A single chiroptical spectroscopic method may not be able to establish the absolute configurations of diastereomers: Dimethylesters of hibiscus and garcinia acids

Article abstract of DOI:10.1021/jp202501y

Electronic circular dichroism (ECD), optical rotatory dispersion (ORD), and vibrational circular dichroism (VCD) spectra of hibiscus acid dimethyl ester have been measured and analyzed in combination with quantum chemical calculations of corresponding spectra. These results, along with those reported previously for garcinia acid dimethyl ester, reveal that none of these three (ECD, ORD, or VCD) spectroscopic methods, in isolation, can unequivocally establish the absolute configurations of diastereomers. This deficiency is eliminated when a combined spectral analysis of either ECD and VCD or ORD and VCD methods is used. It is also found that the ambiguities in the assignment of absolute configurations of diastereomers may also be overcome when unpolarized vibrational absorption is included in the spectral analysis.

Full text of DOI:10.1021/jp202501y

ARTICLE
pubs.acs.org/JPCA
A Single Chiroptical Spectroscopic Method May Not Be Able To
Establish the Absolute Configurations of Diastereomers:
Dimethylesters of Hibiscus and Garcinia Acids
Prasad L. Polavarapu,†,* Emily A. Donahue,† Ganesh Shanmugam,† Giovanni Scalmani,‡
Edward K. Hawkins,† Carmelo Rizzo,† Ibrahim Ibnusaud,§ Grace Thomas,§ Deenamma Habel,§ and
Dellamol Sebastian§
†Department of Chemistry, Vanderbilt University, Nashville, Tennessee 37235, United States;
‡Gaussian, Inc., 340 Quinnipiac Street, Building 40, Wallingford, Connecticut 06492, United States
§Institute for Intensive Research In Basic Sciences, Mahatma Gandhi University, P.D. Hills PO, Kottayam, Kerala, India 686 560
S Supporting Information
b
ABSTRACT: Electronic circular dichroism (ECD), optical
rotatory dispersion (ORD), and vibrational circular dichroism
(VCD) spectra of hibiscus acid dimethyl ester have been
measured and analyzed in combination with quantum chemical
calculations of corresponding spectra. These results, along with
those reported previously for garcinia acid dimethyl ester, reveal
that none of these three (ECD, ORD, or VCD) spectroscopic
methods, in isolation, can unequivocally establish the absolute
configurations of diastereomers. This deficiency is eliminated when a combined spectral analysis of either ECD and VCD or ORD
and VCD methods is used. It is also found that the ambiguities in the assignment of absolute configurations of diastereomers may
also be overcome when unpolarized vibrational absorption is included in the spectral analysis.
’ INTRODUCTION
solvents (i.e., water or dimethylsulfoxide), and therefore pose
serious challenges for spectral interpretations. These challenges
arise from the difficulties in theoretical modeling of molecular
(soluteꢀsolvent) complexes, for which much progress remains to
be realized. In the meantime, conversion of these acids to the
corresponding dimethyl esters (see Chart 1), hibiscus acid dimethyl
ester (HADE), (2S,3R)-3and garcinia acid dimethyl ester (GADE),
(2S,3S)-3, allay these complications and make more straightforward
the conformational search and the overall interpretation of the
solvent effects, especially when an implicit solvent model is used.
The absolute configurations of chiral compounds containing
heavy atoms can be determined using X-ray crystal structures.
Chiroptical spectroscopic methods provide a convenient alter-
nate route for determining the absolute configurations when a
sample under investigation cannot be crystallized or a crystalline
sample does not contain a heavy atom. Four chiroptical spectro-
scopic methods that have gained increased attention5ꢀ8 in recent
years include electronic circular dichroism (ECD),5 optical
rotatory dispersion (ORD),5,6 vibrational circular dichroism
(VCD),5,7 and vibrational Raman optical activity (ROA).5,8
The increasing interest in these methods is due to the reliability
of their applications to chiral molecular structure determination
Hydroxycitric acids are widely utilized in medicines and food
additives.1 1,2-Dihydroxypropane-1,2,3-tricarboxylic acid, or
2-hydroxycitric acid, 1 (see Chart 1) has two adjacent chiral
centers and may exist in four stereoisomeric forms. The isolation
of stereoisomers of 1 in their open chain forms is extremely
difficult because the γ-hydroxyl group promotes spontaneous
lactonization to a γ-butyrolactone ring system during isolation.1
One of the lactone stereoisomers (see Chart 1), (2S,3R)-tetra-
hydro-3-hydroxy-5-oxo-2,3-furandicarboxylic acid, [(2S,3R)-2]
is called hibiscus acid, and its diastereomer, (2S,3S)-tetrahydro-
3-hydroxy-5-oxo-2,3-furandicarboxylic acid [(2S,3S)-2] is called
garcinia acid. Both (2S,3R)-2 and (2S,3S)-2 are extensively distrib-
uted in nature,2 but to the best of our knowledge no report is available
on the existence of their enantiomers, (2R,3S)-2 and (2R,3R)-2.
Although hibiscus acid is extracted from the calyxes/leaves of Hibiscus
sabdariffa (Mathippuli-Roselle plant), which is a common ingredient
in many herbal tea blends, garcinia acid is extracted from the dried
rind of the fruit of Garcinia cambogia, popularly known as “Malabar
tamarind”, and the latter has been credited for the unique regulatory
role on fatty acid synthesis and weight loss.3
Since the absolute configurations of hibiscus and garcinia acids
have been reported in the literature,4 their known absolute confi-
gurations can be used to test the ability of chiroptical spectroscopic
methods for predicting the absolute configurations. However,
carboxylic acids are typically soluble in a limited number of polar
Received: March 16, 2011
Revised:
April 28, 2011
Published: May 13, 2011
r
2011 American Chemical Society
5665
dx.doi.org/10.1021/jp202501y J. Phys. Chem. A 2011, 115, 5665–5673
|

The Journal of Physical Chemistry A
ARTICLE
Chart 1. Stereoisomers of 2-Hydroxycitric Acid, Corresponding Lactones and Their Dimethyl Esters
and the ease with which they can be applied to solution phase
measurements. The availability of commercial instruments has
made the routine experimental chiroptical spectroscopic mea-
surements feasible, and recent advances in quantum chemical
methods5ꢀ8 have resulted in achieving improved accuracy for
reliable predictions of chiroptical properties. Wider access to
chiroptical spectroscopic instruments and to quantum chemical
software that can calculate chiroptical properties, combined with
improved computer technology for faster quantum chemical
calculations, has made it convenient to apply chiroptical methods
for determining the structures of chiral molecules ranging from
small organic molecules to large natural products.9
If the absolute configuration of a chiral compound with a
single source of chirality (such as a center of chirality or helical
segment in the chiral molecule) is already known and the goal is
just to verify if the predicted chiroptical spectra concur with the
experimental spectra, one can calculate the chiroptical spectra for
the known absolute configuration, using reliable quantum chemi-
cal methods and verify if the results compare well with the
corresponding experimental spectra. Even if the absolute con-
figuration of a chiral compound with a single element of stereo-
chemistry is not known and the goal is to determine its absolute
configuration, the quantum chemical calculation of chiroptical
spectra for a chosen configuration is sufficient because those for
its enantiomer are obtained by multiplying the calculated spectra
by ꢀ1.
to have verified, or determined, the absolute configurations
unequivocally if the quantum chemically predicted chiroptical
spectra for more than one of the possible absolute configurations
have a resemblance to the experimental spectrum. Therefore, if
the goal is to determine the unknown absolute configurations of
diastereomers unequivocally or to assess if a given chiroptical
method can unequivocally verify the absolute configurations of
diastereomers, then it is necessary to compare the predicted
chiroptical spectra for all possible diastereomers with the experi-
mental spectrum of a given enantiomer and ascertain that only
one unequivocal match exists between predicted and experimen-
tal spectra. Unfortunately, this criterion is not strictly followed in
the literature, because the predictive ability of a chiroptical
spectrosopic method is often assessed by comparing only the
predicted spectrum for the known configuration with the corre-
sponding experimental spectrum. The presence/absence of any
correlation between the predicted spectra for the other possible
diastereomers and the experimental spectrum is not often verified.
For chiral compounds with multiple stereochemical elements,
it is often the case that all of the diastereomers, or even more than
a single diastereomer, are not available for experimental investi-
gations. In such cases, it is not possible to experimentally deter-
mine if the diastereomers can even be distinguished from their
chiroptical spectra. Thus, the compounds investigated here
provide a unique opportunity, both experimentally and compu-
tationally, because two of the four diastereomeric compounds
with known absolute configurations, namely HADE and GADE,
are available for experimental investigations, and the remaining two
are their enantiomers, which are expected to give mirror image
chiroptical spectra. Thus, the experimental chiroptical data for all
four diastereomers can be inferred and used to test the predictive,
or discriminating, abilities of chiroptical spectroscopic methods.
We demonstrate in this study, using HADE and GADE as test
cases, that a given chiroptical spectroscopic method by itself may
not be able to unequivocally establish the absolute configurations
of all diastereomers, but a combined analysis of the results from
different chiroptical spectroscopic methods has a greater poten-
tial for a definite assignment of the absolute configurations. The
role of unpolarized vibrational absorption spectra for discrimina-
tion among diastereomers is also analyzed.
When a chiral compound has multiple elements of stereo-
chemistry (multiple stereogenic centers or a combination of
stereogenic centers and helical segments), one has to consider 2n
possible stereoisomers, where n is the total number of stereo-
chemical elements. In such cases, the procedure for determining
or verifying the chiral molecular structures using chiroptical
spectra is dependent on the following implicit criterion: If the
experimentally observed chiroptical spectrum for an enantiomer
of a chiral compound is satisfactorily reproduced by a quantum
chemically predicted chiroptical spectrum for only one of all
possible absolute configurations of that compound, then the
absolute configuration used to calculate that satisfactorily pre-
dicted chiroptical spectrum can be assigned to the enantiomer
used for experimental measurements. However, one cannot claim
5666
dx.doi.org/10.1021/jp202501y |J. Phys. Chem. A 2011, 115, 5665–5673

The Journal of Physical Chemistry A
ARTICLE
’ PROCEDURES
∼1000ꢀ900 cmꢀ1 region has been excluded from presentation
for the spectra measured in CD2Cl2 due to interference from
CD2Cl2 absorption. The VA and VCD spectra are presented,
respectively, as molar extinction and differential molar extinction,
in units of L molꢀ1 cmꢀ1, as a function of wavenumber (in units
of cmꢀ1).
Computational Details. Since there are two stereogenic
centers (each connected to a COOCH3 group), four stereo-
isomers are possible for 3, with two of them being mirror images
of the other two. Calculations were performed for (2S,3R)-3 and
(2S,3S)-3. The chiroptical spectra for (2R,3S)-3 and (2R,3R)-3
were then generated from those of (2S,3R)-3 and (2S,3S)-3 by
multiplying the latter by ꢀ1.
Experimental Section. Hibiscus acid, isolated1,10 from fresh
leaves of Hibiscus furcatus/Hibiscus sabdariffa, was converted to
HADE using diazomethane. An ethereal solution of diazo-
methane was prepared according to a literature procedure start-
ing from N-nitroso-N-methylurea (1.6 g).11 To a stirred solution
of hibiscus acid (450 mg, 2.4 mmol) in methanol (3 mL), cooled
in an ice bath, was added the ethereal diazomethane solution in
small portions.1 The gas evolution was allowed to subside
between additions. The addition of diazomethane was continued
until a faint yellow color persisted, indicating the reaction was
complete, which was confirmed by TLC analysis (1:1 hexanes/
ethyl acetate). The mixture was allowed to stand at room
temperature until the yellow color faded, then the solvents were
removed under reduced pressure. Purification was achieved by
medium pressure liquid chromatography with a Biotage SP1
apparatus (Biotage, Charlottesville, VA) using a 25 ꢁ 75 mm
25þM cartridge with a flow rate of 25 mL/min, employing the
following gradient: 100% methylene chloride for 1 column
volume, 0ꢀ6% ethyl acetate in methylene chloride over 10 column
volumes, 6% ethyl acetate in methylene chloride for 6 column
volumes. HADE (278 mg, 1.3 mmol, 62%) was obtained as a
yellow oil. NMR spectral data12,13 were used to confirm the
product. The isolation and preparation of garcinia acid and
GADE were reported previously.14
The HADE and GADE samples used for experimental mea-
surements will be referred to, respectively, as (þ)-HADE and
(þ)-GADE, where (þ) indicates positive optical rotation (at
589 nm in CH3CN solvent). Since the absolute configurations of
hibiscus acid and garcinia acid are known4 to be, respectively,
(2S,3R) and (2S,3S), those of the corresponding esters will also
be the same; that is, (2S,3R) for (þ)-HADE and (2S,3S) for (þ)-
GADE. Thus, although (2S,3R)-3 is the same as (þ)-HADE and
(2S,3S)-3 is same as (þ)-GADE, we will designate their experi-
mental data as belonging, respectively, to (þ)-HADE and (þ)-
GADE and calculated data for individual stereoisomers as
belonging to (2S,3R)-3, (2S,3S)-3, (2R,3S)-3, or (2R,3R)-3.
The optical rotations at six discrete wavelengths (633, 589,
546, 436, 405, 365 nm) were measured at a concentration of
3 mg/mL in CH3CN and with a 0.5 dm cell using an Autopol IV
polarimeter. The ORD spectra are presented as specific rotations
(in units of deg cc gꢀ1 dmꢀ1)) as a function of wavelength (in
units of nm). The electronic absorption (EA) and ECD spectra
were recorded on a Jasco J720 spectrometer at a concentration of
7 mg/mL in CH3CN using a 0.01 cm path length quartz cell. The
solvent spectrum has been subtracted from the experimental
ECD spectrum of sample solution. The CH2Cl2 absorption pre-
vents ECD measurements below ∼230 nm; therefore, this
solvent was not used for ECD measurements. The EA and
ECD spectra are presented, respectively, as molar extinction
and differential molar extinction, in units of L molꢀ1 cmꢀ1, as a
function of wavelength (in units of nm).
VA and VCD Spectra. For gas-phase VA and VCD calcula-
tions, optimized geometries were obtained using the B3LYP15
functional and 6-31G*16 and aug-cc-pVDZ basis sets.17 VA and
VCD calculations were performed at the respective optimized
geometries and are designated as B3LYP/6-31G* and B3LYP/aug-
cc-pVDZ. For VA and VCD calculations in CD2Cl2, geometries
were reoptimized at the B3LYP/aug-cc-pVDZ level using the
polarizable continuum model (PCM),18 as recently modified19
and implemented in the Gaussian 09 program.20 VCD calculations
usingPCM were performed at the respective optimized geometries
and are designated as B3LYP/aug-cc-pVDZ-PCM. Thus, the VA
and VCD calculations performed at B3LYP/6-31G* and B3LYP/
aug-cc-pVDZ represent gas phase results; those at B3LYP/aug-cc-
pVDZ-PCM represent CH2Cl2 solution phase results. The theo-
retical VA and VCD spectra were simulated with Lorentzian band
shapes of 10 cmꢀ1 half-width at half-peak height. Populations of
conformers derived from Gibbs free energies were used to obtain
the population-weighted VCD spectra.
EA and ECD Spectra. The predicted electronic transition
energies obtained with the B3LYP functional are generally
lower21 than the corresponding experimental values. It has been
shown that this deficiency may be corrected using the CAM-
B3LYP functional,22 which provides a better description of the
tails of electronic density, by providing the correct asymptotic
behavior of the exchange contribution. B3LYP and CAM-B3LYP
functionals have similar accuracy for valence-type excitations,
and CAM-B3LYP is clearly superior for Rydberg-type excita-
tions.21 Therefore, the CAM-B3LYP functional was also used
with the aug-cc-pVDZ basis set for EA and ECD calculations. It is
not necessary to use the fully optimized geometry for EA and
ECD calculations at a given theoretical level, unlike for VCD
calculations. Instead, for EA and ECD calculations at a higher
level (including larger basis sets with diffuse functions), a geo-
metry optimized at a lower level is deemed adequate. For this
reason, the geometries were not reoptimized for CAM-B3LYP
calculations, but those optimized in the B3LYP calculation were
used. To provide a balanced view, it should be added that in using
the CAM-B3LYP functional, although a larger exact-exchange
contribution eliminates the ghost states,23 the range-separated
contribution may or may not work favorably,24 depending on the
size of the molecule (compared with the length scale that
separates the short and long ranges). Calculations performed
for EA and ECD are designated as B3LYP/6-31G*, B3LYP/aug-
cc-pVDZ, and CAM-B3LYP/aug-cc-pVDZ to represent gas
phase results and as B3LYP/aug-cc-pVDZ-PCM and CAM-
B3LYP/aug-cc-pVDZ-PCM to represent CH3CN solution
phase results. The theoretical EA and ECD spectra were simu-
lated from the first 25 singlet f singlet electronic transitions
using Gaussian band shapes and 15 nm half-width at 1/e of peak
The vibrational absorption (VA) and VCD spectra were
recorded in the 2000ꢀ900 cmꢀ1 region using a ChiralIR Fourier
transform VCD spectrometer (BioTools, USA). The VCD spectra
were recorded with 1 h data collection time at 8 cmꢀ1 resolution.
VA and VCD spectra were measured in CD2Cl2 at a concentra-
tion of 1.8 mg/100 μL. The samples were held in a 100 μm fixed
path length cell with BaF2 windows. In the VA spectrum, the
solvent absorption was subtracted. Similarly, the solvent VCD
spectrum was subtracted from that of the sample solution. The
5667
dx.doi.org/10.1021/jp202501y |J. Phys. Chem. A 2011, 115, 5665–5673

The Journal of Physical Chemistry A
ARTICLE
Table 1. Dihedral Anglesa of Optimized Conformers of (2S,3R)-3 (HADE) Ring Puckering
B3LYP populations from internal energiesb
6-31G*
aug-cc-pVDZ
ring-puckering angle
c
conformer
(OringꢀCring(=O)ꢀCringꢀCring
)
OdCꢀCringꢀOring
OdCꢀCringꢀOH
gas
gas
CH3CNcc
CH2Cl2 c
1
2
3
4
5
6
7
ꢀ19.2
15.2
175
ꢀ36.5
ꢀ153
139
ꢀ0.47
12.7
0.36
0.25
0.16
0.15
0.07
0.01
0.00
0.37
0.26
0.12
0.18
0.07
0.01
0.00
0.25
0.13
0.03
0.05
0.51
0.00
0.04
0.30
0.16
0.04
0.08
0.39
0.00
0.03
ꢀ19.6
14
178
11.9
ꢀ17.9
12.4
ꢀ3.4
151
ꢀ0.6
ꢀ144
ꢀ177
ꢀ17.9
7.2
a Dihedral angles provided here are those obtained from geometry optimizations at the B3LYP/6-31G* level. b Energies were obtained from reoptimized
geometries with each basis set. c Using PCM.
height. Rotational strength values, calculated in the velocity
representation, have been used for ECD spectral simulations.
The EA spectral intensities have been derived from dimensionless
oscillator strengths. Populations of conformers derived from the
Gibbs free energies were used to obtain the population-weighted
ECD spectra in the gas phase, and those derived from internal
energies of PCM calculation using CH3CN as solvent were used to
obtain the population-weighted solution phase ECD spectra.
ORD Spectra. The functionals and basis sets used for ORD
calculation are identical to those used for ECD calculations.
Optical rotations at six different wavelengths were calculated at
the B3LYP/6-31G*, B3LYP/aug-cc-pVDZ, and CAM-B3LYP/
aug-cc-pVDZ levels to represent gas phase results; and at the
B3LYP/aug-cc-pVDZ-PCM and CAM-B3LYP/aug-cc-pVDZ-
PCM levels to represent solution phase results.
search programs as described below. (2) Conformational search
using SPARTAN program.25 When default options are used,26
this program searches for conformations by changing the tor-
sional angles of rotatable bonds, but does not look for conforma-
tions that differ in the signs of ring-puckering torsional angles27
in the conformational search. As a consequence, the inversion of
the ring-puckering angle (that is, a change in the sign of the ring-
puckering torsional angle) could not be realized. For this reason,
one geometry with a positive puckering torsional angle and a
second geometry with a negative puckering torsional angle were
separately used as starting geometries to find low-energy con-
formers within 20 kcal/mol energy range. This procedure resul-
ted in 11 conformations with positive puckering torsional angles
and 13 conformations with negative puckering torsional angles.
These 24 conformations were further optimized at the B3LYP/6-
31G* level, and seven lowest energy conformations (see Table 1)
were selected for further consideration. (3) Conformational
search using the CONFLEX program.28 This program by default
looks for conformations that differ in the signs of ring-puckering
torsional angles.29 Three different starting geometries of HADE,
each with a different positive or negative puckering torsional
angle, were used for the conformational search within a 20 kcal/
mol energy range. All three cases yielded the same set of 19
conformations. These 19 conformations were optimized at the
B3LYP/6-31G* level, which led to the same seven lowest energy
conformations that are included in Table 1. For solution phase
predictions, the low-energy conformers found in the gas phase
were reoptimized using PCM.19,20
It is important to note that there are significant differences in
populations of gas-phase and solution-phase conformers. For
example, at the B3LYP/aug-cc-pVDZ level, conformer 5 has a
population of 7% in gas phase and 51% in CH3CN. This observa-
tion emphasizes the importance of using PCM,19,20 as noted
previously,14 to interpret the solution phase chiroptical data.
(2S,3S)-3. The conformers of GADE were determined previ-
ously14 using a manual search method. Four different structures
with positive puckering angles were built and optimized at the
B3LYP/631G* level. The resulting geometries were reoptimized
at the B3LYP/aug-cc-pVDZ level, then the puckering angle in
each of the conformers was manually changed to a negative value
and reoptimized, leading to a total of eight conformers in the gas
phase. Among these eight conformers, six conformers are within
2 kcal/mol. For solution phase predictions, the low-energy conformers
VCD and ORD spectral calculations were performed using
gauge-invariant atomic orbitals (GIAOs), and ECD spectral
calculations were performed without using the GIAOs because
ECD was calculated using the velocity formalism, which is origin-
independent.
’ RESULTS AND DISCUSSION
Conformational Analysis. Finding all possible low-energy
conformations for cyclic ring molecules is a nontrivial task. For
five-member ring molecules, in particular, it is necessary to
undertake conformational analysis carefully.
(2S,3R)-3. We used three different procedures to identify the
low-energy conformers: (1) Manual search method. We built the
plausible structures and began the geometry optimizations using
the Gaussian 09 program20 at the STO-2G level. To obtain all
possible conformations, the positive puckering angles of four
different optimized conformations were changed to negative
values, and eight geometries were reoptimized at the B3LYP/6-
31G* level. However, the negative puckering values reverted back
to the positive puckering angles during optimization. Therefore,
this type of conformational search required maintaining the ring-
puckering value as a constant (or performing a potential energy
scan by varying the rest of the structural parameters) or
determining the remaining plausible conformations by other
means (see below). In the end, we arrived at the conformations
listed in Table 1. To ensure that all possible conformations
had been identified, we then used two different conformational
5668
dx.doi.org/10.1021/jp202501y |J. Phys. Chem. A 2011, 115, 5665–5673

The Journal of Physical Chemistry A
ARTICLE
Table 2. Absolute Configurations Inferred from Individual
Spectroscopic Methods
method
(þ)-HADEa
(þ)-GADEa
ECD
ORD
VCD
VA
(2S,3R)
(2S,3S)
(2S,3S)
(2R,3R)
(2R,3S)
(2S,3R)
(2S,3S)
(2S,3S)
(2S,3S)
(2S,3S)
(2S,3R)
(2S,3R)
(2S,3R)
(2S,3R)
(2R,3S)
(2R,3R)
a Bold entries represent common configrations predicted by all spectro-
scopic methods.
CAM-B3LYP/aug-cc-pVDZ-PCM levels display one well-
defined positive ECD band in the 190ꢀ250 nm region. In the
predicted spectrum at the CAM-B3LYP/aug-cc-pVDZ level (see
Supporting Information), this positive ECD band occurs below
190 nm, and only the long wavelength tail of this positive
ECD band can be seen in the 190ꢀ250 nm region. Thus, the
sign of the predicted ECD band in the 190ꢀ250 nm region for
(2S,3R)-3 concurs with experimental observation for (þ)-HADE
at all levels of calculations undertaken. Quantum chemical calcula-
tions of EA and ECD spectra for (2S,3S)-3 were reported,14 and
the importanceof usingthe recentimplementation19,20 of PCM for
predicting the correct ECD sign was discussed. The predicted
population-weighted EA and ECD spectra for (2S,3S)-3 at the
CAM-B3LYP/aug-cc-pVDZ-PCM level are also shown in Figure 1.
In summarizing this section, the predicted ECD spectra of
(2S,3R)-3 and (2S,3S)-3 in the 190ꢀ250 nm region are similar,
as are the experimental ECD spectra of (þ)-HADE and (þ)-
GADE. Except for minor differences in band positions and peak
intensities, these ECD spectra are not very useful for differentiat-
ing between the (2S,3R)-3 and (2S,3S)-3 diastereomers. Since
the enantiomers, (2R,3S)-3 and (2R,3R)-3, would have ECD
spectra with signs opposite to those of (2S,3R)-3 and (2S,3S)-3,
respectively, ECD spectra are also not useful for differentiating
between the (2R,3S)-3 and (2R,3R)-3 diastereomers.
The EA spectra of enantiomers will be the same, but those of
diastereomers need not be the same. However, the predicted EA
spectra of (2S,3R)-3 and (2S,3S)-3 diastereomers, as also seen in
the experimental spectra of (þ)-HADE and (þ)-GADE, are
similar. Therefore, the EA spectra in the 190ꢀ250 nm region are
also not useful for differentiating between (2S,3R)-3 and (2S,3S)-
3 diastereomers. Therefore, if the absolute configurations of (þ)-
HADE and (þ)-GADE were not known, then on the basis of
quantum chemical calculations of ECD, one could associate
either (2S,3R) or (2S,3S) configurations to each of them (as
summarized in Table 2). As a consequence, ECD spectra (either
alone or in combination with EA spectra) in the 190ꢀ250 nm
region cannot unequivocally determine the absolute configura-
tions of diasteromers of 3.
Figure 1. Experimental EA (bottom) and ECD (top) spectra for
dimethylesters of hibiscus acid (HADE) and garcinia acid (GADE) in
CH3CN and population-weighted corresponding spectra obtained at
the CAM-B3LYP/aug-cc-pVDZ-PCM level for (2S,3R)-3 and (2S,3S)-3
diastereomers.
found in the gas phase were reoptimized using PCM. In addition
to this method, we also used a conformational search using the
CONFLEX program28 (with the same options29 as used for
HADE). The conformational search for GADE yielded 13
conformations. These 13 conformations were further optimized
at the B3LYP/6-31G* level, which led to the same six low-energy
conformers14 as found in the above-mentioned manual search
method.
In summarizing the conformational analysis section, the use of
multiple procedures that use different criteria for a conforma-
tional search, and augmenting a conformational search with
potential energy scans, provides confidence that an employed
conformational search procedure may not have missed any low-
energy conformations. Therefore, this multiprong approach
appears suitable to being adopted as a part of the standard
protocol for conformational analysis.
It is to be cautioned that, even though the conformational
search description given above for HADE and GADE may appear
to favor the use of the CONFLEX program,28 we found this
program not to be suitable for conducting the conformational
search for salts of carboxylic acids. The conformational analysis of
chiral carboxylic acid salts will be communicated in the near future.
EA and ECD Spectra. The experimental ECD spectra for (þ)-
HADE and (þ)-GADE in the 190ꢀ250 nm region are shown in
Figure 1. The experimental EA spectra of (þ)-HADE and (þ)-
GADE are similar in the 190ꢀ250 nm region: both have a weak
and broad absorption band spanning the 200ꢀ230 nm region. In
the ECD spectra of (þ)-HADE and (þ)-GADE, there is only
one positive ECD band each (at 216 nm for HADE and at
212 nm for GADE). The magnitude of observed Δε is larger for
(þ)-HADE than that for (þ)-GADE.
ORD Spectra. The experimental ORD spectra for (þ)-HADE
and (þ)-GADE are shown in Figure 2. The optical rotations are
positive for both HADE and GADE at all six wavelengths
examined. The differences among them are only in magnitude,
with specific rotations of HADE being slightly larger than those
of GADE. These experimental data indicate that experimental
ORD is not very useful to differentiate between the (2S,3R) and
(2S,3S) diastereomers (and also between the corresponding
enantiomers).
The predicted population-weighted EA and ECD spectra
for (2S,3R)-3 at the CAM-B3LYP/aug-cc-pVDZ-PCM level
are shown in Figure 1, and the rest are provided as the Sup-
porting Information. The predicted ECD spectra at the B3LYP/
631G*, B3LYP/aug-cc-pVDZ, B3LYP/aug-cc-pVDZ-PCM, and
The predicted population-weighted ORD spectra for (2S,3R)-
3 at the CAM-B3LYP/aug-cc-pVDZ-PCM level are shown in
Figure 2, and the rest are provided as Supporting Information.
5669
dx.doi.org/10.1021/jp202501y |J. Phys. Chem. A 2011, 115, 5665–5673

The Journal of Physical Chemistry A
ARTICLE
Figure 2. Experimental ORD spectra for (þ)-HADE and (þ)-GADE
in CH3CN and population-weighted ORD spectra obtained at the
CAM-B3LYP/aug-cc-pVDZ-PCM level for (2S,3R)-3 and (2S,3S)-3
diastereomers.
The predicted ORD spectra at all theoretical levels display
positive rotations at all six wavelengths calculated. Thus, the
signs of predicted rotations for (2S,3R)-3 at all levels of
calculations undertaken concur with experimental observa-
tion for (þ)-HADE. The same observation holds for (þ)-
GADE; quantum chemical calculations of ORD spectra for
(2S,3S)-3 were reported and discussed previously.14 The
predicted population-weighted ORD spectrum for (2S,3S)-3
at the CAM-B3LYP/aug-cc-pVDZ-PCM level is shown in
Figure 2.
In summarizing this section, the predicted ORD spectra of
(2S,3R)-3 and (2S,3S)-3 in the 365ꢀ633 nm region are similar,
as are the experimental ORD spectra of (þ)-HADE and (þ)-
GADE. Except for minor differences in the magnitude of the
specific rotations, ORD spectra are not useful for differentiating
between the (2S,3R)-3 and (2S,3S)-3 diastereomers. Since
(2R,3S)-3 and (2R,3R)-3 are enantiomeric to (2S,3R)-3 and
(2S,3S)-3, respectively, their ORD spectra would have opposite
signs and are not useful for differentiating between the (2R,3S)-3
and (2R,3R)-3 diastereomers. Therefore, if the absolute config-
urations of (þ)-HADE and (þ)-GADE were not known, then
on the basis of the quantum chemical calculations of ORD, in
the 365ꢀ633 nm region, one could associate either (2S,3R) or
(2S,3S) configurations to each of them (as summarized in
Table 2). As a consequence, ORD spectra in the 365ꢀ633 nm
region, by themselves, cannot unequivocally determine the
absolute configurations of diasteromers of 3.
VA and VCD Spectra. The experimental VA and VCD spectra
for (þ)-HADE and (þ)-GADE in the 1900ꢀ1000 cmꢀ1 region
are shown in Figure 3. Unlike EA spectra, the experimental VA
spectra of (þ)-HADE and (þ)-GADE are quite different.
Similarly, the VCD spectra of (þ)-HADE and (þ)-GADE are
also quite different, which is in contrast to the ECD spectra. The
experimental VCD bands originating from the CdO stretching
modes (1800ꢀ1700 cmꢀ1 region) have opposite signs for (þ)-
HADE and (þ)-GADE (although the magnitudes of VCD
intensities in (þ)-GADE are much smaller). Similarly, the
experimental VCD bands at ∼1300 and 1254 cmꢀ1 have
opposite signs in (þ)-HADE and (þ)-GADE, as are the signs
of the VCD bands in the 1120ꢀ1090 cmꢀ1 region. Thus,
discrimination between (þ)-HADE and (þ)-GADE (hence,
between the diastereomers (2S,3R)-3 and (2S,3S)-3) can be
realized from experimental VCD spectra, unlike the experimental
ECD and ORD spectra. However, the discrimination between
two of the diastereomers, (2S,3R)-3 and (2S,3S)-3, alone is not
sufficient to yield unequivocal absolute configuration assignment
from VCD because it is also necessary to verify if VCD can
Figure 3. Experimental VA (bottom panel) VCD (top panel) spectra
for (þ)-HADE and (þ)-GADE in CH2Cl2.
discriminate between the diastereomeric pair (2S,3R)-3 and
(2R,3R)-3 and between (2S,3S)-3 and (2R,3S)-3 (vide infra).
The predicted population-weighted VA and VCD spectra for
(2S,3R)-3 at the B3LYP/aug-cc-pVDZ-PCM level are shown in
Figure 4, and those obtained in the B3LYP/aug-cc-pVDZ
calculation are provided as Supporting Information. The pre-
dicted VA and VCD spectra for (2S,3R)-3 at the B3LYP/aug-cc-
pVDZ-PCM level are found to have excellent correlation to
the corresponding experimental spectra of (þ)-HADE (see
Figure 4). This correlation becomes poor at the B3LYP/aug-
cc-pVDZ level (see Supporting Information) because gas phase
predicted vibrational band positions deviate significantly from
the observed band positions in CD2Cl2 solution. This observa-
tion reemphasizes our previous observation11 that the recent
implementation of PCM19,20 significantly enhances the correla-
tion between solution-phase calculated VA and VCD spectra and
the corresponding experimental spectra in solution.
The excellent correlation, as seen in Figure 4, between the
experimental VA/VCD spectra of (þ)-HADE and calculated
VA/VCD spectra for (2S,3R)-3 may be used to confirm the
known configuration, (þ)-(2S,3R)-HADE. A similar correlation
between experimental VA/VCD spectra of (þ)-GADE and
calculated VA/VCD spectra for (2S,3S)-3 was used recently14
to confirm the known configuration as (þ)-(2S,3S)-GADE.
However, to validate these conclusions, it is necessary to show
that a similar correlation, or even a resemblance of possible corre-
lation, does not exist between the experimental VCD spectra and
those calculated for the two other diastereomers; namely, the
5670
dx.doi.org/10.1021/jp202501y |J. Phys. Chem. A 2011, 115, 5665–5673

The Journal of Physical Chemistry A
ARTICLE
Figure 4. Comparison of experimental VA (bottom panel) and VCD
(top panel) spectra for (þ)-HADE in CH2Cl2 with predicted popula-
tion-weighted spectra at the B3LYP/aug-cc-pVDZ-PCM level for
(2S,3R)-3.
(2R,3S) and (2R,3R) stereoisomers. For this purpose, the
experimental VA/VCD spectra of (þ)-HADE and (þ)-GADE
are compared with those calculated at the B3LYP/aug-cc-pVDZ-
PCM level for all four diastereomers in Figure 5. Note that the
calculated spectra for the (2R,3S) and (2R,3R) configurations
can be obtained by multiplying the calculated spectra for (2S,3R)
and (2S,3S), respectively, by ꢀ1. The correlation between
experimental and calculated VCD bands is displayed using
double-headed arrows in Figure 5. From this correlation, it can
be noted that the experimental VCD spectra of (þ)-HADE can
be correlated to the calculated spectra of not only the (2S,3R)
configuration but also the (2R,3R) configuration. Although the
correlation of the VCD spectrum of (þ)-HADE with that of the
(2R,3R) configuration is not as strong as that with the (2S,3R)
configuration, the correlation shown in Figure 6 cannot be ruled
out, given the uncertainties in quantum chemical calculations for
vibrational band positions and intensities. The uncertainties in
quantum chemical predictions often arise from variations in
predicted energies, populations, band positions and band inten-
sities due to their dependence on the rigor of the quantum
chemical method used. Similarly, the experimental VCD spectra
of (þ)-GADE can be correlated to the calculated spectra of not
only the (2S,3S) configuration but also the (2R,3S) configura-
tion. Here again, one cannot rule out the possible correlation
between the VCD spectra of (þ)-GADE and the (2R,3S) confi-
guration. On the basis of the VCD spectral analysis alone, a safer
conclusion to draw is that there are two possibilities for the
absolute configuration of (þ)-HADE—namely, (2S,3R) and
(2R,3R)—and two possibilities for the absolute configuration
of (þ)-GADE—namely, (2S,3S) and (2R,3S). These possibili-
ties can be narrowed into unequivocal assignments using the
analysis of multiple spectral methods as discussed below.
Figure 5. Comparison of experimental VA (bottom panel) and VCD
(top panel) spectra for (þ)-HADE and (þ)-GADE in CH2Cl2 with
predicted population-weighted spectra for all four diastereomers
((2S,3R)-3, (2S,3S)-3, (2R,3S)-3, and (2R,3R)-3) at the B3LYP/aug-
cc-pVDZ-PCM level.
(þ)-HADE were not known, then the uncertainty in assigning
(2S,3R) or (2R,3R) configuration to (þ)-HADE on the basis of
VCD results alone could be eliminated by noting that the
(2R,3R) configuration is not applicable to (þ)-HADE because
(þ)-HADE exhibits positive ECD in the 190ꢀ250 nm region,
which is opposite to that predicted for the (2R,3R) configuration.
Similarly, if the absolute configuration of (þ)-GADE were not
known, the uncertainty in assigning (2S,3S) or (2R,3S) config-
urations to (þ)-GADE on the basis of VCD results alone could
be eliminated by noting that the (2R,3S) configuration is not
applicable to (þ)-GADE because (þ)-GADE exhibits positive
ECD in the 190ꢀ250 nm region, which is opposite to that
predicted for the (2R,3S) configuration. Thus, a combined
analysis of ECD and VCD results yields a (2S,3R) configuration
for (þ)-HADE and a (2S,3S) configuration for (þ)-GADE,
providing a better approach for assigning the absolute config-
urations of diastereomers (see Table 2). (b) Combined analysis
of ORD and VCD: Just as in the combined analysis of ECD and
VCD, the (2R,3R) configuration is not applicable to (þ)-HADE
because (þ)-HADE exhibits positive ORD in the 365ꢀ633 nm
region, which is opposite in sign to that predicted for the (2R,3R)
configuration. Similarly, the (2R,3S) configuration is not appli-
cable to (þ)-GADE because (þ)-GADE exhibits positive ORD
Combined Analysis of Spectral Methods. The inability of a
given chiroptical method to determine the absolute configura-
tions of stereoisomers, as noted above, can be overcome if
combined analyses of these methods are undertaken. (a) Com-
bined analysis of ECD and VCD: If the absolute configuration of
5671
dx.doi.org/10.1021/jp202501y |J. Phys. Chem. A 2011, 115, 5665–5673

The Journal of Physical Chemistry A
ARTICLE
(in the 365ꢀ633 nm region), which is opposite in sign to that
predicted for the (2R,3S) configuration. Thus, a combined
analysis of ORD and VCD results yields a (2S,3R) configuration
for (þ)-HADE and a (2S,3S) configuration for (þ)-GADE,
providing a better approach for assigning the absolute config-
urations of diastereomers. (c) Special role of unpolarized VA
spectra: Even though the unpolarized VA spectra of enantiomers
are identical, those of diastereomers need not be. This is evident
in Figures 3 and 5. The VA spectra predicted for the enantiomers
(2S,3S) and (2R,3R) match well with the experimental VA
spectrum of (þ)-GADE. Similarly, the VA spectra predicted
for the enantiomers (2S,3R) and (2R,3S) match well with the
experimental VA spectrum of (þ)-HADE, but the predicted VA
spectra of diastereomers (2S,3S) and (2S,3R) can be seen in
Figure 5 to be significantly different, as also seen in Figure 4 for
the experimental VA spectra of (þ)-HADE and (þ)-GADE.
Thus, VA spectra do provide discrimination among diastereo-
mers (2S,3S) and (2S,3R) or among their enantiomers (2R,3R)
and (2R,3S). Then the ambiguities noted during the assignment
of absolute configurations of diastereomers based on VCD alone
can be avoided (Table 2) if criteria for matching the VA spectral
pattern are also imposed along with matching of VCD band signs.
Similar analysis by combining ECD and VA results or ORD
and VA results can also be used to discriminate between diaste-
reomers (see Table 2). Thus, we recommend the use of VA
spectra, in combination with chiroptical spectra, for differentia-
tion between the diastereomers of chiral compounds.
properties when analyzed with a single chiroptical method.
Combined analyses of different chiroptical spectroscopic meth-
ods may eliminate the ambiguities in assigning the absolute
configurations of diastereomers. It has also been noted that some
ambiguities in the assignment of absolute configurations of
diasteromers using chiroptical methods can be avoided if the
spectral analysis also includes unpolarized VA spectra.
’ ASSOCIATED CONTENT
S
Supporting Information.
EA and ECD spectra of
b
(2S,3R)-HADE at the B3LYP/6-31G*, B3LYP/aug-cc-pVDZ,
CAM-B3LYP/aug-cc-pVDZ, and B3LYP/aug-cc-pVDZ-PCM
levels; ORD spectra of (2S,3R)-HADE at the B3LYP/6-31G*,
B3LYP/aug-cc-pVDZ, CAM-B3LYP/aug-cc-pVDZ, and B3LYP/
aug-cc-pVDZ-PCM levels; VA and VCD spectra of (2S,3R)-
HADE at the B3LYP/aug-cc-pVDZ level. This material is available
free of charge via the Internet at http://pubs.acs.org
’ AUTHOR INFORMATION
Corresponding Author
*Phone: (615)322-2836. Fax: (615)322-4936. E-mail: Prasad.L.
Polavarapu@vanderbilt.edu.
’ ACKNOWLEDGMENT
It should be noted that the combined analysis of ECD and
ORD spectra, however, would not have provided diastereomeric
discrimination (see Table 2) for the molecules considered here,
which is not surprising, since ECD and ORD are not indepen-
dent methods and are related through KramersꢀKronig trans-
form. Nevertheless, simultaneous investigation of ECD and
ORD is recommended for various other reasons.30
P.L.P gratefully acknowledges NSF funding (CHE-0804301)
and helpful communications with Dr. Phil Klunzinger of Wave-
functions Inc., in the use of the SPARTAN program. E.A.D was
supported by the Vanderbilt NSF-REU program (CHE
0850976), and E.K.H. was supported by predoctoral traineeships
(T32 ES07028). I.I. and D.H. acknowledge the Department of
Science and Technology, Government of India, New Delhi, for
financial assistance (Project No. SR/S1/OC/54-2007). This
work was conducted in part using the resources of the Advanced
Computing Center for Research and Education at Vanderbilt
University, Nashville, TN.
If the absolute configurations of (þ)-HADE and (þ)-GADE
were not known, then the possible absolute configurations that
could be inferred for these two compounds from ECD, ORD,
VCD, and VA are summarized in Table 2. It can be seen that the
only absolute configuration that is predicted commonly by all of
the methods for (þ)-HADE is (2S,3R), and that for (þ)-GADE
is (2S,3S). Thus, combined analyses of multiple chiroptical
spectroscopic methods can eliminate the uncertainties in assign-
ing the absolute configurations of diastereomers. These observa-
tions support our previously made recommendation30 for the use
of more than one chiroptical spectroscopic method for chiral
molecular structural determination. In addition, we conclude
from the present study that the unpolarized VA spectra can also
be quite helpful when uncertainties in the assignment of absolute
configurations of diastereomers are not resolved from chiroptical
spectral analysis. Although the Raman (and ROA) spectra are not
investigated here, it is reasonable to estimate that an advantage
similar to that demonstrated here with VA spectra can be realized
from the combined analysis of the Raman and chiroptical spectra
of diastereomers.
’ REFERENCES
(1) Ibnusaud, I; Thomas, P. T.; Rani, R. N.; Sasi, P. V.; Beena, T.;
Hisham, A. Tetrahedron 2002, 58, 4887–4892.
(2) (a) Lewis, Y. S.; Neelakantan, S.; Anjanamurthy Curr. Sci. 1964,
3, 82–83. (b) Lewis, Y. S.; Neelakantan, S. Phytochemistry 1965,
4, 619–625.
(3) Jena, B. S.; Jayaprakasha, G. K.; Singh, R. P.; Sakariah, K. K.
J. Agric. Food Chem. 2002, 50, 10–22.
(4) Boll, P. M.; Sorensen, E.; Balieu, E. Acta Chem. Scand. 1969,
23, 286–293. Glusker, J. P.; Minkin, J. A.; Soule, F. B. Acta Crystallogr.
1972, B28, 2499–2505.
(5) Autschbach, J. Chirality 2009, 21, E116–E152.
(6) Wilson, S. M.; Wiberg, K. B.; Murphy, M. J.; Vaccaro, P. H.
Chirality 2008, 20, 357–369.
(7) (a) Devlin, F. J.; Stephens, P. J.; Cheeseman, J. R.; Frisch, M. J.
J. Am. Chem. Soc. 1996, 118, 6327–6328. (b) Freedman, T. B.; Cao, X;
Dukor, R. K.; Nafie, L. A. Chirality 2003, 15, 743–758.
(8) (a) Barron, L. D.; Buckingham, A. D. Chem. Phys. Lett. 2010,
492, 199–213. (b) Ruud, K.; Thorvaldsen, A. J. Chirality 2009, 21,
E54–E67.
(9) For a recent review on VCD applications for natural
products, see: Polavarapu, P. L. In Comprehensive Chiroptical Spectros-
copy; Berova, N., Polavarapu, P. L., Nakanishi, K., Woody, R. W., Eds.;
John Wiley, New York, 2011.
’ CONCLUSIONS
Using chioptical spectroscopic investigations on dimethyl
esters of hibiscus acid and garcinia acid, it has been noted that
a single chiroptical spectroscopic method by itself may not be
able to determine the absolute configurations of diastereomers.
This is because more than one diastereomer can have similar
5672
dx.doi.org/10.1021/jp202501y |J. Phys. Chem. A 2011, 115, 5665–5673

The Journal of Physical Chemistry A
ARTICLE
(10) The concentrated aqueous extract was washed with hexane. To
the aqueous layer, acetone was added to precipitate the insoluble
materials. After the evaporation of acetone, NaOH solution (8 N) was
added to adjust the pH to 12. Addition of alcohol resulted in the
precipitation of the sodium salt of the acid. The pH was further adjusted
to 2.0 with HCl (2 N), followed by concentration, which leads to a
syrupy mass that was extracted with dry acetone. The acetone extract was
concentrated. The thick syrup triturated was further extracted with ether
to yield the lactone in the pure form.
(28) CONFLEX Program; CONFLEX Corporation: AIOS Meguro
6F, 2-15-19, Kami-Osaki, Shinagawa-ku, Tokyo 141ꢀ0021, Japan; www.
conflex.us.
(29) The options used with the CONFLEX program are MMFF94S
force field, SEL=20.0, SEARCH=ENERGY.
(30) Polavarapu, P. L. Chirality 2008, 20, 664–672.
(11) Arndt, F. Org. Synth 1943, Coll. vol. II, 165–167.
(12) Yoshihara, T.; Ichihara, A.; Nuibe, H.; Sakamura, S.; Sugita, M.;
Imamoto, S.; Senoh, S. Agric. Biol. Chem. 1974, 38, 121–126.
1
(13) NMR data of synthesized HADE are as follows: H NMR12
(600 MHz, DMSO-d6) δH: 6.80 (1H, s), 5.36 (1H, s), 3.76 (3H, s),
3.68 (3H, s), 3.21 (1H, d, J = 17.4 Hz), 2.68 (1H, d, J = 17.4 Hz). 13
C
NMR (150 MHz, DMSO-d6) δC: 172.6, 170.7, 166.0, 81.8, 77.8, 53.0,
52.3, 41.4.
(14) Polavarapu, P. L.; Scalmani, G.; Hawkins, E. K.; Rizzo, C.;
Jeirath, N.; Ibnusaud, I.; Habel, D.; Nair, D. S.; Haleema, S. J. Nat. Prod.
2011, 74, 321–328.
(15) Becke, A. D. J. Chem. Phys. 1993, 98, 5648–5652.
(16) Ditchfield, R.; Hehre, W. J.; Pople, J. A. J. Chem. Phys. 1971,
54, 724–728.
(17) Kendall, R. A.; Dunning, T. H., Jr.; Harrison, R. J. J. Chem. Phys.
1992, 96, 6796–6806.
(18) Tomasi, J.; Mennucci, B.; Cammi, R. Chem. Rev. 2005,
105, 2999–3094. Cossi, M.; Scalmani, G.; Rega, N.; Barone, V. J. Chem.
Phys. 2002, 117, 43–54.
(19) Scalmani, G.; Frisch, M. J. J. Chem. Phys. 2010, 132, 114110–
114124.
(20) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.;
Robb, M. A.; Cheeseman, J. R.; Scalmani, G.; Barone, V.; Mennucci, B.;
Petersson, G. A.; Nakatsuji, H.; Caricato, M.; Li, X.; Hratchian, H. P.;
Izmaylov, A. F.; Bloino, J.; Zheng, G.; Sonnenberg, J. L.; Hada, M.;
Ehara, M.; Toyota, K.; Fukuda, R.; Hasegawa, J.; Ishida, M.; Nakajima,
T.; Honda, Y.; Kitao, O.; Nakai, H.; Vreven, T.; Montgomery, J. A., Jr.;
Peralta, J. E.; Ogliaro, F.; Bearpark, M.; Heyd, J. J.; Brothers, E.; Kudin,
K. N.; Staroverov, V. N.; Kobayashi, R.; Normand, J.; Raghavachari, K.;
Rendell, A.; Burant, J. C.; Iyengar, S. S.; Tomasi, J.; Cossi, M.; Rega, N.;
Millam, J. M.; Klene, M.; Knox, J. E.; Cross, J. B.; Bakken, V.; Adamo, C.;
Jaramillo, J.; Gomperts, R.; Stratmann, R. E.; Yazyev, O.; Austin, A. J.;
Cammi, R.; Pomelli, C.; Ochterski, J. W.; Martin, R. L.; Morokuma, K.;
Zakrzewski, V. G.; Voth, G. A.; Salvador, P.; Dannenberg, J. J.; Dapprich,
€
S.; Daniels, A. D.; Farkas O.; Foresman, J. B.; Ortiz, J. V.; Cioslowski, J.;
Fox, D. J. Gaussian 09; Gaussian, Inc.: Wallingford, CT, 2009.
(21) Caricato, M.; Trucks, G. W.; Frisch, M. J.; Wiberg, K. B. J. Chem.
Theor. Comput. 2010, 6, 370–383.
(22) Yanai, T.; Tew, D. P.; Handy, N. C. Chem. Phys. Lett. 2004,
393, 51–57.
(23) Georgik, L.; Kruse, H.; Grimme, S. In Comprehensive Chiroptical
Spectroscopy; Berova, N.; Polavarapu, P. L.; Nakanishi, K.; Woody, R. W.
Eds.; John Wiley: New York, 2011.
(24) Rohrdanz, M. A.; Herbert, J. M. J. Chem. Phys. 2008,
129, 034107.
(25) SPARTAN ’08, Wavefunction, Inc., Irvine, CA; www.wavefun.
com/products/spartan.html,
(26) Options used with SPARTAN program are CONF_SELEC-
TION_RULE=5, WINDOW=20.0, SEARCHMETHOD=THOROUGH,
MMFF94 force field.
(27) Dr. Phil Klunzinger of Wavefunction, Inc. advised that a trick
that is sometimes useful with the SPARTAN program to find twisted
boat conformations is to select all “waggable” carbon atoms (sp3 carbons
not attached to the sp3 centers) and choose a parameter 3 (in the torsion
window) for them. This procedure did find conformations with both
positive and negative puckering angles for HADE, but not for some
other five-membered ring systems. Several literature papers that dealt
with five-membered ring systems using the SPARTAN program did not
seem to have made use of this option.
5673
dx.doi.org/10.1021/jp202501y |J. Phys. Chem. A 2011, 115, 5665–5673