Subnanomolar inhibitor of cytochrome bc1 complex designed by optimizing interaction with conformationally flexible residues

Article abstract of DOI:10.1021/ja905756c

Cytochrome bc₁ complex (EC 1.10.2.2, bc₁), an essential component of the cellular respiratory chain and the photosynthetic apparatus in photosynthetic bacteria, has been identified as a promising target for new drugs and agricultural fungicides. X-ray diffraction structures of the free bc₁ complex and its complexes with various inhibitors revealed that the phenyl group of Phe274 in the binding pocket exhibited significant conformational flexibility upon different inhibitors binding to optimize respective π-π interactions, whereas the side chains of other hydrophobic residues showed conformational stability. Therefore, in the present study, a strategy of optimizing the π-π interaction with conformationally flexible residues was proposed to design and discover new bc1 inhibitors with a higher potency. Eight new compounds were designed and synthesized, among which compound 5c, with a K_i value of 570 pM, was identified as the most promising drug or fungicide candidate, significantly more potent than the commercially available bc1 inhibitors, including azoxystrobin (AZ), kresoxim-methyl (KM), and pyraclostrobin (PY). To our knowledge, this is the first bc₁ inhibitor discovered from structure-based design with a potency of subnanomolar K_i value. For all of the compounds synthesized and assayed, the calculated binding free energies correlated reasonably well with the binding free energies derived from the experimental Ki values, with a correlation coefficient of r² ) 0.89. The further inhibitory kinetics studies revealed that 5c is a noncompetitive inhibitor with respect to substrate cytochrome c, but it is a competitive inhibitor with respect to substrate ubiquinol. Due to its subnanomolar K_i potency and slow dissociation rate constant (k_-0) 0.00358 s^-1), 5c could be used as a specific probe for further elucidation of the mechanism of bc₁ function and as a new lead compound for future drug discovery.

Full text of DOI:10.1021/ja905756c

Published on Web 11/23/2009
Subnanomolar Inhibitor of Cytochrome bc1 Complex
Designed by Optimizing Interaction with Conformationally
Flexible Residues
Pei-Liang Zhao,† Le Wang,‡ Xiao-Lei Zhu,† Xiaoqin Huang,§ Chang-Guo Zhan,*,§
Jia-Wei Wu,*,‡ and Guang-Fu Yang*,†
Key Laboratory of Pesticide & Chemical Biology, College of Chemistry, Ministry of Education,
Central China Normal UniVersity, Wuhan 430079, P.R. China, MOE Key Laboratory of
Bioinformatics, Department of Biological Sciences and Biotechnology, Tsinghua UniVersity,
Beijing 100084, P.R. China, and Department of Pharmaceutical Sciences, College of Pharmacy,
UniVersity of Kentucky, 725 Rose Street, Lexington, Kentucky 40536
Received July 12, 2009; E-mail: gfyang@mail.ccnu.edu.cn; jiaweiwu@mail.tsinghua.edu.cn; zhan@uky.edu
Abstract: Cytochrome bc1 complex (EC 1.10.2.2, bc1), an essential component of the cellular respiratory
chain and the photosynthetic apparatus in photosynthetic bacteria, has been identified as a promising target
for new drugs and agricultural fungicides. X-ray diffraction structures of the free bc1 complex and its
complexes with various inhibitors revealed that the phenyl group of Phe274 in the binding pocket exhibited
significant conformational flexibility upon different inhibitors binding to optimize respective π-π interactions,
whereas the side chains of other hydrophobic residues showed conformational stability. Therefore, in the
present study, a strategy of optimizing the π-π interaction with conformationally flexible residues was
proposed to design and discover new bc1 inhibitors with a higher potency. Eight new compounds were
designed and synthesized, among which compound 5c, with a Ki value of 570 pM, was identified as the
most promising drug or fungicide candidate, significantly more potent than the commercially available bc1
inhibitors, including azoxystrobin (AZ), kresoxim-methyl (KM), and pyraclostrobin (PY). To our knowledge,
this is the first bc1 inhibitor discovered from structure-based design with a potency of subnanomolar Ki
value. For all of the compounds synthesized and assayed, the calculated binding free energies correlated
reasonably well with the binding free energies derived from the experimental Ki values, with a correlation
coefficient of r2 ) 0.89. The further inhibitory kinetics studies revealed that 5c is a noncompetitive inhibitor
with respect to substrate cytochrome c, but it is a competitive inhibitor with respect to substrate ubiquinol.
Due to its subnanomolar Ki potency and slow dissociation rate constant (k-0 ) 0.00358 s-1), 5c could be
used as a specific probe for further elucidation of the mechanism of bc1 function and as a new lead compound
for future drug discovery.
Introduction
known to catalyze the electron transfer from quinol to a soluble
cytochrome c (cyt c) and couple this electron transfer to the
Understanding drug-receptor interactions is the molecular
basis of designing new compounds with higher potency.1 In
fact, the interaction between drug and receptor is the interaction
between the bioactive conformation of a drug molecule and the
residues in the binding pocket of its receptor. A drug molecule
will display the highest potency when its bioactive conformation
matches very well with the binding pocket. Generally speaking,
some residues in the binding pocket of a receptor protein might
be conformationally flexible, while the others are rigid. There-
fore, optimizing the interactions with the conformationally
flexible residues in the binding pocket should be an effective
way to improve the potency of a drug.
translocation of protons across the membrane.2-5 The bc1
complex has been found in the plasma membrane of bacteria
and in the inner mitochondrial membrane of eukaryotes. Due
to its crucial role in the life cycle, the bc1 complex has been
identified as a promising target for new drugs and agricultural
fungicides.6,7 For example, malaria, especially chloroquine- and
multidrug-resistant Plasmodium falciparum malaria, remains a
major threat to over 40% of the world’s population and is
(2) Kim, H.; Xia, D.; Yu, C. A.; Xia, J. Z.; Kachurin, A. M.; Zhang, L.;
Yu, L.; Deisenhofer, J. Proc. Natl. Acad. Sci. U.S.A. 1998, 95, 8026–
8033.
(3) Gao, X. G.; Wen, X. L.; Yu, G. A.; Esser, L.; Tsao, S.; Quinn, B.;
Zhang, L.; Yu, L. D.; Xia, D. Biochemistry 2002, 41, 11692–11702.
(4) Berry, E. A.; Huang, L. S. FEBS Lett. 2003, 555, 13–20.
(5) Esser, L.; Quinn, B.; Li, Y. F.; Zhang, M. Q.; Elberry, M.; Yu, L. D.;
Yu, C. A.; Xia, D. J. Mol. Biol. 2004, 341, 281–302.
(6) Sauter, H.; Steglich, W.; Anke, T. Angew. Chem., Int. Ed. 1999, 38,
1328–1349.
Cytochrome bc1 complex (EC 1.10.2.2, bc1) is well known
as an essential component of the cellular respiratory chain and
the photosynthetic apparatus in photosynthetic bacteria. It is
† Central China Normal University.
‡ Tsinghua University.
(7) Xiang, H.; McSurdy-Freed, J.; Moorthy, G. S.; Hugger, E.; Bambal,
R.; Han, C.; Ferrer, S.; Gargallo, D.; Davis, C. B. J. Pharm. Sci. 2006,
95, 2657–2672.
§ University of Kentucky.
(1) Ladbury, J. E. Thermochim. Acta 2001, 380, 209–215.
9
10.1021/ja905756c  2010 American Chemical Society
J. AM. CHEM. SOC. 2010, 132, 185–194 185

A R T I C L E S
Zhao et al.
Scheme 1. Chemical Structures of Some Representative bc1 Inhibitors
associated with about a hundred million clinical cases every
year. Atovaquone (Scheme 1) has been used in treating
multidrug-resistant malaria and for prophylaxis in areas with
chloroquine resistance for many years.8-12 The action mecha-
nism of Atovaquone is to inhibit the bc1 complex by binding to
the ubiquinol oxidation pocket of the bc1 complex, where it
interacts with the Rieske iron-sulfur protein.11,12 In addition,
a number of Qo-specific inhibitors of the bc1 complex, always
named QoI fungicides, have been introduced into the agricultural
fungicide market.13,14 The first QoI fungicides entered the market
in 1996, including azoxystrobin (AZ) and kresoxim-methyl
(KM) (Scheme 1), two synthetic analogues of the natural
antifugal compound strobilurin A. With a distributor sale value
of over US$900 million apiece in 2008, AZ currently represents
one of the most important fungicides in the world.
bc1 complex. Therefore, discovery of new bc1 inhibitors with a
higher potency is of great interest.15-17
X-ray diffraction structures of the free bc1 complex and its
complexes with various MOA-type inhibitors revealed that the
conformation of the MOA moiety in the binding pocket is well
conserved,3,5 especially for the methoxy and carbonyl oxygen
atoms, which makes it possible to identify common structural
features that are important for binding. In addition, the Q0 pocket
is hydrophobic and exceptionally rich in highly conserved
residues (Supporting Information, Figure 1s). Most importantly,
the Ar-Ar stacking interactions with the hydrophobic pocket
formed by the side chains of Phe274, Phe128, Ile146, Pro270,
Ala277, Leu294, Met124, and Ile298 (numbered according to
the cyt b subunit of the bovine bc1) have an important
contribution to the binding affinity of inhibitors.3,5 It is
noteworthy that the side chains of all these hydrophobic residues
showed conformational stability during the simulation, except
for the phenyl group of Phe274, which exhibited significant
conformational flexibility in different complexes. The confor-
mational flexibility of the Phe274 side chain allows us to
optimize the corresponding Ar-Ar interactions. However, due
to the existence of the linear cyano group in AZ, the phenyl
ring of the cyanophenoxyl group is almost vertical to the
pyrimidyl ring. As a result, the pyrimidyl ring does not parallel
well with the phenyl ring of Phe274 to reach an ideal Ar-Ar
interaction. In contrast, myxothiazol has a higher potency than
AZ because its thiazole ring parallels well with the phenyl group
of Phe274. These facts suggest that improving the Ar-Ar
interactions between the side-chain of inhibitors and the phenyl
of Phe274 may be an effective way to design and discover new,
more potent inhibitors of bc1 complex.
According to structural information determined by crystal-
lographic studies, Xia et al.5 suggested that the existing bc1
inhibitors could be grouped into three classes: class P, class N,
and class PN. Class P inhibitors, binding at the Q0 site, include
famoxadone, stigmatellin, 5-undecyl-6-hydroxy-4,7-dioxoben-
zothiazole (UHDBT), methoxyacrylate stilbene (MOAS),
myxothiazol, AZ, KM, pyraclostrobin (PY), and many others.
Class N inhibitors, targeting the Qi site, include antimycin A
and diuron. Class PN inhibitors, with the capability of binding
to both the Q0 and Qi sites, include 2-n-nonyl-4-hydroxyquino-
line N-oxide (NQNO) and possibly funiculosin. Among these
inhibitors, MOA-type inhibitors, such as strobilurin, oudemansin,
and AZ, are very important due to their great commercial
success and potential role in elucidating the molecular mech-
anism of bc1 function. However, these commercial inhibitors
as agricultural fungicides or drugs are suffering from the rapid
development of resistance associated with site mutations of the
On the basis of the above premises, the development of new,
more potent bc1 inhibitors by improving the Ar-Ar interactions
of AZ is the main strategy of this work. Without altering the
main structural features of MOA-type inhibitors, we first
designed the molecular framework I (Scheme 2). As an
important kind of five-membered aromatic scaffold with broad-
spectrum biological activities, the pyrazole ring appears in many
(8) Yeates, C. L.; et al. J. Med. Chem. 2008, 51, 2845–2852.
(9) Fry, M.; Pudney, M. Biochem. Pharmacol. 1992, 43, 1545–1553.
(10) Kessl, J. J.; Meshnick, S. R.; Trumpower, B. L. Trends Parasitol.
2007, 23, 494–501.
(11) Mather, M. W.; Darrouzet, E.; Valkova-Valchanova, M.; Cooley, J. W.;
McIntosh, M. T.; Daldal, F.; Vaidya, A. B. J. Biol. Chem. 2005, 280,
27458–27465.
(12) Kessl, J. J.; Lange, B. B.; Merbitz-Zahradnik, T.; Zwicker, K.; Hill,
P.; Meunier, B.; Palsdottir, H.; Hunte, C.; Meshnick, S.; Trumpower,
B. L. J. Biol. Chem. 2003, 278, 31312–31318.
(15) Bolgunas, S.; Clark, D. A.; Hanna, W. S.; Mauvais, P. A.; Pember,
S. O. J. Med. Chem. 2006, 49, 4762–4766.
(16) Ding, M. G.; di Rago, J. P.; Trumpower, B. L. J. Biol. Chem. 2006,
281, 36036–36043.
(13) Bartlett, D. W.; Clough, J. M.; Godwin, J. R.; Hall, A. A.; Hamer,
M.; Parr-Dobrzanski, B. Pest Manag. Sci. 2002, 58, 649–662.
(14) Balba, H. J. EnViron. Sci. Health B 2007, 42, 441–451.
(17) Kolesova, G. M.; Belyakova, M. M.; Mamedov, M. D.; Yaguzhinsky,
L. S. Biochemistry-Moscow 2000, 65, 578–582.
9
186 J. AM. CHEM. SOC. VOL. 132, NO. 1, 2010

Subnanomolar Inhibitor of Cytochrome bc1
A R T I C L E S
Scheme 2. Proposed Structures of New bc1 Inhibitors with an Intramolecular Hydrogen Bond
commercial drugs.18-20 Hence, we replaced the pyrimidine ring
of AZ with a methylpyrazole ring, as shown in Scheme 2.
Meanwhile, an ortho-hydroxy group was designed to replace
the cyano group, with the expectation that an intramolecular
hydrogen bond between the hydroxy group and the N(1) atom
of pyrazole ring could be formed, making it possible for the
pyrazole ring to be perfectly parallel with the phenyl group of
Phe274, improving the Ar-Ar interactions. In addition, in order
to adjust the hydrophobic property of the designed compounds,
some functional groups, such as Cl and CH3, were also
introduced into the R-position of the phenyl moiety.
The strategy we proposed for the selection of compounds to
be synthesized makes use of computational simulation protocols.
These allow us to analyze the key interactions at the atomic
level between the selected compounds and the bc1 complex.
As a result, one compound with a Ki value of 570 pM against
porcine bc1 complex was discovered. To our knowledge, it is
so far the first bc1 inhibitor discovered through structure-based
design with subnanomolar Ki potency. The synthesis, in Vitro
enzymatic assay, and subsequent enzymatic kinetic study of this
bc1 inhibitor are reported in this work. Its excellent potency
and interesting characteristics of enzymatic kinetic behavior
indicate that this compound could be used as a specific probe
for further elucidating the mechanism of bc1 function and as a
new lead compound for future drug discovery.
to remove any possible steric overlap (or hindrance) with the
neighboring conserved residues. The structures of other subunits
of the porcine cyt bc1 complex were not modeled due to the shortage
of their amino acid sequences of some of the subunits. In order to
view the structural integrity, the structures of other subunits of the
porcine cyt bc1 complex were directly copied from the template,
proving their high homology to the bovine heart cyt bc1 complex.
After initial coordinates were generated, all the ionizable residues
of cyt b were set to the standard protonation states. A careful check
of the structural model allowed the proton to be assigned to the
ND1 atoms for all the His residues under the physiological condition
(pH ∼7.4), including His83, His97, His182, and His196 coordinated
with the central irons of the two heme groups. As the His residues
are coordinated optimally with the two heme groups, as seen from
the crystal structure, these residues and heme groups were always
fixed in the following energy minimization steps. The initial 3D
model of the porcine cyt b was energy-minimized by using the
Sander module of the Amber8 program21 suite with a nonbonded
cutoff of 12 Å and a conjugate gradient minimization method. The
heme groups were treated at its oxidized state, and the input
parameters of these heme groups were directly adopted from Izrailev
et al.22 in their steered molecular dynamics simulations. The energy
minimization was performed in the gas phase, first for 5000 steps
with the backbone atoms fixed while the side-chain atoms were
relaxed, and then for another 1000 steps with the side-chain atoms
constrained in order to relax the backbone. After two rounds of
these partial energy-minimization runs, the convergence criterion
of 0.001 kcal mol-1 Å-1 was quickly achieved in a full energy
minimization.
Materials and Methods
Molecular Docking. As observed from the crystal structure, AZ
is bound in the Qo site, which is formed by loop ef, helices cd1
and cd2, and part of the helices B, C, E, and F. Considering the
structural similarity of AZ with other inhibitors, this Qo site was
also selected as a binding site for other inhibitors. Based on the
3D model of the porcine cyt b, the AutoDock 4.0 program23 was
applied to dock these inhibitors into the Qo binding site. The
Gasteiger charges were used for these inhibitors. To select the best
set of docking parameters and to test the reliability of the docking
results, the AZ was first docked into the Qo binding site. In the
docking process, a conformational search was performed for the
AZ molecule using the Solis and Wets local search method, and
the Lamarkian genetic algorithm (LGA) was applied for the
conformational search of the binding complex of AZ with porcine
cyt b. Among a series of docking parameters, the grid size was set
to be 40 × 40 × 40, and the used grid space was the default value
of 0.375 Å. The interaction energy that resulted from probing the
porcine cyt b with the AZ molecule was assessed by the standard
AutoDock scoring function. Among a set of 50 candidates of the
docked complex structures, the best one was first selected according
1. Computational Methods. Homology Modeling of the
Porcine Cytochrome b Subunit (cyt b). In order to study the
binding of a set of representative inhibitors with the porcine
mitochondrial cytochrome bc1 complex, a homology model of
the cyt b was built by using the complete crystal structure of bovine
heart cytochrome bc1 in complex with AZ (PDB entry 1SQB) as
the template and the Homology module of the InsightII program
(version 2000, Accelrys, Inc., San Diego, CA). As one of the
available cytochrome bc1 crystal structures containing inhibitors,
the AZ-bound bc1 complex was selected as the template because
of its structural similarity to the compounds used in this study. The
sequence alignment between the porcine cyt b and the same subunit
of the template was generated by ClusterW with the Blosum scoring
function. The best alignment was selected according to both the
alignment score and the reciprocal positions of the conserved
residues, especially those in or close to the inhibitor-binding sites
of the template. The coordinates of the conserved regions, including
the two heme groups at the bL and bH sites, were transformed
directly from the template structure, whereas the nonequivalent
residues were mutated from the template to the corresponding ones
in the porcine cyt b. The side chains of these nonconserved residues
were relaxed by using the Homology module of InsightII, in order
(21) Pearlman, D. A.; Case, D. A.; Caldwell, J. W.; Ross, W. S.; Cheatham,
T. E.; Debolt, S.; Ferguson, D.; Seibel, G.; Kollman, P. Comput. Phys.
Commun. 1995, 91, 1–41.
(18) McDonald, E.; Jones, K.; Brough, P. A.; Drysdale, M. J.; Workman,
P. Curr. Top. Med. Chem. 2006, 6, 1193–1203.
(22) Izrailev, S.; Crofts, A. R.; Berry, E. A.; Schulten, K. Biophys. J. 1999,
77, 1753–1768.
(19) Lamberth, C. Heterocycles 2007, 71, 1467–1502.
(23) Morris, G. M.; Goodsel, D. S.; Halliday, R. S.; Huey, R.; Hart, W. E.;
Belew, R. K.; Olson, A. J. J. Comput. Chem. 1998, 19, 1639–1662.
(20) Hansen, J.; et al. Bioorg. Med. Chem. Lett. 2008, 18, 4692–4695.
9
J. AM. CHEM. SOC. VOL. 132, NO. 1, 2010 187

A R T I C L E S
Zhao et al.
to the interaction energy and was then compared with the
conformation of AZ extracted from the crystal structure. By tuning
the docking parameters, the final AZ-cyt b complex was obtained,
based on the smallest root-mean-square deviation (rmsd) of the
docked conformation of AZ from its conformation in the crystal
structure. The same set of docking parameters were adopted in the
molecular docking of other inhibitors, including KM, PY, 5a-d,
and 6a-d, into the Qo binding site of cyt b.
All the complex structures derived from molecular docking were
used as starting structures for further energy minimizations using
the Sander module of the Amber8 program before the final binding
structures were achieved. The atomic charges used for these
inhibitors were the restrained electrostatic potential (RESP) charges,
determined by using the standard RESP procedure implemented in
the Antechamber module of the Amber8 program following the
electronic structure and electrostatic potential calculations at the
HF/6-31G* level. The energy minimization process was similar to
that used for modeling the porcine cyt b, i.e., first fixing the
backbone atoms of the protein, in order to relax the side chains
and the docked inhibitor molecule. Energy minimization was then
performed on the whole complex until the convergence criterion
of 0.001 kcal mol-1 Å-1 was reached.
Binding Energy Calculations. Due to the fact that only the
structure of cyt b of the porcine bc1 complex was modeled, and
due to the expected geometric symmetry of the dimer structure of
the porcine bc1 complex, the binding free energy for each inhibitor
with the porcine bc1 complex was represented by the binding free
energy for each inhibitor with the cyt b. Such an approximation
should be reasonable, as all the other subunits are far away from
the Qo binding site and have no direct interactions with the inhibitors
binding at the Qo site. Their effects on the inhibitor binding are on
the long-range interactions; therefore, such long-range interactions
can be treated as background, without significant influence on the
calculated binding free energy (∆Gbind). On the basis of the modeled
complex structure of porcine cyt b bound with inhibitors, the binding
free energy for each of the minimized complexes was estimated
by using the molecular mechanics Possion-Boltzmann surface area
(MM-PBSA) method.24
free energy was calculated at T ) 300 K by using an in-house
program that has proven to be reliable in our other studies.27-33
2. Syntheses of the Title Compounds. Unless noted otherwise,
reagents and starting materials 1a-c were purchased from com-
mercial suppliers and used without further purification whereas all
solvents were redistilled before use. 1H NMR spectra were recorded
on a Mercury-Plus 400 spectrometer in CDCl3 with TMS as the
internal reference. MS spectra were determined using a Trace MS
2000 organic mass spectrometry. Elementary analyses were per-
formed on a Vario EL III elemental analysis instrument. Melting
points were taken on a Buchi B-545 melting point apparatus and
are uncorrected. Intermediates 2a-c, 3a-c, and 4a,b were prepared
according to the reported methods.34-36
Preparation of 4-Hydroxycoumarins 2a-c. In a 250 mL two-
neck round-bottom flask, 36 mmol of o-hydroxyacetophenone 1
was added to 6.3 g of NaH (180 mmol, 70%, dispersion in mineral
oil) in anhydrous toluene (100 mL). After the evolution of hydrogen
had ceased, 6 mL of diethyl carbonate in anhydrous toluene (20
mL) was added dropwise during 30 min, and the mixture was
refluxed with stirring for 3 h. After the mixture cooled to room
temperature, the precipitate was filtered off and washed with 20
mL of toluene. The combined toluene solutions were concentrated
on a rotary evaporator, and the residue was poured into 100 mL of
water and acidified with 2 N hydrochloric acid. The resulting
precipitate was filtered off and recrystallized from ethanol to give
a white solid. Yields for 2a, 2b, and 2c are 85%, 72%, and 75%,
respectively.
Preparation of 3-(2-Hydroxy-5-substituted phenyl)-1-methyl-
1H-pyrazol-5(4H)-ones 3a-c. In a 100 mL three-neck round-
bottom flask, 6 mmol of 6-substituted 4-hydroxycoumarins 2, 5.6 g
(12 mmol) of methylhydrazine, and 40 mL of ethanol were added
gradually. The reaction was refluxed under N2 flow until the material
was completely transformed by TLC. The resulting mixture was
concentrated on a rotary evaporator to give the crude products. The
white crystal 3 could be obtained by recrystallization of the crude
products from ethanol. Yields for 3a, 3b, and 3c are 81%, 92%,
and 90%, respectively.
General Procedure for the Synthesis of Target Compounds
5a-d. A mixture of 5.0 mmol of 3-(2-hydroxy-5-substituted
phenyl)-1-methyl-1H-pyrazol-5(4H)-one 3 and 0.82 g (6.0 mmol)
of anhydrous K2CO3 in dry acetone (20 mL) was stirred and refluxed
for 0.5 h, followed by the addition of 5.0 mmol of intermediate 4.
The reaction was continued for 5-8 h under reflux. The resulting
mixture was cooled to room temperature and filtered off by suction,
and the solvent was evaporated to give the crude product, followed
by chromatography purification on silica using a mixture of
petroleum ether and acetone (12:1) as eluent to give the target
compounds 5a-d in yields of 67-84%.
In the MM-PBSA calculations, the free energy of inhibitor
binding, ∆Gbind, was calculated from the difference between the
free energies of the complex (Gcomplex) and the free cyt b (Gcytb
)
and free inhibitor (Ginhibitor) as
∆Gbind ) Gcomplex - (Gcytb + Ginhibitor
)
(1)
∆Gbind was evaluated as a sum of the changes in the MM gas-
phase binding energy (∆Ebind), solvation free energy (∆Gsolv), and
entropy contribution (-T∆S):
∆Gbind ) ∆Ebind - T∆S
∆Ebind ) ∆EMM + ∆Gsolv
∆Gsolv ) ∆GPB + ∆Gnp
∆Gnp ) γSASA
(2)
(3)
(4)
(5)
Data for 5a: yield, 74%; mp 129-131 °C; 1H NMR (400 MHz,
CDCl3) δ 3.69 (s, 3H, NCH3), 3.73 (s, 3H, COOCH3), 3.85 (s, 3H,
(27) Ji, F. Q.; Niu, C. W.; Chen, C. N.; Chen, Q.; Yang, G. F.; Xi, Z.;
Zhan, C. G. ChemMedChem 2008, 3, 1203–1206.
(28) Pan, Y. M.; Gao, D. Q.; Yang, W. C.; Cho, H.; Yang, G. F.; Tai,
H. H.; Zhan, C. G. Proc. Natl. Acad. Sci. U.S.A. 2005, 102, 16656–
16661.
The electrostatic solvation free energy was calculated by the finite-
difference solution to the Possion-Boltzmann (PB) equation (∆GPB)
as implemented in the Delphi program.25,26 The dielectric constants
used were 1 for the solute and 78.5 for the solvent. The SASA
was calculated by the default surface area calculation program in
the MM-PBSA module of the Amber8 program with the default γ
) 0.0072 kcal Å-2. The entropy contribution, -T∆S, to the binding
(29) Gao, D. Q.; Cho, H.; Yang, W. C.; Pan, Y. M.; Yang, G. F.; Tai,
H. H.; Zhan, C. G. Angew. Chem., Int. Ed. 2006, 45, 653–657.
(30) Pan, Y. M.; Gao, D. Q.; Yang, W. C.; Cho, H.; Zhan, C. G. J. Am.
Chem. Soc. 2007, 129, 13537–13543.
(31) Zhan, C. G.; Zheng, F.; Landry, D. W. J. Am. Chem. Soc. 2003, 125,
2462–2474.
(32) Xiong, Y.; Li, Y.; He, H. W.; Zhan, C. G. Bioorg. Med. Chem. Lett.
2007, 17, 5186–5190.
(24) Kollman, P.; Massova, I.; Reyes, C.; Kuhn, B.; Huo, S.; Chong, L.;
Lee, M.; Duan, Y.; Wang, W.; Donini, O.; Cieplak, P.; Srnivasan, J.;
Case, D.; Cheatham, T. E. Acc. Chem. Res. 2000, 33, 889–897.
(25) Jayaram, B.; Sharp, K. A.; Honig, B. Biopolymers 1989, 28, 975–
993.
(26) Sanner, M. F.; Olson, A. J.; Spehner, J. C. Biopolymers 1996, 38,
305–320.
(33) Pan, Y. M.; Gao, D. Q.; Zhan, C. G. J. Am. Chem. Soc. 2008, 130,
5140–5149.
(34) Desai, N. J.; Suresh, S. J. Org. Chem. 1957, 22, 388–390.
(35) Froggett, J. A.; Hockley, M. H.; Titman, R. B. J. Chem. Res. (S) 1997,
30–31.
(36) Kim, B. T.; Park, N. K.; Choi, G. J.; Kim, J. C.; Pak, C. S. Patent
WO 0112585, 2001; Chem. Abstr. 2001134193207.
9
188 J. AM. CHEM. SOC. VOL. 132, NO. 1, 2010

Subnanomolar Inhibitor of Cytochrome bc1
A R T I C L E S
dCH-OCH3), 5.07 (s, 2H, CH2), 5.82 (s, 1H, HetH), 6.87 (t, J )
8.0 Hz, 1H, ArH), 6.99 (d, J ) 8.0 Hz, 1H, ArH), 7.15-7.23 (m,
3H, ArH), 7.38-7.43 (m, 2H, ArH), 7.53 (dd, J ) 3.6 Hz, J ) 4.8
Hz, 1H, ArH), 7.63 (s, 1H, dCH-OCH3); EI-MS m/z (relative
intensity) 395 ([M + 1]+, 17), 394 (M+, 100), 362 (19), 330 (14),
205 (60), 204 (42), 188 (13), 145 (83), 131 (23), 117 (38), 115
(25), 102 (51). Anal. Calcd for C22H22N2O5: C, 66.99; H, 5.62; N,
7.10. Found: C, 67.19; H, 5.38; N, 7.21.
7.60 (s, 1H, dCH-OCH3), 7.71 (s, 1H, ArH); EI-MS m/z (relative
intensity) 423 ([M + 1]+, 8), 422 (M+, 9), 217 (13), 206 (18), 205
(91), 173 (14), 146 (13), 145 (100), 131 (19), 130 (20), 116 (24),
103 (12). Anal. Calcd for C24H26N2O5: C, 68.23; H, 6.20; N, 6.63.
Found: C, 68.43; H, 5.99; N, 6.65.
1
Data for 6c: yield, 82%; mp 83-85 °C; H NMR (400 MHz,
CDCl3) δ 3.65 (s, 3H, NCH3), 3.67 (s, 3H, COOCH3), 3.70 (s, 3H,
dCH-OCH3), 4.91 (s, 2H, Ph-OCH2), 5.15 (s, 2H, Het-OCH2),
6.13 (s, 1H, HetH), 6.99-7.04 (m, 2H, ArH), 7.18 (dd, J ) 4.0
Hz, J ) 5.2 Hz, 1H, ArH), 7.31-7.39 (m, 6H, ArH), 7.43-7.48
(m, 3H, ArH), 7.51 (s, 1H, dCH-OCH3), 7.94 (d, J ) 7.6 Hz,
1H, ArH); EI-MS m/z (relative intensity) 484 (M+, 9), 280 (13),
204 (100), 177 (21), 174 (14), 173 (23), 144 (98), 131 (32), 117
(16), 114 (54), 103 (37), 101 (35). Anal. Calcd for C29H28N2O5: C,
71.88; H, 5.82; N, 5.78. Found: C,71.72; H, 5.62; N, 5.62.
Data for 5b: yield, 84%; mp 143-145 °C; 1H NMR (400 MHz,
CDCl3) δ 2.29 (s, 3H, Ar-CH3), 3.68 (s, 3H, NCH3), 3.70 (s, 3H,
COOCH3), 3.84 (s, 3H, dCH-OCH3), 5.06 (s, 2H, CH2), 5.81 (s,
1H, HetH), 6.88 (d, J ) 8.0 Hz, 1H, ArH), 6.98 (d, J ) 8.4 Hz,
1H, ArH), 7.20-7.24 (m, 2H, ArH), 7.37-7.41 (m, 2H, ArH), 7.52
(dd, J ) 3.2 Hz, J ) 4.8 Hz, 1H, ArH), 7.62 (s, 1H, dCH-OCH3);
EI-MS m/z (relative intensity) 409 ([M + 1]+, 6), 408 (M+, 44),
205 (23), 204 (24), 175 (10), 146 (11), 145 (100), 131 (28), 103
(34). Anal. Calcd for C23H24N2O5: C, 67.63; H, 5.92; N, 6.86.
Found: C, 67.36; H, 5.83; N, 6.90.
1
Data for 6d: yield, 81%; mp 76-77 °C; H NMR (400 MHz,
CDCl3) δ 3.61 (s, 3H, NCH3), 3.75 (s, 3H, COOCH3), 3.87 (s, 3H,
dCH-OCH3), 5.39 (s, 2H, CH2), 5.76 (s, 1H, HetH), 7.11-7.19
(m, 2H, ArH), 7.28-7.33 (m, 3H, ArH), 7.34-7.40 (m, 3H, ArH),
7.61 (d, J ) 7.2 Hz, 1H, ArH), 7.69 (s, 1H, dCH-OCH3), 7.87
(d, J ) 7.8 Hz, 1H, ArH), 7.99 (d, J ) 7.8 Hz, 1H, ArH), 8.23 (t,
J ) 7.2 Hz, 1H, ArH); EI-MS m/z (relative intensity) 517 ([M +
1]+, 6), 516 (M+, 8), 484 (20), 456 (24), 147 (14), 145 (43), 144
(23), 131 (17), 123 (100), 115 (22), 103 (23), 102 (14). Anal. Calcd
for C29H25FN2O6: C, 67.43; H, 4.88; N, 5.42. Found: C, 67.15; H,
4.69; N, 5.12.
Data for 5c: yield, 81%; mp 189-191 °C;1H NMR (600 MHz,
CDCl3) δ 3.68 (s, 3H, NCH3), 3.73 (s, 3H, COOCH3), 3.88 (s, 3H,
dCH-OCH3), 5.06 (s, 2H, CH2), 5.73 (s, 1H, HetH), 6.88 (d, J )
8.8 Hz, 1H, ArH), 7.08 (dd, J ) 1.4 Hz, J ) 9.0 Hz, 1H, ArH),
7.21 (dd, J ) 3.6 Hz,J ) 5.2 Hz, 1H, ArH), 7.33 (d, J ) 2.4 Hz,
1H, ArH), 7.37 (dd, J ) 3.2 Hz, J ) 5.6 Hz, 2H, ArH), 7.50 (dd,
J ) 3.6 Hz, J ) 5.2 Hz, 1H, ArH), 7.64 (s, 1H, dCH-OCH3);
EI-MS m/z(relative intensity) 428 (M+, 2), 204 (36), 189 (14), 160
(32), 144 (100), 128 (29), 115 (26), 101 (29), 91 (34). Anal. Calcd
for C22H21ClN2O5: C, 61.61; H, 4.94; N, 6.53. Found: C, 61.35; H,
5.15; N, 6.59.
X-ray Diffraction. A colorless block of 5b (0.20 mm × 0.20
mm × 0.10 mm) was mounted on a quartz fiber. Cell dimensions
and intensities were measured at 299 K on a Bruker SMART CCD
area detector diffractometer with graphite-monochromated Mo KR
radiation (λ ) 0.71073 Å); θmax ) 22.13; 6204 measured reflections;
3658 independent reflections (Rint ) 0.0229), of which 2701 had
|Fo| > 2|Fo|. Data were corrected for Lorentz and polarization effects
and for absorption (Tmin ) 0.9819; Tmax ) 0.9909). The structure
was solved by direct methods using SHELXS-97;37 all other
calculations were performed with Bruker SAINT System and Bruker
SMART programs.38 Full-matrix least-squares refinement based on
F2 using the weight of 1/[σ2(Fo 2) + (0.0934P)2 + 0.1767P] gave
final values of R ) 0.0570, ωR ) 0.1576, and GOF(F) ) 1.055
for 289 variables and 1634 contributing reflections. Maximum shift/
error ) 0.000(3), max/min residual electron density ) 0.209/-0.169
e Å-3. Hydrogen atoms were observed and refined with a fixed
value of their isotropic displacement parameter.
Data for 5d: yield, 67%; mp 150-152 °C;1H NMR (600 MHz,
CDCl3) δ 3.66 (s, 3H, NCH3), 3.86 (s, 3H, COOCH3), 4.08 (s, 3H,
dN-OCH3), 5.07 (s, 2H, CH2), 5.82 (s, 1H, HetH), 6.86 (t, J )
4.4 Hz, 1H, ArH), 6.98 (d,J ) 7.8 Hz, 1H, ArH), 7.15-7.25 (m,
2H, ArH), 7.40-7.54 (m, 4H, ArH); EI-MS m/z (relative intensity)
395 (M+, 42), 206 (10), 205 (20), 189 (28), 160 (41), 146
(31),131(96), 118 (37), 116 (100), 115 (19), 89 (27). Anal. Calcd
for C21H21N3O5: C, 63.79; H, 5.35; N, 10.63. Found: C, 63.52; H,
5.58; N, 10.44.
General Procedure for the Synthesis of Target Compounds
6. A reaction system with 1.0 mmol of compounds 5a,b and 0.16 g
(1.2 mmol) of anhydrous K2CO3 or 0.12 g (1.2 mmol) of
triethylamine in dry N,N-dimethylformamide (DMF) (6 mL) was
stirred at room temperature for 0.5 h, followed by the addition of
1.1 mmol of halogenated compounds or substituted benzoyl
chloride. The reaction was continued for 2-24 h. The resulting
mixture was poured into 20 mL of water, extracted with ethyl
acetate (10 mL × 3), washed with saturated sodium chloride (10
mL × 3), dried with anhydrous magnesium sulfate, and filtered
off with suction. The solvent was evaporated to give the crude,
which was purified by chromatography on silica using a mixture
of petroleum ether and acetone (20:1) as eluent to give the target
compounds 6a-d in yields of 46-82%.
3. Inhibitory Kinetics of Porcine bc1 Complex. The preparation
of succinate-cytochrome c reductase (SCR, mixture of respiratory
complex II and bc1 complex) from porcine heart was essentially as
reported.39 The activity of SCR in catalyzing the oxidation of
succinate by cytochrome c was determined in 1.8 mL of reaction
mixture containing 100 mM PBS (pH 7.4), 0.3 mM EDTA, 20 mM
succinate, 60 µM oxidized cytochrome c, and an appropriate amount
of SCR at 23 °C. The enzymatic assay for succinate-ubiquinone
reductase (complex II)-catalyzed oxidation of succinate by 2,6-
dichloroindophenol (DCIP) was carried out in essentially the same
reaction mixture, except for the replacement of the oxidized
cytochrome c with 53 µM DCIP. The reaction was initiated by the
addition of enzyme and monitored continuously by following the
absorbance change at certain wavelengths on a Perkin-Elmer
Lambda 45 spectrophotometer equipped with a magnetic stirrer.
Data for 6a: yield, 78%; mp 117-119 °C; 1H NMR (400 MHz,
CDCl3) δ 3.68 (s, 3H, NCH3), 3.70 (s, 3H, COOCH3), 3.73 (s, 3H,
dCH-OCH3), 3.86 (s, 3H, Ar-OCH3), 5.06 (s, 2H, CH2), 5.81
(s, 1H, HetH), 6.85 (t, J ) 8.0 Hz, 1H, ArH), 6.98 (d, J ) 7.2 Hz,
1H, ArH), 7.07 (dd, J ) 3.6 Hz, J ) 5.6 Hz, 1H, ArH), 7.16-7.24
(m, 2H, ArH), 7.38-7.42 (m, 2H, ArH), 7.54 (dd, J ) 3.6 Hz, J )
5.6 Hz, 1H, ArH), 7.64 (s, 1H, dCH-OCH3); EI-MS m/z (relative
intensity) 410 ([M + 1]+, 3), 408 (M+, 2), 394 (100), 378 (10),
376 (31), 375 (83), 361 (17), 318 (13), 161 (16), 145 (26), 118
(21), 115 (14), 102 (29). Anal. Calcd for C23H24N2O5: C, 67.63; H,
5.92; N, 6.86. Found: C, 67.83; H, 6.01; N, 6.76.
The extinction coefficients used were 18.5 mM-1 cm-1 for A
550
red-ox
600
for cytochrome c reduction and 21 mM-1 cm-1 for A
for DCIP
red-ox
reduction. For the inhibitory kinetic studies, the reaction was carried
out in the presence of varying concentrations of the inhibitor.
The ubiquinol-cytochrome c reductase (bc1 complex) activity
in catalyzing the oxidation of DBH2 by cytochrome c was assayed
in 100 mM PBS (pH 6.5), 2 mM EDTA, 750 µM lauryl maltoside
(n-dodecyl-ꢀ-D-maltoside), 20-115 µM DBH2, 100 µM oxidized
Data for 6b: yield, 46%; mp 152-153 °C; 1H NMR (400 MHz,
CDCl3) δ 2.32 (s, 3H, Ar-CH3), 3.70 (s, 3H, NCH3), 3.77 (s, 3H,
COOCH3), 3.84 (s, 3H, dCH-OCH3), 3.87 (s, 3H, Ar-OCH3),
5.06 (s, 2H, CH2), 6.08 (s, 1H, HetH), 6.85 (d, J ) 8.4 Hz, 1H,
ArH), 7.09 (t, J ) 7.2 Hz, 1H, ArH), 7.21 (d, J ) 7.8 Hz, 1H,
ArH), 7.36-7.40 (m, 2H, ArH), 7.53 (t, J ) 7.2 Hz, 1H, ArH),
(37) Sheldrick, G. M. SHELXTL, Version 5.0; University of Go¨ttingen:
Go¨ttingen, Germany, 2001.
9
J. AM. CHEM. SOC. VOL. 132, NO. 1, 2010 189

A R T I C L E S
Zhao et al.
Scheme 3
k+0Km
Competitive:
A )
,
B ) k-0
(7)
(8)
Km + [S]
Noncompetitive:
A ) k+0, B ) k-0
k+′ 0[S]
Km + [S]
k-′ 0[S]
*
Km + [S]
Uncompetitive:
A )
,
B )
(9)
Results and Discussion
Homology Modeling and Computational Prediction. Because
the Ki values of all compounds were tested on the basis of the
porcine bc1 complex in this work, we first used the crystal
structure of bovine heart cytochrome bc1 in complex with AZ
(PDB entry 1SQB at resolution of 2.69 Å) as a template to
establish the three-dimensional (3D) structural model of the
porcine bc1 complex. The sequence alignment between the
porcine cyt b and the same subunit of the template indicates
that only 38 out of 379 residues were found to be different from
each other, showing very high homology of the porcine cyt b
with the template. The 3D structural models for porcine bc1 in
complex with inhibitors were simulated by integrating homology
modeling, molecular docking, and energy minimizations as
described in the Materials and Methods. As shown in Figure
1A, the porcine bc1 complex is a dimer and has C2 symmetry.
Each monomer contains one cyt b subunit in which the Qo site
is situated. This Qo site is located near the intermembrane space
and partially interacts with the bL heme group. The (E)ꢀ-
methoxy methyl acrylate group and the phenyl group of AZ
are orientated at the deep bottom of the binding pocket. The
(E)ꢀ-methoxy methyl acrylate group interacts with residues
Phe128, Tyr131, Val132, Gly142, and Ala143 from helix C and
helix cd1, and residues Glu271 and Tyr 273 from the ef loop.
This headgroup of the inhibitor blocks the way to the bL heme
group, with the shortest distances around 6 Å, while the distance
from the carbonyl oxygen atom of this headgroup of the inhibitor
to the heme iron is about 10.5 Å. The carbonyl oxygen atom of
this headgroup also forms a hydrogen bond with the backbone
-NH group of residue Glu271 from the ef loop, further
stabilizing the orientation of the headgroup of the inhibitor inside
the Qo binding pocket. The phenyl group of the inhibitor packs
residues Gly142, Val145, Ile146, Pro270, Ile268, and Tyr278
through extensive hydrophobic interactions. Compared with the
conformation of AZ in the crystal structure (blue in Figure 1B),
the rmsd for its docked conformation (red in Figure 1B) is only
0.77 Å. In addition, the calculated binding free energies for KM,
AZ, and PY are in good agreement with the corresponding
experimental data (Table 1, see below), indicating that our
computational protocol is reliable.
The above computational protocol was then used to predict
the potency of the designed compounds. The calculated binding
free energies (∆Gcalcd), summarized in Table 1, indicate that
compounds 5a-d should have a significantly higher inhibitory
potency than AZ and KM. In addition, compounds 5a, 5b, and
5d should have an inhibitory potency similar to that of PY,
whereas compound 5c should have an inhibitory potency higher
than that of PY. An ideal computational protocol can be used
to predict not only compounds with a higher activity but also
compounds with a lower activity. Therefore, in order to validate
the predictability of the computational protocol, some hydroxy-
protected compounds 6a-d were also predicted to have a lower
potency due to the lack of the intramolecular hydrogen bond.
Table 1 also summarizes the Ki values against the porcine
bc1 complex for the inhibition by the synthesized compounds
and also by the commercial compounds that were used as the
cytochrome c, and an appropriate amount of SCR at 23 °C. The
preparation of DBH2 from DB was carried out according to the
procedure described in previous publications,40,41 and the concen-
tration of DBH2 was determined by measuring the absorbance
difference between 288 and 320 nm using an extinction coefficient
of 4.14 mM-1cm-1 for the calculation.42,43 The nonionic detergent
lauryl maltoside was used to decrease the interfering nonenzymatic
activity,42-45 though it was expected to affect the Km value of DBH2
to bc1 complex.45 For each reaction, the nonenzymatic rate for
cytochrome c reduction was followed for at least 100 s before
enzyme was added to initiate the reaction.
The reaction mechanism could be considered as shown in Scheme
3, where E, S, and I represent enzyme, substrate, and inhibitor,
respectively. According to the substrate reaction kinetic theory, the
concentration of product formed at time t is calculated by46
V0 - Vs
kobs
obst
[P] ) Vst +
(1 - e-k
)
(6)
where V0 and Vs are the initial and steady-state velocities of the
reaction in the presence of inhibitor, and kobs is the observed first-
order rate constant:
kobs ) A[I]0 + B
k+0Km + k+′ 0[S]
A )
Km + [S]
*
k-0Km + k-′ 0[S]
B )
*
Km + [S]
We have Km and K* as the Michaelis-Menten constants:
m
k-1 + k2
Km
)
k1
*
k-1
*
Km
)
*
k1
The slow, tight-binding inhibitors can also be classified as
competitive, noncompetitive, and uncompetitive on the basis of
similar considerations, as in the case of the classical inhibitors.
Experimentally, the type of inhibition can be ascertained by
studying the effect of [S] on the apparent rate constants A
and B:
(38) SMART V5.628, SAINT V6.45, and SADABS; Bruker AXS Inc.:
Madison, WI, 2001.
(39) Yu, L.; Yu, C. A. J. Biol. Chem. 1982, 257, 2016–2021.
(40) Rieske, J. S. Methods Enzymol. 1967, 10, 239–245.
(41) Luo, C.; Long, J.; Liu, J. Clin. Chim. Acta 2008, 395, 38–41.
(42) Fisher, N.; Bourges, I.; Hill, P.; Brasseur, G.; Meunier, B. Eur.
J. Biochem. 2004, 271, 1292–1298.
(43) Fisher, N.; Brown, A. C.; Sexton, G.; Cook, A.; Windass, J.; Meunier,
B. Eur. J. Biochem. 2004, 271, 2264–2271.
(44) Fisher, N.; Meunier, B. Pest Manag. Sci. 2005, 61, 973–978.
(45) Chretien, D.; Slama, A.; Brie`re, J.; Munnich, A.; Ro¨tig, A.; Rustin,
P. Curr. Med. Chem. 2004, 11, 233–239.
(46) Tsou, C. L. AdV. Enzymol. 1988, 61, 381–436.
9
190 J. AM. CHEM. SOC. VOL. 132, NO. 1, 2010

Subnanomolar Inhibitor of Cytochrome bc1
A R T I C L E S
Figure 1. (A) Modeled 3D structure of porcine bc1 complex with bound inhibitor. (B) Superimposition of the conformations of AZ in the crystal (blue) and
modeled (red) structures.
Table 1. Ki Values of Compounds 5, 6, KM, AZ, and PY against Porcine bc1 Complex
∆G (kcal/mol)
redisue interaction energies (kcal/mol)
compd
R
R′
X
Ki (nM)
calcda
exptlb
F274
P270
5a
5b
5c
5d
6a
6b
6c
6d
AZ
KM
PY
H
H
H
H
H
CH
CH
CH
N
CH
CH
CH
CH
1.20
3.10
0.57
-11.11
-11.40
-12.50
-10.17
-9.08
-9.59
-9.95
-9.09
-8.57
-8.90
-10.95
-12.19
-11.62
-12.63
-11.50
-10.29
-10.52
-10.79
-9.56
-10.55
-10.24
-10.28
-10.49
-10.30
-11.46
-8.17
-4.97
-4.91
-4.86
-4.88
-4.95
-5.06
-4.70
-4.95
-4.53
-4.54
-6.23
CH3
Cl
H
H
CH3
H
3.80
CH3
CH3
C6H5CH2
29.10
19.90
12.70
97.70
297.60
159.60
3.30
H
p-FC6H4CO
-8.57
-8.92
-9.48
-9.28
-11.59
-9.13
-7.92
a ∆Gcalcd was calculated according to our modified MM-PBSA method. b ∆Gexptl ) -RT ln Ki.
fold improved binding affinity compared to AZ, KM, and PY,
respectively.
Structural Basis of the Higher Binding Affinity of Com-
pounds 5a-d. As shown in Figure 3A,B, similar to that of AZ
(Figure 3B), the methoxy methyl acrylate group of compounds
5a-5d form a hydrogen bond with the backbone amide group
of Glu271 (Figure 3A,B). However, compared to that of AZ
(Figure 3C), the π-π interactions between Phe274 and the
pyrazole ring of compound 5 are improved greatly because the
pyrazole ring and the 2-hydroxyphenyl are totally coplanar due
to the intramolecular hydrogen bond between the hydroxyl group
and the N(1) atom of pyrazole moiety (Figure 3D). In other
words, compared with the conformation in the AZ-complexed
bc1, the phenyl side chain of Phe274 undergoes a rotation of
∼4° in order to enhance the π-π interactions with compound
5c. Figure 3E shows the results of the superimposition of
compounds 5a-d based on the binding pocket. As a result, with
about 522-, 280-, and 5.8-fold improvement of the binding
affinity compared respectively with AZ, KM, and PY, compound
5c is clearly the most potent inhibitor. In contrast, the hydroxy-
protected derivatives 6a-d all have a lower binding affinity
due to the lack of such an intramolecular hydrogen bond.
Figure 2. Correlation between the calculated and experimental binding
free energies of compounds 5, 6, KM, AZ, and PY.
control. Interestingly, the calculated binding free energies
correlated reasonably well with the binding free energies derived
from the corresponding experimental Ki values, with a correla-
tion coefficient of r2 ) 0.89 (Figure 2), further confirming the
reliability of our computational models. As shown in Table 1,
compound 5c distinguishes itself as the most promising candi-
date with the highest potency, with about 522-, 280-, and 5.8-
9
J. AM. CHEM. SOC. VOL. 132, NO. 1, 2010 191

A R T I C L E S
Zhao et al.
Figure 3. Comparison of the binding modes of compound 5c and AZ. (A) Simulated binding model of bc1 in complex with 5c. (B) Simulated binding
model of bc1 in complex with AZ. (C) π-π stacking interaction between the side chain of F274 and the pyrimidyl ring of AZ (the interaction energy
between F274 and AZ is -9.48 kcal/mol). (D) π-π stacking interaction between the side chain of F274 and the enlarged aromatic system of 5c (the
interaction energy between F274 and 5c is -10.28 kcal/mol). (E) Superimposition of the binding conformation of compounds 5a (green), 5b (white), 5c
(magenta), and 5d (blue).
Scheme 4
Synthesis of the Designed Compounds. The synthetic route
for the designed compounds 5 and 6 is shown in Scheme 4.
4-Hydroxy-2H-chromen-2-one (also named 4-hydroxycoumarin)
compounds 2 were prepared easily from 2′-hydroxyacetophe-
nones 1a-c as starting material in good overall yields (85%
for 2a, 72% for 2b, and 75% for 2c). Compounds 2 were then
reacted with methylhydrazine to afford the key intermediates
3a-c, which are 3-(2-hydroxyphenyl)-1-methyl-1H-pyrazol-
5(4H)-ones. It should be noted that compounds 3 underwent
tautomerization under basic conditions to produce the enol
isomers. Fortunately, maybe due to the existence of an intramo-
lecular hydrogen bond, the resulted enol-hydroxy was apparently
more active than the phenol-hydroxy group and reacted with
(E)-methyl 2-(2-bromomethyl-phenyl)-3-methoxyacrylates 4
under basic conditions to produce the target compounds 5 in
yields of 67-84%. In addition, in order to examine the
importance of the intramolecular hydrogen bond, we also
synthesized some hydroxy-protected compounds 6a-d by the
reactions of compounds 5a and 5b with various alkylation or
acetylation reagents. The structures of compounds 2-6 were
1
characterized by H NMR, EI-MS spectrum, and elemental
analyses. In addition, the crystal structure (CCDC 709734) of
9
192 J. AM. CHEM. SOC. VOL. 132, NO. 1, 2010

Subnanomolar Inhibitor of Cytochrome bc1
A R T I C L E S
Figure 4. Crystal structure of compound 5b. The dotted line shows the
intramolecular H-bond.
5b was determined by X-ray diffraction analyses. As shown in
Figure 4, the ꢀ-methoxyacrylate group adopts E-configuration.
Most interestingly, an intramolecular hydrogen bond between
the hydroxyl group and the N(1) atom was formed, as expected.
As a result, the pyrazole ring and the 2-hydroxyphenyl planes
are coplanar to form a bigger aromatic system.
Figure 5. (A) Inhibitory kinetics of bc1 complex by compound 5c. Each
reaction mixture contains 100 mM PBS (pH 7.4), 0.3 mM EDTA, 20 mM
succinate, 60 µM cytochrome c, 0.1 nM SCR, and a certain amount of
compound 5c (1, 0 nM; 2, 1 nM; 3, 1.5 nM; 4, 2 nM; and 5, 5 nM).
Experimental data are shown as colored dots and theoretical values as black
solid lines. Inset: Plot of kobs against concentration of compound 5c. (B) Effect
of cytochrome c concentration on the inhibition of bc1 complex by compound
5c. The assays were carried out in the presence of 5 nM compound 5c and
various concentrations of cytochrome c (1, 4 µM; 2, 12 µM; and 3, 48 µM).
Each reaction was initiated by adding 0.1 nM SCR. Inset: Plot of kobs against
concentration of cytochrome c.
Enzymatic Kinetics. In order to understand the action
mechanism, we further studied the inhibitory kinetics of
succinate-cytochrome c reductase (SCR, mixture of respira-
tory complex II and bc1 complex) with compounds 5a-d.
Complex II first passes electrons from succinate to ubiquino-
ne, and then the cytochrome bc1 complex passes electrons
from reduced ubiquinone to cytochrome c. We measured the
succinate-ubiquinone reductase (complex II) activity of SCR
using succinate and DCIP as substrates and the succinate-
cytochrome c reductase (both complex II and bc1 complex)
activity of SCR using succinate and cytochrome c as
substrates. All of the compounds exhibited no effect on the
activity of complex II (data not shown), whereas they
inhibited the activity of succinate-cytochrome c reductase
considerably (Figure 5). Thus, the inhibition of these
compounds on the activity of SCR may be attributed to their
effects on the function of bc1 complex, and the assay for the
SCR-catalyzed oxidation of succinate by cytochrome c was
used to elucidate the inhibitory kinetics of the bc1 complex.
As is the case for many slow-binding inhibitors, the progress
curves in the presence of compounds 5a-d did not display
a simple linear product versus time relationship, as seen for
classical reversible inhibitors like AZ and KM. Rather, the
product formation versus time shows a curvilinear function.
For example, Figure 5A shows the time courses of the bc1
complex-catalyzed reaction in the presence of different
concentrations of compound 5c. The progress curves start
off linear (the initial-rate phase) but fall off with increasing
time. When time (t) approaches infinity, the concentration
of product ([P]) approaches an asymptote with a positive
slope, which decreases with increasing concentration of
compound 5c. The initial-rate phase reflects the slow onset
of inhibition by these compounds, since it can be simply
eliminated by preincubating the inhibitor with the enzyme
(data not shown). Under the conditions employed in the
present study, the time-dependent behavior of the enzyme-
catalyzed reaction can be well described according to the
substrate reaction kinetic theory.46 The best-fit curves in
Figure 5A were obtained by nonlinear regression analysis,
from which the kinetic parameters, the initial (V0) and steady-
state (Vs) velocities of the reaction in the presence of inhibitor,
and the observed first-order rate constant (kobs) can be
determined. The inset of Figure 5A shows the plot of kobs
against the concentration of compound 5c ([5c]). Noticeably,
kobs is proportional to the inhibitor concentration. From the
slope and intercept of this plot, the apparent rate constants
A ) 0.00631 ( 0.00036 s-1 nM-1 and B ) 0.00358 (
0.00110 s-1 were then determined.
Subsequently, we monitored the time courses of the bc1
complex-catalyzed reaction in the presence of different con-
centrations of cytochrome c and at a fixed concentration of
compound 5c (Figure 5B). Using the procedures described
above, the values of kobs can be obtained by fitting the progress
curves according to the substrate reaction kinetic theory and
are plotted against the concentration of cytochrome c in the inset
of Figure 5B. The data clearly revealed that kobs was independent
of the concentration of substrate cytochrome c under the
experimental conditions used, indicating that compound 5c is a
noncompetitive inhibitor with respect to cytochrome c. Thus,
the values of A and B determined from Figure 5A are equal to
the true association and dissociation rate constants k+0 and k-0
,
respectively, and the inhibition constant for compound 5c can
be calculated as Ki ) k-0/k+0 ) 0.57 ( 0.21 nM.
To study the effect of substrate ubiquinol on the inhibition
kinetics of compound 5c, we directly measured the ubiquinol-
cytochrome c reductase (bc1 complex) activity of SCR using
decylubiquinol (DBH2) and cytochrome c as substrates in the
absence and in the presence of compound 5c. This reaction
exhibited Michaelis-Menten kinetics toward substrate DBH2, and
the kinetic parameters were determined (Km ) 58 ( 4 µM and
kcat ) 238 ( 8 s-1). In the presence of compound 5c, with succinate
used as substrate, the progress curves exhibited a similar curvilinear
9
J. AM. CHEM. SOC. VOL. 132, NO. 1, 2010 193

A R T I C L E S
Zhao et al.
Figure 6. (A,B) Inhibitory kinetics of porcine bc1 complex by compound 5c. Each reaction mixture contains 100 mM PBS (pH 6.5), 2 mM EDTA, 750 µM
lauryl maltoside, 100 µM oxidized cytochrome c, 0.05 nM SCR, a certain amount of DBH2 (A, 20 µM; B, 115 µM), and compound 5c (1, 7.5 nM; 2, 15
nM; and 3, 25 nM). Experimental data are shown as colored dots and theoretical values as black solid lines. Insets: Plots of kobs against concentration of
compound 5c. (C) Plot of the apparent rate constant A against concentration of DBH2. Inset: Plot of 1/A against concentration of DBH2. (D) Plot of the
apparent rate constant B against concentration of DBH2.
pattern due to the slow onset of inhibition (Figure 6). However,
when the experiments were performed at a fixed concentration of
compound 5c, the kobs value decreased with increasing concentration
of DBH2 (Supporting Information, Figure 2s), which is different
from what is observed in Figure 5B. To further elucidate the
inhibitory mechanism, we carried out six sets of inhibitory
experiments in the presence of varying concentrations of DBH2
and compound 5c (Figure 6A,B). Similarly, from these progress
curves of product formation at fixed concentrations of DBH2, the
values of V0, Vs, and kobs can be determined as described previously.
Plots of kobs against the inhibitor concentration give a straight line
(insets of Figure 6A,B), from which the values of A (slope) and B
(intercept) at indicated DBH2 concentrations were obtained. Figure
6C,D shows the dependence of the A and B values on the DBH2
concentration. Clearly, A deceased with increasing the concentration
of DBH2, while B was not affected, suggesting that compound 5c
is a competitive inhibitor with respect to substrate ubiquinol. Thus,
with the Km value of 58 µM, the microscopic inhibition rate
constants k+0 ) 0.00185 ( 0.00003 s-1 nM-1 and k-0 ) 0.00480
( 0.00018 s-1 can be obtained by fitting eq 2 to the experimental
data, and the inhibition constant Ki ) k-0/k+0 ) 2.6 ( 0.1 nM can
be derived. The Ki value determined is about 5-fold higher than
that measured from the succinate-cytochrome c reductase activity
inhibition, which is mainly due to the presence of lauryl maltoside.
This nonionic detergent was reported to affect the Km value of
DBH2 to bc1 complex,45 and it possibly has an impact on the
microscopic inhibition rate constants and the Ki value. Indeed, we
found that the kobs value decreased with increasing concentration
of lauryl maltoside at fixed DBH2 and compound 5c concentrations
(Supporting Information, Figure 3s).
complex. By introducing an intramolecular hydrogen bond to form
a big aromatic system, the π-π interactions with the phenyl of
Phe274, an important residue with conformational flexibility located
in the binding pocket, were significantly improved. As a result, a
new bc1 inhibitor with subnanomolar Ki potency was discovered.
Compound 5c, with a Ki value of 570 pM, was found to display
about 522-, 280-, and 5.8-fold improved binding affinity compared
to the commercial inhibitors AZ, KM, and PY, respectively. The
further inhibitory kinetics studies showed that compound 5c is a
noncompetitive inhibitor with respect to substrate cytochrome c
but a competitive inhibitor with respect to substrate ubiquinol. Due
to its subnanomolar Ki potency and slow dissociation rate (k-0
)
0.00358 s-1), compound 5c could be used as a specific probe for
further elucidating the mechanism of bc1 function and as a new
lead compound for future drug discovery. In addition, the present
work provides a successful case study concerning how to improve
binding affinity of a ligand with a receptor by optimizing the
interactions of the ligand with conformationally flexible residues
of the receptor.
Acknowledgment. The research was supported in part by the
National Basic Research Program of China (No. 2010CB126103),
the NSFC (No. 20925206, 20932005, and 20872045), and the NIH
(RC1MH088480). The authors also acknowledge Dr. Z. X. Wang
for critical discussions and reading of the manuscript.
Supporting Information Available: Complete refs 8 and 20;
Figure 1s, sequence alignment of the Q0 pocket residue;
Figure 2s, inhibitory kinetics of porcine bc1 complex by
compound 5c; Figure 3s, effect of lauryl maltoside on kobs; and
Figure 4s, enzyme kinetics of AZ. This material is available
free of charge via the Internet at http://pubs.acs.org.
Conclusion
In summary, we have successfully designed and synthesized a
series of pyrazole derivatives as inhibitors of cytochrome bc1
JA905756C
9
194 J. AM. CHEM. SOC. VOL. 132, NO. 1, 2010