and washed twice with 1 N
HCl (15 mL). The organic layer was dried over Na2SO4,
concentrated to dryness, and the residue was purified by
flash chromatography (1:2 EtOAc–hexanes) to yield
compound 10b (44.8 mg, 55.8 lmol, 81%) as a colorless
4.3. Representative procedure for the synthesis of com-
pounds 11–14
4.3.1. Compound 12. A solution of 0.25 M NaOH (1 mL)
was added to compound 10b (11.1 mg, 13.8 lmol) and
the mixture was stirred at room temperature for 1 h. The
solution was washed twice with Et2O and the aqueous
layer was acidified with 0.5 M HCl to approximately
pH 2. The precipitate was collected by centrifugation
and then resuspended in water (1 mL). Addition of ethyl
acetate (1 mL) dissolved the precipitate and after the
layers were separated, the aqueous layer was extracted
once with ethyl acetate (1 mL). The ethyl acetate layer
was dried over Na2SO4 and concentrated to dryness to
yield compound 12 as a yellow solid (5.8 mg, 8.99 lmol,
1
solid: H NMR (300 MHz, acetone-d6) d 5.37 (s, 2H),
5.57 (s, 2H), 6.81 (m, 2H), 7.01 (d, J ¼ 8:8 Hz, 2H), 7.03
(d, J ¼ 8:9 Hz, 2H), 7.12 (t, J ¼ 7:5 Hz, 1H), 7.23 (d,
J ¼ 9:0 Hz, 2H), 7.29 (d, J ¼ 9:0 Hz, 2H), 7.38 (d,
J ¼ 7:5 Hz, 1H), 7.40 (d, J ¼ 7:5 Hz, 1H), 7.71 (dd,
J ¼ 7:9, 1.3 Hz, 1H), 7.76 (d, J ¼ 7:9 Hz, 1H), 7.83 (d,
J ¼ 9:0 Hz, 2H), 7.89 (s, 1H), 8.37 (d, J ¼ 9:0 Hz, 2H),
8.43 (d, J ¼ 9:0 Hz, 2H), 9.76 (s, 1H); 13C NMR
(75 MHz, acetone-d6) 68.0, 69.6, 92.4 (q, JCF ¼ 34:3 Hz),
115.8 (two overlapping resonances), 118.5, 119.8, 121.0,
121.1, 121.5 (q, JCF ¼ 280 Hz), 122.0, 123.4, 127.3,
129.4, 129.9, 130.2, 131.2, 131.3, 135.5 (two overlapping
resonances), 136.8, 137.1, 153.4, 158.2, 159.8, 163.5,
163.6, 163.7, 167.0; HRMS-FAB (M+Naþ) calcd for
C41H27F6NaN3O8 826.1600, found 826.1580.
1
65%): H NMR (300 MHz, acetone-d6) d 5.37 (s, 2H),
5.57 (s, 2H), 7.01 (d, J ¼ 8:6 Hz, 2H), 7.03 (d,
J ¼ 8:9 Hz, 2H), 7.12 (t, J ¼ 7:4 Hz, 1H), 7.20 (d,
J ¼ 8:9 Hz, 2H), 7.26 (d, J ¼ 8:9 Hz, 2H), 7.38 (d,
J ¼ 8:3 Hz, 1H), 7.40 (d, J ¼ 8:5 Hz, 1H), 7.70 (dd,
J ¼ 7:9, 1.4 Hz, 1H), 7.78 (d, J ¼ 8:0 Hz, 1H), 7.82 (d,
J ¼ 8:9 Hz, 2H), 7.88 (s, 1H), 8.02 (d, J ¼ 8:8 Hz, 2H),
8.07 (d, J ¼ 8:9 Hz, 2H), 9.76 (s, 1H); 13C NMR
(75 MHz, acetone-d6) d 68.1, 69.7, 115.7 (two overlap-
ping resonances), 118.5, 119.8, 121.9, 123.4, 126.0 (two
overlapping resonances), 127.4, 129.4, 129.9, 130.2,
132.6, 132.7, 135.4, 135.5, 136.7, 137.0, 153.4, 158.2,
164.5 (two overlapping resonances), 165.3, 167.0, 186.0;
LRMS-ESI (M)Hþ, negative ion mode) calcd for
C37H26NO10 644, found 644.
4.2.3. Compound 10c. Compound 10c was prepared from
compound 9 and 4-aminobenzophenone by procedures
analogous to the preparation of 10b except that flash
chromatography was performed with 1:3 EtOAc–hex-
1
anes. 10c (29.4 mg, 36 lmol, 71%): H NMR (300 MHz,
acetone-d6) d 5.37 (s, 2H), 5.58 (s, 2H), 6.81 (m, 2H),
7.22 (d, J ¼ 9:0 Hz, 2H), 7.29 (d, J ¼ 9:0 Hz, 2H), 7.57
(t, J ¼ 7:3 Hz, 1H), 7.67 (t, J ¼ 7:3 Hz, 1H), 7.74 (d,
J ¼ 8:1 Hz, 1H), 7.79 (m, 3H), 7.83 (d, J ¼ 8:7 Hz, 2H),
7.99 (d, J ¼ 8:7 Hz, 2H), 7.94 (s, 1H), 8.37 (d,
J ¼ 8:9 Hz, 2H), 8.43 (d, J ¼ 8:9 Hz, 2H), 10.09 (s, 1H);
13C NMR (75 MHz, acetone-d6) d 68.0, 69.6, 92.4 (q,
JCF ¼ 34:3 Hz), 115.8, 119.4, 121.1 (two overlapping
resonances), 121.5 (q, JCF ¼ 280 Hz), 127.4, 128.7, 129.5,
129.9, 130.2, 131.2, 131.3, 131.5, 132.5, 133.1, 135.7,
136.4, 137.3, 138.5, 143.6, 159.8, 163.5, 163.7, 167.6,
1
4.3.2. Compound 11. (6.3 mg, 10 lmol, 70%): H NMR
(300 MHz, acetone-d6) d 5.43, (s, 2H), 5.62 (s, 2H), 7.27
(m, 4H), 7.51 (m, 3H), 7.79 (m, 3H), 7.90 (d, J ¼ 7:4 Hz,
1H), 7.96 (d, J ¼ 7:4 Hz, 1H), 8.07 (m, 5H), 8.20 (d,
J ¼ 7:9 Hz, 1H), 9.78 (s, 1H); 13C NMR (75 MHz, ace-
tone-d6) d 68.3, 69.8, 115.7, 122.8, 123.0, 125.9 (two
overlapping resonances), 126.0, 126.4, 126.5, 127.7,
128.7, 128.9, 129.7, 130.0, 132.7 (two overlapping reso-
nances), 133.8, 134.7, 135.5, 136.8, 137.2, 164.5, 164.6,
165.2, 167.9, 186.0; LRMS-ESI (M)Hþ, negative ion
mode) calcd for C35H24NO9 602, found 602.
194.9;
C42H27F6NaN3O8 838.1600, found 838.1582.
HRMS-FAB
(M+Naþ)
calcd
for
4.3.3. Compound 13. (12.3 mg, 18.7 lmol, 52%): 1H
NMR (300 MHz, acetone-d6) d 5.39 (s, 2H), 5.59 (s, 2H),
7.20 (d, J ¼ 9:0 Hz, 2H), 7.27 (d, J ¼ 9:0 Hz, 2H), 7.57
4.2.4. Compound 10d. Compound 10d was prepared
from compound 9 and 1-aminopyrene by procedures

Y. T. Chen, C. T. Seto / Bioorg. Med. Chem. 12 (2004) 3289–3298
3297
(t, J ¼ 7:3 Hz, 2H), 7.67 (t, J ¼ 7:3 Hz, 1H), 7.78 (m,
4H), 7.83 (d, J ¼ 8:8 Hz, 2H), 7.93 (s, 1H), 8.00 (m, 4H),
8.07 (d, J ¼ 9:0 Hz, 2H), 10.08 (s, 1H); 13C NMR
(75 MHz, acetone-d6) d 68.1, 69.7, 115.7 (two overlap-
ping resonances), 119.4 (two overlapping resonances),
126.0, 126.1, 127.5, 128.7, 129.5, 129.9, 130.2, 131.5,
132.5, 132.6, 132.7, 133.1, 135.6, 136.4, 137.2, 164.4,
165.3, 167.6, 186.1, 194.9; LRMS-ESI (M)Hþ, negative
ion mode) calcd for C38H26NO10 656, found 656.
the ligand that was present in the original structure as
the basis for initial placement. This system was mini-
mized with 1000 iterations of conjugate gradient mini-
mization.
Acknowledgements
This research was supported by the NIH NIGMS
(Grant R01 GM057327). We thank Professor David
Cane for use of the SGI workstation.
4.3.4. Compound 14. (21.1 mg, 31.1 lmol, 53%): 1H
NMR (300 MHz, DMSO-d6) d 5.43 (s, 2H), 5.57 (s, 2H),
7.30 (m, 4H), 7.74 (m, 2H), 7.95 (m, 3H), 8.09 (m, 3H),
8.18 (m, 3H), 8.31 (m, 5H), 10.99 (s, 1H); 13C NMR
(75 MHz, DMSO-d6) d 68.6, 70.0, 116.3, 116.4, 123.6,
124.6, 125.2, 125.5, 125.7, 125.8, 125.9, 126.1, 126.2,
127.3, 127.8, 128.0, 128.1, 128.7, 129.8, 130.3, 130.5,
131.3, 131.6, 132.2, 133.0, 135.2, 137.3, 137.4, 159.8,
164.5, 164.6, 167.3, 168.2, 168.7, 187.9; LRMS-ESI
(M)Hþ, negative ion mode) calcd for C41H26NO9 676,
found 676.
References and notes
1. Hunter, T. Cell 2000, 100, 113.
2. Zhang, Z.-Y. Curr. Opin. Chem. Biol. 2001, 5, 416.
3. Goldstein, B. J.; Li, P.-M.; Ding, W.; Ahmad, F.; Zhang,
W.-R. Vitam. Horm. 1998, 54, 67.
4. Elchebly, 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.
4.4. PTPase assays
5. 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.; Shul-
man, G. I.; Neel, B. G.; Kahn, B. B. Mol. Cell. Biol. 2000,
20, 5479.
6. van Huijsduijnen, R. H.; Bombrun, A.; Swinnen, D. Drug
Discov. Today 2002, 7, 1013.
7. Thomas, M. L.; Brown, E. J. Immunol. Today 1999, 20,
406.
8. Bliska, J. B.; Guan, K. L.; Dixon, J. E.; Falkow, S. Proc.
Natl. Acad. Sci. U.S.A. 1991, 88, 1187.
9. Burke, T. R., Jr.; Zhang, Z.-Y. Biopolymers 1998, 47, 225.
10. Jia, Z.; Barford, D.; Flint, A. J.; Tonks, N. K. Science
1995, 268, 1754.
11. Vetter, S. W.; Keng, Y.-F.; Lawrence, D. S.; Zhang, Z.-Y.
J. Biol. Chem. 2000, 275, 2265.
12. 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.
The phosphatase activities of the Yersinia PTPase,
PTP1B, LAR, TCPTP, and CD45 (all purchased from
Calbiochem) were assayed using p-NPP as the substrate
at 25 ꢁC and the reaction progress was monitored by UV
spectroscopy. Initial rates were determined by moni-
toring the hydrolysis of p-NPP at 405 nm, from 30 to
150 s after mixing. Percent inhibition assay solutions
contained 100 mM acetate at pH 5.5, 1 mM EDTA,
100 mM NaCl, and 10% DMSO. Substrate concentra-
tions were kept at 2.5 and 1.2 mM for the Yersinia
PTPase and PTP1B assays, respectively. The Km values
under these conditions for the Yersinia PTPase and
PTP1B were found to be 2.5 and 0.62 mM, respectively.
IC50 assay solutions contained substrate at Km concen-
tration, 50 mM 3,3-dimethylglutarate at pH 7.0, 1 mM
EDTA, 50 mM NaCl, and 10% DMSO. The Km values
under these conditions for the Yersinia PTPase, PTP1B,
LAR, TCPTP, and CD45 were found to be 2.7, 2.0, 2.3,
2.1, and 7.0 mM, respectively. IC50 values were calcu-
lated using a Dixon analysis. Data analysis was per-
formed with the commercial graphing package Grafit
(Erithacus Software Ltd).
13. Salmeen, A.; Andersen, J. N.; Myers, M. D.; Tonks, N.
K.; Barford, D. Mol. Cell 2000, 6, 1401.
14. Taing, M.; Keng, Y.-F.; Shen, K.; Wu, L.; Lawrence, D.
S.; Zhang, Z.-Y. Biochemistry 1999, 38, 3793.
15. Shen, K.; Keng, Y.-F.; Wu, L.; Guo, X.-L.; Lawrence, D.
S.; Zhang, Z.-Y. J. Biol. Chem. 2001, 276, 47311.
16. Jia, Z.; Ye, Q.; Dinaut, A. N.; Wang, Q.; Waddleton, D.;
Payette, P.; Ramachandran, C.; Kennedy, B.; Hum, G.;
Taylor, S. D. J. Med. Chem. 2001, 44, 4584.
17. Guo, X.-L.; Shen, K.; Wang, F.; Lawrence, D. S.; Zhang,
Z.-Y. J. Biol. Chem. 2002, 277, 41014.
18. 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.
19. Asante-Appiah, E.; Patel, S.; Dufresne, C.; Roy, P.; Wang,
Q.; Patel, V.; Friesen, R. W.; Ramachandran, C.; Becker,
J. W.; Leblanc, Y.; Kennedy, B. P.; Scapin, G. Biochem-
istry 2002, 41, 9043.
4.5. Molecular modeling
Molecular modeling studies were performed using In-
sight II version 2000.1 (Accelrys, 9685 Scranton Rd.,
San Diego, CA 92121) running on a SGI workstation.
The coordinates from the crystal structure of PTP1B in
complex with [4-{4-[4-(difluorophosphonomethyl)-
phenyl]butyl}phenyl)difluoromethyl]phosphonic
acid
(PDB entry 1KAV) were used as the starting point for
calculations. Hydrogen atoms were added where neces-
sary, charges were fixed, and the calculations were per-
formed using the Discover module and the cvff force
field. Compound 2 was constructed using the Builder
module and manually docked into the active site using
20. Szczepankiewicz, B. G.; Liu, G.; Hajduk, P. J.; Abad-
Zapatero, C.; Pei, Z.; Xin, Z.; Lubben, T. H.; Trevillyan,
J. M.; Stashko, M. A.; Ballaron, S. J.; Liang, H.; Huang,
F.; Hutchins, C. W.; Fesik, S. W.; Jirousek, M. R. J. Am.
Chem. Soc. 2003, 125, 4087.

3298
Y. T. Chen, C. T. Seto / Bioorg. Med. Chem. 12 (2004) 3289–3298
21. Chen, Y. T.; Onaran, M. B.; Doss, C. J.; Seto, C. T.
Bioorg. Med. Chem. Lett. 2001, 11, 1935.
22. Chen, Y. T.; Xie, J.; Seto, C. T. J. Org. Chem. 2003, 68,
4123.
23. Chen, Y. T.; Seto, C. T. J. Med. Chem. 2002, 45, 3946.
24. Barnish, I. T.; Cross, P. E.; Danilewicz, J. C.; Dickinson,
R. P.; Stopher, D. A. J. Med. Chem. 1981, 24, 399.
25. Benniston, A. C.; Harriman, A. Synlett 1993, 3, 223.
26. van Huijsduijnen, R. H. Gene 1998, 225, 1.
27. Xin, Z.; Oost, T. K.; Abad-Zapatero, C.; Hajduk, P. J.;
Pei, Z.; Szczepankiewicz, B. G.; Hutchins, C. W.; Ballaron,
S. J.; Stashko, M. A.; Lubben, T.; Trevillyan, J. M.;
Jirousek, M. R.; Liu, G. Bioorg. Med. Chem. Lett. 2003,
13, 1887.
28. Pei, Z.; Li, X.; Liu, G.; Abad-Zapatero, C.; Lubben, T.;
Zhang, T.; Ballaron, S. J.; Hutchins, C. W.; Trevillyan, J.
M.; Jirousek, M. R. Bioorg. Med. Chem. Lett. 2003, 13,
3129.
29. Tsai, A. Y. M.; Itoh, M.; Streuli, M.; Thai, T.; Saito, H.
J. Biol. Chem. 1991, 266, 10534.
30. Saito, H.; Streuli, M.; Krueger, N. X.; Itoh, M.; Tsai,
A. Y. M. Biochem. Soc. Trans. 1992, 20, 165.