Conception of myeloperoxidase inhibitors derived from flufenamic acid by computational docking and structure modification

Available online at www.sciencedirect.com

Bioorganic & Medicinal Chemistry 16 (2008) 1702–1720

Conception of myeloperoxidase inhibitors derived

from ﬂufenamic acid by computational docking

and structure modiﬁcation

a

b

c

´

Pierre Van Antwerpen, Martine Prevost, Karim Zouaoui-Boudjeltia, Sajida Babar,

I. Legssyer,^cPatrick Moreau,^aNicole Moguilevsky,^dMichel Vanhaeverbeek,^c

e

a

a,

*

`

Jean Ducobu, Jean Neve and Franc¸ois Dufrasne

^aLaboratory of Pharmaceutical Organic Chemistry, Institut de Pharmacie, Universite´ Libre de Bruxelles, 1050 Brussels, Belgium

^bLaboratory of Structure and Function of Biological Membranes, Universite´ Libre de Bruxelles, 1050 Brussels, Belgium

^cLaboratory of Experimental Medicine, CHU Charleroi Hospital A. Ve´sale, Universite´ Libre de Bruxelles,

6110 Montigny-le-Tilleul, Belgium

^dTechnology Transfer Oﬃce, FUNDP, 5000 Namur, Belgium

`

Department of Internal Medicine, CHU Tivoli, Universite´ Libre de Bruxelles, 7100 La Louviere, Belgium

e

Received 5 September 2007; revised 30 October 2007; accepted 7 November 2007

Available online 13 November 2007

Abstract—The development of myeloperoxidase (MPO) inhibitors has been conducted using ﬂufenamic acid as a lead compound.

Computational docking of the drug and its analogs in the MPO active site was ﬁrst attempted. Several molecules were then synthe-

sized and assessed using three procedures for the measurement of their inhibiting activity: (i) the taurine assay, (ii) the accumulation

of compound II, and (iii) the LDL oxidation by ELISA. Most of the synthesized molecules had an activity in the same range as

ﬂufenamic acid but none of them were able to inhibit the MPO-dependent LDL oxidation. The experiments however gave some

useful indications for a rational conception of MPO inhibitors.

1. Introduction

matory syndromes, atherosclerosis, end-stage renal dis-

ease or neurodegenerative diseases.^2–4

Myeloperoxidase (MPO) is a key enzyme belonging to

the complex defense system against exogenous aggres-

sions. It is mostly localized in the azurophil (primary)

granules of neutrophils and, in minor proportions, in

monocytes. Its role is to ensure the production of hypo-

chlorous acid (HOCl) in presence of hydrogen peroxide

(H₂O₂) and chlorine anions (Cl^ꢀ). During phagocytosis,

the MPO/H₂O₂/Cl^ꢀsystem promotes the elimination of

microorganisms by HOCl-mediated oxidations.¹How-

ever, MPO is liable to be released in extracellular ﬂuids

during an oxidative stress and this so-called ‘circulating’

MPO has been suggested to be involved in diﬀerent

pathogenic conditions such as those of chronic inﬂam-

In this context, several authors have pointed out that the

development of eﬃcient MPO inhibitors could be a

valuable way to block the potentially harmful enzymatic

activity. Some members of our research group already

devoted attention to the involvement of MPO in athero-

sclerosis. Indeed, the enzyme is able to bind to LDL and

speciﬁcally oxidize apolipoprotein B-100 (ApoB-100), a

mechanism generally admitted as a pro-atherogenic fac-

tor.^5,6For example, the oxidized LDL (Mox-LDL) in-

duces pro-inﬂammatory responses in endothelial cells

and monocytes.⁷Furthermore, MPO can oxidize the

apolipoprotein AI of HDL, leading to a decrease in

the cholesterol eﬄux.⁸However, the development of

MPO inhibitors has so far been undertaken by a few

groups with only a limited success.^9,10

Keywords: Myeloperoxidase; Flufenamic acid; Molecular docking;

LDL; Chronic inﬂammatory diseases.

In the present paper, ﬂufenamic acid (1, (2-[(3-triﬂuoro-

methylphenyl)amino]benzoic acid, Fig. 1)) was selected

*

Corresponding author. Tel.: +32 26505235; fax: +32 26505249;

e-mail: dufrasne@ulb.ac.be

doi:10.1016/j.bmc.2007.11.025

P. Van Antwerpen et al. / Bioorg. Med. Chem. 16 (2008) 1702–1720

1703

a success rate of about 80%. To assess the reliability of

the selected procedure, salicylhydroxamic acid (SHA),

cyanide (CN^ꢀ), and thiocyanate (SCN^ꢀ) ligands were

docked in the binding site of MPO after removal of all

water molecules and ligands. The docked poses were

compared with the positions of the crystal ligands by

computing the root-mean-square deviation (rmsd)

which measures the distance between them. A crystal

structure of SHA bound to MPO was published in

1996.¹⁴The structure was not deposited in the PBD

but was made available to us by one of the authors (Da-

vey C., private communication). By comparing the pre-

dicted poses with the crystal structure and its description

in the publication (Fig. 2), a rmsd of the atomic posi-

tions of the low-energy pose and the experimental struc-

COOH

H

N

CF₃

1

Figure 1. Flufenamic acid structure.

as a lead compound in a process aiming to design MPO

inhibitors. This drug that has anti-inﬂammatory proper-

ties linked to cyclooxygenase (COX) inhibition has in-

deed been reported as an eﬃcient inhibitor of the

chlorinating activity of MPO at a micromolar range

during the screening of compounds belonging to the

same therapeutic class.¹¹Van Antwerpen et al.

conﬁrmed this observation and established that the

molecule was directly oxidized by the enzyme in its 5-hy-

droxy- and 5-chloro-derivatives.¹²As several X-ray

crystallographic structures for MPO with various

ligands are now available in the scientiﬁc literature, it

becomes possible to investigate the way ﬂufenamic acid

inhibits MPO activity and prevents the interaction of

H₂O₂and/or Cl^ꢀwith the heme moiety.^13,14

˚

ture of 0.49 A was obtained, and indicated a high

similarity. In accordance with the experimental struc-

ture, the hydrophobic moiety of SHA binds in a pocket

lined by the pyrrole ring D, Arg239 and the Phe99, 366

and 407. The plane of the aromatic ring is almost paral-

lel to the heme pyrrole ring. This aromatic moiety also

interacts with the neighboring Phe99, 366, 407 whose

˚

distances vary between 3.9 and 4.8 A. The hydroxamic

acid group lies at the center of the distal cavity between

the distal His95 and the heme iron. The carbonyl and

hydroxyl oxygens form hydrogen bonds with the pro-

tein. The carbonyl oxygen hydrogen bonds to Gln91

side chain and the hydroxyl oxygen of the hydroxamic

function forms a hydrogen bond with His95. The sali-

cylic ring hydroxyl group hydrogen bonds to Arg239

side chain. All these interactions are in agreement with

the experimental description of the crystal structure.

Interestingly, the oxygens of the two hydroxyl and of

the carbonyl groups occupy the same positions as the

water molecules in the native structure (PDB code:

A rational drug design process based on the three-

dimensional (3D) structure of the target gives the possi-

bility to design and evaluate large series of ligands in

terms of aﬃnity. X-ray crystal structures of ligand–pro-

tein complexes along with computational analysis tools

have been successful in the discovery and optimization

of drug candidates¹⁵and a large number of structure-

based drug design programs have been developed in

the last 20 years based on a variety of search algorithms

according to two main approaches: the fragment-based

design and the full-molecule docking. The latest often

resorts to compound libraries which are screened for li-

gands in agreement with the binding site requirements,

while the ﬁrst one uses small molecules with few chemi-

cal functionalities positioned in the receptor binding

site, which are subsequently grown or assembled into

an entire molecule. As the structure of human MPO

complexed with cyanide and thiocyanate has already

been resolved to 1.9 A by X-ray crystallography,¹³com-

˚

putational docking of ﬂufenamic acid and analogs was

performed in the MPO active site after validation of

the docking procedure. Several molecules were then syn-

thesized and assessed, using three procedures able to

measure the potential activity of the inhibitors: (i) the

taurine assay,^11,12(ii) the accumulation of compound

II,^11,12and (iii) the LDL oxidation by ELISA.¹⁶

2. Results and discussion

2.1. Docking

Figure 2. Best pose of the SHA molecule generated by the Glide^ꢁ

docking program. For sake of comparison the experimental structure

of SHA (Davey CA, private communication and Ref. 26) is also

shown. The ligands are depicted as sticks (the experimental structure

with gray carbons and the docked structure with green carbons);

neighboring residues and the heme are represented as thinner lines. The

small blue sphere indicates the location of the Fe ion at the center of

the heme.

2.1.1. Validation. The software Glide^ꢁis widely used to

dock complete molecules into a protein binding site. It

has been shown to reproduce experimentally binding

poses for a large number of protein targets.¹⁷However,

the most successful docking programs generally produce

1704

P. Van Antwerpen et al. / Bioorg. Med. Chem. 16 (2008) 1702–1720

1CXP). Only one docking pose for CN^ꢀis predicted by

Glide. This position reproduces well the crystal position

two crystal water molecules in the native structure. In

the latter pose, the stacking with the pyrrole D ring is

more pronounced with an almost face-to-face stacked

interaction.

of CN^ꢀwith a rmsd of 0.37 A. In contrast, Glide was

˚

not able to reproduce the crystal position of SCN^ꢀ.

The docked poses are located at the crystal positions

of CN^ꢀinstead of that of SCN^ꢀ. To more accurately

In an attempt to generate a ﬂufenamic acid derivative

featuring a better aﬃnity, an ortho-diacid compound

(compound 2) was designed. The rationale behind this

compound arises from the poses obtained for ﬂufenamic

acid and from the observation that MPO binding site

can accommodate two negatively charged small ligands,

namely CN^ꢀand SCN^ꢀ. Another reason for proposing

compound 2 comes from the fact that SHA, which binds

to the MPO active site, contains two adjacent hydrogen

bond acceptors. The diacid compound was thus docked

as three forms which diﬀer by the protonation state of

one of the two acidic functions. The ﬁrst two forms bear

a single negative charge. But one adopts several geome-

tries in which the carboxylate group matches the posi-

tion of the crystal SCN^ꢀand the carboxylic acid that

of the CN^ꢀ. In these poses, three oxygens of the carbox-

ylic functions occupy the positions of three crystal water

molecules. The other strikingly diﬀers with the carboxyl-

ate located between SCN^ꢀand CN^ꢀcrystal positions

and the carboxylic acid group hydrogen bonded to one

of the propionate functions of the heme. The poses of

the third form, which comprises two negative charges,

ﬁt very well the charges of SCN^ꢀand CN^ꢀand exhibits

a better aﬃnity than the other two forms (Table 1).

ꢀ

˚

reproduce the crystal position of SCN (rmsd = 0.77 A),

the ligand had to be docked in the presence of two crys-

tal water molecules present in the crystal structure of

protein–SCN^ꢀcomplex (water molecules numbered

209 and 210 in the crystal structure; PDB code:

1DNU). Thus, the binding mode of SHA and CN^ꢀ

was identiﬁed by docking as well as that of SCN^ꢀpro-

viding the presence of two crystal water molecules.

The docking procedure could therefore be applied to ﬂu-

fenamic acid and analogs.

2.1.2. Docking of ﬂufenamic acid and analogs. Eight out

of 10 poses of ﬂufenamic acid in the MPO binding site

correspond to a conﬁguration in which one of the oxy-

gens of the carboxyl group of ﬂufenamic acid matches

the crystal position of the sulfur atom in SCN^ꢀ, featur-

ing the coincidence of the negative charge of the two li-

gands (Fig. 3A). The aromatic moiety of the ﬂufenamic

acid, bearing the carboxylic group, features an oﬀset

stacked geometry with the pyrrole D cycle of the heme

and one edge-to-face geometry with Phe99. In these

poses, the oxygen of the carboxyl group occupies the po-

sition of one crystal water molecule in the native struc-

ture. One slightly less energetically favorable pose, due

to a few unfavorable van der Waals contacts, corre-

sponds to a location of the carboxyl group shifted more

toward the CN^ꢀin which the oxygens of the carboxylate

group occupy the crystal positions of the nitrogens of

CN^ꢀand SCN^ꢀ, respectively (Fig. 3B). Interestingly in

that pose, both oxygens reside close to the positions of

Flufenamic acid derivatives, featuring a halogen atom

(Cl, F, Br) or a hydroxyl group at the position 5 (com-

pounds 3a–d), have also been docked in the binding site.

This series was designed from the observation that a

molecular surface calculated on the crystal structure re-

veals a small pocket occupied by a crystal water mole-

Figure 3. Two best poses of the ﬂufenamic acid molecule. In the ﬁrst pose (A) the negative charge of the carboxylic acid moiety matches that of the

SCN^ꢀin the complexed structure (PDB code: 1DNW). In the second pose (B) the negative charge is located between the positions of SCN^ꢀand CN^ꢀ

and the two oxygens of the carboxylic function match the locations of two crystal water molecules. The ligand is depicted as sticks; neighboring

residues and the heme are represented as thinner lines. The small blue sphere indicates the location of the Fe ion at the center of the heme.

P. Van Antwerpen et al. / Bioorg. Med. Chem. 16 (2008) 1702–1720

1705

Table 1. DG is the predicted free energy of binding from docking experiments; m/z is the ratio of the mass (m) and charge (z); Acc II is the relative

value of compound II lifetime in the presence and the absence of chloride, compared to ﬂufenamic, acid and the last column lists the experimental

IC₅₀for the inhibition of the MPO/H₂O₂/Cl^ꢀsystem

Compound

DG (kcal/mol)

m/z

Acc II

Without Cl^ꢀ

IC₅₀(lM)

Cl^ꢀ

1

Flufenamic acid (1)

Diacid (2)

ꢀ4.9

ꢀ6.8 to ꢀ4

ꢀ5.7

ꢀ4.4

ꢀ5.0

281

1

1.8 0.1

>100

184 or 369

—

Compound 3a (5OH^ꢀ)

Compound 3b (5F^ꢀ)

Compound 3c (5Cl^ꢀ)

Compound 3d (5Br^ꢀ)

297

298

0

n.d.

2.21

2.42

1.01

0.91

0.92

0.97

1.8 0.5

1.6 0.6

1.8 0.8

315.5

360

Compound 4a

Compound 4b

ꢀ4.2

326

296

7.05

0

8.26

0

5

n.d.

1

ꢀ5.9

Compound 5a

Compound 5b

Compound 5c

Compound 5d

Compound 5e

ꢀ5.0

ꢀ4.6

ꢀ7.7

ꢀ5.9

ꢀ5.7

338

304

197

382

381

0.42

0.2

1.00

0.79

2.53

0.35

0.63

14

2

58 15^*

>100

0.36

0.10

0.72

>100

5.5 0.8

n.d. means that the value could not be calculated as a prooxidant eﬀect was observed. ^*Signiﬁcantly diﬀerent from ﬂufenamic acid (P < 0.001,

Bonferroni t-test). The range in DG given for the diacid compound corresponds to the aﬃnity values obtained for the three forms (see text).

cule and shaped by one of the propionate group of the

heme, Arg333, and Arg424. The geometry of interaction

of those four compounds is close to that of the ﬂufen-

amic acid molecule with a coincidence of one of the car-

boxyl oxygens onto the location of the negative charge

of SCN^ꢀ(Fig. 4). The halogen atom or the hydroxyl

group ﬁts well into the pocket previously described

(Fig. 4). Only one pose obtained for the ﬂuoro com-

pound is similar to the CN^ꢀ-shifted pose obtained with

ﬂufenamic acid. This might be due to the size of F com-

pared to Cl, Br, and OH.

The compound 4 and 5 series were designed following

the observation that the para position in most of the ﬂu-

fenamic acid poses faces a rather large empty pocket.

The idea was then to substitute at this para position dif-

ferent groups of variable lengths so as to provide addi-

tional favorable interactions with the protein. The

derivative of ﬂufenamic acid with the nitro and amino

groups at the para position of the carboxyl group (com-

pounds 4a and 4b) docked into the binding site adopts

geometries again close to those of ﬂufenamic acid.

Two other poses feature a geometry shifted to CN^ꢀwith

a correspondence of the two oxygens of the carboxylate

group onto two crystal water molecules closer to CN^ꢀ.

The two oxygens of the nitro group ﬁt onto two other

crystal water molecules. The amino group in compound

4a makes a hydrogen bond with one propionate group

of the heme and with Glu102.

Compound 5 series was designed by substituting longer

chains at the para position of ﬂufenamic acid. The ratio-

nale behind it comes from the observation that residues

with diﬀerent physicochemical properties line the large

pocket aforementioned and thus that longer chains,

bearing diﬀerent functional groups at their extremities,

could interact favorably with some of these residues.

Two types of geometries adopted by compounds 5 can

be identiﬁed in the binding site. One class of geometries

corresponds closely to the geometries of ﬂufenamic acid,

that is, most of the poses position the carboxylate group

onto SCN^ꢀ. Compounds 5a, 5b, and 5e belong to this

class. Compound 5e however features a larger variety

of poses than the other two. One reason might be that

Figure 4. Best pose of compound 3c. The ligand is depicted as sticks;

neighboring residues and the heme are represented as thinner lines. The

small blue sphere indicates the location of the Fe ion at the center of

the heme. For sake of comparison the SCN^ꢀand CN^ꢀions from the

complexed structure (PDB code: 1DNW) are also shown and depicted

as sticks.

þ

this molecule bears a NH₃at one extremity. In all

þ

poses, the NH₃group of the ligand forms a salt bridge

with the side chain of Glu116 which seems to inﬂuence

1706

P. Van Antwerpen et al. / Bioorg. Med. Chem. 16 (2008) 1702–1720

the position of the benzoic acid moiety. In 5 out of 10

poses, the geometry of interaction for the carboxylate

group of compound 5e matches that of the ﬂufenamic

acid. The other class of geometries adopted by com-

pounds 5c and 5d strongly diﬀers from those of the other

ﬂufenamic acid derivatives. For compound 5d, the OH

group matches the location of the carbon of the CN^ꢀ

while the carbonyl oxygen coincides with the crystal po-

sition of SCN^ꢀin 6 out of 10 poses. Both oxygens match

the locations of two crystal water molecules in the native

structure. In these poses, the OH group hydrogen bonds

to His95. This geometry for the OH group agrees with

the location expected for H₂O₂.¹⁸For compound 5c,

all poses correspond to geometry with the negatively

charged carboxylate of the oleﬁnic chain matching that

of CN^ꢀ. In these geometries, the carboxylate group of

the oleﬁnic chain hydrogen bonds to Gln91 or Arg239.

2.2. Chemistry

Methods used for the synthesis of ﬂufenamic acid deriv-

atives (Fig. 5) were mostly based on Hartwig–Buchwald

coupling reactions.¹⁹As an example for the diacids, the

synthesis of compound 2 started from 2-nitrophthalic

acid (6, Scheme 1). Since esteriﬁcation using SOCl₂in

MeOH failed,²⁰most probably as a consequence of the

presence of a nitro group, a nucleophilic substitution

with MeI was carried out, which was only feasible using

DBU and not weaker bases such as K₂CO₃.²¹After

reduction of the nitro group, the aminophthalate 8

COOH

H

COOH

H

N

CF₃

HOOC

N

CF₃

R

2

3a: R =- OH

3b: R = -F

3c: R = -Cl

3d: R = -Br

4a: R = -NO₂

4b: R = -NH₂

COOH

H

N

CF₃

5a: R = -CH₃

5b: R = -C₆H₅

5c: R = -CH=CHCOOH

5d: R = -(CH₂)₃OH

HN

O

5e: R = -(CH₂)₃NH₂

R

Figure 5. Structures of synthesized compounds.

COOH

COOCH₃

HOOC

NO₂

H₃COOC

i

NO₂

H₃COOC

NH₂

ii

6

7

8

COOCH₃

Br

COOCH₃

H

H₃COOC

N

CF₃

iii

iv

8

9

10

COOH

H

HOOC

N

CF₃

v

10

2

Scheme 1. Synthesis of compound 2. Reagents: (i) MeI, DBU, MeCN; (ii) H₂, Pd/C 5%; (iii) NaNO₂, HCl, CuBr; (iv) 3-triﬂuoromethylaniline,

Pd(CH₃COO)₂, DPePhos, Cs₂CO₃; (v) KOH, EtOH, H₂O.

P. Van Antwerpen et al. / Bioorg. Med. Chem. 16 (2008) 1702–1720

1707

was subjected to a Sandmeyer reaction and transformed

into the bromo derivatives 9.^22,23The Hartwig–Buch-

wald coupling of 9 with 3-triﬂuoromethylaniline deliv-

ered the ester 10 which was further hydrolyzed into

the phthalic acid derivative 2.

anol because acetic acid and dioxane gave higher

proportions of other bromination products.

The series of ﬂufenamic acid derivatives with substitu-

ents at position 4 of the anthranilic acid ring was ob-

tained from a common precursor 22 (Scheme 5) with

2-bromo-4-nitrotoluene (19) as the starting material.

After oxidation of the methyl group in a carboxylic

acid,²⁶followed by esteriﬁcation via a nucleophilic sub-

stitution with MeI,²¹a C–N coupling was carried out to

obtain the nitro derivative 22. A fraction of this com-

pound was hydrolyzed to the acid 4a and after a cata-

lytic hydrogenation, the resulting amine was

transformed into the amino-acid 4b. From compound

23, a series of ﬁve compounds was obtained. The ﬁrst

three include the amides 5a–c which were prepared by

amidiﬁcation with the corresponding acid chlorides in

a mixture of pyridine/CH₂Cl₂, followed by alkaline

hydrolysis. The synthesis of compounds 5d and 5e was

carried out from the bromo derivative 27, either by reac-

tion with potassium acetate and hydrolysis of both ester

functions or by nucleophilic substitution with NaN₃,

catalytic hydrogenation, and ﬁnally hydrolysis.

For the synthesis of compound 3a, methyl gentisate (11)

was used as a starting compound (Scheme 2). The ﬁrst

step was a regioselective protection of the phenol at po-

sition 5 by a benzyl group.²⁴The remaining phenol was

then activated by transformation into a triﬂate which

was subjected to the Hartwig–Buchwald coupling. The

ester was subsequently hydrolyzed and the benzyl pro-

tecting group removed by hydrogenolysis.²⁵The method

used for the syntheses of compounds 3b and 3c was sim-

ilar but started from the corresponding methyl 5-halo-2-

hydroxybenzoate (Scheme 3). The synthesis of 5-bromo

ﬂufenamic acid, derivative 3d, appeared more compli-

cated as it could not be undertaken from methyl 5-bro-

mo-2-hydroxybenzoate (Scheme 4). Indeed, the

Hartwig–Buchwald coupling with the triﬂate derivative

gave a complex mixture of compounds which could be

attributed to the higher reactivity of the bromoarenes

in comparison with the corresponding chloro- and ﬂuo-

roarenes. As the Sandmeyer reaction on paraphenylen-

ediamines was not possible due to the high

oxidizability of the compounds, the direct bromination

of ﬂufenamic acid with bromine was attempted using

several conditions (Br₂in acetic acid, methanol or diox-

ane, respectively). The best solvent appeared to be meth-

2.3. Biological activity

The biological activity of ﬂufenamic acid and its analogs

was ﬁrst assessed by the inhibition of the chlorinating

activity of MPO with the taurine assay. The amount

of MPO was chosen to produce 60 lM of HOCl in

COOCH₃

OH

CF₃

O

OH

S

i

ii

O

HO

O

11

12

13

COOCH₃

COOH

H

N

CF₃

iv

N

CF₃

O

iii

13

15

14

COOH

H

N

CF₃

v

15

HO

3a

Scheme 2. Synthesis of compound 3a. Reagents: (i) benzyl bromide, K₂CO₃, acetone; (ii) (CF₃SO₂)₂O, TEA, CH₂Cl₂; (iii) 3-triﬂuoromethylaniline,

Pd(CH₃COO)₂, DPePhos, Cs₂CO₃; (iv) KOH, EtOH, H₂O; (v) H₂, HCl, Pd/C (5%).

1708

P. Van Antwerpen et al. / Bioorg. Med. Chem. 16 (2008) 1702–1720

COOCH₃

OH

COOCH₃

H

CF₃

O

N

i

ii

S

O

R

16a : R = F

16b: R = Cl

18a: R = F

18b: R = Cl

17a: R = F

17b: R = Cl

iii

COOH

H

N

CF₃

R

3b: R = F

3c: R = Cl

Scheme 3. Synthesis of compounds 3b and 3c. Reagents: (i) (CF₃SO₂)₂O, TEA, CH₂Cl₂; (ii) 3-triﬂuoromethylaniline, Pd(CH₃COO)₂, DPePhos,

Cs₂CO₃; (iii) KOH, EtOH, H₂O.

COOH

H

N

H

CF₃

N

CF₃

i

Br

3d

1

Scheme 4. Synthesis of compound 3d. Reagents: (i) Br₂, MeOH.

5 min at 37 ꢂC in the presence of 100 lM of H₂O₂and

the amount of the produced taurine chloramine was

measured by the thionitrobenzoic acid (TNB) reagent

as previously described.^11,12The percentage of the inhi-

bition was calculated taking the blank as 100% of inhi-

bition and the chlorinating activity in absence of the

compounds as the 0% of inhibition. The concentration

that inhibits at least 50% (IC₅₀) of the MPO activity

was calculated by establishing sigmoid curves of the

MPO activity as a function of compound concentration.

The curves were ﬁtted with a three-parameter sigmoid

model. The scavenging activity of compounds was

simultaneously assessed by replacing the MPO system

with 60 lM of HOCl and by measuring the produced

taurine chloramines with the same procedure.

and the accumulation of compound II, a reduced form

of compound I (Fe^IV@Oꢁ ꢁ ꢁPorꢁ ꢁ ꢁaa).^27–30As a conse-

quence, the accumulation of compound II is a semi-quan-

titative spectral method suitable to document the

inhibiting eﬀect of the molecules by using the speciﬁc spec-

tral patterns of MPO (k_max430 and 575 nm), compound I

(k_max430 nm), and compound II (k_max456 and 625 nm).

As the MPO system is relatively complex, it is of impor-

tance to classify and separate molecules acting as electron

donors (AH) with the oxidized forms of the enzyme (com-

pounds I and II) or as antagonists of chloride in the cata-

lytic site of MPO. It is generally admitted that a molecule

which rapidly converts compounds I and II into native

MPO is an electron donor (AH) that quickly consumes

H₂O₂and also generates unwanted oxidized by-products

during the reaction.¹²However, a molecule that rapidly re-

duces compound I but slowly reacts with compound II is

considered as a chloride antagonist that promotes the

accumulation of compound II. This last case is more suit-

able for therapeutic purposes as it depletes the pool of na-

tive MPO and generates less oxidation adducts during the

inhibition (Fig. 6). In terms of compound II accumulation,

the ﬁrst molecules are characterized by a short lifetime of

compound II in the absence or the presence of chloride

while the second ones are characterized by a longer lifetime

of compound II in both the conditions.

In the meantime, the accumulation of compound II was

measured for the diﬀerent compounds in the absence or

the presence of chloride. Figure 6 shows that MPO, a fer-

ric-porphyrine (Por) enzyme form (Fe^IIIꢁ ꢁ ꢁPorꢁ ꢁ ꢁaa), is

oxidized by H₂O₂to give compound I, an oxy-ferryl inter-

mediate characterized by a structure which is deﬁcient in

two electrons (Fe^IV@Oꢁ ꢁ ꢁPor^Å+ꢁ ꢁ ꢁaa), one removed from

iron and one from porphyrine (Por^Å+). Compound I is able

to produce HOCl in the presence of chloride ions (Cl^ꢀ)

returning to the native enzyme state. However, in condi-

tions such as the absence of Cl^ꢀ, the presence of an excess

of H₂O₂or the presence of a molecule able to inhibit the

formation of HOCl, the system evolves to the promotion

Finally, the inhibition of LDL oxidation was evaluated

in the presence of several concentrations of the synthe-

P. Van Antwerpen et al. / Bioorg. Med. Chem. 16 (2008) 1702–1720

1709

COOH

COOCH₃

Br

H

Br

i

Br

N

CF₃

ii

iii

NO₂

19

NO₂

22

20

21

COOCH₃

COOH

H

N

CF₃

N

CF₃

v

iv

22

iv

23

NH₂

CF₃

4b

NH₂

vi

COOR₁

COOH

H

N

CF₃

iv

5a: R₁= H, R₂= Me

5b: R₁= H, R₂= C₆H₅

24: R₁= Me, R₂= Me

25: R₁= Me, R₂= C₆H₅

26: R₁= Me, R₂= CH=CHCOOEt

27: R₁= Me, R₂= (CH₂)₃Br

4a

iv

NO₂

HN

O

5c : R₁= H, R₂= CH=CHCOOH

vii

R2

29: R₁= Me, R₂= (CH₂)₃N₃

30: R₁= Me, R₂= (CH₂)₃NH₂

5e: R₁= H, R₂= (CH₂)₃NH₂

v

iv

28: R₁= Me, R₂= (CH₂)₃OOCH₃

5d: R₁= H, R₂= (CH₂)₃OH

iv

Scheme 5. Synthesis of compounds 4a–b and 5a–e. Reagents: (i) KMnO₄, pyridine; (ii) MeI, DBU, MeCN; (iii) 3-triﬂuoromethylaniline,

Pd(CH₃COO)₂, DPePhos, Cs₂CO₃; (iv) KOH, EtOH, H₂O; (v) H₂, Pd/C (5%), MeOH; (vi) R₂COCl, pyridine, CH₂Cl₂; (vii) NaN₃or CH₃COOK,

DMF.

HOCl

Cl^-

Chlorination

Pathway

MP³⁺+ H₂O₂

(Fe^III...Por...aa)

Compound I + H₂O

(Fe^IV=O...Por^.+...aa)

AH

-

O₂

H⁺+ A

AH

Compound II

(Fe^IV=O...Por...aa)

Figure 6. Representation of MPO redox transformation catalytic pathways. MP³⁺: native enzyme; compound I: MP³⁺H₂O₂; compound II: MP²⁺

H₂O₂; AH: reducing agent; O₂^ꢀ: superoxide anion. Adapted from Kettle and Winterbourn.³⁰

sized molecules by a recently developed ELISA.⁷The

LDL oxidation in absence of candidate inhibitors was

considered as the 100% and the native LDL as the 0%.

This assessment takes into account the binding capacity

1710

P. Van Antwerpen et al. / Bioorg. Med. Chem. 16 (2008) 1702–1720

of this enzyme on negatively charged macrostructure

like LDLs and actually improves the relevance of the re-

sults in a physiological context.⁶

pound II is 2.21 and 2.41 times superior to ﬂufenamic

acid with compounds 3b (5-F) and 3c (5-Cl), respec-

tively, suggesting a reduced turn-over of native MPO

by the peroxidase cycle. However, compound 3d (5-Br)

has a similar interaction with the MPO/H₂O₂/Cl^ꢀsys-

tem as ﬂufenamic acid. Compound 3a is characterized

by the absence of compound II formation in the exper-

imental conditions. The redox properties of this mole-

cule that is liable to be oxidized in a quinonimine

could explain this phenomenon and the probable pro-

oxidant eﬀect. Indeed, the TNB reagent used in the tau-

rine assay could form a Michael adduct with the quinon-

imine form that increased the consumption of the

reagent. Finally, the addition of a chemical group at po-

sition 5 causes a decreased inhibition of LDL oxidation.

Flufenamic acid inhibits LDL oxidation in a dose-

dependent manner but compounds 3a, 3b, and 3d only

show a signiﬁcant inhibition at 300 lM (respectively,

53 8%, 9 5% and 29 5%) while no inhibition was

observed with compound 3c (112 6%).

The experimental data are summarized in Table 1 that

assembles the results of the docking, the accumulation

of compound II, and the inhibiting eﬀect measured by

the taurine assay. Table 2 compares the percentage of

HOCl scavenging to the percentage of MPO inhibition

at four relevant physiological concentrations in order

to illustrate the enzyme inhibition capacity of com-

pounds independently of their scavenging activity to-

ward HOCl. The inhibition of LDL oxidation is

illustrated in Figures 7 and 8. Except for compounds

5c and 2, the docking free energies (DG kcal/mol) were

very close to ﬂufenamic acid. Furthermore, a similar

inhibiting eﬀect was noted (Table 1) as far as the 5-

halogenous substituted compounds (3b–d) are con-

cerned. However, none of the synthesized molecules

were able to eﬃciently inhibit LDL oxidation (Figs. 7

and 8) or to have an increased activity as compared to

ﬂufenamic acid.

Compounds in group 4 are the precursors of the com-

pound 5 family and again have geometries close to the

one of ﬂufenamic acid. Moreover, the 4-nitro derivative

demonstrates the same inhibiting eﬀect as ﬂufenamic

acid (P > 0.05, Bonferroni t-test) without any reactivity

toward HOCl. However as demonstrated by the lifetime

of compound II, this molecule deeply accumulates the

compound II, depleting the amount of native MPO for

the HOCl production. This molecule is considered as a

real antagonist of chloride in the catalytic site. The pres-

ence of the nitro group is probably responsible for an

electroattractant eﬀect, decreasing the oxidation in posi-

The results with the phthalic derivatives of ﬂufenamic

acid (diacid compounds) clearly show that the addition

of a carboxylic function suppresses the interaction with

MPO. More interesting is the substitution of ﬂufenamic

acid on position 5. According to the docking, molecules

adopt a position similar to the one of ﬂufenamic acid.

However, compound 3a (5-OH derivative) induces a

pro-oxidative eﬀect on the system, while compounds

3b–d have an activity similar to ﬂufenamic acid without

any scavenging toward HOCl. The lifetime of com-

Table 2. Percentages of the scavenging activity toward HOCl at relevant pharmacological concentrations in comparison to the percentages of MPO

inhibition for the synthesized compounds

Compound

% of scavenging/% of inhibition

0.5 lM

1 lM

1.1 0.2

2 lM

4 lM

1.7 0.6

Flufenamic acid

2

25

2

1.6 0.6

52

2.5 0.1

8

1

3

40

2

1

2

64

2

1

2

1

3

Compound 3a (5OH^ꢀ)

Compound 3b (5F^ꢀ)

Compound 3c (5Cl^ꢀ)

Compound 3d (5Br^ꢀ)

2

ꢀ70

ꢀ73

ꢀ72 1

1.9 0.5

ꢀ73

3

3.3 0.7

2.0 0.5

37

1

5

2

1

49

0.4 0.4

2

63

2

3

1

4

1

2

76

1

29

3

42

2

1

57

3

74

4.9 0.7

79

36

51

65

1

Compound 4a

Compound 4b

0

29

1

2

0

29

2

1

2

1

43

2

1

3

2

55

3

1

2

4

1

ꢀ18

ꢀ28.6 0.7

ꢀ30.7 0.3

ꢀ31.5 0.1

Compound 5a

Compound 5b

Compound 5c

Compound 5d

Compound 5e

1

12

1

2

1

2

24

0

3

8

2

1

2

33

0

5

7

1

2

1

42

2

1

11

ꢀ1

15

ꢀ1

25

0

35

2

0.3 0.4

0.6 0.2

0.1 0.6

ꢀ1

12.2 0.6

0.3 0.2

0

8.5

1

3

5

0

12

2

1

3

4

7

1

20

1

4

3

1

4

2

35

1

3

12

21

49

P. Van Antwerpen et al. / Bioorg. Med. Chem. 16 (2008) 1702–1720

1711

5 µM

180

160

140

120

100

80

15 µM

30 µM

300 µM

60

40

20

0

-20

), 30 lM ( ), and

) is, respectively, considered equal to the 0 and

^{Figure 7. LDL oxidation (in %) in relation to diﬀerent concentrations of ﬂufenamic acid and analogs: 5 lM (}h^{), 15 lM (}

300 lM (

100%. In absence of H₂O₂, % = 2 1% (controls, ); in presence of catalase (400 U/mL) (controls, ), % = 5 2%. Results are means SD for

^{). Absorbance of LDL (Blank; controls,}h^{) and oxidized LDL without drugs (controls,}

n = 3. Percentages of LDL oxidation in the presence of ﬂufenamic acid: 97 4%, 5 lM; 82 14, 15 lM; 23 3%, 30 lM; 10 1%, 300 lM.

addition of a chemical group in position 4, such as for

compounds 3a–d, decreases the inhibition of LDL oxi-

dation as compared to ﬂufenamic acid (compound 4a:

121 7%, 143 11%, 135 10%, 16 2%). Finally,

compound 4b inhibits the oxidation in a dose-dependent

manner (79 9%, 70 5%, 56 6 lM, 29 4%).

5 µM

15 µM

30 µM

300 µM

140

120

100

80

Compounds from series 5 include oleﬁnic or alkyl 4-car-

boxamido substituted ﬂufenamic acid derivatives. They

all have a decreased activity as compared to ﬂufenamic

acid, which is related to a short lifetime of compound

II in the presence of chloride. They also do not react

with HOCl. Compound 5c is particularly interesting: in-

deed, the docking procedure predicts an eﬃcient inhibi-

tion (DG = ꢀ7.74 kcal/mol) which contrasted with the

experimental data. The discrepancy is illustrated by

the accumulation of compound II in which the lifetime

of compound II is decreased in the presence of chloride

and increased in its absence, illustrating a weaker inter-

action with the enzyme compared to ﬂufenamic acid.

Possibly another explanation lies in the mass/charge ra-

tio (m/z), such as demonstrated by Burner et al.³¹As a

matter of fact, the catalytic site is embedded in a hydro-

philic pocket accessible via a narrow oval shaped hydro-

phobic channel and the addition of a negative charge to

the structure of the ligand could hamper an inhibition

process.¹³With the exception of compound 5c, the other

molecules of group 5 more rapidly react with compound

II than ﬂufenamic acid; however, this is not related to a

better inhibition as they are unable to suﬃciently accu-

mulate compound II in presence of chloride. It is inter-

esting to note that the interaction with compound II

would be privileged as the 5 position is free and could

be exposed to oxidation. According to the ELISA, only

60

40

20

0

-20

Figure 8. LDL oxidation (in %) in relation to diﬀerent concentrations

), 30 lM ( ),

^{of ﬂufenamic acid and analogs: 5 lM (}h^{), 15 lM (}

and 300 lM ( ). Absorbance of LDL (Blank; controls,

h^{) and}

oxidized LDL without drugs (controls, ) is, respectively, considered

equal to the 0 and 100%. In absence of H₂O₂, % = 2 1% (controls,

); in presence of catalase (400 U/mL) (controls,

), % = 5 2%.

Results are means SD for n = 3. Percentages of LDL oxidation in the

presence of ﬂufenamic acid: 97 4%, 5 lM; 82 14, 15 lM; 23 3%,

30 lM; 10 1%, 300 lM.

tion 5. Compound 4b is the 4-amino derivative which

seems to react with the MPO/H₂O₂/Cl^ꢀsystem in the

same way of the 5-OH derivative (compound 3a). The

1712

P. Van Antwerpen et al. / Bioorg. Med. Chem. 16 (2008) 1702–1720

compound 5a was able to inhibit LDL oxidation in a

dose-dependent manner (122 12%, 86 10%,

61 8%, 29 5%). The addition of more voluminous

chemical groups or substituents at position 4 seems to

have a negative inﬂuence on the inhibiting eﬀect.

like N-acetylcystein. In this context, the addition of big-

ger atoms at position 5 or of small linear hydrocarbon

chains at position 4 dramatically reduces the inhibition

of LDL oxidation in spite of a better inhibiting eﬀect.

The rational conception of MPO inhibitors should

therefore move to molecules that better ﬁt to these

observations. For example, negatively charged mole-

cules have been considered in a mass/charge ratio but

the impact of positive charge on the inhibitor structure

should be assessed. Small molecules with only one aro-

matic ring could also ﬁt to the inhibition of MPO as this

reduces the mass of the inhibitor and also brings about

more planar structure that could properly sneak be-

tween LDL and MPO. The substitution on the aromatic

ring will have to be chosen to avoid electrophil attack

and the oxidation by the compound I of MPO. Finally,

a virtual modeling of the channel that brings the drugs

to the catalytic site would be a convenient tool to en-

hance the predictions. The MPO inhibition would bene-

ﬁt from all these careful considerations, reducing the

discrepancies arising from the considered experimental

models.

3. Conclusions

The docking process of ﬂufenamic acid and of its ana-

logs in the catalytic site of the enzyme provided interest-

ing data which helped to discriminate active molecules

and to predict the eﬃciency of several compounds.

However, the information carried out was in some cases

rather limited as a consequence of the mechanistic com-

plexity of MPO. Indeed, only a non-signiﬁcant correla-

tion coeﬃcient of 0.63 could be derived from the curve

of the percentage of inhibition plotted as a function of

the predicted free energy (DG, excluding the prooxidant

and inactive molecules). This could be attributed to the

narrow range of free energy values, which are within the

error value of 2 kcal/mol usually accepted for such a

parameter. Moreover, MPO preferably reacts with small

substrates and the enzyme is rapidly oxidized in com-

pound I by H₂O₂(K_s= ꢂ10⁷M^ꢀ1s^ꢀ1).³²Therefore,

the actual inhibitors intensively react with the peroxi-

dase cycle of the enzyme and to diﬀerent extent with

compounds I and II. As a matter of fact, the reactivity

of the molecules with compounds I and II cannot be pre-

dicted by a conventional docking. Moreover, the virtual

procedure does not take into account the penetration of

the drugs through the oval shaped hydrophobic channel

which connects the catalytic site to the outer surface of

the enzyme.

4. Experimental

4.1. Molecular modeling

4.1.1. Protein preparation. The X-ray structure of human

MPO complexed to cyanide and thiocyanate (PDB code:

1DNW) was used as the target structure to endeavor the

docking studies.¹³The X-ray water and the CN^ꢀand

SCN^ꢀmolecules were removed from the active site.

In spite of these limitations, the present investigation

permits to draw some conclusions useful for a rational

conception of MPO inhibitors. First, the mass/charge

ratio is a determinant parameter to discriminate the

molecules able to enter the catalytic site, such as already

demonstrated by Burner et al. for thiol-containing mol-

ecules.³¹Second, the nucleophilic properties of the mol-

ecules have to be carefully considered. Indeed,

ﬂufenamic acid, which is an eﬃcient inhibitor of the

MPO/H₂O₂/Cl^ꢀsystem, acts as an electron donor. The

5-position which is free on the molecule, is favorable

to a rapid oxidation by compounds I and II. A substitu-

tion at this position with a halogen slightly increases the

inhibiting eﬀect and decreases the reactivity with com-

pound II (its lifetime in the presence of chloride in-

creases). This statement is conﬁrmed by the addition

of a nitro group at position 4 that dramatically decreases

the reaction rate with compound II by exerting an elec-

troattractant eﬀect on position 5. Third, the MPO inhi-

bition is strongly dependent on the used model. This

cationic enzyme is liable to be adsorbed on negatively

charged macrostructures like LDLs that could modify

the accessibility to the catalytic site.⁵Moreover, a previ-

ous study demonstrated that the size of the molecule is a

crucial parameter for the inhibition of MPO-dependent

LDL oxidation.¹⁶Indeed, ﬂufenamic acid that inhibits

MPO was unable to inhibit the MPO-dependent LDL

oxidation contrary to small thiol-containing molecules

4.1.2. Ligand preparation. The ligand input ﬁles were

prepared according to the following procedure. The ini-

tial 3D structures of the ligands were generated using

Corina³³and the ligand partial charges were ascribed

using the OPLS force-ﬁeld as performed by Glide.

4.1.3. Docking. The Glide (Grid-Based Ligand Docking

With Energetics) algorithm approximates a systematic

search of positions, orientations, and conformations of

the ligand in the receptor binding site using a series of

hierarchical ﬁlters (www.schrodinger.com).¹⁷The shape

and properties of the receptor are represented on a grid

by several diﬀerent sets of force-ﬁelds that provide pro-

gressively more accurate scoring of the ligand pose. The

ﬁelds are computed prior to docking. The binding site is

deﬁned by a rectangular box conﬁning the translations

of the mass center of the ligand. A set of initial ligand

conformations is generated through exhaustive search

of the torsional minima augmented by a heuristic ap-

proach that eliminates conformations unsuitable for

binding to a receptor. The conformers are clustered in

a combinatorial fashion. Each cluster, characterized by

a common conformation of the core and an exhaustive

set of side-chain conformations, is docked as a single ob-

ject in the ﬁrst stage. The search begins with a rough

positioning and scoring phase that signiﬁcantly narrows

the search space and reduces the number of poses to be

further considered to a few hundred. In the following

P. Van Antwerpen et al. / Bioorg. Med. Chem. 16 (2008) 1702–1720

1713

stage, the selected poses are minimized on precomputed

OPLS-AA van der Waals and electrostatic grids for the

receptor. In the ﬁnal stage, the 5–10 lowest-energy poses

obtained in this fashion are subjected to a Monte Carlo

procedure in which nearby torsional minima are exam-

ined, and the orientation of peripheral groups of the li-

gand is reﬁned. In the present work, the binding region

was deﬁned by a 18 A · 18 A · 18 A box centered on the

central position of the cyanide in the crystal complex. At

the most 10 poses were generated for each molecule. De-

fault settings were used for all the remaining parameters.

1H, H5), 8.30 (m, 2H, H4, H6); analyses were similar

to those previously published.³⁵

4.3.2. Methyl 2-bromo-4-nitrobenzoate (21). Yellow solid

(99%); ¹H NMR (CDCl₃): d 3.99 (s, 3H, OCH₃), 7.92 (d,

1H, 8.4 Hz, H6), 8.21 (dd, 1H, J = 2.4, 8.7 Hz, H5), 8.52

(d, J = 2.4 Hz, 1H, H3); analyses were similar to those of

the literature.³⁶

˚

4.4. General procedure for catalytic hydrogenation

Pd/C 5% (0.78 g) was added to a suspension of the acid

(47.4 mmol) in acetonitrile (50 mL). The suspension was

placed in a hydrogenation ﬂask, connected to a Parr

hydrogenation apparatus, and purged with hydrogen.

The mixture was stirred at room temperature for 18 h

under hydrogen (414,000 Pa). After ﬁltration on Celite,

the solvent was evaporated under reduced pressure.

The products were not puriﬁed.

4.1.4. Evaluation of scoring. The minimized poses gener-

ated by Glide are scored using the GlideScore function

(www.schrodinger.com),¹⁷which is a more sophisticated

version of the empirical ChemScore function with addi-

tional force ﬁeld-based components and terms account-

ing for solvation and repulsive interactions.^17,34

Contributions arise from lipophilic and hydrogen bond-

ing interactions. Metal–ligand interactions are described

by a speciﬁc term. Contributions from force-ﬁeld Cou-

lomb and van der Waals interaction energies are incor-

porated in the empirical-based function. Various

databases of ligands were used to parameterize the coef-

ﬁcients of the functional form. The choice of the best

pose is made using a model energy score (Emodel) that

combines the energy grid score, GlideScore, and the

internal strain of the ligand.

4.4.1. Dimethyl 3-aminophthalate (8). Yellowish oil that

1

crystallized at 0 ꢂC (quantitative); H NMR (CDCl₃):

d 3.84 (s, 3H, OCH₃), 3.86 (s, 3H, OCH₃), 5.21 (ls,

2H, NH₂), 6.78 (dd, J = 1, 8.3 Hz, 1H, H4), 6.89 (dd,

1H, H6), 7.23 (t, 1H, J = 7.5 Hz, H5).

4.4.2. Methyl 4-amino-2-[(3-triﬂuoromethylphenyl)amino]

benzoate (23). Yellow oil which crystallized after tritur-

1

ation (98%); H NMR (CDCl₃): d 3.81 (s, OCH₃, 3H),

4.2. Chemistry

4.02 (br, NH₂, 2H), 6.06 (dd, J = 2.1, 8.6 Hz, 1H, H5),

6.41 (d, J = 2.2 Hz, 1H, H3), 7.27 (m, 1H, H4⁰), 7.36

(m, 2H, ArH), 7.46 (s, 1H, H2⁰), 7.77 (d, J = 8.7 Hz,

1H, H6), 9.71 (br, 1H, NH); ¹³C NMR (CDCl₃): d

51.4 (CH₃), 97.5 (C3), 103.4 (C1), 106 (C5), 118.6 (q,

J = 4 Hz, C2⁰), 119.5 (q, J = 4 Hz, C4⁰), 124.1 (q,

J = 271 Hz, CF₃), 125.1 (C1⁰), 129.8 (C6⁰), 131.7 (q,

J = 32 Hz, C3⁰), 133.8 (C5⁰), 141.7 (C6), 148.8 (C2),

152.2 (C4), 168.8 (COO); IR (KBr): 3455 (N–H), 3305

(N–H), 1675 (C@O), 1632, 1602, 1572, 1522, 1468,

Starting compounds and solvents for reactions or NMR

analysis were purchased from Sigma–Aldrich. Except

when indicated, compounds were not puriﬁed and sol-

vents were not dried before use. NMR spectra (at

1

300 MHz for H and at 75 MHz for ¹³C) were recorded

on a Bruker Avance 300 at 293 K. Tetramethylsilane

(TMS) was used as internal standard and d are expressed

in ppm. The IR spectra were taken on a Shimadzu IR-

470 infrared spectrophotometer and m are given in

cm^ꢀ1. Solid compounds were dispersed in KBr (1%

(m/m)) and liquid compounds were spread on a NaCl

plate. Column chromatography (CC) separations were

performed with Kieselgel 100^ꢁ(Merck) and TLC analy-

ses were carried out on Kieselgel GF₂₅₄^ꢁplates (Merck).

1362, 1268, 1169, 1114, 1068, 846, 774, 691 cm^ꢀ1

.

4.4.3. Dimethyl 3-bromophthalate (9). A 48% (m/m) solu-

tion of HBr in water (13.5 mL), water (50 mL), and ami-

no ester 8 (6.7 g, 32 mmol) were poured in an

Erlenmeyer ﬂask. The suspension was cooled with an

ice-bath and a cold (1–5 ꢂC) solution of NaNO₂

(2.21 g, 32 mmol) in water (10 mL) was added dropwise

in 5 min. The solid dissolved and, when the solution be-

came almost clear, the reaction mixture was placed at

room temperature. Then, a solution of cuprous bromide

(2.53 g, 17.6 mol) in a 48% (m/m) solution of HBr in

water (3.2 mL) was added dropwise. After 16 h of stir-

ring at room temperature, the aqueous mixture was ex-

tracted three times with diethyl ether (100 mL). The

organic layers were washed with 2 M NaOH (100 mL)

and brine (50 mL), dried with Na₂SO₄, ﬁltered, and

evaporated under reduced pressure. The orange oil ob-

tained (7.85 g) was puriﬁed by CC (mobile phase:

CH₂Cl₂) to aﬀord 6.41 g of white solid (yield: 73%);

¹H NMR (CDCl₃): d 3.81 (s, 3H, OCH₃), 3.90 (s, 3H,

OCH₃), 7.26 (t, J = 8.1 Hz, 1H, H5), 7.67 (d,

J = 8.1 Hz, 1H, H4), 7.89 (d, J = 7.8 Hz, 1H, H6); ¹³C

NMR (CDCl₃) d 52.8 (OCH₃), 120.1 (C3), 128.9 (C6),

4.3. General procedure for the synthesis of esters

DBU (47.4 mmol for compound 20 and 94.8 mmol for

compound 6) and MeI (94.8 mmol for acid 20 and

0.142 mol for acid 6) were added to a suspension of

the acid (47.4 mmol) in acetonitrile (50 mL). The solid

slowly dissolved and the solution was stirred for 16 h.

The solvent was evaporated under reduced pressure

and the residue was dissolved in diethyl ether

(150 mL). The organic phase was washed with 2 M

HCl (100 mL), 2 M NaOH (100 mL), and brine

(50 mL). After drying with Na₂SO₄and ﬁltration, the

solvent was evaporated under reduced pressure.

4.3.1. Dimethyl 3-nitrophthalate (7). Yellowish oil that

slowly crystallized (64%); H NMR (CDCl₃): d 3.92 (s,

3H, OCH₃), 3.99 (s, 3H, OCH₃), 7.67 (t, J = 8.9 Hz,

1

1714

P. Van Antwerpen et al. / Bioorg. Med. Chem. 16 (2008) 1702–1720

129.2 (C1), 130.2 (C5), 136.7 (C4), 137.1 (C2), 164.6

(COO), 167.6 (COO).

115.8 (d, J = 8 Hz, C3), 117.6 (d, J = 22Hz, C6),

118.4 (q, J = 4 Hz, C2⁰), 120.1 (q, J = 4 Hz, C4⁰),

120.7 (d, J = 22 Hz, C4), 123.7 (q, J = 271 Hz, CF₃),

124.4 (C6⁰), 130.5 (C5⁰), 131.8 (q, J = 32 Hz, C3⁰),

138.6 (d, J = 2 Hz, C2), 143.5 (C1⁰), 151 (d,

J = 235 Hz, C5), 167.9 (d, J = 2Hz, COO); IR (KBr):

3320 (N–H), 1692 (C@O), 1586, 1520, 1431, 1334,

1225, 1187, 1161, 1122, 1069, 981, 910, 821, 788,

4.5. General procedure for the Hartwig–Buchwald

coupling

Finely powdered 3-(triﬂuoromethyl)aniline (1.42 g,

8.8 mmol), palladium(II) acetate (33 mg, 150 lmol),

and bis[2-(diphenylphosphino)phenyl]ether (DPEphos)

(20 mg, 220 lmol) was added in a oven-dried Shlenk-

tube cooled under anhydrous N₂, equipped with a mag-

netic stirrer. The tube was purged with anhydrous N₂.

Then the bromo or triﬂate derivative (7.3 mmol) and tol-

uene (10 mL) (freshly distilled on sodium) were poured

into the ﬂask. The mixture was stirred at room temper-

ature for 5 min. Then, Cs₂CO₃(3.34 g, 10.3 mmol) was

added, the tube was purged with anhydrous N₂, and

the reaction medium was heated to 100 ꢂC for 4 h. After

cooling, the brown solution was diluted with toluene

(50 mL) and washed with 2 M HCl (50 mL) and brine

(50 mL). The organic layer was dried with Na₂SO₄, ﬁl-

tered, and evaporated under reduced pressure. The com-

pounds were puriﬁed by CC (silica gel). The mobile

phases were CH₂Cl₂for esters 10 and 22, and petroleum

ether_40–60/acetone (acetone gradient 0% to 2% to 5%) for

compounds 14, 18a, and 18b.

697 cm^ꢀ1

.

4.5.4. Methyl 5-chloro-2-[(3-triﬂuoromethylphenyl)amino]

1

benzoate (18b). Yellow oil (82%); H NMR (CDCl₃): d

3.90 (s, 3H, OCH₃), 7.17 (d, 1H, J = 9 Hz, H3), 7.25–7.44

(5H, ArH), 7.93 (d, J = 2.7 Hz, 1H, H6), 9.53 (br, 1H,

NH); ¹³C NMR (CDCl₃): d 52.1 (OCH₃), 113.8 (C1),

115.5 (C3), 118.2 (q, J = 4 Hz, C2⁰), 120 (q, J = 4 Hz,

C4⁰), 122.7 (C5), 123.8 (q, J = 271 Hz, CF₃), 124.7 (C6⁰),

130 (C5⁰), 131 (C6), 131.9 (q, J = 32 Hz, C3⁰), 134.1 (C4),

141.1 (C2), 145.3 (C1⁰), 167.7 (COO); IR (KBr): 3315

(N–H), 3060, 2940, 1691 (C@O), 1574, 1514, 1451, 1335,

1250, 1217, 1163, 1121, 969, 924, 898, 819, 790, 729,

696 cm^ꢀ1

.

4.5.5. Methyl 4-nitro-2-[(3-triﬂuoromethylphenyl)amino]

benzoate (22). Yellowish solid (quantitative); H NMR

1

(CDCl₃): d 3.97 (s, 3H, OCH₃), 7.44–7.49 (m, 3H,

ArH), 7.54 (m, 2H, ArH), 7.98 (d, J = 2.2 Hz, 1H,

H3), 8.14 (d, J = 8.8 Hz, 1H, H6), 9.78 (br, 1H, NH);

¹³C NMR (CDCl₃): d 52.8 (OCH₃), 108.5 (C3), 111.7

(C5), 116.6 (C1), 119.8 (q, J = 4 Hz, C2⁰), 121.7 (q,

J = 4 Hz, C4⁰), 123.8 (q, J = 271 Hz, CF₃), 125.8 (C5⁰),

130.6 (C6⁰), 132.6 (q, J = 32 Hz, C3⁰), 133.4 (C6),

140.2 (C1⁰), 148.1 (C2), 151.7 (C4), 167.8 (COO); IR

(KBr): 3290 (N–H), 3090, 2940, 1686 (C@O), 1599,

1527, 1442, 3129, 1218, 1178, 1121, 1091, 1065, 870,

4.5.1. Dimethyl 3-[(3-triﬂuoromethylphenyl)amino]phthal-

1

ate (10). Yellow solid (91%); H NMR (CDCl₃): d 3.88

(s, 3H, OCH₃), 3.90 (s, 3H, OCH₃), 7.21 (dd, J = 1.6,

7 Hz, 1H, H4), 7.26–7.43 (m, 6H, ArH), 8.04 (br, 1H,

NH); ¹³C NMR (CDCl₃):

d 52.7 (OCH₃), 52.8

(OCH₃), 116.7 (q, J = 4 Hz, C2⁰), 117.8 (C2), 119.2

(C6⁰), 119.3 (q, J = 4 Hz, C4⁰), 120.8 (C4), 123.2 (C2),

124 (q, J = 271 Hz, CF₃), 130.1 (C6), 131.9 (C5), 132.1

(q, J = 32 Hz, C3⁰), 134.2 (C1), 142 (C3), 143.6 (C1⁰),

168.5 (COO), 168.6 (COO); IR (KBr): 3365 (N–H),

2935, 1732 (C@O), 1704 (C@O), 1572, 1520, 1535,

820, 794, 734, 710, 676 cm^ꢀ1

.

4.6. 5-Bromo-2-[(3-triﬂuoromethylphenyl)amino]benzoic acid

(3d)

1277, 1250, 1121, 1004, 870, 790, 503, 477 cm^ꢀ1

.

4.5.2. Methyl 5-phenylmethoxy-2-[(3-triﬂuoromethylphenyl)

amino]benzoate (14). Yellow oil which slowly crystallized

(67%); H NMR (CDCl₃): d 3.90 (s, 3H, CH₃), 5.04 (s,

To a stirred solution of compound 1 (0.5 g, 1.8 mmol) in

MeOH (15 mL) was added 100 ll (1.8 mmol) of bromine.

The reaction was checked by TLC (mobile phase:

CH₂Cl₂/MeOH 9:1) and, after 1 h, water (15 mL) was

added. The yellow precipitate was ﬁltered, dried at

60 ꢂC under vacuum (27,000 Pa), and recrystallized in

acetone/petroleum ether_40–60(2:25) to aﬀord a yellow so-

lid (340 mg, 53%); ¹H NMR (DMSO-d₆): d 7.16 (d,

J = 9 Hz, 1H, H3), 7.31 (d, 1H), 7.47–7.54 (4H, ArH),

7.95 (d, J = 2.1 Hz, 1H, H6), 9.65 (s, 1H, NH), 13.2 (br,

1H, COOH); ¹³C NMR (DMSO-d₆): d 109.2 (C5), 115.9

(C1), 116.9 (C3), 117.1 (q, J = 4 Hz, C2⁰), 119.3 (q,

J = 4 Hz, C4⁰), 124.1 (q, J = 271 Hz, CF₃), 124.3 (C6⁰),

130.5 (q, J = 31 Hz, C3⁰), 130.7 (C5⁰), 133.9 (C6), 136.6

(C4), 141.4 (C1⁰), 145 (C2), 168.5 (COO); IR (KBr):

3305 (N–H), 3050, 2900, 1661 (C@O), 1571, 1513, 1429,

1

2H, CH₂), 7.09 (dd, J = 3, 9.1 Hz, 1H, H4), 7.29 (d,

J = 9 Hz, 1H, H3), 7.33–7.46 (9H, ArH), 7.59 (d,

J = 3.1 Hz, 1H, H6), 9.20 (br, 1H, NH); ¹³C NMR

(CDCl₃) d 52.2 (CH₃), 70.9 (CH₂), 114.4 (C1), 116.1

(C3), 116.4 (q, J = 4 Hz, C2⁰), 117.2 (C6), 118.6 (q,

J = 4 Hz, C4⁰), 123 (C4), 124.2 (q, J = 271 Hz, CF₃),

127.7 (C2⁰⁰,C6⁰⁰), 128.2 (C4⁰⁰), 128.7 (C3⁰⁰, C5⁰⁰), 130

(C5⁰,C6⁰), 131.9 (q, J = 32 Hz, C3⁰), 137 (C1⁰⁰), 140.7

(C2), 142.7 (C1⁰), 151.3 (C5), 168.5 (COO); IR (KBr):

3340 (m N–H), 3015, 2880, 2725, 2615, 1659 (m C@O),

1617, 1582, 1522, 1497, 1458, 1418, 1389, 1334, 1286,

1264, 1218, 1170, 1125, 1150, 1065, 923, 892, 818, 787,

728, 704 cm^ꢀ1

.

1339, 1241, 1157, 1120, 883, 820, 789, 696 cm^ꢀ1

.

4.5.3. Methyl 5-ﬂuoro-2-[(3-triﬂuoromethylphenyl)

amino]benzoate (18a). Yellowish solid (75%); ¹H

NMR (CDCl₃): d 3.93 (s, 3H, OCH₃), 7.19 (d,

J = 8.9 Hz, 1H, H3), 7.15–7.44 (5H, ArH), 7.64 (dd,

J = 1.9, 9.4 Hz, 1H, H6), 9.35 (br, 1H, NH); ¹³C

NMR (CDCl₃): d 52.3 (OCH₃), 114 (d, J = 7 Hz, C1),

4.7. General procedure for the hydrolysis of esters

Water (25 mL) and KOH were added to a solution of

the ester (6.6 mmol) in ethanol (25 mL), placed in a

round-bottomed ﬂask. The amount of KOH was deter-

P. Van Antwerpen et al. / Bioorg. Med. Chem. 16 (2008) 1702–1720

1715

mined by the number of ester functions to hydrolyze:

13.2 mmol for compounds 4a, 4b, 5a, 5b, 5e, 14, 18a,

and 18b; 26.4 mmol for compounds 5c, 5d, and 10.

The mixture was reﬂuxed for 2 h and let to cool. After

evaporation of ethanol, the acid was precipitated by

addition of 2 M H₂SO₄(pH ꢃ 2). For compounds 4b

and 5e, acid was added until maximal precipitation

has been obtained (pH ꢃ 5–6). The solid was ﬁltered,

washed three times with water (15 mL), and dried under

reduced pressure (27,000 Pa) at 80 ꢂC.

¹³C NMR (CDCl₃+ 10% CD₃OD): d 97.3 (C3), 103.1

(C1), 106 (C5), 118.3 (q, J = 4 Hz, C2⁰), 119.3 (q,

J = 4 Hz, C4⁰), 124 (q, J = 271 Hz, CF₃), 125 (C6),

129.8 (C6⁰), 131.2 (q, J = 32 Hz, C3⁰), 134.5 (C5⁰),

141.7 (C1⁰), 148.9 (C2), 152.6 (C4), 171 (COO); IR

(KBr): 3433 (N–H), 3290 (N–H), 3025 (m O–H),

2905, 2565, 1651 (C@O), 1622, 1591, 1570, 1528, 1444,

1415, 1333, 1291, 1259, 1168, 1068, 887, 835, 789,

693 cm^ꢀ1

.

4.7.6. 4-Acetamido-2-[(3-triﬂuoromethylphenyl)amino]

benzoic acid (5a). Beige solid (quantitative); H NMR

1

4.7.1. 3-[(3-Triﬂuoromethylphenyl)amino]phthalic acid (2).

Yellow solid (88%); H NMR (D₂O + 1 equivalent of

1

(CDCl₃+ 10% CD₃OD): d 2.11 (s, 3H, CH₃), 6.93 (dd,

J = 1.2, 8.6 Hz, 1H, H5), 7.26–7.50 (m, 4H, ArH), 7.69

(s, 1H, H3), 7.97 (d, J = 8.7 Hz, 1H, H6); ¹³C NMR

(CDCl₃+ 10% CD₃OD): d 23.8 (CH₃), 103.8 (C3),

109.5 (C1,C5), 117.8 (q, J = 4 Hz, C2⁰), 119.3 (q,

J = 4 Hz, C4⁰), 123.9 (q, J = 271 Hz, CF₃), 124.2 (C1⁰),

129.7 (C6⁰), 131.6 (q, J = 32 Hz, C3⁰), 133.4 (C5⁰),

141.4 (C6), 143.7 (C2), 147.6 (C4), 170.3 (CON,COO);

IR (KBr): 3280 (N–H), 3105 (O–H), 2905, 1677

(C@O), 1651, 1578, 1546, 1502, 1424, 1337, 1243,

NaOH): d 7.24–7.40 (m, 5H, ArH), 7.47 (t, J = 9 Hz,

1H, H6).

4.7.2. 5-Fluoro-2-[(3-triﬂuoromethylphenyl)amino]benzoic

1

acid (3b). White solid (80%); H NMR (DMSO-d₆) d

7.30–7.38 (3H, ArH), 7.41–7.53 (2H, ArH), 7.66 (dd,

J = 1.8, 9.6 Hz 1H, H6), 9.43 (br, 1H, NH), 13.52 (br,

1H, COOH); ¹³C NMR (DMSO-d₆) d 115.6 (q,

J = 4 Hz, C2⁰), 116.2 (d, J = 7 Hz, C1), 117 (d,

J = 23 Hz, C6), 117.8 (d, J = 7 Hz, C3), 118.1 (q,

J = 8 Hz, C4⁰), 121.3 (d, J = 23 Hz, C4), 122.8 (C6⁰),

124 (q, J = 271 Hz, CF₃), 129.8 (q, J = 31 Hz, C3⁰),

130.4 (C5⁰), 141.7 (C2), 142.4 (C1⁰), 154.7 (d,

J = 235 Hz, C5), 168.3 (d, J = 2 Hz, COO); IR (KBr):

3315 (N–H), 3060 (O–H), 2900, 1661 (C@O), 1580,

1158, 1118, 1073, 893, 789, 691, 669 cm^ꢀ1

.

4.7.7. 4-Benzamido-2-[(3-triﬂuoromethylphenyl)amino]

benzoic acid (5b). Beige solid (quantitative); H NMR

1

(CDCl₃+ 10% CD₃OD): d 7.11 (dd, J = 1.9, 8.7 Hz,

1H, H5), 7.27 (d, 1H, J = 7.7 Hz), 7.38–7.56 (m, 6H),

7.85 (3H, ArH); 8.02 (d, J = 8.7 Hz, 1H, H6); ¹³C

NMR (CDCl₃+ 10% CD₃OD): d 104.5 (C3), 108.7

(C5), 110.2 (C1), 118.1 (q, J = 4 Hz, C2⁰), 119.3 (q,

J = 4 Hz, C4⁰), 123.9 (q, J = 271 Hz, CF₃), 124.1 (C1⁰),

127.3 (C2⁰⁰,C6⁰⁰), 128.4 (C3⁰⁰,C5⁰⁰), 129.8 (C6⁰), 131.6 (q,

J = 32 Hz, C3⁰), 131.9 (C4⁰⁰), 133.4 (C5⁰), 134.6 (C1⁰⁰),

141.4 (C6), 143.7 (C2), 147.6 (C4), 167.3 (CON), 170.3

(COO); IR (KBr): 3305 (N–H), 3015 (O–H), 2630,

1656 (C@O), 1579, 1534, 1507, 1425, 1333, 1257, 1222,

1517, 1339, 1227, 1179, 1140, 1104, 819, 699 cm^ꢀ1

.

4.7.3. 5-Chloro-2-[(3-triﬂuoromethylphenyl)amino]benzoic

1

acid (3c). Yellow solid (86%); H NMR (DMSO-d₆): d

7.26 (d, J = 8.7 Hz, 1H, H3), 7.36 (m, 1H, ArH), 7.44

(dd, J = 2.4, 9 Hz, 1H, H4), 7.57 (3H, ArH), 7.85 (d,

J = 2.4 Hz, 1H, H6), 9.64 (s, 1H, NH), 13.4 (br, 1H,

COOH); ¹³C NMR (DMSO-d₆): d 115.8 (C3), 116.7

(C1), 117 (q, J = 4 Hz, C2⁰), 119.1 (q, J = 4 Hz, C4⁰),

121.9 (C6⁰), 124 (q, J = 271 Hz, CF₃), 124.2 (C5),

130.4 (q, J = 29 Hz, C3⁰), 130.6 (C5⁰), 130.8 (C6),

133.8 (C4), 141.5 (C1⁰), 144.5 (C2), 168.5 (COO); IR

(KBr): 3305 (N–H), 3055 (O–H), 2905, 1654 (C@O),

1569, 1510, 1425, 1337, 1243, 1158, 1119, 883, 787,

1158, 1118, 1069, 901, 790, 704 cm^ꢀ1

.

4.7.8. (2E)-4-[(3-Carboxy-1-oxopropenyl)amino]-2-[(3-

triﬂuoromethylphenyl)amino]benzoic acid (5c). Orange

1

solid (quantitative); H NMR (DMSO-d₆): d 6.64 (d,

694 cm^ꢀ1

.

J = 15.4 Hz, 1H,CH@CHCOOH), 7.08–7.13 (2H, ArH,

CH@CHCOOH), 7.38 (m, 1H, ArH), 7.58–7.5 (3H,

ArH), 7.83 (s, 1H, H3), 7.91 (d, J = 8.8 Hz, 1H, H6),

9.83 (s, 1H, ArNHAr), 10.64 (s, 1H, OCNH); ¹³C

NMR (DMSO-d₆): d 103.9 (C3), 109.3 (C5), 109.9

(C1), 117.2 (q, J = 4 Hz, C2⁰), 119.0 (q, J = 4 Hz, C4⁰),

124.1 (q, J = 271 Hz, CF₃), 124.3 (C1⁰), 130.2 (C6⁰),

130.4 (CH@CHCOOH), 131.3 (C3⁰), 133.1 (C5⁰), 136.8

(CH@CHCOOH); 141.6 (C6), 143.6 (C2), 146.7 (C4),

162.2 (CON); 166.2 (CH@CHCOOH), 169.3 (Ar-

COOH); IR (KBr): 3300 (N–H), 3030 (O–H), 1667

(C@O), 1582 (C@O), 1536, 1504, 1413, 1328, 1247,

4.7.4. 4-Nitro-2-[(3-triﬂuoromethylphenyl)amino]benzoic

acid (4a). Orange solid (77%); H NMR (CDCl₃+ 10%

1

CD₃OD): d 7.37–7.46 (ArH, 4H), 7.49 (dd, J = 2.3,

8.7 Hz, 1H, H5), 7.94 (d, J = 2.2 Hz, 1H, H3), 8.14 (d,

J = 8.7 Hz, 1H, H6); ¹³C NMR (CDCl₃+ 10%

CD₃OD): d 108.2 (C3), 111.6 (C5), 117.2 (C1), 119.5

(q, J = 4 Hz, C2⁰), 121.3 (q, J = 4 Hz, C4⁰), 121.3 (q,

J = 274 Hz, CF₃), 125.4 (C6⁰), 130.4 (C5⁰), 132.3 (q,

J = 29 Hz, C3⁰), 134 (C6⁰), 141.7 (C1⁰), 148.9 (C2),

151.5 (C4), 169.4 (COO); IR (KBr): 3310 (N–H), 3045

(O–H), 2905, 1662 (C@O), 1533, 1433, 1351, 1322,

1251, 1226, 1132, 820, 797 cm^ꢀ1; analyses were similar

to those of the literature.³⁷

1223, 1154, 1114, 781, 659 cm^ꢀ1

.

4.7.9. 4-[(4-Hydroxy-1-oxobutyl)amino]-2-[(3-triﬂuoro-

methylphenyl)amino]benzoic acid (5d). Orange solid

(quantitative); ¹H NMR (DMSO-d₆): d 1.69 (m, 2H,

O@CCH₂CH₂CH₂OH), 2.35 (t, J = 7 Hz, 2H,

O@CCH₂CH₂CH₂OH), 3.41 (t, J = 6.2 Hz, 2H,

O@CCH₂CH₂CH₂OH), 7.06 (dd, J = 1.9, 8.7 Hz, 1H,

H5), 7.36 (m, 1H, ArH), 7.56–7.57 (3H, ArH); 7.78 (d,

4.7.5. 4-Amino-2-[(3-triﬂuoromethylphenyl)amino]benzoic

acid (4b). Beige solid (68%); H NMR (CDCl₃+ 10%

1

CD₃OD): d 6.05 (dd, J = 2.2, 8.7 Hz, 1H, H5), 6.37 (d,

J = 2.2 Hz, 1H, H3), 7.20 (m, H6⁰, 1H), 7.34 (m, 2H,

ArH), 7.40 (s, 1H, ArH), 7.77 (d, J = 8.6 Hz, 1H, H6);

1716

P. Van Antwerpen et al. / Bioorg. Med. Chem. 16 (2008) 1702–1720

J = 1.9 Hz, 1H, H3), 7.84 (d, J = 8.7 Hz, 1H, H6), 9.83

(br, 1H, O@CNH), 10.08 (s, 1H, ArNHAr); ¹³C NMR

4.8. General procedure for the synthesis of triﬂates (13),

(17a), and (17b)

(DMSO-d₆):

d

28.2 (O@CCH₂CH₂CH₂OH), 33.3

(O@CCH₂CH₂CH₂OH), 60.1 (O@CCH₂CH₂CH₂OH),

103.3 (C3), 108.4 (C5), 109.5 (C1), 116.8 (q, J = 4 Hz,

C2⁰), 118.7 (q, J = 4 Hz, C4⁰), 124.1 (q, J = 271 Hz,

CF₃), 124.4 (C1⁰), 130.1 (C6⁰), 130.5 (q, J = 32 Hz,

C3⁰), 132.9 (C5⁰), 141.7 (C6), 144.5 (C2), 146.6 (C4),

169.4 (COO), 172 (CON); IR (KBr): 3425 (O–H), 3280

(N–H), 2915, 1661 (C@O), 1601, 1577, 1546, 1506,

A stirred solution of the phenol (13.2 mmol) and trieth-

ylamine (3.7 mL, 26.4 mmol) in dry CH₂Cl₂(100 mL)

was cooled with an ice-bath. To this solution, triﬂuoro-

methanesulfonic anhydride (2.5 mL, 14.6 mmol) was

added dropwise. Then, the solution was let to warm to

room temperature and the reaction was carried out for

8 h. The organic phase was washed with 1 M HCl

(50 mL), 2 M NaOH (25 mL), and brine (25 mL). The

solution was dried with Na₂SO₄, ﬁltered, and evapo-

rated under reduced pressure.

1420, 1332, 1228, 1168, 1133, 1099, 1065, 782, 693 cm^ꢀ1

.

4.7.10. 4-[(4-Amino-1-oxobutyl)amino]-2-[(3-triﬂuorometh-

ylphenyl)amino]benzoic acid (5e). White solid (31%); H

1

NMR (DMSO-d₆+ 10% of 2 M NaOH in D₂O): d 1.75

(m, 2H, O@CCH₂CH₂CH₂NH₂), 2.37 (t, 2H, J = 7 Hz,

O@CCH₂CH₂CH₂NH₂), 2.68 (t, 2H, J = 6.2 Hz,

O@CCH₂CH₂CH₂NH₂), 6.90 (dd, J = 1.9, 8.7 Hz, 1H,

H5), 7.13 (m, 1H, ArH), 7.36–7.47 (3H, ArH); 7.77 (d,

J = 1.9 Hz, 1H, H3), 7.85 (d, J = 8.4 Hz, 1H, H6); ¹³C

NMR (DMSO-d₆+ 10% of 2 M NaOH in D₂O): d 26.3

(O@CCH₂CH₂CH₂NH₂), 33.9 (O@CCH₂CH₂CH₂NH₂),

39.7 (O@CCH₂CH₂CH₂NH₂),³⁸104.3 (C3), 109.2 (C5),

113.6 (C1), 116.2 (q, J = 4 Hz, C2⁰), 119.1 (q, J = 4 Hz,

C4⁰), 124.1 (q, J = 271 Hz, CF₃), 121.1 (C6⁰), 130.1

(C6⁰), 130.5 (q, J = 32 Hz, C3⁰), 132.8 (C5⁰), 140.9 (C1⁰),

143.7 (C2), 144.3 (C4), 171.2 (COO,CON); IR (KBr):

3420 (O–H), 3265 (N–H), 3060, 1649 (C@O), 1600,

1579, 1532, 1495, 1466, 1406, 1362, 1337, 1281, 1221,

4.8.1. Methyl 5-phenylmethoxy-2-[((triﬂuoromethyl)sulfonyl)

oxy]benzoate (13). The compound was puriﬁed by CC

(mobile phase = CH₂Cl₂); yellow solid (96%); ¹H

NMR (CDCl₃): d 3.96 (s, 3H, CH₃), 5.10 (s, 2H, CH₂),

7.16 (dd, J = 3, 9.1 Hz, 1H, H4); 7.22 (d, J = 9.1 Hz,

1H, H3), 7.34–7.42 (m, 5H, ArH), 7.67 (d, J = 3 Hz, 1H,

H6); ¹³C NMR (CDCl₃): d 52.8 (CH₃), 70.9 (CH₂),

117.8 (C1), 118.8 (CF₃, J = 319 Hz), 120.6 (C3), 124

(C6), 125.3 (C4), 127.7 (C2⁰,C6⁰), 128.5 (C4⁰), 128.9

(C3⁰,C5⁰), 135.8 (C1⁰), 141.9 (C2), 158 (C5), 164.2 (COO).

4.8.2. Methyl 5-ﬂuoro-2-[((triﬂuoromethyl)sulfonyl)oxy]

benzoate (17a). The product was not puriﬁed and used di-

rectly in the next step. Reddish oil; IR (KBr): 3188, 3070,

2940, 1735 (C@O), 1685, 1488, 1421 (S@O), 1300, 1272,

1116, 1065, 825, 790 cm^ꢀ1

.

1241, 1205 (S@O), 1140, 1068, 901, 865, 832, 770 cm^ꢀ1

.

4.7.11. 5-Phenylmethoxy-2-[(3-triﬂuoromethylphenyl)amino]

benzoic acid (15). Yellow solid (90%); ¹H NMR

(CDCl₃): d 5.06 (s, 2H, CH₂), 7.15 (dd, J = 3, 9.1 Hz,

1H, H4), 7.26–7.46 (ArH, 10H), 7.66 (d, J = 3 Hz, 1H,

H6), 9 (bs, 1H, NH), 10.1 (bs, 1H, COOH); ¹³C NMR

(CDCl₃); d 70.8 (CH₂), 112.6 (C6), 113.9 (C1), 116.2

(C3), 117.1 (C4), 117.4 (q, J = 4 Hz, C2⁰), 119.4 (q,

J = 4 Hz, C4⁰), 124.1 (q, J = 277 Hz, CF₃), 123.9 (C6⁰),

127.7 (C2⁰⁰, C6⁰⁰), 128.3 (C4⁰⁰), 128.8 (C3⁰⁰,C5⁰⁰),³⁹131.9

(q, J = 32 Hz, C3⁰), 136.8 (C1⁰⁰), 142 (C1⁰), 142.2 (C2),

151.3 (C4), 173 (COO); analyses were similar to those

of the literature.³⁵

4.8.3. Methyl 5-chloro-2-[((triﬂuoromethyl)sulfonyl)oxy]

benzoate (17b). The product was not puriﬁed and used

directly in the next step. Yellow oil; IR (KBr): 3085,

2940, 1732 (C@O), 1478, 1425 (S@O), 1285, 1260,

1205 (S@O), 1134, 969, 895, 846, 782 cm^ꢀ1

.

4.8.4. 5-Hydroxy-2-[(3-triﬂuoromethylphenyl)amino]ben-

zoic acid (3a). A solution of compound 15 (630 mg,

1.6 mmol) in MeOH (35 mL) was poured into a hydroge-

nation ﬂask. Then, 5% Pd/C (0.1 g) and 12 M HCl

(0.15 mL 1.8 mmol) were added. The ﬂask was connected

to a Parr hydrogenation apparatus and ﬁlled with hydro-

gen (400,000 Pa). The reaction was carried out at room

temperature for 12 h. The suspension was ﬁltered on Cel-

ite and MeOH was evaporated under reduced pressure.

The solid obtained was recrystallized in ethyl acetate/

petroleum ether_40–60(5:15) to yield yellow crystals

(340 mg, 70%); ¹H NMR (CDCl₃+ 10% CD₃OD): d

6.99 (dd, J = 2.9, 8.9 Hz, 1H, H7), 7.16–7.39 (ArH,

5H), 7.5 (d, J = 2.9 Hz, 1H, H6); ¹³C NMR

(CDCl₃+ 10% CD₃OD): d 115.5 (q, J = 3 Hz, C2⁰),

115.5 (C6), 117.5 (C3), 117.8 (C4), 117.8 (q, J = 4 Hz,

C4⁰), 122.1 (C1), 122.4 (C6⁰), 124.1 (q, J = 271 Hz,

CF₃), 129.8 (C5⁰), 131.6 (q, J = 32 Hz, C3⁰), 139.2 (C2),

143.2 (C1⁰), 149.3 (C5), 170.3 (COO); IR m 3330 (N–H),

3050 (m O–H), 1711 (C@O), 1667, 1587, 1528, 1459,

4.7.12. Methyl 5-phenylmethoxy-2-hydroxybenzoate (12).

Methyl gentisate (5 g, 29.7 mmol) and benzyl bromide

(3.5 mL, 29.7 mmol) were added to a suspension of

K₂CO₃(16.4 g, 0.119 mol) in a mixture of CHCl₃

(160 mL) and MeOH (80 mL). The mixture was reﬂuxed

for 4 h. After cooling, the suspension was ﬁltered and

the solvents were evaporated under reduced pressure.

The residue was puriﬁed by CC (mobile phase: CH₂Cl₂)

1

to give a white solid (5.8 g, 75%); H NMR (CDCl₃): d

3.5 (s, 3H, OCH₃), 5.02 (s, 2H, CH₂), 3.94 (d,

J = 9 Hz, 1H, H6), 7.16 (dd, J = 3.2, 9 Hz, 1H, H4),

7.34–7.46 (ArH, 6H), 10.41 (s, 1H, OH); ¹³C NMR

(CDCl₃): d 52.4 (CH₃), 70.9 (CH₂), 112 (C1), 113.6

(C3), 118.6 (C6), 124.9 (C4), 127.7 (C2⁰,C6⁰), 128.1

(C4⁰), 128.7 (C3⁰,C5⁰), 136.9 (C1⁰), 151.2 (C2), 156.3

(C5), 170.3 (COO); IR (KBr): 3300 (O–H), 1683

(C@O), 1516, 1335, 1226, 1160, 1128, 1071, 1026, 791,

745, 696 cm^ꢀ1, analyses were similar to those of the

literature.²⁴

1432, 1335, 1273, 1247, 1212, 1114, 921, 820, 785 cm^ꢀ1

analyses were similar to those of the literature.³⁵

;

4.8.5. 2-Bromo-4-nitrobenzoic acid (20). Nitro com-

pound 19 (24.4 g, 0.113 mol), pyridine (120 mL), and

water (240 mL) were poured in a round-bottomed ﬂask

P. Van Antwerpen et al. / Bioorg. Med. Chem. 16 (2008) 1702–1720

1717

equiped with a magnetic stirrer and a reﬂuxed con-

denser. The suspension was heated to 100 ꢂC, then re-

moved from the oil bath and KMnO₄(98.2 g,

0.62 mol) was added carefully in small portions.⁴⁰The

reaction was thereby maintained at reﬂux temperature

by addition of the reagent. When addition was com-

plete, the mixture was stirred until the temperature

reached 60–70 ꢂC. Then heating was carried out at

100 ꢂC for 5 h. The ﬂask was removed from the oil bath

and KMnO₄(26.8 g, 0.17 mol) was added in small por-

tions as described above. After 16 h of stirring at 100 ꢂC

and cooling to room temperature, 13 M NaOH (15 mL)

was added and the resulting suspension was ﬁltered on

Celite. The solution was washed with diethyl ether

(100 mL) and then acidiﬁed to pH ꢂ 1 with 12 M HCl.

The yellow suspension was extracted two times with

diethyl ether (150 mL). The organic phase was washed

with brine (50 mL), dried with Na₂SO₄, ﬁltered, and

evaporated under reduced pressure to aﬀord a yellowish

solid (19.35 g, 70%). Analyses were similar to those of

the literature.²⁶

(q, J = 4 Hz, C2⁰), 119.9 (q, J = 4 Hz, C4⁰), 124.1 (q,

J = 271 Hz, CF₃), 124.8 (C1⁰), 127.2 (C2⁰⁰,C6⁰⁰), 128.9

(C3⁰⁰,C5⁰⁰), 130 (C6⁰), 132 (q, J = 32 Hz, C3⁰), 132.2

(C4⁰⁰), 133.1 (C5⁰), 134.6 (C1⁰⁰), 141.3 (C6), 143.4 (C2),

148 (C4), 166.1 (CON), 168.5 (COO); IR (KBr): 3295

(N–H), 1680 (C@O), 1645, 1620, 1589, 1507, 1424,

1235, 1150, 1113, 1081, 770, 709, 687 cm^ꢀ1

.

4.9.3. Methyl (2E)-4-[(4-ethoxy-1,4-dioxo-2-butenyl)

amino]-2-[(3-triﬂuoromethylphenyl)amino]benzoate (26).

1

Yellow solid (79%); H NMR (DMSO-d₆): d 1.25 (t,

J = 7.1 Hz, 3H, CH₂CH₃), 3.83 (s, 3H, OCH₃), 4.20

(q, J = 7 Hz, 2H, CH₂), 6.67 (d, J = 15.5 Hz, 1H,

7.14–7.15

CH@CHCOOEt),

(m,

2H,

H5,

CH@CHCOOEt), 7.39 (m, 1H, ArH), 7.57–7.58 (m, 3H,

ArH), 7.80 (d, J = 1.7 Hz, 1H, H3), 7.89 (d, J = 8.8 Hz,

1H, H6), 9.48 (s, 1H, ArNHAr), 10.69 (s, 1H, CONH);

¹³C NMR (DMSO-d₆) d 14 (CH₂CH₃), 51.9 (OCH₃),

60.9 (CH₂), 104.2 (C3), 108.9 (C5), 110.1 (C1), 117.3

(q, J = 4 Hz, C2⁰), 119.1 (q, J = 4 Hz, C4⁰), 124 (q,

J = 271 Hz, CF₃), 124.4 (C6), 130 (C6⁰); 130.6

(CH@CHCOOEt), 131.6 (C3⁰),⁴¹132.6 (C5⁰), 137.2

(CH@CHCOOEt), 141.6 (C1⁰), 143.8 (C2), 146.4 (C4),

161.9 (O@CN); 164.8 (CH@CHCOOEt), 167.3 (Ar-

COOH); IR (KBr): 3325 (N–H), 1707 (C@O), 1668

(C@O), 1604, 1545, 1509, 1427, 1335, 1296, 1264,

4.9. General procedure for synthesis of amides

Amine 23 (1.2 mmol) was dissolved in a mixture of dry

CH₂Cl₂(5 mL) and TEA (1.2 mmol). The solution was

cooled with an ice-bath and the corresponding acyl chlo-

ride (1.35 mmol) was added dropwise. The ice-bath was

removed and the reaction was carried out at room tem-

perature for 10 h. Then, the solution was washed with

2 M HCl (40 mL), a 10% (w/v) solution of Na₂CO₃

(40 mL), and brine (25 mL). After drying with Na₂SO₄

and ﬁltration, the solvent was evaporated under reduced

pressure. Compounds 24 and 17 were puriﬁed by CC

(mobile phase: CH₂Cl₂). Amides 25 and 26 were recrys-

tallized, respectively, in diethyl ether/hexane (1:4) and

MeOH.

1189, 1157, 1122, 1066, 1027, 969, 867, 794, 687 cm^ꢀ1

.

4.9.4. Methyl 4-[(4-bromo-1-oxobutyl)amino]-2-[(3-tri-

ﬂuoromethylphenyl)amino]benzoate (27). Light brown so-

lid (96%); ¹H NMR (CDCl₃):

d 2.21 (m, 2H,

CH₂CH₂CH₂Br), 2.52 (t, J = 7 Hz, 2H, CH₂CH₂CH₂Br),

3.47 (t, J = 6.2 Hz, 2H, CH₂CH₂CH₂Br), 3.89 (s, 3H,

OCH₃), 6.89 (dd, J = 2, 8.8 Hz, 1H, H5), 7.28 (m, 1H,

ArH), 7.39–7.50 (m, 4H, ArH), 7.55 (d, J = 1.7 Hz, 1H,

H3), 7.92 (d, J = 8.7 Hz, 1H, H6), 9.70 (s, 1H, ArNHAr);

¹³C NMR (CDCl₃): d 27.8 (CH₂CH₂CH₂Br), 33.3

(CH₂CH₂CH₂Br), 33.6 (CH₂CH₂CH₂Br), 51.9 (OCH₃),

103.7 (C3), 108.6 (C5), 109.5 (C1), 118.6 (q, J = 4 Hz,

C2⁰), 120 (q, J = 4 Hz, C4⁰), 124.1 (q, J = 271 Hz, CF₃),

124.8 (C6), 130 (C6⁰), 131.9 (q, J = 32 Hz, C3⁰), 133.1

(C5⁰), 141.3 (C1⁰), 143.1 (C2), 147.9 (C4), 168.5 (COO),

170.3 (CON); IR (KBr): 3285 (N–H), 1675 (C@O),

1617, 1586, 1545, 1509, 1414, 1335, 1262, 1224, 1150,

4.9.1. Methyl 4-acetamido-2-[(3-triﬂuoromethylphenyl)

amino]benzoate (24). White solid (35%); ¹H NMR

(CDCl₃): d 2.06 (s, 3H, O@CCH₃), 3.87 (s, 3H, OCH₃),

6.89 (dd, J = 1.7, 8.7 Hz, 1H, H5), 7.17 (d, J = 7.5 Hz,

1H, ArH), 7.30 (t, J = 7.7 Hz, 1H, H5⁰), 7.38 (d,

J = 8.4 Hz, 1H, ArH), 7.42 (s, 1H, ArH), 7.59 (s, 1H,

ArH), 7.87 (d, J = 8.8 Hz, 1H, H6), 8.46 (br, 1H, CONH),

9.66 (s, 1H, ArNHAr); ¹³C NMR (CDCl₃) d 24.5 (CH₃),

51.4 (OCH₃), 103.7 (C3), 108.4 (C5), 109.6 (C1), 118.2 (q,

J = 4 Hz, C2⁰), 119.6 (q, J = 4 Hz, C4⁰), 124 (q,

J = 271 Hz, CF₃), 124.5 (C1⁰), 129.8 (C6⁰), 131.6 (q,

J = 32 Hz, C3⁰), 132.9 (C5⁰), 141.3 (C6), 143.5 (C2),

147.7 (C4), 168.5 (COO), 169.5 (CON); IR (KBr): 3275

(N–H), 1684 (C@O), 1602, 1545, 1504, 1426, 1336,

1117, 1095, 780, 697 cm^ꢀ1

.

4.10. General procedure for the nucleophilic substitutions

on compound 27

The bromoalkane 27 (1.1 mmol) was dissolved in DMF

(5 mL) and the nucleophile (NaN₃or potassium acetate,

3.3 mmol) was added. The suspension was stirred over-

night. Then, 50 mL of water was added and the mixture

was extracted two times with toluene (50 mL). The or-

ganic solution was washed with brine (50 mL), dried

with Na₂SO₄, ﬁltered, and evaporated under reduced

pressure. The solid obtained was recrystallized in diethyl

ether/petroleum ether_40–60(1:1).

1255, 1226, 1161, 1122, 1090, 1067, 880, 780, 694 cm^ꢀ1

.

4.9.2. Methyl 4-benzamido-2-[(3-triﬂuoromethylphenyl)

amino]benzoate (25). Slightly orange solid (78%); ¹H

NMR (CDCl₃+ 10% CD₃OD): d 3.89 (s, 3H, OCH₃),

7.02 (dd, J = 2.1, 8.7 Hz, 1H, H5), 7.26 (d, J = 6.3 Hz,

1H, ArH), 7.37–7.53 (m, 6H, ArH), 7.71 (d, J = 2 Hz,

1H, H3); 7.7 (d, J = 7.1 Hz, 2H, H2⁰⁰,H6⁰⁰), 7.95 (d,

J = 8.8 Hz, 1H, H6), 8.02 (br, 1H, CONH), 9.72 (s,

1H, ArNHAr); ¹³C NMR (CDCl₃+ 10% CD₃OD): d

51.9 (OCH₃), 104.1 (C3), 108.7 (C5), 109.9 (C1), 118.7

4.10.1. Methyl 4-[(4-acetoxy-1-oxobutyl)amino]-2-[(3-tri-

ﬂuoromethylphenyl)amino]benzoate (28). White solid

(63%); ¹H NMR (DMSO-d₆):

CH₂CH₂CH₂OOCMe), 1.97 (s, 3H, CH₃COO), 2.38 (t,

d

1.85 (m, 2H,

1718

P. Van Antwerpen et al. / Bioorg. Med. Chem. 16 (2008) 1702–1720

J = 7.5 Hz, 2H, CH₂CH₂CH₂OOCMe), 3.82 (s, 3H,

OCH₃), 4.01 (t, J = 6.4 Hz, 2H, CH₂CH₂CH₂OOCMe),

7.09 (dd, J = 1.9, 8.8 Hz, 1H, H5), 7.37 (m, 1H, ArH),

7.54–7.57 (m, 3H, ArH); 7.75 (d, J = 1.9 Hz, 1H, H3),

7.86 (d, J = 8.8 Hz, 1H, H6), 9.46 (s, 1H, ArNHAr),

10.14 (s, 1H, CONH); ¹³C NMR (DMSO-d₆): d 20.7

(CH₂CH₂CH₂OOCMe), 23.9 (CH₂CH₂CH₂OOCMe),

33 (CH₂CH₂CH₂OOCMe), 51.8 (OCH₃), 63.3

(CH₃COO), 103.7 (C3), 108 (C5), 109.8 (C1), 117 (q,

J = 4 Hz, C2⁰), 118.9 (q, J = 4 Hz, C4⁰), 124.1 (q,

J = 271 Hz, CF₃), 124.2 (C6), 130.5 (C6⁰), 131.9 (q,

J = 32 Hz, C3⁰), 132.4 (C5⁰), 141.7 (C1⁰), 144.6 (C2),

146.3 (C4), 167.4 (ArCOO), 170.3 (CON), 171.3 (Me-

COO); IR (KBr): 3340 (N–H), 1706 (C@O), 1676

(C@O), 1605, 1551, 1509, 1445, 1424, 1403, 1349,

1271, 1206, 1164, 1130, 1096, 1067, 1039, 869, 794,

124.6 (C6), 130 (C6⁰), 131.7 (q, J = 32 Hz, C3⁰), 133

(C5⁰), 141.5 (C1⁰), 144.1 (C2), 147.7 (C4), 168.6

(COO), 172.2 (CON); IR (KBr): 3290 (N–H), 2925,

1692 (C@O), 1599, 1532, 1504, 1420, 1324, 1243, 1154,

1116, 1084, 772, 695 cm^ꢀ1

.

4.11. Methods for measurement of the biological activity

4.11.1. Preparation of solutions for the diﬀerent assays.

Hydrogen peroxide (H₂O₂), KI, HCl, NaCl, KOH,

KH₂PO₄, EDTANa₂H₂, MgCl₂Æ6H₂O, NaHCO₃, dieth-

anolamine, polysorbate 80, and NaOH were obtained

from VWR (Leuven, Belgium). Bovine serum albumin

(BSA), 5,5⁰-dithio-bis-nitrobenzoic acid (DTNB),

NaN₃, paranitrophenyl phosphate, Tris(hydroxy-

methyl)aminomethane (Tris), methionine, and catalase

were purchased from Sigma (St. Louis, USA). Sodium

hypochlorite (NaOCl) was obtained from Acros Organ-

ics (New Jersey, USA).

696 cm^ꢀ1

.

4.10.2. Methyl 4-[(4-azido-1-oxobutyl)amino]-2-[(3-tri-

ﬂuoromethylphenyl)amino]benzoate (29). White solid

(quantitative); ¹H NMR (CDCl₃): d 1.95 (m, 2H,

CH₂CH₂CH₂N₃), 2.42 (t, J = 7.1 Hz, 2H, CH₂CH₂CH₂

N₃), 3.37 (t, J = 6.4 Hz, 2H, CH₂CH₂CH₂N₃), 3.89 (s,

3H, OCH₃), 6.89 (dd, J = 2, 8.7 Hz, 1H, H5), 7.28 (m,

1H, ArH), 7.39–7.46 (m, 4H, ArH), 7.54 (d, J = 1.7 Hz,

1H, H3), 7.93 (d, J = 8.8 Hz, 1H, H6), 9.70 (s, 1H, ArN-

HAr); ¹³C NMR (CDCl₃): d 24.5 (CH₂CH₂CH₂N₃), 34.3

(CH₂CH₂CH₂N₃), 50.7 (CH₂CH₂CH₂N₃), 51.9 (OCH₃),

103.7 (C3), 108.6 (C5), 109.5 (C1), 118.6 (q, J = 4 Hz,

C2⁰), 120 (q, J = 4 Hz, C4⁰), 124.1 (q, J = 271 Hz, CF₃),

124.8 (C6), 130 (C6⁰), 131.9 (q, J = 32 Hz, C3⁰), 133.2

(C5⁰), 141.3 (C1⁰), 143.1 (C2), 147.9 (C4), 168.5 (COO),

170.5 (CON); IR (KBr): 3285 (N–H), 2090 (N₃), 1675

(C@O), 1652 (C@O), 1617, 1583, 1546, 1509, 1439, 1414,

A PBS buﬀer (pH 7.4) was prepared at a ﬁnal concentra-

tion of 10 mM phosphate ions (KH₂PO₄/KOH) and

150 mM NaCl. The same phosphate buﬀer (pH 7.4)

was also prepared without NaCl. For LDL preparation

and oxidation, a PBS buﬀer at pH 7.2 was prepared with

a ﬁnal concentration of 2.8 mM of EDTANa₂H₂. A pH

7.5 Tris buﬀer saline (TBS 80) containing 50 mM of

Tris, 300 mM of NaCl, and 0.1% of polysorbate 80

was used during the ELISA. Finally, a pH 9.8 diethanol-

amine buﬀer was extemporarily made up by dissolving

0.101 g of MgCl₂.6H₂O and 0.2 g of NaN₃in water with

97 mL of diethanolamine. The pH was adjusted to 9.8

with HCl and the solution was diluted to 1 l. These

chemicals were of pro-analysis quality. A stock solution

of H₂O₂(30% w/w) was prepared and kept at 4 ꢂC. Be-

fore use, a working solution was prepared by dilution

(60 ll/50 mL water) and the concentration was mea-

sured using an extinction coeﬃcient of 43.6 M^ꢀ1cm^ꢀ1

at 240 nm.⁴²A stock solution of NaOCl (13% w/w)

was prepared and kept at 4 ꢂC. Before use, a working

solution was prepared by dilution (60 ll/50 mL water)

and HOCl concentration was determined at pH 6.2 by

iodometry with a 20 mM KI solution. Iodine formed

was measured at 350 nm and the concentration deter-

1349, 1261, 1222, 1165, 1115, 857, 801, 778, 698 cm^ꢀ1

.

4.10.3. Methyl 4-[(4-amino-1-oxobutyl)amino]-2-[(3-tri-

ﬂuoromethylphenyl)amino]benzoate (30). Compound 29

(830 mg, 2 mmol) was dissolved in MeOH (35 mL),

and 12 M HCl (0.16 mL, 2 mmol) and 5% Pd/C (0.1 g)

were added. The ﬂask was connected to a Parr hydroge-

nation apparatus and purged with hydrogen. The reac-

tion was carried out for 8 h under hydrogen

(400,000 kPa). Then the suspension was ﬁltered on Cel-

ite and the solvent evaporated under reduced pressure.

The residue was dissolved in water (50 mL) and this

solution was made alkaline (pH ꢂ 14) with a 30% (w/

v) solution of NaOH. After two extractions with CH₂Cl₂

(75 mL), the organic phase was washed with brine

(50 mL), dried with Na₂SO₄, ﬁltered, and evaporated

under reduced pressure. A yellow and viscous oil was

obtained (720 mg, 93%). The product was not puriﬁed;

¹H NMR (CDCl₃): d 1.62 (br, 2H, NH₂), 1.80 (m, 2H,

CH₂CH₂CH₂NH₂), 2.45 (t, J = 7 Hz, 2H, CH₂CH₂CH₂

NH₂), 2.81 (t, J = 6.2 Hz, 2H, CH₂CH₂CH₂NH₂), 3.88

(s, 3H, OCH₃), 6.98 (dd, J = 1.9, 8.7 Hz, 1H, H5), 7.26

(m, 1H, ArH), 7.39–7.78 (m, 3H, ArH); 7.58 (d,

J = 1.9 Hz, 1H, H3), 7.91 (d, J = 8.7 Hz, 1H, H6), 9.54

(br, 1H, CONH), 9.72 (s, 1H, ArNHAr); ¹³C NMR

(CDCl₃): d 27.9 (CH₂CH₂CH₂NH₂), 36.3 (CH₂CH₂CH₂

NH₂), 41.4 (CH₂CH₂CH₂NH₂), 51.9 (OCH₃), 103.5

(C3), 108.2 (C5), 109.6 (C1), 118.1 (q, J = 4 Hz, C2⁰),

119.6 (q, J = 4 Hz, C4⁰), 124.1 (q, J = 271 Hz, CF₃),

mined using an extinction coeﬃcient of 22,900 M^ꢀ1

cm^ꢀ1

.

A solution of 2-nitro-5-thiobenzoate (TNB)

43

was prepared daily as previously described and diluted

to 0.45 mM after the concentration was determined

using an extinction coeﬃcient of 13,600 M^ꢀ1cm^ꢀ1at

412 nm.⁴⁴A phosphate buﬀer (pH 7.4), containing a ﬁ-

nal concentration of 10 mM phosphate ions (KH₂PO₄/

KOH) and 0 or 300 mM NaCl, and an acetate buﬀer

(pH 3.0), containing 5.0% v/v acetic acid and 0.20%

m/v ammonium acetate, were prepared. De-oxygenated

milli-Q water served for the preparation of all solutions.

4.11.2. Preparation of the recombinant enzyme and

obtaining of LDL. Recombinant MPO was prepared as

previously described.⁴⁵Each batch of solution is charac-

terized by its protein concentration (mg/mL), its activity

(U/mL), and its speciﬁc activity (U/mg). The chlorina-

tion activity was determined according to Hewson and

Hager.⁴⁶Human plasma served for the isolation of

P. Van Antwerpen et al. / Bioorg. Med. Chem. 16 (2008) 1702–1720

1719

LDL by ultracentrifugation according to Havel et al.⁴⁷

Before oxidation, the LDL fraction (1.019 < d <

1.067 g/mL) was desalted by two consecutive passages

through PD10 gel-ﬁltration columns (Amersham Bio-

sciences, The Netherlands) using PBS buﬀer. The

diﬀerent steps were carried out in the dark and the

protein concentration was measured by the Lowry assay

for both MPO and LDL.⁴⁸

ferent conditions consisting in the presence or the ab-

ꢀ

`

sence of Cl . The method was adapted from Neve

et al.¹¹In a 1.5 mL quartz cell, the following reagents

were introduced, at a ﬁnal concentration stated between

brackets, in a ﬁnal volume of 1.0 mL: 100 ll of MPO

(ꢂ767 nM), 400 ll of pH 7.4 KH₂PO₄/KOH phosphate

buﬀer (10 mM) with or without NaCl (300 mM or

0 mM), 250 ll of a drug solution (500 lM), and 50 ll

of water. The reaction was initiated by addition of

100 ll of H₂O₂(30 lM) and absorbances were simulta-

neously monitored with a diode-array spectrophotome-

ter (Agilent 8453, Palo Alto, CA, USA) at wavelength

characteristic for compound II (456/575 nm) and of

the native enzyme (430/625 nm). Compound II lifetime

was measured as the time for intersection of absorbance

curves at 430 nm and 456 nm or 575 nm and 625 nm in

the presence of a drug inducing an inhibition of MPO-

chlorinating activity. This time corresponds to the com-

plete conversion of compound II to native enzyme,

which is demonstrated by a shift in the enzyme spec-

trum. The results are the means of three-independent

values and are expressed as a relative value compared

to the coumpound II lifetime in the presence of ﬂufen-

amic acid. In this assay, a MPO batch with the following

characteristics was used: 0.644 mg/mL, 70.8 U/mg,

46 U/mL.

4.11.3. Inhibition of myeloperoxidase-chlorinating activ-

ity. The measurement of the inhibition of MPO-chlori-

nating activity was performed as previously

described.¹²The method quantiﬁes the amount of tau-

rine chloramine produced by the MPO/H₂O₂/Cl^ꢀsystem

in the presence of several concentrations of compounds.

The reaction mixture contained the following reagents,

at the concentrations stated between brackets, in a ﬁnal

volume of 1.0 mL: 5 ll of recombinant myeloperoxidase

(ꢂ25 nM), pH 7.4 phosphate buﬀer (PO₄^3ꢀ10 mM/NaCl

300 mM), taurine (15 mM), and ﬂufenamic acid analogs

(0–20 lM). When necessary, the volume was adjusted to

1.0 mL with water. The mixture was incubated at 37 ꢂC

and the reaction was initiated by the addition of H₂O₂

(100 lM). After 5 min, the reaction was stopped by the

addition of 100 ll of catalase (4 U/ll). Finally, the quan-

tity of taurine chloramine was measured by the addition

of 750 ll of 0.45 mM TNB and 2150 ll of water. The re-

sults were expressed as means SD for at least 3 diﬀerent

determinations for each examined drug and by taking the

absence of inhibitor as the 0% and the absence of H₂O₂

as the 100%. The IC₅₀was calculated by establishing a

sigmoid curve of the % of MPO inhibition as a function

of compound concentration. The signiﬁcance of the

mean inhibitory curve for the inhibition in the MPO-

chlorinating activity assay was tested by ﬁtting the curve

with a three-parameter sigmoid model. Results were con-

sidered as statistically signiﬁcant at P < 0.05. In this as-

say, a myeloperoxidase batch with the following

characteristics was used: 0.42 mg/mL, 60 U/mg, 25.2 U/

mL. The scavenging activity toward HOCl was simulta-

neously monitored at relevant pharmacological concen-

trations. The reaction mixture contained the following

reagents, at the concentrations stated between brackets,

in a ﬁnal volume of 1.0 mL: pH 7.4 phosphate buﬀer

4.11.5. Inhibition of LDL oxidation. The LDL oxidation

was carried out at 37 ꢂC in a ﬁnal volume of 500 ll. The

reaction mixture contained the following reagents at the

ﬁnal concentrations indicated between brackets: pH 7.2

PBS buﬀer, MPO (1 lg/mL), LDL (1000 lg/mL), 2 ll of

HCl 1 N (4 mM), one of the drugs at diﬀerent concen-

trations (5, 15, 30, 300 lM), and H₂O₂(100 lM). The

reaction was stopped after 5 min by cooling the tubes

in ice.

In order to speciﬁcally measure the LDL oxidation due

to the MPO/H₂O₂/Cl^ꢀsystem, a recently developed

ELISA was used, based on the binding between a mouse

monoclonal antibody (Mab AG9) speciﬁcally recogniz-

ing the oxidized APO B-100 on LDL and an anti-mouse

immunoglobulin G (Ig G) coupled with alkaline phos-

phatase.⁷However, until now, the chemical structure

of the epitope recognized by our antibodies has not been

elucidated. The amount of oxidized LDL is quantiﬁed

by measurement of the phosphatase activity and ex-

pressed as the absorbance value of the reaction product

of the alkaline phosphatase, that is, para-nitrophenol

(405 nm).

3ꢀ

(PO₄10 mM/NaCl 300 mM), taurine (15 mM), and

ﬂufenamic acid analogs (0.5, 1, 2, 4 lM). When neces-

sary, the volume was adjusted to 1.0 mL with water.

The mixture was incubated at 37 ꢂC and the reaction

was initiated by the addition of HOCl (60 lM). After

5 min, 100 ll of catalase (4 U/ll) was added. Finally,

the quantity of taurine chloramine was measured by

the addition of 750 ll of 0.45 mM TNB and 2150 ll of

water. The percentages of HOCl scavenged were calcu-

lated by taking the absence of inhibitor as the 0% and

the absence of HOCl as the 100%. The results were ex-

pressed as means SD of the percentage of HOCl scav-

enged for at least ﬁve diﬀerent determinations for each

examined drug.

The assay was performed as described by Moguilevsky

et al.⁶in a NUNC maxisorp plate (VWR, Zaventem,

Belgium): 200 ng/well of LDL was coated overnight at

4 ꢂC in a sodium bicarbonate pH 9.8 buﬀer (100 ll).

Afterward, the plate was washed with TBS 80 buﬀer

and then saturated during 1 h at 37 ꢂC with the PBS buf-

fer containing 1% of BSA (150 ll/well). After washing

the wells twice with the TBS 80 buﬀer, the monoclonal

antibody Mab AG9 (200 ng/well), obtained according

to a standard protocol and as previously described,⁶

was added as a diluted solution in PBS buﬀer with

0.5% of BSA and 0.1% of Polysorbate 20. After incuba-

4.11.4. Accumulation of compound II in the presence or

the absence of chloride. The interaction of the drugs with

the native enzyme and its diﬀerent oxidized forms was

assessed by measuring compound II lifetime in two dif-

1720

P. Van Antwerpen et al. / Bioorg. Med. Chem. 16 (2008) 1702–1720

17. Friesner, R. A.; Banks, J. L.; Murphy, R. B.; Halgren, T.

A.; Klicic, J. J.; Mainz, D. T.; Repasky, M. P.; Knoll, E.

H.; Shelley, M.; Perry, J. K.; Shaw, D. E.; Francis, P.;

Shekin, P. S. J. Med. Chem. 2004, 47, 1739.

18. Poulos, T. L.; Kraut, J. J. Biol. Chem. 1980, 255, 8199.

19. Sadighi, J. P.; Harris, M. C.; Buchwald, S. L. Tetrahedron

Lett. 1998, 39, 5327.

20. Hosangadi, B. D.; Dave, R. H. Tetrahedron Lett. 1996, 37,

6375.

21. Su, Q.; Panek, J. S. J. Am. Chem. Soc. 2004, 126,

2425.

22. Chang, J.; Guo, X.; Cheng, S.; Guo, R.; Chen, R.; Zhao,

K. Bioorg. Med. Chem. Lett. 2004, 14, 2131.

23. Hartwell, J. L. In Organic Syntheses; Wiley: New York,

1944; Vol. 24, pp 22–24.

24. Perez-Alvarez, M.; Raymo, F. M.; Rowan, S. J.; Shiraldi,

D.; Fraser Stoddart, J.; Wang, Z.-H.; White, A. J. P.;

Williams, D. J. Tetrahedron 2001, 57, 3799.

25. Nishimura, S. In Handbook of Heterogeneous Catalytic

Hydrogenation for Organic Synthesis; Wiley: New York,

2001; p 584.

tion during 1 h at 37 ꢂC, the plate was washed four times

with the TBS 80 buﬀer and a 3000 times diluted solution

of Ig G anti-mouse Alkaline Phosphatase (Promega,

Leiden, The Netherlands) in the same buﬀer was added

(100 ll/well). The wells were washed again four times

and a revelation solution (150 ll/well) containing 5 mg

of paranitrophenyl phosphate in 5 mL of diethanola-

mine buﬀer was added for 30 min at room temperature.

The reaction was stopped with 60 ll/well of NaOH 3 N

solution. The measurement of the absorbance was per-

formed at 405 nm with a background correction at

655 nm with a Bio-Rad photometer for a 96-well plate

(Bio-Rad laboratories, CA, USA). Results were ex-

pressed as means SD of the percentage of LDL oxida-

tion for six-independent measurements.⁶

Acknowledgment

´

Martine Prevost is Maıtre de recherche au FNRS (Bel-

gian Research National Fund).

ˆ

26. Carrico, D.; Ohkanda, J.; Kendrick, H.; Yokoyama, K.;

Blaskovitch, M. A.; Bucher, C. J.; Buckner, F. S.; Van

Voorhis, W. C.; Chakrabarti, D.; Croft, S. L.; Gelb, M.

H.; Sebti, S. M.; Hamilton, A. D. Bioorg. Med. Chem.

2004, 12, 6517.

References and notes

27. Marquez, L. A.; Huang, J. T.; Dunford, H. B. Biochem-

istry 1994, 33, 1447.

28. Arnhold, J.; Furtmuller, P. G.; Obinger, C. Redox Rep.

2003, 8, 179.

29. Furtmuller, P. G.; Burner, U.; Jantschko, W.; Regelsber-

¨

1. Klebanoﬀ, S. J. J. Leukoc. Biol. 2005, 77, 598.

2. Nurcombe, H. L.; Bucknall, R. C.; Edwards, S. W. Ann.

Rheum. Dis. 1991, 50, 237.

¨

ger, G.; Obinger, C. FEBS Lett. 2000, 484, 139.

30. Kettle, A. J.; Winterbourn, C. C. Biochem. J. 1988, 252,

529.

31. Burner, U.; Jantschko, W.; Obinger, C. FEBS Lett. 1999,

443, 290.

32. Furtmuller, P. G.; Obinger, C.; Huanyu, Y.; Dunford, H.

¨

3. Malle, E.; Waeg, G.; Schreiber, R.; Gro¨ne, E. F.; Sattler,

W.; Gro¨ne, H.-J. Eur. J. Biochem. 2000, 267, 4495.

4. Choi, D.-K.; Pennathur, S.; Perier, C.; Tieu, K.; Teis-

mann, P.; Wu, D.-C.; Jackson-Levis, V.; Vila, M.;

Vonsattel, J. P.; Heinecke, J. W. J. Neurosci. 2005, 25,

6594.

B. Eur. J. Biochem. 2000, 267, 5858.

33. Sadowski, J.; Gasteiger, J. Chem. Rev. 1993, 93, 2567.

34. Eldridge, M. D.; Murray, C. W.; Auton, T. R.; Paolini, G.

V.; Mee, R. P. J. Comput. Aided Mol. Des. 1997, 11, 425.

35. von Boltze, K.-H.; Ba¨cker, U.; Kreisfeld, H. Arzneim.-

Forsch. 1982, 32, 183.

5. Carr, A. C.; Decker, E. A.; Park, Y.; Frei, B. Free Radic.

Biol. Med. 2001, 31, 62.

6. Moguilevsky, N.; Zouaoui Boudjeltia, K.; Babar, S.;

´

Delree, P.; Legssyer, I.; Carpentier, Y.; Vanhaeverbeek,

M.; Ducobu, J. Biochem. Biophys. Res. Commun. 2004,

323, 1223.

36. Williams, F. J.; Relles, H. M.; Manello, D. P. E. J. Org.

Chem. 1977, 42, 3419.

37. Burkartsmaier, A.; Mutschler, E. Arch. Pharm. 1978, 311, 161.

38. Superimposed with solvent signal.

39. Signal for C5⁰overlapped by another one between 126 and

130 ppm.

40. Since this reaction was strongly exothermic, it is recom-

mended to carefuly follow the described procedure.

41. J_C–Fimpossible to determine because of superimposed

signals.

42. Beers, R. J.; Sizer, I. W. J. Biol. Chem. 1952, 195,

133.

43. Gressier, B.; Lebegue, N.; Brunet, C.; Luyckx, M.; Dine,

T.; Cazin, M.; Cazin, J. C. Pharm. World Sci. 1995, 3, 76.

44. Ching, T. L.; De Jong, J.; Bast, A. Anal. Biochem. 1994,

218, 377.

45. Moguilevsky, N.; Garcia-Quintana, L.; Jacquet, A.;

Tournay, C.; Fabry, L.; Pierard, L.; Bollen, A. Eur. J.

Biochem. 1991, 197, 605.

7. Zouaoui Boudjeltia, K.; Legssyer, I.; Van Antwerpen, P.;

Kisoka, R. L.; Babar, S.; Moguilevsky, N.; Delree, P.;

Ducobu, J.; Remacle, C.; Vanhaeverbeek, M.; Brohee, D.

Biochem. Cell Biol. 2006, 84, 805.

8. Zheng, L.; Settle, M.; Brubaker, G.; Schmitt, D.; Hazen,

S. L.; Smith, J. D.; Kinter, M. J. Biol. Chem. 2005, 280, 38.

9. Soyer, Z.; Bas, M.; Pabuccuoglu, A.; Pabuccuoglu, V.

Arch. Pharm. 2005, 338, 405.

10. Kettle, A.; Tiden, A.-K. PCT Int. Pat. 2005037835; 2005.

`

11. Neve, J.; Parij, N.; Moguilevsky, N. Eur. J. Pharmacol.

2001, 417, 37.

12. Van Antwerpen, P.; Dufrasne, F.; Lequeux, M.; Zouaoui

Boudjeltia, K.; Legssyer, I.; Babar, S.; Moreau, P.;

Moguilevsky, N.; Vanhaeverbeek, M.; Ducobu, J.; Neve,

J. Eur. J. Pharmacol. 2007, 570, 235.

13. Blair-Johnson, M.; Fiedler, T.; Fenna, R. Biochemistry

2001, 40, 13990.

14. Davey, C. A.; Fenna, R. E. Biochemistry 1996, 35, 10967.

15. Hardy, L. W.; Malikayil, A. Curr. Drug Discov. 2003, 3,

15.

16. Van Antwerpen, P.; Zouaoui Boudjeltia, K.; Babar, S.;

Legssyer, I.; Moreau, P.; Moguilevsky, N.; Vanhaever-

46. Hewson, W. D.; Hager, L. P. J. Biol. Chem. 1979, 254,

3175.

47. Havel, R. J.; Eder, H. A.; Bragdom, J. H. J. Clin. Invest.

1955, 34, 1345.

48. Peterson, G. L. Anal. Biochem. 1977, 83, 346.

`

beek, M.; Ducobu, J.; Neve, J. Biochem. Biophys. Res.

Commun. 2005, 337, 82.

Article Doi

Conception of myeloperoxidase inhibitors derived from flufenamic acid by computational docking and structure modification

DOI: 10.1016/j.bmc.2007.11.025

Source and publish data:

Authors:

Article abstract of DOI:10.1016/j.bmc.2007.11.025

Full text of DOI:10.1016/j.bmc.2007.11.025

Products guided by the article

R&D Labs maybe for 1019854-39-7

Relevant to this article

Hot Product