From serendipity to rational antituberculosis drug discovery of mefloquine-isoxazole carboxylic acid esters

Article abstract of DOI:10.1021/jm900340a

Both in vitro and in vivo metabolism studies suggested that 5-(2,8-bis(trifluoromethyl)quinolin-4-yloxymethyl)isoxazole-3-carboxylic acid ethyl ester (compound 3) with previously reported antituberculosis activity is rapidly converted to two metabolites 3

Full text of DOI:10.1021/jm900340a

6966 J. Med. Chem. 2009, 52, 6966–6978
DOI: 10.1021/jm900340a
From Serendipity to Rational Antituberculosis Drug Discovery of
Mefloquine-Isoxazole Carboxylic Acid Esters
Jialin Mao,†, ,f Hai Yuan,§, ,ꢀ Yuehong Wang,† Baojie Wan,† Marco Pieroni,§ Qingqing Huang,§ Richard B. van Breemen,‡
Alan P. Kozikowski,*,§ and Scott G. Franzblau*,†
†Institute for Tuberculosis Research, College of Pharmacy and ‡Department of Medicinal Chemistry and Pharmacognosy and §Drug Discovery
Program, Department of Medicinal Chemistry and Pharmacognosy, University of Illinois at Chicago, 833 South Wood Street, Chicago, Illinois
60612. Equal contribution to this work. f Current address: Department of Drug Disposition, Eli Lilly and Company, Indianapolis, IN 46285.
ꢀ Current address: Department of Medicinal Chemistry, PsychoGenics Inc., Tarrytown, NY 10591.
Received March 17, 2009
Both in vitro and in vivo metabolism studies suggested that 5-(2,8-bis(trifluoromethyl)quinolin-
4-yloxymethyl)isoxazole-3-carboxylic acid ethyl ester (compound 3) with previously reported anti-
tuberculosis activity is rapidly converted to two metabolites 3a and 3b. In order to improve the metabolic
stability of this series, chemistry efforts were focused on the modification of the oxymethylene linker of
compound 3 in the present study. Compound 9d with an alkene linker was found to be both more
metabolically stable and more potent than compound 3, with a minimum inhibitory concentration
(MIC) of 0.2 μM and 2.6 μM against replicating and nonreplicating Mycobaterium tuberculosis,
respectively. These attributes make 9d an interesting lead compound. A number of modifications were
made to the structure of 9d, and a series of active compounds were discovered. Although some
neurotoxicity was observed at a high dosage, this new series was endowed with both improved in vitro
anti-TB activity and metabolic stability in comparison to compound 3.
Introduction
0.9 μM and 12.2 μM, respectively.4 Its activity appeared to be
specific for organisms of the Mtb complex, and it effected
significant reductions in colony forming units (CFU) in
infected macrophages with an EC90 of 4.1 μM.5 In this work,
rational modifications of compound 3, based on in vitro and in
vivo assessment of its metabolic stability, are reported.
Tuberculosis (TBa) is presently regarded as one of the most
dangerous infectious diseases worldwide and one of the major
AIDS-associated infections. It has been estimated that more
than 9.2 million people worldwide develop active TB, and 1.7
million die every year.1 Despite these figures, no new anti-TB
drugs havebeenintroduced overthe past40years.2 Moreover,
the current six-month treatment regimen required to cure the
disease is accompanied by significant toxicity, making patient
compliance difficult, and leading to the development of drug
resistance. Successful treatment of multidrug resistant
(MDR)-TB and extensively drug resistant (XDR)-TB is even
more challenging and requires even longer term treatment.3
Therefore, new drugs are urgently needed for the treatment of
TB. 5-(2,8-Bis(trifluoromethyl)quinolin-4-yloxymethyl)iso-
xazole-3-carboxylic acid ethyl ester (compound 3, structure
shown in Figure 1), reported earlier by us as a promising
antituberculosis lead, exhibited excellent activity against both
replicating and nonreplicating Mycobacterium tuberculosis
(Mtb), with minimum inhibitory concentrations (MIC) of
Metabolism of Compound 3 in Mouse and Human Liver
Microsomes. After 30 min incubation of compound 3 with
mouse and human liver microsomes in the presence of
NADPH, two metabolite peaks were observed eluting at
16.1 and 24.7 min (3a and 3b, named according to their
retention times; Figure 1). Both were characterized by LC-
MS-MS. The protonated form of 3a was observed at m/z 282.
On the basis of the nitrogen rule, 3a was suspected to have
only one nitrogen atom. A 20 mass unit loss was observed
several times in the tandem mass spectrum of 3a (Figure 2,
282 to 262, 214 to 194, 140 to 120), which is a characteristic
loss of HF. This suggested that the structure of 3a was related
to the quinoline. Thus, the structure of 3a was proposed to be
2,8-bis(trifluoromethyl)quinolin-4-ol (Figure 2). The proto-
nated molecule of 3b was observed at m/z 407, which is 28
mass units less than compound 3, suggesting the loss of an
ethyl group. The only possible site to lose an ethyl group in
the structure is from the ester portion. The structure of 3b
was thus proposed to be the corresponding acid (Figure 2).
After comparing the retention times and fragment patterns
with those of the authentic compounds, the structures were
confirmed as proposed. That these two metabolites were
generated during the incubation of compound 3 in both
mouse and human liver microsomes in the absence of
NADPH (Figure 1) suggests the involvement of NADPH-
independent enzymes. It was also noted that the acid 3b can
*To whom correspondence should be addressed. For S.G.F.: phone,
þ1-312-355-1715; fax, þ1-312-355-2693. E-mail: sgf@uic.edu. Address:
Institute for Tuberculosis Research, College of Pharmacy, University of
Illinois at Chicago, 833 South Wood Street, Chicago IL 60612, USA.
For A.P.K.: phone, þ1-312-996-7577; fax, þ1-312-996-7107; E-mail:
kozikowa@uic.edu. Address: Drug Discovery Program, Department of
Medicinal Chemistry and Pharmacognosy, University of Illinois at
Chicago, 833 South Wood Street, Chicago, IL 60612, USA.
a Abbreviations: CMC, carboxymethyl cellulose; INH, isoniazid;
LORA, low oxygen recovery assay; MABA, microplate Alamar Blue
assay; MIC, minimum inhibitory concentration; MDR-TB, multidrug-
resistant tuberculosis; Mtb, Mycobacterium tuberculosis; RMP, rifam-
pin; TB, tuberculosis; XDRTB, extensively drug-resistant tuberculosis.
r
pubs.acs.org/jmc
Published on Web 10/28/2009
2009 American Chemical Society

Article
Journal of Medicinal Chemistry, 2009, Vol. 52, No. 22 6967
Figure 1. Metabolic profile of compound 3 in mouse and human liver microsomes.
Figure 2. CID product ion mass spectra of metabolites 3a and 3b.
be metabolized into 3a in both human and mouse liver
microsomes. Interestingly, both metabolites were also ob-
served after 30 min incubation of compound 3 in human
serum (data not shown).
Plasma Concentrations of Compound 3 and Metabolites.
BALB/c mice were dosed orally with a suspension of com-
pound 3 at 400 mg/kg in carboxymethyl cellulose (CMC),
and the plasma concentration of 3 and its metabolites were
analyzed by LC-MS-MS. Consistent with the in vitro liver
microsomal data, 3a was the major metabolite detected in the
mouse plasma. The Cmax and Tmax of 3a were 2.71 μg/mL
and 30 min, respectively. The acid 3b could be detected at
5 min postdose. Unlike 3a, the plasma concentration of 3b
(about 0.1 μg/mL) did not change appreciably over the 3 h
time course, attaining a peak concentration of 0.40 μg/mL at
30 min. The highest plasma concentration of the parent
compound 3 (0.026 μg/mL) occurred at 90 min postdose
(Figure 3).

6968 Journal of Medicinal Chemistry, 2009, Vol. 52, No. 22
Mao et al.
Figure 3. Compound 3 and its metabolites in plasma after oral dosing of mice.
Scheme 1. Synthesis of Analogues in Table 1
the low oxygen recovery assay (LORA)8 were poorly affected
by the introduction of a racemic methyl group (compound
6a) and a dimethyl group (compound 6b). Insertion of more
bulky groups such as a butyl group (compound 6c) and a
benzyl group (compound 6d) led to a remarkable loss of
activity.
Replacement of the Oxymethylene Linker. Removal of the
oxymethylene linker between the quinoline and the isoxazole
ring suggested the synthesis of compound 9a. The intermedi-
ate 8 was prepared from 4-bromo-2,8-bis(trifluoromethyl)-
quinoline (7) by reaction with trimethylsilylacetylene in the
presence of Pd(PPh3)4 and CuI as catalysts, followed by
treatment with KOH in methanol. The NOC reaction
afforded 5-(2,8-bis(trifluoromethyl)quinolin-4-yl)isoxazole-
3-carboxylic acid ethyl ester (9a) (Scheme 2). Although
compound 9a was stable as its corresponding acid form in
mouse liver microsomes (Figure 5), it exhibited a 17-fold
decrease in its anti-TB potency (MABA MIC = 16.5 μM,
Table 2).
Next, the oxymethylene linker was replaced by an amide
linker. 5-Hydroxymethylisoxazole-3-carboxylic acid ethyl
ester was oxidized to its corresponding acid form (com-
pound 11) by use of Jone’s reagent. The carbonyl chloride
(compound 12) was formed from compound 11 by treatment
with oxalyl chloride. Compound 9b was then generated by
condensation of compound 12 and 2,8-bis(trifluoromethyl)-
quinolin-4-ylamine (Scheme 2). Unfortunately, compound
9b was not metabolically stable, as 2,8-bis(trifluoromethyl)-
quinolin-4-ylamine was observed as the major metabolite
after 30 min incubation in mouse liver microsomes(Figure 5).
In addition, it showed less potency with a MABA MIC of
29.5 μM (Table 2).
a. subsitituted propargyl bromides, K2CO3, acetone, reflux, over-
night. b. ethyl chlorooximidaocetate, Et2O, Et3N, 6 h through syringe
pump.
Neither of these metabolites exhibited antituberculosis
activity in microplate Alamar Blue assay (MABA)6 at pH
6.8. Therefore, in order to obtain in vivo efficacy in this series,
chemistry efforts were focused on preventing the formation
of 3a and 3b.
In order to avoid the biotransformation of 3b, the ethyl
ester was replaced by various ester bioisosteres, including the
isomer ester, amides, and the oxadiazole analogues. None of
them showed any antituberculosis activity (unpublished
data).5 These observations suggested that the ester may play
a role as a prodrug, and the ethyl ester moiety was retained in
the following modifications.
Metabolite 3a was suspected to form through an oxidation
step (Figure 2). Presumably, after the oxidation at the R-
position of the isoxazole, 3a is formed by breakdown of the
hemiacetal. Therefore, several modifications on the oxy-
methylene linker of compound 3 were made with the intention
of (1) increasing the steric hindrance at the soft spot and (2)
replacing the oxymethylene linker with surrogate moieties.
Increasing Steric Hindrance at the Soft Spot. Various
alkynes (5a-d) were prepared from 2,8-bis(trifluoromethyl)-
quinolin-4-ol by reaction with substituted propargyl bro-
mides in the presence of K2CO3 in anhydrous acetone. Next,
nitrile oxide cycloaddition chemistry (NOC)7 was performed
by reacting the dipolarophiles 5a-d with ethyl chlorooximi-
doacetate in the presence of triethylamine to afford com-
pounds 6a-d (Scheme 1). Remarkably, these modifications
precluded the formation of 3a when compounds 6a-d were
incubated in mouse liver microsomes for 30 min (Figure 4).
However, the MICs (Table 1) evaluated with the MABA and
A further modification involved replacement of the oxy-
methylene linker by an all carbon linker. Initially, the quino-
line ring without any additional substitution was used to
quickly compare the SAR with our previous SAR of the
oxymethylene linker series.5 The triphenylphosphonium
ylide (14) was prepared from triphenylphosphine by reaction
with 5-bromomethylisoxazole-3-carboxylic acid ethyl ester
(13), a bromination product of 5-hydroxymethylisoxazole-
3-carboxylic acid ethyl ester. Compounds 9c and 9d were
then generated through the Wittig reaction. Compound 9d

Article
Journal of Medicinal Chemistry, 2009, Vol. 52, No. 22 6969
Figure 4. Metabolic profile of compounds 6a-d in mouse liver microsomes.
was catalytically hydrogenated to afford compound 9e
(Scheme 2).
performed for 9d. Although the parent compound was
hydrolyzed to the acid within 5 min during the incubation
with mouse liver microsomes, 80% of the acid remained after
30 min incubation (Figure 6). In contrast to the lack of
antituberculosis activity of compound 3 acid, compound 9d
acid was active against M. tuberculosis H37Rv in MABA at
pH = 6.8 (MIC = 12.5 μM) (Table 2).
In order to evaluate the pharmacokinetic behavior of the
corresponding acid of compound 9d, the parent compound was
administered using the same formulation and oral dosage as
compound 3 (Table 3). Although the Cmax of 9d (0.03 μg/mL)
was similar to that found for compound 3, the Cmax of the
corresponding acid of compound 9d (16.4 μg/mL, 61.7 μM,
Surprisingly, compounds 9c and 9d (Table 2), endowed
with an alkene linker between the 4-position of the quinoline
ring and the 5-position of the isoxazole, had markedly
improved antituberculosis activity compared with com-
pound 3, and the trans isomer 9d (MABA MIC = 0.2 μM)
was found to have a MIC only 2-fold higher than rifampin,
the most active anti-TB drug used today. Furthermore, 9d
showed better activity than the corresponding cis isomer 9c
(MABA MIC = 0.6 μM), while the hydrogenation of the
alkene (compound 9e) resulted in a sharp decrease in potency
with an MIC of 2.7 μM. A metabolic stability assay was

6970 Journal of Medicinal Chemistry, 2009, Vol. 52, No. 22
Mao et al.
Table 1. In Vitro Activity of Mefloquine-Based Analogues Against M. tuberculosis H37Rv
MIC (μM)
IC50 (μM)
SI
Compd
Structure R =
MABA
LORA
Vero
MABA
LORA
ClogP
1
RMP
3
Mefloquine
Rifampin
H
13
0.1
7
2
11
122
0.9
1220
>144
>30.5
0.6
1.6
61
3.6
-
0.9
4.2
12.2
20.7
22.1
11.9
>128
>128
>128
>10.5
>6.2
1.0
4.4
4.8
5.3
6.3
6.0
6a
6b
CH3
(CH3)2
(CH2)3CH3
C6H5
6.8
115.8
>128
23.0
42.7
>128
6c
6d
0.3
-
3.6
-
Figure 5. Metabolic profile of compounds 9a,b in mouse liver microsomes.
almost 4-fold higher than its MIC at pH 6.8) was 60-fold
higher than that of the corresponding acid of compound 3
(Table 3). The acid of compound 9d also had a slightly
longer half-life compared to that of compound 3 (1.9 h vs
1.6 h). Overall, the area under the curve (AUC) of the acid
derived from 9d was 70-fold higher than that of the acid
derived from 3.
These results indicate that, in addition to its improved
in vitro anti-TB activity, compound 9d is more metabolically
stable than compound 3, thus achieving our initial goals in
modifying the linker of the latter.
New Hybrids. Since the new alkene linker demonstrated
both reasonable metabolic stability and an improved in
vitro antituberculosis activity, we expected to design even
more potent anti-TB ligands than compound 9d in this
alkene linker series by applying insights gained from the
previous SARs studies derived from the oxymethylene
linker series.5
Various alkene linker analogues 9f-h and 9j-l were
generated through application of the Wittig reaction. Since
some of required aldehydes were not commercially available,
they were synthesized by two different routes with relatively
low yields. Quinoline-6-carbaldehyde (16) was prepared
from 6-methylquinoline by the Riley reaction with selenium
dioxide at a high temperature, while aldehydes 17, 20 and 21
were prepared by treatment of their corresponding bromides
with n-butyllithium and then DMF (Scheme 3).
Based on the previously observed SAR with the oxy-
methylene linker,5 the alkene linker was moved from the 4-
position to the 6-position. The resulting cis isomer 9f
(MABA MIC = 1.3 μM) was slightly less active than the
trans isomer 9g (MABA MIC = 0.9 μM); however, both
were about 4-6-fold less active than their corresponding
4-substituted compounds 9c (MABA MIC = 0.6 μM) and
9d (MABA MIC = 0.2 μM), respectively. These results
suggested that the SAR gained from the oxymethylene
linker might not be fully translatable to the new alkene
linker series.
In order to confirm this, with the linker attached to the
4-position of the quinoline ring, the corresponding bis-
(trifluoromethyl) analogue 9h was prepared. According to
our previous SAR, addition of two trifluoromethyl groups at

Article
Journal of Medicinal Chemistry, 2009, Vol. 52, No. 22 6971
Scheme 2. Synthesis of Analogues in Table 2
a. (i) Me3Si-CCH, Pd (PPh3)4, CuI, i-Pr2NH, (ii) KOH/MeOH, 70%. b. ethyl chlorooximidaocetate, Et2O, TEA, 6 h through syringe pump, 50%.
c. Jone’s reagent, acctone, 55 °C, 2 h. d. oxalyl chloride, CH2Cl2, rt, 3 h. e. 2,8-bis(trifluoromethyl)quinolin-4-amine, DIEA, CH2Cl2, 0 °C to rt,
overnight. f. CBr4/PPh3, CH2Cl2, -15 °C, 2 h. g. PPh3, acetonitrile, reflux, 2 h. h. NaHDMS, quinoline-4-carbaldehyde, THF, -78 °C to rt, 4 h. i. H2, Pd/
C, EtOAc, rt, 8 h.
the 2- and 8-positions of the quinoline ring boosted the
antituberculosis activity more than 20-fold in the oxymethy-
lene linker series.5 In distinct contrast to the previous results,
compound 9h (MABA MIC = 2.7 μM) was 13-fold less
active than the unsubstituted compound 9d (MABA MIC =
0.2 μM). This trend was further observed in the saturated
linker series as compound 9i with 2,8-bis(trifloromethyl)
groups (MABA MIC = 5.7 μM) was less active than the
unsubstituted compound 9e (MABA MIC = 2.7 μM).
These results indicate that the new series of analogues with
carbon linkers exhibit a considerably different SAR compared
with what we obtained from the previous oxymethylene linked
analogues. To further explore the SAR of this new series,
compounds 9j-k with an electron-donating methoxyl substitu-
tion and compound 9l with 2,6-bis(trifluoromethyl) substitutes
were prepared and tested. The MABA MIC values of
compound 9k and 9l (both at 0.4 μM) were on the same
order as derivative 9d, which were 7-fold more potent than
the 2,8-bis(trifluoromethyl) compound 9h (MABA MIC =
2.7 μM). In addition, compound 9k was shown to be very
potent in the LORA assay as well (LORA MIC = 1.1 μM).
On the basis of the SAR obtained from 9c-l, it is suggested
that an alkene linker is important for activity, and a trans-
alkene is more favored. The position(s) of the substitution(s)
on the quinoline ring may also affect the activity to some
extent, while the electron effect of the substitution(s) is less
important. A pharmacophore model with the target protein
structure may help to provide a more explicit SAR. On the
basis of a literature search, peptide deformylase (PDF) and
ATPase were considered as promising targets for the series.
However, our preliminary studies suggested that neither
of these are the target for the series, and there is no evidence
to suggest an alternative target at this time. In addition,
transposon mutants with resistance to compound 3 all
mapped to genes with no known function (S. Warit, personal
communication). Once the target has been determined, a
systematic exploration of this carbon linker series will be
carried out with the guidance of a pharmacophore model.
The most promising compound 9d was chosen for an in
vivo efficacy test. On the basis of previous work with other
compounds in this series, maximum tolerance of compound
9d was set to 400 mg/kg in CMC. At this dosage, neurotoxi-
city was observed in BALB/c mice. This suggested that in vivo
toxicity assessment should be addressed earlier for later
investigation in this series.
Summary
Compound 3, previously reported as a promising antitu-
berculosis lead, was found to be converted to metabolites 3a
and 3b both in vitro and in vivo. Upon the basis of data gleaned
frombothmetabolismandpharmacokinetic studiesinconcert
with rational drug design principles, metabolically more
stable derivatives endowed with potent anti-TB activity
were synthesized. Thus, after various modifications of the
oxymethylene linker, compound 9d with an alkene linker was
found to be both metabolically more stable and more
potent than compound 3, with a MABA MIC of 0.2 μM,
only 2-fold higher than that of rifampin. Moreover, com-
pound 9d acid turned out to be active against M. tubercu-
losis H37Rv in MABA at pH 6.8 (MIC = 12.5 μM). The
AUC of the acid derived from compound 9d is 70-fold
higher than the acid derived from 3 (3b) when dosed orally
at 400 mg/kg in CMC, and the Cmax of compound 9d acid
is almost 4-fold higher than its MIC. The fact that pyr-
azinamide (the first-line antituberculosis drug) and its
active form pyrazinoic acid (POA) are only active in vitro
at pH 5.6 but not at pH 6.8 suggests that a localized acidic

6972 Journal of Medicinal Chemistry, 2009, Vol. 52, No. 22
Mao et al.
Table 2. In Vitro Activity of Analogues with Different Linkers Against M. tuberculosis H37Rva
a NA: not available.
Table 3. Comparison of PK Profile of Compound 9d and 3
acid of compound
T1/2 (h)
Cmax (μM)
oral AUC (μM h)
9d
3
1.9
1.6
61.7
1.0
156.4
2.2
On the basis of these results, compound 9d was identified as
a new lead. A new series with an alkene linker was found to
contain several potent compounds and SAR that differed
from that observed with the oxymethylene linker. The new
SAR suggested that, for this alkene series, the position(s) of
substitution(s) plays a role in the antituberculosis activity. The
target protein for these compounds is under investigation, and
a systematic exploration of the alkene linker series will be
carried out with early in vivo toxicity assessment.
Figure 6. Mouse liver microsomal stability study of compound 9d.
pH environment, which facilitates the penetration of
POA, is likely produced in the lungs of TB patients as a
consequence of inflammation.9,10 As it is easier for acids to
penetrate the cell wall under acidic conditions, the anti-
tuberculosis activity of compound 9d acid may be superior
in vivo. We have an ongoing effort to develop a robust low
pH MIC assay to further test this hypothesis.
Experimental Section
1. Drug Metabolism and Pharmacokinetics. General Informa-
tion. Reference compounds were synthesized in-house. Pooled
mouse and human liver microsomes (protein concentration:
20 mg/mL) were purchased from XenoTech (Lenexa, KS).

Article
Journal of Medicinal Chemistry, 2009, Vol. 52, No. 22 6973
Table 4. In Vitro Activity of Mefloquine Analogues with Carbon Linkers Against M. tuberculosis H37Rv
Scheme 3. Synthesis of Analogues in Table 4
a. SeO2, 160 °C, 12 h, 33%. b. NaHDMS, 14, THF, -78 °C to rt, 4 h. c. (i) n-BulI, Et2O, -78 °C, 30 min; (ii) DMF, Et2O, -78 °C, 3 h. d. H2, Pd/C,
EtOAc, rt, 8 h. e. PBr3, DMF, rt, 2 h.

6974 Journal of Medicinal Chemistry, 2009, Vol. 52, No. 22
Mao et al.
Table 5. SRMs and Retention Times of Mefloquine-Based Analogues Detected by LC-MS-MS
compound
SRM ion pairs (m/z)
CE (eV)
tube lens
RT (min)
6a
Parent
Acid
449 to 262
421 to 262
282 to 262
463 to 262
435 to 262
282 to 262
491 to 262
463 to 262
282 to 262
509 to 262
481 to 262
282 to 262
405 to 293
377 to 293
282 to 262
448 to 286
420 to 286
281 to 261
25
25
21
53
27
21
35
35
21
40
40
21
32
37
21
47
47
33
87
87
24.49
13.36
-
3a
Parent
96
6b
6c
6d
9a
9b
113
92
20.96
12.90
-
Acid
3a
96
Parent
Acid
100
100
96
26.85
17.33
-
3a
Parent
94
94
28.63
19.65
-
Acid
3a
96
96
Parent
Acid
25.75
16.78
-
92
96
3a
Parent
70
70
19.10
11.23
9.18
Acid
2,8-Bis(trifluoromethyl)-quinolin-4-amine
96
NADPH was purchased from Sigma Aldrich (St. Louis, MO).
All solvents were HPLC grade and purchased from Fisher
Scientific (Hanover Park, IL).
was 320 °C. Under these conditions, compound 3 and its most
abundant metabolites eluted at 28.9, 24.9, and 16.1 min, respec-
tively. Initial Q1 scans were performed between m/z 150 and
490. Tandem mass spectra were acquired at a collision energy of
30 eV using nitrogen as the collision gas at a pressure of
1.5 mTorr.
Microsomal Incubation. Compound stock solutions were
prepared in acetonitrile. The final concentration of acetonitrile
in the incubation media was 0.2% (v/v). The 100 μL incubation
mixture contained a test compound (10 μM), 1 mM NADPH,
and 2.0 mg/mL mouse or human liver microsomes in 50 mM
potassium phosphate buffer (pH 7.4). Controls consisted of the
same reaction mixtures without either NADPH or microsomes.
The reaction mixture was prewarmed at 37 °C for 5 min before
adding NADPH (1 mM). Incubations were carried out in a
37 °C water bath for 30 min, and the reaction was stopped by
adding 300 μL acetonitrile. The samples were centrifuged at
4000g for 10 min before analysis using liquid chromatography-
tandem mass spectrometry (LC-MS-MS).
Microsomal Stability Assay. Compound stock solutions were
prepared in acetonitrile. The final concentration of acetonitrile
in the incubation media was 0.2% (v/v). The stability of
compound 9d in microsomes was determined in triplicate after
its incubation at 1 μM with mouse liver microsomes (0.5 mg/mL)
in 50 mM potassium phosphate buffer (pH 7.4) at 37 °C. The
total incubation volume was 0.8 mL. The reaction mixture was
prewarmed at 37 °C for 5 min before adding NADPH (1 mM).
Aliquots (100 μL) of the reaction mixture at 0, 5, 10, 15, 20, and
30 min were added to acetonitrile (300 μL) containing the
internal standard. 5-(2-Quinolin-4-yl-ethyl)isoxazole-3-car-
boxylic acid ethyl ester at 0.01 μM. The samples were centri-
fuged at 4000g for 10 min before liquid chromatography/
tandem mass spectrometry (LC-MS-MS) analysis of the parent
compound. For control experiments, NADPH was omitted
from these incubations.
Method B. A binary gradient consisted of a mixture of a 0.1%
formic acid in water (solvent A) and acetonitrile (solvent B) at a
flow rate of 0.2 mL/min. The LC gradient was programmed as
follows: 50% to 65% B over 25 min, then followed by an
isocratic hold at 95% B over the next 5 min. The column was
re-equilibrated for 7 min between injections. The column tem-
perature was 30 °C, and the autosampler was maintained at 4 °C.
Ionization was conducted in the positive ionization mode. The
electronspray voltage was 3 kV, and the capillary temperature is
320 °C. Initial Q1 scans were performed between m/z 150 and
550. The limit of detection (LOD) of 2,8-bis(trifluoromethyl)-
quinolin-4-ol and quinolin-4-ol was 120 pM and 160 pM,
respectively, and the retention times for these two compounds
were 8.92 and 4.35 min, respectively. The ion transitions with the
collision energy, tube lens, and retention times during selected
reaction monitoring (SRM) for compounds 6a-d and 9a,b are
listed in Table 5. Three SRMs were set up for each compound,
including the parent, the corresponding acid, and the corre-
sponding quinoline metabolite.
Dosing and in Vivo Sample Collection. Groups of twelve
female BALB/c mice (20-22 g) were administered a single oral
400 mg/kg dose of compound 3 or 9d. The dose was formulated
in 0.5% carboxymethyl cellulose (CMC) at the target concen-
tration of 20 mg/mL. Whole blood samples (250 μL) were
collected aseptically from the retro orbital sinus at 0, 15, 30,
45, 90, and 120 min postdose. Plasma was prepared from the
whole blood by centrifugation at 2000g for 30 min. Three
volumes of acetonitrile containing the internal standard (IS1:
4-amino-2,8-bis(trifluoromethyl)quinoline 0.01 μM for com-
pound 3; IS2: 5-(2-quinolin-4-yl-ethyl)isoxazole-3-carboxylic
acid ethyl ester 0.01 μM for compound 9d) were added to each
sample, and the samples were centrifuged at 4000g for 10 min
before LC-MS-MS analysis.
LC-MS-MS Assay for Metabolite Identification. Qualitative
assessments of compound metabolism were conducted using a
Finnigan TSQ quantum triple quadrapole mass spectrometer
(Thermo Fisher Scientific, San Jose, CA). Reversed-phase
HPLC separations during LC-MS-MS were carried out using
an XTerra MS C18 column (2.1 ꢀ 150 mm, 2.5 μm, Waters)
connected to a Waters 2695 HPLC system.
Method A. A binary gradient consisted of a mixture of a 0.1%
formic acid in water (solvent A) and acetonitrile (solvent B) at a
flow rate of 0.2 mL/min. The LC gradient was programmed as
follows: 15% to 70% B over 30 min, followed by an isocratic
hold at 95% B over the next 5 min. The column was re-
equilibrated for 7 min between injections. The column tempera-
ture was 30 °C, and the autosampler was maintained at 4 °C.
Ionization was conducted in the positive ionization mode. The
electronspray voltage was 3 kV, and the capillary temperature
LC-MS-MS Quantitation. Quantitation of compounds and
their metabolites was carried out using LC-MS-MS with posi-
tive ion electrospray with selected reaction monitoring on a
Finnigan TSQ quantum triple quadrupole mass spectrometer
coupled with a Waters 2695 HPLC system. Separations were
carried out using a Phenomenex (Torrance, CA) Luna C18
column 2.0 mm ꢀ 30 mm, 4 μm particle size. A binary gradient
consisted of a mixture of 0.1% formic acid in water (solvent A)
and 0.1% formic acid in acetonitrile (solvent B) at a flow rate of

Article
Journal of Medicinal Chemistry, 2009, Vol. 52, No. 22 6975
Table 6. SRMs, Retention Times, and Standard Curves of Compound 3 and Metabolites Detected by LC-MS-MS
compd.
SRM ion pairs (m/z)
CE (eV)
tub lens
RT (min)
standard curve
3
3a
435 to 154
282 to 262
407 to 262
281 to 261
295 to 154
267 to 154
297 to 156
27
21
27
33
25
27
25
96
96
92
96
90
92
90
3.2
2.4
1.8
1.5
2.8
2.0
2.6
Y = 0.7515x þ 0.0007 (R = 0.9953)
Y = 0.6067x þ 0.0576 (R = 0.9980)
Y = 0.1715x - 0.0029 (R = 0.9946)
3b
IS1
9d
Y = 0.6438x þ 0.005 (R = 0.9938)
Y = 0.0571x þ 0.1702 (R = 0.9988)
9d acid
IS2
0.4 mL/min. The HPLC gradient was programmed as follows:
40% to 65% B over 2.5 min, followed by an isocratic hold at
95% B over the next 1 min. The column was re-equilibrated for
1.5 min between injections. The column temperature was 30 °C,
and the autosampler was maintained at 4 °C. The collision
energy, tube lens, and retention time during selected reaction
monitoring for compound 3, 3a, 3b, IS1, compound 9d, the acid
of compound 9d, and IS2 are included in Table 6. The ratio of
peak area responses of analytes to IS was used to construct a
standard curve.
solution of ethyl chlorooximidoacetate (1.57 g, 10.34 mmol)
and 4-(1-methylprop-2-ynyloxy)-2,8-bis(trifluoromethyl)quin-
oline (1.14 g, 3.4 mmol) in Et2O (15 mL) was added triethyla-
mine (1.45 mL, 10.34 mmol) in Et2O (12 mL) via syringe pump
over a 6 h period. The mixture was diluted with Et2O (20 mL),
the organic layer separated, washed with H2O (10 mL), dried
(MgSO4), and concentrated in vacuo to afford an oil, which was
purified by column chromatography (EtOAc/hexane = 1: 4) to
give the product as a white solid with a yield of 50%. 1H NMR
(400 MHz, CDCl3): δ 8.50 (d, J = 7.6 Hz, 1H), 8.19 (d, J =
7.2 Hz, 1H), 7.72 (t, J = 7.6 Hz, 1H), 7.17 (s, 1H), 6.74 (s, 1H),
5.96 (q, J = 6.4 Hz, 1H), 4.45 (q, J = 7.2 Hz, 2H), 1.98 (d, J =
6.4 Hz, 3H), 1.43 (t, J = 7.2 Hz, 3H). 13C NMR (100 MHz,
CDCl3): δ 171.8, 160.9, 159.3, 156.6, 149.2, 144.3, 129.7, 128.9,
126.6, 126.1, 124.8, 122.2, 119.4, 102.7, 98.3, 69.2, 62.5, 19.8,
14.1. Electrospray MS: m/z 449.0 ([M þ H]þ, 100). Electrospray
HRMS calcd for [C19H14F6N2O4 þ H]þ: 449.0936; found:
449.0914.
5-[1-(2,8-Bis(trifluoromethyl)quinolin-4-yloxy)-1-methylethyl]-
isoxazole-3-carboxylic Acid Ethyl Ester (6b). The synthesis of 6b
is similar to that of 6a. Yield: 54%. 1H NMR (400 MHz,
CDCl3): δ 8.49 (d, J = 8.0 Hz, 1H), 8.16 (d, J = 7.2 Hz, 1H),
7.90 (s, 1H), 7.76 (t, J = 7.2 Hz, 1H), 6.78 (s, 1H), 4.46 (q, J =
7.2 Hz, 2H), 2.06 (s, 6H), 1.43 (t, J = 7.2 Hz, 3H). 13C NMR
(100 MHz, CDCl3): δ 174.0, 159.1, 158.9, 156.3, 148.6, 144.6,
129.1, 129.0, 128.1, 126.1, 124.5, 123.3, 121.7, 101.4, 62.2, 26.6,
13.7. Electrospray MS: m/z 461 ([M - H]-, 100). Electrospray
HRMS calcd for [C20H16F6N2O4 - H]-:461.0936; found:
461.0935.
5-[1-(2,8-Bis(trifluoromethyl)quinolin-4-yloxy)pentyl]isoxazole-
3-carboxylic Acid Ethyl Ester (6c). The synthesis of 6c is similar
to that of 6a. Yield: 52%. 1H NMR (400 MHz, CDCl3): δ 8.53
(d, J = 7.6 Hz, 1H), 8.20 (d, J = 7.2 Hz, 1H), 7.73 (t, J = 7.6 Hz,
1H), 7.10 (s, 1H), 6.98 (s, 1H), 5.78 (t, J = 5.6 Hz, 1H), 4.45 (q,
J = 7.2 Hz, 2H), 2.18 (m, 2H), 1.67 (m, 2H), 1. 49 (m, 2H), 1.41
(t, J = 7.2 Hz, 3H), 1.01 (t, J = 7.2 Hz, 3H). 13C NMR
(100 MHz, CDCl3): δ 170.9, 160.9, 158.9, 156.2, 148.6, 144.5,
129.2, 128.6, 128.1, 126.3, 125.6, 125.6, 121.9, 102.7, 97.9, 72.8,
62.1, 33.6, 29.3, 26.5, 21.8, 13.7. Electrospray MS: m/z 491.0
([M þ H]þ, 100). Electrospray HRMS calcd for [C22H20F6N2-
O4 þ H]þ: 491.1405; found: 491.1400.
PK Parameter Calculations. The half-life T1/2 was calculated
by the equation T1/2 = 0.693/k, and the areas of all trapezoids
were calculated and summed to give AUC from t = 0 to t = µ.
2. Chemistry. General Information. All starting materials
were purchased from Sigma-Aldrich. 1H NMR and 13C NMR
spectra were recorded on Bruker spectrometer at 400 or
300 MHz and 100 or 75 MHz, respectively, with TMS as an
internal standard. Positive or negative ion mass spectra were
measured using electrospray ionization (ESI) or atmospheric
pressure chemical ionization (APCI) at an ionization potential
of 70 eV; high resolution mass spectrometry (HRMS) was
carried out using Micromass (Manchester, UK) hybrid quadru-
pole time-of-flight mass spectrometer (Q-TOF-2TM). Thin-layer
chromatography (TLC) was performed with Merck 250 mm
60F254 silica gel plates. Column chromatography was performed
using Merck silica gel (40-60 mesh). The purity of the target
compounds was determined to be >98% by analytical HPLC.
4-(1-Methylprop-2-ynyloxy)-2,8-bis(trifluoromethyl)quinoline
(5a, general procedure for 5a-d). To a solution of the 2,8-bis-
(trifluoromethyl)quinolin-4-ol (1.000 g, 3.6 mmol) in dry ace-
tone (25 mL) was added excess anhydrous K2CO3 (5 g) and the
mixture was refluxed for 0.5 h. To the mixture, 3-bromobut-1-
yne (0.47 g, 3.6 mmol) was added dropwise. The resulting
mixture was refluxed for an additional 14 h and then cooled
and filtered, and the filtrate was evaporated and then dried in
1
vacuo to obtain a white solid 5a (1.1 g, 92%). H NMR (400
MHz, CDCl3): δ 8.46 (d, J = 7.6 Hz, 1H), 8.14 (d, J = 7.2 Hz,
1H), 7.64 (t, J = 7.6 Hz, 1H), 7.09 (s, 1H), 5.21 (q, J = 6.4 Hz,
1H), 2.64 (s, 1H), 1.94 (d, J = 6.4 Hz, 3H). 13C NMR (100 MHz,
CDCl3): δ 160.9, 149.2, 144.3, 128.9, 128.8, 125.9, 125.7, 124.6,
122.1, 119.4, 98.8, 79.8, 75.8, 64.6, 21.4.
4-(1,1-Dimethylprop-2-ynyloxy)-2,8-bis(trifluoromethyl)quinoline
(5b). The synthesis is similar to compound 5a. 1H NMR
(400 MHz, CDCl3): δ 8.51 (d, J = 8.0 Hz, 1H), 8.16 (d, J =
7.2 Hz, 1H), 7.90 (s, 1H), 7.76 (t, J = 7.2 Hz, 1H), 2.82 (s, 1H),
1.93 (s, 1H).
4-(1-Ethynylpentyloxy)-2,8-bis(trifluoromethyl)quinoline (5c).
The synthesis of 5c is similar to compound 5a. 1H NMR
(400 MHz, CDCl3): δ 8.47 (d, J = 7.6 Hz, 1H), 8.15 (d, J =
7.2 Hz, 1H), 7.68 (t, J = 7.6 Hz, 1H), 7.28 (s, 1H), 5.07 (t, J =
5.6 Hz, 1H), 2.64 (s, 1H), 2.15 (m, 2H), 1.69 (m, 2H), 1.45 (m,
2H), 1.01 (t, J = 7.2 Hz, 3H).
5-[(2,8-Bis(trifluoromethyl)quinolin-4-yloxy)phenylmethyl]iso-
xazole-3-carboxylic Acid Ethyl Ester (6d). The synthesis of 6d is
similar to that of 6a. Yield: 50%. 1H NMR (400 MHz, CDCl3): δ
8.45 (d, J = 7.6 Hz, 1H), 8.15 (d, J = 7.2 Hz, 1H), 7.66 (t, J =
7.6 Hz, 1H), 7.39 (m, 5H), 7.09(s, 1H), 6.54 (s, 1H), 5.56 (s, 1H),
4.40 (q, J = 7.2 Hz, 2H), 1.41 (t, J = 7.2 Hz, 3H). 13C NMR
(100 MHz, CDCl3): δ 167. 6, 159.1, 158.9, 156.3, 148.6, 144.6,
129.1, 129.0, 128.6, 128.5, 128.1, 126.5, 125.9, 125.6, 124.5,
123.3, 121.7, 102.4, 99.6, 73.0, 61.9, 13.8. Electrospray MS:
m/z 509.0 ([M - H]-, 100). Electrospray HRMS calcd for [C24-
H16F6N2O4 - H]-: 509.0936; found: 509.0955.
4-Ethynyl-2,8-bis(trifluoromethyl)quinoline (8). A mixture of
4-bromo-2,8-bis(trifluoromethyl)quinoline (0.52 g, 1.5 mmol),
(trimethylsilyl) acetylene (0.45 g, 4.5 mmol), CuI (17.1 mg,
0.09 mmol), and Pd(PPh3)4 (102.9 mg, 0.09 mmol) in 15 mL of
toluene-i-Pr2NH (7:3) mixture was stirred at 80 °C in a sealed
flask for 36 h. After allowing to cool to room temperature, the
reaction mixture was diluted with chloroform and passed
4-(1-Phenylprop-2-ynyloxy)-2,8-bis(trifluoromethyl)quinoline
(5d). The synthesis of 5d is similar to compound 5a. H NMR
1
(400 MHz, CDCl3): δ 8.46 (d, J = 7.6 Hz, 1H), 8.14 (d, J =
7.2 Hz, 1H), 7.64 (t, J = 7.6 Hz, 1H), 7.36 (m, 5H), 7.09 (s, 1H),
5.21 (s, 1H), 2.65 (s, 1H).
5-[1-(2,8-Bis(trifluoromethyl)quinolin-4-yloxy)ethyl]isoxazole-
3-carboxylic acid Ethyl Ester (6a). To a vigorously stirred

6976 Journal of Medicinal Chemistry, 2009, Vol. 52, No. 22
Mao et al.
through a short silica gel column eluting with chloroform. After
concentrating in vacuo, the oily residue was purified by column
chromatography on silica gel (EtOAc/hexanes 1:4) to afford a
crude product as a dark-red crystalline material. It was dissolved
in 40 mL of THF and added dropwise into a stirred solution of
KOH (0.25 g, 4.5 mmol) in 100 mL of methanol. The resulting
mixture was stirred at room temperature for 40 min, poured into
water, and extracted with ether. The organic fraction was
washed successively with water, conc. NH4Cl solution, water,
and brine, and dried over Na2SO4. Concentration in vacuo
afforded a solid material, which was purified by column chro-
matography on silica gel (EtOAc/hexanes 1:10) to afford pro-
(400 MHz, CDCl3): δ 10.71 (br s, 1H), 8.82 (d, 1H, J =
8.0 Hz), 8.68 (d, 1H, J = 4.0 Hz), 8.39 (d, 1H, J = 8.0 Hz),
7.99 (t, 1H, J = 8.0 Hz), 4.48 (q, 2H, J = 6.7 Hz), 1.42 (t, 3H,
J = 8.0 Hz). 13C NMR (100 MHz, CDCl3): δ 165.0, 159.6, 158.3,
155.5, 149.1 (q, J = 34.7 Hz), 145.4, 144.4, 130.9 (q, J = 5.5 Hz),
129.0 (q, J = 29.9 Hz), 128.4, 128.3, 124.7 (q, J = 253 Hz),
126.1, 122.5 (q, J = 273 Hz), 111.4, 109.1, 63.1, 14.3. Electro-
spray HRMS: calcd for 446.0581 [C18H11F6N3O4 - H]-, found
446.0590.
Ethyl 5-Bromomethyl-3-isoxazolecarboxylate (13). To a solu-
tion of ethyl 5-(hydroxymethyl)-3-isoxazolecarboxylate 10
(1.54 g, 90%, 8.12 mmol) in dichloromethane (20 mL) at
-15 °C was added triphenylphosphine (2.44 g, 9.35 mmol)
and tetrabromomethane (3.03 g, 9.15 mmol). The reaction was
stirred at the same temperature for 2 h. The resulting mixture
was concentrated under reduced pressure, and the residue was
purified by column chromatography on silica gel (EtOAc/
1
duct with 70% yield. H NMR (400 MHz, CDCl3): δ, 8.23 (d,
J = 7.2 Hz, 1H), 8.58 (d, J = 8.0 Hz, 1H), 7.82 (t, J = 7.6 Hz,
1H), 7.98 (s, 1H), 3.85 (s, 1H). 13C NMR (100 MHz, CDCl3): δ
144.2, 132.7, 130.3, 129.0, 128.2, 127.1, 121.3, 108.2, 98.3, 88.0,
85.9.
1
5-(2,8-Bis(trifluoromethyl)quinolin-4-yl)isoxazole-3-carboxylic
Acid Ethyl Ester (9a). The synthesis of 9a is similar to that of 6a.
Yield: 50%. 1H NMR (400 MHz, CDCl3): δ 8.62 (d, J = 8.0 Hz,
1H), 8.31 (d, J = 7.2 Hz, 1H), 8.14 (s, 1H), 7.90 (t, J = 7.6 Hz,
1H), 7.11 (s, 1H), 4.46 (q, J = 7.2 Hz, 2H), 1.42 (t, J = 7.2 Hz,
3H). 13C NMR (100 MHz, CDCl3): δ 171.7, 163.5, 162.8, 157.0,
148.7, 144.6, 129.9, 128.6, 127.8, 125.8, 124.3, 123.4, 118.0, 99.8,
97.6, 62.4, 14.2. Electrospray MS: m/z 405.2 ([M þ H]þ, 100).
Electrospray HRMS calcd for [C17H10F6N2O3 þ H]þ: 405.0674;
found: 405.0689.
3,5-Isoxazoledicarboxylic Acid, 3-Ethyl Ester (11). Potassium
dichromate (0.95 g, 3.2 mmol) was added dropwise to concen-
trated sulfuric acid (5 mL), and the resulting mixture was diluted
with water (5 mL). The diluted solution was slowly added to a
solution of ethyl 5-(hydroxymethyl)-3-isoxazolecarboxylate
(361 mg, 90%, 1.9 mmol) in acetone (40 mL). The reaction
was stirred at 55 °C for 2 h, and the resulting mixture was
extracted with ethyl acetate (40 mL ꢀ 2). Evaporation of the
solvents provided the crude product (480 mg), which was
dissolved in 0.5 M NaOH solution (20 mL) and washed with
ethyl acetate (20 mL ꢀ 2). The aqueous phase was acidified with
concentrated HCl to pH 2 and was extracted with ethyl acetate
(20 mL ꢀ 2). The combined organic phase was dried over
Na2SO4. Evaporation of the solvents provided 11 as an oil
(310 mg, 88%). 1H NMR (400 MHz, CDCl3): δ 9.71 (br s,
1H), 7.42 (s, 1H), 4.50 (q, 2H, J = 6.7 Hz), 1.45 (t, 3H, J =
6.0 Hz). 13C NMR (100 MHz, CDCl3): δ 161.4, 159.4, 159.1,
157.3, 111.1, 63.1, 14.3.
5-Chlorocarbonyl-isoxazole-3-carboxylic Acid Ethyl Ester
(12). To a solution of 11 (37 mg, 0.2 mmol) in dichloromethane
(0.5 mL) was added a solution of oxalyl chloride in dichlor-
omethane (2.0 M, 0.5 mL, 1.0 mmol). A solution of DMF in
dichloromethane was added (1 drop of DMF was dissolved in
1.5 mL of dichloromethane, and 0.5 mL of the solution was
used). The reaction was stirred at room temperature for 3 h.
Evaporation of the solvents provided the crude 12 as an oil
(40 mg, 98%). 1H NMR (400 MHz, CDCl3): δ = 7.54 (s, 1H),
4.50 (q, 2H, J = 6.7 Hz), 1.45 (t, 3H, J = 6.0 Hz). 13C NMR
(100 MHz, CDCl3): δ 162.6, 158.4, 157.5, 154.8, 113.4, 63.2,
14.2.
Ethyl 5-[[[2,8-bis(trifluoromethyl)-4-quinolinyl]amino]carbon-
yl]-3-isoxazolecarboxylate (9b). To a solution of 2,8-bis-(triflu-
oromethyl)-4-quinolinamine (56 mg, 0.2 mmol) and diisopro-
pylethylamine (78 μL, 0.4 mmol) in dichloromethane (1 mL) at
0 °C was slowly added a solution of crude 12 (40 mg, ∼0.2 mmol)
in dichloromethane (2 mL). The reaction was allowed to warm
to room temperature and was stirred overnight. The resulting
mixture was diluted with dichloromethane (20 mL), washed
with 10% aqueous KHSO4 (10 mL), saturated NaHCO3
(10 mL), and brine (10 mL). The organic phase was dried
over Na2SO4 and concentrated, and the residue was puri-
fied by column chromatography on silica gel (EtOAc/hexanes
1:6) to afford 9b as a pale solid (50 mg, 56%). 1H NMR
hexanes 1:12) to give 13 as a colorless oil (1.56 g, 82%). H
NMR (400 MHz, CDCl3): δ 6.74 (s, 1H), 4.51 (s, 2H), 4.45 (q,
2H, J = 7.2 Hz), 1.42 (t, 3H, J = 7.0 Hz). 13C NMR (100 MHz,
CDCl3): δ 169.5, 159.7, 156.9, 104.6, 62.6, 18.0, 14.3.
[[3-(Ethoxycarbonyl)-5-isoxazolyl]methyl]triphenylphosphonium
Bromide (14). To a solution of 13 (1.56 g, 6.7 mmol) in
acetonitrile (20 mL) was added triphenylphosphine (1.91 g,
7.25 mmol). The solution was refluxed for 2 h. The resulting
mixture was allowed to cool to room temperature and was
further cooled to 0 °C using an ice bath. The formed white solid
was filtered and washed with ice-cool acetonitrile (1 mL ꢀ 2) to
1
give 14 (2.76 g, 83%). H NMR (400 MHz, CDCl3): δ 7.8-
8.0 (m, 9H), 7.6-7.8 (m, 6H), 7.12 (s, 1H), 6.20 (d, 2H, J =
12.0 Hz), 4.45 (d, 2H, J = 4.0 Hz), 1.42 (s, 3H). 13C NMR (100
MHz, CDCl3): δ 162.8 (J = 9.8 Hz), 159.2, 157.1, 135.7 (J =
2.7 Hz), 134.3 (J = 10.4 Hz), 130.6 (J = 13.0 Hz), 117.6 (J =
86.7 Hz), 108.4 (J = 7.1 Hz), 62.5, 24.4 (J = 51.8 Hz), 14.2.
Z-Ethyl
5-[2-(4-quinolinyl)ethenyl]-3-isoxazolecarboxylate
(9c) and E-Ethyl 5-[2-(4-quinolinyl)ethenyl]-3-isoxazolecarboxy-
late (9d). To a solution of 14 (39 mg, 0.079 mmol) in THF (1 mL)
at -78 °C was added a solution of sodium hexamethyldisilazide
in THF (1 M, 83 μL, 0.083 mmol). The mixture was stirred for
1 h at -30 °C and was cooled to -78 °C. A solution of 4-
quinolinecarboxaldehyde (12.7 mg, 0.079 mmol) in THF (1 mL)
was added dropwise. The resulting mixture was allowed to warm
to room temperature and was stirred for 3 h. Saturated ammo-
nium chloride solution was added at 0 °C to quench the reaction.
Solvents were removed under reduced pressure, and the residue
was repartitioned in dichloromethane (20 mL) and water
(20 mL). The organic phase was separated, dried over Na2SO4,
and concentrated. The residue was separated by column chro-
matography on silica gel (EtOAc/hexanes 1:1). Evaporation of
the faster eluting fractions gave 9c as a yellow solid (6 mg, 26%).
1H NMR (400 MHz, CDCl3): δ 8.94 (d, 1H, J = 4.0 Hz), 8.19 (d,
1H, J = 8.0 Hz), 7.90 (d, 1H, J = 8.0 Hz), 7.78 (t, 1H, J =
8.0 Hz), 7.57 (t, 1H, J = 8.0 Hz), 7.36 (d, 1H, J = 4.0 Hz), 7.32
(d, 1H, J = 12.0 Hz), 6.96 (d, 1H, J = 12.0 Hz), 6.11 (s, 1H), 4.34
(q, 2H, J = 8.0 Hz), 1.33 (t, 3H, J = 8.0 Hz). 13C NMR
(100 MHz, CDCl3): δ 168.3, 159.8, 156.6, 150.2, 148.3, 142.6,
132.5, 130.4, 130.3, 127.6, 126.0, 124.5, 120.5, 118.9, 104.2, 62.4,
14.3. Electrospray HRMS: calcd 295.1077 for [C17H14N2O3 þ
H]þ, found 295.1088. Evaporation of the slower eluting frac-
1
tions gave 9d as a light yellow solid (12 mg, 52%). H NMR
(400 MHz, CDCl3): δ 8.96 (d, 1H, J = 4.4 Hz), 8.1-8.2 (m, 3H),
7.78 (t, 1H, J = 9.2 Hz), 7.65 (t, 1H, J = 7.4 Hz), 7.59 (d, 1H,
J = 4.8 Hz), 7.21 (d, 1H, J = 16.0 Hz), 6.83 (s, 1H), 4.49 (q, 2H,
J = 8.0 Hz), 1.46 (t, 3H, J = 6.0 Hz). 13C NMR (100 MHz,
CDCl3): δ 168.2, 160.0, 157.2, 150.3, 149.0, 140.7, 130.6, 130.5,
130.0, 127.5, 126.2, 123.4, 118.3, 117.6, 103.7, 62.6, 14.4. Elec-
trospray HRMS: calcd 295.1077 for [C17H14N2O3 þ H]þ, found
295.1088.
E-Ethyl 5-[2-(4-Quinolinyl)ethenyl]-3-isoxazolecarboxylic
Acid (9d Acid). To a solution of compound 9d (0.58 g, 2 mmol)

Article
Journal of Medicinal Chemistry, 2009, Vol. 52, No. 22 6977
in THF/H2O/MeOH (3: 1: 1, 15 mL) was added lithium hydro-
xide monohydrate (0.6 g, 13.6 mmol). The solution was allowed
to stir for 4 h at ambient temperature and the solvent was
removed in vacuo. The residue was dissolved in water and
extracted with EtOAc (3 ꢀ 15 mL). The aqueous layer was
acidified to pH 2 with 6 M HCl and extrated with EtOAc (3 ꢀ
15 mL). The combined organic extracts were dried (MgSO4) and
evaporated in vacuo to give a colorless product (0.5 g, 92%). 1H
NMR (400 MHz, CDCl3): δ 8.96 (d, 1H, J = 4.4 Hz), 8.1-8.2
(m, 3H), 7.78 (t, 1H, J = 9.2 Hz), 7.63 (t, 1H, J = 7.4 Hz), 7.61
separated by filtration, washed with water, and dried under
reduced pressure to give 18 as a white solid (0.83 g, 70%). H
NMR (400 MHz, CDCl3): δ 8.53 (d, 1H, J = 4.8 Hz), 8.00 (d,
1H, J = 8.4 Hz), 7.66 (d, 1H, J = 4.8 Hz), 7.40 (m, 2H), 3.98 (s,
3H). 13C NMR (100 MHz, CDCl3): δ 159.2, 147.4, 145.3, 132.7,
131.7, 129.3, 125.5, 123.5, 104.6, 55.9.
4-Bromo-2,6-bis(trifluoromethyl)quinoline (19). The synthesis
of 19 is similar to that of 18. Yield: 77%. 1H NMR (400 MHz,
CDCl3): δ 8.60 (s, 1H), 8.38 (d, 1H J = 8.8), 8.14 (s, 1H), 8.06
(dd, 1H, J1 = 1.6, J2 = 8.8). 13C NMR (100 MHz, CDCl3): δ
149.2, 148.2, 136.6, 131.2, 127.5, 127.0, 124.5, 121.7.
1
(d, 1H, J = 4.8 Hz), 7.21 (d, 1H, J = 16.0 Hz), 6.85 (s, 1H). 13
C
NMR (100 MHz, CDCl3): δ 168.7, 160.0, 157.2, 150.3, 149.5,
140.7, 130.6, 130.5, 130.0, 128.5, 126.2, 123.4, 118.3, 117.6,
103.7. Electrospray HRMS: calcd 266.0727 for [C15H10N2-
O3 þ H]þ, found 266.0720.
6-Methoxy-4-quinolinecarboxaldehyde (20). The synthesis of
20 is similar to that of 17. 1H NMR (400 MHz, CDCl3): δ 10.39
(s, 1H), 9.02 (d, 1H, J = 4.0 Hz), 8.43 (s, 1H), 8.07 (d, 1H, J =
8.0 Hz), 7.72 (d, 1H, J = 4.0 Hz), 7.43 (d, 1H, J = 8.0 Hz), 3.98
(s, 3H). 13C NMR (100 MHz, CDCl3): δ 193.7, 160.6, 147.6,
145.9, 135.3, 131.4, 127.4, 125.3, 123.3, 102.5, 55.9.
2,6-Bis(trifluoromethyl)quinoline-4-carbaldehyde (21). The
synthesis of 21 is similar to that of 17. The product was used
for the next step without purification.
Ethyl 5-[2-(4-Quinolinyl)ethyl]-3-isoxazolecarboxylate (9e).
To a solution of 9d (7 mg) in ethyl acetate (3 mL) was added
5% palladium on activated carbon (2 mg). The mixture was
stirred at room temperature under a H2 atmosphere for 8 h. The
mixture was extracted by ethyl acetate, and the palladium on
carbon was removed by filtration through a silica gel pad. The
filtrate was concentrated, and the residue was purified by
column chromatography on silica gel (EtOAc/hexanes 1:1) to
give 9e as a light yellow solid (7 mg, 99%). 1H NMR (400 MHz,
CDCl3): δ 8.83 (d, 1H, J = 4.4 Hz), 8.16 (d, 1H, J = 4.4 Hz),
8.03 (d, 1H, J = 4.4 Hz), 7.75 (t, 1H, J = 7.2 Hz), 7.62 (t, 1H,
J = 7.2 Hz), 7.21 (d, 1H, J = 4.4 Hz), 6.41 (s, 1H), 4.43 (q, 2H,
J = 9.3 Hz), 3.53 (t, 2H, J = 8.0 Hz), 3.30 (t, 2H, J = 8.0 Hz),
1.41 (t, 3H, J = 8.0 Hz). 13C NMR (100 MHz, CDCl3): δ 173.6,
160.2, 156.7, 150.4, 148.6, 145.4, 130.7, 129.7, 127.2, 123.0,
120.9, 102.5, 62.4, 29.9, 27.1, 14.4. Electrospray HRMS: calcd
297.1077 for [C17H14N2O3 þ H]þ, found 297.1074.
Z-Ethyl
5-[2-(6-quinolinyl)ethenyl]-3-isoxazolecarboxylate
(9f) and E-Ethyl 5-[2-(6-quinolinyl)ethenyl]-3-isoxazolecarboxy-
late (9g). The synthesis of 9f,g is similar to that of 9c,d. 9f. 1H
NMR (400 MHz, CDCl3): δ 8.95 (d, 1H, J = 2.4 Hz), 8.12 (dd,
2H, J1 = 13.6 Hz, J2 = 8.4 Hz), 7.86 (s, 1H), 7.58 (d, 1H, J =
16.8 Hz), 7.43 (dd, 1H, J1 = 8.4 Hz, J2 = 4.0 Hz), 7.10 (d, 1H,
J = 12.8 Hz), 6.64 (d, 1H, J = 12.4 Hz), 6.46 (s, 1H), 4.39 (q, 2H,
J = 6.7 Hz), 1.37 (t, 3H, J = 8.0 Hz). 13C NMR (100 MHz,
CDCl3): δ 169.2, 160.0, 156.6, 151.4, 148.3, 136.46, 136.41,
133.9, 130.1, 129.8, 128.3, 128.1, 121.9, 115.6, 103.8, 62.4,
14.3. Electrospray HRMS: calcd 295.1077 for [C17H14N2O3 þ
H]þ, found 295.1074. 9g. 1H NMR (400 MHz, CDCl3): δ 8.93
(d, 1H, J = 2.4 Hz), 8.15 (dd, 2H, J1 = 22.0 Hz, J2 = 8.0 Hz),
7.9-8.0 (m, 2H), 7.58 (d, 1H, J = 16.4 Hz), 7.44 (dd, 1H, J1 =
8.4 Hz, J2 = 4.4 Hz), 7.12 (d, 1H, J = 16.4 Hz), 6.74 (s, 1H), 4.48
(q, 2H, J = 7.1 Hz), 1.45 (t, 3H, J = 7.2 Hz). 13C NMR
(100 MHz, CDCl3): δ 170.1, 160.2, 157.0, 151.3, 149.0, 136.5,
135.4, 133.5, 130.5, 128.6, 128.3, 126.9, 122.1, 113.8, 102.3, 62.5,
14.4. Electrospray HRMS: calcd 295.1077 for [C17H14N2O3 þ
H]þ, found 295.1075.
E-Ethyl 5-[2-[2,8-bis(trifluoromethyl)-4-quinolinyl]ethenyl]-3-
isoxazolecarboxylate (9h). The synthesis of 9h is similar to that
of 9c,d. Only the E-isomer was obtained. 1H NMR (400 MHz,
CDCl3): δ 8.46 (d, 1H, J = 4.4 Hz), 8.24 (d, 1H, J = 6.8 Hz),
8.16 (d, 1H, J = 16.4 Hz), 7.99 (s, 1H), 7.83 (t, 1H, J = 8.0 Hz),
7.31 (d, 1H, J = 16.0 Hz), 6.88 (s, 1H), 4.49 (q, 2H, J = 6.7 Hz),
1.46 (t, 3H, J = 8.0 Hz). 13C NMR (100 MHz, CDCl3): δ 168.3,
159.7, 157.3, 148.5 (q, J = 35.1 Hz), 144.5, 143.6, 129.7 (q, J =
5.3 Hz), 129.6 (q, J = 30.3 Hz), 128.7, 128.0, 127.7, 127.1, 123.6
(q, J = 272 Hz), 121.3 (q, J = 234.2 Hz), 120.3, 114.2, 104.9,
62.7, 14.3. Electrospray HRMS: calcd 431.0825 for [C19H12F6-
N2O3 þ H]þ, found 431.0812.
6-Quinolinecarboxaldehyde (16). 6-Methylquinoline (2.0 g,
13.8 mmol) was heated to 160 °C and selenium dioxide (1 g,
9.2 mmol) was added. The mixture was stirred for 12 h, cooled to
room temperature, and diluted with ethyl acetate (10 mL). The
solution was decanted, and the residue was extracted with ethyl
acetate (10 mL ꢀ 2). The combined organic phase was concen-
trated, and the residue was purified by column chromatography
on silica gel (EtOAc/hexanes 1:2) to give 16 as a light gray solid
(715 mg, 33%). 1H NMR (400 MHz, CDCl3): δ 10.18 (s, 1H),
9.04 (dd, 1H, J1 = 4.0 Hz, J2 = 4.0 Hz), 8.3 (m, 2H,), 8.2 (m,
2H), 7.50 (dd, 1H, J1 = 8.0 Hz, J2 = 4.0 Hz). 13C NMR
(100 MHz, CDCl3): δ 191.4, 153.1, 150.8, 137.4, 134.3, 133.7,
130.8, 127.7, 126.6, 122.2.
2,8-Bis(trifluoromethyl)-4-quinolinecarboxaldehyde (17). To a
solution of n-BuLi (2 M in hexanes, 0.49 mL, 0.49 mmol) in
anhydrous diethyl ether (10 mL) at -78 °C was slowly added a
solution of 4-bromo-2,8-bis(trifluoromethyl)quinoline (334 mg,
0.97 mmol) in diethyl ether (4 mL). The resulting mixture was
stirred at -78 °C for 30 min. A solution of anhydrous DMF
(87 μL, 1.2 mmol) was added dropwise. The reaction was stirred
at -78 °C for 3 h and quenched by water. Diethyl ether (50 mL)
was added, and the organic phase was washed with saturated
NH4Cl (25 mL), saturated Na2S2O3 (25 mL) and brine (25 mL),
dried over Na2SO4, and concentrated. The residue was purified
by column chromatography on silica gel (EtOAc/hexanes 1:8)
to give 17 as a light yellow solid (105 mg, 37%). 1H
NMR (400 MHz, CDCl3): δ 10.56 (br s 1H), 9.31 (d, 1H, J =
8.0 Hz), 8.29 (d, 1H, J = 8.0 Hz), 8.23 (s, 1H), 7.94 (t, 1H, J =
8.0 Hz). 13C NMR (100 MHz, CDCl3): δ 191.4, 149.2 (q, J =
36.3 Hz), 146.0, 138.9, 130.4, 130.0 (q, J = 5.3 Hz), 129.4 (q,
J = 30.5 Hz), 129.0, 125.2, 123.5 (q, J = 272 Hz), 122.4, 121.0
(q, J = 274 Hz).
Ethyl 5-[2-[2,8-Bis(trifluoromethyl)-4-quinolinyl]ethyl]-3-iso-
xazolecarboxylate (9i). The synthesis of 9i is similar to that of
9e. 1H NMR (400 MHz, CDCl3): δ 8.30 (d, 1H, J = 8.4 Hz), 8.19
(d, 1H, J = 7.2 Hz), 7.79 (t, 1H, J = 7.8 Hz), 7.66 (s, 1H), 6.46 (s,
1H), 4.44 (q, 2H, J = 7.2 Hz), 3.63 (t, 2H, J = 8.0 Hz), 3.33 (t,
2H, J = 7.8 Hz), 1.41 (t, 3H, J = 7.2 Hz). 13C NMR (100 MHz,
CDCl3): δ 172.6, 159.9, 156.8, 148.5, 148.4 (q, J = 35.2 Hz),
144.0, 129.8 (q, J = 31.0 Hz), 129.3 (q, J = 5.3 Hz), 128.2, 127.8,
127.3, 123.7 (q, J = 272 Hz), 121.2 (q, J = 273.7 Hz), 117.4,
102.7, 62.5, 30.3, 27.1, 14.3. Electrospray HRMS: calcd
433.0981 for [C19H14F6N2O3 þ H]þ, found 431.0975.
Z-Ethyl 5-[2-(6-Methoxy-4-quinolinyl)ethenyl]-3-isoxazole-
carboxylate (9j) and E-Ethyl 5-[2-(6-methoxy-4-quinolinyl)ethen-
yl]-3-isoxazolecarboxylate (9k). The synthesis of 9j,k is similar to
that of 9c,d. 9j. 1H NMR (400 MHz, CDCl3): δ 8.78 (d, 1H, J =
4.4 Hz), 8.07 (d, 1H, J = 9.2 Hz), 7.41 (dd, 2H, J1 = 8.4 Hz,
J2 = 2.4 Hz), 7.32 (s, 1H), 7.26 (d, 1H, J = 13.2 Hz), 7.07 (d,
1H, J = 2.4 Hz), 6.94 (d, 1H, J = 12.4 Hz), 6.13 (s, 1H), 4.34
4-Bromo-6-methoxyquinoline (18). To a solution of 6-meth-
oxy-4-quinolinol (876 mg, 5 mmol) in anhydrous DMF (8 mL)
under a N2 atmosphere was added PBr3 (484 μL, 5.15 mmol)
dropwise. The reaction was stirred at room temperature for 2 h
and quenched with water. The pH of the resulting mixture was
adjusted to 10 by adding NaHCO3. The formed precipitate was

6978 Journal of Medicinal Chemistry, 2009, Vol. 52, No. 22
(q, 2H, J = 6.7 Hz), 3.88 (s, 3H), 1.33 (t, 3H, J = 8.0 Hz). 13
Mao et al.
C
Supporting Information Available: HPLC purity determina-
tions of the target compounds. This material is available free of
charge via the Internet at http://pubs.acs.org
NMR (100 MHz, CDCl3): δ 168.4, 159.7, 158.4, 156.5, 147.6,
144.8, 140.6, 132.8, 131.9, 126.9, 122.8, 120.6, 118.6, 103.9,
102.0, 62.4, 55.7, 14.2. Electrospray HRMS: calcd 325.1183
for [C18H16N2O4 þ H]þ, found 325.1194. 9k. 1H NMR
(400 MHz, CDCl3): δ = 8.78 (d, 1H, J = 4.4 Hz), 8.05 (s,
1H), 8.02 (d, 1H, J = 12.8 Hz), 7.51 (d, 1H, J = 4.4 Hz), 7.41
(dd, 1H, J1 = 9.2 Hz, J2 = 2.0 Hz), 7.32 (s, 1H), 7.17 (d, 1H, J =
16.4 Hz), 6.80 (s, 1H), 4.48 (q, 2H, J = 6.7 Hz), 3.99 (s, 3H), 1.45 (t,
3H, J = 8.0 Hz). 13C NMR (100 MHz, CDCl3): δ 169.2, 159.9,
158.5, 157.1, 147.6, 145.1, 139.2, 131.9, 130.7, 127.2, 122.5, 117.9,
117.7, 103.5, 101.2, 62.5, 55.9, 14.3. Electrospray HRMS: calcd
325.1183 for [C18H16N2O4 þ H]þ, found 325.1192.
References
(1) WHO, WHO Report 2008 Global tuberculosis control - surveil-
lance, planning, financing. 2008.
(2) Tripathi, R. P.; Tewari, N.; Dwivedi, N.; Tiwari, V. K. Fighting
tuberculosis: an old disease with new challenges. Med. Res. Rev.
2005, 25, 93–131.
(3) CDC Trends in Tuberculosis --- United States, 2007. Morbidity &
Mortality Weekly Report 2008, 57, 281-285.
(4) Mao, J.; Wan, B.; Wang, Y.; Franzblau, S. G.; Kozikowski, A. P.
HTS, chemical hybridization, and drug design identify a chemically
unique antituberculosis agent-coupling serendipity and rational
approaches to drug discovery. ChemMedChem 2007, 2, 811–813.
(5) Mao, J.; Yuan, H.; Wang, Y.; Wan, B.; Pak, D.; He, R.; Kozi-
kowski, A. P; Franzblau, S. G. Synthesis and Antituberculosis
Activity of Novel Mefloquine-Isoxazole Carboxylic Esters as Pro-
drugs (in preparation).
(6) Collins, L.; Franzblau, S. G. Microplate alamar blue assay versus
BACTEC 460 system for high-throughput screening of compounds
against Mycobacterium tuberculosis and Mycobacterium avium.
Antimicrob. Agents Chemother. 1997, 41, 1004–1009.
(7) Kozikowski, A. P. Methods for the stereoselective cis cyanohy-
droxylation and carboxyhydroxylation of olefins. J. Org. Chem.
1983, 48, 366–372.
(8) Cho, S. H.;Warit, S.;Wan, B.;Hwang, C. H.; Pauli, G. F.;Franzblau,
S. G. Low-oxygen-recovery assay for high-throughput screening of
compounds against nonreplicating Mycobacterium tuberculosis. Anti-
microb. Agents Chemother. 2007, 51 (4), 1380–1385.
Z-Ethyl-5-[2-(2,6-bis(trifluoromethyl)quinolin-4-yl)vinyl]isox-
azole-3-carboxylate (9l). The synthesis of 9l is similar to that of
9d. 1H NMR (300 MHz, CDCl3): δ 8.84 (s, 1H), 8.42 (d, J = 9.0,
1H), 8.15 (d, J = 16.2, 1H), 8.07 (d, J = 9.0, 1H), 8.01 (s, 1H),
7.36 (d, J = 16.2, 1H), 6.93 (s, 1H), 4.51 (q, J = 7.2, 2H), 1.48 (t,
J = 7.2 Hz, 3H). 13C NMR (75 MHz, CDCl3): δ 168.0, 159.5,
157.1, 150.1, 149.7, 148.8, 144.3, 132.3, 130.7, 128.0, 126.7,
125.7, 121.3, 121.2, 120.5, 114.3, 104.8, 65.5, 14.1. Electrospray
MS: m/z 431.0 ([Mþ H]þ, 100). Electrospray HRMS calcd for
[C19H12F6N2O3 þ H] þ: 431.0825; found 431.0816.
3. Biology. General Information. Minimum inhibitory con-
centrations against replicating and nonreplicating cultures of
Mycobacterium tuberculosis H37Rv were performed using the
microplate Alamar Blue assay (MABA)6 and low oxygen re-
covery assay (LORA)8 as previously described. Cytotoxicity for
green monkey kidney (VERO) cells was determined following
72 h exposure as previously described.11
(9) Zhang, Y.; Permar, S.; Sun, Z. Conditions that may affect the
results of susceptibility testing of Mycobacterium tuberculosis to
pyrazinamide. J. Med. Microbiol. 2002, 51 (1), 42–49.
(10) Zhang, Y.; Scorpio, A.; Nikaido, H.; Sun, Z. Role of acid pH and
deficient efflux of pyrazinoic acid in unique susceptibility of
Mycobacterium tuberculosis to pyrazinamide. J. Bacteriol. 1999,
181 (7), 2044–2049.
Acknowledgment. Support forthisproject was provided by
the Global Alliance for TB Drug Development. We thank Dr.
Zhenkun Ma and Dr. Takushi Kaneko for helpful discus-
sions. We thank Dr. Rong He and Mr. Mike Serby for editing
the manuscript.
(11) Falzari, K.; Zhu, Z.; Pan, D.; Liu, H.; Hongmanee, P.; Franzblau,
S. G. In vitro and in vivo activities of macrolide derivatives against
Mycobacterium tuberculosis. Antimicrob. Agents Chemother. 2005,
49 (4), 1447–1454.