Structural insight into peroxisome proliferator-activated receptor γ binding of two ureidofibrate-like enantiomers by molecular dynamics, cofactor interaction analysis, and site-directed mutagenesis

Article abstract of DOI:10.1021/jm9013899

Molecular dynamics simulations were performed on two ureidofibrate-like enantiomers to gain insight into their different potency and efficacy against PPARγ. The partial agonism of the S enantiomer seems to be due to its capability to stabilize different regions of the receptor allowing the interaction with both coactivators and corepressors as shown by fluorescence resonance energy transfer (FRET) assays. The recruitment of the corepressor N-CoR1 by the S enantiomer on two different responsive elements of PPARγ regulated promoters was confirmed by chromatin immunoprecipitation assays. Cell-based transcription assays show that PPARγ coactivator 1α (PGC-1α) and cAMP response element binding protein-binding protein (CBP) enhance the basal and ligand-stimulated receptor activity acting as coactivators of PPARγ, whereas the receptor interacting protein 140 (RIP140) and the nuclear corepressor 1 (N-CoR1) repress the transcriptional activity of PPARγ. We also tested the importance of the residue Q286 on the transcriptional activity of the receptor by site-directed mutagenesis and confirmed its key role in the stabilization of helix 12. Molecular modeling studies were performed to provide a molecular explanation for the different behavior of the mutants.

Full text of DOI:10.1021/jm9013899

4354 J. Med. Chem. 2010, 53, 4354–4366
DOI: 10.1021/jm9013899
Structural Insight into Peroxisome Proliferator-Activated Receptor γ Binding of Two
Ureidofibrate-Like Enantiomers by Molecular Dynamics, Cofactor Interaction Analysis, and
Site-Directed Mutagenesis
Giorgio Pochetti,†,[ Nico Mitro,‡,§,[ Antonio Lavecchia, ,[ Federica Gilardi,‡,z Neva Besker,^ Elena Scotti,‡,þ
Massimiliano Aschi,# Nazzareno Re,^ Giuseppe Fracchiolla,3 Antonio Laghezza,3 Paolo Tortorella,3
Roberta Montanari,† Ettore Novellino, Fernando Mazza,†,O Maurizio Crestani,*,‡,[ and Fulvio Loiodice*,3,[
†Istituto di Cristallografia, Consiglio Nazionale delle Ricerche, Montelibretti, 00015 Monterotondo Stazione, Roma, Italia, ‡Laboratorio
“Giovanni Galli” di Biochimica e Biologia Molecolare dei Lipidi e Spettrometria di Massa, and §Laboratorio “Giovanni Armenise-Harvard
ꢀ
Foundation”, Dipartimento di Scienze Farmacologiche, Universita degli Studi di Milano, 20133 Milano, Italia, Dipartimento di Chimica
Farmaceutica e Tossicologica, Universita degli Studi di Napoli “Federico II”, 80131 Napoli, Italia, ^Dipartimento di Scienze del Farmaco,
ꢀ
Universita “G. D’Annunzio”, 06100 Chieti, Italia, Dipartimento di Chimica, Ingegneria Chimica e Materiali, Universita di L’Aquila, 67010
#
ꢀ
ꢀ
L’Aquila, Italia, Dipartimento Farmaco-Chimico, Universita degli Studi di Bari “Aldo Moro”, 70126 Bari, Italia, and ODipartimento di Scienze
3
ꢀ
della Salute, Universita di L’Aquila, 67010 L’Aquila, Italia. [ These authors contributed equally to this work. z Current address: University of
ꢀ
Lausanne, Center for Integrative Genomics, Genopode Building, Room 5010, 1015 Lausanne, Switzerland. þ Current address: Department of
Pathology and Laboratory Medicine/Howard Hughes Medical Institutes, University of California at Los Angeles, Los Angeles, California.
Received September 18, 2009
Molecular dynamics simulations were performed on two ureidofibrate-like enantiomers to gain insight
into their different potency and efficacy against PPARγ. The partial agonism of the S enantiomer seems
to be due to its capability to stabilize different regions of the receptor allowing the interaction with both
coactivators and corepressors as shown by fluorescence resonance energy transfer (FRET) assays. The
recruitment of the corepressor N-CoR1 by the S enantiomer on two different responsive elements of
PPARγ regulated promoters was confirmed by chromatin immunoprecipitation assays. Cell-based
transcription assays show that PPARγ coactivator 1R (PGC-1R) and cAMP response element binding
protein-binding protein (CBP) enhance the basal and ligand-stimulated receptor activity acting as
coactivators of PPARγ, whereas the receptor interacting protein 140 (RIP140) and the nuclear
corepressor 1 (N-CoR1) repress the transcriptional activity of PPARγ. We also tested the importance
of the residue Q286 on the transcriptional activity of the receptor by site-directed mutagenesis and
confirmed its key role in the stabilization of helix 12. Molecular modeling studies were performed to
provide a molecular explanation for the different behavior of the mutants.
Introduction
The first insight into the link between PPARγ and diabetes
came from the discovery of PPARγ as the biologic target for
the thiazolidinedione class of antidiabetic drugs (TZDs).5,6
TZDs are effective insulin sensitizers and have been shown to
improve glucose uptake and lower hyperglycaemia and
hyperinsulinaemia.6-9 However, as full agonists they also
stimulate adipocyte differentiation in vitro and weight gain
in vivo, which normally aggravates the diabetic state. Addi-
tional undesirable side effects associated with TZD treatment
include edema/hemodilution, cardiomegaly, anemia, and in-
creased incidence of bone fractures.10,11
Peroxisome proliferator-activated receptor γ (PPARγ)a
belongs to the nuclear receptor (NR) superfamily of ligand-
activated transcription factors. PPARγ is a member of the
NR1C subgroup that includes PPARR and PPARβ/δ form-
ing a subfamily of “lipid sensing” receptors that regulate
important aspects of lipid and glucose homeostasis.1,2 PPARγ
is the most extensively studied among the PPAR subtypes and
plays important roles in the functions of adipocytes, muscles,
and macrophages with a direct impact on type 2 diabetes,
dyslipidemia, atherosclerosis, and cardiovascular diseases.3,4
As a result of the clinical observations mentioned above,
emphasis has shifted to the development of partial agonists or
selective PPARγ modulators (SPPARγMs). These new li-
gands, although displaying reduced transcriptional activity in
reporter assays, exhibit robust antidiabetic efficacy with im-
proved tolerability.12-15
Previously published data seem to indicate that the unique
properties of partial agonists may be due to their distinct
physical interaction with the receptor. Indeed, X-ray cocrys-
tallographic studies of the PPARγ ligand binding domain
(LBD) complexed withfull agonists indicate the critical roleof
*To whom correspondence should be addressed. Phone: þ39 080-
5442798, Fax: þ39 080-5442231, Email: floiodice@farmchim.uniba.it;
Phone: þ39 02-50318393, Fax: þ39 02-50318391, Email: maurizio.crestani@
unimi.it.
a Abbreviations: PPAR, peroxisome proliferator-activated receptor;
SPPARM, selective peroxisome proliferator-activated receptor modu-
lator; NR, nuclear receptor; TZDs, thiazolidinediones; LBD, ligand
binding domain; RXR, retinoid X receptor; MD, molecular dynamics;
FRET, fluorescence resonance energy transfer; HOBT, hydroxy-
benzotriazole; DIC, N,N-diisopropylcarbodiimide; RMSF, root-mean-
square fluctuation; ED, essential dynamics; ChIP, chromatin immuno-
precipitation.
r
pubs.acs.org/jmc
Published on Web 05/12/2010
2010 American Chemical Society

Article
Journal of Medicinal Chemistry, 2010, Vol. 53, No. 11 4355
helix 12 (H12) in transcriptional regulation of these nuclear
receptors. The structures of PPARγ bound to partial agonists
show that there are two classes of ligands: those contacting
directly H12, spanning from H11 beyond H3, and those
showing no direct interaction with H12, spanning from H3
to the β-sheet. For both classes of compounds, the importance
of other regions such as H3 and β-sheet is emerging and their
stabilization can also affect the transcriptional response.16-21
Moreover, the recently solved structure of full-length
PPARγ-RXR heterodimer demonstrated that the β-sheet
region of PPARγ binds directly to the RXR-DNA binding
domain.22 Thus, differential binding of ligands in the region
between β-sheet and H3 of PPARγ may also affect transcrip-
tional activity through modulation of its interaction with
RXR.
In the present paper, we focused our attention to these
regions of PPARγ, particularly to the factors stabilizing H3.
In a previous work,23 we reported a structural study on two
enantiomeric ureidofibrate-like derivatives (Figure 1) com-
plexed, respectively, with the PPARγ-LBD.
The R enantiomer, R-1, behaves as a full agonistof PPARγ,
whereas the S enantiomer, S-1, is a less potent partial agonist.
Comparing the X-ray structures of the two complexes, we
argued that the partial agonist behavior of S-1 could be
ascribed to a destabilization of the active conformation of
H12. We showed that the suboptimal conformation of H12,
observed in the PPARγ-S-1 complex, is probably due to a
steric hindrance between the ethyl group, linked to the asym-
metric carbon atom of the ligand, and the crucial residue Q286
of PPARγ, situated on H3.
MD simulation studies on the PPARγ-LBD complexed with
the two enantiomers S-1 and R-1 to investigate more deeply
the role of H3 stabilization in the presence of partial and full
agonists. In this regard, we also examined, during the simula-
tion, the distances between the two residues of the charge
clamp in the liganded and unliganded forms in order to
evaluate if they fit the right values for the recruitment of the
coactivators. We also compared the dynamic behavior of
PPARγ-LBD with that of PPARR by MD to understand if
the importance of the stabilization of H3 in the mechanism of
partial agonism is a property of PPARγ or may be extended to
PPARR. Furthermore, we investigated how the two enantio-
mers affect the recruitment of different coregulators to
PPARγ by performing fluorescence resonance energy transfer
(FRET) and chromatin immunoprecipitation (ChIP) assays
and assessed the functional relevance of some of these cor-
egulators in cell-based transcriptional assays. In addition, we
clarified the role of Q286, belonging to H3, on the transcrip-
tional activity of the receptor by testing the effect of site-
directed mutagenesis of Q286. Molecular modeling was per-
formed to rationalize the activity of the mutants. Finally, we
proposed a hypothesis of mechanism of recruitment of coac-
tivators and corepressors induced upon binding of partial
agonists.
Chemistry
The synthesis of the S-1 isomer is depicted in Scheme 1 and
involved the key intermediate acid S-2, which was prepared
from the condensation of 4-bromo-phenol with 2-butanone in
the presence of CHBr3 and KOH, followed by fractional
crystallization from ethanol of the diastereomeric salts ob-
tained with (R)-1-phenylethylamine. The esterification of acid
S-2, followed by condensation with N-vinylphthalimide in the
presence of Pd(AcO)2, tri-o-tolyl-phosphine, and N,N-diiso-
propylethylamine in anhydrous CH3CN provided the inter-
mediate S-3. The reduction of compound S-3, which was
achieved by hydrogenation at 4 atm in the presence of
Wilkinson catalyst, was followed by hydrazinolysis of the
phthalimide moiety leading to the amine intermediate, whose
condensation with heptanoic acid in the presence of hydro-
xybenzotriazole (HOBT) and N,N-diisopropylcarbodiimide
(DIC) afforded the intermediate S-4. The amide group of
compound S-4 was reduced with 1 M borane in THF solution
to give the corresponding amine. The condensation of this
intermediate with 2-chloro-benzoxazole, followed by saponi-
fication of the ester function, led to the final acid S-1. The
In another X-ray and molecular dynamics (MD) work, in
which we explained at molecular level the different pharma-
cological profile of the enantiomers of a novel PPARR/γ dual
agonist,24 we confirmed that a differential stabilization of H3
plays an important role in determining the partial agonist
character of a ligand. These results prompted us to perform
Figure 1. Schematic diagram of R-1 and S-1 enantiomers.
Scheme 1. Synthesis of the PPAR Agonist S-1a
a (a) CHBr3, KOH; (b) fractional crystallization from EtOH of the (R)-1-phenylethylamine salts; (c) MeOH, H2SO4 cat.; (d) N-vinylphthalimide,
Pd(AcO)2, tri-o-tolylphosphine, N,N-diisopropylethylamine, CH3CN; (e) H2 (4 atm), Wilkinson cat., EtOH, rt; (f) N2H4 xH2O, EtOH; (g) heptanoic
acid, HOBT, DIC, CH2Cl2; (h) 1 M BH3 in THF; (i) 2-chloro-benzoxazole, (Et)3N, THF; (j) 1 N NaOH, EtOH, rt.
3

4356 Journal of Medicinal Chemistry, 2010, Vol. 53, No. 11
Pochetti et al.
Figure 2. CR rms fluctuation of apo-PPARγ (black), and PPARγ-R-1 (red) and PPARγ-S-1 (green) complexes.
Scheme 2. Synthetic Pathway for the Preparation of R-2
(a) 95% NaH, DMF, reflux; (b) H2 (4 atm), 10% Pd/C; (c) 48% HBr, NaNO2, CuBr, from 0 to 80 ꢀC.
Table 1. Backbone Atom Traces
synthesis of the R-1 isomer was carried out by following the
same synthetic pathway described in Scheme 1 starting from
compound R-2. Both enantiomers of acid 1 had enantiomeric
excesses >98% as determined by HPLC analysis on chiral
stationary phase (see Experimental Section).
apo-PPARγ
PPARγ-R-1
PPARγ-S-1
backbone = 13.153 nm2
backbone = 11.099 nm2
backbone = 8.715 nm2
PPARR and the corepressor SMRT,27 H3 is less fluctuating
in the presence of S-1 than R-1 (Figure 1 of Supporting
Information). (c) A comparison between the CR RMSF
analysis of the apo-form of PPARR, performed in a previous
work,24 and that of PPARγ evidences that in PPARR, unlike
PPARγ, H3 is stable even in the absence of the ligand (Figure
2 of Supporting Information). (d) The fluctuation of H12 is
very high in the apo-form and in the complex with R-1; on the
contrary, H12 results more stabilized upon the binding of
S-1. (e) Helices H4 and H5 are also less fluctuating upon the
binding of S-1.
Analysis of the Trajectories. A better view of dynamical-
mechanical properties of the investigated system has been
obtained by using essential dynamics (ED) analysis.28 This
procedure is based on the construction of the backbone
covariance matrix whose trace, reported in Table 1, quanti-
tatively provides the fluctuation pattern of the system. Our
results clearly indicate that S-1 stabilizes PPARγ-LBD
more than R-1.
At the same time, the diagonalization of the covariance
matrix also provides the set of eigenvectors along which the
system fluctuates and the extent of the fluctuation. The
spectrum of the corresponding eigenvalues (Figure 3 of
Supporting Information) indicates that the fluctuation of
the system is basically confined within the first two eigen-
vectors. The 2D projection of the trajectories onto the first
two apo-form eigenvectors shows that the binding of the two
enantiomers (red and green points in Figure 3) dramatically
changes the configurational space sampled by the complexed
The absolute configuration of R-1 and S-1 was determined
by chemical correlation by preparing one of the enantiomers
of the key intermediate 2 through a synthetic pathway with
assigned stereochemistry as depicted in Scheme 2. For this
purpose, the R-2-hydroxy-2-methyl-butanoic acid, obtained
by fractional crystallization of its diastereomeric salts with
brucine,25,26 was condensed with 4-nitro-fluorobenzene to
give the 4-nitro-phenoxy intermediate 5 with retention of
configuration. The hydrogenation of the nitro group, fol-
lowed by the treatment with48% HBr, NaNO2, andCuBr, led
to the acid (þ)-2 with the same R configuration of the starting
R-hydroxy acid.
Results
MD Simulation Analysis. From the analysis of the root-
mean-square fluctuation (RMSF) of PPARγ-LBD before
(apo-form) and after complexation with the enantiomers R-1
and S-1 (Figure 2), the following main considerations can be
drawn: (a) H3 in the apo-form fluctuates to a larger extent
with respect to both the complexed forms. (b) H3 stabiliza-
tion realized by the binding of the ligands is greater in the
case of S-1; although small, the RMSF difference between
the two complexes can be considered significant taking into
account the maximum error equal to (0.02 nm evaluated by
dividing the overall trajectory in two subportions. Even
performing MD simulation starting from an inactive con-
formation of H12, similar to that adopted in the ternary
complex of the antagonist 6 (GW6471, Figure 8) with

Article
Journal of Medicinal Chemistry, 2010, Vol. 53, No. 11 4357
Figure 3. Trajectory projection on the first two eigenvectors of (A) apo-PPARγ (black dots) and of PPARγ complexed with the full agonist
R-1 (red dots) and the partial agonist S-1 (green dots). For clarity’s sake, each trajectory projection is also shown separately in (B-D).
Table 2. EC50 Values of Peptides Recruited to PPARγ in the Presence of
PPARγ-LBD with respect to that of the apo-LBD (black
Saturating Concentrations of Rosiglitazone (10 μM), R-1 (25 μM), and
points).
Interestingly, while the more restricted area sampled by
the R-1 complexed form is inside the extended area of the
S-1 (50 μM)a
EC50 (μM)
coregulator peptide
rosiglitazone
R-1
S-1
apo-form, that one sampled by the S-1 complexed form lies
in a new region, outside the region of the apo-form. In other
words, the presence of R-1 induces mechanical-dynamical
alterations of the LBD backbone to a lower extent with
respect to S-1.
SRC-1 (NR box 2)
PGC-1R (NR box 1)
RIP140 (NR box 4)
RIP140 (NR box 7)
CBP (NR box 1)
2.0435 ( 0.012
2.37 ( 0.245 3.6155 ( 0.652
0.36 ( 0.075 0.3206 ( 0.055 0.4223 ( 0.078
0.2676 ( 0.041 0.2408 ( 0.051 0.3319 ( 0.061
0.88495 ( 0.296 1.0454 ( 0.207 1.0655 ( 0.043
0.091 ( 0.004 0.1801 ( 0.054 0.1715 ( 0.019
The Charge Clamp. We analyzed the average distance
between the CR atoms of the residues K301 and E471, the
two residues constituting the so-called “charge clamp”,
during the 15 ns of the simulation. The correct positioning
of these two charged residues is very important for the
recruitment of the coactivator, allowing the right orientation
of the two-turn helix dipole of the coactivator in the
hydrophobic cleft of the nuclear receptor. We found
that the average distance between the two residues is 19.0
TIF2 (NR box 2)
N-CoR1 (NR box 1)
0.7996 ( 0.057 0.9132 ( 0.191
>100 >100
1.736 ( 0.414
0.8134 ( 0.073
a Values are mean ( SE.
this case, the amount of N-CoR1 associated to PPARγ is
comparable to that of the unliganded receptor. We did not
detect any interaction between PPARγ and the corepressor
SMRT in FRET assay in all tested conditions (data not
shown). We finally tested two alternative peptides for the
coregulators CBP (NR box 4) and RIP140 (NR box 7). CBP
(NR box 4) does not show association (data not shown),
whereas RIP140 (NR box 7) can interact with PPARγ
(Figure 4 of Supporting Information) although the EC50 in
this case is higher, indicating that this region of RIP140
poorly interacts with the receptor.
Recruitment of N-CoR1 on the Promoters of PPARγ
Target Genes. To assess the relevance of the association of
N-CoR1 to PPARγ in the presence of S-1, we next per-
formed ChIP experiments examining the corepressor occu-
pancy on the promoter of the lipoprotein lipase (LPL) and of
the scavenger receptor CD36 genes, two known PPARγ
targets in RAW264.7 cells. The results show that N-CoR1
is recruited on the PPRE of both LPL and CD36 promoters
in control cells and, at a greater extent, in the presence of the
PPARγ antagonist 7 (GW9662, Figure 5).29 More interest-
ingly, the corepressor is still associated to the promoters in
the presence of S-1, whereas the treatments with rosiglita-
zone or R-1 significantly reduce the recruitment of N-CoR1,
as expected. These observations further indicate that the
partial agonist behavior of S-1 could be due to the particular
conformation adopted by PPARγ that does not allow the
release of the corepressor at the extent observed with full
agonists. The specificity of N-CoR1 recruitment on target
˚
˚
A in the apo-form, 17.7 A in the R-1 complex, and only 17.3
˚
A in the S-1 complex. The standard errors are below
˚
0.01 A.
Interaction of PPARγ with Coregulators in the Presence of
Rosiglitazone, R-1 and S-1. We next analyzed the functional
consequences of R-1 and S-1 binding on the capacity of
PPARγ to associate with coregulators. For this purpose, we
performed FRET assays with a panel of biotinylated pep-
tides for the coregulators SRC-1, PGC-1R, RIP140, CBP,
TIF-2, and N-CoR1 in the presence of saturating concentra-
tions of the ligands and increasing amounts of the peptides.
Under these conditions, we could evaluate the affinity of
PPARγ, activated by R-1, S-1, or rosiglitazone, for each
coregulator (Table 2). As shown in Figure 4, in the presence
of the ligands all the tested coactivators interact with PPARγ
with similar EC50, except CBP (NR box 1) whose affinity is
higher with rosiglitazone (Table 2). This might explain, at
least in part, the higher potency of rosiglitazone as compared
to that of R-1 and S-1 observed in cell-based transcription
assays.23 Notably, the most striking difference is the associa-
tion of the corepressor N-CoR1 in the presence of S-1 as
opposed to rosiglitazone and R-1, providing an explanation
to our previous observations whereby S-1 is a partial agonist,
whereas the R enantiomer is a full agonist of PPARγ.23 In

4358 Journal of Medicinal Chemistry, 2010, Vol. 53, No. 11
Pochetti et al.
Figure 4. Coregulator interaction with PPARγ by FRET assays. The association of six different coregulators SRC-1, PGC-1R, RIP140, CBP,
TIF-2, and N-CoR1 to PPARγ in the presence of 10 μM rosiglitazone, 25 μM R-1 and 50 μM S-1 was studied by FRET assays, using increasing
concentrations of the indicated biotinylated peptides. The curves are expressed as the ratio of fluorescence reading at 665 and 615 nm multiplied
by 10000 and represent the mean ( SE of independent curves.
expression vector with a panel of different coregulators (PGC-
1R, CBP, RIP140, and N-CoR1) along with the luciferase
reporter gene under the control of Gal4 binding sites. As
shown in Figure 6, PGC-1R enhances both the basal and the
induced transcription activity of Gal4-PPARγ in the presence
of all tested ligands about 3-fold as compared to cells with no
overexpressed coregulator (compare activity of wt constructs
in panel A vs B of Figure 6). The overexpression of CBP gives a
similar pattern (compare activity of wt constructs in panel A vs
C of Figure 6). When RIP140 was overexpressed, the basal
activity was not affected, however the stimulation of all ligands
was blunted (compare activity of wt constructs in panel A vs D
of Figure 6), suggesting that this coregulator behaves as a
repressor. Finally, the corepressor N-CoR1 leads to decreased
basal activity of Gal4-PPARγ, while in the presence of rosigli-
tazone and R-1 we observed a clear stimulation of the receptor
activity, but S-1 modestly affected the transcriptional activity
(compare activity of wt constructs in panel A vs E in Figure 6).
Mutagenesis of the Q286 Residue of PPARγ. To validate
the functional relevance of the residue Q286 that is involved
in the steric hindrance with S-1,23 we substituted this residue
Figure 5. N-CoR1 recruitment to PPARγ target gene promoters by
ChIP assay. RAW264.7 cells were treated with 5 μM PPARγ
antagonist 7 or 5 μM rosiglitazone or 5 μM S-1 or 5 μM R-1 for
16 h. ChIP assay was performed with antibody against N-CoR1 or
control IgG. Immunoprecipitated DNA was analyzed by real time
PCR using primers specific for the PPREs of CD36 and LPL
promoters or for a distal region as a negative control. Results are
expressed as mean ( standard deviations of triplicate samples.
Statistical significant differences were assessed by one-way ANOVA
with Tukey’s post test. (*), (**), (***) indicate P < 0.05, P < 0.01,
P < 0.001, respectively.
with amino acids bearing side chains of different length and
polarity. In particular, the following mutants have been
generated: Q286A, Q286D, Q286E, Q286K, and Q286M
(Figure 6). The substitution Q286A reduces the basal activity
of PPARγ and, to a lower extent, the activity in the presence
of rosiglitazone and R-1. Conversely, S-1 is ineffective on the
Q286A mutant, suggesting that with this amino acid sub-
stitution the S enantiomer does not induce a conformational
change of the receptor favorable for its activation. When
Q286 was mutated to an acidic amino acid, we obtained
different behaviors as compared to the wild type receptor.
The basal activity of Q286D mutant decreases, whereas that
of Q286E improves very strongly (4-fold with respect to that
of the wild type). The activity of the mutant Q286D is
virtually unaffected by the ligands. No ligand can also
increase the activity of the Q286E mutant, indicating that
PPRE was confirmed by evaluating its association with a
distal genomic region.
Role of Coregulators on PPARγ Activity. We next deter-
mined the functional relevance of the coregulator interaction
with PPARγ observed in FRET assays in a cellular context.
To this end, we cotransfected cells with Gal4-PPARγ-LBD

Article
Journal of Medicinal Chemistry, 2010, Vol. 53, No. 11 4359
Figure 6. Effect of coregulators on the transcriptional activity of PPARγ and analysis of point mutants. The effect of the overexpression of the
coregulators PGC-1R (B), CBP (C), RIP140 (D), and N-CoR1 (E) on the basal and ligand-stimulated transcription activity of PPARγ was
tested in cotransfection assays in HEK293 cells. Rosiglitazone, S-1, and R-1 were added at a final concentration of 5 μM in dimethyl sulfoxide
(DMSO). The effect of the same coregulators in the absence and in the presence of ligands was also tested on the transcriptional activity of the
indicated mutants of PPARγ. Results are expressed as percentage of wt in the absence of ligands, and each point is the mean ( SE of three
independent experiments each performed in triplicate wells. The fold activation of each mutant as compared with the vehicle (DMSO) is
indicated above the bars.
the introduction of a glutamate residue in this position may
lock the apo-receptor in the active conformation and the
addition of ligands cannot increase the transcription activity
any further. When a lysine residue substituted for Q286, the
basal receptor activity dropped dramatically and cannot be
restored by ligands. Finally, the substitution Q286M reduces the
basal activity and, to a lower extent, the activity in the presence
of rosiglitazone and R-1. Notably,S-1affects the transcriptional
activity of PPARγ Q286M to a lesser extent as opposed to
rosiglitazone and R-1. Interestingly, PGC-1R and CBP inc-
rease both the basal and ligand-dependent transcription acti-
vity of mutant Gal4-PPARγ, thus overcoming the defective
activity of the mutant forms of the receptor (Figure 6A,B,C). On
the other hand, RIP140 improves the basal activity of the
mutants to the level of that of the wt, with the exception of
Q286E mutant, which is reduced; instead, this coregulator
dampens the ligand-dependent transcription activity of the
mutated receptors (Figure 6A,D). The overexpression of
the corepressor N-CoR1 determines a general reduction of
the activities of all the mutated forms of Gal4-PPARγ in
the presence of all tested ligands (Figure 6A,E), except for
the mutant Q286A in the presence of rosiglitazone and R-1,
underlying the importance of the glutamine residue in
determining the behavior of S-1.
forms with respect to the apo-form, but unexpectedly, the S-1
ligand, which behaves as a partial agonist, seems to better
stabilize the PPARγ-LBD compared to the full agonist R-1
(8.7 nm2 vs 11.1 nm2). The analysis of the 2D projection of
trajectory on the two essential eigenvectors demonstrates that
the configurational space sampled by the S-1 complex during
the simulation lies on a different region from that of the apo-
form and the R-1 complexed form (Figure 3). The complex
with the full agonist R-1 lies in a restricted area of the
configurational space, which is inside the extended area of
the apo-form. This common region would correspond to a
population of active conformations able to recruit coactiva-
tors. The conformationally mobile apo-PPARγ-LBD could
also assume, in fact, active conformations, as evidenced by its
rather high basal activity.30 It is conceivable to suppose that
the different region covered by the complex with the partial
agonist S-1 represents a population of LBD conformations
more stable compared to R-1 but less productive for the
coactivator recruitment. In fact, the EC50 values relative to
the interaction of the coactivators SRC-1, PGC-1R, and TIF2
are slightly higher in the complex with S-1 compared to R-1
(Table 2), indicating that the affinity for these coactivators
decreases with the partial agonist. Looking at the RMSF
analysis, it is evident that S-1 better stabilizes H3 and H12
with respect to R-1. These results confirm the hypothesis,
already advanced by our previous X-ray studies on these two
enantiomers, that the partial agonist S-1 is not able to stabi-
lize H12 in the proper position to efficaciously recruit the
Discussion
The essential dynamics traces of PPARγ-LBD with and
without the ligands show a global stabilization of the liganded

4360 Journal of Medicinal Chemistry, 2010, Vol. 53, No. 11
Pochetti et al.
Table 3. CR Distances between K301 and E471 in the Ternary Com-
plexes of PPARγ with Ligands and SRC-1 Coactivator Fragmenta
˚
PDB code
A
1FM9
1K74
19.6
19.9
18.7, 19.6a
19.6, 19.6a
2PRG
1FM6
a These distances are referred to the two molecules of PPARγ in the
asymmetric unit.
coactivator and stabilizes instead a suboptimal conformation
of H12, resulting in an attenuated transcriptional response.
However, the most striking result observed in our FRET and
ChIP analyses is that the corepressor N-CoR1 remains asso-
ciated to PPARγ in the complex with S-1 at levels comparable
to that of samples with no added ligand, suggesting that the
conformation adopted by the receptor in the complex with S-1
does not allow the complete release of the corepressor, as
observed with the full agonists rosiglitazone and R-1. More-
over, as shown in the cotransfection assays, when N-CoR1 is
overexpressed, the treatment with S-1 fails to induce the
transcriptional activity of the wt and mutated Gal4-PPARγ.
Therefore, on the basis of these multiple approaches, we
propose that the retention of the corepressor bound to
PPARγ might be the main explanation for the partial agonist
behavior of S-1 as in the case of other partial agonists
described for the liver X receptor (LXR).31,32 Our data
indicate that RIP140 and N-CoR1 are corepressors of
PPARγ. However, the repressor activity of these two coregu-
lators seems to occur via two different modalities. On one
hand, N-CoR1 is constitutively associated to PPARγ in the
absence of agonists or even in the presence of the partial
agonist S-1. Full agonists like rosiglitazone and R-1 release
N-CoR1 and allow the full activation of the receptor, whereas
in the presence of the partial agonist S-1 N-CoR1 is still
associated to the receptor and prevents its full activation. On
the other hand, the association of RIP140 with PPARγ
increases in the presence of ligands and reduces the activity
of Gal4-PPARγ, thus suggesting a repressive role of this
coregulator on PPARγ as already reported by others.33 We
also confirmed the hypothesis advanced by H/D exchange
experiments18 whereby partial agonists differentially stabilize
other regions of the binding pocket. Particularly, the partial
agonists of the first class, spanning from H11 and H12 beyond
H3, show decreasing efficacy with an increasing stabilization
of H3. The compounds of this class preferentially stabilize H3
through H-bonds or closer hydrophobic contacts made with
residues of this helix, as in the case of S-1 with the Q286 residue
of H3 in PPARγ. This differential stabilization of regions
away from H12 suggests that these regions could be involved
in the interaction with the coactivators, and this interaction
could be mediated by the presence of partial agonists. For
example, the CBP coactivator utilizes a more extensive motif
than the canonical LxxLL in interacting with the LBD of
nuclear receptors, suggesting a binding mode governed by
LxxLL motif but modulated by flanking residues and possibly
fine-tuned by additional interactions with partial agonists.34
In addition, the analysis of average distances between the
two residues of the charge clamp shows a deviation, in the case
Figure 7. PPARγ coactivator binding surface. The two residues
forming the charge clamp are shown in white. The SRC-1 coacti-
vator motif is shown in purple.
˚
between 18 and 20 A in the presence of a coactivator and an
agonistligand (Table 3). Inthe complex withR-1, thisdistance
˚
is almost within this range (17.7 A). The hydrophobic cleft for
the recruitmentof the coactivator is formed by helices3, 4, and
12 (Figure 7). In the complex with S-1, the observed short
average distance between the residues involved in the charge
clamp is probably due to the lower fluctuation of these three
helices upon the S-1 binding, representing additional evidence
corroborating the hypothesis of a receptor adopting a non-
optimal conformation for the binding of coactivators. There-
fore, it is conceivable to suppose that partial agonists like S-1
preferentially select suboptimal less active conformations of
the PPARγ-LBD among those possible existing in solution,
resulting in a slightly decreased affinity for certain coactiva-
tors and do not allow the complete dissociation of corepres-
sors from the receptor.
Could S-1 Be a Selective PPARγ Modulator? The analysis
of the EC50 values of coregulator recruitment (Table 2)
shows that rosiglitazone and R-1 have typical profiles of full
agonists and that R-1 behaves in a very similar way to
rosiglitazone, recruiting all the tested coactivators almost
with the same affinity. Yet, a difference in the EC50 value of
CBP is observed with R-1 versus rosiglitazone. It is possible
that the higher EC50 for this coactivator may explain the
lower potency of this enantiomer as compared to rosiglita-
zone. The partial agonist S-1 shows a more interesting profile
of interaction with coregulators. With this ligand, the EC50
values of all coactivators are higher, in particular that of CBP
and TIF-2, with respect to that obtained with rosiglitazone.
Most strikingly, however, S-1 is able, unlike the other two
ligands, to recruit the corepressor N-CoR1, providing a
functional explanation to the partial agonist behavior of this
ligand. As demonstrated in the only known crystal complex
with a corepressor, the ternary complex PPARR-SMRT-627
(Figure 8), when the corepressor is bound to PPAR-LBD its
three-turn R-helix extends into the space normally occupied
bytheactive conformation of H12. Thishelix, displaced by the
antagonist, is repositioned toward H3 and it is loosely packed
against it. In this way, the corepressor is bound into the
hydrophobic cleft formed by H3, H4, and H12, with the last
one forming a different edge of the hydrophobic cleft com-
pared to the complex with the coactivator. The behavior of
S-1 suggests that this ligand is able to stabilize the H12 for
recruiting the coactivators, although in a suboptimal way.
This is evidenced in the crystal structure of PPARγ-S-123 and
˚
of the complex with the partial agonist S-1 (17.3 A), from the
distance normally observed in the structures of the ternary
complexes of PPARγ with the ligand and coactivator motif.35
It is well-known that in PPARs, as well as in other NRs, the
distance between the CR atoms of these two residues ranges

Article
Journal of Medicinal Chemistry, 2010, Vol. 53, No. 11 4361
Figure 9. Main interactions of Q286 into the apo-form of wt
PPARγ (A) and of E286 into the mutant Q286E (B).
The Key Role of Q286. In Figure 8, it is also possible to
notice the key role of the residue Q277 of PPARR (equivalent
to Q286 in PPARγ) in stabilizing H12 in its inactive con-
formation. On the other hand, as previously reported,23 this
residue was demonstrated to play a determinant role even in
the stabilization of the active conformation of H12. These
observations strengthened further our project to carry out a
site-directed mutagenesis of Q286 in PPARγ. The mutation
analysis made on this residue clearly evidence its key position
in H3 for the stabilization of H12, in some cases also in the
absence of the ligand. As can be seen in Figure 9A, in the apo-
form of wild type PPARγ, the residue Q286 is involved in
two H-bonds: the former between the CO group and a crystal
water molecule and the latter between the NH2 group and the
CO of S464 belonging to the loop 11/12. Upon the binding of
Figure 8. The ternary complex of PPARR-SMRT-6 (PDB code
1KKQ). The ligand is shown in yellow, the corepressor SMRT in
pink, H12 in red, H4 in orange, and H3 in white; the residue Q277 is
also shown in white.
rosiglitazone36 or R-123 and, to a lesser extent, of S-1,23
a
in our MD simulations which show a nonoptimal distance
between the residues forming the charge clamp. At the same
time, S-1 would be also able to stabilize H3 and H4, as
demonstrated by the MD simulations, in such a way to allow,
alternatively, the accommodation of the corepressor upon
rearrangement of H12 against H3. As shown in Figure 8, in
fact, the corepressor is held in its position through several
interactions with H3 and H4, hence a better stabilization of
these helices by a partial agonist like S-1 turns out to be very
important for the recruitment of the corepressor and also for
the stabilization of H12 in its inactive conformation close to
H3. In this regard, we performed a further MD simulation,
with R-1 and S-1, starting from an inactive conforma-
tion of H12, similar to that assumed in the ternary complex
PPARR-SMRT-6. As shown in Figure 1 of Supporting
Information, also in this case H3 is less fluctuating in the
presence of S-1, indicating that its greater stability does not
depend on H12 conformation but only on the ligand. More-
over, the stabilization of H3 seems to favor the inactive
conformation of H12 and also of the loop 11/12, as indicated
by the lower RMSF of these regions when S-1 is present (see
Figure 1 of Supporting Information), helping in this way the
recruitment of corepressor or diminishing its release. All these
observations are corroborated by both FRET analysis with
coregulator peptides and ChIP assays, the latter providing a
snapshot of the actual association of the corepressor N-CoR1
with PPARγ in the presence of the partial agonist S-1.
The behavior of S-1 at molecular level, therefore, seems to
account for its PPARγ partial agonism and makes it a good
candidate as a selective modulator, that is, a molecule able to
function as a full agonist in some tissues where a sufficient
amount of coactivators is available and as a partial agonist
or antagonist in specific tissues with different ratio of
coactivator/corepressor proteins.
network of strong interactions can be realized among the
ligand, H3, and the loop 11/12. These interactions stabilize
the active conformation of H12 better than in the apo-form.
In this network, Q286 plays a crucial role through its
CONH2 group. In the PPARγ-rosiglitazone complex, the
NH2 group makes a H-bond with a CO of the ligand
thiazolidinedione ring and an amino-aromatic interaction
with F282, while the CO group engages a H-bond with a
crystal water, which bridges in turn the backbone NH of
H466 (loop 11/12). S-1 and R-1 complexes with PPARγ are
stabilized by two H-bonds, a stronger one between Q286
NH2 and S464 backbone CO and a weaker one between
Q286 CO and H466 backbone NH.
To seek structural explanations for the effects of Q286
substitutions in PPARγ, we used molecular modeling to
assess how these mutations would affect both receptor basal
activity and ligand binding. Using the Rotamer Explorer
function37 inside MOE,38 Q286 was computationally mu-
tated in the ligand-free PPARγ (PDB code 1PRG)36 and in
rosiglitazone (PDB code 2PRG),36 S-1 (PDB code 2I4P),23
and R-1-bound (PDB code 2I4J)23 structures, and the most
favorable rotamers of each residue were examined. These
rotamers were determined based on the lowest ΔG. ΔG is the
sum of the change in free energy of solvation between the
states where protein and rotamer are separated and when
they form a complex plus the torsional strain of the rotamer.
Figure 6 shows the effect of the mutation of Q286 on the
transcriptional activity. The basal activity of all the mutants,
with the only exception of Q286E, is reduced. The basal
activity of the mutated forms of PPARγ can be restored only
if supraphysiological levels of coactivators are present in the
cells, as in the case of cotransfected PGC-1R and CBP.
Intriguingly, we observed that RIP140 reduces the basal
activity of the Q286E mutant, which is otherwise strongly

4362 Journal of Medicinal Chemistry, 2010, Vol. 53, No. 11
Pochetti et al.
enhanced, thus further supporting its role as a corepressor.
The reduced basal activity of the mutants could be explained
by the loss of the H-bond made by Q286 with the carbonyl
oxygen of S464. None of the mutants has functional groups
able to make this H-bond, with the exception of Q286D and
Q286E. Nevertheless, D has a shorter side-chain with respect
to Q and cannot form H-bonds with the backbone of the loop
11/12. On the contrary, the side-chain of E, equal in size to Q,
is able to form a charge-reinforced H-bond with the back-
bone NH of H466 (Figure 9B). This would account for a
strong interaction between H3 and the loop 11/12, up to the
beginning of H12, able to stabilize this helix also in the
absence of a ligand. This network of interactions would be
efficient to such an extent that the binding of a ligand would
not affect at all the activity. Looking at activities in the
presence of the ligands, it is evident that S-1 does not affect
the mutants activities but only that of the wild type receptor
(6.4-fold the basal activity) and, to a lesser extent, that of
Q286M (3.1-fold). On the contrary, R-1 and rosiglitazone
improve the activity of the hydrophobic mutants Q286A and
Q286M possessing a terminal methylgroup whichcan interact
with the ligands through vdW contacts, partially restoring the
activity of the wild type receptor (Figures 5A-B and 6A-B of
Supporting Information). The activity of the mutants Q286K
and Q286D is impaired also in the presence of the ligands.
Particularly, the long and positively charged side-chain of
lysine adopts a low-energy trans-gauche conformation which
prevents any local interaction with the protein, thus destabi-
lizing the loop 11/12 (Figure 7 of Supporting Information).
to improved basal activity. The strategic position of Q286 in
H3 is also highlighted by the fact that only the two potent
full agonists rosiglitazone and R-1 are still able to activate
the receptor after the mutation of this residue, whereas the
less potent partial agonist S-1 turns out to be practically
ineffective.
Experimental Section
Chemical Methods. Column chromatography was performed
˚
on ICN silica gel 60 A (63-200 μm) as a stationary phase.
Melting points were determined in open capillaries on a
Gallenkamp electrothermal apparatus and are uncorrected.
Mass spectra were recorded with a HP GC/MS 6890-5973
MSD spectrometer, electron impact 70 eV, equipped with HP
chemstation. For GC/MS analysis of acid analytes, the corres-
ponding methyl esters, obtained by reaction with a solution of
diazomethane in Et2O, were used. 1H NMR spectra were
recorded in CDCl3 on a Varian-Mercury 300 (300 MHz)
spectrometer at room temperature (20 ꢀC). Chemical shifts are
expressed as parts per million (δ). Purity of all tested compounds
was >95%, as confirmed by combustion analysis carried out
with an Eurovector Euro EA 3000 model analyzer. Optical
rotations were measured with a Perkin-Elmer 341 polarimeter
at room temperature (20 ꢀC): concentrations are expressed as
g(100 mL)-1. The enantiomeric excesses of the final acids were
determined by HPLC analysis on Chiralcel OD column (4.6 mm
i.d. ꢀ 250 mm, Daicel Chemical Industries, Ltd., Tokyo, Japan).
Analytical liquid chromatography was performed on a PE
chromatograph equipped with a Rheodyne 7725i model injec-
tor, a 785A model UV/vis detector, a series 200 model pump,
and a NCI 900 model interface. Chemicals were obtained from
Aldrich (Milan, Italy), Lancaster (Milan, Italy), or Acros
(Milan, Italy) and were used without any further purification.
Preparation of (þ)-R- or (-)-S-2-(4-Bromo-phenoxy)-2-meth-
yl Butanoic Acid (R-2 and S-2). A solution of KOH (29.21 g,
52.11 mmol) in H2O (75 mL) was added dropwise to a solution
of 4-bromo-phenol (15.1 g, 86.78 mmol) in 2-butanone (130 mL).
After 0.5 h at room temperature, a first amount of CHBr3
(5.5 mL) was added dropwise, during 1 h, to the reaction mixture.
After 1 h at room temperature, a second addition of CHBr3
(11.5 mL) was carried out during 1.5 h. The resulting mixture was
stirred overnight at room temperature, after which the organic
solvent was distilled off. The aqueous phase was carefully acid-
ified with 6 N HCl and extracted with ethyl acetate. The collected
organicphase was washedwithbrine, driedover Na2SO4, filtered,
and evaporated to dryness affording a brown oily residue, which
was dissolved in ethyl acetate and extracted five times with
NaHCO3 saturated solution. The aqueous phase was carefully
acidified with 6 N HCl and extracted four times with Et2O. The
collected organic layer was dried over Na2SO4, filtered, and
evaporated to dryness to give a dark-red oily residue, which
was chromatographed on a silica gel column (petroleum ether/
ethyl acetate/MeOH 7:2:1 as eluent), affording the desired acid 2
as a pale-yellow solid in 75% yield.
Conclusions
The behavior of the ligand S-1 allows one to suppose that
some PPARγ partial agonists would be able to select popula-
tions of the LBD with nonoptimal distances between the
residues of the charge clamp. This could allow the stabiliza-
tion of suboptimal active conformations of H12 which would
be still able to recruit the coactivator, even though with
reduced affinity. At the same time, the suboptimal conforma-
tion of H12 couldfavor its repositioning toward H3 assuming,
in this way, a nonactive conformation able to recruit the
corepressor, as observed in the ternary crystal complex
PPARR-SMRT-6. This recruitment could be further en-
hanced by the better stabilization of H3 and H4 induced by S-
1. In other words, in the presence of partial agonists like S-1,
H12 could function as a switch; its active conformation,
although suboptimal, would allow the recruitment of coacti-
vators, whereas its inactive conformation against H3 could
favor the binding of corepressors.
This behavior, moreover, seems to be characteristic of
PPARγand not of PPARR. Therefore, the capability to induce
both the recruitment of coactivators and corepressors makes
S-1 a good candidate as a selective PPARγ modulator, namely
a ligand that, depending on the balance between coactivators
and corepressor in a given tissue, could behave as an agonist or
antagonist. In fact, the cotransfection experiments confirm
that when N-CoR1 is overexpressed, S-1 behaves almost like
an antagonist, whereas R-1 and rosiglitazone can activate the
receptor activity, being a full agonist of PPARγ.
Both enantiomers of 2 were obtained by fractional crystal-
lization of the diastereomeric salts with the chiral resolving
agent S- or R-1-phenylethylamine. For this purpose, a solution
of R-1-phenylethylamine (10.6 g; 87 mmol) in absolute EtOH
(15 mL) was added to a solution of the racemic acid (21.6 g; 79.1
mmol) in absolute EtOH (30 mL). The resulting mixture was
stirred for 1 h at 40 ꢀC. The EtOH was distilled off, affording a
white solid mixture of the two diastereomeric salts, which was
recrystallized from ethyl acetate until the [R]D had reached a
constant value ([R]D = -8.3, c 2, MeOH).
The optically pure diastereomeric salt, obtained after six
recrystallizations, was suspended in Et2O (20 mL) and 6 N
HCl (15 mL) was added to the suspension. The resulting mixture
was stirred for 0.5 h at room temperature, and the organic layer
was separated, washed with brine, and dried over Na2SO4. The
By site-directed mutagenesis, we also confirmed, as pre-
viously reported, the crucial role of the residue Q286 in the
stabilization of H12 of PPARγ. In addition, the mutation
analysis showed that one mutant of this residue is able to
directlystabilize H12 alsointhe absence ofa ligand, givingrise

Article
Journal of Medicinal Chemistry, 2010, Vol. 53, No. 11 4363
organic solvent was evaporated to dryness, affording the title
acid (-)-S-2 (1.51 g; 5.35 mmol) as a colorless oil. (þ)-R-2 was
obtained following the procedure described above using S-1-
phenylethylamine as a chiral resolving agent.
Preparation of (þ)-R- or (-)-S-Methyl 2-(4-Bromo-phenoxy)-
2-methylbutanoate (Step c, Scheme 1). A solution of acid R- or
S-2 (10 mmol) in MeOH (40 mL) and two drops of conc H2SO4
was stirred for 3 h at reflux, and then the solvent was distilled off
and the residue was dissolved in ethyl acetate. The resulting
solution was washed with NaHCO3 saturated solution and
brine, and then the organic phase was dried over Na2SO4 and
the solvent was removed under reduced pressure to give the title
compounds as pale-yellow oils.
Preparation of R- or S-Methyl 2-[4-(2-Phthalimido-2-yl-ethen)-
phenoxy]-2-methylbutanoate (R-3, S-3). A solution of R- or S-methyl
2-(4-bromo-phenoxy)-2-methylbutanoate (6.19 mmol), tri-o-tolyl-
phosphine (0.48 mmol), N-vinylphthalimide (6.19 mmol), and
N,N-diisopropylethylamine (9.27 mmol) in anhydrous CH3CN
(12 mL) was added, under N2 atmosphere, to a suspension of
Pd(AcO)2 (0.16 mmol) in the same anhydrous solvent (2 mL).
The reaction mixture was stirred for 24 h at reflux, and then the
organic solvent was evaporated in vacuo and CH2Cl2 (20 mL)
was added to the residue. The precipitate was filtered off through
a celite pad, washed four times with CH2Cl2 (20 mL), and the
filtrate was washed with brine and dried over Na2SO4. The
solvent was evaporated to dryness affording a yellow solid
residue, which was chromatographed on a silica gel column
(petroleum ether/ethyl acetate 9:1 as eluent) to give the title
compound as a yellow solid, which was used in the next step
without any further purification.
Preparation of R- or S-Methyl 2-[4-(2-Phthalimido-2-yl-
ethyl)phenoxy]-2-methylbutanoate (Step e, Scheme 1). A solution
of R-3 or S-3 (5.82 mmol) in THF (35 mL) was added to a stirred
suspension of Wilkinson catalyst (350 mg, 0.38 mmol) in abs
EtOH (5 mL). The resulting mixture was stirred at room
temperature under H2 atmosphere (4 atm) for 5 h. The suspen-
sion was filtered through a celite pad to remove the catalyst, and
the solvent was evaporated to dryness, providing a dark solid
residue which was chromatographed on a silica gel column
(petroleum ether/ethyl acetate 7:3 as eluent), affording the
desired compound as a yellow solid.
overnight at room temperature and then was carefully added with
MeOH (15 mL) and stirred for 0.5 h at reflux. The organic solvent
was evaporated to give an oily residue, which was dissolved in THF
(15 mL) and cooled to 0 ꢀC. N(Et)3 (0.2 mL) and a solution of
2-chlorobenzoxazole (2.20 mmol) in anhydrous THF (5 mL) were
added to the so-obtained solution. The resulting reaction mixture
was stirred for 0.5 h at 0 ꢀC, 0.5 h at room temperature, 2 h at reflux,
and 15 h at room temperature. The organic solvent was evaporated
to dryness, affording a solid residue which was chromatographed
on a silica gel column (petroleum ether/ethyl acetate 9:1 as eluent) to
give the title compound as a yellow oil.
Preparation of the Final Acids R-1 and S-1. A solution of the
corresponding methyl ester (0.86 mmol), obtained from the
previous steps, in EtOH (10 mL) and 1 N NaOH (5 mL) was
stirred for 4 h at room temperature. The organic solvent was
removed under reduced pressure and the aqueous phase was
acidified with 3 N HCl and extracted with CH2Cl2. The organic
layer was dried over Na2SO4, filtered, and evaporated to dry-
ness, affording an oily residue which was chromatographed on a
silica gel column (petroleum ether/ethyl acetate 8:2 and then
petroleum ether/ethyl acetate/MeOH 7.5:2:0.5 as eluents) to
give the title compound as a white solid, which was recrystallized
from Et2O/n-hexane.
Preparation of R-2-(4-Nitro-phenoxy)-2-methylbutanoic Acid
(5). R-2-Hydroxy-2-methylbutanoic acid was obtained by frac-
tional crystallization of its diastereomeric salts with brucine,
following a procedure reported in the literature.25,26 Afterward,
a solution of R-2-hydroxy-2-methylbutanoic acid (1.79 g, 15.21
mmol) in anhydrous DMF (10 mL) was added, under N2
atmosphere, to an ice-bath cooled suspension of 95% NaH
powder (1.10 g, 45.6 mmol) in the same anhydrous solvent
(20 mL). After 0.5 h at 0 ꢀC, a solution of 4-nitro-fluoro-benzene
(2.14 g, 15.2 mmol) in anhydrous DMF (15 mL) was added to
the reaction mixture, which was stirred overnight at room
temperature. Ethyl acetate (20 mL) was added to quench the
reaction mixture, and then the organic solvent was distilled off
and 6 N HCl was added to the residue and extracted with ethyl
acetate. The resulting solution was washed with NH4Cl satu-
rated solution and brine, and then the organic phase was dried
over Na2SO4 and the solvent was removed under reduced
pressure to give the title compound as a red oil in 85% yield.
GC/MS of the methyl ester, m/z (%): 253 (8) [Mþ], 115 (100).
This acid was used in the next step without any further purification.
Preparation of (þ)-R-2-(4-Bromo-phenoxy)-2-methylbutanoic
Acid (R-2, Steps b and c, Scheme 2). A solution of acid 5 (3.09 g,
12.21 mmol) in MeOH (160 mL) was added to a stirred suspen-
sion of 5% Pd (0.1 g) in MeOH (5 mL). The resulting mixture
was stirred at room temperature under H2 atmosphere (4 atm)
for 4 h. The suspension was filtered through a celite pad to
remove the catalyst and the solvent was evaporated to dryness,
affording a dark solid residue, which was readily used in the
following step. This intermediate was dissolved in H2O (15 mL)
and 48% HBr (15 mL), and the resulting mixture was cooled to 0
ꢀC and added with a solution of NaNO2 (0.6 g, 8.70 mmol) in
H2O (15 mL).
The reaction mixture was stirred for 0.5 h at 0 ꢀC and then
added with a solution of CuBr (3.42 g, 23.84 mmol) in H2O
(30 mL) and 48% HBr (25 mL). The resulting mixture was
heated for 1.5 h at 80 ꢀC and stirred overnight at room
temperature; afterward, a 50% KOH solution (15 mL) was
carefully added up to pH 10. The suspension was filtered
through a celite pad, and the aqueous filtrate was washed with
CH2Cl2, acidified with 6 N HCl and extracted with CHCl3. The
collected organic layers were dried over Na2SO4, and the solvent
was removed under reduced pressure to give the title compound
as a colorless oil in 70% yield. ([R]D = þ17, c 1, MeOH).
Coregulator Interaction Assays by Fluorescence Resonance
Energy Transfer (FRET) Assay. The ligand binding domain
(LBD) of human PPARγ was expressed as N-terminal His-
tagged protein using a pET28a vector and purified as previously
Preparation of R- or S-Methyl 2-[4-(2-Amino-ethyl)phenoxy]-
2-methylbutanoate (Step f, Scheme 1). N2H4 xH2O (31.50 mmol)
3
was added to a solution of R- or S-methyl 2-[4-(2-phthalimido-2-
yl-ethyl)phenoxy]-2-methylbutanoate (5.25 mmol) in absolute
EtOH (40 mL). The reaction mixture was stirred for 1 h at reflux
and overnight at room temperature. The suspension was filtered,
and the organic solvent was evaporated in vacuo to give a yellow
solid which was dissolved in ethyl acetate. The solution was
washed with brine, dried over Na2SO4, and the organic solvent
was evaporated to dryness, affording the desired compound in
quantitative yield as a yellow oil. The resulting amine was used in
the next step without any further purification.
Preparation of (þ)-R- or (-)-S-Methyl 2-[4-(2-Heptanoylamino-
ethyl)phenoxy]-2-methylbutanoate (R-4, S-4). Heptanoic acid (7.51
mmol), HOBT xH2O (2.48 mmol), and DIC (10.02 mmol) were
3
added to
a
solution of R- or S-methyl 2-[4-(2-amino-
ethyl)phenoxy]-2-methylbutanoate (5.01 mmol) in CH2Cl2 (40 mL).
The reaction mixture was stirred for 15 h at room temperature. The
organic phase was washed with NaHCO3 saturated solution, 1 N
HCl, and brine and then dried over Na2SO4 and filtered. The
solvent was evaporated to dryness, affording a yellow oily residue
which was chromatographed on a silica gel column (petroleum
ether/ethyl acetate from 6:4 to 4:6 as eluents) to give the desired
compound as a yellow oil.
Preparation of R- or S-Methyl 2-{4-[2-(N-heptyl-N-(benzoxazol-
2-yl)amino-ethyl)]phenoxy}-2-methylbutanoate (Steps h and i,
Scheme 1). One M BH3 in THF solution (14 mmol) was added,
under N2 atmosphere, to a stirred solution of R- or S-4 (1.68 mmol)
in anhydrous THF (15 mL). The reaction mixture was stirred

4364 Journal of Medicinal Chemistry, 2010, Vol. 53, No. 11
Pochetti et al.
described.23 In brief, freshly transformed Escherichia coli BL21
DE3 were grown in LB medium with 30 μg of kanamycin ꢀ
mL-1 at 37 ꢀC to an OD of 0.6. The culture was then induced
with 0.1 mM isopropyl-β-D-thiogalactopyranoside and further
incubated at 18 ꢀC for 20 h. Cells were harvested and resus-
pended in a 20 mL ꢀ L-1 culture of buffer A (20 mM Tris, 150 mM
NaCl, 10% glycerol, 1 mM Tris 2-carboxyethylphosphine
HCl, pH 8) in the presence of protease inhibitors (Complete
Mini EDTA-free; Roche Applied Science). Cells were sonicated,
and the soluble fraction was isolated by centrifugation (35000g
for 45 min). The supernatant was loaded onto a Ni2þ-nitrilo-
triacetic acid column (GE Healthcare) and eluted with a gra-
dient of imidazole 0-300 mM in buffer A with a PU980 HPLC
system (Jasco, Lecco, Italy). The fractions containing the pro-
tein were collected, quantitated with a Bradford assay, and
analyzed on 12% SDS-PAGE. The protein was then dialyzed
over buffer C (20 mM Tris, 20 mM NaCl, 10% glycerol, 1 mM
Tris 2-carboxyethylphosphine HCl, pH 8) to remove imidazole.
The identity of the His-tagged protein was determined on the
basis of the molecular weight and of the sequencing of the tryptic
peptides obtained by liquid chromatography-electrospray io-
nization mass spectrometry and tandem mass spectrometry
(data not shown) (LTQ; ThermoElectron Co., San Jose, CA),
respectively. The protein was then loaded onto a Q-Sepharose
HP column (GE Healthcare) and eluted with a gradient of NaCl
0-500 mM in buffer B (20 mM Tris, 10% glycerol, 1 mM Tris
2-carboxyethylphosphine HCl, pH 8). The protein was then
dialyzed over buffer B and kept frozen in aliquots at a concen-
in triplicate. Sixteen hours after the ligand addition, cells were
fixed by adding 1 mL of fixation buffer (500 mM HEPES/KOH,
pH 7.9, 100 mM NaCl, 1 mM EDTA, 0.5 mM EGTA, 11%
formaldehyde) for 10 min and then washed twice with PBS. Cells
were scraped and pelleted for 5 min at 1000g at 4 ꢀC. Cells were
subsequently lysed in 300 μL of lysis buffer (50 mM Tris, pH8,
5 mM EDTA, pH8, 1% SDS, Complete Mini, Roche). The lysates
were homogenized by sonicating on ice with five pulses of 10 s at a
power setting of 10 (SONOPULS Bandelin, Berlin Germany)
sonicator. Samples were then diluted by adding 1.7 mL of dilution
buffer (20 mM TrisHCl, pH8, 100 mM NaCl, 2 mM EDTA, pH8,
0.5% Triton X-100, Complete Mini, Roche). At this point, 200 μL
per samples were kept as input. The lysates were precleared by
adding 50 μL of Protein-G Sepharose (Amersham) with salmon
sperm DNA and incubated 2 h at 4 ꢀC on a rotating shaker.
Supernatants were collected and each sample was split in three
parts: in the first, we added 5 μg of antibody recognizing N-CoR1
(sc-8994 Santa Cruz Biotechnology, CA), the second one was
incubated with 5 μg of general IgG, and the last one with no
antibody. All the samples were rotated overnight at 4 ꢀC. Further,
the complexes were recovered by adding 50 μL of Protein-G
Sepharose with salmon sperm DNA by rotating them for 1 h at
4 ꢀC. The immunoprecipitated DNA was then washed and eluted
twice with 100 μL of elution buffer incubating 20 and 10 min,
respectively, and shaking the beads every 5 min. The cross-linking
was reversed incubating samples at 65 ꢀC for at least 6 h. Samples
were incubated with 1 μL of proteinase K for 1 h at 55 ꢀC. The
DNAs were cleaned by Qiaquick columns (Qiagen) and eluted in
50 μL of H2O. DNA samples were analyzed by real time qPCR
using 1 μL per reaction following the SYBR green protocol. The
primers used are shown in Table 1 of Supporting Information and
were those already described by Lefterova et al. and Araki et al.39,40
Transfection Assays and Mutation Analysis. The expression
plasmids for PPARγ-LBD fused in frame with the Gal4 DNA
binding domain (pGal4-PPARγ-LBD) and the reporter vec-
tor containing five copies of the Gal4 upstream activating
sequences (pGal4UAS-luciferase) driving the transcription of
the luciferase reporter gene were kindly donated by Dr. Krister
tration of 1 mg mL-1
.
3
Fluorescence resonance energy transfer assays to evaluate
peptide recruitment were performed in 384-well plates in a final
volume of 10 μL. A mix of 8 ng of human PPARγ-LBD, 0.8 ng
of europium-labeled anti-His antibody (Perkin-Elmer, Monza,
Italia), and 86 ng of allophycocyanin-labeled streptavidin
(Perkin-Elmer), in a FRET buffer containing 50 mM Tris pH
7.5, 50 mM KCl, 1 mM DTT, and 0.1% free fatty acids BSA was
prepared. The tested ligands were added to the mix at the
concentration ensuring the saturation of the receptor. The
biotinylated peptides (PRIMM, Milano, Italia) were added in
12-point dose response curves starting at 9 μM as the highest
concentration. The reactions were equilibrated for 1 h at room
temperature and then measured in an Envision multiplate
reader (Perkin-Elmer) using 340 nm as excitation and 615 and
665 nm as emission wavelengths. The ratio between 665 (APC
signal) and 615 (europium signal) was used to evaluate the
peptide recruitment on the receptor.
€
Bamberg (AstraZeneca, Molndal, Sweden). The expression
vectors for PGC-1R, RIP140, CBP, and N-CoR1 were kindly
provided by Anastasia Kralli (The Scripps Research Institute,
La Jolla, CA), Iannis Talianidis (Biomedical Sciences Research
Center “Alexander Fleming”, Vari, Greece), and Serena Ghi-
sletti (Istituto FIRC di Oncologia Molecolare, Milano, Italia),
respectively. The mutants Q286A, Q286D, Q286E, Q286K, and
Q286M were prepared using the QuikChange site-directed
mutagenesis kit (Stratagene, La Jolla, CA) and confirmed by
automated sequencing. pCMV-GFP, containing the green
fluorescent protein gene driven by the cytomegalovirus early
promoter/enhancer, was used as an internal standard to normal-
ize for transfection efficiency across wells. For transactivation
assays, HEK293T cells (American Type Culture Collection,
Manassas, VA) were transfected in suspension with 10 ng of
pUAS5XGal4-Luc, 10 ng of receptor expressing vector, 10 ng of
coregulator expression vector, 10 ng of pCMV-GFP, and 160 ng
of pCDNA3 for a total of 200 ng of DNA per well, using Fugene
6 (Roche, Milano, Italia) as a transfection reagent.41 Trans-
fected cells (4 ꢀ 104) were then seeded in 96-well plates. Four
hours after transfection, cells were treated with 5 μM of the
indicated compounds in Dulbecco’s Modified Eagle Medium,
(Invitrogen, Milano, Italy) containing 100 U penicillin G ꢀ
mL-1, 100 μg of streptomycin sulfate ꢀ mL-1, and 10%
charcoal stripped fetal bovine serum (Euroclone, Milano,
Italia). Green fluorescence protein intensity was assessed by
bottom reading of the plated living cells using a plate reader
(Envision, Perkin-Elmer). Then luciferase activity in cell ex-
tracts was determined after 20 h with a luciferase detection kit
(Perkin-Elmer, Monza, Italia) using a plate reader luminometer
(Envision, Perkin-Elmer).37 Transfection experiments were per-
formed in triplicates.
The peptide sequences used are Biotin-CPSSHSSLTERH-
KILHRLLQEGSPS-COOH (NR box 2) for SRC-1 spanning
from amino acid 676 to 700 (reference number NP_671766),
Biotin-DGTPPPQEAEEPSLLKKLLLAPANT-COOH (NR
box 1) for PGC-1R spanning from amino acid 130 to 154 (reference
number NP_037393), Biotin-LERNNIKQAANNSLLLHLLKSQ-
TIP-COOH (NR box 4) spanning from amino acid 366 to 390, and
Biotin-PVSPQDFSFSKNGLLSRLLRQNQDSYL-COOH (NR
box 7) spanning from amino acid 805 to 831 for RIP140 (reference
number NP_003480), Biotin-SGNLVPDAASKHKQLSELLRG-
GSGS-COOH (NR box 1) spanning from amino acid 56 to 80 and
Biotin-SVQPPRSISPSALQDLLRTLKSP-COOH (NR box 4)
spanning from amino acid 2055 to 2078 for CBP (reference number
NP_004371), Biotin-GSTHGTSLKEKHKILHRLLQDSSSPVD-
COOH (NR box 2) for TIF2 spanning from amino acid 676 to 702
(reference number NP_006531), Biotin-SFADPASNLGLEDIIR-
KALMGSFDD-COOH (NR box 1) for N-CoR1 spanning from
amino acid 2253 to 2277 (reference number NP_006302), and Biotin-
APGVKGHQRVVTLAQHISEVITQ-COOH (NR box 1) for
SMRT spanning from amino acid 2123 to 2145 (reference
number NP_006303).
Chromatin Immunoprecipitation Assays. Murine macrophage
cells RAW264.7 were seeded in 10 cm plates and were treated
with the indicated compounds. The treatments were performed

Article
Computational Chemistry. Molecular modeling and graphics
Journal of Medicinal Chemistry, 2010, Vol. 53, No. 11 4365
istrative and financial management. This work was supported
by The Giovanni Armenise-Harvard Foundation Career
Development Grant (NM), by the CARIPLO Foundation
manipulations were performed using the molecular operating en-
vironment (MOE)38 and UCSF-CHIMERA software packages,42
running on a 2 CPU (PIV 2.0-3.0 GHZ) Linux workstation.
Energy minimizations were realized by employing the AMBER 9
program,43 selecting the Cornell et al. force field.44 Molecular
modeling calculations were based on the X-ray structure coordi-
nates of the complexes of the ligand-free PPARγ (PDB code
1PRG)36 and the rosiglitazone- (PDB code 2PRG),36 S-1- (PDB
code 2I4P)23 and R-1-bound (PDB code 2I4J)23 structure, depos-
ited in the Protein Data Bank (PDB). The mutations at position 286
were accomplished by the MOE Rotamer Explorer function. The
resulting mutants Q286E, Q286A, Q286K, Q286M, and Q286D
models were then solvated with water molecules in a truncated
octahedron periodic box. The distance between the box walls and
ꢀ
(FG), by grants from the Universita degli Studi di Milano
(MC) and Universita degli Studi di Bari “Aldo Moro” and by
ꢀ
a donation from the Francesco Del Gaudio family.
Supporting Information Available: Yields, physicochemical
properties, and spectroscopic data for intermediates and final
compounds; molecular dynamics, molecular modeling, and
biological data. This material is available free of charge via
the Internet at http://pubs.acs.org.
References
˚
the protein was set to 10.0 A, which resulted in about 13000 water
(1) Rosen, E. D.; Spiegelman, B. M. PPARgamma: a nuclear regulator
of metabolism, differentiation, and cell growth. J. Biol. Chem.
2001, 276, 37731–37734.
(2) Nuclear Receptors Nomenclature Committee. A unified nomen-
clature system for the nuclear receptor superfamily. Cell 1999, 97,
161-163.
(3) Evans, R. M.; Barish, G. D.; Wang, Y. X. PPARs and the complex
journey to obesity. Nat. Med. 2004, 10, 355–361.
molecules for each system. The TIP3P water model was used.45 The
systems were neutralized by adding the corresponding number of
counterions. Energy minimization was conducted in three steps.
First, only water molecules and ions were allowed to move. Next,
˚
the movement was extended to the residues within 5.0 A around the
mutated residue. Finally, all atoms were allowed to move freely. In
each step, energy minimization was carried out by a combination of
the steepest descent method for 5000 steps and the conjugated
gradient method for another 5000 steps.
(4) Semple, R. K.; Chatterjee, V. K.; O’Rahilly, S. PPAR gamma and
human metabolic disease. J. Clin. Invest. 2006, 116, 581–589.
(5) Forman, B. M.; Tontonoz, P.; Chen, J.; Brun, R. P.; Spiegelman,
B. M.; Evans, R. M. 15-Deoxy-delta 12,14-prostaglandin J2 is a
ligand for the adipocyte determination factor PPAR gamma. Cell
1995, 83, 803–812.
(6) Lehmann, J. M.; Moore, L. B.; Smith-Oliver, T. A.; Wilkison,
W. O.; Willson, T. M.; Kliewer, S. A. An antidiabetic thiazolidi-
nedione is a high affinity ligand for peroxisome proliferator-
activated receptor gamma (PPAR gamma). J. Biol. Chem. 1995,
270, 12953–12956.
(7) Olefsky, J. M.; Saltiel, A. R. PPAR gamma and the treatment of
insulin resistance. Trends Endocrinol. Metab. 2000, 11, 362–368.
(8) Staels, B. PPAR agonists and the metabolic syndrome. Therapie
2007, 62, 319–326.
(9) Willson, T. M.; Cobb, J. E.; Cowan, D. J.; Wiethe, R. W.; Correa,
I. D.; Prakash, S. R.; Beck, K. D.; Moore, L. B.; Kliewer, S. A.;
Lehmann, J. M. The structure-activity relationship between per-
oxisome proliferator-activated receptor gamma agonism and the
antihyperglycemic activity of thiazolidinediones. J. Med. Chem.
1996, 39, 665–668.
(10) Nesto, R. W.; Bell, D.; Bonow, R. O.; Fonseca, V.; Grundy, S. M.;
Horton, E. S.; Le Winter, M.; Porte, D.; Semenkovich, C. F.;
Smith, S.; Young, L. H.; Kahn, R. Thiazolidinedione use, fluid
retention, and congestive heart failure: a consensus statement from
the American Heart Association and American Diabetes Associa-
tion. Diabetes Care 2004, 27, 256–263.
(11) Bodmer, M.; Meier, C.; Kraenzlin, M. E.; Meier, C. R. Risk of
fractures with glitazones: a critical review of the evidence to date.
Drug. Saf. 2009, 32, 539–547.
(12) Berger, J. P.; Petro, A. E.; Macnaul, K. L.; Kelly, L. J.; Zhang,
B. B.; Richards, K.; Elbrecht, A.; Johnson, B. A.; Zhou, G.;
Doebber, T. W.; Biswas, C.; Parikh, M.; Sharma, N.; Tanen,
M. R.; Thompson, G. M.; Ventre, J.; Adams, A. D.; Mosley, R.;
Surwit, R. S.; Moller, D. E. Distinct properties and advantages of a
novel peroxisome proliferator-activated protein [gamma] selective
modulator. Mol. Endocrinol. 2003, 17, 662–676.
(13) Acton, J. J., III; Black, R. M.; Jones, A. B.; Moller, D. E.; Colwell,
L.; Doebber, T. W.; Macnaul, K. L.; Berger, J.; Wood, H. B.
Benzoyl 2-methyl indoles as selective PPARgamma modulators.
Bioorg. Med. Chem. Lett. 2005, 15, 357–362.
PPARγ and the related complexes with R-1 and S-1 have been
investigated in aqueous solution by means of MD simulations
using the GROMACS package.46 Each system was initially placed
in a rectangular box of 866 nm3 with 27480 water molecules
described by the single point charge (SPC)47 and six counterions
(Naþ). The dimension of the box was selected in order to avoid any
interaction of the solute with its replica, which could arise from the
application of periodic boundary conditions. Rototranslational
constraints were applied to the solute for obtaining correct statis-
tical mechanics and thermodynamics.48 For all the investigated
systems, we used the same initial coordinates, taken from the
PPARγ-R-1 crystal complex (PDB code 2I4J).23 The following
protocol was adopted for all the simulations: the systems were
initially minimized by using a standard steepest descent algorithm;
then, after solvation and the initial equilibration of the solvent, each
system was relaxed for 100 ps in an isothermal/isobaric ensemble to
obtain the correct density of the water. Each simulation was finally
carried out in an isothermal/isochoric ensemble at 300 K for about
15.3 ns using an integration step of 2.0 fs. For each simulation, the
first part of trajectory (0.5-2.0 ns) was discarded basing on the CR
root-mean-square deviations and only the last 14.0 ns of each
trajectory were analyzed. The temperature was kept constant by
the isokinetic temperature coupling.49 All bond lengths were con-
strained using LINCS.50 The long-range electrostatics was com-
puted by the particle mesh Ewald method,51 with 34 wave vectors in
each dimension and a fourth-order cubic interpolation. Gromos
force field52 parameters were adopted for the receptor, while the
Lennard-Jones parameters of similar atoms were considered for the
ligands.
Acknowledgment. We thank Dr. Krister Bamberg
€
(AstraZeneca, Molndal, Sweden) for kindly providing the
(14) Dropinski, J. F.; Akiyama, T.; Einstein, M.; Habulihaz, B.; Doebber,
T.; Berger, J. P.; Meinke, P. T.; Shi, G. Q. Synthesis and biological
activities of novel aryl indole-2-carboxylic acid analogs as PPARgam-
ma partial agonists. Bioorg. Med. Chem. Lett. 2005, 15, 5035–5038.
(15) Minoura, H.; Takeshita, S.; Ita, M.; Hirosumi, J.; Mabuchi, M.;
Kawamura, I.; Nakajima, S.; Nakayama, O.; Kayakiri, H.; Oku,
T.; Ohkubo-Suzuki, A.; Fukagawa, M.; Kojo, H.; Hanioka, K.;
Yamasaki, N.; Imoto, T.; Kobayashi, Y.; Mutoh, S. Pharmacolo-
gical characteristics of a novel nonthiazolidinedione insulin sensi-
tizer, FK614. Eur. J. Pharmacol. 2004, 494, 273–281.
(16) Oberfield, J. L.; Collins, J. L.; Holmes, C. P.; Goreham, D. M.;
Cooper, J. P.; Cobb, J. E.; Lenhard, J. M.; Hull-Ryde, E. A.; Mohr,
C. P.; Blanchard, S. G.; Parks, D. J.; Moore, L. B.; Lehmann, J. M.;
Plunket, K.; Miller, A. B.; Milburn, M. V.; Kliewer, S. A.; Willson,
T. M. A peroxisome proliferator-activated receptor gamma ligand
Gal4-PPARγ vector and pET28a-PPARγ-LBD expres-
sion vector and Drs. Anastasia Kralli (The Scripps Research
Institute, La Jolla, CA), Iannis Talianidis (Biomedical
Sciences Research Center “Alexander Fleming”, Vari, Greece),
and Serena Ghisletti (Istituto FIRC di Oncologia Molecolare,
Milano, Italia) for the kind gift of the expression vectors
for PGC-1R, RIP140, CBP, and N-CoR1. We also thank
Dr. Cristina Godio for expression and purification and
Dr. Donatella Caruso for mass spectrometry analysis of
His-tagged PPARγ LBD. We are also grateful to Elda
ꢀ
Desiderio Pinto (Universita degli Studi di Milano) for admin-

4366 Journal of Medicinal Chemistry, 2010, Vol. 53, No. 11
Pochetti et al.
inhibits adipocyte differentiation. Proc. Natl. Acad. Sci. U.S.A.
1999, 96, 6102–6106.
CREB-binding protein and ligand binding domains of nuclear
receptor: beyond the LXXLL motif. J. Biol. Chem. 2005, 280,
5682–5692.
(17) Burgermeister, E.; Schnoebelen, A.; Flament, A.; Benz, J.; Stihle,
M.; Gsell, B.; Rufer, A.; Ruf, A.; Kuhn, B.; Marki, H. P.; Mizrahi,
J.; Sebokova, E.; Niesor, E.; Meyer, M. A novel partial agonist of
peroxisome proliferator-activated receptor-gamma (PPARgamma)
recruits PPARgamma-coactivator-1alpha, prevents triglyceride ac-
cumulation, and potentiates insulin signaling in vitro. Mol. Endocri-
nol. 2006, 20, 809–830.
(18) Bruning, J. B.; Chalmers, M. J.; Prasad, S.; Busby, S. A.;
Kamenecka, T. M.; He, Y.; Nettles, K. W.; Griffin, P. R. Partial
agonists activate PPARgamma using a helix 12 independent
mechanism. Structure 2007, 15, 1258–1271.
(35) Molnar, F.; Matilainen, C.; Carlberg, J. Structural determinants of
the agonist-independent association of human peroxisome prolif-
erator-activated receptors with coactivators. J. Biol. Chem. 2005,
280, 26543–26556.
(36) Nolte, R. T.; Wisely, G. B.; Westin, S.; Cobb, J. E.; Lambert,
M. H.; Kurokawa, R.; Rosenfeld, M. G.; Willson, T. M.; Glass,
C. K.; Milburn, M. V. Ligand binding and co-activator assembly of
the peroxisome proliferator-activated receptor-gamma. Nature
1998, 395, 137–143.
(37) Bower, M. J.; Cohen, F. E.; Dunbrack, R. L., Jr. Prediction of
protein side-chain rotamers from a backbone-dependent rotamer
library: a new homology modeling tool. J. Mol. Biol. 1997, 267,
1268–1282.
(19) Nettles, K. W. Insights into PPARgamma from structures with
endogenous and covalently bound ligands. Nat. Struct. Mol. Biol.
2008, 15, 893–895.
(20) Xu, H. E.; Li, Y. Ligand-dependent and -independent regulation
of PPAR gamma and orphan nuclear receptors. Sci. Signal. 2008,
1, 52.
(38) Molecular Operating Environment (MOE), version 2005.06; Chemi-
cal Computing Group, Inc.: Montreal, 2005.
(39) Lefterova, M. I.; Zhang, Y.; Steger, D. J.; Schupp, M.; Schug, J.;
Cristancho, A.; Feng, D.; Zhuo, D.; Stoeckert, C. J., Jr.; Liu, X. S.;
Lazar, M. A. PPARgamma and C/EBP factors orchestrate adipo-
cyte biology via adjacent binding on a genome-wide scale. Genes
Dev. 2008, 22, 2941–2952.
(40) Araki, O.; Ying, H.; Furuya, F.; Zhu, X.; Cheng, S. Y. Thyroid
hormone receptor beta mutants: dominant negative regulators of
peroxisome proliferator-activated receptor gamma action. Proc.
Natl. Acad. Sci. U.S.A. 2005, 102, 16251–16256.
(21) Li, Y.; Zhang, J.; Schopfer, F. J.; Martynowski, D.; Garcia-Barrio,
M. T.; Kovach, A.; Suino-Powell, K.; Baker, P. R.; Freeman, B. A.;
Chen, Y. E.; Xu, H. E. Molecular recognition of nitrated fatty acids
by PPAR gamma. Nat. Struct. Mol. Biol. 2008, 15, 865–867.
(22) Chandra, V.; Huang, P.; Hamuro, Y.; Raghuram, S.; Wang, Y.;
Burris, T. P.; Rastinejad, F. Structure of the intact PPAR-gamma-
RXR-alpha nuclear receptor complex on DNA. Nature 2008, 456,
350–356.
(23) Pochetti, G.; Godio, C.; Mitro, N.; Caruso, D.; Galmozzi, A.;
Scurati, S.; Loiodice, F.; Fracchiolla, G.; Tortorella, P.; Laghezza,
A.; Lavecchia, A.; Novellino, E.; Mazza, F.; Crestani, M. Insights
into the mechanism of partial agonism: crystal structures of the
peroxisome proliferator-activated receptor gamma ligand-binding
domain in the complex with two enantiomeric ligands. J. Biol.
Chem. 2007, 282, 17314–17324.
(24) Montanari, R.; Saccoccia, F.; Scotti, E.; Crestani, M.; Godio, C.;
Gilardi, F.; Loiodice, F.; Fracchiolla, G.; Laghezza, A.; Tortorella,
P.; Lavecchia, A.; Novellino, E.; Mazza, F.; Aschi, M.; Pochetti, G.
Crystal structure of the peroxisome proliferator-activated receptor
gamma (PPARgamma) ligand binding domain complexed with a
novel partial agonist: a new region of the hydrophobic pocket could
be exploited for drug design. J. Med. Chem. 2008, 51, 7768–7776.
(25) Mattocks, A. R. Novel reactions of some R-acyloxy acid chlorides.
J. Chem. Soc. 1964, 1918–1930.
(26) Christensen, B. W.; Kjaer, A. The Absolute Configuration of
2-Hydroxy-2-methylbutyric Acid. Acta Chem. Scand. 1962, 16,
2466–2467.
(27) Xu, H. E.; Stanley, T. B.; Montana, V. G.; Lambert, M. H.;
Shearer, B. G.; Cobb, J. E.; McKee, D. D.; Galardi, C. M.; Plunket,
K. D.; Nolte, R. T.; Parks, D. J.; Moore, J. T.; Kliewer, S. A.;
Willson, T. M.; Stimmel, J. B. Structural basis for antagonist-
mediated recruitment of nuclear co-repressors by PPARalpha.
Nature 2002, 415, 813–817.
(41) De Fabiani, E.; Nitro, M.; Gilardi, F.; Caruso, D.; Crestani, M.
Coordinated control of cholesterol catabolism to bile acids and of
gluconeogenesis via a novel mechanism of transcription regulation
linked to the fasted-to-fed cycle. J. Biol. Chem. 2003, 278, 39124–
39132.
(42) Huang, C. C.; Couch, G. S.; Pettersen, E. F.; Ferrin, T. E. Chimera:
an extensible molecular modeling application constructed using
standard components. Pac. Symp. Biocomput. 1996, 1, 724 http://
www.cgl.ucsf.edu/chimera.
(43) Case, D. A.; Darden, T. A.; Cheatham, T. E., III; Simmerling,
C. L.; Wang, J.; Duke, R. E.; Luo, R.; Merz, K. M.; Pearlman,
D. A.; Crowley, M.; Walker, R. C.; Zhang, W.; Wang, B.; Hayik,
S.; Roitberg, A.; Seabra, G.; Wong, K. F.; Paesani, F.; Wu, X.;
Brozell, S.; Tsui, V.; Gohlke, H.; Yang, L.; Tan, C.; Mongan, J.;
Hornak, V.; Cui, G.; Beroza, P.; Mathews, D. H.; Schafmeister, C.;
Ross, W. S.; Kollman, P. A. AMBER 9; University of California: San
Francisco, 2006.
(44) Cornell, W. D.; Cieplak, P.; Bayly, C. I.; Gould, I. R.; Merz, K. M.,
Jr.; Ferguson, D. M.; Spellmeyer, D. C.; Fox, T.; Caldwell, J. W.;
Kollman, P. A. A second generation force field for the simulation
of proteins, nucleic acids, and organic molecules. J. Am. Chem. Soc.
1995, 117, 5179–5197.
(45) Jorgensen, W. L.; Chandrasekhar, J.; Madura, J.; Klein, M. L.
Comparison of simple potential functions for simulating liquid
water. J. Chem. Phys. 1983, 79, 926–935.
(28) Amadei, A.; Linssen, A. B.; Berendsen, H. J. Essential dynamics of
(46) Van der Spoel, D.; van Buren, A. R.; Apol, E.; Meulenhoff, P. J.;
Tieleman, D. P.; Sijbers, A. L. T. M.; van Drunen, R.; Berendsen,
H. J. C. Gromacs User Manual, version 1.3, 1996.
protein. Proteins 1993, 17, 412–425.
(29) Davies, S. S.; Pontsler, A. V.; Marathe, G. K.; Harrison, K. A.;
Murphy, R. C.; Hinshaw, J. C.; Prestwich, G. D.; Hilaire, A. S.;
Prescott, S. M.; Zimmerman, G. A.; McIntyre, T. M. Oxidized
alkyl phospholipids are specific, high affinity peroxisome prolif-
erator-activated receptor gamma ligands and agonists. J. Biol.
Chem. 2001, 276, 16015–16023.
(30) Nagy, L.; Schwabe, J. W. Mechanism of the nuclear receptor
molecular switch. Trends Biochem. Sci. 2004, 29, 317–324.
(31) Albers, M.; Blume, B.; Schlueter, T.; Wright, M. B.; Kober, I.;
Kremoser, C.; Deuschle, U.; Koegl, M. A novel principle for partial
agonism of liver X receptor ligands. Competitive recruitment of
activators and repressors. J. Biol. Chem. 2006, 281, 4920–4930.
(32) Phelan, C. A.; Weaver, J. M.; Steger, D. J.; Joshi, S.; Maslany, J. T.;
Collins, J. L.; Zuercher, W. J.; Willson, T. M.; Walker, M.; Jaye,
M.; Lazar, M. A. Selective partial agonism of liver X receptor alpha
is related to differential corepressor recruitment. Mol. Endocrinol.
2008, 22, 2241–2249.
(47) Berendsen, H. J. C.; Postma, J. P. M.; van Gunsteren, W. F.
Hermans interaction models for water in relation to protein
hydration. In Intermolecular Forces; Pullman, B., Ed.; Reidel Pub-
lishing Company: Dordrecht, The Netherlands, 1981; pp 331-342.
(48) Amadei, A.; Chillemi, G.; Ceruso, M.; Grottesi, A.; Di Nola, A.
Molecular dynamics simulations with constrained roto-transla-
tional motions: theoretical basis and statistical mechanical consis-
tency. J. Chem. Phys. 2000, 112, 9–23.
(49) Berendsen, H. J. C.; Postma, J. P. M.; van Gunsteren, W. F.;
Di Nola, A. Molecular dynamics with coupling to an external bath.
J. Chem. Phys. 1984, 81, 3684–3690.
(50) Hess, B.; Bekker, H.; Berendsen, H. J. C.; Frajie, J. G. E. M.
LINCS: a linear constraint solver for molecular simulations.
J. Comput. Chem. 1997, 18, 1463–1472.
(51) Darden, T. A.; York, D. M.; Pedersen, L. G. Particle mesh Ewald:
method for Ewald sums in large systems. J. Chem. Phys. 1993, 98,
10089–10092.
(52) van Gunsteren, W. F.; Billeter, S. R.; Eising, A. A.; Hunemberger,
P. H.; Kruger, P.; Mark, A. E.; Scott, V. R. P.; Tironi, I. G.
Biomolecular Simulations: The GROMOS96 Manual and User
Guide; vdf Hochschlverlag AG an der ETH Zurich: Zurich, 1996.
(33) Treuter, E.; Albrektsen, T.; Johansson, L.; Leers, J.; Gustafsson,
J. A. A regulatory role for RIP140 in nuclear receptor activation.
Mol. Endocrinol. 1998, 12, 864–881.
(34) Klein, F. A.; Atkinson, R. A.; Potier, N.; Moras, D.; Cavarelli, J.
Biochemical and NMR mapping of the interface between