Structural basis for the improved potency of Peroxisome Proliferator-Activated Receptor (PPAR) agonists

Article abstract of DOI:10.1002/cmdc.201000194

The need to develop safer and more effective antidiabetic drugs is essential owing to the growth worldwide of the diabetic population. Targeting the PPAR receptor is one strategy for the treatment of diabetes; the PPAR agonists rosiglitazone and pioglitazone are already on the market. Here we report the identification of a potent PPAR agonist, 15, whose PPARγ activation was more than 20 times better than that of rosiglitazone. Compound 15 was designed to incorporate an indole head with a carboxylic acid group, and 4-phenylbenzophenone tail to achieve a PPARγ EC₅₀ of 10 nm. Compound 15 showed the most potent PPARγ agonist activity among the compounds we investigated. To gain molecular insight into the improved potency of 15, a structural biology study and binding energy calculations were carried out. Superimposition of the X-ray structures of 15 and agonist 10 revealed that, even though they have the same indole head part, they adopt different conformations. The head part of 15 showed stronger interactions toward PPARγ; this could be due to the presence of the novel tail part 4-phenylbenzophenone, which could enhance the binding efficiency of 15 to PPARγ.

Full text of DOI:10.1002/cmdc.201000194

DOI: 10.1002/cmdc.201000194
Structural Basis for the Improved Potency of Peroxisome
Proliferator-Activated Receptor (PPAR) Agonists
Yi-Hui Peng,[a, b] Mohane Selvaraj Coumar,[a] Jiun-Shyang Leou,[a] Jian-Sung Wu,[a] Hui-
Yi Shiao,[a] Chia-Hui Lin,[a] Wen-Hsing Lin,[a] Tzu Wen Lien,[a] Xin Chen,[a] John T.-A. Hsu,[a, c] Yu-
Sheng Chao,[a] Chien-Fu Huang,[d] Ping-Chiang Lyu,[b] Hsing-Pang Hsieh,*[a] and Su-
Ying Wu*[a, b]
The need to develop safer and more effective antidiabetic
drugs is essential owing to the growth worldwide of the dia-
betic population. Targeting the PPAR receptor is one strategy
for the treatment of diabetes; the PPAR agonists rosiglitazone
and pioglitazone are already on the market. Here we report
the identification of a potent PPAR agonist, 15, whose PPARg
activation was more than 20 times better than that of rosiglita-
zone. Compound 15 was designed to incorporate an indole
head with a carboxylic acid group, and 4-phenylbenzophenone
tail to achieve a PPARg EC50 of 10 nm. Compound 15 showed
the most potent PPARg agonist activity among the compounds
we investigated. To gain molecular insight into the improved
potency of 15, a structural biology study and binding energy
calculations were carried out. Superimposition of the X-ray
structures of 15 and agonist 10 revealed that, even though
they have the same indole head part, they adopt different con-
formations. The head part of 15 showed stronger interactions
toward PPARg; this could be due to the presence of the novel
tail part 4-phenylbenzophenone, which could enhance the
binding efficiency of 15 to PPARg.
Introduction
Type 2 diabetes mellitus (T2DM) results from insulin resistance,
which is prevalent among obese individuals who also have hy-
pertension, hyperlipidemia, and atherosclerosis. In the United
States, more than half of adults are overweight; and around
7% have T2DM. According to the World Health Organization
(WHO), the number of people suffering from diabetes world-
wide was 171 million in 2000 and will likely reach 366 million
in 2030 (Diabetes: Facts & Figures. http://www.who.int/diabe-
tes/facts/world_figures/en/, accessed April 30, 2010).
They reduce hepatic insulin resistance and gluconeogenesis,
thereby reducing the glucose and insulin levels in the blood.
Many researchers have found that PPARg agonists are critical
determinants for adipose cell differentiation and that PPARg is
the action site of antidiabetes TZD drugs.[6] Troglitazone, the
first PPARa/g TZD agonist, was approved in 1997, only to be
withdrawn in March 2000 because of hepatotoxicity. Recently,
rosiglitazone, a PPARg selective agonist, was shown to increase
cardiovascular risk in patients with T2DM.[7] These adverse
events pushed the FDA to add a boxed warning in August
2007 to prescription information for TZDs. Muraglitazar, the
first non-TZD/fibrate PPARa/g agonist, was withdrawn after
Peroxisome proliferator-activated receptors (PPARs), which
belong to the orphan nuclear hormone receptor superfamily,[1]
are valid drug targets for modulating the insulin pathway and
increasing insulin sensitivity.[2] PPARs are ligand-activated tran-
scription factors and regulate the transcription of target genes
by forming heterodimers with retinoid X receptors (RXRs). The
ligand binding domain (LBD) of PPARs has a volume of
~1300 ꢀ3, which is larger than those of other nuclear receptors
and can accommodate more diverse ligands.[3] The binding of
lipophilic small molecules in the LBD of PPARs induces confor-
mational changes that recruit coactivator proteins to activate
the PPARs.[4]
[a] Y.-H. Peng, Dr. M. S. Coumar, J.-S. Leou, J.-S. Wu, Dr. H.-Y. Shiao,
Dr. C.-H. Lin, Dr. W.-H. Lin, T. W. Lien, Dr. X. Chen, Dr. J. T.-A. Hsu,
Dr. Y.-S. Chao, Dr. H.-P. Hsieh, Dr. S.-Y. Wu
Division of Biotechnology and Pharmaceutical Research
National Health Research Institutes
35 Keyan Road, Zhunan, Miaoli County 350 (Taiwan)
Fax: (+886)37-586-456
E-mail: hphsieh@nhri.org.tw
suying@nhri.org.tw
[b] Y.-H. Peng, Dr. P.-C. Lyu, Dr. S.-Y. Wu
Institute of Bioinformatics and Structural Biology
National Tsing Hua University
PPARs have three subtypes—PPARa, PPARg, and PPARd—
which play important roles in glucose and lipid homeostasis.
The fibrated drugs are effective ligands in activating PPARa,
which lowers triglyceride and increases HDL cholesterol syn-
thesis. The thiazolidinedione (TZD) class of drugs—including
rosiglitazone (Avandiaꢁ) and pioglitazone (Actosꢁ)—which acti-
vate PPARg within cellular nuclei and regulate the transcription
of more than 100 genes, are used to treat T2DM patients.[5]
101, Sect. 2, Guangfu Road, Hsinchu 300 (Taiwan)
[c] Dr. J. T.-A. Hsu
Department of Biological Science and Technology
National Chiao Tung University
1001 University Road, Hsinchu 300 (Taiwan)
[d] Dr. C.-F. Huang
Department of Biological Science and Technology, I-Shou University
Kaohsiung 82445 (Taiwan)
ChemMedChem 2010, 5, 1707 – 1716
ꢂ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1707

S.-Y. Wu, H.-P. Hsieh, et al.
MED
clinical trials due to increased cardiovascular adverse events,
such as myocardial infarction, stroke, congestive heart failure,
and transient ischemic attack.[8,9] Due to the adverse side ef-
fects of TZDs, especially serious myocardial infarction and
heart failure, there is an urgent need to search for a class of
PPAR agents distinct from TZDs for the treatment of T2DM.
Figure 1a shows the general design of PPAR agonists; they
comprise an acidic group attached to an aromatic part, which
Figure 2. Design of the most potent (in our studies) PPARg agonist 15.
head part of compound 10 and the 4-phenylbenzophenone
tail part of compound 1 confers on compound 15 the most
potent PPARg agonist activity among the compounds we in-
vestigated. Furthermore, we collected and solved the X-ray co-
crystal structure of 15 to gain insight into PPARg interactions,
and compared the results with the X-ray co-crystal structure of
1 and 10 to further understand the improved potency of 15.
Figure 1. Design of PPAR agonists.
Results and Discussion
Synthesis
in turn linked to the aromatic tail through a spacer. Our previ-
ous studies focused on the use of indole[10–12] and 1,2,3,4-tetra-
hydroquinoline as the head parts,[11] along with various aro-
matic tail parts to obtain potent PPAR agonists with varying
levels of selectivity toward a, g and d subtypes. We recently re-
ported the identification of 4-phenylbenzophenone as a novel
tail-part building block that confers selective and potent
PPARg agonist activity when incorporated into both indole and
tetrahydroquinoline series of compounds through N-linkages,
while the acidic head group is linked through an oxygen
bridge (Figure 1b).[11]
The syntheses of tetrahydroquinoline compounds 3–5, 8, and
9 are shown in Scheme 1. Commercially available 5-hydroxy-
quinoline (16) was O-alkylated with 1-bromo-3-chloropropane
under basic conditions in acetone to give 17. The tetrahydro-
quinoline 18 was obtained by partial hydrogenation of the
quinoline ring of 17 by using a NaBH3CN and BF3 ethyrate re-
ducing system.[13] N-alkylation of the tetrahydroquinoline ring
with 3-bromopropionic acid ethyl ester afforded 19. Coupling
of the appropriate tail-part building blocks A, B, C, F, and G
(ROH) with 19 followed by LiOH-mediated ester hydrolysis af-
forded the desired products 3–5, 8, and 9.
The tetrahydroquinoline compounds 6 and 7 were synthe-
sized through an alternative synthetic method, as shown in
Scheme 2. For this purpose intermediate 18 was first coupled
to tail-part building block C ((4-fluorophenyl)-(4-hydroxy-3-pro-
pylphenyl)methanone) to give 20. This was followed by N-alky-
lation by copper-mediated coupling with acrylonitrile under
Here we report the effect of linking a 4-phenylbenzophe-
none tail part to both indole and tetrahydroquinoline through
an O-linkage; while the acidic head group is linked to the ni-
trogen (Figure 1c). Switching the linking of the tail part from N
to O in the indole series resulted in a compound, 15, that
showed the most potent PPARg activity (Figure 2). The most in-
teresting feature of 15 is that the combination of the indole
1708
www.chemmedchem.org
ꢂ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2010, 5, 1707 – 1716

Improved Potency of PPAR Agonists
Scheme 1. a) 1-bromo-3-chloropropane, K2CO3, cat. KI, acetone, reflux, 6 h, 94%; b) NaBH3CN, BF3 ethyrate, CH3OH, reflux, 1 h, 79%; c) 3-bromopropionic acid
ethyl ester, K2CO3, cat. KI, DMF, 1208C, 4 h, 81%; d) i: ROH (tail-part building blocks A, B, C, F and G), K2CO3, cat. KI, DMF, 1208C, 2 h; ii: LiOH, MeOH/H2O (4:1),
reflux, 2 h, 51–79% over two steps.
Scheme 2. a) ROH (tail-part building block C), K2CO3, cat. KI, DMF, 1208C, 2 h, 85%; b) acrylonitrile, CuI, AcOH, microwave irradiation, 1008C, 0.5 h, 94%;
c) KOH, H2O, methanol or 2-methoxyethanol, reflux, 6 h, 72–83%.
microwave irradiation to give 21. Reflux of 21 under basic con-
ditions in the presence of methanol or 2-methoxyethanol as
solvent produced the desired acid 6 and 7, respectively.
Synthesis of indole compound 15 was carried out as shown
in Scheme 3. Commercially available 5-hydroxyindole was O-al-
kylated with 1-bromo-3-chloropropane under basic conditions,
yielding 22.[14] Further O-alkylation of 4-phenylbenzophenone
(tail part B) with compound 22 gave the intermediate 23. The
desired acid 15 was obtained by N-alkylation of 23 with
methyl-2-bromoacetate under basic conditions, followed by
ester hydrolysis with lithium hydroxide.
Structure–activity relationship studies
The newly synthesized compounds 3–9 and 15 were screened
in cell-based transactivation assays for all three PPAR subtypes;
the EC50 results are summarized in Table 1. According to our
recent studies, compound 1 (a 1,2,3,4-tetrahydroquinoline
head part attached to a 4-phenylbenzophenone tail part B
through N-terminal of tetrahydroquinoline ring) showed excel-
lent potency and selectivity toward PPARg.[11] We reversed the
1,2,3,4-tetrahydroquinoline head part to bear the acidic group
at N-terminal of tetrahydroquinoline and linked a diverse set of
tail parts, A–G (Table 1) at the 5-position through a three-
carbon spacer to obtain compounds 3–9 in order to elucidate
the SAR in this new series. We found that 3, containing tail
part A (naphthalen-2-yl-phenyl-methanone), has PPARa/g dual
activity similar to that of 2, which bears the same tail part. Also
4, bearing the 4-phenylbenzophenone (tail part B), showed
PPARg-selective activity similar to 1 bearing the same tail part
B. However, 4 showed a tenfold decrease in activity for PPARg
activation compared to 1. Attempts to replace 4-phenyl group
in tail part B of 4 with fluoro (tail part C), methoxy (D), and 2-
methoxyethoxy (E) groups resulted in compounds 5–7, which
all showed lower PPARg activity than 4. In addition to using 4-
phenylbenzophenone and related tail parts B–E, we used two
other tail parts, F and G, based on the benzo[d]isoxazole core
to explore the activity in this new series. We previously
showed that incorporation of the benzo[d]isoxazole building
block in indole series resulted in the potent PPAR pan agonists
Scheme 3. a) ROH (tail-part building block B), K2CO3, cat. KI, DMF, 808C, 6 h,
80%; b) i: methyl-2-bromoacetate, K2CO3, cat. KI, CH3CN, reflux, 18 h; ii:
LiOH, MeOH/H2O (4:1), reflux, 2 h, 54% over two steps.
ChemMedChem 2010, 5, 1707 – 1716
ꢂ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemmedchem.org
1709

S.-Y. Wu, H.-P. Hsieh, et al.
MED
compounds 2 and 13 resulted in improved PPARg agonist ac-
tivity for compounds 1 and 14,[11] thus suggesting that the 4-
phenylbenzophenone group might also be intrinsically better
suited for PPARg agonist activity. Based on these line of evi-
dence and thought, compound 15 was designed and synthe-
sized by replacing the naphthalen-2-yl-phenyl-methanone (tail
part A) of 10 with 4-phenylbenzophenone (tail part B). Com-
pound 15 with the acid head group attached to the N-terminal
of indole and tail part B attached to the 5-position showed a
sevenfold increase in potency for PPARg and a 12-fold decrease
in activity at PPARa receptor compared to 10, as well as a
complete loss of activity at the PPARd receptor. Compound 15
showed the most potent PPARg agonist activity among the
compounds we investigated, with 12-fold better PPARg activity
than PPARa activity. This result further supported the hypothe-
sis that 4-phenylbenzophenone indeed imparted potent and
PPARg-selective activity when incorporated as the tail part of
the indole series of PPAR agonists. Of particular interest is that
15 showed over 20-fold better PPARg activation than the stan-
dard compound rosiglitazone, which is currently used clinically
for antidiabetes therapy (Table 1).
Table 1. PPAR agonist activity of new analogues 3–9 and 15.
Structure biology studies
Compd
Acid group
+
Tail
part
Transactivation assay (TA)
Based on the excellent activity of compound 15, we co-crystal-
lized PPARg (207–477 a.a.) with 15 and solved the structure to
1.97 ꢀ resolution (Table 2). The data clearly show occupancy of
EC50 [mm][a]
head part
PARa
PPARg
PPARd
1[11]
2[11]
3
4
5
6
7
8
9
10[14]
11[10]
12[10]
13[11]
14[11]
15
I
I
II
II
II
II
II
II
II
III
III
III
IV
IV
III
B
A
A
B
C
D
E
3.99
0.12
0.05
0.05
0.10
0.22
0.54
7.4
>10
>10
>10
>10
–
Table 2. Crystallographic data and refinement characteristics of PPARg
LBD co-crystallized with 15.
>10
–
–
–
>10
1.84
1.04
0.43
0.07
0.23
0.29
0.65
0.12
0.01
0.22
–
–
>10
Parameter
Values
F
0.30
0.03
0.01
0.01
0.02
0.12
1.92
0.12
resolution [ꢀ]
30–1.97
G
A
F
G
A
B
B
–
0.72
unit cell P21
a=55.404, b=87.744, c=57.674
0.50
0.01
0.08
a=g=908, b=90.730
947444
38549
total reflection obs.
unique reflections
multiplicity
>10
24.57
>10
>10
>10
Rmerge [%] (outer shell)
hI/s(I)i (outer shell)
completeness [%] (outer shell)
Rwork [%]
5.5 (36.8)
16.56 (3.16)
97.6 (97.9)
23.07
rosiglitazone
>10
[a] Values are within ꢀ15% (nꢁ2).
Rfree [%]
27.48
0.012
1.378
RMS[a] bonds [ꢀ]
RMS[a] angles [8]
11 and 12.[10] However, we found that incorporation of 3-tri-
fluoromethylbenzo[d]isoxazole (tail part F) in the new tetrahy-
droquinoline series resulted only in the weak dually PPARa/g
active compound 8. Compound 9 with 3-phenylbenzo[d]isoxa-
zole building block (tail part G), however, showed pan-PPAR
agonist activity.
[a] RMS=root-mean-square deviation.
15 in the LBD of PPARg. The carboxylic acid of 15 formed four
hydrogen bonds with Tyr473 (2.50 ꢀ), His449 (2.69 ꢀ), His323
(2.75 ꢀ) and Ser289 (3.12 ꢀ; Figure 3a and b; for clarity, Ser289
is omitted in Figure 3b). According to previous reports,[3a] the
H-bonds contributed to the stabilized interactions of activation
function helix 2 (AF-2 Helix) in PPARg, and this helped to re-
cruit coactivators and activate the PPAR transcription function
of downstream target genes. The indole head of 15—located
around helixes 3, 5, 7, and 10 of the PPARg LBD—formed hy-
drophobic interactions with Phe282 and Cys285 on helix 3 and
Interestingly, we found that compounds 10–12 with an
indole head part had higher PPARg agonist activity than com-
pounds 3, 8, and 9, which bear the same tail parts A, F, and G,
respectively, but have a tetrahydroquinoline head part; this
suggests that the indole head parts of 10–12 might be intrinsi-
cally better suited for PPARg agonist activity. Furthermore, we
previously showed that replacing naphthalen-2-yl-phenyl-
methanone (tail part A) with 4-phenylbenzophenone (B) in
1710
www.chemmedchem.org
ꢂ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2010, 5, 1707 – 1716

Improved Potency of PPAR Agonists
perimposition of 15 with 10 revealed, surprisingly, that the
head part and the linker of these two ligands adopt different
conformations and occupy different binding positions despite
having the same chemical structure. The linker and the indole
ring of 15 moved closer to the AF2 helix by 0.62 ꢀ and 0.66 ꢀ,
respectively, compared with 10; this led to the shift of the car-
boxylic acid group. The movement of the carboxylic acid
group enabled the formation of two additional H-bonds with
Ser289 and His323 (Figure 4). In addition to the stronger H-
Figure 4. Two-dimensional representation of the H-bond (a) and hydro-
phobic (c) interactions of the carboxylic acid, indole and propyl moieties
of compounds a) 10 and b) 15 with the surrounding amino acid residues of
PPARg LBD, based on X-ray co-crystal data. For clarity, the interactions of the
linker and the tail part with the surrounding amino acid residues are omit-
ted.
bond interactions, the indole ring of 15 formed more hydro-
phobic interactions with His449. Moreover, compared with 10,
the propyl moiety of 15 shifted 1.01 ꢀ closer to Arg288 on
helix 3 and 0.92 ꢀ closer to Ile326 on helix 5 to make extra hy-
drophobic interactions with these active-site residues. The dif-
ference in the head part binding and the shift of the propyl
moiety in the 4-phenylbenzophenone tail part might provide
the structural basis for the stronger activity of 15 for PPARg.
We also superimposed the crystal structures of 1 and 15
(Figure 3b), both of which have the same tail parts. We found
that the 4-phenylbenzophenone tail parts occupy similar posi-
tions in the crystal structures. The head parts of these two
compounds contributed to the major difference in their poses.
The indole ring of 15 was rotated 90.038 relative to the tetra-
hydroquinoline ring of 1 along the axis perpendicular to the
plane of the ring. From the superimposition of 1 and 15, we
found that the carboxylic acid head part of 15 got closer to
Tyr473, His449, and His323 of the PPARg LBD, thereby forming
stronger H-bond interactions than in 1. The indole ring head
of 15 also formed stronger hydrophobic interactions with
His449 than the tetrahydroquinoline ring of 1.
Figure 3. Superimposition of crystal structures of a) 10 (orange; PDB ID:
2F4B) and b) 1 (green; PDB ID: 3GBK) on 15 (yellow; PDB ID: 3NOA). For
clarity, Ser289 is omitted in panel b).
His449 on helix 10. The linker of 15 formed hydrophobic inter-
actions with Leu330 on helix 5. The 4-phenylbenzophenone
tail of 15 formed hydrophobic interactions with Leu270,
Cys285, Ile341, Gly284, Leu255, Glu259, and Arg280. The hy-
drophobic interactions of the 4-phenylbenzophenone tail with
the surrounding protein residues play important roles in the
improved PPARg activity of 15. Moreover, the propyl group of
15 interacted hydrophobically with Arg288 and Ile326.
Compound 15—which has the same indole head as 10 but
differs in the tail part—showed sevenfold-improved PPARg
transactivation potency compared to 10. Moreover, the PPARg
binding affinities as measured by scintillation proximity assay
(SPA) also revealed that compound 15 showed stronger PPARg
binding, with an IC50 of 27 nm, than compound 10 (IC50 =
50 nm).[14] To understand the difference in the activity level, we
compared co-crystal structures of 15 with 10 (Figure 3a). Su-
ChemMedChem 2010, 5, 1707 – 1716
ꢂ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemmedchem.org
1711

S.-Y. Wu, H.-P. Hsieh, et al.
MED
In addition, we calculated the binding energy of the three
compounds 1, 10, and 15 in the PPARg LBD as a whole and
also for individual moieties (head/tail parts) in order to get a
better insight into the role played by the component parts in
the interactions of the ligands with the protein (Table 3). Com-
pound 15 showed the most stable conformation with the
PPARg LBD, with ꢂ79.18 kcalmolꢂ1 of binding energy, followed
by 1, with ꢂ60.81 kcalmolꢂ1. Compound 10 showed the least
stability, with ꢂ51.95 kcalmolꢂ1 of binding energy. The total
binding energy correlated with PPARg agonist activity for all
three compounds.
Next, we set out to find the differences in the binding stabil-
ity of distinct parts of the three compounds in order to gain a
better understanding of the role these parts play in ligand
binding. When we compared the binding energy of 1, 10, and
15’s component parts: acid group, head part, and linker
moiety (A+H+L), then 15 showed the lowest binding energy,
with ꢂ42.40 kcalmolꢂ1, compared with ꢂ17.42 kcalmolꢂ1 for
10 and ꢂ26.08 kcalmolꢂ1 for 1. The head parts, along with the
linker of 15, possessed the most binding stability in PPARg,
which corresponded to closer H-bond distances (Table 3) for
the carboxyl acidic head part and stronger hydrophobic inter-
actions with His449 in 15, as discussed above. Further, the
binding energy of the T components of 1, 10, and 15, were
ꢂ37.22 kJmolꢂ1, ꢂ34.43 kJmolꢂ1, and ꢂ39.55 kJmolꢂ1, respec-
tively. Comparison of 10 and 15, which contain the same head,
showed that the 4-phenylbenzophenone tail had a powerful
effect on the overall binding energy of compounds, especially
improving and enhancing the stability of the indole head in
15, thereby imparting a high degree of PPARg potency.
Table 3. Binding energy and H-bond distance of compounds 1, 10 and
15.
Conclusions
There is an urgent need to develop safer and more effective
antidiabetic drugs, owing to the worldwide growth of the dia-
betic population. Targeting the insulin-sensitizing PPARg recep-
tor is an effective strategy, with rosiglitazone and pioglitazone
on the market. Based on general PPAR-agonist design con-
cepts, we used indole and tetrahydroquinoline rings as aro-
matic head groups and linked them to 4-phenylbenzophenone
and other tail groups through spacers to identify several
potent PPAR agonists with varying levels of activity and selec-
tivity toward a, g, and d subtypes. Compound 15, which has
an indole head and 4-phenylbenzophenone tail, displayed
PPARg activation more than 20 times better than that of rosi-
glitazone. Moreover X-ray co-crystal studies and binding-
energy calculations in three structurally related compounds (1,
10, and 15) suggest that the presence of the novel tail part 4-
phenylbenzophenone in 1 and 15 favorably affects com-
pounds’ overall binding efficiency to PPARg. In particular, the
binding stability of 15 with an indole head part is enhanced;
this results in greater PPARg potency for 15 than for 10, which
has the same indole head but a different naphthalen-2-yl-
phenyl-methanone tail. Further efforts are underway to identify
the full potential of 15 for diabetes treatment.
[a]
Ebind [kcalmolꢂ1
]
1
10
15
acid group (A)+
head part (H)+
linker (L)
ꢂ26.08
ꢂ17.42
ꢂ42.40
Experimental Section
tail part (T)
ꢂ37.22
ꢂ60.81
ꢂ34.43
ꢂ51.95
ꢂ39.55
ꢂ79.18
General methods: All commercial chemicals and solvents were re-
agent grade and were used without further treatment unless oth-
erwise noted. All reactions were carried out under dry nitrogen. Re-
actions were monitored by TLC on Merck 60 F254 silica gel glass-
backed plates (5ꢃ10 cm); zones were detected visually under ul-
traviolet irradiation (254 nm) or by spraying with phosphomolybdic
acid reagent (Aldrich) followed by heating at 808C. Flash column
chromatography was performed on silica gel (Merck Kieselgel 60,
total molecule (A+H+L+T)
Y473 (2.70)
H449 (2.88)
H323 (2.97)
S389 (3.12)
Y473 (2.50)
H449 (2.69)
H323 (2.75)
S389 (2.97)
Y473 (3.32)
H449 (2.96)
H-bond length [ꢀ][b]
1
[a] Binding energy was calculated by Insight II. [b] H-bond length be-
tween the carboxylic acid group of compound and the amino acid resi-
due in PPARg LBD.
No. 9385, 230–400 mesh ASTM). H NMR spectra were obtained on
a Varian Mercury-300 or Varian Mercury-400 spectrometer operat-
ing at 300 and 400 MHz, respectively. Chemical shifts are reported
1712
www.chemmedchem.org
ꢂ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2010, 5, 1707 – 1716

Improved Potency of PPAR Agonists
relative to the solvent peak or TMS. High-resolution mass spectra
were measured on a Finnigan (MAT-95XL) electron impact (EI) mass
spectrometer, a Finnigan/Thermo Quest MAT 95XL FAB mass spec-
trometer, or a VG platform Electrospray (ESI) mass spectrometer.
(CH2), 31.4 (CH2), 32.4 (CH2), 41.6 (CH2), 47.5 (CH2), 48.8 (CH2), 60.4
(CH2), 63.9 (CH2), 99.6 (CH), 104.3 (CH), 110.7 (C), 126.7 (CH), 145.5
(C), 156.4 (C), 172.3 ppm (C); HRMS (EI) calcd for C17H24ClNO3:
325.1445 [M]+, found: 325.1429.
5-(3-Chloropropoxy)quinoline (17): A mixture of quinolin-5-ol (16,
1.0 g, 6.89 mmol), 1-bromo-3-chloropropane (1.36 g, 13.7 mmol),
potassium carbonate (1.9 g, 13.7 mmol), and potassium iodide
(0.9 g, 0.1 equiv) in acetone (20 mL) was heated under reflux for
6 h. The mixture was cooled to room temperature and filtered,
then the solvent was removed under vacuum. The residue ob-
tained was partitioned between ethyl acetate (3ꢃ20 mL) and 1n
HCl (20 mL). The organic layer was separated, washed with water
(2ꢃ20 mL) followed by brine (2ꢃ20 mL), and then dried over anhy-
drous sodium sulfate. The solvent was removed in vacuo to give
an oily residue, which was filtered through a short silica column by
eluting with ethyl acetate/hexane (1:5) to give compound 17
(1.43 g, 94%). 1H NMR (400 MHz, CDCl3): d=2.39 (quin, J=6.4 Hz,
2H), 3.84 (t, J=6.4 Hz, 2H), 4.31 (t, J=6.4 Hz, 2H), 6.88 (d, J=
7.6 Hz, 1H), 7.38 (dd, J=8.4, 4.4 Hz, 1H), 7.59 (d, J=8.8 Hz, 1H),
7.61 (d, J=7.6 Hz, 1H), 7.70 (d, J=8.8 Hz, 1H), 8.55 (m, 1H),
8.91 ppm (dd, J=4.4, 1.6 Hz, 1H); 13C NMR (100 MHz, CDCl3): d=
32.1 (CH2), 41.5 (CH2), 64.7 (CH2), 105.1 (CH), 120.2 (CH), 120.7 (C),
121.8 (CH), 129.3 (CH), 130.6 (CH), 149.0 (C), 150.7 (CH), 154.0 ppm
(C); HRMS (EI) calcd for C12H12ClNO: 221.0607 [M]+, found:
221.0611.
3-{5-[3-(6-Benzoyl-1-propylnaphthalen-2-yloxy)propoxy]-3,4-di-
hydro-2H-quinolin-1-yl}propionic acid (3): A solution of 3-[5-(3-
chloropropoxy)-3,4-dihydro-2H-quinolin-1-yl]propionic acid ethyl
ester (19; 0.15 g, 0.462 mmol), (6-hydroxy-5-propylnaphthalen-2-yl)-
phenylmethanone[11] (tail part building block A; 0.106 g,
0.42 mmol), potassium carbonate (0.122 g, 0.92 mmol), and potassi-
um iodide (0.015 g, 0.09 mmol) in DMF (10 mL) were heated at
1208C for 2 h. The mixture was cooled to room temperature,
quenched with 1n HCl (10 mL), and extracted with ethyl acetate
(3ꢃ20 mL). The combined organic layer was dried over anhydrous
sodium sulfate, and the solvent was removed in vacuo to give an
oily residue. A mixture of the above oily residue and LiOH (0.020 g,
0.84 mmol) was held at reflux for 2 h in methanol/water (4:1,
15 mL). The solvent was removed in vacuo, and 1n HCl (10 mL)
was added to the residue, which was extracted with ethyl acetate
(3ꢃ20 mL). The combined organics were dried over anhydrous
sodium sulfate, the solvent was removed in vacuo, and the residue
was chromatographed over a short column of silica gel by eluting
with dichloromethane/methanol (15:1) to give 3 (0.131 g, 57%).
1H NMR (300 MHz, CDCl3): d=1.01 (t, J=7.5 Hz, 3H), 1.64 (sext, J=
7.5 Hz, 2H), 1.88 (m, 2H), 2.22–2.38 (m, J=6.3, 5.7 Hz, 2H), 2.58–
2.69 (m, 4H), 3.07 (t, J=7.6 Hz, 2H), 3.21 (m, 2H), 3.60 (m, 2H),
4.20 (t, J=5.7 Hz, 2H), 4.33 (t, J=6.3 Hz, 2H), 6.30 (d, J=8.4 Hz,
1H), 6.33 (d, J=9.0 Hz, 1H), 7.02 (t, J=8.4 Hz, 1H), 7.32 (d, J=
9.0 Hz, 1H), 7.50 (t, J=7.5 Hz, 2H), 7.61 (t, J=7.5 Hz, 1H), 7.77 (d,
J=9.0 Hz, 1H), 7.83 (d, J=1.8 Hz, 1H), 7.84 (d, J=9.0 Hz, 1H), 7.94
(dd, J=9.0, 1.8 Hz, 1H), 8.03 (d, J=9.0 Hz, 1H), 8.21 ppm (d, J=
1.8 Hz, 1H); 13C NMR (100 MHz, CDCl3): d=14.7 (CH3), 21.5 (CH2),
21.8 (CH2), 23.5 (CH2), 27.3 (CH2), 30.1 (CH2), 48.0 (CH2), 49.2 (CH2),
64.5 (CH2), 65.9 (CH2), 100.5 (CH), 105.0 (CH), 111.6 (C), 114.9 (CH),
124.0 (CH), 124.9 (C), 126.2 (CH), 127.2 (CH), 128.0 (C), 128.5 (2CH),
129.5 (CH), 130.2 (2CH), 132.26 (C), 132.31 (CH), 133.3 (CH), 135.5
(C), 138.5 (C), 145.8 (C), 156.0 (C), 157.0 (C), 177.1 (C), 196.9 ppm
(C); HRMS (ESI) calcd for C35H38NO5: 552.2750 [M]+, found:
552.2682.
5-(3-Chloro-propoxy)-1,2,3,4-tetrahydroquinoline (18): NaCNBH3
(1.3 g, 13.56 mmol) and BF3 ethyrate (1.74 mL, 9.04 mmol) were
added to a solution of 5-(3-chloropropoxy)quinoline (17 1.0 g,
4.52 mmol) in dry methanol (20 mL), and the mixture was heated
under reflux for 1 h. After the reaction mixture had cooled, cold
water was added, then the organics were removed under vacuum.
The residue obtained was partitioned between ethyl acetate (3ꢃ
20 mL) and water; the organic layer was separated and dried over
anhydrous sodium sulfate. The solvent was removed in vacuo, and
the residue was purified over silica column by eluting with ethyl
1
acetate/hexane (1:7) to give compound 18 (0.808 g, 79%). H NMR
(400 MHz, CDCl3): d=1.87–1.94 (m, 2H), 2.21 (quin, J=6.0 Hz, 2H),
2.63 (t, J=6.4 Hz, 1H), 3.23 (t, J=5.6, 2H), 3.74 (t, J=6.4 Hz, 2H),
4.06 (t, J=5.6 Hz, 2H), 6.14 (d, J=8.0 Hz, 1H), 6.19 (d, J=8.0 Hz,
1H), 6.89 ppm (t, J=8.0 Hz, 1H); 13C NMR (100 MHz, CDCl3): d=
20.6 (CH2), 21.8 (CH2), 32.5 (CH2), 41.5 (CH2), 41.7 (CH2), 64.0 (CH2),
100.0 (CH), 107.7 (CH), 110.0 (C), 126.6 (CH), 145.9 (C), 156.9 ppm
(C); HRMS (EI) calcd for C12H16ClNO: 225.0920 [M]+, found:
225.0918.
3-(5-{3-[4-(Biphenyl-4-carbonyl)-2-propylphenoxy]propoxy}-3,4-
dihydro-2H-quinolin-1-yl)propionic acid (4): Compound 4 was
synthesized in
a manner similar to 3 by using 19 (0.15 g,
0.462 mmol) and biphenyl-4-yl-(4-hydroxy-3-propylphenyl)metha-
none[11] (tail part building block B; 0.133 g, 0.42 mmol) as the start-
1
ing materials. Yield: 0.145 g (60%); H NMR (400 MHz, CDCl3): d=
3-[5-(3-Chloropropoxy)-3,4-dihydro-2H-quinolin-1-yl]propionic
acid ethyl ester (19): A solution of 5-(3-chloropropoxy)-1,2,3,4-tet-
rahydroquinoline (18; 0.8 g, 3.58 mmol), 3-bromopropionic acid
ethyl ester (1.73 mL, 10.74 mmol), potassium carbonate (0.99 g,
7.16 mmol), and potassium iodide (0.118 g, 0.72 mmol) in DMF
(10 mL) was heated at 1208C for 4 h. The mixture was cooled to
room temperature, the solvent was removed, 1n HCl (10 mL) was
added, and the resulting mixture was extracted with ethyl acetate
(3ꢃ20 mL). The combined organic phase was dried over anhydrous
sodium sulfate, the solvent was removed in vacuo, and the residue
was purified over silica column by eluting with ethyl acetate/
hexane (1:15) to give compound 19 (1.03 g, 89%). 1H NMR
(400 MHz, CDCl3): d=1.24 (t, J=7.6 Hz, 2H), 1.86–1.93 (m, 2H),
2.20 (sext, J=5.6 Hz, 2H), 2.56 (t, J=7.2 Hz, 2H), 2.64 (t, J=6.4 Hz,
2H), 3.21 (t, J=5.6 Hz, 2H), 3.59 (t, J=7.2 Hz, 2H), 3.72 (t, J=
6.4 Hz, 2H), 4.05 (t, J=5.6 Hz, 2H), 4.13 (quin, J=7.6 Hz, 2H), 6.22
(d, J=8.4 Hz, 1H), 6.31 (d, J=8.4 Hz, 1H), 6.99 ppm (t, J=8.4 Hz,
1H); 13C NMR. (100 MHz, CDCl3): d=14.1 (CH3), 21.0 (CH2), 21.1
0.94 (t, J=7.2 Hz, 2H), 1.62 (sext, J=7.2 Hz, 2H), 1.92 (sext, J=
6.0 Hz, 2H), 2.32 (m, 2H), 2.61–2.68 (m, 6H), 3.23 (t, J=5.6 Hz, 2H),
3.61 (t, J=7.2 Hz, 2H), 4.17 (t, J=6.0 Hz, 2H), 4.26 (t, J=6.0 Hz,
2H), 3.65 (t, J=7.2 Hz, 2H), 4.17 (t, J=6.0 Hz, 2H), 4.25 (t, J=
6.0 Hz, 2H), 6.30 (d, J=8.4 Hz, 1H), 6.34 (d, J=8.4 Hz, 1H), 6.91 (d,
J=9.2 Hz, 1H), 7.03 (t, J=8.4 Hz, 1H), 7.38–7.51 (m, 3H), 7.64–7.72
(m, 6H), 7.84 ppm (d, J=8.0, 2H); 13C NMR (100 MHz, CDCl3): d=
14.0 (CH3), 21.2 (CH2), 21.6 (CH2), 22.8 (CH2), 29.5 (CH2), 31.0 (CH2),
32.2 (CH2), 47.7 (CH2), 48.9 (CH2), 64.2 (CH2), 64.8 (CH2), 100.3 (CH),
104.8 (CH), 110.1 (CH), 111.5 (C), 126.8 (2CH), 127.0 (CH), 127.3
(2CH), 128.0 (C), 128.9 (2CH), 129.8 (C), 130.4 (2CH), 130.5 (CH),
131.2 (C), 132.2 (CH), 137.2 (C), 140.1 (C), 144.6 (C), 145.6 (C), 156.8
(C), 160.5 (C), 176.7 (C), 195.5 ppm (C); HRMS (EI) calcd for
C37H39NO5: 577.2828 [M]+, found: 577.2859.
3-(5-{3-[4-(4-Fluorobenzoyl)-2-propylphenoxy]propoxy}-3,4-dihy-
dro-2H-quinolin-1-yl)propionic acid (5): Compound 5 was synthe-
sized in a manner similar to 3 by using 19 (0.15 g, 0.462 mmol)
ChemMedChem 2010, 5, 1707 – 1716
ꢂ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemmedchem.org
1713

S.-Y. Wu, H.-P. Hsieh, et al.
MED
and (4-fluorophenyl)-(4-hydroxy-3-propylphenyl)methanone[11] (tail
part building block C; 0.108 g, 0.42 mmol) as the starting materials.
Yield: 0.112 g (51%); 1H NMR (400 MHz, CDCl3): d=0.86 (t, J=
7.2 Hz, 3H), 1.53 (sext, J=7.2 Hz, 2H), 1.80–1.87 (m, 2H), 2.24
(quin, J=6.0 Hz, 2H), 2.52–2.61 (m, 4H), 3.15 (t, J=5.4 Hz, 2H),
3.54 (t, J=7.2 Hz, 2H), 4.09 (t, J=6.0 Hz, 2H), 4.18 (t, J=6.0 Hz,
2H), 6.21 (t, J=8.4 Hz, 1H), 6.26 (t, J=8.4 Hz, 1H), 6.81–6.84 (m,
1H), 6.95 (t, J=8.4 Hz, 1H), 7.05–7.10 (m, 2H), 7.54–7.57 (m, 2H),
7.70–7.74 ppm (m, 2H); 13C NMR (100 MHz, CDCl3): d=14.0 (CH3),
21.2 (CH2), 21.5 (CH2), 22.7 (CH2), 29.4 (CH2), 29.7 (CH2), 32.2 (CH2),
47.7 (CH2), 48.9 (CH2), 64.1 (CH2), 64.8 (CH2), 100.2 (CH), 104.8 (CH),
110.1 (CH), 111.3 (C), 115.2 (d, J=21.6 Hz, 2CH), 118.6 (C), 126.9
(CH), 129.4 (C), 130.4 (CH), 131.2 (C), 132.1 (CH), 132.3 (d, J=8.8 Hz,
2CH), 134.6 (d, J=2.9 Hz, C), 145.5 (C), 156.7 (C), 160.6 (C), 164.9
(d, J=251 Hz, CF), 194.6 ppm (C); HRMS (ESI) calcd for C31H35FNO5:
520.2499 [M+H]+, found: 520.2513.
residue was partitioned between ethyl acetate (3ꢃ20 mL) and 1n
HCl (10 mL). The combined organic phases were dried over anhy-
drous sodium sulfate, the solvent was removed in vacuo, and the
residue was chromatographed over a short column of silica gel by
eluting with ethyl acetate/hexane (1:7) to give 20 (0.740 g, 85%).
1H NMR. (400 MHz, CDCl3): d=0.94 (t, J=7.6 Hz, 2H), 1.61 (sext, J=
7.6 Hz, 2H), 1.89 (m, 2H), 2.31 (sext, J=6.0 Hz, 2H), 2.56 (t, J=
7.2 Hz, 2H), 2.62 (t, J=7.6 Hz, 2H), 2.65 (t, J=6.4 Hz, 2H), 3.22 (t,
J=5.6 Hz, 2H), 4.15 (t, J=6.0 Hz, 2H), 4.25 (t, J=6.0 Hz, 2H), 6.14
(d, J=8.0 Hz, 1H), 6.21 (d, J=8.4 Hz, 1H), 6.90 (dd, J=8.4, 8.0 Hz,
1H), 6.91 (d, J=2.4 Hz, 1H), 7.11–7.16 (m, 2H), 7.61–7.64 (m, 1H),
7.64 (s, 1H), 7.76–7.81 ppm (m, 2H); 13C NMR (100 MHz, CDCl3): d=
14.0 (CH3), 20.6 (CH2), 21.8 (CH2), 22.7 (CH2), 29.3 (CH2), 32.2 (CH2),
41.4 (CH2), 63.9 (CH2), 64.7 (CH2), 99.8 (CH), 107.7 (CH), 109.8 (C),
110.0 (CH), 115.1 (d, J=21.6 Hz, 2CH), 126.6 (CH), 129.4 (C), 130.3
(CH), 131.2 (C), 132.0 (C), 132.2 (d, J=8.8 Hz, 2CH), 134.6 (d, J=
2.9 Hz, C), 145.9 (C), 156.9 (C), 160.5 (C), 164.9 (d, J=252 Hz, CF),
194.4 ppm (C); HRMS (ESI) calcd for C28H31FNO3: 448.2288 [M+H]+,
found: 448.2307.
3-{5-[3-(7-Propyl-3-trifluoromethylbenzo[d]isoxazol-6-yloxy)pro-
poxy]-3,4-dihydro-2H-quinolin-1-yl}propionic acid (8): Compound
8 was synthesized in a manner similar to 3 by using 19 (0.15 g,
0.462 mmol) and 7-propyl-3-trifluoromethyl-benzo[d]isoxazol-6-ol[10]
(tail part building block F; 0.103 g, 0.42 mmol) as the starting mate-
3-(5-{3-[4-(4-Fluorobenzoyl)-2-propylphenoxy]propoxy}-3,4-dihy-
dro-2H-quinolin-1-yl)propionitrile (21): A mixture of 20 (0.200 g,
0.45 mmol), CuCl (0.109 g, 1.10 mmol), AcOH (6.6 mL, 0.09 mmol),
and acrylonitrile (3.0 mL) was heated to 1008C for 30 min under
microwave irradiation. Then solvent was removed under vacuum,
and the residue was partitioned between ethyl acetate (3ꢃ20 mL)
and water. The combined organic phases were dried over anhy-
drous sodium sulfate, and the solvent was removed in vacuo to
give an oily residue, which was filtered through a short silica
column by eluting with ethyl acetate/hexane (1:7) to give com-
1
rials. Yield: 0.169 g (79%); H NMR (300 MHz, CDCl3): d=0.94 (t, J=
7.5 Hz, 3H), 1.68 (sext, J=7.5 Hz, 2H), 1.91 (m, 2H), 2.32 (sext, J=
6.0 Hz, 2H), 2.63 (t, J=6.6 Hz, 2H), 2.65 (t, J=6.3 Hz, 2H), 2.90 (t,
J=7.5 Hz, 2H), 3.22 (t, J=5.4 Hz, 2H), 3.60 (t, J=7.2 Hz, 2H), 4.16
(t, J=6.0 Hz, 2H), 4.29 (t, J=6.0 Hz, 2H), 6.28 (d, J=8.4 Hz, 1H),
6.33 (d, J=8.4 Hz, 1H), 7.02 (t, J=8.4 Hz, 1H), 7.07 (d, J=8.7 Hz,
1H), 7.54 ppm (d, J=8.7 Hz, 1H); 13C NMR (100 MHz, CDCl3): d=
14.0 (CH3), 14.9 (CH2), 21.1 (CH2), 21.6 (CH2), 22.7 (CH2), 29.4 (CH2),
32.2 (CH2), 48.0 (CH2), 49.4 (CH2), 64.1 (CH2), 64.7 (CH2), 100.4 (CH),
103.8 (CH), 110.0 (CH), 111.3 (C), 115.2 (d, J=21.6 Hz, 2CH), 118.6
(C), 127.1 (CH), 129.5 (C), 130.3 (CH), 131.2 (C), 132.0 (CH), 132.3 (d,
J=8.8 Hz, 2CH), 134.6 (d, J=2.9 Hz, C), 144,5 (C), 156.9 (C), 160.5
(C), 164.9 (d, J=251 Hz, CF), 194.5 ppm (C); HRMS (EI) calcd for
C26H29FN2O5: 506.2029 [M]+, found: 506.2031.
1
pound 21 (0.213 g, 94%). H NMR (400 MHz, CDCl3): d=0.93 (t, J=
7.6 Hz, 2H), 1.61 (sext, J=7.6 Hz, 2H), 1.94 (m, 2H), 2.32 (sext, J=
6.0 Hz, 2H), 2.58 (t, J=6.8 Hz, 2H), 2.62 (t, J=7.6 Hz, 2H), 2.67 (t,
J=6.8 Hz, 2H), 3.32 (t, J=5.6 Hz, 2H), 3.65 (t, J=7.2 Hz, 2H), 4.17
(t, J=6.0 Hz, 2H), 4.25 (t, J=6.0 Hz, 2H), 6.21 (d, J=8.4 Hz, 1H),
6.31 (d, J=8.4 Hz, 1H), 6.90 (d, J=8.8 Hz, 1H), 7.03 (t, J=8.4 Hz,
1H), 7.12–7.17 (m, 2H), 7.61–7.65 (m, 2H), 7.77–7.81 ppm (m, 2H);
13C NMR (100 MHz, CDCl3): d=14.0 (CH3), 14.9 (CH2), 21.1 (CH2),
21.6 (CH2), 22.7 (CH2), 29.4 (CH2), 32.2 (CH2), 48.0 (CH2), 49.4 (CH2),
64.1 (CH2), 64.7 (CH2), 100.4 (CH), 103.8 (CH), 110.0 (CH), 111.3 (C),
115.2 (d, J=21.6 Hz, 2CH), 118.6 (C), 127.1 (CH), 129.5 (C), 130.3
(CH), 131.2 (C), 132.0 (CH), 132.3 (d, J=8.8 Hz, 2CH), 134.6 (d, J=
2.9 Hz, C), 144,5 (C), 156.9 (C), 160.5 (C), 164.9 (d, J=251 Hz, CF),
194.5 ppm (C); HRMS (EI) calcd for C31H34FN2O3: 501.2553 [M+H]+,
found: 501.2549.
3-{5-[3-(3-Phenyl-7-propylbenzo[d]isoxazol-6-yloxy)propoxy]-3,4-
dihydro-2H-quinolin-1-yl}propionic acid (9): Compound 9 was
synthesized in
a manner similar to 3 by using 19 (0.15 g,
0.462 mmol) and 3-phenyl-7-propyl-benzo[d]isoxazol-6-ol[10] (tail
part building block G) (0.106 g, 0.42 mmol) as the starting materi-
1
als. Yield: 0.176 g (78%); H NMR (400 MHz, CDCl3): d=0.97 (t, J=
7.2 Hz, 2H), 1.72 (sext, J=7.2 Hz, 2H), 1.91 (m, 2H), 2.31 (sext, J=
6.0 Hz, 2H), 2.65 (t, J=7.2 Hz, 2H), 2.67 (t, J=6.8 Hz, 2H), 2.93 (t,
J=7.2 Hz, 2H), 3.23 (t, J=5.6 Hz, 2H), 3.61 (t, J=7.2 Hz, 2H), 4.19
(t, J=6.0 Hz, 2H), 4.29 (t, J=6.0 Hz, 2H), 6.31 (d, J=8.0 Hz, 1H),
6.36 (d, J=8.0 Hz, 1H), 7.01 (d, J=8.8 Hz, 1H), 7.04 (t, J=8.4 Hz,
1H), 7.50–7.57 (m, 3H), 7.66 (d, J=8.8, 2H), 7.92–7.95 ppm (m,
2H); 13C NMR (100 MHz, CDCl3): d=14.1 (CH3), 21.2 (CH2), 21.6
(CH2), 22.3 (CH2), 25.7 (CH2), 29.7 (CH2), 31.1 (CH2), 47.6 (CH2), 48.9
(CH2), 64.1 (CH2), 66.0 (CH2), 100.2 (CH), 104.7 (CH), 110.1 (CH), 111.3
(C), 113.1 (C), 114.2 (C), 119.4 (CH), 127.0 (CH), 128.0 (2CH), 129.0
(2CH), 129.3 (C), 130.0 (CH), 145.5 (C), 156.7 (C), 157.2 (C), 158.0 (C),
164.1 (C), 177.6 ppm (C); HRMS (ESI) calcd for C31H34N2O5: 514.2468
[M]+, found: 514.2469.
3-(5-{3-[4-(4-Methoxybenzoyl)-2-propylphenoxy]propoxy}-3,4-di-
hydro-2H-quinolin-1-yl)propionic acid (6):
A solution of 21
(0.200 g, 0.40 mmol) and KOH (0.09 g, 1.60 mmol) in MeOH
(5.0 mL) and H2O (1.0 mL) was held at reflux for 6 h. The mixture
was cooled to room temperature, the solvents were removed
under vacuo, 1n HCl (10 mL) was added, the mixture was extract-
ed with ethyl acetate (3ꢃ20 mL), and the combined organics were
dried over anhydrous sodium sulfate. The solvent was removed in
vacuo, and the residue was filtered through a short silica column
by eluting with dichloromethane/methanol (15:1) to give com-
pound 6 (0.176 g, 83%). 1H NMR (400 MHz, CDCl3): d=0.82–0.93
(m, 3H), 1.54–1.62 (m, 2H), 1.83–1.94 (m, 2H), 2.27–2.32 (m, 2H),
2.53–2.68 (m, 6H), 3.29 (t, J=5.6 Hz, 2H), 2.63 (t, J=6.8 Hz, 2H),
3.87 (s, 3H), 4.15 (t, J=6.0 Hz, 2H), 4.22 (t, J=6.0 Hz, 2H), 6.19 (t,
J=8.4 Hz, 1H), 6.29 (t, J=8.4 Hz, 1H), 6.94 (d, J=8.8 Hz, 2H), 7.00–
7.24 (m, 1H), 7.59–7.63 (m, 2H), 7.77 ppm (d, J=8.8 Hz, 2H);
13C NMR (100 MHz, CDCl3): d=14.1 (CH3), 21.3 (CH2), 22.7 (CH2),
22.8 (CH2), 29.1 (CH2), 31.3 (CH2), 32.2 (CH2), 47.5 (CH2), 48.5 (CH2),
(4-Fluorophenyl)-{3-propyl-4-[3-(1,2,3,4-tetrahydroquinolin-5-
yloxy)propoxy]phenyl}methanone (20): A solution of 5-(3-chloro-
propoxy)-1,2,3,4-tetrahydroquinoline (18; 0.690 g, 2.13 mmol), (4-
fluorophenyl)-(4-hydroxy-3-propylphenyl)methanone[11] (tail part
building block C) (1.73 mL, 1.94 mmol), potassium carbonate
(0.803 g, 5.82 mmol), and potassium iodide (0.064 g, 0.39 mmol) in
DMF (10 mL) was heated at 1208C for 4 h. The mixture was cooled
to room temperature, DMF was removed under vacuum, and the
1714
www.chemmedchem.org
ꢂ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2010, 5, 1707 – 1716

Improved Potency of PPAR Agonists
46.9 (CH2), 55.5 (CH3), 64.1 (CH2), 64.6 (CH2), 101.6 (CH), 104.6 (CH),
110.0 (CH), 113.6 (2CH), 120.1 (C), 127.6 (CH), 128.3 (CH), 130.4
(CH), 130.7 (CH), 130.9 (C), 132.4 (C),133.5 (2CH), 135.6 (C), 147.9
(C), 155.1 (C), 159.6 (C), 161.1 (C), 176.1 (C), 194.5 ppm (C); MS
(FAB) m/z (relative intensity): 531 ([M]+, 8), 369 (28), 307 (44), 289
(46), 154 (100), 69 (100).
was dried over anhydrous sodium sulfate, and the solvent was re-
moved in vacuo to give an oily residue. LiOH (0.032 g, 1.20 mmol)
was added to a solution of this residue in methanol/water (4:1,
15 mL), and the mixture was held at reflux for 2 h. The solvent was
removed in vacuo, 1n HCl was added to the residue, and the mix-
ture was extracted with ethyl acetate (3ꢃ20 mL). The combined or-
ganic layer was dried over anhydrous sodium sulfate. The solvent
was removed in vacuo, and the residue was chromatographed
over a short column of silica gel by eluting with dichloromethane/
methanol (15:1) to give 15 (0.183 g, 54%). 1H NMR (400 MHz,
CDCl3): d=0.94 (t, J=7.6 Hz, 3H), 1.62 (sext, J=7.6 Hz, 2H), 2.33
(quin, J=6.4 Hz, 2H), 2.63 (t, J=7.6 Hz, 2H), 4.24 (t, J=6.4 Hz, 2H),
4.28 (t, J=6.4 Hz, 2H), 4.85 (s, 2H), 6.48 (d, J=3.2 Hz, 1H), 6.88–
6.92 (m, 2H), 7.05 (d, J=3.2 Hz, 1H), 7.11–7.15 (m, 2H), 7.38–7.50
(m, 3H), 7.64–7.70 (m, 6H), 7.84 ppm (d, J=8.4 Hz, 2H); 13C NMR
(100 MHz, CDCl3): d=14.1 (CH3), 22.8 (CH2), 29.4 (CH2), 32.3 (CH2),
47.5 (CH2), 64.8 (CH2), 65.1 (CH2), 102.3 (CH), 104.3 (CH), 109.6 (CH),
110.2 (CH), 112.9 (CH), 126.8 (2CH), 127.3 (2CH), 128.0 (CH), 128.9
(2CH), 129.0 (CH), 129.1 (C), 129.6 (C), 130.5 (2CH), 130.5 (CH),
131.1 (C), 131.8 (C), 132.2 (C), 137.1 (C), 140.1 (C), 144.6 (C), 153.5
(C), 160.6 (C), 172.6 (C), 195.9 ppm (C); MS (FAB, 70 eV) m/z (relative
intensity): 506 ([M+H]+, 5), 505 ([M]+, 12), 503 (40), 295 (26), 307
(100), 235 (41), 155 (100), 137 (100).
3-[5-(3-{4-[4-(2-Methoxyethoxy)benzoyl]-2-propylphenoxy}pro-
poxy)-3,4-dihydro-2H-quinolin-1-yl]propionic acid (7): A solution
of 21 (0.200 g, 0.40 mmol) and KOH (0.09 g, 1.60 mmol) in 2-me-
thoxyethanol (5.0 mL) and H2O (1.0 mL) was held at reflux for 6 h.
The mixture was cooled to room temperature, 1n HCl (10 mL) was
added, the mixture was extracted with ethyl acetate (3ꢃ20 mL),
and the combined organic phases were dried over anhydrous
sodium sulfate. The solvent was removed in vacuo, and the residue
was filtered through a short silica column by eluting with dichloro-
methane/methanol (15:1) to give compound 7 (0.165 g, 72%).
1H NMR (400 MHz, CDCl3): d=0.93 (t, J=7.6 Hz, 2H), 1.50 (quin, J=
7.2 Hz, 2H), 1.86–1.95 (m, 2H), 3.29 (quin, J=6.0 Hz, 2H), 2.58–2.68
(m, 6H), 3.22 (t, J=6.0 Hz, 2H), 3.47 (s, 3H), 3.61 (t, J=7.2 Hz, 2H),
3.79 (t, J=4.8 Hz, 2H), 4.16 (t, J=6.4 Hz, 2H), 4.20 (t, J=4.8 Hz,
2H), 4.24 (t, J=6.4 Hz, 2H), 6.28 (d, J=8.4 Hz, 1H), 6.33 (d, J=
8.4 Hz, 1H), 6.95–7.05 (m, 3H), 7.64–7.60 (m, 2H), 7.74–7.80 ppm
(m, 2H); 13C NMR (100 MHz, CDCl3): d=14.0 (CH3), 21.5 (CH2), 22.8
(CH2), 22.8 (CH2), 29.4 (CH2), 31.1 (CH2), 32.2 (CH2), 47.7 (CH2), 48.9
(CH2), 59.2 (CH2), 64.5 (CH2), 64.7 (CH2), 67.4 (CH2), 70.8 (CH2), 100.1
(CH), 104.7 (CH), 110.0 (CH), 111.3 (2CH), 120.3 (C), 126.9 (CH),
130.1 (CH), 130.9 (C), 131.1 (C), 132.0 (CH), 132.2 (2CH), 145.6 (C),
156.7 (C), 160.1 (C), 161.9 (C), 176.7 (C), 194.9 ppm (C); HRMS (EI)
calcd for C37H39NO5: 575.2883 [M]+, found: 577.2879.
[10]
Biology: PPARg scintillation proximity assays (SPAs) for IC50 de-
terminations and PPARa/g/d TA assays for EC50 determinations for
the compounds were carried out as previously described.[10,12,14]
X-ray co-crystal structure determination
Protein purification: PPARg LBD (amino acids 207–477) was sub-
cloned into pGEX6P-1 (GE Healthcare) and then transformed into
BL21(DE3) cells. The cells were grown at 208C in 2YT medium in-
duced with isopropyl b-d-1-thiogalactopyranoside (IPTG; 0.2 mm)
at OD600 =0.6, for 6 h. Cells were then harvested and resuspended
in lysis buffer. The lysate was centrifuged at 19000 rpm for 30 min,
and the supernatant was purified by Q anion-exchange column
(HiPrep Q XL 16–10, GE Healthcare; 50 mm Tris-Cl pH 7.5, 1 mm
EDTA 0.1 mm PMSF) by using a salt gradient. The elute was puri-
fied by GST affinity column (GSTPrep FF 16/10 Column, GE Health-
care), and then the buffer was exchanged by Desalting column (Se-
phadex G-25 XK 26/20, GE Healthcare; 50 mm Tris-pH 8.0, 150 mm
NaCl, 1 mm EDTA, 1 mm DTT). The GST-tagged protein was digest-
ed with PreScission protease (GE Healthcare) at 48C overnight and
purified by a second GST affinity column. The flow-through was
then collected. The buffer was changed to a crystallization buffer
(20 mm Tris-Cl, pH 8.0, 100 mm NaCl, 5 mm DTT, 1 mm EDTA) and
concentrated for use, then stored at ꢂ808C.
Biphenyl-4-yl{4-[3-(1H-indol-5-yloxy)propoxy]-3-propylphenyl}-
methanone (23): A solution of 5-(3-chloropropoxy)-1H-indole 22
(0.30 g, 1.44 mmol), (biphenyl-4-yl-(4-hydroxy-3-propylphenyl)me-
thanone[11] (tail part building block B; 0.414 g, 1.31 mmol), potassi-
um carbonate (0.271 g, 1.96 mmol), and potassium iodide (0.022 g,
0.13 mmol) in DMF (10 mL) was heated at 808C for 6 h. The mix-
ture was cooled to room temperature, 1n HCl (10 mL) was added,
and the mixture was extracted with ethyl acetate (3ꢃ20 mL). The
combined organic layer was dried over anhydrous sodium sulfate.
The solvent was removed in vacuo, and the residue was chromato-
graphed over a short column of silica gel by eluting with ethyl ace-
tate/hexane (1:4) to give 22 (0.515 g, 80%). 1H NMR (400 MHz,
CDCl3): d=0.95 (t, J=7.6 Hz, 3H), 1.62 (sext, J=7.6 Hz, 2H), 2.35
(quin, J=6.0 Hz, 2H), 2.65 (t, J=7.6 Hz, 2H), 4.24 (t, J=6.0 Hz, 2H),
4.29 (t, J=6.0 Hz, 2H), 6.46–6.48 (m, 1H), 6.87 (dd, J=8.8, 2.4 Hz,
1H), 6.93 (d, J=9.2 Hz, 1H), 7.14 (d, J=2.4 Hz, 1H), 7.18–7.20 (m,
1H), 7.28 (d, J=8.8 Hz, 1H), 7.38–7.51 (m, 3H), 7.64–7.72 (m, 6H),
7.84 (d, J=8.0 Hz, 2H), 8.11 ppm (br, 1H); 13C NMR (100 MHz,
CDCl3): d=14.1 (CH3), 22.8 (CH2), 29.4 (CH2), 32.3 (CH2), 64.8 (CH2),
65.1 (CH2), 102.4 (CH), 103.5 (CH), 110.1 (CH), 111.7 (CH), 112.8 (CH),
124.9 (CH), 126.8 (2CH), 127.3 (2CH), 128.0 (CH), 128.3 (C), 128.9
(2CH), 129.7 (C), 130.4 (2CH), 130.5 (CH), 131.1 (C), 131.2 (C), 132.1
(CH), 137.2 (C), 140.1 (C), 144.5 (C), 153.3 (C), 160.6 (C), 195.6 ppm
(C); HRMS (EI) calcd for C33H31NO3: 489.2304 [M]+, found: 489.2313.
Crystallization and structure determination: The purified PPARg LBD
was concentrated to 8.0 mgmLꢂ1 and co-crystallized by using the
hanging-drop vapor-diffusion method. A solution of 15 (1 mL,
0.5 mm) in buffer (20 mm Tris-Cl pH 8.0, 100 mm NaCl, 5 mm DTT,
1 mm EDTA) was mixed with concentrated PPARg (9 mL) and incu-
bated on ice for 1 h. The pre-incubated complex solution was then
added to an equal amount of crystallization buffer (27.5%
PEG3350, 16 mm sodium citrate) at 188C for crystallization. Diffrac-
tion data were collected at 100 K at the NSRRC synchrotron facility
on station BL13B1 with an ADSC Quantum 315 system. The data
were processed by DENZO and reduced with SCALPACK.[15] The
structure was solved by molecular replacement with MOLREP[16] by
using a monomer of the published PPARg LBD structure (PDB ID:
2HWR) as the search model. Structure refinement and water-mole-
cule addition were done by the program REFMAC.[17] Model build-
ing was performed with the program O.[18] The completed struc-
(5-{3-[4-(Biphenyl-4-ylcarbonyl)phenoxy]propoxy}-1H-indol-1-yl)-
acetic acid (15): A mixture of biphenyl-4-yl{4-[3-(1H-indol-5-yloxy)-
propoxy]-3-propylphenyl}methanone (23; 0.300 g, 0.65 mmol),
methyl-2-bromoacetate (0.295 g, 1.95 mmol), potassium carbonate
(0.269 g, 1.95 mmol), and potassium iodide (0.012 g, 0.07 mmol) in
CH3CN (10 mL) was held at reflux for 18 h. The mixture was cooled
to room temperature, and the solvents were removed under
vacuo. Then 1n HCl (10 mL) was added, and the mixture was ex-
tracted with ethyl acetate (3ꢃ20 mL). The combined organic layer
ChemMedChem 2010, 5, 1707 – 1716
ꢂ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemmedchem.org
1715

S.-Y. Wu, H.-P. Hsieh, et al.
MED
[6] a) J. M. Lehmann, L. B. Moore, T. A. Smith-Oliver, W. O. Wilkison, T. M.
Willson, S. A. Kliewer, J. Biol. Chem. 1995, 270, 12953–12956; b) P. Ton-
tonoz, E. Hu, B. M. Spiegelman, Curr. Opin. Genet. Dev. 1995, 5, 571–
576.
ture model was optimized by using a development version of
PDB_REDO.[19] The X-ray coordinates of 15 in complex with the
PPARg LBD have been deposited as PDB ID: 3NOA.
Energy calculation: The docking module implemented in the pro-
gram Insight II (Accelrys Inc., San Diego, CA) was used to calculate
the binding energy between the compounds and the PPARg LBD.
The hydrogen atoms were first added to the compounds and pro-
tein. The potentials for the compounds and PPARg LBD were sub-
sequently assigned by using the CFF force field. The parameters
for potentials assignment by using the CFF force field were set at
the default values. The interaction energy, a combination of
van der Waals energy and electrostatic energy, between the com-
pounds and PPARg LBD was finally calculated by the docking
module in the Insight II program.
[7] S. E. Nissen, K. Wolski, N. Engl. J. Med. 2007, 356, 2457–2471.
[8] S. E. Nissen, K. Wolski, E. J. Topol, JAMA J. Am. Med. Assoc. 2005, 294,
2581–2586.
[9] C. R. Waites, M. A. Dominick, T. P. Sanderson, B. E. Schilling, Toxicol. Sci.
2007, 100, 248–258.
[10] N. Mahindroo, C. F. Huang, Y. H. Peng, C. C. Wang, C. C. Liao, T. W. Lien,
S. K. Chittimalla, W. J. Huang, C. H. Chai, E. Prakash, C. P. Chen, T. A. Hsu,
C. H. Peng, I. L. Lu, L. H. Lee, Y. W. Chang, W. C. Chen, Y. C. Chou, C. T.
Chen, C. M. Goparaju, Y. S. Chen, S. J. Lan, M. C. Yu, X. Chen, Y. S. Chao,
S. Y. Wu, H. P. Hsieh, J. Med. Chem. 2005, 48, 8194–8208.
[11] C. H. Lin, Y. H. Peng, M. S. Coumar, S. K. Chittimalla, C. C. Liao, P. C. Lyn,
C. C. Huang, T. W. Lien, W. H. Lin, J. T. Hsu, J. H. Cheng, X. Chen, J. S. Wu,
Y. S. Chao, H. J. Lee, C. G. Juo, S. Y. Wu, H. P. Hsieh, J. Med. Chem. 2009,
52, 2618–2622.
[12] N. Mahindroo, Y. H. Peng, C. H. Lin, U. K. Tan, E. Prakash, T. W. Lien, I. L.
Lu, H. J. Lee, J. T. Hsu, X. Chen, C. C. Liao, P. C. Lyu, Y. S. Chao, S. Y. Wu,
H. P. Hsieh, J. Med. Chem. 2006, 49, 6421–6424.
[13] A. Srikrishna, T. J. Reddy, R. Viswajanani, Tetrahedron 1996, 52, 1631–
1636.
[14] N. Mahindroo, C. C. Wang, C. C. Liao, C. F. Huang, I. L. Lu, T. W. Lien, Y. H.
Peng, W. J. Huang, Y. T. Lin, M. C. Hsu, C. H. Lin, C. H. Tsai, J. T. Hsu, X.
Chen, P. C. Lyu, Y. S. Chao, S. Y. Wu, H. P. Hsieh, J. Med. Chem. 2006, 49,
1212–1216.
Acknowledgements
Financial support from Taiwan’s National Health Research Insti-
tutes and National Science Council is gratefully acknowledged.
We would like to thank the staff of beamline BL13B1 and BL13C1
at the National Synchrotron Radiation Research Center (NSRRC),
Taiwan, for technical assistance.
[15] Z. Otwinowski, W. Minor in Methods in Enzymology, Vol. 276: Macromo-
lecular Crystallography, Humana, Totowa, 1997, pp. 307–326.
[16] A. Vagin, A. Teplyakov, J. Appl. Crystallogr. 1997, 30, 1022–1025.
[17] G. N. Murshudov, A. A. Vagin, E. J. Dodson, Acta Crystallogr. Sect. D. Biol.
Crystallogr. 1997, 53, 240–255.
Keywords: diabetes · PPAR agonists · structural biology ·
structure–activity relationships
[18] T. A. Jones, J. Y. Zou, S. W. Cowan, M. Kjeldgaard, Acta Crystallogr. Sect. A
Found Crystallogr. 1991, 47, 110–119.
[1] S. A. Kliewer, K. Umesono, D. J. Noonan, R. A. Heyman, R. M. Evans,
Nature 1992, 358, 771–774.
[19] R. P. Joosten, J. Salzemann, V. Bloch, H. Stockinger, A.-C. Berglund, C.
Blanchet, E. Bongcam-Rudloff, C. Combet, A. L. Da Costa, G. Deleage, M.
Diarena, R. Fabbretti, G. Fettahi, V. Flegel, A. Gisel, V. Kasam, T. Kervinen,
E. Korpelainen, K. Mattila, M. Pagni, M. Reichstadt, V. Breton, I. J. Tickle,
G. Vriend, J. Appl. Crystallogr. 2009, 42, 376–384.
[2] J. P. Berger, T. E. Akiyama, P. T. Meinke, Trends Pharmacol. Sci. 2005, 26,
244–251.
[3] a) R. T. Nolte, G. B. Wisely, S. Westin, J. E. Cobb, M. H. Lambert, R. Kuroka-
wa, M. G. Rosenfeld, T. M. Willson, C. K. Glass, M. V. Milburn, Nature
1998, 395, 137–143; b) H. E. Xu, M. H. Lambert, V. G. Montana, D. J.
Parks, S. G. Blanchard, P. J. Brown, D. D. Sternbach, J. M. Lehmann, G. B.
Wisely, T. M. Willson, S. A. Kliewer, M. V. Milburn, Mol. Cell 1999, 3, 397–
403.
Received: May 4, 2010
[4] C. K. Glass, M. G. Rosenfeld, Genes Dev. 2000, 14, 121–141.
[5] a) H. Yki-Jarvinen, N. Engl. J. Med. 2004, 351, 1106–1118; b) I. W. Camp-
bell, Curr. Mol. Med. 2005, 5, 349–363.
Revised: July 21, 2010
Published online on August 23, 2010
1716
www.chemmedchem.org
ꢂ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2010, 5, 1707 – 1716