Potent non-peptidyl inhibitors of protein tyrosine phosphatase 1B

Article abstract of DOI:10.1016/S0968-0896(98)00075-3

The development of inhibitors of protein tyrosine phosphatases (PTPs) has recently been the subject of intensive investigation due to their potential as chemotherapeutics and as tools for studying signal transduction pathways. Here we report the evaluatio

Full text of DOI:10.1016/S0968-0896(98)00075-3

BIOORGANIC &
MEDICINAL
CHEMISTRY
Bioorganic & Medicinal Chemistry 6 (1998) 1457±1468
Potent Non-peptidyl Inhibitors of Protein Tyrosine
Phosphatase 1B
Scott D. Taylor,a,* Christopher C. Kotoris,a A. Nicole Dinaut,a
Qingping Wang,b Chidambaram Ramachandranb and Zheng Huangb
aDepartment of Chemistry, Erindale College, University of Toronto, 3359 Mississauga Rd. N., Mississauga, Ontario, Canada, L5L 1C6
bThe Merck-Frosst Center for Therapeutic Research, 16711 Trans Canada Highway, Kirkland, Quebec, Canada, H9H 3L1
Received 3 December 1997; accepted 13 February 1998
AbstractÐThe development of inhibitors of protein tyrosine phosphatases (PTPs) has recently been the subject of
intensive investigation due to their potential as chemotherapeutics and as tools for studying signal transduction path-
ways. Here we report the evaluation of a variety of small molecule, non-peptidyl inhibitors of protein tyrosine phos-
phatase 1B (PTP1B), bearing the a,a-di¯uoromethylenephosphonic acid (DFMP) group, a non-hydrolyzable
phosphate mimetic. A series of phenyl derivatives bearing a single DFMP group were initially surveyed. In general,
these were not signi®cantly more potent inhibitors than the parent compound, a,a-di¯uorobenzylphosphonic acid, with
the exception being the meta-phenyl substituted species which decreased the IC50 by approximately 17-fold relative to
a,a-di¯uorobenzylphosphonic acid. However, certain compounds bearing two DFMP moieties were very potent inhi-
bitors. Some of these are among the most potent small molecule inhibitors of any PTP reported to date with the best
one exhibiting a Ki of 1.5 mM. The structural basis for these results are discussed. One of the bis-DFMP inhibitors was
examined in detail and it was found that the ¯uorines were essential for potent inhibition. Inhibition was independent
of pH between pH 5.5±7.2 suggesting that both the mono and dianionic forms of the individual DFMP groups bind
equally well. The trends observed in the inhibitory potency of these compounds with PTP1B were very similar to the
trends observed by other workers on the Km's of the analogous phenylphosphate substrates with rat PTP1. This indi-
cates that studies of non-peptidyl substrates with rat PTP1 can be used as a guide for the development of human
PTP1B inhibitors. # 1998 Elsevier Science Ltd. All rights reserved.
Introduction
work in this area has focussed on preparing agents that
target the PTKs3,4 while much less e€ort has been
directed towards the PTPs. However, recent reports that
over expression of certain PTPs occurs in a number of
disease states1,5±8 have resulted in an increase in
research in this area, and a number of reports describing
PTP inhibitors have recently appeared.9±27
The phosphorylation and dephosphorylation of tyrosine
residues in proteins is now recognized as an important
cellular regulatory mechanism.1,2 The phosphorylation
event is catalyzed by protein tyrosine kinases (PTKs)
while the dephosphorylation is catalyzed by protein
tyrosine phosphatases (PTPs). It has been known for
several years now that the disruption of this process is at
the root of a variety of disease states including several
human cancers. Consequently, there has been much
interest in the development of therapeutic agents that
target these enzymes. Until fairly recently, most of the
The most common approach to designing PTP inhibi-
tors has been to incorporate non-hydrolyzable phos-
photyrosine (pTyr) mimetics in place of pTyr in
synthetic peptide substrates.9±13 Among the most potent
of this class of inhibitors were those containing
di¯uoromethylenephosphonyl phenylalanine9 (F2Pmp,
1). Certain peptides bearing 1 bind up to 103 times bet-
ter than the analogous peptide bearing methylenephos-
phonyl phenylalanine (Pmp). Some selectivity between
*Corresponding author. Fax: (905) 828-5425;
E-mail: staylor@credit.erin.utoronto.ca
0968-0896/98/$Ðsee front matter # 1998 Elsevier Science Ltd. All rights reserved
PII: S0968-0896(98)00075-3

1458
S. D. Taylor et al./Bioorg. Med. Chem. 6 (1998) 1457±1468
PTPs could be obtained by varying the amino acid
sequence around the F2Pmp group.13 Although this
approach has produced potent and relatively selective
PTP inhibitors, there are a number of drawbacks to
using peptidyl frameworks. Perhaps the most serious
of these shortcomings is their poor bioavailability: pep-
tidyl inhibitors are usually susceptible to proteolytic
degradation and often exhibit poor cellular uptake.
Another potential problem is the possibility that not all
PTPs recognize their in vivo substrates based solely
on the amino acid sequences that surrounds the pTyr
residue.
We are speci®cally interested in designing reversible,
small molecule inhibitors of PTP1B. PTP1B is a cyto-
solic PTP consisting of a single catalytic domain. It is
anchored by its C-terminus to the cytoplasmic face of
the ER. This is one of the most studied of all the PTPs
and several crystal structures of this enzyme complexed
with various ligands30,31 including the naphthyl deriva-
tive 2,23 have been reported. Although the precise in
vivo function of PTP1B has not yet been established,
there is now considerable evidence that this enzyme may
be involved in the down-regulation of insulin signaling
by dephosphorylating speci®c phosphotyrosine residues
on the insulin receptor kinase.32±39 PTP1B is now being
considered as a potential therapeutic target to help
combat certain forms of diabetes. For example, vana-
date, a phosphate analogue and a potent competitive
inhibitor of PTP1B19,21 has been shown to be an insulin-
mimetic and in human clinical trials to be potentially
useful in treating certain forms of diabetes.40 It has been
suggested that part of vanadate's insulin-mimetic e€ect
is due to its inhibition of PTPs.41 However, vanadate is
a non-speci®c phosphatase inhibitor and has a wide
variety of e€ects on biological systems.42 Consequently,
there has been much interest in obtaining speci®c inhi-
bitors of this enzyme since they could have potential
both as therapeutics and as tools for studying insulin-
dependent signal transduction pathways.
Several irreversible, non-peptidyl inhibitors of PTPs
have been reported, however, these are not speci®c for
any particular PTP14±19 and some also inhibit other
types of phosphatases. A number of metal-containing
reversible inhibitors such as gallium nitrate20 and vana-
date19,21 have been reported. These species, which most
likely function as mimics of inorganic phosphate or as
transition state analogues, are also known to inhibit
many other enzymes including several involved in signal
transduction pathways.28,29
Reports describing reversible, organic, small molecule
inhibitors of PTP1B are sparse. Sul®rcin, a sulfate-
bearing marine natural product, was shown to be a non-
speci®c inhibitor of certain phosphatases such as
Cdc25A (IC50=7.8 mM), V HR (IC50=4.7 mM) and
PTP1B (IC50=29.8 mM).27 A variety of sul®rcin analo-
gues have been synthesized and examined for PTP inhi-
bition, the best of which exhibit IC50's of 2.8 mM and
4.4 mM with Cdc25A and PTP1B, respectively.27 Burke
and co-workers examined a series of benzylic phospho-
nates as reversible inhibitors of rat PTP1, the rat analo-
gue of human PTP1B.22 From this series, the naphthyl
derivative, 2 bearing the di¯uoromethylenephosphonic
acid (DFMP) moiety, exhibited a Ki value of 179 mM.
However, 2 was also a good inhibitor of the serine/
threonine phosphatase PP2A, but was not examined
with other PTPs. Surprisingly, the phenyl derivative 3
was a poor inhibitor. On the basis of molecular model-
ing studies with 2 and PTP1B, Burke and co-workers
determined that the addition of a hydroxyl group at the
4-position of 2 would yield a more potent inhibitor.23
Thus, 4 was constructed via a 12-step synthesis,43 and,
consistent with the modeling predictions, found to be a
better inhibitor than 2. The e€ect, however, was very
modest, lowering the Ki by only twofold (93 mM) com-
pared to 2. More recently, we have shown that naphthyl
derivatives bearing two DFMP groups at positions 2
and 7 or positions 2 and 6 in the naphthyl ring are
There have been only a handful of reports describing
reversible, organic, small molecule inhibitors of PTPs.
Frechette and co-workers reported that certain functio-
nalized benzylic a-hydroxyphosphonic acids are good
inhibitors of CD45, the best of which exhibited an IC50
of 1.2 mM.24 Researchers at Merck have reported that
adrenolate (4-amino-1-hydroxybutylidene-1.1-bisphos
phonate), a potent inhibitor of bone resorption, is an
inhibitor of PTPe (IC50=2 mM).25 Nornuciferene, an
aporphine alkaloid has been reported to be an inhibitor
of CD45 with an IC50 of 5.3 mM.26

S. D. Taylor et al./Bioorg. Med. Chem. 6 (1998) 1457±1468
1459
approximately sixfold more potent inhibitors of PTP1B
than 2.44
the DFMP group. Although Burke and co-workers had
shown that compounds 2±4 were not particularly potent
PTP1B inhibitors, Frechette's studies24 with benzylic a-
hydroxyphosphonates and CD45 indicate that relatively
simple aryl phosphonates could be e€ective inhibitors of
PTPs. In addition, there have been a number of reports
describing small molecule aryl-bearing phosphonates as
excellent inhibitors of other enzymes that are capable of
utilizing phosphotyrosine such as osteoclastic acid
phosphatase50 and human prostatic acid phospha-
tase.51,52 We therefore reasoned that an appropriately
functionalized aryl DFMP derivative may exhibit potent
inhibition of PTP1B.
In this report, we examine a series of phenyl derivatives
bearing the DFMP moiety as inhibitors of PTP1B. In
general, we have found that the majority of phenyl
derivatives bearing a single DFMP group were not good
inhibitors of PTP1B and were not signi®cantly more
e€ective than the parent compound 3. However, certain
compounds bearing two aryl DFMP moieties are good
inhibitors of PTP1B and some of these are among the
most potent small molecule inhibitors of PTP1B repor-
ted to date.
Since the above mentioned studies with phenyl phos-
phate substrates and rat PTP1 indicated that ortho and
to a lesser extent meta substitution interferes with ligand
binding, we focussed mainly upon para-substituted
phenyl DFMP derivatives. The ®rst series of compounds
studied were phenyl DFMP derivatives substituted
mainly at the para-position with simple aromatic and
non-aromatic moieties (Table 1). An initial assessment
of the inhibitory e€ects of these derivatives was per-
formed using 500 mM inhibitor with PTP1B and FDP as
substrate at a concentration equal to its Km value
(20 mM) at pH 6.2. The results, in terms of percent inhi-
bition, are summarized in Table 1. Consistent with
Burke's studies,22 the unsubstituted compound, 3, is a
very poor inhibitor of PTP1B, resulting in only 41%
inhibition at 500 mM and has an IC50 of 610 mM with
PTP1B. Substitution at the para-position with non-aro-
matic moieties (5±8) did not signi®cantly improve the
inhibition and, in one instance (5) yielded a worse inhi-
bitor. Similarly, substitution at the para-position with
aromatic groups (9, 10 and 13) did not signi®cantly
improve the inhibition. However, addition of a phenyl
group at the meta-position (12) signi®cantly improved
Results and Discussion
As a starting point for the development of non-peptidyl
PTP1B inhibitors, we consulted the existing literature
regarding the active site speci®city of this enzyme or
related PTPs with both peptidyl and non-peptidyl sub-
strates and ligands. The presence of a phosphate group
appears to be essential for binding of tyrosyl-containing
peptides and proteins: PTPs do not bind tyrosine-con-
taining peptides or proteins that lack a phosphate
group45,46 nor do they bind O-methylphosphotyrosine-
containing peptides.45 Consequently, it is not surprising
that the majority of reversible, organic, PTP inhibitors
reported to date have incorporated into them a non-
hydrolyzable pTyr mimetic such as the DFMP moiety
or others such as thiophosphate,12 sulfate,11 and
methylenephosphonyl9 groups. However, the DFMP
group appears to be the most e€ective.9,13 Montserat et
al. and Zhang have studied the active site speci®city of
rat PTP1 with a number of non-peptidyl substrates.47,48
Rat PTP1 is the rat homologue of human PTP1B and
the catalytic domain of this enzyme (residues 1±322) has
97% sequence homology to the corresponding 322 resi-
dues in human PTP1B.49 In general, aromatic phos-
phates are signi®cantly better substrates than aliphatic
ones in terms of both higher kcat's and lower Km's. b-
naphthylphosphate has a lower Km than phenylphos-
phate (approximately two to ninefold, depending on the
pH). However, substitution of the phenylphosphate ring
with certain substituents can result in a lowering of the
Km to levels considerably less than that of b-naphthyl-
phosphate.47,48 In general, substituents on the phenyl
ring ortho and to a lesser extent meta to the phosphate
group have higher Km's compared to their para-sub-
stituted counterparts, and this has been attributed to
steric hindrance.47 Negatively charged substituents
remote from the phosphate group often signi®cantly
increase the anity of the enzyme for the substrate.
Table 1. Percent inhibitionof PTP1B with 500 mM, a,a-di¯uoro-
methylenephosphonates, 3, and 5±13
% Inhibitiona
3, X=H
41‹2
32‹2
64‹2
71‹1
73‹2
68‹1
66‹2
25‹2
94‹1
67‹1
5, X=p-COOH
6, X=p-Br
7, X=p-N3
8, X=p-NO2
9, X=p-COOCH2Ph
10, X=p-COPh
11, X=o-Ph
12, X=m-Ph
13, X=p-Ph
In light of the above substrate and inhibitor studies, we
decided to focus our e€orts on aryl derivatives bearing
aErrors are reported as‹SD.

1460
S. D. Taylor et al./Bioorg. Med. Chem. 6 (1998) 1457±1468
the inhibitor potency. This appeared to be entirely
unique to the meta-substitution since the para-phenyl
and ortho-phenyl substituted derivatives (11 and 13)
were moderate (para-substituted) to very poor (ortho-
substituted) inhibitors. The meta-phenyl derivative (12)
has an IC50 of 35 mM and is therefore more potent than
naphthyl derivatives 2 and 4. The IC50 for the para-
phenyl derivative (13) was only 210 mM. Since the per-
cent inhibition of compounds 6±10 was very similar to
13, it is likely that they would exhibit similar IC50's.
We noted that the results in Table 1 followed the same
trend as that observed by Montserat et al. for the Km's
of the analogous phosphate substrates with rat PTP1.47
For example, these workers have reported that p-car-
boxyphenylphosphate exhibits a higher Km (2.1 mM)
than p-nitrophenyl phosphate (0.68 mM) with rat PTP1,
while our studies show that the para-carboxy-a,a-
di¯uorobenzylphosphonic acid 5 is a poorer inhibitor of
human PTP1B than the para-nitro-a,a-di¯uoro-
benzylphosphonic acid 8. Although meta-substituted
phenylphosphates were usually poorer substrates than
their para-substituted analogues with rat PTP1,47 meta-
phenyl phenylphosphate was an exception. The Km
reported for para-phenyl phenylphosphate, 2.4 mM, is
approximately seven times greater than that reported
for meta-phenyl phenylphosphate (0.32 mM).47 Simi-
larly, the IC50 determined for para-phenyl-a,a-di¯uoro-
benzylphosphonic acid 13 (210 mM) is six times
substrates. These results are consistent with our earlier
studies where we demonstrated that ¯uoresceindipho-
sphate (16) has a signi®cantly lower Km and hydrolysis
of the ®rst phosphate group occurs much faster than
¯uoresceinmonophosphate (17) with PTP1B.53 Conse-
quently, we reasoned that an additional DFMP group
on inhibitors such as 12 and 13 may yield very potent
PTP1B inhibitors.
We initially tested this hypothesis with compound 12
since it was the most potent inhibitor amongst those
reported in Table 1. However, the symmetrical bis-
DFMP derivative 18 was a slightly poorer inhibitor than
the mono-DFMP compound 12 (Table 2). This was not
greater
than
meta-phenyl-a,a-di¯uorobenzylphos-
phonic acid 12 (35 mM). However, non-peptidyl sub-
strates meta-substituted with aliphatic hydrocarbons
(ethyl, isopropyl) exhibited higher Km's than meta-phenyl
phenylphosphate with rat PTP1.47 This suggests that the
e€ect observed with the meta-phenyl-substituted com-
pounds may involve more than just simple hydrophobic
interactions and may possibly involve pi-stacking inter-
actions with an aromatic residue located close to the
active site.
The above results suggest that studies using phenyl-
based, non-peptidyl substrates with rat PTP1 could be
used as a guide for the development of human PTP1B
inhibitors bearing the DFMP group. We therefore deci-
ded to examine additional DFMP-bearing structures
based on substrate studies with rat PTP1. An important
observation made by Montserat et al. was that phenyl-
phosphate substrates bearing an additional charged
moiety (phosphate or carboxylate) remote from the
phosphate group exhibited very low Km's with rat
PTP1.47 This was especially evident for certain symme-
trical substrates bearing two phosphate groups and
oneof these (14) is the most potent low molecular weight
substrate ever reported for a PTP and exhibits a Km
(16 mM) 19 times lower than its mono-phosphorylated
analogue (15) and is almost as low as the best peptidyl

S. D. Taylor et al./Bioorg. Med. Chem. 6 (1998) 1457±1468
1461
Table 2. IC50 Values for phosphonates, 12, 13 and 18±22, with
PTP1B
Phosphonate
IC50 (mM)
12
13
18
19
20
22
35‹1
210‹5
45‹1
15‹1
2025‹15
51.5‹0.3
aErrors are reported as ‹SD.
the case with the symmetrical bis-DFMP analogues of
10 and 13 (compounds 19 and 20) which completely
inhibited PTP1B at 500 mM and exhibited IC50's of
9.9 mM and 15 mM, respectively (Table 2). Thus, in the
case of the biphenyl derivative (20) the additional
DFMP group decreased the IC50 by about 14-fold. The
inhibition was competitive and a Ki of 4.5‹1 mM
(Fig. 1(a) and (b)) was determined for 20.
To determine whether the presence of the ¯uorines was
absolutely necessary for potent inhibition, the non-
¯uorinated analogue of 20, compound 21, was exam-
ined. 21 was a very poor inhibitor of PTP1B and exhib-
ited an IC50 of about 2 mM (Table 2) approximately
200-fold greater than that of 20. To ascertain whether
the a-¯uorines were essential on both phosphonate
groups the bis-phosphonate 22, in which only one of the
benzylic phosphonates was a-¯uorinated, was exam-
ined. This compound exhibited an IC50 of 51 mM, which
is only ®vefold greater than the fully ¯uorinated species
20 but is approximately 40-fold less than the IC50
obtained with the non-¯uorinated species 21. Thus, for
this biphenyl series of bis-phosphonates, two ¯uorines
alpha to just one of the phosphonate groups enhance
the enzymes anity for the bis-phosphonate by
approximately 40-fold. However, the addition of two
more a-¯uorines on the other phosphonate group
results in a mere ®vefold anity enhancement.
Figure 1. Inhibition of PTP1B by compound 20. (a) The activ-
ity of PTP1B (0.15 mg/mL) was measured at pH 6.2 as described
under `Experimental Procedures' in the presence of the follow-
ing concentrations of 20: (*), 0 mM; (&), 5 mM; (~), 8 mM;
(!), 10 mM; (^), 12 mM; ( ), 15 mM. (b) Replot of the slope
from the double-recipricol plot versus concentrations of com-
pound 20.
require the dianionic form of the phosphonates for
binding. At pH 6±7, the non-¯uorinated phosphonates
exist only partially as the dianions while the ¯uorinated
derivatives have almost completely ionized13 and there-
fore bind tighter. However, Chen et al. have shown that
F2Pmp-bearing peptides inhibit rat PTP1-catalyzed
dephosphorylation of pTyr-containing peptides in a pH
independent manner between pH 5.0±7.0.13 This indi-
cates that the dianionic and monoanionic forms of these
inhibitors bind equally well and that inhibition is not
due to pKa e€ects. The crystal structure of 2 complexed
with PTP1B has recently been reported.23 The two
¯uorine atoms interact with the phenyl ring of Phe182
via Van der Waals contacts and the Pro-R ¯uorine
forms H-bonds with the amido group of Phe182. To
determine whether the ¯uorines in 20 enhance binding
via a pKa e€ect or some other type of interaction such
as H-bonding, we measured the inhibition constant of
20 as a function of pH. The second ionization constant
The enhanced binding e€ect of the ¯uorines alpha to the
phosphonate group in peptidyl and non-peptidyl PTP1B
and rat PTP1 inhibitors varies depending on the inhi-
bitor. With certain F2Pmp-bearing peptides this e€ect is
dramatic and can result in an over 1000-fold decrease in
Ki compared to their Pmp analogues.9 The e€ect seen
here with compounds 19±22 is less dramatic but never-
theless quite signi®cant (approximately 200-fold
di€erence in IC50 between 20 and 21). Why are a-¯uoro-
phosphonates signi®cantly better inhibitors than their
non-¯uorinated counterparts? Two possible explana-
tions have been put forward.13 One is that the ¯uorines
may increase the anity by H-bonding with speci®c
residues. The alternative explanation is that the PTPs

1462
S. D. Taylor et al./Bioorg. Med. Chem. 6 (1998) 1457±1468
(pKa) of a,a-di¯uorobenzylphosphonic acid has been
determined54 to be 5.71 and it is likely that the second
ionization constants of the individual DFMP groups on
20 will be similar. Thus, we determined the Ki of 20 with
PTP1B between pH 5.5 and 7.2. We found that the Ki of
20 changed very little in this pH range (data not shown)
which suggests that both the monoanionic and dianionic
forms of the individual DFMP groups on 20 bind to
PTP1B equally well suggesting that the ¯uorines in 20
enhance binding by interacting with speci®c residues on
the enzyme.
the other occupying the low anity non-catalytic site.
Instead, 14 binds in two mutually exclusive modes. One
of the phosphates of 14 binds in either the catalytic site
or the non-catalytic site while the second phosphate
group is bound by a combination of electrostatic,
hydrophobic, aromatic-aromatic and water-mediated
hydrogen bonds in an area slightly removed from both
the catalytic and non-catalytic sites. It is possible that
one of the DFMP groups in compounds 19, 20 and 23±
26 occupies the high anity catalytic site while the other
DFMP group occupies the low anity site. However,
this is unlikely for all six of these inhibitors since our
results indicate that increasing the distance between the
two DFMP groups does not dramatically increase or
decrease their inhibitory potency relative to one
another. Studies using synthetic peptidyl substrates
indicate that a number of PTPs, including PTP1B, have
Montserat et al. reported that diphenylmethyl derived
phosphates exhibited lower Km's than biphenylphos-
phate derivatives.47 Thus, we examined the bis-DFMP
diphenylmethyl derivative (23) as a PTP1B inhibitor
(Table 3) and obtained an IC50 (23 mM) that was slightly
greater than that of 20. The length of the hydrocarbon
chain linking the two aryl groups was increased and these
species were examined (24±26, Table 3) for PTP1B inhibi-
tion. The 1,2-diphenylethane derivative (24) was a better
inhibitor than 20 yielding an IC50 of 6.8 mM. Any further
increase in the length of the hydrocarbon linker arm
resulted in only slightly better inhibitors than (24) with the
1,4-diphenylbutane derivative (26) being the best, exhi-
biting an IC50 of 4.4 mM and was a potent competitive
inhibitor with a Ki of 1.5mM (Fig. 2(a) and (b)).
It is remarkable that although a,a-di¯uorobenzylphos-
phonic acid 3 is a very poor inhibitor of PTP1B, simply
by joining two such derivatives together, much more
potent inhibitors of PTP1B can be obtained. Why does
the presence of more than one phosphonate or phos-
phate group enhance the binding of both substrates and
inhibitors to PTPs? Very recently, X-ray crystallo-
graphic analysis of PTP1B grown in the presence of
compound 14 or saturating pTyr revealed a second aryl
phosphate-binding site.55 This is a low-anity, non-
catalytic binding site adjacent to the active site, with
Arg-24 and Arg-254 forming important ionic interac-
tions with the phosphate group. These studies55 also
reveal that the high anity of 14 is not a result of one of
the phosphate groups occupying the catalytic site and
Table 3. IC50 Values for bis-phosphonates, 23±26, with
PTP1B
IC50 (mM)
Figure 2. Inhibition of PTP1B by compound 26. (a) The activ-
ity of PTP1B (0.15 mg/mL) was measured at pH 6.2 as described
under `Experimental Procedures' in the presence of the follow-
ing concentrations of 26: (*), 0 mM; (&, 1 mM; (~), 3 mM;
(!), 5 mM; (^), 7 mM; ( ), 9 mM. (b) Replot of the slope from
the double-reciprocal plot versus concentrations of compound
26.
23, n=1
24, n=2
25, n=3
26, n=4
23‹2
6.8‹0.2
5.8‹0.3
4.4‹0.4
aErrors are reported as ‹SD.

S. D. Taylor et al./Bioorg. Med. Chem. 6 (1998) 1457±1468
1463
a strong preference for acidic residues N-terminal to
phosphotyrosine and other physiologically occurring
pTyr proteins exhibit this motif.56±59 The crystal struc-
ture of the Cys215Ser PTP1B mutant complexed with
DADE(pY)L, a high anity peptide substrate, reveals
that the preference for acidic residues at positions P-1
and P-2 is mainly a result of salt bridges between Arg47
and the side chains of the Asp and Glu at the P-1 and P-
2 positions in the peptide.30 It is also likely that there are
additional residues besides Arg47 that are capable of
interacting with anionic groups since it is known that
the presence of acidic residues at positions P-3 and P-4
in peptidyl substrates also enhances their anity for rat
PTP1 and PTP1B.56,58 Thus, the high anity of inhibi-
tors 19, 20 and 23±26 may be a result of interactions of
the second DFMP moiety with residues that do not
constitute either phosphate-binding sites as was the case
with one of the phosphate groups on 14.
PTP1B. Combining these two observations in a single
compound should produce highly potent PTP1B inhibi-
tors. The trends we observed in the inhibitory potency
of the DFMP compounds with PTP1B were very similar
to the trends observed by other workers47,48 on the Km's
of the analogous phenylphosphate substrates with rat
PTP1. This indicates that studies with non-peptidyl
substrates with rat PTP1 can be used as a guide for the
development of human PTP1B inhibitors. Studies by
Kole et al. suggest that inhibitors bearing the DFMP
group are incapable of penetrating cellular mem-
branes.22 Consequently, we are in the process of con-
verting our most potent inhibitors (such as 26) into cell
permeable, fully substituted phosphate esters that can
be activated by light or intracellular enzymes and these
`caged' compounds will be used for cellular studies to
study kinase-dependent signal transduction pathways.
Although we have established that many of the low
molecular weight compounds described here can serve
as potent PTP inhibitors, it still remains to be seen if
inhibitors such as 20 and 23±26 are selective for PTP1B
compared to other PTPs. The crystal structures of
PTP1B complexed with inhibitors 20 and 26 is currently
underway60 and these studies will allow us to determine
how these inhibitors bind to PTP1B and will be invalu-
able for the further development of yet more potent
PTP1B inhibitors.
A number of compounds reported in this study (12, 19,
20, 23±26) are good inhibitors of PTP1B and are sig-
ni®cantly better inhibitors of PTP1B than naphthyl
DFMP inhibitors 2 and 4. Under our assay conditions,
naphthyl derivative 2 exhibits an IC50 of 95‹10 mM
with PTP1B. This is approximately 24-fold greater than
that exhibited by 26. It should also be pointed out that
Burke and co-workers have reported that 2 is a good
inhibitor of the serine/threonine phosphatase PP2A.22
We have also examined 2 with PP2A and found that,
under our assay conditions, it has an IC50 of 753 mM
which indicates that 2 is approximately eight times less
potent with PP2A than with PTP1B. We also examined
26 with PP2A and found that it has an IC50 of
200‹8 mM which is approximately 45 times greater than
26 with PTP1B. Thus, 26 is considerably more potent
and selective than 2 in terms of selectivity between PP2A
and PTP1B. Unfortunately, inhibition studies with 4
and PP2A were not reported. It is also important to
note that the inhibitors obtained in this study are readily
accessible, starting from commercially available starting
materials via synthetic routes that require only three±
®ve steps.
Experimental
General. Unless otherwise noted, all starting materials
for syntheses were obtained from commercial suppliers
(Aldrich or Lancaster) and were used without further
puri®cation. Tetrahydrofuran (THF) and diethylether
(ether) were distilled from sodium/benzophenone ketyl
under argon. Dichloromethane was distilled from calcium
hydride under argon. DMF was distilled under reduced
pressure from calcium hydride and stored over 4-A sieves
under argon. Reactions involving moisture-sensitive
reagents were executed under an inert atmosphere of dry
argon. Flash chromatography was performed using
silica gel 60 (Toronto Research Chemicals, 230±400
mesh ASTM). 1H, 31P, and 19F NMR spectra were
recorded on a Varian 200-Gemini NMR machine at
approximately 200, 80, and 188 MHz, respectively.
Unless stated otherwise, all 13C spectra were recorded
on a Varian-500 or Varian-400 at approximately 125
and 100 MHz, respectively. For 1H NMR spectra run in
CDCl3, chemical shifts (d) are reported in parts per
million relative to the internal standard tetra-
Conclusion
In summary, we have examined a series of phenyl
DFMP derivatives as inhibitors of PTP1B. Several of
the inhibitors examined in this study exhibit Ki values in
the low mM range and are among the most potent,
reversible, organic, non-peptidyl PTP inhibitors repor-
ted to date. The most potent of these compounds were
those bearing two DFMP groups and the ¯uorines were
essential for potent inhibition. In addition to having two
DFMP groups, the presence of a phenyl group meta to a
phenylDFMP moiety signi®cantly enhances binding to
1
methylsilane (TMS). For H NMR spectra run in D2O,
chemical shifts (d) are reported in parts per million
relative to the HDO peak at d 4.68. For 13C spectra run
in CDCl3, chemical shifts are reported in parts per mil-
lion relative to the CDCl3 residual carbons (d 77.0 for

1464
S. D. Taylor et al./Bioorg. Med. Chem. 6 (1998) 1457±1468
the central peak). For 13C spectra run in D2O, chemical
shifts are reported in parts per million relative to the
CH3 peak of 3-(trimethylsilyl)-1-propanesulfonic acid
(external). For 31P NMR spectra, chemical shifts are
reported in parts per million relative to 85% phosphoric
acid (external). For 19F NMR, chemical shifts are
reported in parts per million relative to tri¯uoroacetic
acid (external). Electron impact (EI) and fast atom
bombardment (FAB) mass spectra (negative ion) were
obtained on a Micromass 70-S-250 mass spectrometer.
All inhibitors described in this report were examined
for purity by HPLC (Waters Prep LC 4000 System
equipped with a Vydac 218TP54 C-18 reverse phase
column using 0.1% TFA in water/acetonitrile) and
exhibited a single peak in the HPLC chromatogram.
Fluorescein diphosphate (FDP) substrate was synthe-
sized as described.61 Bu€er chemicals and bovine
serum albumin (BSA) were obtained from Sigma
Chemical Company. PP2A was purchased from
Upstate Biotechnology. Expression and puri®cation of
FLAG-PTP1B has been described elsewhere.19 Enzyme
assay solutions were prepared with deionized/distilled
water.
30 min at which time a white precipitate formed. The
organic layer was evaporated and the aqueous layer
lyophilized several times leaving 5 as a white solid in
near quantitative yields. 1H NMR (D2O): d 7.89 (d,
J=8.8 Hz, 2H, aryl), 7.63 (d, J=8.8 Hz, 2H, aryl); 31P
NMR (D2O): d 4.94 (t, JPF=97.7 Hz); 19F NMR (D2O):
d 8.08 (d, JFP=97.7 Hz); 13C NMR (D2O): d 175.7 (s,
CO), 139.0 (td, JCF=22.0 Hz, JCP=11.7 Hz, aryl), 138.4
(s, aryl), 129.6 (s, aryl), 126.8 (t, JCF=6.6 Hz, aryl),
121.9 (td, JCF=261.8 Hz, JCP=189.2 Hz, CF2);
FABMS: m/z 251 (M 3+2H+).
4-[(Diethylphosphono)di¯uoromethyl]azidobenzene (28).
A
mixture of 4-[(diethylphosphono)di¯uoromethyl]-
nitrobenzene62 (0.53 g, 1.71 mmol) and zinc dust
(5 equiv) in TFA (25 mL) was stirred overnight. The
reaction was ®ltered, cooled to 0 ꢀC in an ice bath,
NaNO2 (0.24 g, 2 equiv) was added and the mixture
stirred in the dark for 30 min. NaN3 (1.11 g, 10 equiv)
was added over a period of 2 min and the mixture stirred
for an additional 15 min. Ether (50 mL) was added and
the mixture stirred for 1 h. The reaction mixture was
then washed with water (1Â50 mL) and the organic
layer dried (Na2SO4) and concentrated leaving an oily
yellow residue. Silica gel chromatography (EtOAc/
CH2Cl2, 1:9, Rf=0.9) gave pure 28 (76% yield) as a
Inhibitor syntheses. The synthesis of inhibitors 3, 6, 8±
13, 19, 20, 23, 25 and 26, have been reported else-
where.62 The synthesis of inhibitors 5, 7, 18, 21, 22 and
24 are described below.
1
yellow oil. H NMR (CDCl3): d 7.59 (d, J=8.1 Hz, 2H,
aryl), 7.09 (d, J=8.4 Hz, 2H, aryl), 4.20 (m, 4H,
2(CH2)), 1.32 (t, J=7.0 Hz, 6H, 2(CH3)); 31P NMR
(CDCl3): d 3.86 (t, JPF=117.0 Hz); 19F NMR (CDCl3):
4-[(Dimethylphosphono)di¯uoromethyl]benzoic acid (27).
A mixture of 962 (300 mg, 0.81 mmol) and approxi-
mately 15 mg of 10% Pd/C, in methanol (5 mL), was
stirred overnight under H2 (1 atm). The mixture was ®l-
tered through Celite and concentrated by rotary eva-
poration to give 27 (90% yield) as a white solid
(mp=124±126 ꢀC). No further puri®cation was neces-
sary. 1H NMR (CDCl3): d 8.18 (d, J=8.8 Hz, 2H, aryl),
7.74 (d, J=8.8 Hz, 2H, aryl), 3.87 (d, J=10.3 Hz, 6H,
2(CH3)); 31P NMR (CDCl3): d 5.99 (t, JPF=114.5 Hz);
19F NMR (CDCl3): d 29.82 (d, JFP=114.5 Hz); 13C
d
31.9 (d, JFP=117.0 Hz); 13C NMR (CDCl3): d 142.7
(d, J=2.2 Hz, aryl), 129.0 (m, aryl), 127.9 (td,
JCF=6.6 Hz, JCP=2.2 Hz, aryl), 118.9 (s, aryl), 117.7
(td, JCF=262.9 Hz, JCP=219.7 Hz, CF2), 64.7 (d,
JCP=6.6 Hz, 2(CH2)), 16.3 (d, JCP=5.1 Hz, 2(CH3)); IR
1
(neat): 2149.1 cm
.
EIMS: m/z 306 (MH+), 169
(MH+-PO(OEt)2), 109 (MH+-197); HRMS(EI): calcd
for C11H14F2N3O3P (M+) m/z 305.0747, found
305.0741.
NMR (CDCl3):
d
167.6 (s, COOH), 136.8 (td,
4-[(Phosphono)di¯uoromethyl]azidobenzene, ammonium
salt (7). TMSBr (0.32 mL, 3 equiv) was added to a
solution of 28 (250 mg, 0.82 mmol) in anhydrous
CH2Cl2 (5 mL) and re¯uxed for 36 h. The reaction was
concentrated by rotary evaporation and the residual oil
placed under high vacuum for 2 h. The product was then
re-dissolved in CH2Cl2 (3 mL) and 3 equiv NH4HCO3
(0.20 g in 3 mL H2O) was added. The biphasic mixture
was rapidly stirred for 30 min at which time a white
precipitate formed. The organic and aqueous layers
were separated and the aqueous layer lyophilized several
times yielding pure 7 as a white solid. 1H NMR (D2O): d
7.60 (d, J=8.4 Hz, 2H, aryl), 7.18 (d, J=8.5 Hz, 2H,
aryl); 31P NMR (D2O): d 4.62 (t, JPF=104.5 Hz); 19F
NMR (D2O): d 29.2 (d, JFP=104.5 Hz); 13C NMR
(D2O): d (142.6, 131.6, 128.4, 119.7 {s, aryl}), 120.8 (td,
JCF=21.8 Hz, JCP=13.7 Hz, aryl), 133.9 (s, aryl), 130.1
(s, aryl), 126.5 (td, JCF=6.6 Hz, JCP=2.2 Hz, aryl),
118.3 (td, JCF=262.9 Hz, JCP=219.7 Hz, CF2), 55.2 (d,
JCP=7.3 Hz, 2(CH3)); EIMS: m/z 280 (M+), 171 (M+-
PO(OMe)2); HRMS(EI): calcd for C10H11F2O5P (M+)
m/z 280.0302, found 280.0312.
4-[(Phosphono)di¯uoromethyl]benzoic acid, ammonium
salt (5). Trimethylsilylbromide (TMSBr, 0.24 mL, 4
equiv) was added to 27 (130 mg, 0.46 mmol) dissolved in
anhydrous CH2Cl2 (5 mL) and stirred for 6 h. The mix-
ture was then concentrated down and placed under high
vacuum for 2 h. The product was then re-dissolved in
CH2Cl2 (3 mL) and 4 equiv NH4HCO3 (0.146 g in 3 mL
H2O) was added. The mixture was stirred vigorously for

S. D. Taylor et al./Bioorg. Med. Chem. 6 (1998) 1457±1468
1465
JCF=256.6 Hz, JCP=199.2 Hz, CF2); IR (KBr disc):
2147.3 cm 1; FABMS: m/z 248 (M 2+1H+).
then concentrated down and placed under high vacuum
for 1.5 h. The product was then re-dissolved in CH2Cl2
(2 mL) and 6.2 equiv NH4HCO3 (0.083 g in 2 mL H2O)
was added. The biphasic mixture was rapidly stirred for
30 min, the organic layer was removed by rotary eva-
poration and the aqueous layer lyophilized several times
to remove excess NH4HCO3, yielding pure 18 as a white
powder (99% yield). 1H NMR (D2O): d 7.91 (s, 2H,
aryl), 7.77 (s, 2H, aryl), 7.58 (m, 4H, aryl); 31P NMR
(D2O): d 6.90 (bs); 19F NMR (D2O): d 27.3 (d,
JFP=94.6 Hz);13C NMR: d 140.78 (s, aryl), 137.40 (btd,
aryl), 129.83 (s, aryl), 129.23 (s, aryl), 126.19 (bt, aryl),
125.33 (bt, aryl); FABMS: m/z 413 (M 4+3H+).
3,30-Bis(phosphonomethyl)biphenyl (29). A solution of
3,30-bis(bromomethyl)biphenyl63 (1.0 g, 2.9 mmol) and
trimethylphosphite (3.5 mL, 29.4 mmol, 10 equiv) in dry
benzene (4 mL) was re¯uxed for 12 h. The excess tri-
methyl-phosphite and dimethylmethylphosphonate
(formed during the reaction) were removed by vacuum
distillation and the residue was puri®ed by silica gel
chromatography (MeOH:CHCl3, 5:95, Rf=0.3) to give
1
pure 29 as a clear, colorless oil (85% yield). H NMR
(CDCl3): d 7.35 (m, 8H, aryl), 3.66 (d, J=11.7 Hz, 12H,
4(OCH3)), 3.19 (d, J=20.5 Hz, 4H, CH2P); 31P NMR
(CDCl3): d 26.75 (s); 13C NMR (CDCl3): d 141.0 (s,
aryl), 131.9 (d, JCP=9.1 Hz, aryl), 128.6 (m, aryl), 125.6
(d, JCP=3.7 Hz, aryl), 52.6 (d, JCP=6.4 Hz, -OCH3),
32.8 (d, JCP=138.2 Hz, -CH2P); d EI(MS) m/z 398
(M+), 289 (M+- PO(OMe)2); HRMS (EI) calcd for
C18H24O6P2 (M+) m/z 398.1048, found 398.1044.
4,40-bis(phosphonomethyl)biphenyl, ammonium salt (21).
TMSBr (0.33 mL, 6 equiv) was added to 4,40-bis(phos-
phonomethyl)biphenyl62 (200 mg, 0.425 mmol) dissolved
in anhydrous CH2Cl2 (3 mL) and stirred for 12 h. The
mixture was then concentrated down and placed under
high vacuum for 1.5 h. The product was then re-dis-
solved in CH2Cl2 (3 mL) and 8 equiv NH4HCO3
(0.209 g in 3 mL H2O) was added. The biphasic mixture
was rapidly stirred for 30 min, the organic layer was
removed by rotary evaporation and the aqueous layer
lyophilized several times to remove excess NH4HCO3,
yielding pure 21 as a white powder (99% yield).
1H NMR (D2O): d 7.69 (d, J=7.7 Hz, 4H, aryl), 7.45
(d, J=8.0 Hz, 4H, aryl), 3.10 (d, J=20.8 Hz, 4H,
CH2P); 31P NMR (D2O): d 21.00 (s); 13C NMR (D2O):
d 151.6 (d, JCP=2.9 Hz, aryl), 148.2 (d, JCP=8.6 Hz,
aryl), 143.7 (d, JCP=5.7 Hz, aryl), 140.2 (s, aryl), 49.1
3,30-Bis[(dimethylphosphono)di¯uoromethyl]biphenyl (30).
To
a
solution of sodium hexamethyldisilazide
(NaHMDS, Aldrich, 4.15 mL, 1.0 M in THF, 5.5 equiv)
in dry THF (4.0 mL) at 78 ꢀC was added a solution of
29 (0.3 g, 0.754 mmol, 1.0 equiv) in dry THF (30 mL)
over a period of 2 min. The resulting orange to dark red
solution was stirred for 1 h at 78 ꢀC and then a solu-
tion of N-¯uorobenzenesulfonimide (NFBS, 1.73 g, 7.3
equiv) in dry THF (4.0 mL) was added over a period of
2 min, during which time the solution turned from dark
red to light yellow. After addition, the solution was
stirred for 1.5 h and then allowed to warm to 30 ꢀC
during which time a precipitate formed. The reaction
was quenched with 0.01 N HCl (50 mL) and the result-
ing solution (precipitate dissolves) was extracted with
EtOAc (3Â50 mL). The organics were combined and
washed with 5% NaHCO3 (1Â75 mL), brine (1Â75 mL),
dried (MgSO4), and concentrated by rotary evaporation
to give a yellow oil that was puri®ed via silica gel chro-
matography (EtOAc:hexanes, 4:1, Rf=0.3) to give 30 as
4
(d, JCP=128.6 Hz, CH2P); FABMS: m/z 341 (M
+3H+).
4-[4-((Diethylphosphono)di¯uoromethyl)benzene]toluene
(31). To 4-[(diethylphosphono)di¯uoromethyl]bromo-
benzene62 (410 mg, 1.21 mmol) in deaerated ethanol
(10 mL), was added 4-methylbenzeneboronic acid
(0.25 g, 1.5 equiv), Pd(OAc)2 (0.011 g, 0.042 equiv), and
solid Na2CO3 (0.19 g, 1.5 equiv). The dark brown reac-
tion mixture was stirred at room temperature for 36 h,
under argon, diluted with ether (100 mL) and ®ltered.
The ®ltrate was washed with 1 N NaOH (1Â100 mL),
brine (1Â100 mL), dried (MgSO4) and concentrated
down by rotary evaporation. The product was puri®ed
using silica gel chromatography (EtOAc:hexane, 3:7,
Rf=0.3) to give 31 as a white solid (71% yield). 1H
NMR (CDCl3): d 7.67 (s, 4H, aryl), 7.51 (d, J=7.3 Hz,
2H, aryl), 7.27 (d, J=5.9 Hz, 2H, aryl), 4.21 (m, 4H,
2(CH2)), 2.41 (s, 3H, CH3Ar), 1.34 (t, J=6.6 Hz, 6H,
2(CH3)); 31P NMR (CDCl3): d 4.39 (t, JPF=117.5 Hz);
19F NMR (CDCl3): d 32.2 (d, JFP=117.5 Hz); 13C
NMR (CDCl3): d (143.7, 137.8, 137.2, 129.6, 127.0,
126.8, 126.7, 126.6 {s, aryl}), 118.3 (weak td, CF2), 64.6
(d, JCP=7.4 Hz, CH2), 20.93 (s, CH3Ar), 16.2 (d,
JCP=4.6 Hz, CH3); EIMS: m/z 354 (M+), 217 (M+
1
a pale-yellow oil (42% yield). H NMR (CDCl3): d 7.30
(m, 8H, aryl), 3.87 (d, J=10.3 Hz, 12H, 4(OCH3)); 31P
NMR (CDCl3): d 6.60 (t, JPF=116 Hz); 19F NMR
(CDCl3): d 31.5 (d, JFP=116 Hz); 13C NMR (CDCl3):
d 140.6 (s, aryl), 136.0 (m, aryl), 129.7 (s, aryl), 129.1 (s,
aryl), 125.5 (btd, aryl), 124.9 (btd, aryl), 118.1 (td,
JCF=263.5 Hz, JCP=215.9 Hz, CF2), 54.6 (d,
JCP=7.3 Hz, OCH3); EI(MS) m/z 470 (M+), 361 (M+-
PO(OMe)2); HRMS (EI) calcd for C18H20F4O6P2 (M+)
m/z 470.0671, found 470.0671.
3,3(-Bis[(phosphono)di¯uoromethyl]biphenyl, ammonium
salt (18). TMSBr (0.19 mL, 8.5 equiv) was added to a
solution of 30 (0.08 g, 0.170 mmol) in anhydrous
CH2Cl2 (2 mL) and stirred for 12 h. The mixture was

1466
S. D. Taylor et al./Bioorg. Med. Chem. 6 (1998) 1457±1468
-PO(OEt)2); HRMS(EI): calcd for C18H21O3F2P (M+)
m/z 354.1194, found 354.1196.
(s, aryl), 136.5 (d, JCP=9.6 Hz, aryl), 136.5 (m, aryl),
131.2 (d, JCP=5.9 Hz, aryl), 127.9 (s, aryl), 127.6 (bt,
aryl), 127.3 (s, aryl), 123.0 (weak td, CF2), 36.6 (d,
JCP=126.7 Hz, CH2P); FABMS: m/z 377 (M 4+3H+).
4-[4-((Diethylphosphono)di¯uoromethyl)benzene]diethyl
benzylphosphonate (32). N-Bromosuccinimide (1 equiv,
0.14 g) was added in several portions to a solution of 31
(275 mg, 0.78 mmol) and benzoyl peroxide (0.0075
equiv, 0.0014 g) in anhydrous CCl4 (25 mL). Another
portion of benzoyl peroxide (0.0014 g) was added
towards the ®nal addition of NBS. The mixture was
re¯uxed for 2 h, cooled to room temperature, washed
with water (2Â15 mL), dried (MgSO4), and con-
centrated down by rotary evaporation. The crude reac-
tion mixture was passed through a silica gel column
(hexane:EtOAc, 6:4, Rf=0.4) yielding a partially pur-
i®ed mixture of monobrominated (ꢁ80% of the total)
and dibrominated (ꢁ5% of the total) as a colorless oil.
This crude mixture (170 mg) was then added to excess
triethyl phosphite (10 mL) and re¯uxed for 10 h. The
excess phosphite and diethyl ethylphosphonate (formed
during the reaction) was removed by vacuum distillation
and the product was puri®ed via silica gel chromato-
graphy (EtOAc, Rf=0.5) to give 32 as a yellow oil (73%
1,2(-Bis[4-((dimethylphosphono)methyl)benzene]ethane (33).
1,2-Bis[4-(bromomethyl)benzene]ethane64 (0.8 g, 2.17
mmol) was added to trimethylphosphite (2.6 mL,
21.7 mmol, 10 equiv) in benzene (5 mL). The resulting
solution was re¯uxed for 12 h. Unreacted phosphite and
dimethyl methylphosphonate (formed during the reac-
tion) were removed by vacuum distillation and the resi-
due was subjected to silica gel chromatography (MeOH/
CHCl3 4:96, Rf=0.4) to give pure 33 as a white solid
(75% yield). 1H NMR (CDCl3): d 7.15 (m, 8H, aryl),
3.66 (d, JHP=10.6 Hz, 12H, 4(OCH3)), 3.13 (d,
JHP=21.6 Hz, 4H, 2(CH2P)), 2.88 (s, 4H, aryl-(CH2)2-
aryl); 31P NMR (CDCl3): d 26.8 (s); 13C NMR (CDCl3):
d 140.35 (d, JCP=3.6 Hz, aryl), 129.61 (d, JCP=6.6 Hz,
aryl), 128.72 (d, JCP=2.9 Hz, aryl), 128.62 (s, aryl),
52.81 (d, JCP=7.3 Hz, OCH3), 37.32 (s, aryl-(CH2)2-
aryl), 32.45 (d, JCP=138.4 Hz, CH2P); EIMS m/z 426
(M+), 213 (M+ -213); HRMS (EI) calcd for
C20H28O6P2 (M+) m/z 426.1361, found 426.1366.
1
yield). H NMR (CDCl3): d 7.68 (s, 4H aryl), 7.56 (d,
J=7.4 Hz, 2H, aryl), 7.39 (d, J=8.8 Hz, 2H, aryl), 4.14
(m, 8H, 4(CH2)), 3.21 (d, JHP=22.0 Hz, 2H, CH2P),
1.31 (m, 12H, 4(CH3)); 31P NMR (CDCl3): d 23.88 (s),
4.30 (t, Jꢁ117 Hz); 19F NMR (CDCl3): d 32.1 (d,
JFPꢁ117 Hz); 13C NMR (CDCl3): d 143.1 (s, aryl), 138.5
(d, JCP=3.7 Hz, aryl), 131.6 (d, JCP=9.6 Hz, aryl),
131.4 (m, aryl), 130.3 (d, JCP=6.6 Hz, aryl), 127.3 (d,
JCP=3.6 Hz, aryl), 126.9 (s, aryl), 126.7 (td, JCF=6.6,
JCP=2.2 Hz, aryl), 118.1 (weak td, JCF=263.3,
1,20 -Bis[4-((dimethylphosphono)di¯uoromethyl)benzene]-
ethane (34). To a solution of NaHMDS (Aldrich, 1.0 M
in THF, 4.26 mL, 4.26 mmol, 5.5 equiv) in dry THF
(5 mL) was added a solution of 33 (0.33 g, 0.775 mmol, 1
equiv) in dry THF (7 mL), dropwise at 78 ꢀC over a
period of 2 min. The color of the solution changed from
light yellow to red, and a ®ne suspension formed. After
1 h of stirring at 78 ꢀC, a solution of NFBS (1.91 g,
5.65 mmol, 7.3 equiv) in dry THF (4.3 mL) was added
over a period of 2 min, during which time the suspension
became a solution and the color changed from red to
yellow. After addition, the solution was stirred for 1±2 h
and was allowed to warm to 30 ꢀC, at which point a
precipitate formed. The reaction was quenched with
0.01 N HCl (50 mL, precipitate dissolves) and was
extracted with EtOAc (3Â50 mL). The organics were
combined and washed with 5% NaHCO3, brine, dried
(MgSO4) and concentrated by rotary evaporation to
give a yellow oil which was puri®ed via ¯ash chromato-
graphy (hexane:EtOAc, 1:4, Rf=0.3) to give pure 34 as
JCP=218.9 Hz, CF2), 64.7 (d, J =7.4 Hz, CH2), 62.1
CP
(d, JCP=6.6 Hz, CH2), 33.5 (d, JCP=138.4 Hz, CH2P),
16.4 (d, JCP=5.1 Hz, 2(CH3)), 16.3 (d, JCP=5.1 Hz,
2(CH3)); EIMS: m/z 490 (M+), 353 (M+ -PO(OEt)2);
HMRS(EI): calcd for C22H30F2O6P2 (M+) m/z
490.1472, found 490.1486.
4-[4-((Phosphono)di¯uoromethyl)benzene]benzylphosphonic
acid, ammonium salt (22). TMSBr (0.08 mL, 6 equiv)
was added to 32 (50 mg, 0.10 mmol) dissolved in anhy-
drous CH2Cl2 (4 mL) and re¯uxed for 36 h. The mixture
was then concentrated down and placed under high
vacuum for 2 h. The product was then re-dissolved in
CH2Cl2 (3 mL) and 8 equiv NH4HCO3 (0.065 g in 3 mL
H2O) was added. The mixture was stirred for 30 min at
which time a white precipitate formed. The organic and
aqueous layers were separated and the aqueous layer
lyophilized several times to remove excess NH4HCO3,
yielding pure 22 as a white powder (88% yield). 1H
NMR (D2O): d 7.66 (bs, 6H, aryl), 7.39 (bs, 2H, aryl),
3.03 (d, JHP=20.5 Hz, 2H, CH2P); 31P NMR (D2O): d
21.21 (s), 6.24 (bs); 19F NMR (D2O): d 37.14 (d,
JFP=94.6 Hz); 13C NMR (D2O): d 142.3 (s, aryl), 138.5
1
a white solid (37% yield). H NMR (CDCl3): d 7.50 (d,
J=8.0 Hz, 4H, aryl), 7.23 (d, J=8.0 Hz, 4H, aryl), 3.79
(d, JHP=10.6 Hz, 12H, 4(OCH3)), 2.97 (s, 4H, aryl-
(CH2)2-aryl); 31P NMR (CDCl3):
d
d
6.64 (t,
31.5 (d,
JPF=117.5 Hz); 19F NMR (CDCl3):
JFP=117.5 Hz); 13C NMR (CDCl3): d 144.20 (s, aryl),
130.10 (btd, aryl), 128.62 (s, aryl), 126.13 (bt, aryl),
118.16 (weak td, CF2), 54.87 (d, JCP=5.9 Hz, OCH3),
37.15 (s, aryl-(CH2)2-aryl); EIMS m/z 498 (M+), 389
(M+ -PO(OMe)2), 249 (M+ -249); HRMS (EI) calcd for
C20H24F4O6P2 (M+) m/z 498.0984, found 498.0973.

S. D. Taylor et al./Bioorg. Med. Chem. 6 (1998) 1457±1468
1467
1,20 -Bis[4-((phosphono)di¯uoromethyl)benzene]ethane,
ammonium salt (24). To solution of 0.633 g
Kinetic studies with PP2A. Kinetic studies with PP2A
were performed using FDP as substrate in 25 mM Tris±
HCl, pH 6.5, containing 5 mM DTT, 0.5 mM MnCl2
and 0.1 mg/mL of BSA. Assay were carried out in 96-
well plates with total volume of 200 mL and rates were
obtained by continuously monitoring the ¯uorescence
from the production of FMP with excitation at 440 nm (slit
width 20 nm) and emission at 530 nm (slit width 25 nm)
using a Cyto¯uor II plate reader (Perseptive Biosystems).
Reactions were initiated by the addition of PP2A (®nal
concentration 0.13 mg/mL based on the commercially
reported protein content). Under these conditions, the
FDP has a Km of 35 mM and Vspec of 0.35 mmol/mg/min.
The Vspec/Km of 0.01 mmol/mg/min/mM is slightly lower
than what was reported for the best known phosphory-
lated protein substrate.65 The IC50's of 2 and 26 were
determined in duplicate at 10 di€erent inhibitor con-
centrations with FDP at Km concentration (35 mM).
a
(0.127 mmol, 1 equiv) of 34 in dry CH2Cl2 (2 mL) was
added TMSBr (0.1 mL, 0.763 mmol, 6 equiv). The solu-
tion was stirred at room temperature for 12 h. Following
removal of CH2Cl2 the residue was subjected to high
vacuum for several hours. The oil was re-dissolved in
CH2Cl2 (2 mL) and a solution of NH4HCO3 (0.062 g in
3 mL water, 6.2 equiv) was added. The biphasic mixture
was stirred vigorously for 30±60 min and the CH2Cl2
layer was removed by rotary evaporation. The aqueous
layer was then lyophilized three to four times, yielding
24 as a white powder in near quantitative yield. 1H
NMR (D2O): d 7.5 (d, J=8.0 Hz, 4H, aryl), 7.31 (d,
J=8.1 Hz, 4H, aryl), 2.99 (s, 4H, aryl-(CH2)2-aryl); 31P
NMR (D2O): d 5.84 (t, JPF=93.9 Hz); 19F NMR (D2O):
d
27.41 (d, JFP=96.1 Hz); 13C NMR (D2O): d 144.06
(s, aryl), 134.87 (btd, aryl), 129.17 (s, aryl), 126.82 (bt,
JCF=6.4 Hz, aryl), 122.71 (weak td, CF2), 36.99 (s, aryl-
(CH2)2-aryl); FABMS: m/z 441 (M 4+3H+).
Acknowledgements
Kinetic studies with PTP1B. Rates of PTP1B-catalyzed
dephosphorylation in the presence or absence of inhibi-
tors were determined using FDP as substrate53 in assay
bu€er containing 50 mM bis-Tris, 2 mM EDTA, 5 mM
DTT and 0.2 mg/mL BSA, at 25 ꢀC. Stock solutions of
inhibitors were prepared in assay bu€er and found to be
stable under these conditions inde®nitely. Assays were
carried out in 1 mL cuvettes with total volumes of
700 mL. Reactions were initiated by the addition of
PTP1B (®nal concentration 0.15 mg/mL). Rates were
obtained by continuously monitoring the production of
FMP at 450 nM using a Varian Cary 1 spectro-
photometer. The percentage inhibition in the presence
of inhibitors were performed in duplicate using FDP at
Km concentration (20 mM) with 500 mM inhibitor at pH
6.2. IC50 determinations were determined in duplicate at
®ve or six di€erent inhibitor concentrations with FDP at
Km concentration at pH 6.2. Ki's for 20 and 26 were
determined by measuring the initial rate (v) using var-
ious FDP concentrations (20, 25, 35, 50 and 100 mM
FDP for 20 and 25, 35, 50 and 100 mM FDP for 26) at
various ®xed concentrations of inhibitor (0, 5, 8, 10, 12
and 15 mM of 20 and 0, 1, 3, 5, 7 and 9 mM of 26). 1/v
versus 1/[S] was plotted as shown in the following
equation:
We would like to thank the Natural Sciences and Engi-
neering Research Council of Canada (NSERC) for
®nancial support for S.D.T. through an operating grant
and for A.N.D. for a postgraduate scholarship. We
would also like to thank the University of Toronto for
an University of Toronto Open Fellowship for C.C.K.
References and Notes
1. Sun, H.; Tonks, N. K. Trends Biochem. Sci. 1994, 19, 480.
2. Tonks, N. K. Adv. Pharmacol. 1996, 36, 91.
3. Burke, T. R. Drugs of the Future 1992, 17, 119.
4. Levitzki, A.; Gazit, L. Science 1995, 267, 1782.
5. Ide, R.; Maegawa, H.; Kikkawa, R.; Shigetta, Y.; Kashi-
wagi, A. Biochem. Biophys. Res. Commun. 1994, 201, 77.
6. Weiner, J. R.; Hurteau, J. A.; Kerns, B. J.; Whitaker, R. S.;
Conaway, M. R.; Berchuck, A.; Bast, R. C. Am. J. Obstet.
Gynecol. 1994, 170, 1171.
7. Zheng, X. M.; Wang, Y.; Pallen, C. J. Nature 1992, 359, 336.
8. Pallen, C. J. Semin. Cell Biol. 1993, 4, 403.
9. Burke, T. R.; Kole, H. K.; Roller, P. P. Biochem. Biophys.
Res. Commun. 1994, 204, 129.
10. Kole, H. K.; Akamatsu, M.; Ye, B.; Yan, X.; Barford, D.;
Roller, P. P.; Burke, T. R. Biochem. Biophys. Res. Commun.
1995, 209, 817.
È
É
1=v ˆ ꢀKm=Vmax†ꢀ1 ‡ ‰IŠ=Ki† 1=‰SŠ ‡ 1=Vmax
11. Liotta, A. S.; Kole, H. K.; Fales, H. M.; Roth, J.; Bernier,
M. J. Biol. Chem. 1994, 269, 22996.
The slopes of these plots ꢂꢂKm=Vmax†ꢂ1 ‡ ‰IŠ=Ki†† were
determined using the program SigmaPlot. These slopes
were replotted against the concentration of inhibitor
using the SigmaPlot program and the Ki's were obtained
from the x-intercepts of these replots. All Ki studies
were determined using bis-Tris as bu€er and were per-
formed in duplicate.
12. Hiriyanna, K. T.; Baedke, D.; Baek, K-H.; Forney, B. A.;
Kordiyak, G.; Ingebritsen, T. S. Anal. Biochem. 1994, 223, 51.
13. Chen, L.; Wu, L.; Otaka, A.; Smyth, M. S.; Roller, P. R.;
Burke, T. R.; den Hertog, J.; Zhang, Z.-Y. Biochem. Biophys.
Res. Comm. 1995, 216, 976.
14. Taylor, W. P.; Zhang, Z.-Y.; Widlanski, T. S. Bioorg. Med.
Chem. 1996, 4, 1515.

1468
S. D. Taylor et al./Bioorg. Med. Chem. 6 (1998) 1457±1468
15. Yu, L.; McGill, A.; Ramirez, J.; Wang, P. G., Zhang, Z.-Y.
Bioorg. Med. Chem. Lett. 1995, 5, 1003.
40. Gold®ne, A. B.; Simonson, D. C.; Folli, F.; Patti, M.-E.;
Kahn, C. R. J. Clin. Endocrinol. Metab. 1995, 80, 3311.
16. Zhang, Z.-Y., Davis, J. P.; Van Etten, R. L. Biochemistry
1992, 31, 1701.
41. Shechter, Y. Diabetes 1990, 39, 1.
42. Shechter, Y.; Meyerovitch, J.; Farfel, Z.; Sack, J.; Bruck,
R.; Bar-Meir, S.; Amir, S.; Denagi, H.; Karlish, S. J. D. In
Vanadium in Biological Systems; Chasteen, N. D., Ed.; Kluwer
Academic: Netherlands; 1990; pp 129±142.
17. Wang, Q.; Cechert, U.; Jirik, F; Withers, S. G. Biochem.
Biophys. Res. Commun. 1994, 200, 577.
18. Ghosh, J.; Miller, R. A. Biochem. Biophys. Res. Commun.
1993, 194, 36.
43. Ye, B.; Burke, T. R. Jr. Tetrahedron 1996, 52, 9963.
19. Huyer, G.; Liu, S.; Kelly, J.; Mo€at, J.; Payette, P.; Ken-
nedy, B.; Tsaprailis, G.; Gresser, M.; Ramachandran, C. J.
Biol. Chem. 1997, 272, 843.
44. Wang, Q.; Huang, Z.; Ramachandran, C.; Dinaut, A. N.;
Taylor, S. D. Bioorg. Med. Chem. Lett. 1998, 8, 345.
45. Ruzzene, M.; Donella-Deana, A.; Marin, O.; Perich, J. W.;
Ruzza, P.; Borin, G,; Calderan, A.; Pinna, L. A. Eur. J. Bio-
chem., 1993, 211, 289.
20. Berggren, M. M.; Burns, L. A.; Abraham, R. T.; Powis, G.
Cancer Res. 1993, 53, 1862.
21. Posner, B.; I., Faure, R.; Burgess, J. N. W.; Bevan, A. P.;
Lachance, D.; Zhang-Sun, G.; Fantus, I. G.; Ng, J. B.; Hall, D.
A.; Lum, B. S.; Shaver, A. J. J. Biol. Chem. 1994, 269, 4596.
46. Zhang, Z.-Y.; Maclean, D.; Thieme-Selfer, A. M.; Roeske,
R.; Dixon, J. E. Anal. Biochem. 1993, 211, 7.
47. Montserat, J.; Chen, L.; Lawrence, D. S.; Zhang, Z.-Y. J.
Biol. Chem. 1996, 271, 7868.
22. Kole, K. H.; Smyth, M. S.; Russ, P. L.; Burke, T. R. Bio-
chem. J. 1995, 311, 1025.
48. Zhang, Z.-Y. J. Biol. Chem. 1994, 270, 11199.
23. Burke, T. R.; Ye. B.; Yan, X.; Wang, S.; Jia, Z., Chen, L.;
Zhang, Z.-Y.; Barford, D. Biochemistry 1996, 35, 15898.
49. Guan, K. L.; Dixon, J. E. J. Biol. Chem. 1994, 266, 17026.
50. Schwender, C. F.; Beers, S. A.; Malloy, E.; Demarest, K.;
Minor, L.; Lau, K. W. H. Bioorg. Med. Chem. Lett. 1995, 5, 1801.
24. Frechette, R. F., Ackerman, C.; Beers, S.; Look, R., Moore
Bioorg. Med. Chem. Lett. 1997, 7, 2169.
51. Schwender, C. F.; Beers, S. A.; Malloy, E.; Cinicola, J. J.;
Wustrow, D. J.; Demarest, K.; Jordan, J. Bioorg. Med. Chem.
Lett. 1996, 6, 311.
25. Schmidt, A.; Rutledge, S. J.; Endo, N.; Opas, N. E.;
Tanaka, H.; Wesolowski, G.; Leu, C. T.; Huang, Z.; Rama-
chandran, C.; Rodan, S.; Rodan, G. A. Proc. Natl. Acad. Sci.
U.S.A. 1996, 93, 3068.
52. Beers, S. A.; Schwender, C. F.; Loughney, D. A.; Malloy,
E.; Demarest, K.; Jordan, J. Bioorg. Med. Chem. 1996, 4, 1693.
26. Miski, M.; Shen, X.; Cooper, R.; Gillum, A. M.; Fisher, D.
K.; Miller, R. W. Higgins, T. J. Bioorg. Med. Chem. Lett. 1995,
5, 1519.
53. Desmarais, S.; Govindarajan, A.; Ramachandran, C;
Zamboni, R.; Abdullah, K.; Huang, Z. FASEB J. 1995, 9,
A1347.
27. Cebula, R. E.; Blanchard, J. L.; Boisclair, M. D.; Pal, K.;
Bockovich, N. J. Bioorg. Med. Chem. Lett. 1997, 7, 2015.
54. Smyth, M. S.; Ford, H. Jr.; Burke, T. R. Jr. Tetrahedron.
Lett. 1992, 33, 4137.
28. Gorden, J. A. Meth. Enzymol. 1991, 264, 7747.
55. 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.
29. Nechay, B. R.; Nanniga, L. B.; Nechay, P. S. E.; Post, R. L.;
Granttham, J. J.; Macara, I. G.; Kubena, L. F.; Phillips, T. D.;
Nielsen, F. H. Fed. Proc. Fed. Am. Soc. Exp. Biol. 1986, 45, 123.
56. Agostinis, P.; Donella-Deana, A.; Van Hoof, C.; Cesaro,
L.; Brunati, A. M.; Ruzzene, M.; Merlevede, W.; Pinna, L. A.,
Goris, J. Eur. J. Biochem. 1996, 236, 548.
30. Jia, Z.; Barford, D.; Flint, A. J.; Tonks, N. K. Science
1995, 268, 1754.
31. Barford, D.; Flint, A. J.; Tonks, N. K. Science 1994, 263,
1397.
57. Happen, K. L.; Jakes, S.; Richards, J.; Jena, B. P.; Beck, B.
L.; Tabatabai, L. B; Ingebritsen, T. S. Biochemistry 1993, 32,
12405.
32. Goldstein, B. J. Receptor 1993, 2, 1.
33. Kusari, J.; Kenner, K. A.; Suh, K.-I.; Hill, D. E; Henry, R.
R. J. Clin. Invest. 1994, 293, 1156.
58. Zhang, Z.-Y.; Thieme-Selfer, A. M.; Maclean, D.; McNa-
mara, D. J.; McNamara, D. J.; Dobrusiin, E. M.; Sawyer, T.
K.; Dixon, J. E. Proc. Natl. Acad. Sci. U.S.A 1993, 90, 4446.
34. Kenner K. A.; Hill, D. E.; Olefsky, J. M.; Kusari, J. J. Biol.
Chem. 1993, 268, 25455.
59. Zhang, Z.-Y.; Maclean, D.; McNamara, D. J.; McNamara,
D. J.; Sawyer, T. K.; Dixon, J. E. Biochemistry, 1994, 33, 2285.
35. Maegawa, H.; Ide, R.; Hasegawa, M.; Ugi, S.; Egawa, K.;
Iwanishi, M.; Kikkawa, R.; Shigeta, Y.; Kashiwagi, A. J. Biol.
Chem. 1995, 270, 7724.
60. Crystal structures of PTP1B complexed with 20 and 26 are
being performed by Professor Zongchao Jia, Department of
Biochemistry, Queen's University.
36. Kenner, K. A.; Anyanwu, E.; Olefsky, J. M; Kusari, J. J.
Biol. Chem. 1996, 271, 19810.
61. Scheigetz, J., Gilbert, M. and Zamboni, R. Org. Prep.
Proc. Intl. 1997, 29, 561.
37. Ahmad, F.; Li, P.-M.; Goldstein, B. J. J. Biol. Chem. 1995,
270, 20503.
62. Taylor, S. D.; Kotoris, C. C.; Dinaut, A. N.; Chen, M.-J.
Tetrahedron 1998, 54, 1691.
38. Bandyopadhyay, D.; Kusari, A.; Kenner, K. A.; Liu, F.;
Cherno€, J.; Gustafson, T. A.; Kusari, J. J. Biol. Chem. 1997,
272, 1639.
63. Wenner, J. J. Org. Chem. 1952, 17, 523.
64. Wagner, J.; Bendig, J; Kreysig, D. Z. Chem. 1983, 11, 407.
39. Rhamachandran, C.; Aebersold, R.; Tonks, N. K.; Pot, D.
A. Biochemistry 1992, 31, 4232.
65. Zhang, Z.; Zhao, S.; Long, F.; Zhang, L.; Bai, G.; Shima,
H.; Nagao, M.; Lee E. Y. C. J. Biol. Chem. 1994, 269, 16997.