European Journal of Medicinal Chemistry 89 (2015) 817e825

Contents lists available at ScienceDirect

European Journal of Medicinal Chemistry journal homepage: http://www.elsevier.com/locate/ejmech

Original article

Structural development studies of PPARs ligands based on tyrosine scaffold Barbara De Filippis a, Pasquale Linciano a, Alessandra Ammazzalorso a, Carmen Di Giovanni c, Marialuigia Fantacuzzi a, Letizia Giampietro a, Antonio Laghezza b, Cristina Maccallini a, Paolo Tortorella b, Antonio Lavecchia c, **, Fulvio Loiodice b, **, Rosa Amoroso a, *
“G. d'Annunzio”, Via dei Vestini 31, 66100 Chieti, Italy Dipartimento di Farmacia, Universita
degli Studi di Bari “Aldo Moro”, Via Orabona 4, 70126 Bari, Italy Dipartimento di Farmacia-Scienze del Farmaco, Universita c
degli Studi di Napoli Federico II, Via Domenico Montesano 49, 80131 Napoli, Italy Dipartimento di Farmacia, “Drug Discovery” Laboratory, Universita a

b

a r t i c l e i n f o

a b s t r a c t

Article history: Received 16 July 2014 Received in revised form 28 October 2014 Accepted 30 October 2014 Available online 31 October 2014

PPARs are nuclear receptors with a critical physiological role in lipid and glucose metabolism. As part of our effort to develop new and selective PPAR agonists containing stilbene and its bioisoster phenyldiazene, novel analogs were synthesized starting from tyrosine and evaluated as PPAR agonists. We tested the effects of phenyloxazole replacement of GW409544, a well-known PPARa/g dual agonist, with stilbene or phenyldiazene moiety, spaced by an ether bridge to tyrosine portion. These structural modifications provided potent and selective PPARg agonists. Molecular docking studies performed on these new compounds complemented the experimental results and allowed to gain some insights into the nature of binding of the ligands. © 2014 Elsevier Masson SAS. All rights reserved.

Keywords: PPARs Stilbene Phenyldiazene PPARg agonist Tyrosine

1. Introduction Peroxisome Proliferator-Activated Receptors (PPARs) belong to the II class of nuclear receptors (NRs) superfamily [1]. They are ligand-dependent transcriptional factors involved in the control and expression of several genes implicated in glucidic and lipidic homeostasis and energetic balance [2]. In humans, three different isoforms have been identified: PPARa, PPARb/d, and PPARg. They have different tissue distribution as well as different binding affinity for ligands and recruitment ability of coactivators and

Abbreviations: PPARs, peroxisome proliferator-activated receptors; NR, nuclear receptor; PPARa, peroxisome proliferator-activated receptor a; PPARb/d, peroxisome proliferator-activated receptor b/d; PPARg, peroxisome proliferator-activated receptor g; VLDL, very low density lipoprotein; LDL, low density lipoprotein; HDL, high density lipoprotein; DCC, dicyclohexylcarbodiimide; HOBt, hydroxybenzotriazole; hPPARa, human PPARa; hPPARg, human PPARg; hPPARd, human PPARd. * Corresponding author. Dipartimento di Farmacia, Universit
a “G. d'Annunzio”, Via dei Vestini 31, 66100 Chieti, Italy ** Corresponding authors. E-mail addresses: [email protected] (A. Lavecchia), fulvio.loiodice@ uniba.it (F. Loiodice), [email protected] (R. Amoroso). http://dx.doi.org/10.1016/j.ejmech.2014.10.083 0223-5234/© 2014 Elsevier Masson SAS. All rights reserved.

corepressors. In their physiological activity, PPARs are activated by long-chain fatty acids, cycloxygenase-derived prostaglandines and their metabolites [3]. PPARa is mostly involved in the control of lipidic catabolism: it induces fatty acid b-oxidation, resulting in decrease of very low density lipoprotein (VLDL), low density lipoprotein (LDL), and increase of high density lipoprotein (HDL) in blood. On vascular cells it exhibits anti-inflammatory and antiaggregating activity. Fibrates are PPARa agonists currently used as lipid-lowering agents [4]. PPARb/d is ubiquitously distributed, with relatively higher levels in brain, adipose tissue and skin. It is involved in lipid metabolism, with a physiological profile similar but non interchangeable with PPARa [5]. No PPARb/d agonists approved by health organizations are available. PPARg is expressed in adipose tissue, where it induces lipogenesis and fat storage, and in skeletal muscle, where it improves insulin sensitivity [6]. It is the target of thiazolidinedione class of insulin-sensitizing drugs, clinically employed in patients with type 2 diabetes [7]. PPARs, therefore, represent valuable therapeutic targets for the treatment of both hyperlipidemia and insulin resistance in metabolic disorders, including metabolic syndrome [8]. For this purpose, in the last years many efforts have been directed toward the combination, in a unique chemical entity, of the insulin-sensitizing effect of PPARg

818

B. De Filippis et al. / European Journal of Medicinal Chemistry 89 (2015) 817e825

activation with the additional lipid-modifying activity of other PPAR subtypes. One of the approaches adopted has been the development of dual PPARa/g agonists, considered a very attractive option in the treatment of dyslipidemic type 2 diabetes [9]. A basically three-module structure for dual PPARa/g agonists is always exactly conserved: a carboxylic acid head group, a simple linker group generally bearing a phenylether, and a lipophylic tail [10]. A number of dual agonists belonging to the class of isoxazolidinediones, N-substituted tyrosine analogs, a-substituted-3phenyl propanoic acid derivatives, thiazolidine-2,4-dione and barbituric acid hybrids, and others have been developed and studied (Fig. 1) [11]. An alternative approach for the treatment of metabolic disorders is represented by the development of partial PPARg agonists or selective PPARg modulators [12]; in fact, ligands endowed with PPARg full agonist activity, despite their proven benefits, possess a number of deleterious side effects such as weight gain, peripheral edema, increased risk of congestive heart failure, and higher rate of bone fracture [13]. Some selected agonists with attenuated PPARg activity showed to retain the beneficial effects while reducing the adverse effects, but none of them has progressed further than preclinical or phase II trials. This necessitates the development of additional pharmacophores for the activation of this PPAR subtype. Natural stilbenes, chalcones, and some of their synthetic derivatives have shown to interact with PPARs, thus showing a potential as therapeutic antilipidemic and antidiabetic agents [14]. As a part of the ongoing research to find an effective PPAR-target based drug candidate, we have recently reported new PPARa agonists derived by the combination of antilipidemic drugs gemfibrozil and clofibric acid (the active metabolite of clofibrate) with natural a-asarone, stilbene, chalcone, and other bioisosteric modifications [15]. The highest agonistic activity was seen with an E-stilbene derivative of gemfibrozil and a phenyldiazene derivative of clofibric acid (Fig. 2). In this study, we describe the synthesis and the evaluation of PPAR activity of the new tyrosine derivatives 1aeo (Fig. 3), based on the combination of GW409544, a potent full agonist on both PPARa and PPARg, and stilbene or phenyldiazene scaffolds. The tyrosine O

a)

b) N

O

N

Fig. 2. a) Trans-stilbene derivative of gemfibrozil and b) phenyldiazene derivative of clofibric acid.

portion of GW409544 remained essentially unchanged, except two derivatives in which the hydrophobic appendix was modified by introducing a methyl in place of phenyl (1n) or inverting the position of the carbonyl and ethylenic groups of the vinylogous amide (1o), respectively. The phenyloxazole moiety was replaced with stilbene or its bioisoster phenyldiazene, spaced by an alkylphenoxy linker from the central phenyl ring. Noteworthy, most compounds were highly potent selective PPARg agonists with efficacy variable in the 30e80% range compared to the reference compound rosiglitazone. To gain more details on the interactions at a molecular level and to propose a binding mode explaining the SAR data, docking experiments were carried out. An isomerization process was hypothesized and confirmed by 1H NMR studies to account for the lower activity of the phenyldiazene derivatives compared to the stilbene ones. 2. Results and discussion 2.1. Chemistry Two series of compounds were synthesized: the N-vinylogous tyrosine derivatives (1aen) and the N-amide tyrosine derivative 1o. Some of them are stilbene derivatives (1aeg and 1neo), the others Cl

O

COOi-Pro

COOH

O

Cl

O Gemfibrozil

Clofibrate

Fenofibrate

S

S

O O

b)

O

N H

HO

O

N

O

O

N Troglitazone

N H

Rosiglitazone

COOH

O

c)

COOH

O N

O

S

COOH

O

COOEt

a)

COOH

O

N

O

HN O

OMe

Aleglitazar GW409544 Fig. 1. Chemical structures of representative a) fibrates, b) thiazolidinediones, and c) dual PPARa/g agonists.

B. De Filippis et al. / European Journal of Medicinal Chemistry 89 (2015) 817e825

819

Hydrophobic tail COOH

O N

HN

O

O

linker

Hydrophobic appendix

GW409544

X A

COOH

A O

NHR

nO

1a-o Fig. 3. GW409544 and designed novel tyrosine derivatives.

are their phenyldiazene bioisosteres (1hem). The suitable stilbene, phenyldiazene and tyrosine building blocks were prepared, and the final compounds were obtained by condensation of these molecular intermediates (Schemes 1e5). Compounds 1ceg and 1jem present a substituent in 4-position of the distal aromatic ring, with different lipophilic and electronic properties. Due to the very low solubility into polar solvents of compound 1e (n ¼ 2), we did not synthesize its homologous with n ¼ 3. Moreover, we could not obtain the ethylenic homologs of compounds 1l and 1m, because of their unexpected instability. The stilbene scaffolds condensed with the linker groups (5aeg) were synthesized according to Scheme 1. The SN2 reaction between 4-hydroxybenzaldehyde and 2-iodoethanol or 3-bromopropanol,

in DMF in the presence of K2CO3, furnished the corresponding aldehydes 2aeb. Phosphonium salts 3aed were obtained in quantitative yields refluxing the appropriate benzylchloride with PPh3, in toluene. The aldehydes 2aeb were then condensed under Wittig's conditions with the phosphonium salts 3aed, in the presence of LiOH and LiCl in H2O at reflux, to achieve the alcohols 4aeg as mixtures of the two isomers E and Z, with a stoichiometric ratio 70:30, determined by NMR spectra. Each E/Z mixture was fully converted to the corresponding E-isomers by reflux in toluene for 1 h with a grain of I2 as a catalyst. The halogenation of 4aeg with PPh3, imidazole and I2 in toluene gave the corresponding alkyl iodides 5aeg [16].

O

O a

H

H O

OH

nOH

2a n = 2 2b n = 3

X c, d O

X

b

X PPh3+ Cl-

Cl 3a 3b 3c 3d

X=H X = Cl X = CH3 X = CF3

4a 4b 4c 4d

X=H X = Cl X = CH3 X = CF3

n=2 n=2 n=2 n=2

nOH

4e X = H n=3 4f X = Cl n = 3 4g X = CF3 n = 3

e X

O 5a 5b 5c 5d

X=H X = Cl X = CH3 X = CF3

n=2 n=2 n=2 n=2

nI

5e X = H n = 3 5f X = Cl n = 3 5g X = CF3 n = 3

Scheme 1. Reagents and conditions: (a) 2-iodoethanol or 3-bromopropanol, K2CO3, DMF, 150  C, 4 h; (b) PPh3, toluene, reflux, 3e6 h; (c) LiCl, 1 N LiOH, H2O, reflux, 6 h; (d) I2 grain, toluene, reflux, 1 h; (e) PPh3, imidazole, I2, toluene, r.t., 6e12 h.

820

B. De Filippis et al. / European Journal of Medicinal Chemistry 89 (2015) 817e825

X

c

N

N

Br

O

X

X

a, b

7a 7b 7c 7d

N

N

NH2

OH 6a 6b 6c 6d

X=H X = Cl X = NO2 X = OCH3

d

X=H X = Cl X = NO2 X = OCH3

X N

N O

OH

8a X = H 8b X = Cl

e

X N

N O

I

9a X = H 9b X = Cl Scheme 2. Reagents and conditions: (a) NaNO2, 6 N HCl, 0  C, 1 h; (b) PhOH, 6 N NaOH, 0  C, 1 h; (c) 1,3-dibromopropane, K2CO3, DMF, r.t. to 100  C, 2 h; (d) 2-iodoethanol, K2CO3, DMF, r.t. to 150  C, 10e15 h; (e) PPh3, imidazole, I2, toluene, r.t., 6e12 h.

HO

COOCH3

a

COOCH3

HO

NH2

HN O R 10a R = Ph 10b R = CH3

Scheme 3. Reagents and conditions: (a) acetylacetone or benzoylacetone, MeOH, reflux, 24 h.

For the synthesis of the phenyldiazene backbones condensed with the linker group, the phenylazophenol intermediates 6aed were obtained by a diazotationecopulation reaction starting from the appropriate 4-substitued anilines and phenol. Compounds 6aed were then reacted with 1,3-dibromopropane, in standard SN2 conditions, to give bromides 7aed, or with 2-iodoethanol to give the alcohols 8aeb which were converted into the corresponding iodides 9aeb as above described, with PPh3, imidazole and I2 in toluene (Scheme 2). The N-substituted tyrosine methylester intermediates were obtained as depicted in Schemes 3e4. The starting material (S)-tyrosine methyl ester was reacted with acetylacetone or benzoylacetone in MeOH at reflux for 24 h, to get the enamines 10 aeb in good yield (Scheme 3). Amide 11 was

COOCH3 HO

COOCH3

a

NH3+ Cl-

HO

HN O 11

Scheme 4. Reagents and conditions: methylmorpholine, DMF, 0  C to r.t., 8 h.

(a)

cinnamic

acid,

DCC,

HOBt,

N-

synthesized by coupling of (S)-tyrosine methyl ester hydrochloride and cinnamic acid in the presence of dicyclohexylcarbodiimide (DCC), N-methylmorpholine and hydroxybenzotriazole (HOBt), in DMF (Scheme 4). The final carboxylic acids 1aeo were obtained according to Scheme 5. Esters 12aeo were synthesized, under standard SN2 conditions, by reacting the alkyl halides 5aeg, 7aed, and 9aeb with the N-substituted tyrosine methyl esters 10aeb, and 11. All esters underwent racemization of Ca of the tyrosine moiety. After hydrolysis of 12aeo, carried out with 1 N LiOH in THF:MeOH 3:1 at room temperature (r.t.), the carboxylic acids 1aeo were obtained.

2.2. Biological activity Compounds 1aeo were evaluated for their agonist activity on the human PPARa (hPPARa), PPARg (hPPARg), and PPARd (hPPARd) subtypes. For this purpose, GAL4-PPAR chimeric receptors were expressed in transiently transfected HepG2 cells according to a previously reported procedure [17]. The results obtained (Table 1) were compared with corresponding data for Wy-14,643, rosiglitazone, and L-165,041 used as reference compounds in the PPARa, PPARg, and PPARd transactivation assays, respectively. The maximum induction obtained with the reference agonist was defined as 100%. The substitution of phenyloxazole moiety of GW409544 with stilbene scaffold and its bioisoster phenyldiazene led to a series of compounds endowed with potent and selective PPARg agonist activity, but inactive on PPARa and PPARd. Only the unsubstituted stilbene 1a displayed a little activity towards PPARa even though its EC50 was not computable; in fact, the activity increased with increasing concentrations up to 10 mM, above which the activity began to decrease due to its probable cytotoxicity. As regards the structureeactivity relationships, some interesting considerations can be drawn. The azo group reduced PPARg activity compared to ethylenic double bond; compounds 1h and 1i, in fact,

B. De Filippis et al. / European Journal of Medicinal Chemistry 89 (2015) 817e825 X A

a

NHR

HO O

X

COOMe

+

A

821

A

COOMe

A

nX

O

NHR

nO

10a-b, 11

5a-g, 7a-d, 9a-b

12 a-o

b

X A

COOH

A O

NHR

nO

1a-o

Scheme 5. Reagents and conditions: (a) K2CO3, DMF, r.t. to 40  C, 12e24 h; (b) 1 N LiOH, THF, MeOH, r.t., 6e12 h.

between the ether oxygen atoms significantly improved the activity when the substituent on the aromatic ring was lacking (1a and 1h had about 4-fold and 8-fold higher potency than 1b and 1i, respectively). By contrast, the length of the linker does not seem to affect the activity following the introduction of a substituent (compare 1c, 1f, and 1j with 1d, 1g, and 1k, respectively). Finally, for the last two compounds of the series, a modification of the group linked to the nitrogen was carried out. The phenyl group was

had about 15-fold and 30-fold lower potency than 1a and 1b, respectively. However, when a substituent was introduced on the distal phenyl ring, phenyldiazene derivatives turned out to be as active as stilbene derivatives (compare compounds 1j and 1k with 1c and 1d, respectively). Except for methoxy derivative 1m, different stereo-electronic properties of the substituent did not significantly affect the activity (see compounds 1ce1g and 1ke1l). This could apply, also, for the linker; in fact, the two carbon chain

Table 1 Biological evaluation through transactivation assay of compounds 1aeo.

X A

COOH

A O

Compound

R1

X

1a

1b 1c 1d 1e 1f 1g 1h 1i 1j 1k 1l 1m 1n

1o

Wy-14,643 Rosiglitazone L-165,041

“ “ “ “ “ “ “ “ “ “ “ “ O

O

A

n

NHR

nO

PPARa

PPARg

PPARd

EC50 (mM)

E (%)a

EC50 (mM)

E (%)

EC50 (mM)

E (%)

75 ± 12

i.ac

i.a.

± ± ± ± ± ± ± ± ± ± ± ± ±

i.a. i.a. i.a. i.a. i.a. i.a. i.a. i.a. i.a. i.a. i.a i.a. i.a.

i.a. i.a. i.a. i.a. i.a. i.a. i.a. i.a. i.a. i.a. i.a. i.a. i.a.

H

CH

2

n.c.b

20 ± 1

0.012 ± 0.008

H Cl Cl CH3 CF3 CF3 H H Cl Cl NO2 OCH3 H

CH CH CH CH CH CH N N N N N N CH

3 2 3 2 2 3 2 3 2 3 3 3 2

n.c. i.a. i.a. i.a. i.a. i.a. i.a. n.c. i.a. i.a. i.a. i.a. i.a.

9±4 i.a. i.a. i.a. i.a. i.a. i.a. 8±2 i.a. i.a. i.a. i.a. i.a.

0.047 0.026 0.012 0.021 0.084 0.032 0.183 1.370 0.039 0.047 0.029 n.c. 0.250

H

CH

2

i.a.

i.a.

0.640 ± 0.02

45 ± 1

i.a.

i.a.

1.56 ± 0.30 i.a. i.a.

100 ± 10 i.a. i.a.

i.a. 0.039 ± 0.003 i.a.

i.a. 100 ± 10 i.a.

i.a i.a. 0.021 ± 0.01

i.a. i.a. 100 ± 4

± ± ± ± ± ± ± ± ± ± ±

0.005 0.006 0.001 0.014 0.022 0.005 0.004 0.06 0.014 0.012 0.007

± 0.050

65 50 56 77 69 46 54 31 52 59 54 24 61

15 8 8 7 18 5 11 4 5 8 3 5 9

a E%, efficacy values calculated as percentage of the maximum obtained fold induction with the reference compounds (Wy-14,643 for PPARa, rosiglitazone for PPARg, and L165,041 for PPARd) [20]. b n.c., not computable; the activity, in fact, increases with increasing concentrations up to 10 mM above which the activity begins to decrease. c i.a., inactive at tested concentrations.

822

B. De Filippis et al. / European Journal of Medicinal Chemistry 89 (2015) 817e825

removed and replaced by a methyl (1n) and the enamine was substituted by an amide (1o). In both cases the activity decreased: 1n and 1o had 50-fold and 20-fold lower potency than reference compound 1a, respectively. 2.2.1. Docking calculations In an attempt to understand the structural basis for the observed PPARg agonist effects of compounds 1aeo and to identify the molecular determinants of their subtype selectivity, in silico docking analysis was performed using Autodock Vina [18]. This generates low-energy binding poses by evaluating the combined energetic contributions of torsion, steric repulsion, H-bonding, and hydrophobic interaction between the ligand and the protein binding pocket. Docking experiments were carried out into the Xray crystal structures of PPARg (PDB code: 1K74) and PPARa (PDB code: 1KTL) complexed to GW409544 [11]. Blind docking was performed on PPARa and PPARg structures to validate its accuracy at reproducing the experimental binding mode of GW409544. The co-crystal ligand GW409544 was removed from the active site and docked back into the binding site. Only the highest ranked poses with strongest binding affinities (kcal/mol) were selected. The results of blind docking indicate that the binding conformation of GW409544 determined by AutoDock Vina matches well with that of the co-crystallized ligand in both PPARa and PPARg binding pockets. The root mean square deviation (rmsd) between the top ranked predicted conformation and the observed X-ray crystallographic conformation was 1.3 Å for both PPARa and PPARg. This result, in addition to previous extensive validation [18,19], demonstrates the ability of AutoDock Vina to accurately predict the binding conformations of ligandeprotein complexes. It has been shown that the S-enantiomers of the tyrosine-based PPARg agonists, such as GW409544, farglitazar, GW1929, and GW7845, have greater binding affinity and functional activity at PPARg than the corresponding R-enantiomers [20]. This is in accord with the X-ray crystal structures of farglitazar [21] and GW409544 [11] bound to PPARg, in which both ligands were found to have Sconfiguration. Based on these evidences, compounds 1ael, which are close analogs of farglitazar and GW409544, were docked as Senantiomers in this study. Ligands 1ael were predicted to bind into the large PPARg pocket in a U-shaped conformation similar to that observed for GW409544 with the carboxylate group oriented toward the AF-2 helix. Fig. 4 shows the complex between PPARg and compounds 1a and 1j chosen as representative examples of stilbene and phenyldiazene derivatives, respectively. H-bonds are made between the carboxylate group of the ligands and the critical residues H449 (H11), Y473 (H12), H323 (H5), and S289 (H3). Ligands contain a vinylogous amide substituent that occupies, without completely fill, the lipophilic benzophenone pocket formed by helices 3, 7, and 10. The terminal phenyl ring of amide substituent forms hydrophobic contacts with F282, C285, L330 and F363 and is further stabilized by a parallel displaced pstacking interaction with the aromatic ring of F363 of helix 7. This result is in consonance with the SAR data displaying that replacement of the terminal phenyl ring of amide substituent with a methyl group, as in compound 1n, decreases the potency towards PPARg. The remainder of the ligands wraps around helix 3 and buries the stilbene tail into a lipophilic pocket formed by H3, the b sheet and the U-loop, that links H20 to H3. The stilbene tail is shifted ~4 Å compared with the phenyloxazole tail of GW409544 into the hydrophobic ligand-binding pocket, while maintaining the critical interactions with the H3/b sheet/U-loop pocket. Compounds 1ael make important hydrophobic interactions with residues G284, F287, C285, R288, L330, I341, S342, T268 and E343. The azo nitrogen of compounds 1he1l accepts a H-bond from the OH group of S342

side chain located within the b-sheet region of the receptor. In addition, the distal ring of stilbene tail appears to be optimally oriented for a favorable T-shaped pep stacking interaction with F287 positioned on the helix 3, which contributes to further increase the H3 stabilization. When the para position of the azobenzene distal benzene ring is replaced by a methoxy group, as in 1m, a steric clash occurs with T268 (U-loop) and E343 (b-sheet), which leads to a repositioning of the entire molecule in the binding pocket. Therefore, the significant reduced PPARg activity of 1m is ascribable to an unfavorable steric interaction with the protein. The fact that compounds 1ael bind the PPARg ligand binding domain in a mode slightly different from that of the full agonist GW409544 raises the possibility that receptor dynamics could be differentially affected by the particular ligand binding mode [22]. It can be assumed that the partial agonist effect of these compounds may derive from a mechanism distinct from full agonists and that the helix 3, the b-sheet and the U-loop participate in these effects [23]. It is apparent that docking calculations do not provide an exhaustive rationalization of the 15- and 30-fold lower potency of the unsubstituted phenyldiazenes 1hei, compared with the corresponding stilbene ones 1aeb. In fact, a very similar binding mode for 1h and 1i was obtained, with a comparable binding affinity. However, it is known that azobenzene derivatives possess a photosensitive group and undergo a reversible isomerization upon light irradiation. Moreover, the effect of substituents, solvents and temperature can strongly influence the rate of thermal E-to-Z relaxation [24]. In the E state, the azo group is in its linear structure (with a molecular length of 9 Å) and its polarity is small (0.52 D) due to the axial symmetry, while in the Z state, the bent structure (with a molecular length of 5 Å) leads to an increased dipolar moment (3.08 D) due to its non-axial symmetry. Therefore, it can be predicted that unsubstituted phenyldiazene derivatives as 1hei undergo E-to-Z isomerization more easily than substituted ones 1je1m, and that, when they are in the Z configuration, they do not bind to the receptor owing to steric repulsive interactions with R288, T268, F287, and E343 side chains, thus explaining the observed difference in PPARg potency. This hypothesis was confirmed performing a 1H NMR investigation on compounds 1h and 1j as reported below. To explain the dramatic difference in PPAR subtype selectivity of ligands 1aeo, we carried out docking calculations of the representative compound 1a into PPARa binding cleft. AutoDock Vina predicted a binding mode associated with a poor estimated binding affinity (4.3 kcal/mol). From a visual inspection of the 1a/PPARa complex, it seems clear that the presence of the stilbene or phenyldiazene scaffold increases the steric hindrance inside the binding cavity and changes the optimal binding mode of the ligand, causing the loss of the H-bonding network between the carboxylate group and the key amino acid residues S280, Y314, H440, and Y464. On the other hand, substitution of the larger Y314 in PPARa to the sterically smaller His323 in PPARg has previously been shown to be largely responsible for subtype selectivity of the tyrosine-based PPARg agonist, farglitazar [9]. Xu et al.'s work postulated that the specificity of ligand binding arises from small changes in the position of ligand when docked into the PPAR pockets that are in various shapes. 2.3.

1

H NMR studies on photoisomerization

Thanks to the biological and photophysical properties, the azocompounds occupy a prestigious position in medicinal chemistry research; so far, they have been shown to exert antimicrobial, antitumor, and antiviral effects, and have been utilized in the construction of molecular switches for the regulation of enzyme activity [25]. Reductive enzymes in the liver and intestinal

B. De Filippis et al. / European Journal of Medicinal Chemistry 89 (2015) 817e825

823

photoisomerization of the unsubstituted analog 1h and the chlorine derivative 1j, chosen as representative examples, through 1H NMR studies (Fig. 5). We acquired the spectra at time zero and after 2, 6, and 24 h; during this time, compounds were maintained in DMSO solution at room temperature and exposed to the day-light. The unsubstituted phenildiazene 1h showed a partial photoisomerization; this phenomenon was clearly visible already after two hours through the appearance of a splitted signal at 4.20 ppm, associated to the methylene of the spacer group linked to the oxygen of the azobenzene moiety, and two set of signals in the aromatic region at 6.81 ppm and at 7.30 ppm, respectively. After 24 h, the relative intensity of these multiplets remained unaltered over time, with a 60:40 ratio of E and Z isomers. Conversely, the same experiment conducted on the chloro-derivative 1j did not show any new signal confirming that the presence of a chlorine in para position of the distal phenyl ring inhibits or delays the photoisomerization process. By contrast, the stilbene derivatives are stable to photoisomerization, as demonstrated by the same NMR studies.

3. Conclusion

Fig. 4. Binding mode of stilbene derivative 1a (a) and phenyldiazene derivative 1j (b) into the PPARg binding site represented as a ribbon model. Only amino acids located within 4 Å of the bound ligand are displayed and labeled. H-bonds discussed in the text are depicted as dashed lines.

microbial azoreductase can catalyze the reductive cleavage of the azo linkage to produce aromatic amines [26]. Phenyldiazene or azobenzene is known to undergo E-to-Z isomerization when irradiated with UVevisible light or subjected to mechanical stress or electrostatic stimulation, and the ring substituents can affect this process through steric and electronic effects [24]. In the case of biologically active molecules, the photoisomerization can modify the interactions into the ligand binding site of the receptor, strongly influencing the biological response [27]. In order to ascertain the possible isomerization process for our phenyldiazenes, we investigated the stability toward

In conclusion, in this paper we reported the synthesis and biological activity of new compounds derived from GW409544, a wellknown PPARa/g dual agonist. The substitution of phenyloxazole moiety of GW409544 with stilbene scaffold and its bioisoster phenyldiazene led to a series of compounds endowed with potent and selective PPARg agonist activity, but inactive on PPARa and PPARd. Docking experiments performed on these new compounds allowed to explain their selectivity on PPARg subtype as well as the lower activity of phenyldiazenes devoid of substituents on the distal phenyl ring. For these compounds, a possible E to Z isomerization process accounting for their reduced potency was predicted by docking experiments. NMR studies confirmed that this process partially occurs in a time frame compatible with that of the bioassay. Further biological and pharmacological investigation is under way to support the hypothesis that these novel PPARg partial agonists may act as selective PPARg modulators representing, therefore, potential leads of a new class of drugs with better therapeutic effects in the treatment of dyslipidemic type 2 diabetes.

Fig. 5. Zooming E-1h and E-1j 1H NMR spectra showing photoisomerization E-to-Z of 1h.

824

B. De Filippis et al. / European Journal of Medicinal Chemistry 89 (2015) 817e825

4. Experimental section

interactions, repulsions, H-bonding, etc.) and intramolecular (torsion, rotational torque) factors.

4.1. Synthetic chemistry 4.3. Biological methods All synthetic data are reported in the Supporting Information. 4.2. Computational chemistry Molecular modeling and graphics manipulations were per€dinger) [28] and UCSF-Chimera formed using Maestro 9.7 (Schro 1.8.1 software packages [29] running on a E4 Computer Engineering E1080 workstation provided of a Intel Core i7-930 Quad-Core processor. Figures were generated using Pymol 1.0 [30]. 4.2.1. Protein and ligand preparation The crystal structures of PPARa (PDB ID: 1K7L) and PPARg (PDB ID: 1K74) complexed with the same ligand GW409544 [9] were download from the PDB Bank [31] and employed for the automated docking studies. The two proteins were processed through the Protein Preparation Wizard in Maestro [32]. All crystallographic water molecules and other chemical components were deleted, the right bond orders as well as charges and atom types were assigned and the hydrogen atoms were added to both proteins. Arginine and lysine side chains were considered as cationic at the guanidine and ammonium groups, and the aspartic and glutamic residues were considered as anionic at the carboxylate groups. Imidazole rings of H440 into PPARa and H449 and H323 into PPARg were set in their Nε 2-H (N tau-H) tautomeric state. Moreover, an exhaustive sampling of the orientations of groups, whose H-bonding network needs to be optimized, was performed. Finally, the protein structures were refined with a restrained minimization with the OPLS2005 force field [33] by imposing a 0.3 Å rmsd limit as the constraint. The structures of 1ae1o ligands were built using the fragment dictionary of Maestro and preprocessed with LigPrep 2.9 [34], which prepares the ligands in multiple protonation and tautomerization states at a neutral pH. Ligands were then optimized by Macromodel 10.2 [35] using the MMFFs force field with the steepest descent (1000 steps) followed by truncated Newton conjugate gradient (500 steps) methods. Partial atomic charges were computed using the OPLS-AA force field. 4.2.2. Docking simulations The computer-simulated automated docking studies were performed using the molecular docking software Autodock Vina 1.1.2 [18]. AutoDockTools 1.5.6 [36] was used to remove crystal waters and add polar hydrogens and to save the protein in the appropriate file formate for docking with AutoDock Vina. The distance between donor and acceptor atoms that form a H-bond was defined as 1.9 Å with a tolerance of 0.5 Å, and the acceptorehydrogenedonor angle was not less than 120 [37]. For both PPARa and PPARg, the search space was a box with XYZ dimensions 50 Å  56 Å  72 Å respectively, and was centered on the protein. The above search space dimension encompasses the entirety of the protein, as is required for blind docking. The AutoDock Vina parameter “Exhaustiveness”, which determines how comprehensively the program searches for the lowest energy conformation, was set to the default value, eight, for both docking setups. Once the blind docking technique was validated, the search space was decreased to encompass only the ligand binding sites. Autodock Vina finds ligand poses with the best fit and strongest binding affinity (global minimums) by a stochastic algorithm to explore surfaces/pockets of the rigid macromolecule, through an iterative series of local optimizations evaluating both intermolecular (hydrophobic

Reference compounds, the medium, and other cell culture reagents were purchased from SigmaeAldrich (Milan, Italy). 4.3.1. Plasmids The expression vectors expressing the chimeric receptor containing the yeast Gal4-DNA binding domain fused to the human PPARa, PPARg, or PPARd ligand binding domain (LBD) and the reporter plasmid for these Gal4 chimeric receptors (pGal5TKpGL3) containing five repeats of the Gal4 response elements upstream of a minimal thymidine kinase promoter that is adjacent to the luciferase gene were described previously [38]. 4.3.2. Cell culture and transfections Human hepatoblastoma cell line HepG2 (Interlab Cell Line Collection, Genoa, Italy) was cultured in minimum essential medium (MEM) containing 10% heat-inactivated fetal bovine serum, 100 U of penicillin G mL1, and 100 mg of streptomycin sulfate mL1 at 37  C in a humidified atmosphere of 5% CO2. For transactivation assays, 105 cells per well were seeded in a 24-well plate and transfections were performed after 24 h with CAPHOS, a calcium phosphate method, according to the manufacturer's guidelines. Cells were transfected with expression plasmids encoding the fusion protein Gal4-PPARa-LBD, Gal4-PPARg-LBD or Gal4-PPARdLBD (30 ng), pGal5TKpGL3 (100 ng), and pCMVbgal (250 ng). Four hours after transfection, cells were treated for 20 h with the indicated ligands in triplicate. Luciferase activity in cell extracts was then determined by a luminometer (VICTOR3 V Multilabel Plate Reader, PerkinElmer). b-Galactosidase activity was determined using ortho-nitro-phenyl-b-D-galactopyranoside as described previously [39]. All transfection experiments were repeated at least twice. Acknowledgments This work was financially supported by the Ministero dell’Is
e della Ricerca Scientifica e Tecnologica truzione, dell'Universita [MIUR-PRIN 2010-2011, grant 2010W7YRLZ_003 (A.L.)] and Cariplo Foundation [grant 2009.2727 (F.L.)]. Appendix A. Supplementary data Supplementary data related to this article can be found at http:// dx.doi.org/10.1016/j.ejmech.2014.10.083. References [1] D.L. Bain, A.F. Heneghan, K.D. Connaghan-Jones, M.T. Miura, Nuclear receptor structure: implications for function, Annu. Rev. Physiol. 69 (2007) 201e220.
gue, N. Renault, C. Furman, R. Millet, S. Yous, S. Speca, [2] C. Pirat, A. Farce, N. Lebe P. Berthelot, P. Desreumaux, P. Chavatte, Targeting peroxisome proliferatoractivated receptors (PPARs): development of modulators, J. Med. Chem. 55 (2012) 4027e4061. [3] G.S. Harmon, M.T. Lam, C.K. Glass, PPARs and lipid ligands in inflammation and metabolism, Chem. Rev. 111 (2011) 6321e6340. [4] B. Staels, J. Dallongeville, J. Auwerx, Mechanism of action of fibrates on lipid and lipoprotein metabolism, Circulation 98 (1998) 2088e2093. [5] A. Tesse, R. Andriantsitohaina, T. Ragot, PPARd activity in cardiovascular diseases: a potential pharmacological target, PPAR Res. (2009) 9. Article ID 745821.  jcik, L.A. Wo [6] A. Zieleniak, M. Wo zniak, Structure and physiological functions of the human peroxisome proliferator-activated receptor g, Arch. Immunol. Ther. Exp. 56 (2008) 331e345. [7] B. Cariou, B. Charbonnel, B. Staels, Thiazolidinediones and PPARg agonists: time for a reassessment, Trends Endocrinol. Metab. 23 (2012) 205e215.

B. De Filippis et al. / European Journal of Medicinal Chemistry 89 (2015) 817e825 [8] P.T. Cheng, R. Mukherjee, PPARs as targets for metabolic and cardiovascular diseases, Mini Rev. Med. Chem. 5 (2005) 741e753. [9] a) C. Fievet, J.-C. Fruchart, B. Staels, PPARa and PPARg dual agonists for the treatment of type 2 diabetes and the metabolic syndrome, Curr. Opin. Pharmacol. 6 (2006) 606e614; b) M. Heald, M.A. Cawthorne, Dual acting and pan-PPAR activators as potential anti-diabetic therapies, Handb. Exp. Pharmacol. 203 (2011) 35e51. [10] a) L.V. Sonawane, S.B. Bari, Ligand-based in silico 3D-QSAR study of PPAR-g agonists, Med. Chem. Res. 20 (2011) 1005e1014; b) A. Farce, N. Renault, P. Chavatte, Structural insight into PPARg ligands binding, Curr. Med. Chem. 16 (2009) 1768e1789. [11] a) H.E. Xu, M.H. Lambert, V.G. Montana, K.D. Plunket, L.B. Moore, J.L. Collins, J.A. Oplinger, S.A. Kliewer, R.T. Gampe Jr., D.D. McKee, J.T. Moore, T.M. Willson, Structural determinants of ligand binding selectivity between the peroxisome proliferator-activated receptors, Proc. Natl. Acad. Sci. U. S. A. 98 (2001) 13919e13924; b) S. Hidalgo-Figueroa, J.J. Ramírez-Espinosa, S. Estrada-Soto, J.C. Almanzarez, R. Rom n-Aguilar, J.V. Hern Pe an-Ramos, F.J. Alarco andez-Rosado, ~ o, G. Navarrete-Va zquez, Discovery of thiaH. Moreno-Díaz, D. Díaz-Coutin zolidine-2,4-dione/biphenylcarbonitrile hybrid as dual PPAR a/g modulator with antidiabetic effect: in vitro, in silico and in vivo approaches, Chem. Biol. Drug Des. 81 (2013) 474e483. [12] S. Azhar, Peroxisome proliferator-activated receptors, metabolic syndrome and cardiovascular disease, Future Cardiol. 6 (2010) 657e691. nardeau, M. Meyer, E. Sebokova, J. Mizrahi, [13] a) B.C. Hansen, X.T. Tigno, A. Be Effects of aleglitazar, a balanced dual peroxisome proliferator-activated receptor a/g agonist on glycemic and lipid parameters in a primate model of the metabolic syndrome, Cardiovasc. Diabetol. 10 (2011); b) A. Rubenstrunk, R. Hanf, D.W. Hum, J.-C. Fruchart, B. Staels, Safety issues and prospects for future generations of PPAR modulators, Biochim. Biophys. Acta 1771 (2007) 1065e1081; c) A. Grey, Skeletal consequences of thiazolidinedione therapy, Osteoporos. Int. 19 (2008) 129e137; d) B.G. Shearer, A.N. Billin, The next generation of PPAR drugs: do we have the tools to find them? Biochim. Biophys. Acta 1771 (2007) 1082e1093. [14] a) C.S. Mizuno, G. Ma, S. Khan, A. Patny, M.A. Averyc, A.M. Rimando, Design, synthesis, biological evaluation and docking studies of pterostilbene analogs inside PPARa, Bioorg. Med. Chem. 16 (2008) 3800e3808; b) P. Iannelli, V. Zarrilli, E. Varricchio, D. Tramontano, F.P. Mancini, The dietary antioxidant resveratrol affects redox changes of PPARa activity, Nutr. Metab. Cardiovasc. Dis. 17 (2007) 247e256; c) W. Li, X. He, W. Shi, H. Jia, B. Zhong, Pan-PPAR agonists based on the resveratrol scaffold: biological evaluation and docking studies, ChemMedChem 5 (2010) 1977e1982. [15] a) B. De Filippis, A. Giancristofaro, A. Ammazzalorso, A. D'Angelo, M. Fantacuzzi, L. Giampietro, C. Maccallini, M. Petruzzelli, R. Amoroso, Discovery of gemfibrozil analogues that activate PPARa and enhance the expression of gene CPT1A involved in fatty acids catabolism, Eur. J. Med. Chem. 46 (2011) 5218e5224; b) L. Giampietro, A. D'Angelo, A. Giancristofaro, A. Ammazzalorso, B. De Filippis, M. Fantacuzzi, P. Linciano, C. Maccallini, R. Amoroso, Synthesis and structure-activity relationships of fibrate-based analogues inside PPARs, Bioorg. Med. Chem. Lett. 22 (2012) 7662e7666. [16] P.R. Skaanderup, C.S. Poulsen, L. Hyldtoft, M.R. Jørgensen, R. Madsen, Regioselective conversion of primary alcohols into iodides in unprotected methyl furanosides and pyranosides, Synthesis (2002) 1721e1727. [17] a) J.M. Lehmann, L.B. Moore, T.A. Smith-Oliver, W.O. Wilkinson, T.M. Willson, S.A. Kliewer, An antidiabetic thiazolidinedione is a high affinity ligand for peroxisome proliferator-activated receptor g (PPARg), J. Biol. Chem. 270 (1995) 12953e12956; b) A. Pinelli, C. Godio, A. Laghezza, N. Mitro, G. Fracchiolla, V. Tortorella, A. Lavecchia, E. Novellino, J.-C. Fruchart, B. Staels, M. Crestani, F. Loiodice, Synthesis, biological evaluation, and molecular modeling investigation of new chiral fibrates with PPARa and PPARg agonist activity, J. Med. Chem. 48 (2005) 5509e5519; c) L. Porcelli, F. Gilardi, A. Laghezza, L. Piemontese, N. Mitro, A. Azzariti, F. Altieri, L. Cervoni, G. Fracchiolla, M. Giudici, U. Guerrini, A. Lavecchia, R. Montanari, C. Di Giovanni, A. Paradiso, G. Pochetti, G.M. Simone, P. Tortorella, M. Crestani, F. Loiodice, Synthesis, characterization and biological evaluation of ureidofibrate-like derivatives endowed with peroxisome proliferator-activated receptor activity, J. Med. Chem. 55 (2012) 37e54. [18] O. Trott, A.J. Olson, AutoDock Vina: improving the speed and accuracy of docking with a new scoring function, efficient optimization, and multithreading, J. Comput. Chem. 31 (2010) 455e461. nyi, D. van der Spoel, Blind docking of drug-sized compounds to [19] C. Hete proteins with up to a thousand residues, FEBS Lett. 580 (2006) 1447e1450. [20] a) B.R. Henke, S.G. Blanchard, M.F. Brackeen, K.K. Brown, J.E. Cobb, J.L. Collins, W.W. Harrington Jr., M.A. Hashim, E.A. Hull-Ryde, I. Kaldor, N-(2Benzoylphenyl)-L-tyrosine PPARgamma agonists. 1. Discovery of a novel series of potent antihyperglycemic and antihyperlipidemic agents, J. Med. Chem. (1998) 5020e5036; b) J.L. Collins, S.G. Blanchard, G.E. Boswell, P.S. Charifson, J.E. Cobb, B.R. Henke, E.A. Hull-Ryde, W.M. Kazmierski, D.H. Lake, L.M. Leesnitzer, N-(2benzoylphenyl)-L-tyrosine PPARg agonists. 2. Structure-activity relationship

[21]

[22]

[23]

[24]

[25]

[26] [27] [28] [29]

[30] [31]

[32]

[33]

[34] [35] [36] [37] [38]

[39]

825

and optimization of the phenyl alkyl ether moiety, J. Med. Chem. (1998) 5037e5054; c) J.E. Cobb, S.G. Blanchard, E.G. Boswell, K.K. Brown, P.S. Charifson, J.P. Cooper, J.L. Collins, M. Dezube, B.R. Henke, E.A. Hull-Ryde, N-(2-benzoylphenyl)-Ltyrosine PPARg agonists. 3. Structure-activity relationship and optimization of the N-aryl substituent, J. Med. Chem. (1998) 5055e5069. R.T.J. Gampe, V.G. Montana, M.H. Lambert, A.B. Miller, R.K. Bledsoe, M.V. Milburn, S.A. Kliewer, T.A. Willson, E.H. Xu, Asymmetry in the PPARgamma/RXR-alpha structure reveals the molecular basis of heterodimerization among nuclear receptors, Mol. Cell (2000) 545e555. a) B.C. Kallenberger, J.D. Love, V.K. Chatterjee, J.W. Schwabe, A dynamic mechanism of nuclear receptor activation and its perturbation in a human disease, Nat. Struct. Biol. 10 (2003) 136e140; b) Y. Hamuro, S.J. Coales, J.A. Morrow, K.S. Molnar, S.J. Tuske, M.R. Southern, P.R. Griffin, Hydrogen/deuterium-exchange (H/D-Ex) of PPARg LBD in the presence of various modulators, Protein Sci. 15 (2006) 1883e1892. J.B. Bruning, M.J. Chalmers, S. Prasad, S.A. Busby, T.M. Kamenecka, Y. He, K.W. Nettles, P.R. Griffin, Partial agonists activate PPARgamma using a helix 12 independent mechanism, Structure 15 (2007) 1258e1271. a) G. Tiberio, L. Muccioli, R. Berardi, C. Zannoni, How does the trans-cis photoisomerization of azobenzene take place in organic solvents? ChemPhysChem 11 (2010) 1018e1028; b) N. Nishimura, T. Sueyoshi, H. Yamanaka, E. Imai, S. Yamamoto, S. Hasegawa, Thermal cis-to-trans isomerization of substituted azobenzenes II. Substituent and solvent effects, Bull. Chem. Soc. Jpn. 49 (1976) 1381e1387; c) R.J.W. Le Fevre, J. Northcott, The effects of substituents and solvents on the cis-trans change of azobenzene, J. Chem. Soc. (1953) 867e870; d) G.S. Hartley, The cis-form of azobenzene and the velocity of the thermal cis/trans-conversion of azobenzene and some derivatives, J. Chem. Soc. (1938) 633e642; e) W.R. Brode, J.H. Gould, G.M. Wyman, The relation between the absorption spectra and the chemical constitution of dyes. Phototropism and cis-trans isomerism in aromatic azo compounds, J. Am. Chem. Soc. 74 (1952) 4641e4646; f) W.R. Brode, J.H. Gould, G.M. Wyman, The relation between the absorption spectra and the chemical constitution of dyes. Effect of solvent and of temperature on the cis-trans isomerization of azo dyes, J. Am. Chem. Soc. 75 (1953) 1856e1859; g) P.P. Birnbaum, D.W.G. Style, The photo-isomerization of some azobenzene derivatives, Trans. Farad. Soc. 50 (1954) 1192e1196; h) G. Gabor, E. Fischer, Spectra and cis-trans isomerism in highly bipolar derivatives of azobenzene, J. Phys. Chem. 75 (1971) 581e583; i) H.M. Dhammika Bandara, S.C. Burdette, Photoisomerization in different classes of azobenzene, Chem. Soc. Rev. 41 (2012) 1809e1825. a) S. Eris¸kin, N. S¸ener, S. Yavuz, I. S¸ener, Synthesis, characterization, and biological activities of 4-imino-3-arylazo-4H-pyrimido[2,1-b][1,3]benzothiazole-2-oles, Med. Chem. Res. 23 (2014) 3733e3743; b) A.A. Deeg, M.S. Rampp, A. Popp, B.M. Pilles, T.E. Schrader, L. Moroder, K. Hauser, W. Zinth, Isomerization- and temperature-jump-induced dynamics of a photoswitchable b-hairpin, Chem. Eur. J. 20 (2014) 694e703. K.-T. Chung, S.E. Stevens, C.E. Cerniglia, The reduction of azo dyes by the intestinal microflora, Crit. Rev. Microbiol. 18 (1992) 175e190. W.A. Velema, W. Szymanski, B.L. Feringa, Photopharmacology: beyond proof of principle, J. Am. Chem. Soc. 136 (2014) 2178e2191. €dinger, LLC, New York, NY, 2014. Maestro, Version 9.7, Schro C.C. Huang, G.S. Couch, E.F. Pettersen, T.E. Ferrin, Chimera: an extensible molecular modelling application constructed using standard components, Pac. Symp. Biocomput. 1 (1996) 724. T. Warren, L. DeLano, The PyMOL Molecular Graphics System, DeLano Scientific LLC, San Carlos, CA, USA. http://www.pymol.org/. H.M. Berman, T. Battistuz, T.N. Bhat, W.F. Bluhm, P.E. Bourne, K. Burkhardt, Z. Feng, G.L. Gilliland, L. Iype, S. Jain, P. Fagan, J. Marvin, D. Padilla, V. Ravichandran, B. Schneider, N. Thanki, H. Weissig, J.D. Westbrook, C. Zardecki, The protein data bank, Acta Crystallogr. D. Biol. Crystallogr. 58 (2002) 899e907. G.M. Sastry, M. Adzhigirey, T. Day, R. Annabhimoju, W. Sherman, Protein and ligand preparation: parameters, protocols, and influence on virtual screening enrichments, J. Comput. Aid. Mol. Des. 27 (2013) 221e234. Z. Song, F.A. Kovacs, J. Wang, J.K. Denny, S.C. Shekar, J.R. Quine, T.A. Cross, Transmembrane domain of M2 protein from influenza A virus studied by solid-state (15)N polarization inversion spin exchange at magic angle NMR, Biophys. J. 79 (2000) 767e775. € dinger Release 2014-1: LigPrep, Version 2.9, Schro €dinger, LLC, New York, Schro NY, 2014. €dinger, LLC, New York, NY, 2013. MacroModel, Version 10.2, Schro http://mgltools.scripps.edu. G.R. Desiraju, T. Steiner, The Weak Hydrogen Bond: in Structural Chemistry and Biology, 9, Oxford University Press, New York, 2001. E. Raspe, L. Madsen, A.M. Lefebvre, I. Leitersdorf, L. Gelman, J. PeinadoOnsurbe, J. Dallongeville, J.-C. Fruchart, R. Berge, B. Staels, Modulation of rat liver apolipoprotein gene expression and serum lipid levels by tetradecylthioacetic acid (TTA) via PPARa activation, J. Lipid Res. 40 (1999) 2099e2110. T. Hollon, F.K. Yoshimura, Variation in enzymatic transient gene expression assays, Anal. Biochem. 182 (1989) 411e418.