Synthesis and biological activity of L-tyrosine-based PPARγ agonists with reduced molecular weight

Article abstract of DOI:10.1016/S0960-894X(01)00649-7

A series of PPARγ agonists were synthesized from L-tyrosine that incorporated low molecular weight N-substituents. The most potent analogue, pyrrole (4e), demonstrated a K_i of 6.9 nM and an EC₅₀ of 4.7 nM in PPARγ binding and functional assays, respectively. Pyrrole (4e), which is readily synthesized from L-tyrosine methyl ester in four steps, also demonstrated in vivo activity in a rodent model of Type 2 diabetes.

Full text of DOI:10.1016/S0960-894X(01)00649-7

Bioorganic & Medicinal Chemistry Letters 11 (2001) 3111–3113
Synthesis and Biological Activity of L-Tyrosine-based PPARꢀ
Agonists with Reduced Molecular Weight
Kevin G. Liu,a,* Millard H. Lambert,a Andrea H. Ayscue,a Brad R. Henke,b
Lisa M. Leesnitzer,a William R. Oliver, Jr.,b Kelli D. Plunket,a H. Eric Xu,a
Daniel D. Sternbachb and Timothy M. Willsona,*
aNuclear Receptor Discovery Research, GlaxoSmithKline, Research Triangle Park, NC 27709, USA
bMetabolic Diseases Drug Discovery, GlaxoSmithKline, Research Triangle Park, NC 27709, USA
Received 15 July 2001; accepted 13 September 2001
Abstract—A series of PPARg agonists were synthesized from l-tyrosine that incorporated low molecular weight N-substituents.
The most potent analogue, pyrrole (4e), demonstrated a Ki of 6.9 nM and an EC50 of 4.7 nM in PPARg binding and functional
assays, respectively. Pyrrole (4e), which is readily synthesized from l-tyrosine methyl ester in four steps, also demonstrated in vivo
activity in a rodent model of Type 2 diabetes. # 2001 Elsevier Science Ltd. All rights reserved.
Peroxisome proliferator-activated receptors (PPARs)
have been the subject of intense research following the
discovery of their physiological roles in the regulation
glucose and lipid homeostasis.1 PPARg is the molecular
target for the thiazolidinedione (TZD) class of insulin
sensitizing antihyperglycemic agents. Two of the TZDs,
pioglitazone and rosiglitazone (Fig. 1), are currently
approved for treatment of Type 2 diabetes.1
The X-ray crystal structure of farglitazar bound to PPARg
showed that the benzophenone group was inserted into a
lipophilic pocket, while the tyrosine–NH and benzophe-
none C¼O formed an intramolecular hydrogen bond.4
This intramolecular hydrogen bond reduces the basicity
and the polarity of the tyrosine amino group.
Through a process of model building into the PPARg
crystal structure,5 alternative N-substituents were selec-
We recently reported a series of potent tyrosine-based
PPARg agonists, which are exemplified by farglitazar
and GW7845 (Fig. 1).2 Although these compounds
contain an asymmetric center, the more potent S-enan-
tiomers are conveniently synthesized from naturally-
occurring l-tyrosine.2 Unlike the TZDs,3 the tyrosine-
based PPARg agonists do not undergo racemization in
vivo.2 Our initial studies2 suggested that an N-benzo-
phenone or N-benzoate group was essential for PPARg
activity. As part of an ongoing program to develop new
l-tyrosine-based PPARg agonists, we decided to re-
evaluate the structural requirements for the tyrosine
N-substituents. In addition, a goal of the current study
was to employ the X-ray crystal structure of PPARg4 in
the design of molecules with reduced molecular weight
compared to the first generation compounds (Fig. 1),
while maintaining robust PPARg agonist and in vivo
antidiabetic activity.
*Corresponding authors. Fax:+1-919-315-0430; e-mail: kliu@
kalypsys.com (K. G. Liu); tmw20653@gsk.com (T. M. Willson).
Figure 1. Structures of PPARg agonists.
0960-894X/01/$ - see front matter # 2001 Elsevier Science Ltd. All rights reserved.
PII: S0960-894X(01)00649-7

3112
K. G. Liu et al. / Bioorg. Med. Chem. Lett. 11 (2001) 3111–3113
ted that would add a small lipophilic substituent while
mimicking some of the effects of the intramolecular
hydrogen bond present in the benzophenone analogue.
Compounds were built within the PPARg binding site
by removal of the benzophenone group from farglita-
zar, and growth of alternative substituents onto the
tyrosine nitrogen, using the MVP program.5 The
growth process, which generates multiple energy-mini-
mized conformations, was carried out twice, once within
the protein binding site and once in aqueous solution. A
basicity adjustment was applied to the calculated bind-
ing energy to account for the fraction of the molecule in
the neutral ionization state at pH 7.6 The model build-
ing predicted that pyrrole, morpholino and n-alkyl
derivatives would bind to PPARg effectively at pH 7.
Other derivatives, such as amide, piperidine and
dimethylpyrrole, were predicted to be less potent due to
their polarity, basicity or poor steric fit, respectively.
aldehydes. N-Cycloalkyl groups (4b–d) were introduced
through alkylation with dibromoalkylating agents.
Compounds 3c and 3d were prepared through acylation
and sulfonylation, respectively. The pyrimidine com-
pound 3e was generated by aromatic nucleophilic sub-
stitution of 2-chloropyrimidine. Reaction of the free
amine from 2 with 2,5-dimethoxytetrahydrofuran or
with hexane-2,5-dione in acetic acid provided the
pyrrole compounds 4e and 4f, respectively (Scheme 1).
Presence of sodium acetate in the reaction mixture
was found to greatly accelerate pyrrole formation and
to be crucial for high yields. Unfortunately, significant
racemization (up to 10%) occurred when the pyrroles
were formed in acetic acid in the presence of sodium
acetate as determined by chiral HPLC. Gratefully, use
of a biphasic reaction conditions reported by Jefford8
led to pyrrole formation without racemization (Scheme
2). Saponification with aqueous LiOH in dioxane or
THF provided enantiomerically pure carboxylic acids.
The enantiomer of 4e was also synthesized from d-tyr-
osine. Use of pyrrole as a primary amine protecting
group was first introduced by Bruekelman.9 Pyrroles
have been shown to be stable to strong bases, nucleo-
philes, and to brief contact with strong acids. Consistent
with these reports, compounds 4e and 4f are stable to
aqueous LiOH, acidic workup, and storage at room
temperature. Notably, structurally related racemic
pyrroles have been claimed to show in vivo antidiabetic
activity.10
The synthesis of tyrosine analogues 3a–e and 4a–f is
depicted in Scheme 1. Mitsunobu reaction between
commercially available t-butoxycarbonyl (BOC)-pro-
tected l-tyrosine methyl ester 1 and 5-methyl-2-phenyl-
4-oxazolylethanol7 using triphenylphosphine and
diisopropylazodicarboxylate in toluene was followed
by removal of the BOC protecting group with 4 M
HCl in dioxane to afford the amine hydrochloride 2
in 70% overall yield. Intermediate 2 was derivatized
on the tyrosine nitrogen to provide the targeted set
of analogues following saponification with LiOH
(Scheme 1). N-Alkyl groups (3a–b and 4a) were
introduced by reductive alkylation with the requisite
Compounds 3a–e and 4a–f were screened against the
three human PPAR subtypes in binding and cell-based
reporter assays (Table 1). The details of both the bind-
ing2,11 and the functional assays2 have been reported
previously. The compounds (3–4) displayed subtype-
selectivity for PPARg over PPARa and PPARd. How-
ever, the selectivity ranged from 3.5-fold with alkyl
amine (3b) to >700-fold with pyrrole (4e) in the cell-
based assay. Variation in the polarity of the N-sub-
stituent was found to have a dramatic effect on the
PPARg activity. As predicted, the acetyl (3c) and sulfo-
nyl (3d) groups were weak PPARg agonists with EC50
>1 mM, suggesting that polar functionality is poorly
tolerated. The N-propyl amine (3b) and N,N-dimethyl
amine (4a) were more potent than the corresponding
mono-methyl analogue 3, showing that increased
lipophilicity was beneficial.
Scheme 1. Reagents and conditions: (a) 5-methyl-2-phenyl-4-oxazolyl-
ethanol, PPh3, DIAD, PhCH3, 79%; (b) 4 M HCl in dioxane, 100%;
(c) 3a: (i) 2-nitrobenzenesulfonyl chloride, Et3N, CH2Cl2, 98%; (ii)
MeI, Cs2CO3, DMF, 33%; (iii) thiophenol, K2CO3, acetonitrile, 62%;
3b: CH3CH2CHO, NaBH(OAc)3, CH2Cl2; 3c: CH3COCl, Et3N,
CH2Cl2, 98%; 3d: MeSO2Cl, Et3N, CH2Cl2, 46%; 3e: 2-chloropyr-
imidine, Na2CO3, EtOH, 15%; 4a: HCHO, NaCNBH3, HOAc,
CH3CN, 26%; 4b: 1,5-dibromopentane, Na2CO3, EtOH, 42%; 4c:
2-bromoethyl ether, Na2CO3, EtOH, 60%; 4d: 1,4-dibromobutane,
Na2CO3, EtOH, 47%; 4e: 2,5-dimethoxytetrahydrofuran, HOAc,
NaOAc, 61%; 4f: hexane-2,5-dione, HOAc, 36%; (d) LiOH, THF,
H2O, MeOH, 90–100%.
Scheme 2. Reagents and conditions: (a) 2,5-dimethoxytetrahydro-
furan, 1,2-dichloroethane, H2O, 77%; (b) LiOH, dioxane, H2O, 98%.

K. G. Liu et al. / Bioorg. Med. Chem. Lett. 11 (2001) 3111–3113
Table 1. Activity of tyrosine-based PPAR agonists
3113
Compd
R1
R2
Binding affinity, Ki (nM)a
Functional activity, EC50 (nM)b
PPARa
PPARg
PPARd
PPARa
PPARg
PPARd
Pioglitazone
Rosiglitazone
Farglitazar
GW7845
3a
3b
3c
3d
3e
4a
4b
4c
4d
4e
ent-4e
4f
—
—
490
1100
—
72
—
—
1200
—
—
—
630
120
1.1
3.7
2100
52
2600
620
280
850
480
158
360
6.9
—
—
1200
1200
—
—
—
—
—
—
—
—
—
2500
—
—
—
—
450
3500
—
830
—
—
—
—
—
6900
—
3600
1300
—
580
43
—
—
—
—
—
—
—
—
—
—
—
—
—
3800
—
H
H
H
H
H
H
H
Me
2-(COPh)Ph
2-(CO2Me)Ph
Me
n-Pr
MeCO
MeSO2
2-pyrimidine
Me
0.34
0.71
5000
240
3500
2600
590
410
138
24
–CH2CH2CH2CH2CH2–
–CH2CH2OCH2CH2–
–CH2CH2CH2CH2–
—
390
4.7
20
–CH=CHCH=CH–
–CH=CHCH=CH–
–(Me)CH=CHCH=CH(Me)–
560
2600
—
51
980
710
—
aCompounds were assayed for their ability to bind to human PPAR ligand binding domains by scintillation proximity assay.2,11 The Ki value was
calculated according to the equation Ki=IC50/(1+[L]/Kd), where IC50 is the concentration of test compound required to inhibit 50% of the specific
binding of the radioligand, [L] is the concentration of the radioligand used, and Kd is the dissociation constant for the radioligand at the receptor.
bCompounds were assayed for agonist activity on human PPAR-GAL4 chimeric receptors in transiently transfected CV-1 cells as described;2
EC50=the concentration of test compound that gave 50% of the maximal reporter activity; all data are ꢁ20% for n=3; —inactive at 3 mM in the
binding assay or 10 mM in the functional assay.
The basicity of the tyrosine nitrogen was a key factor
that affected the potency on PPARg, with compounds
designed to reduce the nitrogen basicity demonstrating
improved potency in the binding and functional assays.
Thus, the pyrimidine 3e had an EC50 of 590 nM, while
the morpholinyl derivative 4c was more potent than the
isosteric cycloalkyl analogue 4b. The pyrrolidinyl
compound 4d showed similar PPARg potency to its
acyclic analogue 4a, but pyrrole 4e, which reduces the
basicity of the tyrosine nitrogen, was a potent PPARg
agonist with EC50 of 4.7 nM in the functional assay and
a Ki of 6.9 nM in the binding assay. 2,5-Dimethyl pyr-
role 4f was less potent than its unsubstituted analogue
4e, most likely due to an unfavorable steric interaction
with the receptor. Finally, pyrrole 4e (synthesized from
l-tyrosine) was 10 times more potent at PPARg and 5
times more potent at PPARa than its enantiomer ent-4e
(synthesized from d-tyrosine). The higher activity of
pyrrole 4e, with the S-configuration, is consistent with
the enantioselectivity of the first generation tyrosine-
based compounds.2
Acknowledgements
The authors would like to thank Melissa Lindsay and
Manon Villeneuve for the chiral HPLC determination
of enantiomeric purity for 4e.
References and Notes
1. Willson, T. M.; Brown, P. J.; Sternbach, D. D.; Henke,
B. R. J. Med. Chem. 2000, 43, 527.
2. Henke, B. R.; Blanchard, S. G.; Brackeen, M. F.; Brown,
K. K.; Cobb, J. E.; Collins, J. L.; Harrington, W. W., Jr.;
Hashim, M. A.; Hull-Ryde, E. A.; Kaldor, I.; Kliewer, S. A.;
Lake, D. H.; Leesnitzer, L. M.; Lehmann, J. M.; Lenhard,
J. M.; Orband-Miller, L. A.; Miller, J. F.; Mook, R. A.;
Noble, S. A.; Oliver, W.; Parks, D. J.; Plunket, K. D.;
Szewczyk, J. R.; Willson, T. M. J. Med. Chem. 1998, 41, 5020.
3. Parks, D. J.; Tomkinson, N. C. O.; Villeneuve, M. S.;
Blanchard, S. G.; Willson, T. M. Bioorg. Med. Chem. Lett.
1998, 8, 3657.
4. Gampe, R. T., Jr.; Montana, V. G.; Lambert, M. H.;
Miller, A. B.; Bledsoe, R. K.; Milburn, M. V.; Kliewer, S. A.;
Willson, T. M.; Xu, H. E. Mol. Cell 2000, 5, 545.
5. Lambert, M. H. In Practical Application of Computer-Aided
Drug Design; Charifson, P. S., Ed.; Marcel Dekker: New
York, 1997; p 243.
The in vivo antihyperglycemic and antihyperlipidemic
efficacy of 4e were evaluated in male Zucker diabetic
fa/fa rats (n=6). When dosed orally at 10 mg/kg bid for
14 days, the pyrrole (4e) reduced fasting plasma glucose
by 40%, and fasting serum triglycerides by 24%, while
HDL-cholesterol was raised by 31% compared to
vehicle-treated animals.
6. To correct for ionization of the amine at pH 7,
rt*ln½1 þ 10ðpK ÀpHÞꢀ was added to the calculated binding
a
energy, where rt=0.596 kcal at 300 K, and pKa is the acidity
constant for the basic amine.
7. Hulin, B.; Clark, D. A.; Goldstein, S. W.; McDermott,
R. E.; Dambek, P. J.; Kappeler, W. H.; Lamphere, C. H.;
Lewis, D. M.; Rizzi, J. P. J. Med. Chem. 1992, 35, 1853.
8. Jefford, C. W.; de Villedon de Naide, F.; Sienkiewicz, K.
Tetrahedron: Asymmetry 1996, 7, 1069.
9. Bruekelman, S. P.; Leach, S. E.; Meakins, G. D.; Tirel,
M. D. J. Chem. Soc., Perkin Trans. 1 1984, 2801.
10. Faller, A. WO Patent 9429302, 1994. Yanagisawa, H,
Takamura, M., Fujita, T., Fujiwara, T. WO Patent 9737970, 1997.
11. Nichols, J. S.; Parks, D. J.; Consler, T. G.; Blanchard,
S. G. Anal. Biochem. 1998, 257, 112.
In summary, the pyrrole (4e, MW=416) was identified
as a potent l-tyrosine-based insulin sensitizer with
significantly lower molecular weight than the 1st
generation compounds farglitazar (MW=546) and
GW7845 (MW=500). The low basicity of the pyrrole
N-substituent in 4e may be key to its potent PPARg
activity.