2-Aryl-3,3,3-trifluoro-2-hydroxypropionic acids: A new class of protein tyrosine phosphatase 1B inhibitors

Article abstract of DOI:10.1016/j.bmcl.2007.09.069

A new series of non-peptidic, mono-acid protein tyrosine phosphatase 1B (PTP1B) inhibitors has been identified by structure-based design. Compounds with 2-(indol-3-yl)- and 2-phenyl-3,3,3-trifluoro-2-hydroxypropionic acid core units targeted at the enzyme's primary site and a hydrophobic chlorophenylthiazole extension in its 2° site exhibit 3-60 μM IC₅₀s for PTP1B inhibition in an Sf9 cell-based assay.

Full text of DOI:10.1016/j.bmcl.2007.09.069

Available online at www.sciencedirect.com
Bioorganic & Medicinal Chemistry Letters 17 (2007) 6579–6583
2-Aryl-3,3,3-trifluoro-2-hydroxypropionic acids: A new class
of protein tyrosine phosphatase 1B inhibitors
David R. Adams,a,* Achamma Abraham,a Jun Asano,b Catherine Breslin,d
Colin A. J. Dick,c Ulrich Ixkes,a Blair F. Johnston,e Derek Johnston,c
Justin Kewnay,c Simon P. Mackay,e Simon J. MacKenzie,c Morag McFarlane,d
Lee Mitchell,c Daniel Spinksc and Yasuo Takanob
aDepartment of Chemistry, School of Engineering & Physical Sciences, Heriot-Watt University, Edinburgh EH14 4AS, UK
bDiscovery Research Laboratories, Kyorin Pharmaceutical Co., Ltd, 2399-1, Nogi, Nogi-machi, Simotsuga-gun,
Tochigi 329-0114, Japan
cKyorin Scotland Research Laboratories, West of Scotland Science Park, Todd Campus, Glasgow G20 OXA, UK
dScottish Biomedical, West of Scotland Science Park, Todd Campus, Glasgow G20 OXA, UK
eStrathclyde Institute of Pharmacy & Biomedical Sciences, Strathclyde University, 27 Taylor Street, Glasgow G4 0NR, UK
Received 4 September 2007; revised 19 September 2007; accepted 20 September 2007
Available online 25 September 2007
Abstract—A new series of non-peptidic, mono-acid protein tyrosine phosphatase 1B (PTP1B) inhibitors has been identified by struc-
ture-based design. Compounds with 2-(indol-3-yl)- and 2-phenyl-3,3,3-trifluoro-2-hydroxypropionic acid core units targeted at the
enzyme’s primary site and a hydrophobic chlorophenylthiazole extension in its 2° site exhibit 3–60 lM IC50s for PTP1B inhibition in
an Sf9 cell-based assay.
Ó 2007 Elsevier Ltd. All rights reserved.
The binding of insulin to the insulin receptor (IR), a
member of the receptor tyrosine kinase family, enables
ATP binding and autophosphorylation of tyrosine resi-
dues in its intracellular domain. This autophosphoryla-
tion promotes the receptor’s kinase activity and
facilitates interaction with its intracellular substrates.1
Protein tyrosine phosphatase 1B (PTP1B) has been
shown to dephosphorylate the insulin receptor and
downregulate insulin signaling.2–5 Consequently inhibi-
tors for PTP1B are of therapeutic interest because of
their potential for suppressing dephosphorylation of
the IR, thereby maintaining the receptor in an activated
state and potentiating insulin signaling. Such com-
pounds may have therapeutic utility for treating type 2
diabetes, which is characterized by insulin resistance.6
Several series of PTP1B inhibitors have been disclosed.
The majority of these act competitively and contain
one or other of a variety of phosphotyrosyl (pTyr)
mimetic subunits: mainly phosphonic and carboxylic
acids, including a-keto, malonic, cinnamic, salicylic,
and oxalylaminobenzoic acids.6 Here, we present the
design and synthesis of a new inhibitor class possessing
a 2-aryl-3,3,3-trifluoro-2-hydroxypropionic acid core
unit targeted at the PTP1B active site.
One of the challenges in developing PTP1B inhibitors is
that minimal pTyr mimetics exhibit only modest affinity
for the enzyme compared to its phosphorylated protein
substrates. Indeed the enzyme makes extensive contact
with its substrate proteins over the surface beyond the
immediate enclosed pTyr-binding catalytic pocket. This
is clearly seen in the PTP1B X-ray co-crystal structure
with a peptide fragment, RDI(pY)ETD(pY)(pY)RK,
taken from the IR (Fig. 1).7 The enzyme’s catalytic site
is formed by a ‘P-loop’ motif, 213VHCSAGIGRSG223
,
which is conserved across all PTPs as a characteristic
(I/V)HCXAGXXR(S/T)G signature sequence.8 During
dephosphorylation the substrate pTyr residue is engaged
by the PTP1B P-loop through a salt bridge to R221 and
by additional hydrogen bonds to donors in the main
chain. This engagement induces the closure of a flexible
Keywords: Protein tyrosine phosphatase 1B; PTP1B; Inhibitor;
Diabetes; Obesity; Structure-based design; Insulin resistance; Indole;
Thiazole; Phospho-tyrosine; Insulin sensitivity.
*
Corresponding author. Tel.: +44 0131 4518021; fax: +44 0131
4513180; e-mail: D.R.Adams@hw.ac.uk
0960-894X/$ - see front matter Ó 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2007.09.069

6580
D. R. Adams et al. / Bioorg. Med. Chem. Lett. 17 (2007) 6579–6583
Figure 1. PTP1B inhibitors and ligand binding modes. The PTP1B co-crystal structure (PDB: 1Q1M) with Abbott inhibitor 1 (red stick) is shown.
The IR fragment RDI(pY)ETD(pY)(pY)RK (thin stick) is superimposed from its co-crystal structure with PTP1B (PDB: 1G1F) and hydrogen bonds
D48 in the alternative conformation marked (yellow stick). Indole 4 (black stick) is shown docked in the 1° site.
loop—the ‘WPD loop’—over the top of the pTyr resi-
due, gripping its benzene ring in a clamp between hydro-
phobic residues (F182 and Y46). From the primary pTyr
binding site (Fig. 1, 1° site) the peptide ligand branches
in two directions. On the C-terminal side a second pTyr
side chain is engaged by R24 and R254 in a depression
in the PTP1B surface that forms a secondary binding
site (2° site). On the N-terminal side the ligand runs over
Y46 and engages PTP1B with a salt bridge to R47 (3°
site). The main chain of the peptide ligand is addition-
ally engaged in hydrogen bonds from the D48 side chain
(Fig. 1, yellow stick) where the ligand passes the mouth
of the 1° site.
because computational docking studies suggested that
its benzo ring should superimpose onto the fluoroben-
zene ring of Abbott’s phenylisoxazole 1 (Fig. 1, red
stick), thus allowing us to predict that extensions from
the indole 5-position could be directed into the 2° site.
Compound 1, as the first published example of a PTP1B
inhibitor with activity in a cell-based assay,10a was a log-
ical template choice for developing an extension into the
2° site in a new cell-permeable inhibitor series.
We commenced by preparing a range of simple substi-
tuted indoles (8–11), as summarized in Scheme 1. Com-
pounds were prepared in racemic form and, as cell-based
activity was a key objective, directly assessed in a
PTP1B-specific assay in Sf9 cells.13 Activity data are
summarized in Table 1. No inhibition was observed with
concentrations up to 400 lM for indole 8. This result is
understandable given the flexibility of the enzyme and
the small enclosed volume of the 1° site in its WPD
loop-closed state. Activity was observed, however, with
the introduction of a benzyloxy side chain at the indole
6- and 7-positions, as in 10 and 11 (IC50 382 and
176 lM, respectively). Clearly the activity observed for
these compounds may be sensitive to differences in cell
permeability, but this result was consistent with the
potential for extended surface contact over the edge of
Y46 and/or F182 in these compounds. The 5-benzyloxy
substituted isomer (9), with the side chain targeted at the
gateway leading to the 2° site, did not exhibit observable
activity at 400 lM. Our binding model for core indole 4
suggested that an extended side chain at the 5-position
should allow more effective exploitation of the enzyme’s
2° site however. We therefore prepared indole 12 with a
salicylate-containing side chain similar to that in the
Abbott compound (1). The activity of this racemic com-
pound (IC50 123 lM) was comparable to that of 1 (IC50
110 lM) in our Sf9 cell assay. For reference purposes the
activity of phosphonic acid 2 was also determined (IC50
1.2 lM) and found to be very similar to the value previ-
ously reported by Merck Frosst using the same assay.11a
We next investigated whether the side chain in indole 12
Zhang and coworkers first proposed exploitation of the
2° site as a paradigm for achieving enhanced ligand
binding affinity with bisphosphonate structures that
bridge to the 1° site.9 Compounds that span these two
sites have since been developed by other groups, notably
by Abbott10 (e.g., 1) and by Merck Frosst 11 (e.g., 2 and
3), where ligand binding modes are supported by X-ray
co-crystal structures. Exploitation of the 2° site also use-
fully confers selectivity for inhibition of PTP1B over
other PTPs, as this site and the channel communicating
to the 1° site appear only to be conserved in one other
phosphatase (TCPTP).12
The potential pharmacokinetic problems associated
with peptidic, phosphonate, and multiply ionized com-
pounds led us to consider development of a new PTP1B
inhibitor series, preferably with a single acid functional-
ity. Inspection of X-ray co-crystal structures suggested
that the 3D space in the 1° site may be sub-optimally
used by many pTyr mimetics. In particular we were
struck by the potential of a CF3 group in an (S)-con-
figured 2-aryl-3,3,3-trifluoro-2-hydroxypropionate to
occupy a frequently under-exploited cavity beneath
F182. The hydroxy and acid functions in such a mole-
cule could maintain interactions with the enzyme’s
P-loop. As a starting point we selected 3-indolyl as the
aryl group of the core structure (4, Fig. 1, black stick)

D. R. Adams et al. / Bioorg. Med. Chem. Lett. 17 (2007) 6579–6583
6581
Scheme 1. Reagents and conditions: (a) MOMCl, NaH, DMF, rt (74%-quant.); (b) BnBr, NaH, DMF, rt (71%); (c) SEMCl, NaH, DMF, rt (98%);
(d) F3CC(O)CO2Et, PhMe, D (59%-quant.) (e) KOH (aq), EtOH, D or 1 M TBAF, THF, 40 °C (48–69%); (f) H2, Pd-C, EtOH, rt (quant.);
(g) ICH2CO2Et, K2CO3, DMF, rt (82%); (h) LiAlH4, THF, 0 °C (78%); (i) methyl 2,6-dihydroxybenzoate, diisopropyl azodicarboxylate, PPh3, THF,
rt (70%); (j) H2, Pd-C, EtOH/EtOAc, rt (67%-quant.); (k) 13 (Ref. 16), diisopropyl azodicarboxylate, PPh3, THF, rt (25–39%).
Table 1. PTP1B inhibitory activity for compounds determined in the
Sf9 cell-based assay13
thiazole nitrogen and methyl group should replace the
salicylate ester of 12, with the chlorophenyl group
Compound
IC50/lM
extending more fully over the surface of M258. Pleas-
ingly the activity for these compounds was found to lie
in the 27–59 lM range in our cell-based assay.
1
2
110
1.2
NIa
NIa
382
176
123
59
8
9
A need to avoid acidic N-deprotection conditions for the
rather sensitive indole core motivated the replacement of
the N(1)-MOM group in indoles 8–11 by SEM in 12 and
15. In practice, cleavage of the SEM group was also
found to require harsh conditions (1 M TBAF/THF at
reflux) to which the indole was rather unstable. A desire
to find a more robust core structure as well as to simplify
the target series led us to prepare the phenyl analogue
(17) of indoles 14–16, Scheme 2. The activity of 17
(IC50 37 lM) was found to be comparable to the indoles.
Our earlier finding that the 6- and 7-benzyloxy groups in
compounds 10 and 11 promoted activity of the parent
indole (8) suggested that an additional benzyloxy group
attached to the core of 17 might enhance activity. We
envisaged that such a group would extend the surface
contact between the compound and the protein through
interaction with Y46 or F182. For synthetic convenience
we chose to introduce the additional substituent at the
position ortho to the long side chain targeted at the 2°
site. We therefore prepared the trisubstituted phenyl ser-
ies 18–25 for comparison, Scheme 2. Phenol 18 (IC50
105 lM) was rather less active than parent 17. However,
the introduction of a hydrophobic group (compounds
19–24) maintained or improved the activity of the par-
ent. The benzyloxy substituted analogue (19) was most
active in this series (IC50 3.2 lM) and found to possess
comparable activity (IC50 3.0 lM) in a cell-free PTP1B
10
11
12
14
15
16
17
18
19
20
21
22
23
24
25
32
27
37
105
3.2
12
23
36
10
12
500
a No inhibition was observed for compounds 8 and 9 with concen-
trations up to 400 lM.
could be altered to make more effective use of hydro-
phobic contact over the surface of M258 (Fig. 1, green
surface). The extended aromatic quinoline nucleus ful-
fills this role in the case of compound 3 for example.
We wished to redesign the side chain without introduc-
tion of an additional acid functionality, however, and
chose to test the chlorophenylthiazole unit in com-
pounds 14–16. Our modeling studies suggested that the

6582
D. R. Adams et al. / Bioorg. Med. Chem. Lett. 17 (2007) 6579–6583
Scheme 2. Reagents and conditions: (a) n-BuLi THF, À78 °C then F3CC(O)CO2Et (50%); (b) 10% CF3CO2H/CH2Cl2 (quant.); (c) 13 (Ref. 16),
diisopropyl azodicarboxylate, PBu3, THF, rt (25%); (d) KOH, EtOH aq, 60 °C (34%-quant.); (e) Ac2O, Et3N, CH2Cl2 (quant.); (f) H2, Pd-C, EtOAc,
rt (quant.); (g) 13 (Ref. 16), diisopropyl azodicarboxylate, PPh3, THF, rt (79%); (h) KOH, EtOH, PhMe, 0 °C (quant.); (i) F3CC(O)CO2Et, Et3N,
PhMe, rt (30%); (j) RBr, K2CO3, DMF, rt [yields for nn-ester precursors: 19 (64%), 20 (75%), 21 (quant.), 22 (84%), 23 (40%), 24 (88%), 25 (89%)].
References and notes
1. For reviews of insulin signaling, see: (a) Ferreira, I. A.;
Akkerman, J. W. N. Vitam. Horm. 2005, 70, 25; (b) Gual,
P.; Le Marchand-Brustel, Y.; Tanti, J. F. Biochimie 2005,
87, 99; (c) Kido, Y.; Nakae, J.; Accili, D. J. Clin.
Endocrinol. Metab. 2001, 86, 972.
2. Goldstein, B. J.; Li, P.-M.; Ding, W.; Ahmad, F.; Zhang,
W.-R. Vitam. Horm. 1998, 54, 67.
3. Elchebly, M.; Payette, P.; Michaliszyn, E.; Cromlish, W.;
Collins, S.; Loy, A. L.; Normandin, D.; Cheng, A.;
Himms-Hagen, J.; Chan, C.-C.; Ramachandran, C.;
Gresser, M. J.; Tremblay, M. L.; Kennedy, B. P. Science
1999, 283, 1544.
4. Klaman, L. D.; Boss, O.; Peroni, O. D.; Kim, J. K.;
Martino, J. L.; Zabolotny, J. M.; Moghal, N.; Lubkin, M.;
Kim, Y.-B.; Sharpe, A. H.; Stricker-Krongrad, A.; Shul-
man, G. I.; Neel, B. G.; Kahn, B. B. Mol. Cell. Biol. 2000,
20, 5479.
5. Iversen, L. F.; Andersen, H. S.; Branner, S.; Mortensen, S.
Figure 2. Docked model of (S)-19 (black stick/mesh surface) in the
B.; Peters, G. H.; Norris, K.; Olsen, O. H.; Jeppesen, C.
PTP1B catalytic site; PTP1B solvent accessible surface is shown (solid;
B.; Lundt, B. F.; Ripka, W.; Møller, K. B.; Møller, N. P.
color coding as Fig. 1).
H. J. Biol. Chem. 2000, 275, 10300.
6. For reviews of PTP1B inhibitors, see: (a) Zhang, S.; Zhang,
Z.-Y. Drug Discovery Today 2007, 12, 373;
assay with para-nitrophenyl phosphate at pH 7.0.14
(b) Lee, S.; Wang, Q. Med. Res. Rev. 2007, 27, 553;
Computational docking15 of compound 19 (Fig. 2) sup-
(c) Bialy, L.; Waldmann, H. Angew. Chem. Int. Ed. 2005,
44, 3814; (d) Liu, G. Drugs Future 2004, 29, 1245;
(e) Taylor, S. D.; Hill, B. Exp. Opin. Invest. Drug. 2004,
ported the proposed binding model in which the chlor-
ophenylthiazole unit occupies the enzyme’s 2° site and
13, 199; (f) Taylor, S. D. Curr. Top. Med. Chem. 2003, 3,
the benzyloxy group folds over Tyr46. Interestingly,
the introduction of a polar oxyacetic acid side chain
759; (g) Zhang, Z.-Y.; Lee, S. Y. Exp. Opin. Invest. Drug.
2003, 12, 223; (h) Zhang, Z.-Y. Annu. Rev. Pharmacol.
(25) was substantially detrimental to activity. This is
Toxicol. 2002, 42, 209.
consistent with the requirement for a hydrophobic
7. Salmeen, A.; Andersen, J. N.; Myers, M. D.; Tonks, N.
extension from the core structure envisaged in our bind-
ing model, but it is also possible that the additional acid
functionality compromises cell permeability.
K.; Barford, D. Mol. Cell 2000, 6, 1401.
8. Pannifer, A. D. B.; Flint, A. J.; Tonks, N. K.; Barford, D.
J. Biol. Chem. 1998, 273, 10454.
9. 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.
In summary, conformational mobility, a small enclosed
catalytic pocket, and extensive substrate contacts out-
with the active site present challenges for the design of
PTP1B inhibitors. Here, we have used a structure-based
approach to design novel 2-aryl-3,3,3-trifluoro-2-
hydroxypropionic acid ligands for the PTP1B 1° site.
These compounds exhibit PTP1B inhibitory activity in
a cell-based assay and provide a starting point for a
new series of non-peptidic, mono-acid PTP1B inhibitors.
10. (a) Liu, G.; Xin, Z. L.; Pei, Z. G.; Hajduk, P. J.; Abad-
Zapatero, C.; Hutchins, C. W.; Zhao, H. Y.; Lubben, T.
H.; Ballaron, S. J.; Haasch, D. L.; Kaszubska, W.;
Rondinone, C. M.; Trevillyan, J. M.; Jirousek, M. R.
J. Med. Chem. 2003, 46, 4232; (b) Zhao, H. Y.; Liu, G.;
Xin, Z. L.; Serby, M. D.; Pei, Z. H.; Szczepankiewicz, B.
G.; Hajduk, P. J.; Abad-Zapatero, C.; Hutchins, C. W.;
Lubben, T. H.; Ballaron, S. J.; Haasch, D. L.; Kaszubska,

D. R. Adams et al. / Bioorg. Med. Chem. Lett. 17 (2007) 6579–6583
6583
W.; Rondinone, C. M.; Trevillyan, J. M.; Jirousek, M. R.
Bioorg. Med. Chem. Lett. 2004, 14, 5543; (c) Xin, Z. L.;
Liu, G.; Abad-Zapatero, C.; Pei, Z. H.; Hajduk, P. J.;
Ballaron, S. J.; Stashko, M. A.; Lubben, T. H.; Trevillyan,
J. M.; Jirousek, M. R. Bioorg. Med. Chem. Lett. 2003, 13,
3947.
Payette, P.; Ramachandran, C.; Kennedy, B. P.; Scapin,
G. Bioorg. Med. Chem. Lett. 2004, 14, 1043.
12. (a) Romsicki, Y.; Kennedy, B. P.; Asante-Appiah, E.
Arch. Biochem. Biophys. 2003, 414, 40; (b) Iversen, L. F.;
Moller, K. B.; Pedersen, A. K.; Peters, G. H.; Petersen, A.
S.; Andersen, H. S.; Branner, S.; Mortensen, S. B.; Moller,
N. P. J. Biol. Chem. 2002, 277, 19982.
11. (a) Dufresne, C.; Roy, P.; Wang, Z.; Asante-Appiah, E.;
Cromlish, W.; Boie, Y.; Forghani, F.; Desmarais, S.;
Wang, Q.; Skorey, K.; Xu, L.; Gordon, R.; Chan, C. C.;
Waddleton, D.; Ramachandran, C.; Kennedy, B. P.;
Leblanc, Y. Bioorg. Med. Chem. Lett. 2004, 14, 1039; (b)
Lau, C. K.; Bayly, C. I.; Gauthier, J. Y.; Li, C. S.; Therien,
M.; Asante-Appiah, E.; Cromlish, W.; Boie, Y.; Forghani,
F.; Desmarais, S.; Wang, Q.; Skorey, K.; Waddleton, D.;
13. Cromlish, W. A.; Payette, P.; Kennedy, B. P. Biochem.
Pharmacol. 1999, 58, 1539.
14. Bergmeyer, H. U. Methods Enzymatic Anal. 1974, 1, 496.
15. Computational docking was performed within the Insight-
II modeling and simulations suite from Accelrys.
16. Binggeli, A.; Grether, U.; Hilpert, H.; Hirth, G.; Maerki,
H.-P.; Meyer, M.; Mohr, P. WO2004020420, 2004.