Determination of absolute configurations of bedaquiline analogs by quantum chemical electronic circular dichroism calculations and an X-ray diffraction study

Article abstract of DOI:10.1002/ejoc.201400135

Two chiral analogs of bedaquiline were selected from a series of compounds designed as anti-Mycobacterium tuberculosis drugs for synthetic and stereochemical research. The compounds were synthesized from chiral precursors for the first time, and the absolute configurations (ACs) were determined by electronic circular dichroism (ECD) and quantum chemical calculations. Single crystals of both compounds were obtained for X-ray analysis, and the crystal structures and ab initio calculated geometries were extremely similar, which permitted a comparison of the relative reliabilities of ACs obtained by ECD analyses and theoretical simulation. In addition, the in vitro antituberculosis activities of the two compounds were investigated. Two chiral analogs of bedaquiline were designed as anti-Mycobacterium tuberculosis drugs. The synthesis and determination of the absolute configurations (ACs) of the two compounds are described. The ACs were determined by electronic circular dichroism with the aid of theoretical calculations and further validated by an X-ray diffraction study. Copyright

Full text of DOI:10.1002/ejoc.201400135

Job/Unit: O40135
/KAP1
Date: 05-05-14 18:10:45
Pages: 9
FULL PAPER
DOI: 10.1002/ejoc.201400135
Determination of Absolute Configurations of Bedaquiline Analogs by Quantum
Chemical Electronic Circular Dichroism Calculations and an X-ray Diffraction
Study
Zhiqiang Wang,[a] Lixia Zhao,[a] Yu Chen,[b] Wei Xu,[b] and Tiemin Sun*[a]
Keywords: Chirality / Circular dichroism / Density functional calculations / X-ray diffraction / Medicinal chemistry
Two chiral analogs of bedaquiline were selected from a series
of compounds designed as anti-Mycobacterium tuberculosis
drugs for synthetic and stereochemical research. The com-
pounds were synthesized from chiral precursors for the first
time, and the absolute configurations (ACs) were determined
by electronic circular dichroism (ECD) and quantum chemi-
cal calculations. Single crystals of both compounds were ob-
tained for X-ray analysis, and the crystal structures and ab
initio calculated geometries were extremely similar, which
permitted a comparison of the relative reliabilities of ACs ob-
tained by ECD analyses and theoretical simulation. In ad-
dition, the in vitro antituberculosis activities of the two com-
pounds were investigated.
the requirement for high-quality crystals has limited its
range of application. Another commonly used chiroptical
method is electronic circular dichroism (ECD), which pro-
vides information such as the origin of optical activity and
the electronic and geometric structures of chiral mo-
lecules.[3] However, ECD was often used as a relative
method based on some empirical chirality rules, which
sometimes involve exceptions and may lead to incorrect as-
signments.
Recently, the application of ab initio predictions of
chiroptical properties in structural analysis has increased
significantly. Quantum chemical ECD calculations offer the
broadest applicability as they have fewer restrictions regard-
ing the compound under analysis. With the aid of quantum
chemical calculations, one can not only verify the deter-
mined AC of a compound with a single source of chirality
by judging whether the predicted spectrum matches well
with the corresponding experimental spectrum but also
identify the undetermined AC when the calculated spec-
trum for a chosen configuration is accordant with the ex-
perimental spectrum. Moreover, ECD calculations can help
with the assignment of the AC of a compound with mul-
tiple stereogenic centers through comparison of the experi-
mental spectrum with the predicted spectra of all possible
absolute configurations of that compound, and the absolute
configuration used to calculate that satisfactorily predicted
chiroptical spectrum can be assigned to the enantiomer
used for experimental measurements.[4]
Introduction
Drug chirality continues to be an important theme in the
design, development, and marketing of new drugs. Chirality
significantly influences the biological and pharmacological
properties of a drug. The antituberculosis (anti-TB) drug
bedaquiline has two stereogenic centers, which lead to four
stereochemically different structures, and the minimal in-
hibitory concentrations (MICs) for the R/S, S/R, S/S, and
R/R stereoisomers are 0.07, 44.13, 8.80, and 2.78 μg/mL,
respectively.[1] The most active R/S stereoisomer (beda-
quiline) was approved by the US Food and Drug Adminis-
tration (FDA) in 2012 as part of combination therapy for
adults with multi-drug-resistant TB when other alternatives
are not possible.[2]
The structure–activity relationships of chiral molecules
tested in biological systems play a vital part in the drug
development process; therefore, the need to offer reliable
analytical techniques to determine the three-dimensional
structures, including the absolute stereochemistry, has in-
creased. Techniques such as X-ray diffraction (XRD) con-
tinue to play an important role in the determination of the
absolute configurations (ACs) of chiral molecules; however,
[a] Key Laboratory of Structure-Based Drug Design and
Discovery, Shenyang Pharmaceutical University, Ministry of
Education,
Shenyang 110016, P. R. China
E-mail: suntiemin@126.com
http://www.syphu.edu.cn/
Our group has focused on the development of potent,
selective, and less-toxic antituberculosis drugs based on be-
daquiline and, at the same time, the study of the molecular
structures and characterization of related compounds.[5–8]
As mentioned earlier, the R/S chirality of bedaquiline
[b] School of Life Science and Biopharmaceutics, Shenyang
Pharmaceutical University,
Shenyang 110016, P. R. China
http://www.syphu.edu.cn/
Supporting information for this article is available on the
WWW under http://dx.doi.org/10.1002/ejoc.201400135.
Eur. J. Org. Chem. 0000, 0–0
© 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1

Job/Unit: O40135
/KAP1
Date: 05-05-14 18:10:45
Pages: 9
Z. Wang, L. Zhao, Y. Chen, W. Xu, T. Sun
FULL PAPER
makes it more active and selective than other isomers. Tak- Scheme 2).[7] At first, the commercially available chiral tem-
ing into account the importance of chirality and on the ba- plate (S)-(–)-1-(4-methoxyphenyl)ethylamine was intro-
sis of the structure–activity relationship, we designed sev- duced to 1 and, thus, a mixture of diastereoisomers 2a and
eral chiral analogs of bedaquiline with two stereogenic cen- 2b showing two distinguishable TLC spots was generated.
ters to develop antituberculosis drugs (Scheme 1). However, Common column chromatography enabled the separation
it is challenging to obtain optically pure products in a usual of the diastereomers into the less-polar 2a and more-polar
way and, the subsequent determination of the absolute 2b in 30% yields. Then, the two intermediates were acylated
stereochemistry is needed for the study of the structure– with chloroacetyl chloride and converted to the desired
activity relationship. In our research, we constructed the products 3a and 3b in the usual way through a subsequent
two stereogenic centers in the designed analogs by introduc- amination reaction with morpholine.
ing a chiral template to the racemic starting material; there-
fore, pairs of diastereoisomers would be generated and
would be separable into optically pure products by common
column chromatography. To identify the absolute configu-
rations, ECD spectroscopy could be considered as an op-
tion. However, the multiple chromophores in the designed
structures would lead to very complex spectra, which would
be difficult to analyze. Quantum chemical calculations
make this analysis feasible. In this paper, two diastereoiso-
mers of the designed compounds were chosen as representa-
tive compounds for the stereostructure investigation. The
ACs of the two compounds were determined by employing
experimental and theoretical ECD spectroscopy. To validate
the assignments of the stereostructures and to obtain
deeper insights into the structural characteristics of the de-
signed analogs, single crystals of both compounds were cul-
tivated in mixed solvents, and X-ray structure analyses were
performed. Moreover, conformational analyses of the two
compounds in the ground state were conducted, and the
structures optimized by a DFT method were compared
with the X-ray structures. Finally, the in vitro antitubercu-
losis activities of the two compounds were investigated.
Scheme 2. Synthesis of 3a and 3b: (a) (S)-1-(4-methoxyphenyl)-
ethylamine, column separation; (b) Et3N, ClCH2CClO; (c) morph-
oline, K2CO3.
Determination of the ACs of 3a and 3b
The two stereogenic centers in the target compounds may
lead to two possible absolute configurations, the S/S config-
uration (3a) and the R/S configuration (3b), as the intro-
duced chiral source is already known to have the S configu-
ration and the other chiral carbon is indefinite. To deter-
mine their ACs, their ECD spectra were predicted by quan-
tum chemical calculations.
Scheme 1. Bedaquiline and the designed derivatives.
Prior to the ECD calculations, a conformational search
must be performed to define the stable conformations.
Theoretically, the conformational characteristic of a mol-
ecule can affect its physical properties at the working tem-
perature. The energy differences among the conformers can
Results and Discussion
Preparation of the Target Compounds
The target compounds were synthesized from 6- usually be overcome by the energy contributions and, con-
bromo-3-[chloro(phenyl)methyl]-2-methoxyquinoline
(1, sequently, the ECD spectra are influenced by the particular
2
www.eurjoc.org
© 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 0000, 0–0

Job/Unit: O40135
/KAP1
Date: 05-05-14 18:10:45
Pages: 9
Determination of Absolute Configurations of Bedaquiline Analogs
energy population distribution.[9] In the cases of 3a and 3b, thus, the compound is conformationally flexible. Newman
the single bonds that connect the three independent projections of the conformers along the C–N bond axis
chromophores, namely, the substituted quinoline ring and show that the substituents of the front carbon atom and the
the two phenol rings, may rotate at room temperature and back nitrogen atom are staggered in conformers 3a-1 and
lead to conformational changes. As a result, the position of 3a-2, whereas the phenyl ring and morpholine ring gen-
the chromophores may change, and the shape of the ECD erally eclipse each other in 3a-3; which forces more interac-
spectrum may be influenced significantly. Therefore, the ob- tions between the front and back substituents and creates
served room-temperature ECD spectrum is a Boltzmann more strain (Figure 1). The ground-state energies of 3a-1
average over this wide ensemble. Hence, all conformers and 3a-2 are similar, and only a slight difference between
present in significant proportions should be considered in the orientation of the methoxy group is observed.
calculations to obtain accurate predictions. On the other
hand, solvation can also influence the experimental chirop-
tical properties significantly, even in the absence of solvent
hydrogen-bonding interactions. Vaccaro and co-workers
have disclosed significant differences between gas-phase op-
tical rotations and solution-phase measurements by cavity
ring-down polarimetry.[10] Therefore, it is necessary to in-
clude the factor of solvation for the theoretical simulation
of chiroptical properties. The polarizable continuum model
(PCM) is by far the most widely used method for studying
the influence of solvation.[11]
Initial conformational searches for 3a and 3b were per-
formed by Monte Carlo methods together with the
MMFF94 molecular mechanics force field. The structures
found were then optimized in methanol by using the PCM
at the DFT/B3LYP/6-31G** level in the Gaussian 09 pro-
gram package.[12] The relative free energies of the conforma-
tions of 3a and 3b in methanol are shown in Table 1. From
the relative free energies, the percentage population of each
conformation in a room-temperature equilibrium mixture
can be predicted. Generally, the molecules have conforma-
tions with high strain as well as conformations with low
strain. High-strain conformations are when certain parts of
the molecule that repel are forced to be close to one an-
Figure 1. Relatively stable conformers of 3a and 3b. All conformers
are shown as Newman projections viewed along the C–N bond
axis. The substituted quinoline and phenyl rings are at the front,
and the morpholine ring and the (S)-1-(4-methoxyphenyl)ethyl
group are at the back.
In the case of 3b, the conformational search led to eleven
other. Therefore, low-strain molecular conformations are conformations. Of these, the major contributions, which
favored. Newman projections provide one way to look at have a Boltzmann weighting factor larger than 5%, are 3b-
the conformation of a noncyclical molecule.
1 (52.9%), 3b-2 (18.7%), 3b-3 (7.3%), and 3b-4 (5.0%). It
is clear that two biggest groups in 3b-1, namely, the substi-
tuted quinoline ring and the (S)-1-(4-methoxyphenyl)ethyl
group, adopt a staggered conformation to reduce strain in
the molecule. The structural difference between 3b-2 and
3b-4 is caused by the twisting of flexible side chain, which
changes the position of the morpholine ring. Compared
with 3b-2, conformer 3b-4 is less favorable because it forces
more interactions between the two groups [i.e., the morph-
oline ring and the (S)-1-(4-methoxyphenyl)ethyl group]
attached to the back nitrogen atom (Figure 1).
Additionally, the theoretically calculated dipole moments
(Table 1) show that 3a is less polar than 3b, in agreement
with the experimentally observed tendency.
The ECD spectra of the stable conformers of 3a and 3b
in methanol were calculated by using the PCM with time-
dependent density functional theory (TD-DFT) method at
the CAM-B3LYP/TZVP level. Compared with other hybrid
functionals, TD-DFT has been recognized as the high-level
Table 1. Relative free energies, dipole moments (μ), and pop-
ulations of the conformations of 3a and 3b as determined in meth-
anol.
Con-
former
ΔG[b]
P
μ
Con-
former
ΔG[b]
P
μ
[%][c] [D]
[%][c] [D]
3a-1[a]
3a-2[a]
3a-3[a]
3a-4
0.00 35.2
0.08 30.7
0.12 28.7
1.78 1.7
1.97 1.3
2.01 1.2
2.21 0.8
2.98 0.2
3.78 0.1
3.31
6.37
5.25
3.77
6.76
4.65
4.07
6.15
3.16
4.88
6.60
3b-1[a]
3b-2[a]
3b-3[a]
3b-4[a]
3b-5
0.00 52.9
0.62 18.7
1.17 7.3
1.39 5.0
1.45 4.6
1.51 4.1
1.53 4.0
2.08 1.6
2.13 1.4
3.32 0.2
3.68 0.1
7.00
6.88
6.26
6.27
8.15
3.60
6.02
2.17
2.26
3.26
4.60
3a-5
3a-6
3b-6
3a-7
3b-7
3a-8
3b-8
3a-9
3b-9
Weighted μ for 3a
3b-10
3b-11
Weighted μ for 3b
[a] See Figure 1 for the structures of the conformers. [b] ΔG in
kcalmol–1. [c] Population based on ΔG value.
In the case of 3a, the conformational search led to nine method that leads to the best accuracy/cost compromise in
conformations, and 3a-1 (35.2%), 3a-2 (30.7%), and 3a-3 ECD calculations.[13] Many instances have indicated that
(28.7%) are significantly populated at room temperature; the use of the CAM-B3LYP functional led to better agree-
Eur. J. Org. Chem. 0000, 0–0
© 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
3

Job/Unit: O40135
/KAP1
Date: 05-05-14 18:10:45
Pages: 9
Z. Wang, L. Zhao, Y. Chen, W. Xu, T. Sun
FULL PAPER
ment between predicted and experimental ECD spectra, shows a positive CE at 187 nm and the two negative CEs
such as the simulation of the ECD spectra of dimethyl gar- at 196 and 232 nm.
cinia,[14] oxysporone,[15] regiolone, and isosclerone.[16] To
The theoretical ECD spectra of 3a and 3b are also de-
generate the final predicted spectra, all of the simulated picted in Figure 2. A UV correction has been applied to the
spectra of the lowest-energy conformations were averaged theoretical spectra (Figure S8). In the so-called UV correc-
according to the Boltzmann distribution theory by the tion, one looks for a match between the computed and ex-
Specdis program.[17]
perimental UV/Vis spectra and then applies the same shift
The experimental ECD spectra of 3a and 3b in MeOH to the ECD ones.[18]
(1.99ϫ10–5 m) in the 180–300 nm range are shown in Fig-
The calculated ECD spectra of the S/S conformer repro-
ure 2. In the ECD spectrum of 3a, a positive cotton effect duced well the sign of the experimental CEs of 3a at 188,
(CE) is visible at 196 nm, a negative CE is visible at 215 nm, 196, 215, and 232 nm. Meanwhile, the CEs at 187, 195, 210,
and a positive band at 232 nm displays vibrational struc- and 232 nm in the experimental ECD spectrum of 3b were
ture. In the case of 3b, the experimental ECD spectrum satisfactorily reproduced by the simulated ECD spectrum
shows a small positive CE at 187 nm and three negative for the R/S configuration. Notwithstanding some small dis-
CEs at 195, 210, and 232 nm. As the conformational flexi- crepancies, this result supports the assignment of an S/S
bility of the two isomers and the shape of the ECD spectra configuration to 3a and an R/S configuration to 3b.
may change as the solvent changes, the ECD spectra of 3a
The molecular conformations and stereochemistry affect
and 3b were further measured in acetonitrile (1.32ϫ10–5 m) the CEs greatly. The calculated ECD spectra of individual
in the 180–300 nm range. As shown in Figure 2, the ECD conformers of 3a and 3b are shown in Figure 3. The calcu-
spectrum of 3a in acetonitrile shows the same pattern as lated ECD spectrum of 3a-1 shows CEs at 200, 209, and
that recorded in methanol. In the ECD spectrum of 3b, al- 232 nm with the same signs as those of 3a-2 and only
though the acetonitrile ECD spectrum presents a shoulder slightly larger magnitude at 209 and 232 nm. The calculated
peak at 210 nm, remarkable similarities were observed in ECD spectrum of 3a-3 shows a blueshift of 7 nm compared
the two different solvents, and the acetonitrile spectrum with that of 3a-2 and a CE of opposite sign at 209 nm. In
the case of 3b, all of the calculated ECD spectra of the
individual conformers show CEs of the same signs over the
spectral range under 200 nm. Conformers 3b-1 and 3b-3
show the same negative CEs at 209 and 232 nm, whereas
3b-2 and 3b-4 show the opposite signs at these positions.
As 3b-1 is the preferred conformation in methanol and oc-
cupies the highest Boltzmann population among the con-
formers, the weighted ECD spectrum shows similar features
to that of 3b-1. From the aforementioned analyses, the
ECD spectra of individual conformers vary greatly because
of conformational and configurational changes.
The origin of the CEs in the experimental ECD spectra
of 3a and 3b could be explained by molecular orbital (MO)
analysis at the same level as the ECD calculation. Figure 4
shows the MOs of 3a mainly involved in the electronic tran-
sitions used to assign the experimental bands. The positive
CE at 196 nm has contributions from the electronic transi-
tion from MO155 and MO156 to MO159 and from
MO155ǞLUMO158 (LUMO = lowest unoccupied MO).
In addition, the negative CE at 215 nm in the experimental
spectrum might be caused by the transition from MO151
to MO160 (πǞπ*) with
a
contribution from
a
HOMO157ǞLUMO158 (πǞπ*; HOMO = highest occu-
pied MO) excitation. The electronic absorption peak at
232 nm is mainly described by the transitions from MO154,
MO155, and HOMO157 to LUMO158 with an important
contribution from MO156 to LUMO159. This peak is
mainly dominated by transitions from the three separate
rings to the substituted quinoline ring. The above results
suggest that the CEs at 196 nm in the ECD spectra are
mainly dominated by the spatial position of the quinoline
and morpholine ring, yet the CEs at 215 and 232 nm are
significantly influenced by the three separate aryl rings. In
the case of 3b (Figure S12), the CEs at 210 and 232 nm in
Figure 2. Experimental and calculated (TD-DFT/CAM-B3LYP/
TZVP) ECD spectra of (a) 3a and (b) 3b.
4
www.eurjoc.org
© 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 0000, 0–0

Job/Unit: O40135
/KAP1
Date: 05-05-14 18:10:45
Pages: 9
Determination of Absolute Configurations of Bedaquiline Analogs
might reflect its behavior in solution and could serve to
validate our assignments of the ACs. Thus, we further per-
formed a structural determination of the two compounds
by the X-ray diffraction approach. The crystals were grown
by slow evaporation of water/ethanol mixtures under ambi-
ent conditions. Compound 3a crystallizes in the ortho-
rhombic crystal system with the P212121 space group [unit
cell: a = 13.6246(18) Å, b = 13.7652(18) Å, c = 15.763(2) Å].
Compound 3b also crystallizes in the orthorhombic crystal
system with the P212121 space group [unit cell: a =
9.9808(12) Å, b = 10.9712(14) Å, c = 27.042(3) Å]. The
crystallographic and refinement data are shown in Table S1.
ORTEP diagrams of the molecular structures and the
atomic numbering schemes of 3a and 3b are shown in Fig-
ures 5 and 6, respectively. The absolute configurations at
both stereogenic centers in the molecular structure 3a are
indeed S/S, and those in 3b are R/S. This is consistent with
our conclusion obtained by ECD analyses.
Figure 3. Calculated ECD spectra of individual conformers of (a)
3a and (b) 3b.
the experimental ECD spectrum are dominated by the same
transitions as the CEs at 215 and 232 nm for 3a. The nega-
tive CE at 195 nm is caused by a combination of electronic
transitions from MO152ǞLUMO158, MO156ǞMO159,
and MO157ǞMO161 and MO163, which are slightly dif-
ferent from the transitions in 3a responsible for the CE at
196 nm.
Figure 5. DFT-optimized and crystal structure of 3a; displacement
ellipsoids are drawn at the 30% probability level. The hydrogen
atoms are omitted for clarity.
Some selected experimental and calculated geometry pa-
rameters for 3a and 3b are listed in Table 2 (all parameters,
including bond lengths and angles are shown in Table S2).
The N(2)–C(18)–C(19) and N(2)–C(18)–O(2) bond angles
of 120.5 and 120.9°, respectively, indicate the sp2-hybridized
nature of the C(18) atom. The morpholine system exists in
a “chair” conformation, and the bond angles indicate the
sp3-hybridized nature of those atoms. The C(8)–C(9)–N(1)–
C(4) torsion angle of 0.85° indicates that the substituted
quinoline ring remains almost planar. The dihedral angle
between the atoms of the substituted quinoline and benzene
(A) planes is 87.9°. They are almost perpendicular with re-
spect to one another. The dihedral angle between the atoms
of the substituted quinoline and benzene (B) planes is 21.7;
hence, the benzene (B) plane and the substituted quinoline
are nearly coplanar.
Figure 4. Molecular orbitals involved in the electronic transitions
of 3a.
X-ray Structure Analysis for 3a and 3b
The ring centroid–centroid distance of 3.930 Å between
The conformational and stability analyses of the two the substituted quinoline and benzene (B) plane is indica-
compounds revealed that they show molecular flexibility. tive of a weak π-stacking stabilization. These interactions
On the basis of these considerations, we assumed that the are important in view of the stability of the crystal struc-
conformation of the compound in the crystalline state ture.
Eur. J. Org. Chem. 0000, 0–0
© 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
5

Job/Unit: O40135
/KAP1
Date: 05-05-14 18:10:45
Pages: 9
Z. Wang, L. Zhao, Y. Chen, W. Xu, T. Sun
FULL PAPER
angle is 77.8°. Among all the DFT-optimized conformers of
3b, conformer 3b-1 was in accord with its relative crystalline
structure. The optimized geometry of 3b-1 shows no bond
length differences of more than 0.04 Å from the X-ray val-
ues (Table S2), whereas the calculated bond angles deviate
from the X-ray values by –2 and 2.5° at N(2)–C(11)–C(12)
and N(2)–C(18)–C(20), respectively. The results of the DFT
calculations show excellent agreement with the experiment
values and further confirm the ACs obtained by ECD
analyses.
We have analyzed the hydrogen-bond interactions in the
molecular crystal structures of 3a and 3b. No conventional
hydrogen bonds were found at 293(2) K. The molecular
packing is mainly determined by van der Waals interactions
and a few weak hydrogen bonds. There are weak intermo-
lecular hydrogen bonds C(32)–H(32a)···O(2)i (i = –x, 1/2 +
y, 3/2 – z) in 3a and C(1)–H(1)···O(4)ii (ii = 1 – x, –1/2 + y,
3/2 – z) in 3b (Table S3). In addition to the interactions
described above, the crystal structure can be stabilized by
weak π-stacking interactions between the planes of the π-
electron systems of the substituted quinoline and benzene
(B) rings.
Figure 6. DFT-optimized and crystal structure of 3b; displacement
ellipsoids are drawn at the 30% probability level. The hydrogen
atoms are omitted for clarity.
Table 2. Selected experimental and calculated geometry parameters
for 3a and 3b.
Antituberculosis Activity
Bond angle [°]
Exp. (3a)
Calcd.[a]
Difference
In the antitubercular activity studies, both 3a and 3b
were found to be biologically active; the inhibitory activity
of 3b was a little better than that of 3a, and the MIC values
were 20.3 and 50 mgmL–1, respectively. This is consistent
with our hypothesis that the inhibitory activity of the R/S
isomer is better than that of the S/S stereoisomer.
C(8)–C(11)–C(12)
C(8)–C(11)–N(2)
N(2)–C(11)–C(12)
N(2)–C(25)–C(24)
N(2)–C(25)–C(26)
C(24)–C(25)–C(26)
N(2)–C(18)–C(19)
N(2)–C(18)–O(2)
113.9
114
113.6
115.1
113.1
112.4
111.5
115
–0.3
1.1
–0.7
2
113.8
110.4
110.4
117.3
120.5
120.9
1.1
–2.3
–0.8
1.1
119.7
122
Exp. (3b)
Calcd.[b]
Difference
Conclusions
C(8)–C(11)–C(12)
C(8)–C(11)–N(2)
N(2)–C(11)–C(12)
N(2)–C(18)–C(19)
N(2)–C(18)–C(20)
C(19)–C(18)–C(20)
N(2)–C(27)–C(28)
N(2)–C(27)–O(3)
112.3
113.2
115.4
111.5
110.4
115.4
118.1
122.9
113.6
115
113.4
111.1
112.9
114
1.3
1.8
–2
–0.4
2.5
–1.4
1.5
–0.8
We have synthesized two chiral analogs of bedaquiline
for the first time. The ACs of the isomers were determined
from experimental and theoretical ECD spectra. X-ray
structure analyses further validated the assignments of the
ACs; therefore ECD analysis aided by quantum chemistry
calculations can provide a great help to identify the absolute
configurations of the designed compounds. The theoretical
calculations combined with the experimental results permit-
ted the unambiguous assignment of the S/S configuration
119.6
122.1
[a] Calculated geometry parameters for conformer 3a-3. [b] Calcu-
lated geometry parameters for conformer 3b-1.
The crystal structure of 3a was compared with the DFT- to less-polar 3a and the R/S configuration to more-polar
optimized structures (Table 2). Among all the conformers 3b. The results of the optimization of 3a-3 and 3b-1 by DFT
of 3a, conformer 3a-3 was in accord with the crystalline calculations show excellent agreement with experimental re-
structure. It is not surprising that the most-populated con- sults and further proved the accuracy of the AC assignment.
former 3a-1 does not match the crystalline structure as the In addition, tests of the anti-TB activity of the compounds
structures were calculated in menthol; in the solid state, the have verified our assumption that the activity of the R/S
overcrowded space increases intermolecular forces and mo- isomer is better than that of S/S stereoisomer.
lecular energy. As expected, most of the calculated geome-
try parameters for the two compounds are close to the X- search. The S/S isomers are less polar than the R/S ones,
ray values. and the theoretical calculations of the molecular dipole mo-
An interesting phenomenon was observed in our re-
Similarly, in the structure of 3b, the dihedral angle be- ments well reproduced the experimental results. This shed
tween the atoms of the substituted quinoline and benzene a light on the possibility that theoretical calculations of the
(A) planes is 85.7°. The benzene (B) plane is also almost dipole moments may also provide a meaningful reference
perpendicular to the benzene (A) plane, and the dihedral to assign the AC of a molecule.
6
www.eurjoc.org
© 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 0000, 0–0

Job/Unit: O40135
/KAP1
Date: 05-05-14 18:10:45
Pages: 9
Determination of Absolute Configurations of Bedaquiline Analogs
7.33–7.35 (m, 2 H), 7.27–7.28 (m, 1 H), 7.15–7.16 (m, 2 H), 6.85–
6.87 (m, 2 H), 4.95 (s, 1 H), 3.93 (s, 3 H), 3.79 (s, 3 H), 3.47 (q, J
= 7.0 Hz, 1 H), 1.36 (d, J = 6.7 Hz, 3 H) ppm. 13C NMR
(100 MHz, CDCl3): δ = 160.4, 158.7, 144.5, 144, 137.3, 134.9,
132.2, 129.8, 129.7, 129.3, 128.8, 128.5, 128.4, 128.3, 128.1, 128,
127.9, 127.8, 127.6, 113.8, 113.8, 58.4, 55.4, 55.4, 53.8, 24.3 ppm.
MS (ESI+): m/z = 477.1 [M + H]+, 479.1 [M + 2 + H]+.
Experimental Section
General Remarks: IR spectra were recorded by using a Perkin–
Elmer one FTIR spectrometer with KBr pellets. H and 13C NMR
1
spectra were recorded at 400 and 100 MHz, respectively, with tet-
ramethylsilane (TMS) as an internal standard and CDCl3 as sol-
vent. Mass spectra were recorded with a Hewlett–Packard 1100
LC/MSD spectrometer (Hewlett–Packard, USA). ECD spectra
were recorded with a JASCO CD-2095 spectropolarimeter. The X-
ray diffraction data of the selected crystals of 3a with dimensions
of 0.31ϫ 0.24ϫ0.16 mm and 3b with dimensions of
0.35ϫ0.23ϫ0.20 mm were recorded with a Bruker P4 X-dif-
fractometer; the data were collected by using graphite-monochro-
mated Mo-Kα radiation (λ = 0.71073 Å) at 293 K. For 3a and 3b,
data collection: APEX2;[19] cell refinement: SAINT;[20] program
used to solve structure: SHELXS-97;[21] program used to refine
structure and draw molecular figures: SHELXTL-97;[22] program
used to measure centroid–centroid distance: Mercury 2.3.[23]
General Procedure for the Synthesis of 3a and 3b: To a solution
of 2a (1.18 g, 2.8 mmol) and triethylamine (0.610 g, 8.4 mmol) in
dichloromethane (50 mL) at 0 °C was added 2-chloroacetyl chlor-
ide (0.332 g, 2.94 mmol) diluted with dry dichloromethane (10 mL)
under a nitrogen atmosphere. After stirring for 2 h at –5 °C and
1 h at room temp., the mixture was treated with water (30 mL), and
the organic layer was washed with brine (30 mLϫ2). The com-
bined organic phase was then dried with anhydrous sodium sulfate.
After concentration in vacuo, the crude product (no further purifi-
cation was necessary) was dissolved in acetonitrile (50 mL), and
then morpholine (0.245 g, 2.8 mmol) and potassium carbonate
(1.35 g, 9.8 mmol) were added to the above solution. After stirring
for 8 h at 75 °C, the reaction mixture was quenched with water
and then extracted with ethyl acetate. The organic phase was then
dehydrated with anhydrous sodium sulfate and concentrated at re-
duced pressure, and the residue was purified by flash column
chromatography (40% ethyl acetate in hexane) to afford a white
powder in a yield of 60% (0.907 g).
Materials: Triethylamine and acetonitrile were purchased from the
Aldrich, Co. Ltd. and used after dehydration with 4 Å molecular
sieves. Morpholine, sodium iodide, (S)-(–)-1-(4-methoxyphenyl)-
ethylamine, and 2-chloroacetyl chloride were purchased from the
Wako Co. Ltd. All reagents were of analytical grade and are com-
mercially available. For TLC analysis, precoated plates of silica gel
60 F254 were used, and spots were visualized with UV light.
CCDC-916936 (for 3a) and -916935 (for 3b) contain the supple-
mentary crystallographic data for this paper. These data can be
obtained free of charge from The Cambridge Crystallographic
Data Centre via www.ccdc.cam.ac.uk/data_request/cif.
N-[(S)-(6-Bromo-2-methoxy-3-quinolinyl)(phenyl)methyl]-N-[(S)-1-
(4-methoxyphenyl)ethyl]-2-morpholinoacetamide (3a): White solid.
[α]D = +18.8 (c = 0.21, CH Cl ). IR (KBr pellets): ν = 3442 (NH,
˜
2
2
C=O), 2951, 2852 (OCH3), 1649 (C=O), 1569, 1513, 1493
(ArC=C), 1462 (CH2), 1401 (CH3), 1343 (–CH), 1252 (=C–O),
1180 (C–N), 1116, 1010 (–C–O), 1074 (C–Br), 867, 828, 739, 698
(ArH) cm–1. 1H NMR (400 MHz, CDCl3): δ = 7.67–7.72 (m, 1 H),
7.59–7.62 (m, 1 H), 7.55–7.52 (m, 1 H), 7.49–7.47 (m, 1 H), 7.33–
7.42 (m, 3 H), 7.22–7.24 (m, 2 H), 7.06–7.10 (m, 2 H), 6.61–6.83
(m, 2 H), 5.57 (br s, 1 H), 3.83 (s, 3 H), 3.75 (s, 3 H), 3.55 (br s, 4
H), 3.34 (s, 2 H), 2.56 (br s, 4 H), 1.90 (d, J = 6.8 Hz, 3 H) ppm.
13C NMR (100 MHz, CDCl3): δ = 173.7, 159.1, 155, 144.6, 144.3,
137.6, 134.6, 132.6, 129.4, 129.2, 129, 128.7, 128.5, 128.3, 128.1,
127.5, 127.4, 127.4, 127.2, 126.1, 112.9, 112.9, 66.5, 66.5, 63.1, 55.1,
55.1, 53.1, 53.7, 53.7, 53.4, 20.9 ppm. MS (ESI+): m/z = 604.1 [M
+ H]+, 606.1 [M + 2 + H]+.
General Procedure for the Synthesis of 2a and 2b: To a stirred solu-
tion of 6-bromo-3-[chloro(phenyl)methyl]-2-methoxyquinoline
(3.27 g, 9 mmol; prepared according to the literature method),[7]
potassium carbonate (4.35 g, 31.5 mmol), and sodium iodide
(0.132 g, 0.9 mmol) in acetonitrile (180 mL) at 20 °C was added
(S)-(–)-1-(4-methoxyphenyl)ethylamine (1.50 g, 9.9 mmol) under a
nitrogen atmosphere until no starting material could be detected
by TLC. The reaction mixture was then processed with dichloro-
methane in the usual manner. After drying and evaporation of the
solvent, the residue was purified by silica gel chromatography with
0.5% ether in petroleum ether to give less-polar 2a (1.30 g, 30%)
and more-polar 2b (1.28 g, 30%).
N-[(R)-(6-Bromo-2-methoxy-3-quinolinyl)(phenyl)methyl]-N-[(S)-1-
(4-methoxyphenyl)ethyl]-2-morpholinoacetamide (3b): White solid.
(S)-N-[(S)-(6-Bromo-2-methoxy-3-quinolinyl)(phenyl)methyl]-1-(4-
methoxyphenyl)ethanamine (2a): White solid. [α]D = –43.7 (c = 0.21,
[α]D = –127.3 (c = 0.22, MeOH). IR (KBr pellets): ν = 3439 (NH,
˜
CH Cl ). IR (KBr pellets): ν = 3438 (NH), 2951 (–OCH ), 1600,
˜
2
2
3
C=O), 2953, 2852 (OCH3), 1649 (C=O), 1569, 1513, 1493
(ArC=C), 1461 (CH2), 1401 (CH3), 1343 (–CH), 1252 (=C–O),
1180 (C–N), 1117, 1011 (–C–O), 1075 (C–Br), 867, 828, 740, 698
(ArH) cm–1. 1H NMR (400 MHz, CDCl3): δ = 8.18 (s, 1 H), 7.86–
7.89 (m, 1 H), 7.79–7.81 (m, 1 H), 7.63–7.69 (m, 1 H), 7.42–7.56
(m, 2 H), 7.30–7.34 (m, 1 H), 7.08 (br s, 2 H), 6.88 (br s, 2 H),
6.57–6.70 (m, 2 H), 5.61–5.79 (m, 1 H), 3.98 (s, 3 H), 3.82 (s, 3 H),
3.65–3.73 (m, 4 H), 3.01–3.36 (m, 2 H), 2.53 (br s, 4 H), 1.46 (br
s, 3 H) ppm. 13C NMR (100 MHz, CDCl3): δ = 173.5, 159.8, 155,
144.6, 144.3, 138.9, 132.7, 131.4, 130, 129.6, 129.6, 129.6, 128.8,
128.5, 128.3, 127.9, 127.9, 127.9, 127.5, 126.5, 113.8, 113.8, 66.7,
66.7, 63.3, 55.3, 55.3, 55.3, 53.9, 53.9, 19.0 ppm. MS (ESI+): m/z
= 604.1 [M + H]+, 606.1 [M + 2 + H]+.
1512, 1566, 1458 (C=C), 1400 (CH3), 1332 (CH), 1178 (C–N), 1251
(=C–O), 1106, 1008 (C–O), 1073 (C–Br), 824, 748, 700 (ArH) cm–1.
1H NMR (400 MHz, CDCl3): δ = 8.10 (s, 1 H), 7.74–7.76 (m, 1
H), 7.63–7.70 (m, 1 H), 7.50–7.53 (m, 1 H), 7.26–7.28 (m, 2 H),
7.23–7.25 (m, 2 H), 7.21 (s, 1 H), 7.15–7.17 (m, 2 H), 6.86–6.89 (m,
2 H), 4.91 (1 H, S), 3.91 (s, 3 H), 3.81 (s, 3 H), 3.65 (q, J = 6.6 Hz,
1 H), 1.37 (d, J = 6.6 Hz, 3 H) ppm. 13C NMR (100 MHz, CDCl3):
δ = 160.9, 158.8, 144.5, 144.2, 135.3, 135.2, 132.3, 130.7, 129.7,
129.6, 129.5, 128.7, 128.4, 128.4, 128.4, 127.8, 127.6, 127.1, 126.4,
114, 114, 58.8, 55.4, 55.2, 53.6, 24.6 ppm. MS (ESI+): m/z = 477.1
[M + H]+, 479.1 [M + 2 + H]+.
(S)-N-[(R)-(6-Bromo-2-methoxy-3-quinolinyl)(phenyl)methyl]-1-(4-
methoxyphenyl)ethanamine (2b): White solid. [α]D = –58.3 (c = 0.21,
CH Cl ). IR (KBr pellets): ν = 3439 (NH), 2952 (OCH ), 1600, Computational Details: The conformational analysis was performed
˜
2
2
3
1511, 1569, 1492 (C=C), 1401 (CH3), 1341 (CH), 1178 (C–N), 1244
by arbitrarily fixing the absolute configuration of C-11 for 3a and
(=C–O), 1117, 1010 (C–O), 1075 (C–Br), 827, 732, 699 (ArH) cm–1. 3b and using the Spartan 08 package with the MMFF94 molecular
1H NMR (400 MHz, CDCl3): δ = 8.18 (s, 1 H), 7.78–7.79 (m, 1
H), 7.67–7.69 (m, 1 H), 7.45–7.48 (m, 1 H), 7.33–7.35 (m, 2 H),
mechanics force field and Monte Carlo searching. Conformers
within a 4 kcal/mol energy window from the particular global mini-
Eur. J. Org. Chem. 0000, 0–0
© 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
7

Job/Unit: O40135
/KAP1
Date: 05-05-14 18:10:45
Pages: 9
Z. Wang, L. Zhao, Y. Chen, W. Xu, T. Sun
FULL PAPER
[7] Y. Bai, L. Wang, Y. Chen, L. Yuan, W. Xu, T. Sun, J. Mol.
Struct. 2011, 1005, 113–120.
[8] L. Li, T. Cai, Z. Wang, Z. Zhou, Y. Geng, T. Sun, Spectrochim.
Acta Part A 2014, 120, 106–118.
[9] G. Barone, D. Duca, A. Silvestri, L. Gomez-Paloma, R. Riccio,
G. Bifulco, Chem. Eur. J. 2002, 8, 3240–3245.
[10] T. Mueller, K. B. Wiberg, P. H. Vaccaro, J. Phys. Chem. A 2000,
104, 5959–5968.
mum were chosen for geometrical optimization and energy calcula-
tion at the DFT/B3LYP/6-31G** level of theory in the program
package Gaussian 09. The relative population of each conformer
was valued on the basis of a Boltzmann weighting factor at 298 K.
Vibrational analysis was formed at the same level to confirm the
stability of minima. TD-DFT/CAM-B3LYP/TZVP was employed
to calculate the excitation energies (denoted by wavelength in nm)
and rotatory strength R. ECD spectra were calculated by Specdis
on the basis of rotatory strengths by using a half-bandwidth of
0.2 eV with conformers. The spectra were constructed on the basis
of the Boltzmann weighting according to their population contri-
bution.
[11] B. Mennucci, J. Tomasi, R. Cammi, J. R. Cheeseman, M. J.
Frisch, F. J. Devlin, S. Gabriel, P. J. Stephens, J. Phys. Chem. A
2002, 106, 6102–6113.
[12] M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria,
M. A. Robb, J. R. Cheeseman, G. Scalmani, V. Barone, B.
Mennucci, G. A. Petersson, H. Nakatsuji, M. Caricato, X. Li,
H. P. Hratchian, A. F. Izmaylov, J. Bloino, G. Zheng, J. L. Son-
nenberg, M. Hada, M. Ehara, K. Toyota, R. Fukuda, J. Hase-
gawa, M. Ishida, T. Nakajima, Y. Honda, O. Kitao, H. Nakai,
T. Vreven, J. A. Montgomery, Jr., J. E. Peralta, F. Ogliaro, M.
Bearpark, J. J. Heyd, E. Brothers, K. N. Kudin, V. N. Starov-
erov, T. Keith, R. Kobayashi, J. Normand, K. Raghavachari,
A. Rendell, J. C. Burant, S. S. Iyengar, J. Tomasi, M. Cossi, N.
Rega, J. M. Millam, M. Klene, J. E. Knox, J. B. Cross, V.
Bakken, C. Adamo, J. Jaramillo, R. Gomperts, R. E. Strat-
mann, O. Yazyev, A. J. Austin, R. Cammi, C. Pomelli, J. W.
Ochterski, R. L. Martin, K. Morokuma, V. G. Zakrzewski,
G. A. Voth, P. Salvador, J. J. Dannenberg, S. Dapprich, A. D.
Daniels, O. Farkas, J. B. Foresman, J. V. Ortiz, J. Cioslowski,
D. J. Fox, Gaussian 09, revision C.01, Gaussian, Inc., Wall-
ingford, 2010.
Test of the Antituberculosis Activities of 3a and 3b: A series of broth
tubes containing diluted chemical synthetic compound concentra-
tions were prepared and inoculated with a 48 h liquid culture of
Mycobacterium phlei 1180. The tubes were then incubated at 37 °C
for 20 h. The MIC is defined as the lowest concentration that pre-
vented the mycobacterial growth, and the inhibition of the com-
pounds against other strains is under test. Rifampicin was chosen
as the standard drug, and the MIC was 10 mgmL–1. All com-
pounds were tested at concentration of 1.0 mgmL–1.
Supporting Information (see footnote on the first page of this arti-
cle): IR, ECD, and 1H and 13C NMR spectra of 2a, 2b, 3a, and
3b, detailed crystal and DFT-optimized geometry parameters for
3a and 3b, and hydrogen-bond geometries.
[13] J. Autschbach, Chirality 2009, 21, 116–152.
[14] P. L. Polavarapu, G. Scalmani, E. K. Hawkins, C. Rizzo, N.
Jeirath, I. Ibnusaud, D. Habel, D. S. Nair, S. Haleema, J. Nat.
Prod. 2011, 74, 321–328.
Acknowledgments
[15] G. Mazzeo, E. Santoro, A. Andolfi, A. Cimmino, P. Troselj,
A. G. Petrovic, S. Superchi, A. Evidente, N. Berova, J. Nat.
Prod. 2013, 76, 588–599.
[16] A. Evidente, S. Superchi, A. Cimmino, G. Mazzeo, L. Mugnai,
D. Rubiales, A. Andolfi, A.-M. Villegas-Fernández, Eur. J. Org.
Chem. 2011, 5564–5570.
This work was supported by the Program for Innovative Research
Team of the Ministry of Education, and the Program for Liaoning
Innovative Research Team in University. Theoretical calculations
were conducted on the ScGrid and Deepcomp7000 at the Super-
computing Center, Computer Network Information Center of the
Chinese Academy of Sciences.
[17] T. Bruhn, A. Schaumlöffel, Y. Hemberger, G. Bringmann, Chi-
rality 2013, 25, 243–249.
[18] G. Bringmann, K. Maksimenka, T. Bruhn, M. Reichert, T.
Harada, R. Kuroda, Tetrahedron 2009, 65, 5720–5728.
[19] APEX2 (v. 1.0–27), Bruker AXS Inc., Madison, 2005.
[20] SADABS (v. 2.01), SMART (v. 5.603), and SAINT (v. 6.36a),
Bruker AXS Inc., Madison, 2000.
[21] G. M. Sheldrick, SHELXTS-97 and SHELXTL-97, University
of Göttingen, Germany, 1997.
[22] G. M. Sheldrick, SHELXTL, v. 5.1, Bruker AXS, Inc., Madi-
son, 1997.
[23] C. F. Macrae, I. J. Bruno, J. A. Chisholm, P. R. Edgington, P.
McCabe, E. Pidcock, L. Rodriguez-Monge, R. Taylor, S. J.
van de Streek, P. A. Wood, J. Appl. Crystallogr. 2008, 41, 466–
470.
[1] M. R. de Jonge, L. H. M. Koymans, J. E. G. Guillemont, A.
Koul, K. Andries, Proteins Struct., Funct., Bioinf. 2007, 67,
971–980.
[2] FDA Approves Bedaquiline for Resistant TB Treatment, Med-
scape, Dec. 31, 2012.
[3] Y.-M. Sang, L.-K. Yan, J.-P. Wang, Z.-M. Su, J. Phys. Chem.
A 2012, 116, 4152–4158.
[4] P. L. Polavarapu, E. A. Donahue, G. Shanmugam, G. Scal-
mani, E.-K. Hawkins, C. Rizzo, I. Bnusaud, G. Thomas, D.
Habel, D. Sebastian, J. Phys. Chem. A 2011, 115, 5665–5673.
[5] Z.-Q. Cai, C.-J. Zhu, Y.-X. Cui, B. Jiang, T.-M. Sun, Sci. China
Ser. B 2009, 52, 2051.
[6] Z.-Q. Cai, S.-R. Li, T.-M. Sun, Chin. J. Struct. Chem. 2010, 29,
796.
Received: January 26, 2014
Published Online: ᭿
8
www.eurjoc.org
© 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 0000, 0–0

Job/Unit: O40135
/KAP1
Date: 05-05-14 18:10:45
Pages: 9
Determination of Absolute Configurations of Bedaquiline Analogs
Circular Dichroism
Two chiral analogs of bedaquiline were de-
signed as anti-Mycobacterium tuberculosis
drugs. The synthesis and determination of
the absolute configurations (ACs) of the
two compounds are described. The ACs
were determined by electronic circular
dichroism with the aid of theoretical calcu-
lations and further validated by an X-ray
diffraction study.
Z. Wang, L. Zhao, Y. Chen, W. Xu,
T. Sun* ............................................. 1–9
Determination of Absolute Configurations
of Bedaquiline Analogs by Quantum
Chemical Electronic Circular Dichroism
Calculations and an X-ray Diffraction
Study
Keywords: Chirality / Circular dichroism /
Density functional calculations / X-ray
diffraction / Medicinal chemistry
Eur. J. Org. Chem. 0000, 0–0
© 0000 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
9