Structure of protein tyrosine phosphatase 1B in complex with inhibitors bearing two phosphotyrosine mimetics

Article abstract of DOI:10.1021/jm010266w

Protein tyrosine phosphatases (PTPases) are signal-transducing enzymes that dephosphorylate intracellular proteins that have phosphorylated tyrosine residues. It has been demonstrated that protein tyrosine phosphatase 1B (PTP1B) is an attractive therapeut

Full text of DOI:10.1021/jm010266w

4584
J . Med. Chem. 2001, 44, 4584-4594
Str u ctu r e of P r otein Tyr osin e P h osp h a ta se 1B in Com p lex w ith In h ibitor s
Bea r in g Tw o P h osp h otyr osin e Mim etics
Zongchao J ia,† Qilu Ye,† A. Nicole Dinaut,‡ Qingping Wang,§ Deena Waddleton,§ Paul Payette,§
Chidambaram Ramachandran,§ Brian Kennedy,§ Gabriel Hum,‡ and Scott D. Taylor*,‡
Department of Chemistry, University of Waterloo, Waterloo, Ontario, N2L 3G1, Canada, Department of Biochemistry,
Queen’s University, Kingston, Ontario, K7L 3N6, Canada, and Department of Biochemistry and Molecular Biology,
Merck-Frosst Center for Therapeutic Research, Pointe-Claire, Dorval, Quebec, H9R 4P8, Canada
Received J une 14, 2001
Protein tyrosine phosphatases (PTPases) are signal-transducing enzymes that dephosphorylate
intracellular proteins that have phosphorylated tyrosine residues. It has been demonstrated
that protein tyrosine phosphatase 1B (PTP1B) is an attractive therapeutic target because of
its involvement in regulating insulin sensitivity (Elcheby et al. Science 1999, 283, 1544-1548).
The identification of a second binding site in PTP1B (Puius et al. , Proc. Natl. Acad. Sci. U.S.A.
1997, 94, 13420-13425) suggests a new strategy for inhibitor design, where appropriate
compounds may be made to simultaneously occupy both binding sites to gain much higher
affinity and selectivity. To test this hypothesis and gain further insights into the structural
basis of inhibitor binding, we have determined the crystal structure of PTP1B complexed with
two non-peptidyl inhibitors, 4 and 5, both of which contain two aryl difluoromethylenephos-
phonic acid groups, a nonhydrolyzable phosphate mimetic. The structures were determined
and refined to 2.35 and 2.50 Å resolution, respectively. Although one of the inhibitors seems
to have satisfied the perceived requirement for dual binding, it did not bind both the active
site and the adjacent noncatalytic binding site as expected. The second or distal phosphonate
group instead extends into the solvent and makes water-mediated interactions with Arg-47.
The selectivity of the more potent of these two inhibitors, as well as four other inhibitors bearing
two such phosphate mimetics for PTP1B versus seven other PTPases, was examined. In general,
selectivity was modest to good when compared to PTPases Cdc25a, PTPmeg-1, PTPâ, and CD45.
However, selectivity was generally poor when compared to other PTPases such as SHP-1, SHP-
2, and especially TCPTP, for which almost no selectivity was found. The implications these
results have concerning the utility of dual-binding inhibitors are discussed.
Protein tyrosine phosphatases (PTPases) are signal-
transducing enzymes that dephosphorylate intracellular
proteins phosphorylated at tyrosine residues. Together
with protein tyrosine kinases, which catalyze the op-
posing protein phosphorylation reactions, protein ty-
rosine phosphatases regulate the level and extent of
protein tyrosine phosphorylation and play vital roles in
regulating many and diverse cellular processes. These
include cell growth, proliferation and differentiation,
metabolism, immune responses, cell-cell adhesion, and
cell-matrix contacts.1-5
Protein tyrosine phosphatases are classified as intra-
cellular PTPases or receptor-like PTPases depending on
the cellular localization of these enzymes.1-5 Among
PTPs, human protein tyrosine phosphatase 1B (PTP1B)
has been the most extensively studied. The enzyme is
widely expressed in insulin-sensitive tissues and has
been implicated in the attenuation of insulin signaling.6
A recent transgenic experiment has demonstrated that
PTP1B plays a key regulatory role in modulating both
insulin sensitivity and resistance to diet-induced obe-
sity.7 Thus, it is being increasingly recognized as a
potential therapeutic target for the treatment of type 2
diabetes and obesity. In addition, PTP1B may also be a
key regulatory component of other pathways. For ex-
ample, overexpression of PTP1B has been shown to
occur in a significant percentage of breast cancer
patients.8 It has also been demonstrated that PTP1B
dephosphorylates p130Cas, suggesting that it may have
a regulatory role in mitogen-mediated signal transduc-
tion pathways.9 Consequently, there is tremendous
interest in obtaining inhibitors of this enzyme.10,11
Over the past several years, numerous reports have
appeared in the literature describing inhibitors of
PTP1B.11 Although a number of highly potent or selec-
tive inhibitors of PTP1B have been prepared, combining
both of these features in a single molecule is still
problematic. Early attempts to prepare both potent and
selective PTP1B inhibitors involved preparing peptidyl
structures in which the labile phosphotyrosine (pTyr)
residue is replaced with a nonhydrolyzable pTyr mi-
metic.10 The most effective of these mimetics was the
difluoromethylphosphonic acid (DFMP) group, first
utilized by Burke and co-workers.12,13 Certain peptides
bearing difluoromethylphosphonophenylalanine (F2-
Pmp, 1) exhibited affinities for PTP1B that were more
than an order of magnitude greater than the analogous
peptide substrates and exhibited some selectivity for
PTP1B over other PTPases examined. Later, Moran et
* To whom correspondence should be addressed: phone, (519)
888-4567, ext 3325; fax, (519) 746-0435; e-mail, s5taylor@sciborg.
uwaterloo.ca.
† Queen’s University.
‡ University of Waterloo.
§ Merck-Frosst Center for Therapeutic Research.
10.1021/jm010266w CCC: $20.00 © 2001 American Chemical Society
Published on Web 11/28/2001

Tyrosine Phosphatase 1B
J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 26 4585
al reported that certain tripeptides bearing a cinnamic
acid moiety were potent inhibitors of PTP1B, exhibiting
phosphate group binds by forming ionic interactions
with Arg-254 and Arg-24 and water-mediated hydrogen
bonds with Met-258 and Gly-259. Although with satu-
rating pTyr both sites could be occupied simultaneously,
only 50% of the pTyr was bound as pTyrB. These results
suggest that PTP1B contains one high-affinity catalytic
aryl phosphate binding site and, adjacent to it, one low-
affinity noncatalytic aryl phosphate binding site in
which Arg-254 and Arg-24 are key residues. Compound
2 was found to bind to PTP1B in two mutually exclusive
modes. In one mode, the one phosphate group binds in
the catalytic site while the other phosphate makes a
water-mediated hydrogen bond with Gln-262. In the
other mode, one phosphate again makes a water-
mediated hydrogen bond with Gln-262. However, the
other phosphate group binds in the second aryl phos-
phate binding site. Each site was approximately 50%
occupied. Two molecules of 2 are incapable of binding
simultaneously to the enzyme as a result of steric
repulsion between the two distal phosphate groups. On
the basis of these studies, it was suggested that a
compound that simultaneously occupies both aryl phos-
phate binding sites may be a potent inhibitor of PTP1B.32
The possibility that some of the residues involved in the
second noncatalytic site may be less conserved among
PTPases than those found in the catalytic site suggests
that such compounds could also be selective for PTP1B.32
Ki’s as low as 79 nM.14 However, there is some question
as to whether these compounds are acting as reversible
inhibitors or as irreversible affinity labels. Peptides
bearing sulfotyrosine,15,16 O-malonyltyrosine,17 fluoro-
O-malonyltyrosine,18,19 and other pTyr mimetics20 are
also inhibitors of PTP1B though none are as effective
phosphate mimetics as the DFMP group. Because of the
problems associated with the development of peptide-
based therapeutics, more recent efforts have focused on
developing small-molecule PTP1B inhibitors. Burke and
co-workers21,22 and later ourselves23-25 reported that
even relatively simple aryl derivatives bearing a single
DFMP group were modest inhibitors of PTP1B. Rice et
al. have examined a series of triamides functionalized
with an isoxazole ring, a carboxylic acid group, and
various hydrophobic moieties.26 One of these compounds
was a noncompetitive inhibitor of PTP1B with a Ki of
850 nM. Modest selectivity was found between PTP1B
and dual specificity phosphatases Cdc25A, -B, and -C.
Its selectivity with other PTPases was not reported.
Very recently, researchers at Wyeth-Ayerst reported
that certain compounds from a series of benzofuran and
benzothiophene derivatives, as well as a series of
azolidinediones, were potent inhibitors of PTP1B, ex-
hibiting IC50 values in the 11-300 nM range.27-29
Selectivity for PTP1B over other PTPases varied any-
where from 5- to 175-fold. Also reported very recently
was the inhibition of PTP1B with a series of oxalylami-
nobenzoic acid derivatives.30,31 The most potent com-
pound from this series exhibited an IC50 of 5 µM with
PTP1B at physiological pH; however, it was highly
selective (>150-fold) for PTP1B compared to other
PTPases examined (SHP-1, PTPR D1, PTPꢀ D1, PTPâ,
CD45 D1D2, LAR D1D2).
Several years ago we reported that compounds such
as 4 and 5, bearing two DFMP groups, are in general
more effective inhibitors than compounds bearing a
single DFMP group.23-25 The effect of the additional
DFMP group on binding could be modest, such as with
compound 5 (IC50 ) 26 µM) versus compound 7 (IC50
)
95 µM), or it could be quite significant, such as with
compound 4, which exhibited a Ki of 1.5 µM with PTP1B
and is a 2-3 orders of magnitude more effective inhibi-
tor than compound 6. Later, Taing et al. reported a wide
range of non-peptidyl bis-DFMP compounds, several of
which exhibited Ki’s with PTP1B similar to that of
compound 4.34 Very recently, Desmarais and co-workers
reported that the tripeptide E-F2Pmp-F2Pmp is a
potent inhibitor of PTP1B, exhibiting an IC50 of 40 nM,
while the tripeptide F2Pmp-E-F2Pmp exhibits an IC50
of 400 nM.35 This indicates that the high inhibitory
potency of the E-F2Pmp-F2Pmp peptide is a result of
its sequence and not its overall charge.
The presence of a second aryl phosphate binding site
seems to have offered an explanation as to why com-
pounds bearing two DFMP groups are more effective
inhibitors than the analogous compounds bearing only
a single DFMP group. Indeed, it has been suggested by
Taing et al. and Desmarais et al. that the potencies of
their bis-DFMP compounds were a result of the inter-
action of the two phosphonate groups with the two
phosphate binding sites, although no structural evidence
was given to support this supposition. In the case of the
naphthyl inhibitor 5, the distance separating the two
DFMP groups is too short and the naphthyl ring too
rigid to allow both DFMP groups to interact with both
phosphate binding sites simultaneously, which suggests
that the second DFMP group on 5 enhances inhibitory
potency by interacting with residues that are not part
of the second phosphate binding site. However, modeling
studies suggested to us that the linker arm separating
In addition to the primary binding pocket or the active
site, a second noncatalytic aryl phosphate binding site
with lower binding affinity has been identified in
PTP1B.32 The second site was discovered by crystallizing
PTP1B in the presence of the high-affinity bis-phospho-
nate substrate 2 (Km ) 16 µM)33 or saturating (53 mM)
amounts of phosphotyrosine (3, pTyr). pTyr was found
to bind in two different modes. In one mode (pTyr A),
the phosphate group of pTyr binds in the catalytic site
in the usual manner. In the other mode (pTyr B), the

4586 J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 26
J ia et al.
the two DFMP groups in 4 was of sufficient length and
flexibility to allow them to interact with the two
phosphate binding sites simultaneously. Does the sec-
ond DFMP group on inhibitor 4 occupy the second
phosphate binding site? Why is compound 4 a better
inhibitor than 5? Is the second site an important motif
for the development of potent and selective PTP1B
inhibitors? In an attempt to answer these questions and
to learn more about those features that are important
for the rational design of small-molecule PTP1B inhibi-
tors, we determined the X-ray crystal structures of
PTP1B complexed with compounds 4 and 5. In addition,
the selectivity of compound 4, as well as several other
bis-DFMP-bearing inhibitors, for PTP1B versus seven
other PTPs is also reported.
Resu lts a n d Discu ssion
Cr ysta lliza tion a n d Cr ysta l P a ck in g. Extensive
soaking experiments using various conditions including
using various buffers, pH, inhibitor concentrations, and
soaking time failed to produce crystals with the bound
inhibitor, as evidently shown by the lack of difference
in the electron density. This, however, is not surprising
because previous experiments had shown that those
ligands larger than inorganic phosphate could not be
soaked in but rather they have to be cocrystallized.32,36
Modification of the crystallization conditions of the
native PTP1B led to the reproducible production square-
plate cocrystals, the largest being ∼0.15 mm × 0.15 mm
× 0.05 mm in size. The only significant changes of
crystallization conditions are pH (now 6.5 instead of 7.5)
and the increased concentration of PEG 8000. The
cocrystals were usually too small for diffraction experi-
ments, although occasionally a diffraction-size crystal
could be obtained by chance. The cocrystals also have
external morphology different from that of the native
PTP1B crystals. Although they both have the same
space group and are very similar in a and b dimensions,
there is an elongation (∼2-5 Å) in the c axis in both
complex crystals. Interestingly, c axis elongation has
similarly occurred in another PTP1B complex involving
a larger ligand such as phosphotyrosine peptide, where
the c axis increased ∼20 Å.36 However, the elongation
in the c axis does not simply create more space between
molecules along the c axis to accommodate substrate
or inhibitor; it changes the crystal packing completely,
hence making solution of the structure by molecular
replacement necessary.
Molecular replacement calculations gave a single
sharp and unambiguous solution in both cases. For
instance, in the case of the PTP1B-4 complex, the
correlation coefficient of the top peak is 64.6%, compared
to the second peak of 37.6%. At the active site, a clear
and unambiguous difference in electron density corre-
sponding to the inhibitor was evident (Figure 1). In
contrast, there was no observable difference density at
the second binding site. The lack of the difference
density remains unchanged after further refinement of
the complex structure with the inclusion of the inhibitor
and gradual inclusion of water molecules. The final
models exhibit excellent geometry as examined using
the program Procheck,37 and the final refinement
statistics are summarized in Table 1 for both complex
structures.
F igu r e 1. Electron density map of inhibitor 4 (contoured at
σ), together with the final models of the inhibitor and PTP1B.
Ta ble 1. Crystallographic Data and Refinement Statistics
compound 4
compound 5
space group
P3121
P3121
a (Å)
89.22
87.52
b (Å)
89.22
87.52
c (Å)
108.78
50.00-2.35
89,257
20,176
0.039
106.29
50.00-2.50
61,055
13,400
0.054
resolution ranges (Å)
no. total reflns
no. unique reflns
Rsym
completeness (%)
refinement
92.2%
80.0%
resolution rangesa (Å)
R (Rfree, 10% random)
no. protein atoms
no. inhibitor atoms
no. water atoms
rmsd of bond length (Å)
rmsd of bond angles (deg)
6.00-2.35
0.208 (0.250)
2,427
30
244
6.00-2.50
0.189 (0.254)
2,427
24
144
0.011
1.41
0.008
1.47
a
σ
All data used with no cut.
Bin d in g of In h ibitor s to P TP 1B. The complex
structure of inhibitor 4 with PTP1B clearly shows that
one of the DFMP groups (the proximal group) of the
inhibitor occupies the active site (Figures 1-3). The
binding mode is similar to those of all other complexes
where the inhibitor/substrate has a similar aryl phos-
phate/phosphonate group.22,38 As shown in Figure 4, the
three terminal oxygen atoms accept hydrogen bonds
from the main-chain amide groups of the PTP loop and
form electrostatic interactions with Arg-221. There is a
water molecule that is bound in a cavity that forms H
bonds with the Pro-R fluorine, the N-H of Phe-182, and

Tyrosine Phosphatase 1B
J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 26 4587
second site. Therefore, we conclude that this inhibitor
does not bind to the second aryl phosphate-binding site.
The hydrocarbon linker chain does not bend as is
required to direct the distal phenyl DFMP group toward
the second phosphate binding site. Instead, it extends
into the solvent and is involved in water-mediated
H-bonding interactions with Arg-47.
As fully expected, the proximal phosphate group of
compound 5 binds to PTP1B similarly to the binding in
PTP1B-4 and other reported phosphonate complex
structures.22,38 Because of the shorter length of inhibitor
5, it was not expected that it would be able to reach the
second site. Indeed, the second phosphate group, ex-
tending straight from the aromatic moiety, makes little
or no contact with the enzyme. Thus, the shorter
inhibitor 5 is not able to make the critical water-
mediated interaction between the second phosphate
group and Arg-47 as observed in the inhibitor 4 complex,
which is likely the main factor contributing to the
different affinity between these two compounds. Inter-
actions between Arg-47 and acidic residues in ligands
and inhibitors have been noted by other workers. Arg-
47 has been shown to provide key interactions with the
Glu residue of the D-A-D-E-pY-L-NH2 peptide.36 Burke
and co-workers have examined this interaction as a
basis of rational inhibitor design.39 These workers found
that certain dipeptidyl and non-peptidyl difluorometh-
ylphosphonic acids bearing carboxylic acid functional-
ities were better inhibitors of PTP1B than the parent
compounds 6 and 7. These results were consistent with
molecular dynamics studies that suggested that some
of this increase in affinity could be attributed to interac-
tions between Arg-47 and the carboxyl group.
F igu r e 2. Ribbon diagram of PTP1B-4 complex. Compound
4 is shown in ball and stick model. The extended conformation
of compound 4 (shown in green) is evident. Side chains of
contact residues of PTP1B are also shown.
one of the oxygens of the phosphonate group. The
conformational change of the WPD loop takes place to
close the active site. Phe-182, Tyr-46, and Val-49 may
form hydrophobic interactions with the proximal phenyl
ring and the proximal section of the hydrocarbon linker
arm. Because of the exposed position of the protein
region in the vicinity of the second aryl phosphate group,
the number of contacts between the enzyme and 4 is
limited and higher temperature factors of both the
inhibitor and enzyme are observed, as evident by the
weak and noncontinuous density in the solvent-exposed
region (Figure 1). However, even though with the weak
density, the extended conformation of the inhibitor,
which only binds to the primary or active site, is beyond
any doubt. Clearly, there is no difference density in the
Bis-DFMP In h ibitor s Bear in g Alter n ative Lin ker
Ar m s. Although our modeling studies suggested that a
linker arm consisting of four CH2 groups seems to be
sufficient for allowing both DFMP groups to interact
with the two phosphate binding sites, dual occupancy
may require a longer linker that has more optimal
F igu r e 3. Stereodiagram of a closeup view of the PTP1B-4 complex. The inhibitor and enzyme residues are represented in ball
and stick model. The inhibitor is dark green. Residues that make immediate contact with the inhibitor are yellow, and those
farther away are pink. Contact residues are also labeled.

4588 J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 26
J ia et al.
exception being meg-1 for which it exhibited a 35-fold
preference of PTP1B. With the exception of compound
9, all of the compounds exhibited good selectivity for
PTP1B when compared to Cdc25a and meg-1. For the
other four compounds, selectivity for PTP1B versus
PTPâ ranged from good (45- to 77-fold for compounds
4, 8, and 11) to modest (15-fold for compound 10).
Similar trends in PTP1B selectivity with these four
compounds were also found using CD45. The selectivity
of all of the compounds for PTP1B versus TCPTP, SHP-
1, and SHP-2 was modest to very poor. The only
exception to this was compound 8, which exhibited a
39-fold preference for PTP1B versus SHP-2. Compound
11 had previously been shown by Taing et al. to inhibit
PTP1B with a Ki of 2.7 µM, which is similar to the IC50
value (1.7 µM) that we have obtained for 11 with
PTP1B. Taing et al. also reported that compound 11
displayed little inhibitory potency with VHR or PTPR
and was 18-fold less effective with LAR compared to
PTP1B.34 Our studies with compound 11 show that its
selectivity is good compared to cdc25A and PTP-meg1,
modest with SHP-2, CD45, and PTPâ, but very poor
when compared to SHP-1 and TCPTP.
F igu r e 4. Schematic of inhibitor 4-PTP1B interactions. A
distance cutoff of 3.2 Å was used except for those involving
certain aromatic rings.
In h ibitor P oten cy a n d th e Secon d P h osp h a te
Bin d in g Site. As mentioned earlier, Taing et al.
recently reported PTP1B inhibition studies with a wide
variety of compounds bearing two DFMP groups.34 The
best inhibitor from this series, compound 21, exhibited
length and, more importantly, afford more variation in
torsion angle freedom. Because the two DFMP groups
in 4 clearly do not interact with both binding sites, we
prepared compound 8 (Scheme 1) in which the two aryl
DFMP groups were joined by a longer yet still flexible
linker, as well as compounds 9-11 (Scheme 2), in which
the linker arm joining the two DFMP groups was longer
yet was more rigid than the linker arms in inhibitors 4
and 8. However, these compounds exhibited IC50 values
(1.7-5.3 µM) with PTP1B that were similar to that
obtained with compound 4 (Table 2). Thus, altering the
length and flexibility of the linker arm made little
difference in inhibitory potency. Analysis of the data
obtained by Taing and co-workers on the inhibition of
PTP1B with a wide variety of compounds bearing two
DFMP groups34 reveals a similar phenomenon.
a Ki of 0.93 µM with PTP1B. Although it was suggested
that the two DFMP groups in 21 interacted with the
two phosphate binding sites in PTP1B, the possibility
of less specific interactions with residues outside the
second site was also raised. It is of significance that
compound 4 (Ki ) 1.5 µM), which exhibits an affinity
for PTP1B that is only slightly less than compound 21,
interacts with Arg-47 and not the second phosphate
binding site. It is also of significance that compounds
8-11 exhibit little difference in inhibitory potency with
Compounds 4 and 8-11 were examined for selectivity
by comparing the IC50 values obtained for PTP1B with
those obtained with seven other PTPases (Table 2).
Compound 9 demonstrated poor selectivity for PTP1B
versus all of the PTPases examined with the one
Sch em e 1

Tyrosine Phosphatase 1B
J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 26 4589
Sch em e 2
Ta ble 2. IC50 Values of Inhibitors 4 and 8-11
IC50 (µM)
compound
PTP1B
TCPTP
SHP-1
SHP-2
CD45
Cdc25a
PTPâ
PTPmeg-1
4
8
9
10
11
4.4
3.9
4.2
5.3
1.7
6.4
6.1
4.6
7.9
3.2
31
61
6.3
16
15
91
152
16
32
60
198
236
14
108
42
16% at 250 µM
342
300
15
77
77
9% at 250 µM
2% at 250 µM
146
30% at 380 µM
30% at 500 µM
13% at 250 µΜ
30
163
163
PTP1B compared to compound 4, yet the length and
flexibility of the units linking the two phenyl DFMP
groups in compounds 8-11 differ significantly from that
found in inhibitor 4. Indeed, the potency of compounds
8-11 with PTP1B differs little from compound 21,
which has an even longer spacer separating the two
phenyl DFMP groups. There are two possible explana-
tions for these results. One possibility is that, despite
the differences in the length and flexibility of the spacers
separating the two phenyl DFMP groups in both our
and Taing et al.’s studies, none of the compounds
reported here or by Taing et al. interact with both
phosphate binding sites simultaneously. The second
phenyl DFMP increases inhibitory potency by interact-
ing with other positive residues that are within the
vicinity of the active site, of which there are several
(Arg-43, Arg-45, Arg-47, Lys-36, Lys-116, Lys-120). The
other possibility is that some of the compounds reported
here and by Taing et al. do interact with the second site.
However, if this is the case, then bis-DFMP inhibitors
that interact with both phosphate binding sites may not
be significantly more potent than bis-DFMP inhibitors
that interact with both the catalytic site and other
residues that are not part of the second site, such as
Arg-47.
(D181A-PTP1B) and a >10-fold lower Km for PTP1B
than the analogous peptide bearing a single pTyr
residue (ETDYpYRKGGKGGL). Arg-47 forms a salt
bridge with the Asp residue in the peptide. One of the
pTyr residues binds in the catalytic site, while the other
occupies the second phenylphosphate binding site. The
phenyl ring of the substrate in the noncatalytic site is
poorly ordered and may adopt a variety of conforma-
tions. This observation, along with mutagenesis studies,
indicates that Arg-254 in the noncatalytic site is the
main contributor to the enhanced affinity.40 The dis-
tance between the two phosphate groups in the diphos-
phorylated IRK peptides reported by Salmeen et al. is
greater than in compound 4, which may explain why 4
does not exhibit dual binding. However, an array of non-
peptidyl compounds bearing two DFMP groups attached
by linker arms of varying length and/or flexibility,
reported here and by Taing et al.,34 did not differ
significantly from 4 in terms of inhibitory potency. We
have shown that a significant increase in inhibitor
potency, by over 2 orders of magnitude, can result from
interaction of the second phosphate group by a water-
mediated H bond with Arg-47. This is greater than the
difference in affinity between the mono- and diphos-
phorylated IRK peptides.40 This suggests that interac-
tion of a phosphate group with Arg-254 may not be any
more advantageous, in terms of inhibitor potency, than
interaction with other positively charged residues in the
vicinity of active site, such as Arg-47.
Although dual binding has yet to be established with
PTP1B inhibitors, it has very recently been unequivo-
cally demonstrated with pTyr-bearing substrates.40
Salmeen et al. reported the crystal structure of C215A
PTP1B complexed with the insulin receptor kinase
(IRK) peptide, ETDpYpYRKGGKGGL, which bears two
tandem pTyr residues.40 This peptide exhibits a 70-fold
greater affinity for a PTP1B substrate-trapping mutant
Selectivity a n d th e Secon d P h osp h a te Bin d in g
Site. It is significant to note that all of the compounds
reported in this study exhibited very poor selectivity for
PTP1B versus TCPTP. TCPTP is abundantly and widely

4590 J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 26
J ia et al.
expressed.41 TCPTP is capable of dephosphorylating an
18-amino acid phosphopeptide bearing three phospho-
tyrosines that correspond to the autophosphorylated
sequence of the insulin receptor kinase (IRK), and an
Asp-Ala mutant of TCPTP was capable of trapping the
entire intracellular IRK.42 Unlike the PTP1B knock-out
mice, TCPTP null mice die soon after birth.43 This
underscores the importance of obtaining inhibitors that
are specific for PTP1B as far as their therapeutic
potential is concerned. TCPTP and PTP1B exhibit a
high degree of sequence homology.44 Like PTP1B,
TCPTP has an Arg at position 47 as well as all of the
key residues in the second phosphate binding site such
as Arg-24, Arg-254, Gly-259, and Met-258.44 Thus, it is
unlikely that Arg-47 or the second phosphate binding
site will be useful motifs for obtaining inhibitors that
are capable of differentiating between PTP1B and
TCPTP. Nevertheless, it is possible that these motifs
may be useful for obtaining inhibitors that are selective
for PTP1B versus PTPs other than TCPTP. SHP-1 and
SHP-2 have a Lys at the position corresponding to that
of Arg-47 in PTP1B.44,45 Several, but not all, of the key
residues that constitute the second phosphate binding
site, such as Arg-254, Gly-259, and Met-258, are present
in SHP-1 and SHP-2.44,45 These residues are less highly
conserved in the receptor-like PTPs such as PTPâ and
CD45.44 The catalytic domain of cdc25a exhibits little
structural homology to PTP1B.46 Compound 21 reported
by Taing et al. exhibited very good selectivity (100- to
3600-fold) for PTP1B versus the dual specificity PTPase
VHR and the receptor-like PTPases LAR and PTPR.
However, its selectivity with TCPTP was not reported.
The tripeptide E-F2Pmp-F2Pmp, which is a potent
inhibitor of PTP1B (Ki ) 40 nM), was found to be at
least 100-fold more selective for PTP1B versus CD45,
LAR, PTPâ, and SHP-1.35 Its selectivity with TCPTP
was not reported.
tilled before use. Solvents were purchased from Caledon
Laboratories (Georgetown, Ontario, Canada), Lancaster Syn-
thesis Incorporated, and BDH Canada (Toronto, Canada).
Tetrahydrofuran (THF) was distilled from sodium metal in the
presence of benzophenone under argon. CH2Cl2 was distilled
from calcium hydride under argon. Dimethylformamide (DMF)
was distilled under reduced pressure from calcium hydride and
stored over 4 Å sieves under argon. Reactions involving
moisture-sensitive reagents were executed under an inert
atmosphere of dry argon or nitrogen. All glassware was
predried prior to use, and all liquid transfers were performed
using dry syringes and needles. Silica gel chromatography was
1
performed using silica gel 60A (Silicycle, 230-400 mesh). H,
19F, 31P, and 13C NMR spectra were recorded on a Varian 200-
Gemini, Bruker AC-200, and Bruker AC-300 NMR spectrom-
eters. The abbreviations s, d, t, q, m, dd, dt, and br are used
for singlet, doublet, triplet, quartet, multiplet, doublet of
doublets, doublet of triplets, and broad, respectively. Chemicals
shifts (δ) for 1H NMR spectra run in CDCl3 are reported in
ppm relative to the internal standard tetramethylsilane (TMS).
Chemical shifts (δ) for 1H NMR spectra run in CD3OD are
reported in ppm relative to residual solvent protons (δ 3.30).
Chemical shifts (δ) for 1H NMR spectra run in D2O are
reported in ppm relative to residual solvent protons (δ 4.79).
For 13C NMR spectra run in CDCl3, chemical shifts are
reported in ppm relative to the CDCl3 residual carbons (δ 77.0
for central peak). For 13C NMR spectra run in CD3OD,
chemical shifts are reported in ppm relative to the CD3OD
residual carbons (δ 49.0 for central peak). For 13C NMR spectra
run in D2O, chemical shifts are reported in ppm relative to
CH3OH in D2O (external). For 31P NMR spectra, chemical
shifts are reported in ppm relative to 85% phosphoric acid
(external). For 19F NMR spectra, chemical shifts are reported
in ppm relative to trifluoroacetic acid (external). Low-resolu-
tion and high-resolution electron impact mass spectra (EIMS)
were obtained on a Micromass 70-S-250 mass spectrometer.
Low-resolution electrospray ionization mass spectra (LRES-
IMS) were obtained on a Micromass Quatro II mass spectrom-
eter. High-resolution electrospray ionization mass spectra
(HRESMS) were obtained on a Micromass ZabSpec oaTOF
spectrometer. All melting points were taken on a Mel-Temp
melting point apparatus and are uncorrected. Analytical HPLC
was performed on a Waters 600E LC system using a Vydac
218TP54 analytical C-18 reverse-phase column and a Waters
2487 absorbance detector set at 254 nm. Enzyme assay
solutions were prepared with deionized/distilled water. Com-
pounds 4 and 5 were prepared as previously described.49,50
T-cell PTP was obtained from Calbiochem (human, recombi-
nant E. coli, catalog no. 539732, lot no. B21267).
Con clu sion s
Dual binding of PTP1B inhibitors has yet to be
verified. Nevertheless, our results suggest that interac-
tion of bis-DFMP-bearing inhibitors with Arg-47 can
result in a significant increase in inhibitor potency and
that interaction with this residue may be just as
effective a means of increasing inhibitor potency as
interaction with the second phosphate binding site. Our
results from selectivity experiments clearly demonstrate
a further challenge in developing highly specific inhibi-
tors toward PTP1B as reflected by the virtual nonexist-
ence of selectivity toward PTP1B and TCPTP. Given the
extremely high sequence identity between these two
enzymes, it is not surprising that poor selectivity was
observed. It is unlikely that inhibitors that target either
the second phosphate binding site or Arg47 will be
selective for PTP1B vs TCPTP. Finally, the limited
membrane permeability of DFMP-bearing compounds22
must also be addressed. The use of less highly charged
phosphate mimetics is currently being examined as a
means of avoiding this problem.47,48
1-Iod o-4-(2ben zyloxyeth oxym eth yl)ben zen e (13). To a
solution of p-iodobenzyl bromide (12, 2.1 g, 7.07 mmol, 3 equiv)
in anhydrous CH2Cl2 (5 mL) was added ethylene glycol (0.13
mL, 2.33 mmol, 1 equiv) and silver(I) oxide (1.64 g, 7.08 mmol,
3 equiv). The reaction mixture was stirred for 48 h at room
temperature and filtered, and the filtrate was concentrated
by rotary evaporation. Pure 13 was obtained as a white solid
(0.60 g, 52%) after purification via flash chromatography
(EtOAc/hexane, 20:80, Rf ) 0.3) and recrystallization in CH2-
Cl2/hexane: mp 75-77 °C; 1H NMR (200 MHz, CDCl3) δ 7.67
(d, J ) 8.7 Hz, 4H, aryl), 7.09 (d, J ) 7.3 Hz, 4H, aryl), 4.51
(s, 4H, 2(CH2Ar)), 3.64 (s, 4H, OCH2CH2O); 13C NMR (50 MHz,
CDCl3) δ 138.2, 137.5, 129.5, 92.9, 72.7, 69.9; LREIMS m/z
494 (M+, trace), 277 (loss of CH2ArI, 100%); HRMS(EI) calcd
for C16H16O2I2 493.9239, found 493.9229.
{[4-(2-{4-[(Dieth oxyp h osp h or yl)d iflu or om eth yl]ben z-
yloxy}eth oxym eth yl)p h en yl]d iflu or om eth yl}ph osp h on ic
Acid Dieth yl Ester (14). Diethylbromodifluoromethane phos-
phonate (2.7 mL, 15.2 mmol, 9.5 equiv) was added to a stirring
suspension of cadmium powder (1.8 g, 16.0 mmol, 10 equiv)
in anhydrous DMF (10 mL), and the mixture was stirred for
4 h in a round-bottom flask fitted with a condenser (exothermic
reaction). The mixture was then filtered quickly under an inert
atmosphere through oven-dried Celite into a round-bottom
flask containing 13 (0.8 g, 1.6 mmol, 1.0 equiv) and copper(I)
Exp er im en ta l Section
Gen er a l. All starting materials for syntheses were obtained
from commercial suppliers (Aldrich Chemical Co., Oakville,
Ontario, Canada, or Lancaster Synthesis Incorporated, Wind-
ham, NH). Trimethylsilyl bromide (TMSBr) was freshly dis-

Tyrosine Phosphatase 1B
J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 26 4591
chloride (0.8 g, 8.1 mmol, 5.1 equiv), and the mixture was
stirred for 12 h. The mixture was filtered, and the filtrate was
transferred to a separatory funnel, to which was added CH2-
Cl2 (30 mL). The organic layer was washed with water (2 ×
25 mL), 5% NaHCO3 (1 × 25 mL), and brine (1 × 25 mL), dried
(MgSO4), filtered, and concentrated by rotary evaporation. 14
was purified via silica gel chromatography (EtOAc/hexane, 55:
45, Rf ) 0.4) and was obtained in pure form as a colorless oil
(0.78 g, 79%). 1H NMR (200 MHz, CDCl3) δ 7.59 (d, J ) 8.0
Hz, 4H, aryl), 7.43 (d, J ) 8.1 Hz, 4H, aryl), 4.62 (s, 4H, 2(CH2-
Ar)), 4.17 (m, 8H, 4(OCH2CH3)), 3.69 (s, 4H, OCH2CH2O)), 1.30
(t, J ) 7.0 Hz, 12H, 4(OCH2CH3)); 13C NMR (50 MHz, CDCl3)
4.35 (t, J PF ) 117.5 Hz); LREIMS m/z 278 (M+, 16%), 141 (M+
- P(O)(OEt)2, 100%); HRMS(EI) calcd for C12H17O3F2P 278.0883,
found 278.0889.
Syn th esis of 18-20. Gen er a l P r oced u r e. A solution of
15 (1.0 equiv) and N-bromosuccinimide (1.1 equiv) in benzene
(60 mL per gram of 15) was irradiated using a reflector IR
heat lamp (250 W, 120 v) for 1.5 h. The solution was washed
with water, 5% NaHCO3, and brine, dried (MgSO4), and
concentrated by rotary evaporation. The desired product was
partially purified by silica gel chromatography (EtOAc/toluene,
30:70, Rf ) 0.4) to yield a clear, colorless oil. 1H NMR analysis
of the crude material indicated a mixture of the desired product
16, the undesired dibrominated material 17, and unreacted
15 in a 68:6.5:25.5 ratio. This mixture was used without any
further purification for the synthesis of 18-20. To a solution
of the diol (hydroquinone, resorcinol, or catechol, 0.3 equiv)
and K2CO3 (1.0 equiv) in anhydrous DMF (12 mL/mmol diol)
was added a solution of crude 16 (approximately 1.0 equiv of
16) and anhydrous DMF (5 mL of DMF/g of crude material).
After 12 h, EtOAc (90 mL/mmol of diol) was added and the
organic layer was washed twice with water and once with brine
(90 mL/mmol of diol of each aqueous solution). The organic
layer was dried (MgSO4), filtered, and concentrated by rotary
evaporation. Pure 18-20 was obtained by silica gel chroma-
tography as a white crystalline solid or as a colorless oil. Yields
are based on the diol as the limiting reagent.
δ 141.4, 132.0, 127.3, 126.4, 118.1 (td, J CF ) 263.5 Hz, J CP
)
217.8 Hz), 72.7, 70.0, 64.6 (d, J CP ) 6.4 Hz), 16.2 (d, J CP ) 5.5
Hz); 19F NMR (188 MHz, CDCl3) δ -32.25 (d, J FP ) 117.5 Hz);
31P NMR (81 MHz, CDCl3) δ 4.13 (t, J PF ) 117.5 Hz); LREIMS
m/z 616 (M+ + H+, trace), 278 (CH2ArCF2P(O)(OEt)2 + H+,
77%), 140 (100%).
{[4-(2-{4-[(Dieth oxyp h osp h or yl)d iflu or om eth yl]ben z-
yloxy}eth oxym eth yl)p h en yl]d iflu or om eth yl}p h osp h on ic
Acid (8). To a solution of ester 14 (0.100 g, 0.16 mmol, 1 equiv)
in dry CH2Cl2 (4 mL) was added TMSBr (0.17 mL, 1.3 mmol,
8 equiv), and the solution was refluxed for 2 days under an
argon atmosphere. The solvent was removed by rotary evapo-
ration, the resulting oil was dissolved in dry CH2Cl2, and the
solution was concentrated again by rotary evaporation. This
process was repeated two more times. The resulting colorless
oil was then placed under high vacuum for 16 h to remove
any traces of unreacted TMSBr. The oil was redissolved in
benzene (3 mL), and an aqueous solution (3 mL) of NH4HCO3
(0.077 g, 0.97 mmol, 6 equiv) was added. After the mixture
was vigorously stirred for 30 min, the organic layer was
evaporated and the aqueous solution was lyophilized repeat-
edly to yield 8 as a pure, white solid in near-quantitative yield
[(4-{4-[4-(Diflu or op h osp h on om eth yl)ben zyloxy]p h en -
oxym eth yl}p h en yl)d iflu or om eth yl]p h osp h on ic a cid d i-
eth yl ester (18): white solid; yield, 44%; mp 91-93 °C; TLC
Rf ) 0.3 (EtOAc/CH2Cl2, 12:88); 1H NMR (200 MHz, CDCl3) δ
7.64 (d, J ) 7.6 Hz, 4H, aryl), 7.51 (d, J ) 8.4 Hz, 4H, aryl),
6.90 (s, 4H, aryl), 5.07 (s, 4H, 2(CH2Ar)), 4.20 (m, 8H, 4(OCH2-
CH3)), 1.32 (t, J ) 7.0 Hz, 12H, 4(OCH2CH3)); 13C NMR (50
MHz, CDCl3) δ 153.3, 140.3, 132.3 (m) 127.1, 126.5, 118.0
1
(weak td), 116.1, 70.3, 64.7 (d, J CP ) 6.4 Hz), 16.3 (d, J CP
)
(0.093 g). H NMR (200 MHz, D2O) δ 7.63 (d, J ) 7.7 Hz, 4H,
4.6 Hz); 19F NMR (188 MHz, CDCl3) δ -32.30 (d, J FP ) 116.0
Hz); 31P NMR (81 MHz, CDCl3) δ 4.07 (t, J PF ) 116.0 Hz);
LREIMS m/z 662 (M+, 40%), 277 (CH2ArCF2P(O)(OEt)2, 100%);
HREIMS calcd for C30H36O8F4P2 662.1822, found 662.1788.
aryl), 7.49 (d, J ) 8.0 Hz, 4H, aryl), 4.65 (s, 4H, 2(CH2Ar)),
3.74 (s, 4H, OCH2CH2O); 13C NMR (50 MHz, D2O) δ 139.8,
133.7 (m), 127.9, 126.0 (t, J ) 6.6 Hz), 119.8 (weak td), 72.0,
68.9; 19F NMR (188 MHz, D2O) δ -28.36 (d, J FP ) 96.1 Hz);
31P NMR (202.5 MHz, D2O) δ 4.65 (t, J PF ) 96.4 Hz); LRESIMS
m/z 501.1 (monoanion, loss of 4 × NH4+, gain of 3H+), 250.1
(dianion, loss of 4 × NH4+, gain of 2H+); HRESMS calcd for
C18H19F4O8P2 501.0485, found 501.0485. Analytical HPLC
analysis of the ammonium salt was performed using acetoni-
trile (solvent A) and 0.1% TFA in water (solvent B) as eluants
under the following conditions: linear gradient of 5% A/95%
B to 100% A (5 min), 100% A (10 min), linear gradient of 100%
A to 5% A/95% B (5 min), 100% B (10 min). Only a single peak
was evident on the HPLC chromatogram. Retention time was
7.6 min.
[(4-{3-[4-(Diflu or op h osp h on om eth yl)ben zyloxy]p h en -
oxym eth yl}p h en yl)d iflu or om eth yl]p h osp h on ic a cid d i-
eth yl ester (19): colorless oil; yield, 76%; TLC Rf ) 0.5
1
(EtOAc/hexane, 70:30); H NMR (200 MHz, CDCl3) δ 7.64 (d,
J ) 7.3 Hz, 4H, aryl), 7.51 (d, J ) 7.3 Hz, 4H, aryl), 7.18 (t, J
) 8.8 Hz, 1H, aryl), 6.58 (m, 3H, aryl), 5.08 (s, 4H, 2(CH2Ar)),
4.20 (m, 8H, 4(OCH2CH3)), 1.30 (t, J ) 7.3 Hz, 12H,
4(OCH2CH3)); 13C NMR (50 MHz, CDCl3) δ 160.0, 140.1, 133.1
(m), 130.0, 127.1, 126.6 (br t), 118.1 (td, J CF ) 263.5 Hz, J CP
)
218.0 Hz), 107.9, 102.8, 69.6, 64.6 (d, J CP ) 6.4 Hz), 16.2 (d,
J CP ) 5.5 Hz); 19F NMR (188 MHz, CDCl3) δ -32.30 (d, J FP
)
116.7 Hz); 31P NMR (81 MHz, CDCl3) δ 4.13 (t, J PF ) 117.5
Hz); LREIMS m/z 662 (M+, trace), 642 (loss of HF, 100%), 277
(CH2ArCF2P(O)(OEt)2, 50%); HREIMS calcd for C30H36O8F4P2
662.1822, found 662.1806.
(Diflu or o-p-tolylm eth yl)p h osp h on ic Acid Dieth yl Es-
ter (15). Diethylbromodifluoromethane phosphonate (14.5 mL,
82.0 mmol, 4.6 equiv) was added to a stirring suspension of
cadmium (10.0 g, 89.0 mmol, 5.0 equiv) in anhydrous DMF
(50 mL), and the mixture was stirred for 4 h in a round-bottom
[(4-{2-[4-(Diflu or op h osp h on om eth yl)ben zyloxy]p h en -
oxym eth yl}p h en yl)d iflu or om eth yl]p h osp h on ic a cid d i-
eth yl ester (20): colorless oil; yield, 69%; TLC Rf ) 0.3
(EtOAc/CH2Cl2, 12:88); 1H NMR (200 MHz, CDCl3) δ 7.59 (m,
8H, aryl), 6.90 (m, 4H, aryl), 5.21 (s, 4H, 2(CH2Ar)), 4.18 (m,
flask fitted with
a condenser (exothermic reaction). The
mixture was then filtered quickly under an inert atmosphere
through oven-dried Celite into a round-bottom flask containing
p-iodotoluene (3.90 g, 18.0 mmol, 1.0 equiv) and copper(I)
chloride (4.39 g, 44.0 mmol, 2.5 equiv), and the mixture was
stirred for 12 h at room temperature. The mixture was filtered,
and the filtrate was transferred to a separatory funnel, to
which was added CH2Cl2 (100 mL). The solution was washed
with water (2 × 50 mL), 5% NaHCO3 (1 × 50 mL), and brine
(1 × 50 mL), dried (MgSO4), filtered, and concentrated by
rotary evaporation. Pure 15 was obtained by silica gel chro-
matography (EtOAc/hexane, 30:70, Rf ) 0.5) as a colorless oil
(4.05 g, 82%). 1H NMR (200 MHz, CDCl3) δ 7.51 (d, J ) 7.3
Hz, 2H, aryl), 7.25 (d, J ) 8.8 Hz, 2H, aryl), 4.20 (m, 4H,
2(OCH2CH3)), 2.40 (s, 3H, CH3), 1.31 (t, J ) 7.3 Hz, 6H,
2(OCH2CH3)); 13C NMR (50 MHz, CDCl3) δ 140.9, 129.6 (m),
129.0, 126.1 (br t), 118.2 (weak td, CF2), 64.5 (d, J CP ) 6.4
Hz), 21.1, 16.1 (d, J CP ) 3.7 Hz); 19F NMR (188 MHz, CDCl3)
δ -32.09 (d, J FP ) 119.0 Hz); 31P NMR (81 MHz, CDCl3) δ
8H, 4(OCH2CH3)), 1.31 (t, J ) 7.1 Hz, 12H, 4(OCH2CH3)); 13
C
NMR (50 MHz, CDCl3) δ 149.1, 140.4, 132.4 (m), 127.1, 126.5
(t, J ) 6.0 Hz), 122.1, 118.0 (weak td), 115.9, 71.0, 64.6 (d, J CP
) 6.4 Hz), 16.2 (d, J CP ) 4.6 Hz); 19F NMR (188 MHz, CDCl3)
δ -32.31 (d, J FP ) 116.0 Hz); 31P NMR (81 MHz, CDCl3) δ
4.05 (t, J PF ) 116.0 Hz); LREIMS m/z 662 (M+, 7%), 642 (loss
of HF, 50%), 277 (CH2ArCF2P(O)(OEt)2, 100%); HREIMS calcd
for C30H36O8F4P2 662.1822, found 662.1799.
Syn th esis of 9-11. Gen er a l P r oced u r e. To a solution of
the ethyl ester 18, 19, or 20 (1.0 equiv) in dry CH2Cl2 (9 mL/
0.1 mmol ester) was added TMSBr (8 equiv), and the solution
was refluxed for 2 days under an argon atmosphere. The
solvent was removed by rotary evaporation, the resulting oil
was dissolved in dry methylene chloride, and the solution was
concentrated again by rotary evaporation. This process was

4592 J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 26
J ia et al.
repeated two more times. The resulting colorless oil was then
placed under high vacuum for 16 h. The oil was redissolved
in benzene (3 mL), and water (3 mL) was added. After the
mixture was vigorously stirred for 30 min, the organic layer
was removed by rotary evaporation and the aqueous solution
was lyophilized to yield pure 9-11 as white solids in near-
quantitative yield. The acids were converted to their am-
monium salts by addition of an aqueous solution of 2.5 equiv
of ammonium bicarbonate followed by repeated lyophilization
until a constant weight was obtained. The compounds were
analyzed by analytical HPLC as described above for compound
8. All compounds showed only a single peak in the HPLC
chromatogram.
centrifuged at 10 000g in a SS34 rotor for 15 min. The
supernatant was then centrifuged at 43 000g in a 50.2 Ti rotor
in a Beckman ultracentrifuge for 50 min. The supernatant was
applied to a 12 mL M2 flag (Sigma) monoclonal affinity column
that was prepared as follows. The resin was washed with three
column volumes of low-salt buffer (20mM Tris-HCl, pH 8, 150
mM NaCl, 1 mM DTT, and 2% Tween-20), then was activated
with two column volumes of 0.1 M glycine pH 3, and finally
was washed with three column volumes of low-salt buffer. The
flow through was collected, and Tween-20 was added to a final
concentration of 1% (v/v). This was reapplied to the same anti-
flag affinity resin, and the mixture was washed with 10 column
volumes of high-salt buffer (20 mM Tris-HCl, pH 8, 500 mM
NaCl, 1 mM DTT, and 2% Tween-20) followed by 5 column
volumes of low-salt buffer. The flag-CDC25a was eluted with
4 column volumes of low-salt buffer containing 0.17 mg/mL
flag peptide. DTT to a final concentration of 5 mM was added
to each fraction. Approximately 12 mg of flag-CDC25a (>80%
pure as estimated by SDS-PAGE) were obtained.
In h ibition Stu d ies. PTPase assays were performed using
3,6-fluorescein diphosphate (FDP) as substrate as previously
described.35 IC50 values were obtained using 10 different
concentrations of inhibitor and at FDP concentrations ap-
proximately equal to the KM concentration for each enzyme.
Data were fit to the four-parameter equation16 with errors less
than 10% for all IC50 values obtained.
Cr ysta lliza tion of P TP 1B-4 a n d P TP 1B-5 Com p lexes.
All crystallization experiments were carried out at 4 °C using
the hanging-drop vapor diffusion method. Native PTP1B
crystals were grown according to the published conditions54
with one modification, namely, using 18-20% instead of 12-
16% PEG 8000. In soaking experiments, PTP1B native crystals
were placed in the inhibitor solution over a range of concentra-
tions (from 5 to 20 mM) and time periods (from 2 to 36 h). For
cocrystallization, the protein stock solution consists of 10 mg/
mL PTP1B in a mixture of 10 mM Tris-HCl, 25 mM NaCl, 0.2
mM EDTA, 3.0 mM DTT, and 0.1 mM inhibitor 4 or 5. For
the PTP1B-4 complex, a 2 µL drop of this stock solution was
mixed with an equal volume of precipitating solution consisting
of 19% PEG 8000, 0.1 M Hepes (pH 7.5), and 0.2 M magnesium
acetate and was equilibrated against 1 mL of the precipitating
solution. The conditions for PTP1B-5 are identical except the
buffer pH was changed to 6.7.
Diffr a ction Da ta Collection . Diffraction data were col-
lected at the F1 station of Cornell High Energy Synchrotron
Source (CHESS). For cryoprotection, crystals were serially
transferred to crystallization solutions supplemented with 5%,
10%, 20%, and 25% glycerol for 2 min at each concentration
followed by immediate flash cooling in liquid propane. Data
collection was carried out in a stream of nitrogen gas at 100
K. The PTP1B-4 and PTP1B-5 crystals diffracted to 2.35 and
2.50 Å, respectively. Data were processed using Denzo/
Scalepack,55 and the data statistics are summarized in Table
1.
Str u ctu r e Deter m in a tion a n d Refin em en t. Molecular
replacement calculations were carried out using the program
EPMR,56 with the probing model being the native PTP1B
peptide complex structure devoid of the peptide.36 After the
first round of refinement using CNS,57 clear and unambiguous
difference electron density was evident in the active-site pocket
in both PTP1B-4 and PTP1B-5. A model of the inhibitor was
then built and fitted in the difference density. The structure
was refined, and water molecules were gradually added.
Diagrams were prepared using Xtalview,58 Molscript,59 and
Raster3D.60 Coordinates of PTP1B-4 and PTP1B-5 complex
structures have been deposited in the Protein Data Bank
(accession codes: 1KAK and 1KAV).
[(4-{4-[4-(Diflu or op h osp h on om eth yl)ben zyloxy]p h en -
oxym et h yl}p h en yl)d iflu or om et h yl]p h osp h on ic
a cid
(9): 1H NMR (200 MHz, CD3OD) δ 7.63 (d, J ) 8.0 Hz, 4H,
aryl), 7.53 (d, J ) 8.0 Hz, 4H, aryl), 6.93 (s, 4H, aryl), 5.10 (s,
4H, CH2Ar); 13C NMR (125 MHz, CD3OD) δ 154.4, 141.5, 134.7
(m), 128.1, 127.5 (t, J ) 6.8 Hz), 116.9, 70.9; 19F NMR (188
MHz, D2O) δ -29.82 (d, J FP ) 105.2 Hz); 31P NMR (81 MHz,
D2O) δ 4.47 (t, J PF ) 105.2 Hz); LRESMS m/z 549 (monoanion,
loss of 1H+), 274 (dianion, loss of 2H+); HRESMS calcd for
C22H19F4O8P2 549.0485, found 549.0480. HPLC retention time
was 8.14 min.
[(4-{3-[4-(Diflu or op h osp h on om eth yl)ben zyloxy]p h en -
oxym eth yl}-h en yl)d iflu or om eth yl]p h osp h on ic a cid (10):
1H NMR (200 MHz, D2O) δ 7.52 (d, J ) 7.3 Hz, 4H, aryl), 7.37
(d, J ) 8.8 Hz, 4H, aryl), 7.06 (t, J ) 8.0 Hz, 1H, aryl), 6.50
(m, 3H, aryl), 4.92 (s, 4H, 2(CH2Ar)); 13C NMR (75 MHz, D2O)
δ 159.2, 137.9, 136.1 (m), 130.7, 127.8, 126.5 (t, J ) 6.8 Hz),
108.6, 102.8, 70.1; 19F NMR (188 MHz, D2O) δ -30.00 (d, J FP
) 105.3 Hz); 31P NMR (81 MHz, D2O) δ 4.56 (t, J PF ) 106.0
Hz); LRESMS m/z 549 (monoanion, loss of 1H+), 274 (dianion,
loss of 2H+); HRESMS calcd for C22H19F4O8P2 549.0485, found
549.0491. HPLC retention time ) 8.02 min.
[(4-{2-[4-(Diflu or op h osp h on om eth yl)ben zyloxy]p h en -
oxym eth yl}ph en yl)diflu or om eth yl]ph osph on ic Acid (11):
1H NMR (200 MHz, D2O) δ 7.47 (d, J ) 8.8 Hz, 4H, aryl), 7.29
(d, J ) 7.4 Hz, 4H, aryl), 6.70 (bdt, 4H, aryl), 4.87 (s, 4H, 2(CH2-
Ar)); 13C NMR (75 MHz, D2O) δ 147.7, 139.0, 134.0 (m), 127.7,
126.2 (t, J ) 6.8 Hz), 122.3, 119.9 (td, J CF ) 261.0 Hz, J CP
)
201.0 Hz), 115.4, 70.6; 19F NMR (188 MHz, D2O) δ -29.98 (d,
J FP ) 106.9 Hz); 31P NMR (81 MHz, D2O) δ 4.64 (t, J PF ) 106.1
Hz); LRESMS m/z 549 (monoanion, loss of 1H+), 274 (dianion,
loss of 2H+); HRESMS calcd for C22H19F4O8P2 549.0485, found
549.0485. HPLC retention time was 7.91 min.
P u r ifica tion of P TP a ses. The flag-PTP1B (1-320) and
flag-CD45 (intracellular domain) were purified as described
by Huyer et al.51 and Wang et al.52 The GST-PTPâ, GST-SHP1,
and GST-SHP2 were purified as previously reported.35 CDC25a
cDNA was obtained from osteoblast total RNA using the
GeneAmp RNA PCR kit (Perkin-Elmer Cetus, CT). BamH1
and EcoR1 were added at the 5′ and 3′ end, respectively, during
amplification. The recombinant baculovirus was prepared
using the Bac-to-Bac baculovirus expression system (Gibco
BRL, MD). The CDC25a DNA fragment was ligated in a
pFastBac1 plasmid that was engineered to include a FLAG
sequence at the 5′ end of the cDNA. Recombinant bacmid was
obtained by transformation into competent cells DH10BAC
Escherichia coli cells. The resultant was transfected into
Spodoptera frugiperda Sf9 cells, and the virus found in the
supernatant medium was amplified two times up to total viral
stock volume of 500 mL. Production of the recombinant
CDC25a was done essentially as described by Summers and
Smith.53 Sf9 cells were infected in Graces supplemented
medium (TNM-FH) at a density of (1-3) × 106 cells/mL with
the appropriate amount of viral stock (moi 5). The cells were
cultured in suspension at 27 °C and were harvested at 72 h
postinfection. The cell pellet was washed twice in ice-cold PBS
and frozen at -80 °C until purification. All subsequent steps
were carried out at 4 °C. The cell pellet (40 g) was lysed in 40
mL of lysis buffer (40 mM Tris-HCl, pH 7.5, 0.2% Triton X-100,
4 mM DTT, 4 mM EDTA, and one complete protease inhibitor
tablet) and sonicated four times for 30 s on ice. The lysate was
Ack n ow led gm en t. This work was supported in part
by the Natural Sciences and Engineering Research
Council of Canada (NSERC) in the form of operating
grants to Z.J . and S.D.T. and a postgraduate scholarship
to A.N.D. Z.J . holds a Canada Research Chair in Struc-
tural Biology. The assistance from CHESS staff during

Tyrosine Phosphatase 1B
J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 26 4593
(22) Burke, T. R., J r.; Ye, B.; Yan, X.; Wang, S.; J ia, Z.; Chen, L.;
Zhang, Z.-Y.; Barford, D. Small molecule interactions with
protein-tyrosine phosphatase PTP1B and their use in inhibitor
design. Biochemistry 1996, 35, 15989-15986.
data collection is greatly appreciated. Brent Wathen
helped with the diagrams.
(23) Taylor, S. D.; Kotoris, C. C.; Dinaut, A. N.; Wang, Q.; Ram-
achandran, C.; Huang, Z. Potent non-peptidyl inhibitors of
protein tyrosine phosphatase 1B. Bioorg. Med. Chem. 1998, 6,
1457-1468.
(24) Taylor, S. D.; Kotoris, C. C.; Dinaut, A. N.; Wang, Q.; Ram-
achandran, C.; Huang, Z. Potent non-peptidyl inhibitors of
protein tyrosine phosphatase 1B (correction). Bioorg. Med. Chem.
1998, 6, 2235.
(25) Wang, Q.; Huang, Z.; Ramachandran, C.; Dinaut, A. N.; Taylor,
S. D. Naphthalenebis[R,R-difluoromethylenephosphonates] as
potent inhibitors of protein tyrosine phosphatases. Bioorg. Med.
Chem. Lett. 1998, 8, 345-350.
Refer en ces
(1) Neel, B. G.; Tonks, N. K. Protein tyrosine phosphatases in signal
transduction. Curr. Opin. Cell Biol. 1997, 9, 193-204.
(2) Denu, J . M.; Stuckey, J . A.; Saper, M. A.; Dixon, J . E. Form and
function in protein dephosphorylation. Cell 1996, 87, 361-364.
(3) Li, L.; Dixon, J . E. Form, function, and regulation of protein
tyrosine phosphatases and their involvement in human diseases.
Semin. Immunol. 2000, 12, 75-84.
(4) J ia, Z. Protein phosphatases: structures and implications.
Biochem. Cell Biol. 1997, 75, 17-26.
(5) Zhang, Z. Y. Protein-tyrosine phosphatases: biological function,
structural characteristics, and mechanism of catalysis. Crit. Rev.
Biochem. Mol. Biol. 1998, 33, 1-5.
(6) Kennedy, B. P.; Ramachandran, C. Protein tyrosine phos-
phatase-1B in diabetes. Biochem. Pharmacol. 2000, 60, 877-
883.
(7) 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. Increased insulin sensitivity and obesity
resistance in mice lacking the protein tyrosine phosphatase-1B
gene. Science 1999, 283, 1544-1548.
(8) Weiner, J . R.; Kerns, B. J .; Harvey, E. L.; Conaway, M. R.;
Iglehart, J . D.; Berchuck, A.; Bast, R. C. Overexpression of the
protein tyrosine phosphatase PTP1B in human breast cancer:
association with p185c-erbB-2 protein expression. J . Natl.
Cancer Inst. 1994, 86, 372-378.
(9) Liu, F.; Hill, D. E.; Chernoff, J . Direct binding of the proline-
rich region of protein tyrosine phosphatase 1B to the Src
homology 3 domain of p130(Cas). J . Biol. Chem. 1996, 271,
31290-31295.
(26) Rice, R. L.; Rusnak, J . M.; Yokokawa, F.; Yokokawa, S.; Messner,
D. J .; Boynton, A. L.; Wipf, P.; Lazo, J . S. A targeted library of
small-molecule, tyrosine, and dual-specificity phosphatase in-
hibitors derived from a rational core design and random side
chain variation. Biochemistry 1997, 36, 15965-15974.
(27) Malamas, M. S.; Sredy, J .; Gunawan, I.; Mihan, B.; Sawicki, D.
R.; Seestaller, L.; Sullivan, D.; Flam, B. R. New azolidinediones
as inhibitors of protein tyrosine phosphatase 1B with antihy-
perglycemic properties. J . Med. Chem. 2000, 43, 995-1010.
(28) Malamas, M. S.; Sredy, J .; Moxham, C.; Katz, A.; Xu, W.;
McDevitt, R.; Adebayo, F. O.; Sawicki, D. R.; Seestaller, L.;
Sullivan, D.; Taylor, J . R. Novel benzofuran and benzothiophene
biphenyls as inhibitors of protein tyrosine phosphatase 1b with
antihyperglycemic properties. J . Med. Chem. 2000, 43, 1293-
1310.
(29) Wrobel, J .; Sredy, J .; Moxham, C.; Dietrich, A.; Li, Z.; Sawicki,
D. R.; Seestaller, L.; Wu, L.; Katz, A.; Sullivan, D.; Tio, C.; Zhang,
Z.-Y. PTP1B inhibition and antihyperglycemic activity in the ob/
ob mouse model of novel 11-arylbenzo[b]naphtho[2,3-d]furans
and 11-arylbenzo[b]naphtho[2,3-d]thiophenes. J . Med. Chem.
1999, 42, 3199-3202.
(30) Iverson, L. F.; Anderson, H. S.; Branner, S.; Mortensen, S. B.;
Peters, G. H.; Norris, K.; Olsen, O. H.; J eppesen, C. B.; Lundt,
B. F.; Ripka, W. B.; Moller, K. B.; Moller, N. P. H. Structure-
based design of a low molecular weight, nonphosphorus, non-
peptide, and highly selective inhibitor of protein-tyrosine phos-
phatase 1B. J . Biol. Chem. 2000, 275, 10300-10307.
(31) Andersen, H. S.; Iversen, L. F.; J eppesen, C. B.; Branner, S.;
Norris, K.; Rasmussen, H. B.; Moller, K. B.; Moller, N. P.
2-(Oxalylamino)-benzoic acid is a general, competitive inhibitor
of protein-tyrosine phosphatases. J . Biol. Chem. 2000, 275,
7101-7108.
(32) Puius, Y. A.; Zhao, Y.; Sullivan, M.; Lawrence, D. S.; Almo, S.
C.; Zhang, Z. Y. Identification of a second aryl phosphate-binding
site in protein-tyrosine phosphatase 1B: a paradigm for inhibitor
design. Proc. Natl. Acad. Sci. U.S.A. 1997, 94, 13420-13425.
(33) Montserat, J .; Chen, L.; Lawrence, D. S.; Zhang, Z.-Y Potent low
molecular weight substrates for protein-tyrosine phosphatase.
J . Biol. Chem. 1996, 271, 7868-7872.
(34) Taing, M.; Keng, Y. F.; Shen, K.; Wu, L.; Lawrence, D. S.; Zhang,
Z. Y. Potent and highly selective inhibitors of the protein tyrosine
phosphatase 1B. Biochemistry 1999, 38, 3793-3803.
(35) Desmarais, S.; Friesen, R. W.; Zamboni, R.; Ramachandran, C.
[Difluro(phosphono)methyl]phenylalanine-containing peptide in-
hibitors of protein tyrosine phosphatases. Biochem. J . 1999, 337,
219-223.
(36) J ia, Z.; Barford, D.; Flint, A. J .; Tonks, N. K. Structural basis
for phosphotyrosine peptide recognition by protein tyrosine
phosphatase 1B. Science 1995, 268, 1754-1758.
(37) Laskowski, R. A.; MacArthur, M. W.; Moss, D. S.; Thronton, J .
M. PROCHECK: a program to check the stereochemical quality
of protein structures. J . Appl. Crystallogr. 1993, 27, 283-291.
(38) Groves, M. R.; Yao, Z.-J .; Roller, P. P.; Burke, T. R.; Barford, D.
Structural basis for inhibition of the protein tyrosine phos-
phatase 1B by phosphotyrosine peptide mimetics. Biochemistry
1998, 37, 17773.
(39) Yao, Z. J .; Ye, B.; Wu, X. W.; Wang, S.; Wu, L.; Zhang, Z. Y.;
Burke, T. R., J r. Structure-based design and synthesis of small
molecule protein-tyrosine phosphatase 1B inhibitors. Bioorg.
Med. Chem. 1998, 6, 1799-1810.
(10) Burke, T. R., J r.; Zhang, Z.-Y. Protein-tyrosine phosphatases:
structure, mechanism, and inhibitor discovery. Biopolymers
1998, 47, 225-241.
(11) Ripka, W. C. Protein Tyrosine Phosphatase Inhibitors Annu.
Rep. Med. Chem. 2000, 35, 231-250.
(12) Burke, T. R.; Kole, H. K.; Roller, P. P. Potent inhibition of insulin
receptor dephosphorylation by a hexamer peptide containing the
phosphotyrosyl mimetic F2Pmp. Biochem. Biophys. Res. Com-
mun. 1994, 204, 129-134.
(13) Chen, L.; Wu, L.; Otaka, A.; Smyth, M. S.; Roller, P. R.; Burke,
T. R.; den Hertog, J .; Zhang, Z.-Y. Why is phosphonodifluorom-
ethyl phenylalanine
a more potent inhibitory moiety than
phosphonomethyl phenylalanine toward protein-tyrosine phos-
phatases? Biochem. Biophys. Res. Commun. 1995, 216, 976-
984.
(14) Moran, E. J .; Sarshar, S.; Cargill, J . F.; Shahbaz, M. M.; Lio,
A.; Mjalli, A. M. M.; Armstrong, R. W. Radio frequency tag
encoded combinatorial library method for the discovery of
tripeptide-substituted cinnamic acid inhibitors of protein ty-
rosine phosphatases. J . Am. Chem. Soc. 1995, 117, 10787-
10788.
(15) Liotta, A. S.; Kole, H. K.; Fales, H. M.; Roth, J .; Bernier, M. A
synthetic tris-sulfotyrosyl dodecapeptide analogue of the insulin
receptor 1146-kinase domain inhibits tyrosine dephosphorylation
of the insulin receptor in situ. J . Biol. Chem. 1994, 269, 22996.
(16) Desmarais, S.; J ia, Z.; Ramachandran, C. Inhibition of protein
tyrosine phosphatases PTP1B and CD45 by sulfotyrosyl pep-
tides. Arch. Biochem. Biophys. 1998, 345, 225-231.
(17) Kole, H. K.; Akamatsu, M.; Ye, B.; Yan, X.; Barford, D.; Roller,
P. P.; Burke, T. R. Protein-tyrosine phosphatase inhibition by a
peptide containing the phosphotyrosyl mimetic, L-O-malonylty-
rosine. Biochem. Biophys. Res. Commun. 1995, 209, 817-822.
(18) Burke, T. R.; Ye, B.; Akamatsu, M.; Ford, H.; Yan, X.; Kole, H.
K.; Wolf, G.; Shoelson, S. E.; Roller, P. P. 4′-O-[2-(2-fluoroma-
lonyl)]-L-tyrosine: a phosphotyrosyl mimic for the preparation
of signal transduction inhibitory peptides. J . Med. Chem. 1996,
39, 1021-1027.
(19) Akamatsu, M.; Roller, P. P.; Chen, L.; Zhang, Z.-Y.; Ye, B.; Burke,
T. R., J r. Potent inhibition of protein-tyrosine phosphatase by
phosphotyrosine-mimic containing cyclic peptides. Bioorg. Med.
Chem. 1997, 5, 157-163.
(20) Gao, Y.; Wu, L.; Luo, J . H.; Guo, R.; Yang, D.; Zhang, Z.-Y.;
Burke, T. R., J r. Examination of novel non-phosphorus-contain-
ing phosphotyrosyl mimetics against protein-tyrosine phos-
phatase-1B and demonstration of differential affinities toward
Grb2 SH2 domains. Bioorg. Med. Chem. Lett. 2000, 10, 923-
927.
(40) Salmeen, A.; Anderson, J . N.; Myers, M. P.; Tonks, N. K.;
Barford, D. Molecular basis for the dephosphorylation of the
activation segment of the insulin receptor by protein tyrosine
phosphatase 1B. Mol. Cell 2000, 6, 1401-1412.
(41) Ibarra-Sanchez, M. de J .; Simoncic, P. D.; Nestel, F. R.; Duplay,
P.; Lapp, W. S.; Tremblay, M. L. The T-cell protein tyrosine
phosphatase. Semin. Immunol. 2000, 12, 379-386.
(42) Walchli, S.; Curchod, M.-L.; Gobert, R. P.; Arkinstall, S.; Hooft
van Huijsduuijnen, R. Identification of tyrosine phosphatases
that dephosphorylate the insulin receptor. A brute force ap-
proach based on “substrate-trapping” mutants. J . Biol. Chem.
2000, 275, 9792-9796.
(21) Kole, H. K.; Smyth, M. S.; Russ, P. L.; Burke, T. R., J r.
Phosphonate inhibitors of protein-tyrosine and serine/threonine
phosphatases. Biochem. J . 1995, 311 (3), 1025-1031.

4594 J ournal of Medicinal Chemistry, 2001, Vol. 44, No. 26
J ia et al.
(43) You-Ten, K. E.; Muise, E. S.; Itie, A.; Michaliszyn, E.; Wagner,
J .; J othy, S.; Lapp, W. S.; Tremblay, M. L. Impaired bone marrow
microenvironment and immune function in T cell protein ty-
rosine phosphatase-deficient mice. J . Exp. Med. 1997, 186, 683-
693.
(44) Hooft van Huijsduuijnen, R. Protein tyrosine phosphatases:
counting the trees in the forest. Gene 1998, 225, 1-8.
(45) Yang, J .; Liang, X.; Niu, T.; Meng, W.; Zhao, Z.; Zhou, G. W.
Crystal Structure of the Catalytic Domain of Protein-Tyrosine
Phosphatase SHP-1 J . Biol. Chem. 1998, 273, 28199-28207.
(46) Fauman, E. B.; Cogswell, J . P.; Lovejoy, B.; Rocque, W. J .;
Holmes, W.; Montana, V. G.; Piwinica-Worms, H.; Rink, M. J .;
Saper, M. A. Crystal structure of the catalytic domain of the
human cell cycle control phosphatase, Cdc25A. Cell 1998, 93,
617-662.
(52) Wang, Q.; Scheigetz, J .; Gilbert, M.; Snider, J .; Ramachandran,
C. Fluorescein monophosphates as fluorogenic substrates for
protein tyrosine phosphatases. Biochim. Biophy. Acta 1999,
1431, 14-23.
(53) Summers, M. D.; Smith, G. E. A Manual for Methods for
Baculovirus Vector and Insect Culture Procedures; Bulletin No.
1555; Texas A & M University: Texas Agricultural Experiment
Station, College Station, TX, 1987.
(54) Barford, D.; Keller, J . C.; Flint, A. J .; Tonks, N. K. Purification
and crystallization of the catalytic domain of human protein
tyrosine phosphatase 1B expressed in Escherichia coli. J . Mol.
Biol. 1994, 239, 726-730.
(55) Otwinowski, Z.; Minor, W. Processing of X-ray diffraction data
collected in oscillation mode. Methods Enzymol. 1997, 276, 307-
326.
(47) Burke, T. R.; Yao, Z.-J .; Liu, D.-G.; Voight, J .; Gao, Y. Phospho-
tyrosylmimetics in the Design of Peptide-Based Signal Trans-
duction Inhibitors. Biopolymers (Pept. Sci.) 2001, 60, 32-44.
(48) Liu, S.; Dockendorf, C.; Taylor, S. D. Synthesis of protected L-4-
[Sulfono(difluoromethyl)]phenylalanine and its incorporation
into a peptide. Org. Lett. 2001, 3, 1571.
(49) Taylor, S. D.; Dinaut, A. N.; Thadani, A.; Huang, Z. Synthesis
of Benzylic Mono(R,R-difluoromethylphosphonates) and Benzylic
Bis(R,R-difluoromethylphosphonates) via Electrophilic Fluorina-
tion. Tetrahedron Lett. 1996, 45, 8089-8092.
(50) Taylor, S. D.; Kotoris, C. C.; Dinaut, A. N.; Chen, M.-J . Synthesis
of Aryl(difluoromethylenephosphonates) via Electrophilic Fluo-
rination of R-Carbanions of Benzylic Phosphonates with N-
Fluorobenzenesulfonimide. Tetrahedron 1998, 54, 1691-1714.
(51) Huyer, G.; Liu, S.; Kelly, J .; Moffat, J .; Payette, P.; Kennedy,
B.; Tsaprilis, G.; Gresser, M. J .; Ramachandran, C. Mechanism
of inhibition of protein-tyrosine phosphatases by vanadate and
pervanadate. J . Biol. Chem. 1997, 272, 843-851.
(56) Kissinger, C. R.; Gehlhaar, D. K. EPMR: A program for
crystallographic molecular replacement by evolutionary search,
October 1997.
(57) Bru¨nger, A. T.; Adams, P. D.; Clore, G. M.; DeLano, W. L.; Gros,
P.; Grosse-Kunstleve, R. W.; J iang, J . S.; Kuszewski, J .; Nilges,
M.; Pannu, N. S.; Read, R. J .; Rice, L. M.; Simonson, T.; Warren,
G. L. Crystallography and NMR system: a new software system
for macromolecular structure determination. Acta Crystallogr.
1998, D54, 905-921.
(58) McRee, D. E. A visual protein crystallographic software system
for X11/XView. J . Mol. Graphics 1992, 10, 44-46.
(59) Kraulis, P. J . MOLSCRIPT: a program to produce both detailed
and schematic plots of protein structures. J . Appl. Crystallogr.
1991, 24, 946-950.
(60) Merritt, E. A.; Bacon, D. J . Raster3D: Photorealisticmolecular
graphics. Methods Enzymol. 1997 277, 505-524.
J M010266W