Discovery and development of the covalent hydrates of trifluoromethylated pyrazoles as riboflavin synthase inhibitors with antibiotic activity against Mycobacterium tuberculosis

Article abstract of DOI:10.1021/jo900768c

(Chemical Equation Presented) A high-throughput screening (HTS) hit compound displayed moderate inhibition of Mycobacterium tuberculosis and Escherichia coli riboflavin synthases. The structure of the hit compound provided by the commercial vendor was rea

Full text of DOI:10.1021/jo900768c

pubs.acs.org/joc
Discovery and Development of the Covalent Hydrates of
Trifluoromethylated Pyrazoles as Riboflavin Synthase Inhibitors with
Antibiotic Activity Against Mycobacterium tuberculosis
Yujie Zhao,† Adelbert Bacher,‡ Boris Illarionov,§ Markus Fischer,§ Gunda Georg,z
Qi-Zhuang Ye,z Phillip E. Fanwick,# Scott G. Franzblau,^ Baojie Wan,^ and
Mark Cushman*,†
†Department of Medicinal Chemistry and Molecular Pharmacology, School of Pharmacy and
Pharmaceutical Sciences, and The Purdue Cancer Center, Purdue University, West Lafayette, Indiana
‡
§
€
47907, Ikosatec GmbH, Konigsbergerstrasse 74, 85748 Garching, Germany, Institute of Biochemistry and
Food Chemistry, Food Chemistry Division, University of Hamburg, D-20146 Hamburg, Germany,
zDepartment of Medicinal Chemistry, University of Kansas, Lawrence, Kansas 66047, #Department of
Chemistry, Purdue University, West Lafayette, Indiana 47907, and ^Institute for Tuberculosis Research,
College of Pharmacy, University of Illinois at Chicago, Chicago, Illinois 60612
cushman@pharmacy.purdue.edu
Received April 13, 2009
A high-throughput screening (HTS) hit compound displayed moderate inhibition of Mycobacterium
tuberculosis and Escherichia coli riboflavin synthases. The structure of the hit compound provided by the
commercial vendor was reassigned as [3-(4-chlorophenyl)-5-hydroxy-5-(trifluoromethyl)-4,5-dihydro-
1H-pyrazol-1-yl](o-tolyl)methanone (18). The hit compound had a kis of 8.7 μM vs. M. tuberculosis
riboflavin synthase and moderate antibiotic activity against both M. tuberculosis replicating phenotype
and nonreplicating persistent phenotype. Molecular modeling studies suggest that two inhibitor
molecules bind in the active site of the enzyme, and that the binding is stabilized by stacking between
the benzene rings of two adjacent ligands. The most potent antibiotic in the series proved to be [5-
(4-chlorophenyl)-5-hydroxy-3-(trifluoromethyl)-4,5-dihydro-1H-pyrazol-1-yl](m-tolyl)methanone (16),
which displayed a minimum inhibitory concentration (MIC) of 36.6 μM vs. M. tuberculosis replicating
phenotype and 48.9 μM vs. M. tuberculosis nonreplicating phenotype. The HTS hit compound and its
analogues provide the first examples of riboflavin synthase inhibitors with antibiotic activity.
Introduction
two enzymes in the riboflavin biosynthetic pathway (Scheme 1),
are therefore targets for the design of potential antibiotics.1-5
Riboflavin (4), also known as vitamin B2, is the central
component of the cofactors FAD and FMN, and is therefore
required for a wide variety of cellular processes. It plays a key
role in energy production, and is required for the metabolism of
fats, carbohydrates, and proteins. Whereas animals acquire
riboflavin completely from dietary sources, plants and patho-
genic microorganisms rely on biosynthesis. Since the riboflavin
biosynthetic pathway is absent in the human host, inhibition of
the pathway should be selectively toxic to the pathogen and not
to the host. Riboflavin synthase and lumazine synthase, the last
(1) Bacher, A.; Eberhardt, S.; Fischer, M.; Kis, K.; Richter, G. Annu.
Rev. Nutr. 2000, 20, 153–167.
(2) Kis, K.; Volk, R.; Bacher, A. Biochemistry 1995, 34, 2883–2892.
(3) Gerhardt, S.; Schott, A.-K.; Kairies, N.; Cushman, M.; Illarionov, B.;
Eisenreich, W.; Bacher, A.; Huber, R.; Steinbacher, S.; Fischer, M. Structure
2002, 10, 1371–1381. Figure 1 has been reproduced from this reference with
permission from Elsevier .
(4) Illarionov, B.; Eisenreich, W.; Bacher, A. Proc. Natl. Acad. Sci. U.S.A.
2001, 98, 7224–7229.
(5) Illarionov, B.; Kemter, K.; Eberhardt, S.; Richter, G.; Cushman, M.;
Bacher, A. J. Biol. Chem. 2001, 276, 11524–11530.
DOI: 10.1021/jo900768c
r
Published on Web 06/22/2009
J. Org. Chem. 2009, 74, 5297–5303 5297
2009 American Chemical Society

Zhao et al.
JOCArticle
SCHEME 1. The Last Two Steps in the Riboflavin Biosynthetic
Pathway
SCHEME 2. Hypothetical Mechanism of the Riboflavin Synthase-
Catalyzed Reaction (X-, Undefined Nucleophile; R, Ribityl Chain)
The final step in the biosynthesis of riboflavin is catalyzed by
riboflavin synthase. A hypothetical mechanism for the forma-
tion of riboflavin is outlined in Scheme 2. The early steps in this
proposal involve the addition of a nucleophile to the lumazine 3
that will function as the donor of the four-carbon unit to form
6, and the deprotonation of the C-7 methyl group of the
lumazine 3 that will function as the acceptor of the four-carbon
unit to form the anion 5.6-8 Although the identity of the
nucleophile has not been rigorously established, likely candi-
dates include water or one of the ribityl hydroxyl groups.9,10
Nucleophilic addition of the anion 5 to the imine 6 affords
intermediate 7, which tautomerizes to yield intermediate 8.
Elimination of the anion X- from 9 results in the iminium ion
10, which is attacked intramolecularly by the enamine to
produce the pentacyclic intermediate 11. The pentacyclic com-
pound 11 has been isolated and shown to be a kinetically
competent intermediate, and its structure has been established
by multinuclear NMR spectroscopy.4 Two sequential elimina-
tion reactions then produce the final products 1 and 4.
Riboflavin synthase is characterized by an internal sequence
repeat within each homotrimer subunit, which suggests that each
riboflavin monomer contains two topologically similar domains,
(6) Beach, R. L.; Plaut, G. W. E. J. Am. Chem. Soc. 1970, 92, 2913–2916.
(7) Beach, R. L.; Plaut, G. W. E. J. Am. Chem. Soc. 1971, 36, 3937–3943.
(8) Paterson, T.; Wood, H. S. C. J. Chem. Soc., Perkin Trans. 1 1972,
1051–1056.
(9) Cushman, M.; Patrick, D. A.; Bacher, A.; Scheuring, J. J. Org. Chem.
1991, 56, 4603–4608.
(10) Bown, D. H.; Keller, P. J.; Floss, H. G.; Sedlmaier, H.; Bacher, A.
J. Org. Chem. 1986, 51, 2461–2467.
5298 J. Org. Chem. Vol. 74, No. 15, 2009

Zhao et al.
JOCArticle
SCHEME 3. Syntheses of Compounds 15 and 16
However, those compounds are very polar substrate, inter-
mediate, and product analogues that cannot be expected to be
able to penetrate cells membranes and exert antibiotic activity.
High-throughput screening platforms were therefore developed
to provide alternative methods to obtain inhibitors with drug-
like properties.31,32 The method involving riboflavin synthase
measures the inhibition of riboflavin formation in the presence
of substrate 3. High-throughput screening of a commercial
library of about 100 000 compounds using this assay resulted in
127 hit compounds having IC50 values of 16.7 μg/mL or less,
and the tentative identification of [3-(4-chlorophenyl)-5-hydro-
xy-5-(trifluoromethyl)-4,5-dihydro-1H-pyrazol-1-yl](m-tolyl)-
methanone (15), based on the vendor’s structure, as one of
the most potent riboflavin synthase inhibitors of the series.
The present paper describes the assignment of the correct
structure 18 to the hit compound that had been identified as
15 by the vendor, as well as the confirmation of the riboflavin
synthase inhibitory activity of 15, 18, and related trifluoro-
methylated, covalently hydrated pyrazoles. This series of
compounds provides the first examples of riboflavin synthase
inhibitors with antibiotic activity.
FIGURE 1. Hypothetical structure of the active site dimer of
S. pombe riboflavin synthase with bound substrate 3.3
each of which can accommodate one substrate molecule.9,11-14
The active site has been proposed to be located at the interface of
adjacent subunits (Figure 1).3 More specifically, two substrate
lumazine molecules are assumed to be sandwiched in a tight
intermolecular interface between the N-terminal domain from
one subunit and the C-terminal domain from an adjacent subunit.
Figure 1 displays the two barrel domains of adjacent subunits
viewed along a pseudo-2-fold symmetry axis and two molecules of
bound substrate 3 in the active site, where two molecules of 3 are
ideally positioned for the dismutation reaction.3
We have previously reported the structure-based design
and synthesis of lumazine synthase and riboflavin synthase
inhibitors that are useful mechanism probes.9,12,13,15-30
(11) Otto, M.; Bacher, A. Eur. J. Biochem. 1981, 115, 511–517.
(12) Cushman, M.; Patel, H. H.; Scheuring, J.; Bacher, A. J. Org. Chem.
1992, 57, 5630–5643.
(13) Cushman, M.; Mavandadi, F.; Kugelbrey, K.; Bacher, A. Bioorg.
Med. Chem. 1998, 6, 409–415.
(14) Harvey, R. A.; Plaut, G. W. E. J. Biol. Chem. 1966, 241, 2120–2136.
(15) Cushman, M.; Patel, H. H.; Scheuring, J.; Bacher, A. J. Org. Chem.
1993, 58, 4033–4042.
Results and Discussion
In an attempt to confirm the vendor0s structure 15 of the hit
compound, samples were obtained both commercially (Chem-
Div, Inc., CD 3852-0429) and also by synthesis from m-toluic
hydrazide (13) and 1-(4-chlorophenyl)-4,4,4-trifluoro-1,3-butane-
dione (14) in refluxing ethanol containing a catalytic amount of
hydrochloric acid (Scheme 3).33 Compounds 15 [mp 94-95 ꢀC,
negative ion ESIHRMS m/z calcd for C18H13ClF3N2O2 (M -
Hþ)- 381.0618, found 381.0624] and 16 [mp 172-173 ꢀC,
ESIHRMS m/z calcd for C18H14ClF3N2O2Na (MNaþ)
405.0594, found 405.0601] were obtained by flash column chro-
matography of our synthetic material. The mass spectral data of
both compounds are consistent with a molecular weight of 382.
The ratio of the products 15 and 16 was approximately 10:1,
respectively.
(16) Scheuring, J.; Lee, J.; Cushman, M.; Pater, H.; Patrick, D. A.;
Bacher, A. Biochemistry 1994, 33, 7634–7640.
(17) Scheuring, J.;Cushman, M.;Bacher, A.J. Org. Chem. 1995, 60, 243–245.
(18) Cushman, M.; Mavandadi, F.; Kugelbrey, K.; Bacher, A. J. Org.
Chem. 1997, 62, 8944–8947.
(19) Cushman, M.; Mavandadi, F.; Yang, D.; Kugelbrey, K.; Kis, K.;
Bacher, A. J. Org. Chem. 1999, 64, 4635–4642.
(20) Kendall,P.M.;Johnson,J.V.;Cook,C.E.J. Org. Chem. 1979,44, 1421–1424.
(21) Cushman, M.; Mihalic, J. T.; Kis, K.; Bacher, A. J. Org. Chem. 1999,
64, 3838–3845.
(22) Cushman,M.;Yang,D.;Kis,K.;Bacher,A.J. Org. Chem. 2001,66, 8320–8327.
(23) Cushman, M.; Yang, D.; Mihalic, J. T.; Chen, J.; Gerhardt, S.; Huber,
R.; Fischer, M.; Kis, K.; Bacher, A. J. Org. Chem. 2002, 67, 6871–6877.
(24) Cushman, M.; Yang, D.; Gerhardt, S.; Huber, R.; Fischer, M.; Kis,
K.; Bacher, A. J. Org. Chem. 2002, 67, 5807–5816.
(25) Cushman, M.; Sambaiah, T.; Jin, G.; Illarionov, B.; Fischer, M.;
Bacher, A. J. Org. Chem. 2004, 69, 601–612.
(26) Chen, J.; Sambaiah, T.; Illarionov, B.; Fischer, M.; Bacher, A.;
Cushman, M. J. Org. Chem. 2004, 69, 6996–7003.
(27) Zhang, Y. L.; Jin, G. Y.; Illarionov, B.; Bacher, A.; Fischer, M.;
Cushman, M. J. Org. Chem. 2007, 72, 7176–7184.
(28) Cushman, M.; Jin, G.; Illarionov, B.; Fischer, M.; Ladenstein, R.;
Bacher, A. J. Org. Chem. 2005, 70, 8162–8170.
(29) Talukdar, A.; Illarionov, B.; Bacher, A.; Fischer, M.; Cushman, M.
J. Org. Chem. 2007, 72, 7167–7175.
(31) Craig, J. C.; Naik, A. R. J. Am. Chem. Soc. 1962, 84, 3410.
(32) Kaiser, J.; Illarionov, B.; Rohdich, F.; Eisenreich, W.; Saller, S.;
Van den Brulle, J.; Cushman, M.; Bacher, A.; Fischer, M. Anal. Biochem.
2007, 365, 52–61.
(30) Zhang, Y. L.; Illarionov, B.; Morgunova, E.; Jin, G. Y.; Bacher, A.;
Fischer, M.; Ladenstein, R.; Cushman, M. J. Org. Chem. 2008, 73, 2715–2724.
(33) Umarov, B. B.; Gaibullaev, K. S.; Parpiev, N. A. O’zb. Resp. Fanlar
Akad. Ma’ruzalari 2000, 9, 42–45.
J. Org. Chem. Vol. 74, No. 15, 2009 5299

Zhao et al.
JOCArticle
SCHEME 4. Mechanism Proposed for Formation of Compound 15
SCHEME 5. Mechanism Proposed for Formation of Compound 16
Mechanisms for the formation of the major product 15 and
the minor product 16 are proposed in Schemes 4 and 5. The
ketone of 14 bearing the trifluoromethyl group is expected to
exist mainly as the hemiketal derived from addition of ethanol,
the solvent.34 Consistent with this expectation was the appear-
ance of a major trifluoromethyl signal -6.71 ppm upfield from
CF3COOH (external standard) and a minor signal -1.15 ppm
upfield from CF3COOH in a 19F NMR spectrum recorded
after 14 was dissolved in dilute ethanolic HCl. The trifluoro-
methylated hemiketal could explain the preferred regiochem-
istry of addition of the hydrazine amino group to the carbonyl
attached to the p-chlorophenyl ring (Scheme 4). On the other
hand, the trifluoromethylated carbonyl group of the β-diketone
14, which is present in the reaction mixture as a minor
component, is expected to be more reactive due to the attached
highly electronegative trifluoromethyl substituent. Nucleophi-
lic attack of the terminal amino group of the hydrazine 13 on
the trifluoromethylated carbonyl of 14 would eventually lead to
the minor product 16 (Scheme 5). In both cases, the covalently
hydrated pyrazoles are stabilized by the electron-attracting
amide moiety and trifluoromethyl substituent. Both 15 and
16 are stable carbinolamides.
values as compound 16 employing different mobile systems.
However, their melting points were significantly different:
CD 3852-0429, mp 76-78 ꢀC; compound 16, mp 172-173 ꢀC.
1
Moreover, the H and 13C NMR spectral data from the
commercial sample (Chemdiv Co., CD 3852-0429), which
purportedly had structure 15, did not match with either one
of the compounds. The 1H NMR spectra of compounds 15
and 16 indicated that they both contain a methylene group in
a 4,5-dihydropyrazole ring. The structure assigned to syn-
thetic 15 was confirmed by X-ray crystallography (Figure S4,
Supporting Information). At this point, it became clear that
CD 3852-0429 has a different structure than advertised.
To assign the structures of compounds 15 and 16, 19F NMR
spectra of the synthesized compounds and the commercial com-
pound (Chemdiv Co., CD 3852-0429) were analyzed. The triflu-
oromethyl groups of CD 3852-0429 and compound 15 both
appeared in their 19FNMRspectraat-3.11 to -3.56 ppm upfield
from CF3CO2H, supporting their attachment to sp3-hybridized
carbon atoms. Meanwhile, the trifluoromethyl group of com-
pound 16 appeared at 10.33 ppm downfield from CF3CO2H,
indicating its attachment to an sp2-hybridized carbon atom.35,36
Thin layer chromatography indicated that the commercial
compound (Chemdiv Co., CD 3852-0429) had the same Rf
(35) Cushman, M.; Wong, W. C.; Bacher, A. J. Chem. Soc., Perkin Trans.
I 1986, 1043–1050.
ꢀ
(34) Fustero, S.; Roman, R.; Sanz-Cervera, J. F.; Simon-Fuentes, A.;
Cunat, A. C.; Villanova, S.; Murguıa, M. J. Org. Chem. 2008, 73, 3523–3529.
´
ꢀ
(36) Cushman, M.; Wong, W. C.; Bacher, A. J. Chem. Soc., Perkin Trans.
I 1986, 1051–1053.
~
5300 J. Org. Chem. Vol. 74, No. 15, 2009

Zhao et al.
JOCArticle
SCHEME 6. Synthesis of Compounds 18 and 19
SCHEME 7. Synthesis of Compounds 21 and 22
TABLE 1. Inhibition Constants vs. Riboflavin Synthases from M. tuber-
culosis and E. coli
The synthetic and commercial materials were studied in
more detail to explain the observed differences. The com-
mercial sample CD 3852-0429 had the same molecular
weight, similar methylene 1H NMR chemical shifts, and
similar 19F NMR chemical shifts as compound 15. However,
the 2D HMBC NMR spectrum of CD 3852-0429 indicated
that the methyl group was coupled to only one adjacent
proton, whereas that of 15 provided evidence for two ad-
jacent protons. To elucidate the structure of the commercial
sample, a series of isomeric compounds were synthesized
and characterized (Schemes 6 and 7). This led to the assign-
ment of the structure of the commercial sample from Chem-
div as [3-(4-chlorophenyl)-5-hydroxy-5-(trifluoromethyl)-
4,5-dihydro-1H-pyrazol-1-yl](o-tolyl)methanone (18) since
the melting point, IR, 1H NMR, 13C NMR, and 19F NMR
of our synthetic compound 18 and the commercial com-
pound were identical. Furthermore, a mixed mp 79-81 ꢀC of
18 and the commercial sample CD 3852-0429 was not
depressed, whereas the mixed mp 79-95 ꢀC of 15 and the
commercial sample CD 3852-0429 was depressed. These
results emphasize the necessity to confirm the structures of
hit compounds obtained from commercial libraries.
M. tuberculosis
riboflavin synthase
E. coli riboflavin
synthase
compd
parameter
15
Kia (μM)
Kisb (μM)
mechanism
Ki (μM)
Kis (μM)
mechanism
Ki (μM)
Kis (μM)
Mechanism
Ki (μM)
Kis (μM)
mechanism
Ki (μM)
14 ( 3
38 ( 11
partial
106 ( 31
91 ( 17
partial
50 ( 14
120 ( 28
partial
31 ( 10
36 ( 12
Partial
312 ( 182
57 ( 12
partial
16
18
19
21
22
22 ( 4
noncompetitive
8.7 ( 0.9
Noncompetitive
6.7 ( 1.6
16 ( 4
partial
16 ( 8
10 ( 2
partial
23 ( 14
10 ( 2
partial
135 ( 20
Kis (μM)
mechanism
Ki (μM)
Kis (μM)
mechanism
competitive
61 ( 13
104 ( 22
partial
a Ki is the inhibitor dissociation constant for the process E þ I h EI.
b Kis is the inhibitor dissociation constant for the process ES þ I h ESI.
Since the folding topologies and the amino acids of E. coli
and S. pombe enzymes in direct contact with the bound ligands
are closely similar, the structure of the E. coli active site with
bound ligands can be proposed by combining elements of the
two structures. The two ligand molecules of 6-carboxyethey-7-
oxo-8-ribityllumazine (23) in the N-terminal barrel (acceptor
site) and C-terminal barrel (donor site) were removed from the
enzyme structure, and two molecules of compound 15 were
then docked sequentially in the two sites by using GOLD
software (BST, version 3.0, 2005), which docks a flexible ligand
into a semiflexible protein structure.38 Energy minimization
was performed by using Sybyl 8.0 and the MMFF94s force
field. All possible docking patterns for different ligand config-
uration combinations (RR, SS, RS, SR in acceptor site and
donor site, respectively) were analyzed and the RS combination
had the highest binding energy (Supporting Information, Table
S1). The structure of the RS combination having the highest
GOLD score is displayed in Figure 2. The binding is stabilized
Enzyme kinetics revealed that all six compounds in this series
inhibit M. tuberculosis riboflavin synthase and E. coli riboflavin
synthase (Table 1) with Ki and Kis values in the micromolar
range. Compound 19 was the most potent inhibitor of M.
tuberculosis riboflavin synthase with a Ki of 6.7 ( 1.6 μM, while
compound 18 was the most potent inhibitor of E. coli riboflavin
synthase, with a Ki of 31 ( 10 μM in Tris buffer, pH 7.0.
Molecular modeling was performed in order to investigate
the binding of this series of compounds to the enzyme. The
crystal structure of E. coli riboflavin synthase, a homotrimer
that consists of an asymmetric assembly of monomers (mole-
cules A, B, and C), has been published along with a proposed
active site between the C-terminal barrel of molecule A and
the N-terminal barrel of molecule C (PDB code 1i8d).37
A
crystal structure has also been determined of monomeric S.
pombe riboflavin synthase with bound substrate analogue
6-carboxyethyl-7-oxo-8-ribityllumazine (PDB code 1kzl).3
(37) Liao, D.-I.; Wawrzak, Z.; Calabrese, J. C.; Viitanen, P. V.; Jordan,
D. B. Structure 2001, 9, 399–408.
(38) Jones, G.; Willett, P.; Glen, R. C.; Leach, A. R.; Taylor, R. J. Mol.
Biol. 1997, 267, 727–748.
J. Org. Chem. Vol. 74, No. 15, 2009 5301

Zhao et al.
JOCArticle
FIGURE 2. Hydrogen bonding network existing in the hypothetical active site binding model of E. coli riboflavin synthase with compound 15
(R in the acceptor site and S in the donor site). The amino acid residues are red and labeled by chain/residue names, and the ligands are colored
by atom type. The diagram is programmed for wall-eyed (relaxed) viewing.
TABLE 2. In vitro Activity of Compounds 15-22 Against M. tuberculosis
MIC (μM)a
by stacking interactions between one benzene ring of each com-
pound 15 in the binding pocket, with the remaining two benzene
rings pointing in opposite directions, which is consistent with the
stacking pattern of the substrate shown in Figure 1. In addition,
the binding of compound 15 to the enzyme could be mediated
by hydrogen bonding interactions involving the backbone
nitrogen of Thr50 and the backbone oxygen of Asp62 from
the acceptor site, as well as the backbone nitrogen, backbone
oxygen, and side chain oxygen of Ser146 from the donor site.
compd
MABAb
LORAc
15
16
18
19
21
22
>128
>128
36.6 ( 3.7
36.5 ( 0.8
>128
48.9 ( 0.9
63.7 ( 3.4
>128
>128
>128
50.5 ( 3.6
0.08
0.47
56.0 ( 3.5
0.85
>128
rifampin (RMP)
isoniazid (INH)
nitroimidazopyran (PA 824)
0.49
1.31
a MIC=minimum inhibitory concentration. b MABA=microplate
Alamar Blue assay. c LORA=low oxygen recovery assay (luciferase
readout).
aromatic substituent positions were prepared and evaluated
for inhibition of M. tuberculosis and E. coli riboflavin
synthases. All of the compounds were evaluated for activity
against both R-TB and NRP-TB M. tuberculosis phenotypes
by making use of both the MABA and LORA methods.
All six compounds showed moderate inhibition in enzyme
kinetic studies and three of them exhibit activity against
M. tuberculosis. This study provides possible leads for the
design of new riboflavin synthase inhibitors as potential
antitubercular agents. This is the first report of riboflavin
synthase inhibitors with antibiotic activity. However, the
evidence that inhibition of M. tuberculosis riboflavin
synthase is responsible for the antibiotic activity is circum-
stantial at the present time, and off-target effects may con-
tribute. On the basis of the structures of the pyrazoles
reported in this study, other targets that should be consid-
ered include M. tuberculosis CYP51,43,44 M. tuberculosis
CYP121,43,44 and M. tuberculosis UDP-galactopyranose
mutase (UGM).45 Structurally related pyrazoles that have
antibiotic activity vs. M. tuberculosis have also recently been
reported without target identification.46
Since all six compounds had been shown to inhibit
M. tuberculosis riboflavin synthase, they were evaluated for
their activity against the M. tuberculosis strain H37Rv (re-
plicating phenotype, R-TB) by using MABA39,40 (microplate
alamar blue assay) and M. tuberculosis strain H37Rv-CA-lux
AB (nonreplicating persistent phenotype, NRP-TB) by using
LORA (low oxygen recovery assay) methods.41,42 Among
the compounds shown in Table 2, three riboflavin synthase
inhibitors 16, 18, and 22 exhibited moderate antibiotic
activity against replicating M. tuberculosis (MIC=36.6,
36.5, and 50.5 μM, respectively, vs. 0.08 μM for rifampin and
0.47 μM for isoniazid) and nonreplicating M. tuberculosis
(MIC = 48.9, 63.7, and 56.0 μM vs. 0.85 μM for rifampin
and >128 μM for isoniazid).
In conclusion, a novel inhibitor of riboflavin synthase was
discovered by high-throughput screening and its structure
was corrected, and a series of its analogues with different
(43) McLean, K. J.; Marshall, K. R.; Richmond, A.; Hunter, I. S.;
Fowler, K.; Kieser, T.; Gurcha, S. S.; Besra, G. S.; Munro, A. W. Micro-
biology (Reading, U.K.) 2002, 148, 2937–2949.
(44) Menozzi, G.; Merello, L.; Fossa, P.; Schenone, S.; Ranise, A.; Mosti,
L.; Bondavalli, F.; Loddo, R.; Murgioni, C.; Mascia, V.; La Colla, P.;
Tamburini, E. Bioorg. Med. Chem. 2004, 12, 5465–5483.
(45) Carlson, E. E.; May, J. F.; Kiessling, L. L. Chem. Biol. 2006, 13, 825–
837.
(39) Collins, L. A.; Franzblau, S. G. Antimicrob. Agents Chemother.
1997, 41, 1004–1009.
(40) Falzari, K.; Zhu, Z. H.; Pan, D. H.; Liu, H. W.; Hongmanee, P.;
Franzblau, S. G. Antimicrob. Agents Chemother. 2005, 49, 1447–1454.
(41) Jayaprakash, S.; Iso, Y.; Wan, B.; Franzblau, S. G.; Kozikowski,
A. P. ChemMedChem 2006, 1, 593–.
(42) Bate, A. B.; Kalin, J. H.; Fooksman, E. M.; Amorose, E. L.; Price,
C. M.; Williams, H. M.; Rodig, M. J.; Mitchell, M. O.; Cho, S. H.; Wang,
Y. H.; Franzblau, S. G. Bioorg. Med. Chem. Lett. 2007, 17, 1346–1348.
(46) Castagnolo, D.; De Logu, A.; Radi, M.; Bechi, B.; Manetti, F.;
Magnani, M.; Supino, S.; Meleddu, R.; Chisu, L.; Botta, M. Bioorg. Med.
Chem. 2008, 16, 8587–8591.
5302 J. Org. Chem. Vol. 74, No. 15, 2009

Zhao et al.
JOCArticle
Experimental Section
[5-(4-Chlorophenyl)-5-hydroxy-3-(trifluoromethyl)-4,5-dihydro-
1H-pyrazol-1-yl](o-tolyl)methanone (19). Compound 19 was sepa-
rated from 18 by silica gel column chromatography, eluting with
dichloromethane-hexane 3:1, to afford a light yellow powder
(28.6 mg, 7.5%): mp 201-202 ꢀC. IR (KBr) 3401, 1666 cm-1; 1H
NMR (300 MHz, CDCl3) δ 7.46-7.36 (m, 6 H), 7.28-7.23 (m,
2 H), 5.29 (s, 1 H), 3.55 (d, J=18.57 Hz, 1 H), 3.21 (d, J=19.45 Hz,
1 H), 2.37 (s, 3 H); 13C NMR (75 MHz, CDCl3) δ 171.0, 143.9 (q,
J=38.9 Hz), 140.7, 136.5, 134.8, 133.0, 130.7, 129.2, 128.4, 125.4,
125.3, 121.2, 119.4 (q, J=269.7 Hz), 94.9, 48.00, 19.8; 19F NMR
(300 MHz, CDCl3) δ 10.28 (s, 3 F); negative ion ESIMS m/z 381
{[M(Cl35) - Hþ]-, 100}, 383 {[M(Cl37) - Hþ]-, 29}, negative ion
HR ESIMS calcd for (M - Hþ)- 381.0618, found 381.0618.
[3-(4-Chlorophenyl)-5-hydroxy-5-(trifluoromethyl)-4,5-dihydro-
1H-pyrazol-1-yl](p-tolyl)methanone (21). 1-(4-Chlorophenyl)-4,4,
4-trifluoro-1,3-butanedione (250.6 mg, 1 mmol) was added to a
stirred solution of p-toluic hydrazide (150.2 mg, 1 mmol) in
ethanol (17 mL). The mixture was heated under reflux for 10 h.
Removal of the solvent produced a yellow oil, which was sub-
jected to silica gel chromatography with mobile phase dichloro-
methane-hexane 3:1 to afford 21 as white crystals (61.5 mg,
16.1%): mp 125-126 ꢀC. IR (KBr) 3369, 1647 cm-1; 1H NMR
(300 MHz, CDCl3) δ 7.90 (d, J=8.20 Hz, 2 H), 7.57 (dt, J1=
8.58 Hz, J2=1.87 Hz, 2 H), 7.40 (dt, J1 = 8.52 Hz, J2=1.76 Hz,
2 H), 7.29 (d, J=8.02 Hz, 2 H), 6.80 (s, 1 H), 3.70 (d, J=18.55 Hz,
1 H), 3.53 (d, J=18.58 Hz, 1 H), 2.45 (s, 3 H); 13C NMR
(75 MHz, CDCl3) δ 171.1, 151.8, 143.1, 137.1, 130.5, 129.8,
129.1, 128.6, 128.4, 127.9, 123.4 (q, J=285.45 Hz), 93.1 (q, J=
33.9 Hz), 43.00, 21.6; 19F NMR (300 MHz, CDCl3) δ -3.35 (s,
3 F); ESIMS m/z 405 (M(Cl35)Naþ, 100), 407 (M(Cl37)Naþ, 30);
negative ion ESIMS 381 {[M(Cl35) - Hþ]-, 100}, 383 {[M(Cl37) -
Hþ]-, 27}, negative ion HR ESIMS calcd for (M-Hþ)- 381.0618,
found 381.0620. Anal. Calcd for C18H14ClF3N2O2: C, 56.48; H,
3.69; N, 7.32. Found: C, 56.42; H, 3.65; N, 7.21.
19F NMR Spectroscopy. The 19F NMR chemical shifts are
reported relative to CF3COOH (external standard).
[3-(4-Chlorophenyl)-5-hydroxy-5-(trifluoromethyl)-4,5-dihydro-
1H-pyrazol-1-yl](m-tolyl)methanone (15). 1-(4-Chlorophenyl)-4,4,
4-trifluoro-1,3-butanedione (751.8 mg, 3 mmol) was added to a
stirred solution of m-toluic hydrazide (450.54 mg, 3 mmol) in
ethanol (50 mL). The mixture was heated at reflux for 10 h.
Concentrated HCl (2 mL) was added to the reaction mixture,
which was then stirred for another 2 h. Removal of the solvent
produced a yellow oil, which was subjected to silica gel column
chromatography, eluting with dichloromethane-hexane 3:1, to
afford white crystals (702.3 mg, 61.2%): mp 94-95 ꢀC. The mixed
mp of 15 and the commercial sample CD 3852-0429 was 79-
95 ꢀC. IR (KBr) 3379, 1648 cm-1; 1H NMR (300 MHz, CDCl3)
δ 7.84-7.81 (m, 2 H), 7.57-7.54 (m, 2 H), 7.40-7.35 (m, 4 H),
6.87 (s, 1 H), 3.72 (d, J=18.57 Hz, 1 H), 3.54 (d, J=18.57 Hz, 1 H),
2.45 (s, 3 H); 13C NMR (75 MHz, CDCl3) δ 171.2, 152.00, 137.6,
137.00, 133.0, 132.8, 130.7, 129.4, 128.4, 127.9, 127.7, 127.5, 123.4
(q, J=285.2 Hz), 93.05 (q, J=33.7 Hz), 42.9, 21.2; 19F NMR
(300 MHz, CDCl3) δ -3.11 (s, 3 F); negative ion ESIMS m/z 381
{[M(Cl35) - Hþ]-, 100}, 383 {[M(Cl37) - Hþ]-, 27}, negative ion
HR ESIMS calcd for (M - Hþ)- 381.0618, found 381.0624. The
crystal structure was solved by X-ray diffraction analysis and the
data are summarized as follows: C18H14ClF3N2O2; FW=382.77;
a=15.9749(2) A; b=16.4925(3) A; c=27.4117(6) A; R=81.5285
(7)ꢀ; β=82.2566(7)ꢀ; γ=89.0766(14)ꢀ; vol=7078.1(2) A3; triclinic;
space group P1 (No. 2); Z = 16; crystal size = 0.35 ꢀ 0.28 ꢀ
2
0.13 mm3; GOF=1.111; R(Fo)=0.075; Rw(Fo )=0.123. Anal.
Calcd for C18H14ClF3N2O2: C, 56.48; H, 3.69; N, 7.32. Found: C,
56.33; H, 3.66; N, 7.22.
[5-(4-Chlorophenyl)-5-hydroxy-3-(trifluoromethyl)-4,5-dihydro-
1H-pyrazol-1-yl](m-tolyl)methanone (16). Compound 16 was se-
parated from 15 by silica gel column chromatography, eluting
with dichloromethane-hexane 3:1, to afford a light yellow pow-
[5-(4-Chlorophenyl)-5-hydroxy-3-(trifluoromethyl)-4,5-dihydro-
1H-pyrazol-1-yl](p-tolyl)methanone (22). Compound 22 was sepa-
rated from 21 by silica gel column chromatography, eluting with
dichloromethane-hexane 3:1, to afford yellow crystals (10 mg,
der (72.4 mg, 6.3%): mp 172-173 ꢀC. IR (KBr) 3391, 1652 cm-1
;
1H NMR (300 MHz, CDCl3) δ 7.72-7.69 (m, 2 H), 7.43-7.32 (m,
6 H), 5.22 (s, 1 H), 3.49 (d, J=19.14 Hz, 1 H), 3.17 (d, J=19.14 Hz,
1 H), 2.39 (s, 3 H); 13C NMR (75 MHz, CDCl3) δ 169.4, 143.6 (q,
J=38.72 Hz), 140.6, 137.9, 134.6, 133.4, 131.9, 130.7, 129.2, 127.9,
127.4, 125.5, 119.53 (q, J = 269.4 Hz), 95.6, 47.4, 21.3; 19F NMR
(300 MHz, CDCl3) δ 10.33 (s, 3 F); ESIMS m/z 405 [M(Cl35)Naþ,
100], 407 [M(Cl37)Naþ, 33], negative ion ESIMS 381 {[M(Cl35) -
Hþ]-, 100}, 383 {[M(Cl37) - Hþ]-, 25}, HR ESIMS calcd for
[M(Cl35)Naþ] 405.0594, found 405.0601.
1
2.6%): mp 159-160 ꢀC. IR (KBr) 3436, 1655 cm-1; H NMR
(300 MHz, CDCl3) δ 7.85 (d, J = 8.22 Hz, 2 H), 7.41 (dq, J1=
8.72 Hz, J2=2.20 Hz, 4 H), 7.27 (t, J1=J2=4.54 Hz, 2 H), 5.22 (s,
1 H), 3.50 (d, J=19.05 Hz, 1 H), 3.19 (d, J=19.13 Hz, 1 H), 2.43
(s, 3 H); 13C NMR (75 MHz, CDCl3) δ 169.1, 143.4, 143.4
(q, J = 38.93 Hz), 140.6, 134.6, 130.4, 129.3, 129.1, 128.9,
128.72, 125.46, 119.49 (q, J = 269.65 Hz), 95.6, 47.2, 21.6;
19F NMR (300 MHz, CDCl3) δ 10.32 (s, 3 F); ESIMS m/z 405
[M(Cl35)Naþ, 100], 407 [M(Cl37)Naþ, 32], negative ion ESIMS
381 {[M(Cl35) - Hþ]-, 100}, 383 {[M(Cl37) - Hþ]-, 26}, HR
ESIMS calcd for (MNaþ) 405.0594, found 405.0602.
[3-(4-Chlorophenyl)-5-hydroxy-5-(trifluoromethyl)-4,5-dihydro-
1H-pyrazol-1-yl](o-tolyl)methanone (18). 1-(4-Chlorophenyl)-4,4,
4-trifluoro-1,3-butanedione (250.6 mg, 1 mmol) was added to a
stirred solution of o-toluic hydrazide (150.2 mg, 1 mmol) in
ethanol (17 mL). The mixture was heated at reflux for 10 h.
Removal of the solvent produced a yellow oil, which was sub-
jected to silica gel column chromatography, eluting with dichloro-
methane-hexane 3:1, to afford a white powder (147.7 mg,
38.7%): mp 77-78 ꢀC. The mixed mp of 18 and the commercial
Acknowledgment. This research was made possible by
NIH grant GM51469, the Fonds der Chemischen Industrie,
and the Hans Fischer Gesellschaft. Some of this research was
conducted in a facility constructed with the financial support
of a Research facilities Improvement program grant C06-
14499 from the National Institutes of Health.
sample CD 3852-0429 was 79-81 ꢀC. IR (KBr) 3392, 1652 cm-1
;
1H NMR (300 MHz, CDCl3) δ 7.49-7.26 (m, 8 H), 6.59 (s, 1 H),
3.72 (d, J=18.57 Hz, 1 H), 3.57 (d, J=18.62 Hz, 1 H), 2.41 (s, 3 H);
13C NMR (125 MHz, CDCl3) δ 172.4, 152.4, 137.2, 136.2, 134.2,
130.47, 130.4, 129.1, 128.3, 128.2, 127.9, 125.3, 123.4 (q, J =
285.88 Hz), 92.4 (q, J=34.04 Hz), 43.3, 19.6; 19FNMR(300 MHz,
CDCl3) δ -3.55 (s, 3 F); ESIMS m/z 381 {[M(Cl35)Hþ]-, 100},
383 {[M(Cl37)Hþ]-, 33}, negative ion HR ESIMS calcd for
(M-Hþ)- 381.0618, found 381.0610. Anal. Calcd for C18H14-
ClF3N2O2: C, 56.48; H, 3.69; N, 7.32. Found: C, 56.40; H, 3.67;
N, 7.27.
Supporting Information Available: Figures of hypothetical
E.coli riboflavin synthase active site models with other config-
urations of compound 15, the NMR spectra of all compounds,
the CIF file for the crystal structure of dihydropyrazole 15, the
molecular modeling procedure for compound 15 (Figure 2),
the riboflavin synthase inhibition assay procedure, and the
M. tuberculosis inhibition assay procedure. This material is
available free of charge via the Internet at http://pubs.acs.org.
J. Org. Chem. Vol. 74, No. 15, 2009 5303