Structure based design of a series of potent and selective non peptidic PTP-1B inhibitors

Article abstract of DOI:10.1016/j.bmcl.2003.11.076

A series of benzotriazole phenyldifluoromethylphosphonic acids were found to be potent PTP-1B inhibitors. Molecular modeling on the X-ray crystal structure of the lead structure led to the design of potent PTP-1B inhibitors that show moderate selectivity against TC-PTP, a very closely related protein tyrosine phosphatase.

Full text of DOI:10.1016/j.bmcl.2003.11.076

Bioorganic & Medicinal Chemistry Letters 14 (2004) 1043–1048
Structure based design of a series of potent and selective non
peptidic PTP-1B inhibitors
Cheuk K. Lau,a,* Christopher I. Bayly,a,* Jacques Yves Gauthier,a Chun Sing Li,a
Michel Therien,a Ernest Asante-Appiah,b Wanda Cromlish,b Yves Boie,b
Farnaz Forghani,b Sylvie Desmarais,b Qingping Wang,b Kathryn Skorey,b
Deena Waddleton,b Paul Payette,b Chidambaram Ramachandran,b
Brian P. Kennedyb and Giovana Scapinc
aDepartment of Medicinal Chemistry, Merck Frosst Centre for Therapeutic Research, PO Box 1005, Pointe Claire-Dorval,
Quebec H9R 4P8, Canada
bDepartment of Biochemistry and Molecular Biology, Merck Frosst Centre for Therapeutic Research, PO Box 1005,
Pointe Claire-Dorval, Quebec H9R 4P8, Canada
cDepartment of Medicinal Chemistry, Merck Research Laboratories, PO Box 2000, Rahway, NJ 07065, USA
Received 8 May 2003; accepted 14 November 2003
Abstract—A series of benzotriazole phenyldifluoromethylphosphonic acids were found to be potent PTP-1B inhibitors. Molecular
modeling on the X-ray crystal structure of the lead structure led to the design of potent PTP-1B inhibitors that show moderate
selectivity against TC-PTP, a very closely related protein tyrosine phosphatase.
# 2004 Elsevier Ltd. All rights reserved.
Protein tyrosine phosphatase 1B (PTP-1B) has been
implicated in the negative regulation of insulin signaling
pathway.1À3 Recent studies showed that PTP-1B knock-
out mice exhibit increased insulin sensitivity and are
resistant to diet-induced obesity.4 A correlation between
insulin resistance states and levels of PTP-1B expression in
muscle and adipose tissues in human has been reported.5À7
Treatment with anti-sense oligonucleotide specific for
PTP-1B results in normalization of blood glucose and
insulin levels in animal models of type 2 diabetes.8 These
results taken together suggest that PTP-1B may play a
role in the insulin resistance associated with diabetes and
obesity. Specific PTP-1B inhibitors may thus be therapeutic
beneficial in the treatment of these conditions.9,10
related Tcell protein tyrosine phosphatase (TC-PTP).
Even though the implication of inhibiting TC-PTP is
not yet elucidated, TC-PTP knockout mice are not
viable 3–5 weeks after birth.11 Initial attempts to obtain
X-ray crystal structures of PTP-1B complexed with this
series of compounds were not successful. Lacking
knowledge of how these compounds bind to the active
site, it was difficult to design selective inhibitors against
TC-PTP. In an attempt to develop a PTP-1B inhibitor
that is selective against a broad range of PTPs including
TC-PTP, alternative structures based on the deoxy-
benzoin template were investigated. Herein, we report
the development of a series of potent PTP-1B inhibitors
based on a tetrasubstituted 4-[2-(1H-1,2,3-benzotriazol-
1-yl)ethyl]phenyl(difluoro)methyl-phosphonate scaffold.
The tetrasubstituted methine structure allowed the
exploration of SAR in four directions in space. In
addition, X-ray crystal structures of PTP-1B complexed
with several compounds in this series were obtained.
Molecular modeling based on the structural information
obtained from these crystal structures has led to the
design of potent PTP-1B inhibitors that exhibit moderate
selectivity against TC-PTP. The observed selectivity was
corroborated by X-ray crystal structure of the PTP-1B-
inhibitor complex.
In the preceding paper, colleagues from our laboratory
reported the development of a series of potent and
selective non-peptidic PTP-1B inhibitors based on the
deoxybenzoin scaffold. This series of compounds is
selective for PTP-1B versus a number of other protein
tyrosine phosphatases (PTPs) but not against the closely
* Corresponding author. Tel.: +1-514-428-8622; fax: +1-514-428-
4900; e-mail: lau@merck.com; christopher_bayly@merck.com
0960-894X/$ - see front matter # 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2003.11.076

1044
C. K. Lau et al. / Bioorg. Med. Chem. Lett. 14 (2004) 1043–1048
The tetrasubstituted benzotriazole compounds based on
1 were prepared quite readily as shown in Scheme 1.
Treatment of benzotriazole with benzyl bromide in the
presence of NaH in DMF gave the 1-benzyl-1H-1,2,3-
benzotriazole. Deprotonation of the latter with n-butyl
lithium followed by sequential alkylation with the
appropriate electrophiles gave either the mono- or bis-
phenyldifluoromethylphosphonic acid (PDFMP) deri-
vatives. The same reaction sequence can be applied to
other heterocyclic benzylated benzotriazoles.
these compounds show excellent selectivity against
other phosphatases such as CD-45 but not against
TC-PTP (Table 1).
As noted earlier, the tetrahedral oriented scaffold facil-
itates the exploration of SAR in a wider area of space.
Presumably, in this series of compounds, one of the
phenyldifluoromethylphosphonic acids is anchored into
the phosphate-binding loop of the active site of PTP-1B.
The benzotriazole serves as a second anchor while the
lipophilic chain provides lipophilic interaction with the
enzyme. These three substituents are optimized for
potency but not selectivity against TC-PTP. Attention
was then turned to the study of the SAR of the remaining
phenyl substituent. The results are summarized in Table 2.
The ability of these compounds to inhibit PTP-1B was
assessed with the fluorescein diphosphate (FDP) assay12
and the PTP-1B overexpressed Sf-9 cell assay.13 The
results are summarized in Tables 1–3.
The trisubstituted benzotriazole compound 1 is much less
active as a PTP-1B inhibitor than the tetrasubstituted
analogue 2 which differs by an extra bromobenzyl sub-
stituent in a tetrahedral arrangement. Varying the chain
length of the lipophilic bromobenzyl substituent (3–6)
has little effect on potency. A similar result was reported
in the deoxybenzoin series. This indicates that the rigid
tetrahedral orientation as well as the presence of all four
substituents is essential for the increase in potency.
However replacing the benzylic linkage with an ester
renders the compound (7) less potent. Polar substituents
such as sulfonamide and tetrazole (8 and 9) at the end of
the lipophilic chain have no effect on the potency except
when it is substituted by another difluoromethylphos-
phonic acid (10). Compound 10 is more potent than the
exact analogue of the benzoin series with an IC50=
0.016 mM in the PTP-1B assay. The compound is also
active in the intact Sf9 cell assay with an IC50=0.12
mM. The results suggest that the benzotriazole ring
might have a more significant interaction with the
enzyme at the active site. The ortho-bromo analogue 11
was even more potent with an IC50=0.005 mM. All
Substitution of the phenyl ring with electron with-
drawing or donating group has no effect on potency or
selectivity against TC-PTP (examples 12, 13). Replacing
the phenyl group with nitrogen bearing heterocycles
(14–17) had no effect on potency or selectivity either
while the benzothiazole 14, tetrazole 16 and the phe-
nylthiazole 17 have lost some selectivity against CD-45
(Table 2). Based on SAR at this juncture, the quest for
selectivity against TC-PTP appeared futile. Fortunately,
the X-ray crystal structure of PTP-1B complexed with
compound 12 provided the crucial information on how
this series of compounds binds to the active site. This
allowed us to design selective compounds against TC-
PTP based on the small difference between the two
enzymes in the active site.
Figure 1 shows how compound 12 binds in the active
site of PTP-1B. The phenyldifluoromethylphosphonate
buried in the primary binding site is only partly visible,
mostly obscured by the flap Phe182. The difluorophenyl
makes weak interactions with the protein surface, while
the benzotriazole makes a hydrogen bond to the peptide
backbone between Arg47 and Asp48 (which make a salt
bridge) as well as hydrophobic interactions with the
Arg47 side-chain. A pathway to selectivity lay in the
part of the active site that binds the other phenyldi-
fluoromethylphosphonate. This fragment lies exposed
on the surface with the phenyl ring lying in a shallow
lipophilic shelf where the lipophilic chain of compounds
2–6 presumably also lie. The increased potency over the
hydrogen substituted 1 can be attributed to this lipo-
philic interaction. The difluoromethylphosphonate moiety
does not reach the two arginines (Arg24 andArg254)
associated with the secondary phosphotyrosine (pY)
binding site.14 Based on sequence alignment of the
highly-homologous PTP-1B and TC-PTP, there are
only two amino acid residue differences between the two
enzymes in the first shell of the active site, located at the
periphery of the secondary binding site. Phenylalanine
52 (F52) and the alanine 27 (A27) in PTP-1B (both in
yellow in Fig. 1) are replaced by a tyrosine and a serine
respectively in TC-PTP.
Based on this information, molecular modeling suggested
that extending the lipophilic chain to the secondary binding
site could be achieved readily with a rigid biphenyl linker.
Optimum binding to the two arginines (R24 and R254)
Scheme 1.

C. K. Lau et al. / Bioorg. Med. Chem. Lett. 14 (2004) 1043–1048
1045
Table 1.
Compd
R
H
PTP-1B (IC50 mM)
CD-45 (IC50 mM)
>50
TC-PTP (IC50 mM)
SF9 (IC50 mM)
1
8.04
7.26
—
2
3
4
5
0.089
0.046
0.049
0.038
17
30
50
69
0.092
0.072
0.049
0.024
1.60
0.65
0.46
1.60
6
7
0.098
0.24
64
50
0.062
0.29
4.07
1.7
8
0.069
50
0.082
0.93
9
0.074
0.016
50
50
0.085
0.010
0.37
0.12
10
11
0.005
25
0.004
0.16
at the secondary binding site could be attained by placing
a phosphonic acid at the meta-position of the terminal
phenyl ring. With both phosphonic acids anchored into
the primary and secondary binding site, and the benzo-
triazole to provide the third anchor, the molecule would
be rigidly locked into the active site. Modeling suggested
that with a rigid substituent, like that of a methyl-
quinoline with disubstitution on the methyl substituent
protruding towards the F52 (a slightly more bulky tyrosine
in TC-PTP in this position), some selectivity could be
obtained (Fig. 2). The biphenylphosphonic acid (19)
was prepared and it was found to be the most potent
PTP-1B inhibitor with an IC50=0.003 mM. The binding
mode was confirmed by X-ray of the PTP-1B complexed
with the compound. With the X-ray and modeled active
sites superimposed based on active site Cas, the X-ray
inhibitor was within 1 A RMSD from the modeled
inhibitor, and both secondary phosphonate binding sites
were occupied as predicted. The corresponding methyl-
quinoline phosphonic acid 20, providing the scaffolding
for the designed selective compounds, was less potent.
Both compounds were not selective against TC-PTP.

1046
C. K. Lau et al. / Bioorg. Med. Chem. Lett. 14 (2004) 1043–1048
Table 2.
Compd
R
PTP-1B (IC50 mM)
CD-45 (IC50 mM)
TC-PTP (IC50 mM)
SF9 (IC50 mM)
12
0.012
50
0.008
0.12
13
14
15
16
0.011
0.018
0.020
0.032
50
11
50
7.2
0.016
0.020
0.028
0.016
0.084
0.088
0.062
1.2
17
0.019
50
0.018
0.19
Figure 1. X-ray structure of compound 12 in the active site of PTP-1B.
Active-site residues important for binding and conserved between
PTP-1B and TC-PTP are labeled in white. The two active-site residues
that differ between the two isoforms, Ala27 and Phe52, are depicted in
yellow, with the exposed active site surface of these residues also in
yellow.
Figure 2. Compound 20 modeled into PTP-1B. The ligand parent
scaffold is in green, the designed methylquinolinephosphonate frag-
ment is in orange. Compared to compound 12, the second phospho-
nate is positioned to take advantage of the two arginines in the
secondary binding site (Arg24 and Arg254, see Fig. 1), and the methyl
is poised close to Phe52 so that substitution on the methyl will intro-
duce selectivity against TC-PTP.
However, disubstituting the methyl by making the
methoxyisobutylquinoline analogue 21 showed moderate
selectivity. The four diastereomers of 21 were resolved.
The most potent isomer 21a has an IC50=0.005 mM in
the PTP-1B enzyme assay and an IC50=0.036 mM in
the TC-PTP assay. The X-ray structure of PTP-1B
complexed with this isomer has been obtained and the
predicted structure was confirmed (Fig. 3).
The slightly less potent diastereomer 21b with an
IC50=0.007 mM in the PTP-1B enzyme assay and an
IC50=0.041 mM in the TC-PTP assay is presumed to be
isomeric at the methoxybutyl side chain of the quinoline.

C. K. Lau et al. / Bioorg. Med. Chem. Lett. 14 (2004) 1043–1048
1047
Figure 3. X-ray structure of compound 21a in the active site of
PTP-1B. With a virtually identical binding mode to that predicted
for compound 20 (see Fig. 2), the introduction of the methoxy and
isobutyl substituents on the methylquinoline forced an interaction
with Phe52 in PTP-1B but which was less tolerated by the larger
Tyr at this position in TC-PTP, leading to modest selectivity against
TC-PTP.
Scheme 2.
Table 3.
Compd
R
PTP-1B (IC50 mM)
CD-45 (IC50 mM)
TC-PTP (IC50 mM)
SF9 (IC50 mM)
10
0.016
50
0.010
0.12
18
19
0.016
0.003
50
50
0.010
0.003
0.12
0.24
20
0.012
50
0.020
0.22
21a
21b
21c
21d
0.005
0.007
0.28
50
50
50
32
0.036
0.041
1.13
0.058
0.10
5.34
1.79
0.20
0.93

1048
C. K. Lau et al. / Bioorg. Med. Chem. Lett. 14 (2004) 1043–1048
Since the other two diastereomers are much less potent,
it is reasonable to assume that they are isomeric at the
methine where the benzotriazole and the difluorophenyl
group switched position. The result provides an estimate
on how much binding energy the benzotriazole nitrogen
provides. The results are summarized in Table 3. T he
quinoline derivatives are prepared from 6,8-dibromo-2-
methylquinoline using standard Suzuki coupling reac-
tions and simple functional group transformations as
shown in Scheme 2.
Gresser, M. J.; Tremblay, M. L.; Kennedy, B. P. Science
(Washington, D.C.) 1999, 283, 1544.
5. Ahmad, F.; Considine, R. V.; Goldstein, B. J. J. Clin.
Invest. 1995, 95, 2806.
6. Ahmad, F.; Azevedo, J. L.; Cortright, R.; Dohm, G. L.;
Goldstein, B. J. J. Clin. Invest. 1997, 100, 449.
7. Kusari, J.; Kenner, K. A.; Suh, K. I.; Hill, D. E.; Henry,
R. R. J. Clin. Invest. 1994, 93, 1156.
8. Zinker, B. A.; Rondinone, C. M.; Trevillyan, J. M.; Gum,
R. J.; Clampit, J. E.; Waring, J. F.; Xie, N.; Wilcox, D.;
Jacobson, P.; Frost, L.; Kroeger, P. E.; Reilly, R. M.;
Koterski, S.; Opgenorth, T. J.; Ulrich, R. G.; Crosby, S.;
Butler, M.; Murray, S. F.; McKay, R. A.; Bhanot, S.;
Monia, B. P.; Jirousek, M. R. Proc. Natl. Acad. Sci.
U.S.A 2002, 99, 11357.
In summary, structure based design has provided a series
of very potent PTP-1B inhibitors. Moderate selectivity
against TC-PTP can be obtained by design.
9. Kennedy, B. P.; Ramachandran, C. Biochem. Pharmacol.
2000, 60, 877.
10. Johnson, T. O.; Ermolieff, J.; Jirousek, M. R. Nat. Rev.
Drug Discov. 2002, 1, 696.
References and notes
11. You, T.; Muise, E. S.; Itie, A.; Michaliszyn, E.; Wagner,
J.; Jothy, S.; Lapp, W. S.; Tremblay, M. L. J. Exp. Med.
1997, 186, 683.
12. Huang, Z.; Wang, Q.; Ly, H. D.; Gorvindarajan, A.;
Scheigetz, J.; Zamboni, R.; Desmarais, S.; Ramachandran,
C. J. Biomol. Screen. 1999, 4, 327.
1. Bandyopadhyay, D.; Kusari, A.; Kenner, K. A.; Liu, F.;
Chernoff, J.; Gustafson, T. A.; Kusari, J. J. Biol. Chem.
1997, 272, 1639.
2. Kenner, K. A.; Anyanwu, E.; Olefsky, J. M.; Kusari, J. J.
Biol. Chem. 1996, 271, 19810.
3. Ahmad, F.; Li, P. M.; Meyerovitch, J.; Goldstein, B. J. J.
Biol. Chem. 1995, 270, 20503.
13. Cromlish, W. A.; Payette, P.; Kennedy, B. P. Biochem.
Pharmacol. 1999, 58, 1539.
4. Elchebly, M.; Payette, P.; Michaliszyn, E.; Cromlish, W.;
Collins, S.; Loy, A. L.; Normandin, D.; Cheng, A.;
Himms-Hagen, J.; Chan, C. C.; Ramachandran, C.;
14. 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.