Solvent-induced reversed stereoselectivity in reciprocal resolutions of mandelic acid and erythro -2-amino-1,2-diphenylethanol

Article abstract of DOI:10.1021/jo401517q

Solvent-induced chirality switching in reciprocal optical resolution between mandelic acid (1) and erythro-2-amino-1,2-diphenylethanol (2) has been demonstrated. The stereochemistry of the deposited salts was controlled by changing the crystallization solvent from 1-PrOH or 1-BuOH to 1,4-dioxane. It was revealed from ¹H NMR spectra, thermogravimetric analysis, and X-ray crystallography of the salts that an equimolar amount of the crystallization solvent was incorporated in each diastereomeric salt. On the basis of the crystal structures, it was found that both the hydrogen-bonding ability and the size of the solvent molecule played an important role. Differences in the formed hydrogen-bonding networks (columnar or sheetlike structure) and their packing manner were found to be crucial for the reversed stereoselectivity. Furthermore, pseudopolymorphic salt crystals that incorporated 1,4-dioxane were obtained during the enantioseparation of racemic 2, and their solid-state properties were examined by measurement of their IR spectra. This solvent-induced dual stereocontrol technique was successfully applied to the successive resolution process, eliminating the need to change the resolving agent for access to both enantiomers of 1 and 2.

Full text of DOI:10.1021/jo401517q

Article
pubs.acs.org/joc
Solvent-Induced Reversed Stereoselectivity in Reciprocal Resolutions
of Mandelic Acid and erythro-2-Amino-1,2-diphenylethanol
Hiroaki Shitara,† Toshiki Shintani,‡ Koichi Kodama,§ and Takuji Hirose*,§
§
†Technical Support Center and Graduate School of Science and Engineering, Saitama University, Shimo-Ohkubo, Sakura-ku,
Saitama City, Saitama 338-8570, Japan
‡Saitama Prefectural Koshigayakita Senior High School, Ohdomari, Koshigaya City, Saitama 343-0044, Japan
S
* Supporting Information
ABSTRACT: Solvent-induced chirality switching in reciprocal
optical resolution between mandelic acid (1) and erythro-2-
amino-1,2-diphenylethanol (2) has been demonstrated. The
stereochemistry of the deposited salts was controlled by
changing the crystallization solvent from 1-PrOH or 1-BuOH
to 1,4-dioxane. It was revealed from 1H NMR spectra,
thermogravimetric analysis, and X-ray crystallography of the
salts that an equimolar amount of the crystallization solvent
was incorporated in each diastereomeric salt. On the basis of
the crystal structures, it was found that both the hydrogen-
bonding ability and the size of the solvent molecule played an important role. Differences in the formed hydrogen-bonding
networks (columnar or sheetlike structure) and their packing manner were found to be crucial for the reversed stereoselectivity.
Furthermore, pseudopolymorphic salt crystals that incorporated 1,4-dioxane were obtained during the enantioseparation of
racemic 2, and their solid-state properties were examined by measurement of their IR spectra. This solvent-induced dual
stereocontrol technique was successfully applied to the successive resolution process, eliminating the need to change the
resolving agent for access to both enantiomers of 1 and 2.
improved by repeated recrystallization, with its subsequent
decomposition affording the pure enantiomer. When the other
enantiomer of a target racemate is sought, it is common to
recover the residual mother liquor and use the antipode of the
initial resolving agent to produce the other less-soluble
diastereomeric salt. These are labor-intensive and time-
consuming processes, making them economically unfavorable.
The solvent has been recognized as being one of the crucial
factors involved in the process of optical resolution. The
particular solvent used affects the incidence and efficiency of
the crystallization, and even the stereochemistry, of the less-
soluble diastereomeric salt. A number of reports have been
published that demonstrate the important role that solvents
play in the isolation of both enantiomers.4 The key role of the
resolving solvent is mostly regarded to be its ability to solubilize
the two diastereomeric salts to differing degrees. On the other
hand, the more qualitative feature of its ability to reverse the
chirality, has been generally overlooked to date.2,5
INTRODUCTION
■
Enantiopure compounds are extremely valuable in both
academic and industrial fields as they are key intermediates in
the synthesis of functional compounds and materials. In the
pharmaceutical industry in particular, the demand for single
enantiomers has been steadily increasing over the last two
decades.1 One commonly used method for obtaining
enantiopure acids and amines is optical resolution via
diastereomeric salt formation with a chiral resolving agent.2
Although it is many years since this technique was first
described, it remains the dominant approach in manufacturing
pharmaceutical products that contain chiral centers. This
methodology has the advantage of operational simplicity with
standard equipment and facilities. Recently, a modified version,
termed Dutch resolution, was reported to give improved
results, working by the application of a mixture of a small
number of structurally related compounds as resolving agents.3
However, a major issue associated with diastereomeric
resolution is achieving the optimal conditions, with selection
of a suitable resolving agent being the most significant factor. A
screening process is generally required in order to identify
potential resolving agents that crystallize to form a salt with a
target racemate in a specific solvent. Further optimization to
achieve preferential crystallization of one diastereomeric salt
(the less-soluble salt) is then necessary, with conditions such as
solvent, crystallization temperature, and amount of resolving
agent needing to be screened. The purity of this salt can be
In 2004, Sakai et al. recognized the significance of such a
solvent effect and proposed a strategic method called
dielectrically controlled resolution (DCR). By employing this
technique, both enantiomers of a target molecule could be
obtained using a single enantiomer of the resolving agent under
controlling the permittivity of the solvent.6a,b This method-
Received: July 12, 2013
© XXXX American Chemical Society
A
dx.doi.org/10.1021/jo401517q | J. Org. Chem. XXXX, XXX, XXX−XXX

The Journal of Organic Chemistry
Article
Table 1. Diastereomeric Resolution of ( )-1 with Enantiopure (−)-2
1
a
b
c
d
entry
solvent
ε
yield (%)
[α]D (deg)
OP (%ee)
resolution efficiency
less-soluble salt
1
2
3
4
EtOH
24
22
17
2
12
54
94
57
95.2
−139.6
−112.0
135.8
62
90
73
88
7
(+)-1·(−)-2
(−)-1·(−)-2
(−)-1·(−)-2
(+)-1·(−)-2
1-PrOH
1-BuOH
1,4-dioxane
48
68
50
a
b
Yield of the less-soluble diastereomeric salt. See the Experimental Section. Specific rotation values of resolved 1. Solvent: H2O, c 0.8, 20 °C.
Optical purity calculated from the specific rotation value based on the literature.13 [α]D 153.5° (c 1, H2O). Resolution efficiency = yield (%) ×
c
d
OP (%ee)/100.
ology saves a great deal of time and effort as it removes the
need to change the resolving agent to obtain the second target
molecule, and it is especially attractive in the cases where only
one enantiomer is available as a resolving agent because owing
to their derivation from natural chiral sources.6 A key factor of
the DCR method lies in control of the hydration of one of the
diastereomeric salt pair, aided by hydrogen-bond formation.
This methodology was particularly interesting for its
uniqueness, in addition to its expected merit in industrial
applications, and prompted us to further investigate novel
systems. In our previous studies, as well as a hydrogen-bond
forming effect,7a a space-filling effect7b,c of organic solvent
molecules was demonstrated. In the resolution of α-
methylbenzylamine (MBA) with N-(p-toluenesulfonyl)-(S)-
phenylalanine (TPA), (S)-MBA·(S)-TPA crystallized as the
less-soluble salt from aqueous alcohol. On the other hand, the
(R)-MBA·(S)-TPA salt was obtained from a mixture of 2-
PrOH/1,4-dioxane or 2-PrOH/cyclohexane.7b,c The latter salt
contained an equimolar amount of the 1,4-dioxane or
cyclohexane molecules, with no hydrogen bonds contributing
to the structure; that is, the solvent molecules worked only as a
space filler. To the best of our knowledge, such a phenomenon
has not been reported to date. However, the efficiency of
resolving each enantiomer was not satisfactory using this
method. It is noteworthy that these chirality switching systems
are based on the structural change of only one diastereomeric
salt induced by solvent incorporation. This is disadvantageous
from the viewpoint of obtaining both enantiomers in high-
resolution efficiency.
BuOH, as well as 1,4-dioxane, were examined as the fractional
crystallization solvent of ( )-1·(−)-2. The deposited salt was
filtered and decomposed by addition of an aqueous NaOH
solution. After (−)-2 was recovered by extraction, the aqueous
solution was acidified, and optically active 1 was extracted. The
optical purity of resolved 1 was determined by comparison of
its specific rotation value with literature data.13 The results of
the optical resolution are summarized in Table 1, along with the
permittivity of each solvent.
It was observed that the resolving solvent significantly
affected the resolution. In the cases of EtOH and 1,4-dioxane,
(+)-1·(−)-2 was obtained as the less-soluble salt (Table 1,
entries 1 and 4), whereas for 1-PrOH and 1-BuOH, (−)-1·
(−)-2 was obtained preferentially (entries 2 and 3). The
present system is therefore a novel example in which both
enantiomers can be obtained using a single enantiomer of a
resolving agent by solvent switching. While the result using
EtOH was not satisfactory, (+)-1 was obtained in high yield
and purity by crystallization from 1,4-dioxane. On the other
hand, remarkably high optical purity of (−)-1 (90% ee) was
achieved from 1-PrOH, while its resolution efficiency was
higher for 1-BuOH. From these results, the permittivity of the
solvents cannot be regarded as a determining parameter for
reversed stereoselectivity, although resolution using a combi-
nation of different solvents to control the permittivity was not
examined.14
In this work, we present a diastereomeric resolution system
involving mandelic acid (1) and erythro-2-amino-1,2-dipheny-
lethanol (2), which shows chirality switching induced by
solvent incorporation into both diastereomeric salts, as well as
reciprocal resolution. The structural features of the solvent
molecules and the crystal packing are discussed on the basis of
comparison of the crystallographic data of the diastereomeric
salts. The potential of this stereocontrol technique is also
demonstrated by the successive resolution of 1 and 2.
RESULTS AND DISCUSSION
■
Optical Resolution of Mandelic Acid (1) with
Enantiopure erythro-2-Amino-1,2-diphenylethanol (2)
via Diastereomeric Salt Formation. Enantiopure 2 has
been applied as a chiral auxiliary,8 an organic fluorescent
material,9 and in supramolecular gel formation.10 Furthermore,
2 has been previously reported to be a suitable resolving agent
for various 2-arylalkanoic acids.11 However, resolution of α-
hydroxycarboxylic acids has not been frequently reported.12 In
the present study, an equimolar amount of (−)-2 and ( )-1
were mixed to form the diastereomeric salts. In order to study
the effects of the solvent, three alcohols, EtOH, 1-PrOH, and 1-
The composition of the precipitated salts was estimated from
1
the measured H NMR spectra, and it was suggested that the
salts contained solvent molecules (molar ratio of solvent to 1·
(−)-2 was 0.8−1.0), and crystallized as a 1·(−)-2·solvent salt,
with the exception of that obtained from EtOH. Such an
inclusion of solvents could change the solubility of the
diastereomeric salt crystals, resulting in the switching of
chirality. It is known that EtOH can be also incorporated
into the salts; however, in the present case, only a small amount
of EtOH was detected in the 1H NMR analysis, probably due to
its desorption, even at ambient temperature.15
B
dx.doi.org/10.1021/jo401517q | J. Org. Chem. XXXX, XXX, XXX−XXX

The Journal of Organic Chemistry
Article
Table 2. Diastereomeric Resolution of ( )-2 with Enantiopure (−)-1
2
a
b
c
d
entry
solvent
ε
yield (%)
[α]D (deg)
OP (%ee)
resolution efficiency
less-soluble salt
1
2
3
1-PrOH
22
17
2
50
93
−8.5
−6.1
7.8
>99
44
50
40
66
(−)-1·(−)-2
(−)-1·(−)-2
(−)-1·(+)-2
1-BuOH
1,4-dioxane
e
71
94
a
b
Yields of the less-soluble diastereomeric salt. See the Experimental Section. Specific rotation values of resolved 2. Solvent: EtOH, c 0.6, 20 °C.
c
Optical purity determined by chiral HPLC analysis of the 2-oxazolidinone derivative. (Chiralcel OJ-3 column, 4.6 mm × 250 mm; 2-PrOH/hexane
d
e
= 3:7 v/v). Resolution efficiency = yield (%) × OP (%ee)/100. Calculated based on half the amount of ( )-2.
The thermal stability of the incorporation of solvent
molecules in the diastereomeric salts was examined using
thermogravimetric analysis (TGA). Heating of the three salts of
1·(−)-2 obtained by recrystallization from 1-PrOH, 1-BuOH,
and 1,4-dioxane showed a gradual weight loss at 80−120 °C,
which mostly corresponded to the amounts of included
suggested in previous literature,15 both salts incorporated
EtOH molecules, with compositions of 1·(−)-2/EtOH = 1:1.
1
solvents estimated from H NMR measurements (Supporting
Information, Figure S1). These results indicate that the
interactions responsible for the incorporation of the solvent
molecules were comparable for all of the solvents, even though
they gave rise to reversed enantioselectivity.
Optical Resolution of erythro-2-Amino-1,2-diphenyle-
thanol (2) with Enantiopure Mandelic acid (1) via
Diastereomeric Salt Formation. Considering the impact of
solvent on enantioselectivity in the resolution of ( )-1 with
(−)-2, its reciprocal resolution was next examined, with (−)-1
used as the resolving agent and ( )-2 used as the target
racemic substrate. It is known that reciprocal resolution is not
always successful because it is not enantiomeric with the
original resolution system.2 The diastereomeric resolution of
racemic 2 with enantiopure 1 in EtOH has been previously
shown to be unsuccessful, with both salts being deposited.15 On
consideration of the results shown in Table 1, 1-PrOH, 1-
BuOH, and 1,4-dioxane were examined as fractional crystal-
lization solvents (Table 2). As the specific rotation value of
enantiopure 2 was very small, the optical purity of resolved 2
was determined by chiral high-performance liquid chromatog-
raphy (HPLC) analysis of its 2-oxazolidinone derivative.16 The
solvent dependency of the enantioselectivity was found to
correspond well with the original resolution shown in Table 1;
(−)-2 was efficiently obtained from 1-PrOH and 1-BuOH
(entries 1 and 2), and (+)-2 was obtained from 1,4-dioxane,
Figure 1. Crystal structures of the diastereomeric salts of 1 with (−)-2
crystallized from EtOH. (a) (−)-1·(−)-2·EtOH (more-soluble salt)
viewed from the a axis. (b) (+)-1·(−)-2·EtOH (less-soluble salt)
viewed from the a axis. Hydrogen atoms are omitted for clarity. The
dotted lines and shadows represent hydrogen bonds and hydrogen-
bonding networks, respectively.
1
with higher efficiency (entry 3). It was observed in the H
NMR spectra that the deposited salts from 1-PrOH and 1-
BuOH were a 1:1 mixture of (−)-1·2 and the solvent, which
was the same as was observed in the original resolution.
However, the composition of the salt obtained from 1,4-
dioxane was estimated to be (−)-1/2/solvent = 2:1:1.8 from
the 1H NMR analysis. This result indicates that the crystallized
salt was different from that obtained in the original resolution,
in spite of the same solvent being used for the crystallization.
Although the molar ratio was different, high efficiency for (+)-2
was still achieved by crystallization from 1,4-dioxane, and facile
solvent-induced stereocontrol was possible.
Comparison of the Crystal Structures of the Diaster-
eomeric Salts (−)-1·(−)-2 and (+)-1·(−)-2. In order to
elucidate the mechanism of the reversed enantioselectivity in
the reciprocal resolutions of 1 and 2 that were dependent on
the crystallization solvents, we carried out X-ray crystallo-
graphic analysis of both diastereomeric salts. The structures of
the more-soluble (−)-1·(−)-2 salt and less-soluble (+)-1·(−)-2
salts prepared from EtOH solution are shown in Figure 1. As
In both diastereomeric salt crystals, 1 and 2 formed charge-
assisted hydrogen bonds, producing hydrogen-bonding supra-
molecular networks. In addition, the EtOH molecules were
included by formation of hydrogen bonds, thereby reinforcing
the networks. However, the particular type of network for each
diastereomer was totally different. In the (−)-1·(−)-2·EtOH
salt, (−)-1 and (−)-2 formed one-dimensional (1D) hydrogen-
bonding ladderlike networks, linked together by EtOH
molecules to afford a two-dimensional (2D) sheetlike structure
(Figure 1a). The hydroxy group of each EtOH molecule
formed two hydrogen bonds to link a hydroxy oxygen atom of
(−)-1 and an ammonium nitrogen atom of (−)-2. On the other
hand, in the (+)-1·(−)-2·EtOH salt crystal, 1D columnar
hydrogen-bonding networks with a 2-fold screw axis were
constructed (Figure 1b). Channel-like cavities were formed
along the column, and the EtOH molecules were incorporated
within the channels via the formation of two hydrogen bonds,
C
dx.doi.org/10.1021/jo401517q | J. Org. Chem. XXXX, XXX, XXX−XXX

The Journal of Organic Chemistry
Article
one with a carboxylate group of (+)-1 and the other with an
ammonium group of (−)-2. Formation of such a 1D columnar
structure is often observed in salts consisting of carboxylic acids
and enantiopure 2 when crystallized from alcoholic solvents. In
addition, there have been a number of reports regarding solvent
alcohol molecules being included in the channels of columnar
structures.9,11 We have previously successfully applied such an
inclusion phenomenon to the efficient enantioseparation of
various alcohols and sulfoxides.17 Such a high ability of the salts
of 2 to incorporate solvent molecules could be due to the two
bulky and rigid aromatic rings of 2 positioned in a gauche
conformation.18
The low efficiency of the optical resolution in EtOH
demonstrated that the (−)-1·(−)-2 salt was slightly more
soluble than the (+)-1·(−)-2 salt, while the two diastereomeric
salts had similar stabilities. From a comparison of these crystal
structures, it appears that the (+)-1·(−)-2·EtOH salt was
preferable from the viewpoint of the stability of the 1D
networks. In this case, the molecules were related by the 21-
screw symmetry and combined more tightly, while the (−)-1·
(−)-2·EtOH salt formed less symmetrical 1D ladderlike
networks. However, the latter is preferable due to the higher
dimensionality of the hydrogen-bonding network connected by
EtOH molecules. It is indicated that the opposite influence of
these two factors resulted in the comparable stability of the two
diastereomeric salts and low resolution efficiency.
When optical resolution was carried out in 1-BuOH or 1-
PrOH, (−)-1·(−)-2 was precipitated as the less-soluble salt
with high efficiency. The structures of both diastereomeric salts
prepared from 1-BuOH are shown in Figure 2. It was found
that 1-BuOH was included in these crystals, with compositions
of 1·(−)-2/1-BuOH = 1:1. Unlike the salts that incorporated
EtOH, similar 1D columnar hydrogen-bonding networks were
constructed in both diastereomeric salt crystals, with 1-BuOH
embedded in the channels of the columns. However, there are
several differences between the two diastereomeric salts. First,
the orientations of 1 in the columnar structure were different,
and therefore, 1-BuOH molecules became associated by
hydrogen bonds in a different fashion. The hydroxy group of
1-BuOH formed hydrogen bonds with an ammonium group of
(−)-2 and a hydroxy group of (−)-1 in the (−)-1·(−)-2·1-
BuOH crystal (Figure 2a). On the other hand, 1-BuOH was
captured by an ammonium group of (−)-2 and a carboxylate
group of (+)-1 in the (+)-1·(−)-2·1-BuOH crystal (Figure 2b).
Second, the manner of column packing was found to be
different. All of the columnar structures were uniform and
arranged in a highly symmetrical fashion (space group:
P212121) in the (−)-1·(−)-2·1-BuOH crystal. However, the
more-soluble salt crystal, (+)-1·(−)-2·1-BuOH, was composed
of two kinds of columns with slightly different molecular
orientations, which should be unfavorable from the viewpoint
of the lower symmetry. In addition, the columns were packed in
a less symmetrical P21 space group. Less dense molecular
packing in (+)-1·(−)-2·1-BuOH was suggested by the crystal
densities, which were 1.239 and 1.207 for the less and more
soluble diastereomers, respectively. These disadvantages would
result in the lower stability of the (+)-1·(−)-2·1-BuOH salt,
and as a result, improved resolution efficiency.
Figure 2. Crystal structures of the diastereomeric salts of 1 and (−)-2
crystallized from 1-BuOH. (a) (−)-1·(−)-2·1-BuOH (less-soluble salt)
viewed from the a axis. (b) (+)-1·(−)-2·1-BuOH (more-soluble salt)
viewed from the b axis. Hydrogen atoms are omitted for clarity. The
dotted lines and shadows represent hydrogen bonds and hydrogen-
bonding networks, respectively.
of 1D columnar structures; however, the situation was different
from the case of 1·(−)-2·1-BuOH, with both crystallized in the
same space group P212121. The crystal structure of the less-
soluble (−)-1·(−)-2·1-PrOH salt was observed to be almost the
same as that of the (−)-1·(−)-2·1-BuOH crystal (Figure 3a),
whereas the structure of the more-soluble (+)-1·(−)-2·1-PrOH
was very similar to that of (+)-1·(−)-2·EtOH (Figure 3b). The
positions of the 1-PrOH molecules included in the more-
soluble salt could not be accurately determined owing to the
structural disorder, which was probably due to the shape of the
channel being unsuitable for the inclusion of 1-PrOH.
Comparison of the densities of both salt crystals demonstrated
that (+)-1·(−)-2·1-PrOH was less dense than (−)-1·(−)-2·1-
PrOH (1.216 and 1.237, respectively). These differences
indicated that (+)-1·(−)-2·1-PrOH was less stable than
(−)-1·(−)-2·1-PrOH, which led to the predominant precip-
itation of the (−)-1·(−)-2·1-PrOH salt during the optical
resolution.
The three linear aliphatic alcohols were each incorporated via
two hydrogen bonds in both diastereomeric salts. In the case of
(+)-1·(−)-2, columnar hydrogen-bonding networks were
commonly constructed. However, by increasing the alkyl
chain length of the solvent, from EtOH to 1-PrOH and 1-
BuOH, it appears that the molecule became too long to be
settled in the channel, with symmetrical packing of the resulting
columns prevented for the (+)-1·(−)-2·1-BuOH salt. On the
other hand, the molecular arrangement of the (−)-1·(−)-2 salt
changed from ladderlike to a columnar structure on increasing
the alkyl chain length, with dense packing of columns achieved
The crystal structures of the diastereomeric salts prepared
from 1-PrOH are shown in Figure 3. The stoichiometry was the
same as for the cases of the other alcohols, with 1·(−)-2/1-
PrOH = 1:1 found for both salt crystals. In the two
diastereomeric salts, the molecules were arranged in the form
D
dx.doi.org/10.1021/jo401517q | J. Org. Chem. XXXX, XXX, XXX−XXX

The Journal of Organic Chemistry
Article
Figure 3. Crystal structures of the diastereomeric salts of 1 with (−)-2
crystallized from 1-PrOH. (a) (−)-1·(−)-2·1-PrOH (less-soluble salt)
viewed from the a axis. (b) (+)-1·(−)-2·1-PrOH (more-soluble salt)
viewed from the b axis. Hydrogen atoms are omitted for clarity. The
dotted lines and shadows represent hydrogen bonds and hydrogen-
bonding networks, respectively.
Figure 4. Crystal structures of the diastereomeric salts of 1 with (−)-2
crystallized from 1,4-dioxane. (a) (−)-1·(−)-2·1,4-dioxane (more-
soluble salt) viewed from the a axis. (b) (+)-1·(−)-2·1,4-dioxane (less-
soluble salt) viewed from the a axis. Hydrogen atoms are omitted for
clarity. The dotted lines and shadows represent hydrogen bonds and
hydrogen-bonding networks, respectively.
by effectively filling the channels between the columns with the
solvent molecules. Such an influence of the molecular length of
alcohols was further supported by the fact that (+)-1·(−)-2
crystallized from 2-PrOH afforded a solvated (+)-1·(−)-2·2-
PrOH crystal, in which the molecular arrangement was very
similar to that of the (+)-1·(−)-2·EtOH crystal (Supporting
Information, Figure S2). Because the effective molecular
lengths of EtOH and 2-PrOH are comparable, it is probable
that the manner of packing in the columnar structures was
mainly determined by molecular length of the solvent alcohols.
A similar effect of the molecular length of the alcohols on the
packing of columnar structures was also observed in our
previous study.17c
columnar structures with a 2-fold screw axis were formed, and
the conformation of (+)-1 was fixed by a bifurcated hydrogen
bond involving an ammonium hydrogen atom of (−)-2
interacting with both a hydroxy oxygen atom and a carboxylate
oxygen atom of (+)-1. The two phenyl groups of (−)-2 were
present in a trans conformation, and a hydroxy group of (−)-2
faced outward from the columns. The hydroxy hydrogen atom
of (−)-2 formed a hydrogen bond with the carboxylate oxygen
atom of (+)-1 in the neighboring column, and the columns
were tightly connected to give a 2D sheetlike hydrogen-
bonding network with a void space. The 1,4-dioxane molecules
were incorporated in the void by formation of a hydrogen bond
between an ether oxygen atom of 1,4-dioxane and an
ammonium group of (−)-2. The molecular length of 1,4-
dioxane was comparable with the size of the void space;
therefore, it appeared that 1,4-dioxane was appropriate for
filling the void. In addition, the resulting fairly planar surface of
the 2D sheet-like networks contributed to efficient van der
Waals packing. All of these factors presumably resulted in the
high stability of the (+)-1·(−)-2·1,4-dioxane salt, and the
resulting high-resolution efficiency with reversed stereo-
selectivity compared to the cases where alcohols were applied
as the solvent.
Finally, the structures of both diastereomeric salts prepared
from 1,4-dioxane solution of the 1:1 mixture of 1 and (−)-2 are
1
shown in Figure 4. As suggested from the H NMR spectra,
both salts included 1,4-dioxane molecules, with their
composition identified as being 1·(−)-2/1,4-dioxane = 1:1.
The more-soluble (−)-1·(−)-2·1,4-dioxane salt formed 1D
columnar structures, packed in a P212121 space group (Figure
4a). The molecular arrangement was almost the same as those
of the less-soluble (−)-1·(−)-2·1-PrOH and (−)-1·(−)-2·1-
BuOH salts. The intercolumnar channels were occupied with
1,4-dioxane molecules that were connected by one hydrogen
bond between an ether oxygen atom of 1,4-dioxane and an
ammonium group of (−)-2. The density of the crystal was as
large as 1.309, and it was suggested that the molecules were
tightly packed and the salt crystal was relatively stable. The
solubility of the salt crystallized from 1,4-dioxane was found to
be lower than those of 1-PrOH or 1-BuOH. On the other hand,
a rather different structure was observed for the less-soluble
(+)-1·(−)-2·1,4-dioxane salt crystal (Figure 4b). Symmetric 1D
Furthermore, during the optical resolution of 2 with
enantiopure (−)-1 in 1,4-dioxane, the precipitated crystals
consisted of (−)-1 and 2 in a molar ratio of 2:1, with high
stereoselectivity for (+)-2 still achieved. In order to elucidate
the detailed mechanism, a single crystal was prepared from a
1,4-dioxane solution of a 2:1 mixture of (+)-1 and (−)-2, and
its structure is shown in Figure 5. The obtained crystal was
E
dx.doi.org/10.1021/jo401517q | J. Org. Chem. XXXX, XXX, XXX−XXX

The Journal of Organic Chemistry
Article
a
Table 3. Successive Resolution of ( )-1 with (−)-2 by the Solvent Switch Method
first resolution ( )-1 from 1-BuOH
second resolution (+)-enriched 1 from 1,4-dioxane
b
c
d
e
b
c
d
e
yield (%)
[α]D (deg)
OP (%ee)
resolution efficiency
yield (%)
[α]D (deg)
OP (%ee)
resolution efficiency
70
−136.7
89
62
71
116.6
75
53
a
b
c
d
( )-1/(−)-2 = 1.0. Yield of the less-soluble salt. Specific rotation of resolved 1. Solvent: H2O, c 0.8, 25 °C. Optical purity calculated from the
specific rotation value based on the literature; [α]D 153.5° (c 1, H2O). Resolution efficiency = yield (%) × OP(%)/100.
e
a
Table 4. Successive Resolution of ( )-2 with (−)-1 by the Solvent Switch Method
first resolution ( )-2 from 1-BuOH
second resolution (+)-enriched 2 from 1,4-dioxane
b
c
d
e
b
,
f
c
d
e
yield (%)
[α]D (deg)
OP (%ee)
resolution efficiency
yield (%)
40
[α]D (deg)
OP (%ee)
resolution efficiency
93
−6.1
44
40
7.8
95
37
a
b
c
d
( )-2/(−)-1 = 1.0. Yield of the less-soluble salt. Specific rotation of resolved 2. Solvent: EtOH, c 0.6, 20 °C. Optical purity determined by chiral
HPLC for the 2-oxazolidinone derivative. Resolution efficiency = yield (%) × OP(%)/100. Calculated based on half the amount of ( )-2.
e
f
possible reason for such a polymorphism in our reciprocal
resolution system is the difference in the amount of (−)-1 (or
(+)-1) and (+)-2 (or (−)-2) during crystallization. In the
resolution of ( )-2 using (−)-1, two equivalents of (−)-1 for
every (+)-2 were present, while a half amount of (+)-1 to (−)-2
was present in the original resolution involving ( )-1.
Successive Resolution of 1 and 2 by Solvent Switch-
ing. The present data and the aforementioned high utility of 1
and 2 prompted us to successively resolve both enantiomers of
each compound by solvent switching, using a previously
reported method.6b Successive resolution of ( )-1 with (−)-2
was first examined. As can be seen in Table 1, the less-soluble
diastereomeric salts changed from (+)-1·(−)-2 to (−)-1·(−)-2
Figure 5. Crystal structure of 2((+)-1)·(−)-2·2(1,4-dioxane) viewed
by selecting 1-BuOH for the first resolution and 1,4-dioxane for
from the a axis. Hydrogen atoms are omitted for clarity. The dotted
the second resolution.
lines and shadows represent hydrogen bonds and hydrogen-bonding
networks, respectively.
After the (−)-1·(−)-2·1-BuOH salt was collected by
filtration in the first resolution, the (+)-1·(−)-2 enriched salt
was recovered by condensation of the mother liquor. The
residue was crystallized from 1,4-dioxane to obtain (+)-1·
(−)-2·1,4-dioxane as the second crop. Although the procedure
was not optimized, the results shown in Table 3 corresponded
well to those in Table 1. In addition, as a reciprocal resolution
system, resolution of ( )-2 with (−)-1 was performed.
Considering that resolution from 1,4-dioxane produced the
nonstoichiometric salt, as seen in Table 2, 1-BuOH was
selected as the first resolving solvent, with good results
obtained, as summarized in Table 4.
found to be a pseudopolymorph of the (+)-1·(−)-2·1,4-dioxane
mentioned above, with a composition of (+)-1/(−)-2/1,4-
dioxane = 2:1:2. The molecular arrangement of 2((+)-1)·
(−)-2·2(1,4-dioxane) was apparently similar to that of (+)-1·
(−)-2·1,4-dioxane, with the two phenyl groups of (−)-2
present in the trans position, and 2D sheetlike hydrogen-
bonding networks constructed. One of the two molecules of
(+)-1 formed charge-assisted hydrogen bonds with (−)-2, and
the resultant 1D ladderlike hydrogen-bonding networks were
combined with the other neutral (+)-1 molecules. Two 1,4-
dioxane molecules were also independently included, filling the
void space. One formed a hydrogen bond with an ammonium
hydrogen atom of (−)-2, while the other interacted with a
hydroxy hydrogen atom of (−)-2. Though they crystallized in a
less symmetric space group (P1), it appeared that the efficient
packing of the planar 2D networks made this pseudopolymo-
phic crystal stable enough to precipitate preferentially. The
solid-state behavior of the salts was further elucidated by
measurement of their IR spectra (Supporting Information,
Figure S3). The salt, 2((+)-1)·(−)-2·2(1,4-dioxane), showed
an absorption band at 1719 cm−1, derived from the CO
stretching vibration of neutral 1. This absorption band
disappeared on cogrinding of this salt with another equivalent
of (−)-2 in the presence of a few drops of 1,4-dioxane. The
spectrum of the resultant solid was consistent with that of
(+)-1·(−)-2·1,4-dioxane, which suggested that the 1:1 salt
formation was enhanced by grinding. There have been some
reports on similar liquid-assisted cogrinding being useful for
controlling the polymorphic forms of cocrystals or salts.19 One
CONCLUSION
■
In conclusion, we have demonstrated highly efficient reciprocal
optical resolution between mandelic acid, 1, and erythro-2-
amino-1,2-diphenylethanol, 2. The stereoselectivity was dra-
matically switched by changing the crystallization solvent from
1,4-dioxane to 1-PrOH or 1-BuOH, while EtOH gave
disappointing results. The technique was applied to the
successive resolution without the requirement for a change in
resolving agent, and both enantiomers of 1 combined with 2
were obtained efficiently. Crystallographic analysis of the
diastereomeric salts showed that they generally incorporated
the crystallization solvent, affording inclusion crystals with
different compositions and structures. Differences in the
hydrogen-bonding networks and their packing manner were
found to be crucial for the solvent-dependent switching of
stereoselectivity in the diastereomeric resolution. This is the
first example of a solvent switching system where the stabilities
of both diastereomers are altered by incorporation of solvent
F
dx.doi.org/10.1021/jo401517q | J. Org. Chem. XXXX, XXX, XXX−XXX

The Journal of Organic Chemistry
Article
The white solid (0.7291 g) recovered from the filtrate was
recrystallized from 1,4-dioxane (30 mL) to obtain the less-soluble
(+)-1·(−)-2 salt as platelet crystals (0.5693 g, 71%; [α]D24 = −44.9° (c
1.0, MeOH); mp 147−152 °C). The apparent yield was calculated
molecules. The insight obtained from this study will potentially
contribute to clarification of the mechanisms of solvent
switching. Further application of this method to other
hydroxycarboxylic acids or aminoalcohols is now under
investigation.
1
based on the H NMR data, which indicated a ratio of (+)-1·(−)-2/
1,4-dioxane = 1:1. Decomposition of the crystals (0.2572 g) afforded
25
(+)-1 (0.0800 g, 65%) in 75% ee ([α]D = 116.6° (c 0.80, H2O)).
EXPERIMENTAL SECTION
Resolution of ( )-erythro-2-Amino-1,2-diphenylethanol
(( )-2) by (−)-Mandelic Acid ((−)-1) from 1-PrOH. The
diastereomeric salts of ( )-2 (0.2927 g, 1.37 mmol) and (−)-1
(0.2119 g, 1.39 mmol) were recrystallized from 1-PrOH (20 mL) to
obtain needlelike crystals (0.1867 g). Further recrystallization from 1-
PrOH (5 mL) afforded the less-soluble (−)-1·(−)-2 salt (0.1420 g,
■
Resolution of ( )-Mandelic Acid (( )-1) by erythro-(1R,2S)-
(−)-2-Amino-1,2-diphenylethanol ((−)-2) from EtOH. To ( )-1
(0.5310 g, 3.49 mmol) in MeOH (5 mL) was added (−)-2 (0.7414 g,
3.47 mmol) in MeOH (15 mL). After concentration, the resulting
diastereomeric salts (1.3251 g) were recrystallized from EtOH (12
mL) and filtered to obtain needlelike crystals (0.7225 g). The crystals
were recrystallized once from 90% EtOH (10 mL) to afford the less-
soluble (+)-1·(−)-2 salt (0.0795 g, 12%; [α]D = −55.6° (c 0.97,
MeOH); mp 147−150 °C). The apparent yield was calculated
assuming the salt composition was 1:1 for 1/(−)-2.
23
50%; [α]D = −115.6° (c 1.0, MeOH); mp 157−159 °C). The
apparent yield was calculated based on the 1H NMR data, which
indicated a ratio of (−)-1·(−)-2/1-PrOH = 1:0.8.
18
The salt crystals were dissolved in H2O at 60 °C and decomposed
by addition of 28% aqueous ammonia (0.2 mL) to precipitate (−)-2
24
(0.0621 g, 42%) ([α]D = −8.5° (c 0.48, EtOH)).
The salt was dissolved in H2O (5 mL) and decomposed by addition
of 1 M aqueous NaOH (5 mL), followed by extraction with Et2O. The
aqueous layer was acidified by addition of 1 M aqueous HCl (8 mL)
and extracted with Et2O. After drying over anhydrous Na2SO4, the
organic layer was concentrated to obtain (+)-1 (0.0256 g, 9.6%). From
the measurement of the specific rotation, the optical purity was
The optical purity of (−)-2 was determined to be greater than 99%
ee by chiral HPLC analysis of its oxazolidinone derivative, which was
prepared by reaction with triphosgene (Chiralcel OJ-3, 4.6 mm × 250
mm; eluent: 2-PrOH/hexane = 3:7 v:v; flow rate: 0.5 mL min−1;
detection wavelength: 254 nm; retention time: derivative of (+)-2, 14
min, derivative of (−)-2, 20 min).
25
calculated to be 62% ee ([α]D = 95.2° (c 0.51, H2O)).
Resolution of ( )-Mandelic Acid (( )-1) by erythro-(1R,2S)-
(−)-2-Amino-1,2-diphenylethanol ((−)-2) from 1-PrOH. The
diastereomeric salts of ( )-1 (0.5003 g, 3.28 mmol) and (−)-2
(0.7025 g, 3.29 mmol) were recrystallized from 1-PrOH (60 mL) to
obtain needlelike crystals (0.5739 g). Further recrystallization of the
crystals from 1-PrOH (35 mL) afforded the less-soluble (−)-1·(−)-2
salt (0.3686 g, 54%; [α]D18 = −114.0° (c 0.97, MeOH); mp 146−151
Resolution of ( )-erythro-2-Amino-1,2-diphenylethanol
(( )-2) by (−)-Mandelic Acid ((−)-1) from 1-BuOH. The
diastereomeric salts of ( )-2 (0.2912 g, 1.36 mmol) and (−)-1
(0.2094 g, 1.37 mmol) were recrystallized from 1-BuOH (25 mL) to
obtain needlelike crystals (0.3599 g), and the filtrate was concentrated
to obtain a white solid (0.2136 g). Recrystallization of the crystals from
1-BuOH (10 mL) afforded the less-soluble (−)-1·(−)-2 salt (0.2787 g,
1
20
°C). The apparent yield was calculated based on the H NMR data,
93%; [α]D = −87.8° (c 1.0, MeOH); mp 141−149 °C). The
apparent yield was calculated based on the 1H NMR data, which
which indicated a ratio of (−)-1·(−)-2/1-PrOH = 1:0.8. Decom-
position of the salt (0.2468 g) afforded (−)-1 (0.0783 g, 46%) in 90%
indicated a ratio of (−)-1·(−)-2/1-BuOH = 1:1. Decomposition of the
20
20
ee ([α]D = −139.6° (c 0.78, H2O)).
crystals (0.2062 g) afforded (−)-2 (0.0621 g, 57%) in 44% ee ([α]D
= −6.1° (c 0.6, EtOH)).
Resolution of ( )-Mandelic Acid (( )-1) by erythro-(1R,2S)-
(−)-2-Amino-1,2-diphenylethanol ((−)-2) from 1-BuOH. The
diastereomeric salts of ( )-1 (0.5926 g, 3.89 mmol) and (−)-2
(0.8325 g, 3.90 mmol) were recrystallized from 1-BuOH (70 mL) to
obtain needlelike crystals (1.1238 g). Recrystallization of the crystals
from 1-BuOH (50 mL) afforded the less-soluble (−)-1·(−)-2 salt
(0.8063 g, 94%; [α]D20 = −101.3° (c 1.0, MeOH); mp 149−154 °C).
Resolution of ( )-erythro-2-Amino-1,2-diphenylethanol
(( )-2) by (−)-Mandelic Acid ((−)-1) from 1,4-Dioxane. The
diastereomeric salts of ( )-2 (0.3030 g, 1.42 mmol) and (−)-1
(0.2159 g, 1.41 mmol) were recrystallized from 1,4-dioxane (30 mL)
16
to obtain platelike crystals (0.3855 g, 91%; [α]D = −12.7° (c 0.9,
MeOH); mp 109−114 °C). The apparent yield was calculated based
1
The apparent yield was calculated based on the H NMR data, which
1
on the H NMR data, which indicated a ratio of (−)-1/(+)-2/1,4-
indicated a ratio of (−)-1·(−)-2/1-BuOH = 1:0.94. Decomposition of
dioxane = 2:1:1.8. Decomposition of the salt crystals (0.3855 g)
the crystals (0.2562 g) afforded (−)-1 (0.0786 g, 83%) in 73% ee
21
afforded (+)-2 (0.1107 g, 72%) in 94% ee ([α]D = 7.8° (c 0.6,
21
([α]D = −112.0° (c 0.78, H2O)).
EtOH)).
Resolution of ( )-Mandelic Acid (( )-1) by erythro-(1R,2S)-
(−)-2-Amino-1,2-diphenylethanol ((−)-2) from 1,4-Dioxane.
The diastereomeric salts of ( )-1 (0.5961 g, 3.91 mmol) and (−)-2
(0.8393 g, 3.93 mmol) were recrystallized from 1,4-dioxane (100 mL)
to obtain platelike crystals (0.6530 g). Further recrystallization from
1,4-dioxane (20 mL) afforded the less-soluble (+)-1·(−)-2 salt crystals
Successive Resolution of ( )-erythro-2-Amino-1,2-dipheny-
lethanol (( )-2) by (−)-Mandelic Acid ((−)-1) by Solvent Switch
between 1-BuOH and 1,4-Dioxane. The white solid (0.2136 g)
obtained from the mother liquor after resolution of ( )-2 by (−)-1
from 1-BuOH was recrystallized from 1,4-dioxane (20 mL) to obtain
the less-soluble (−)-1·(+)-2 salt crystals as platelets (0.1755 g, 40%;
[α]D20 = 31.4° (c 1.0, MeOH); mp 157−160 °C). The apparent yield
20
(0.5072 g, 57%; [α]D = −43.4° (c 1.0, MeOH); mp 149−154 °C).
1
The apparent yield was calculated based on the H NMR data, which
1
was calculated based on the H NMR data, which indicated a ratio of
indicated a ratio of (+)-1·(−)-2/1,4-dioxane = 1:1. Decomposition of
(−)-1/(+)-2/1,4-dioxane = 1.8:1:1.8. Decomposition of the crystals
(0.1755 g) afforded (+)-2 (0.0575 g, 39%) in 95% ee ([α]D19 = 7.8° (c
0.57, EtOH)).
the crystals (0.2481 g) afforded (+)-1 (0.0775 g, 53%) in 88% ee
22
([α]D = 135.8° (c 0.77, H2O)).
Successive Resolution of ( )-Mandelic Acid (( )-1) by
erythro-(1R,2S)-(−)-2-Amino-1,2-diphenylethanol ((−)-2) by
Solvent Switch between 1-BuOH and 1,4-Dioxane. The
diastereomeric salts of ( )-1 (0.5406 g, 3.55 mmol) and (−)-2
(0.7546 g, 3.53 mmol) were recrystallized from 1-BuOH (90 mL) to
obtain needlelike crystals (0.7538 g). Further recrystallization from 1-
BuOH (40 mL) afforded the less-soluble (−)-1·(−)-2 salt crystals
(0.5473 g, 70%; [α]D24 = −103.7° (c 1.0, MeOH); mp 157−159 °C).
Single-Crystal X-ray Analyses of the Solvated Diastereo-
meric Salt Crystals. Single crystals suitable for X-ray diffraction
analysis were successfully prepared by slow evaporation of the
saturated solutions of pure diastereomeric salts, except for (+)-1·
(−)-2·EtOH. The single crystal of (+)-1·(−)-2·EtOH was obtained
during optical resolution of ( )-1 with (−)-2. X-ray crystallographic
data were collected using a CCD diffractometer with graphite
monochromated Mo Kα radiation. Data collections were carried out
at low temperature (150 K). The structures were solved by a direct
methods SIR 9720 and refined by SHELXL-97 programs.21 Crystallo-
graphic information files have been deposited with the Cambridge
1
The apparent yield was calculated based on the H NMR data, which
indicated a ratio of (−)-1·(−)-2/1-BuOH = 1:1. Decomposition of the
27
crystals (0.2535 g) afforded (−)-1 (0.0748 g, 59%) in 89% ee ([α]D
= −136.7° (c 0.74, H2O)).
G
dx.doi.org/10.1021/jo401517q | J. Org. Chem. XXXX, XXX, XXX−XXX

The Journal of Organic Chemistry
Article
Structural Database. Crystallographic parameters and CCDC numbers
are summarized in Table S1 in the Supporting Information.
1641−1646. (c) Kodama, K.; Kimura, Y.; Shitara, H.; Yasukate, M.;
Sakurai, R.; Hirose, T. Chirality 2011, 23, 326−332.
(8) (a) Gamsey, S.; DeLaTorre, K.; Singaram, B. Tetrahedron:
Asymmetry 2005, 16, 711−715. (b) Fang, T.; Du, D.-M.; Lu, S.-F.; Xu,
J. Org. Lett. 2005, 7, 2081−2084. (c) Abe, A. M. M.; Sauerland, S. J. K.;
Koskinen, A. M. P. J. Org. Chem. 2007, 72, 5411−5413. (d) Mahadik,
G. S.; Knott, S. A.; Szczepura, L. F.; Peters, S. J.; Standard, J. M.;
Hitchcock, S. R. J. Org. Chem. 2009, 74, 8164−8173. (e) Haddad, T.
D.; Hirayama, L. C.; Singaram, B. J. Org. Chem. 2010, 75, 642−649.
(f) Haddad, T. D.; Hirayama, L. C.; Buckley, J. J.; Singaram, B. J. Org.
Chem. 2012, 77, 889−898.
ASSOCIATED CONTENT
* Supporting Information
Summary of crystallographic data; additional figures including
TGA and IR spectra of the diastereometric salts and the crystal
structure of (+)-1·(−)-2·2-PrOH; and X-ray crystallographic
data (in CIF format) of the salt crystals. This material is
available free of charge via the Internet at http://pubs.acs.org.
■
S
(9) (a) Imai, Y.; Murata, K.; Kawaguchi, K.; Sato, T.; Kuroda, R.;
Matsubara, Y. Org. Lett. 2007, 9, 3457−3460. (b) Shiota, N.; Kinuta,
T.; Sato, T.; Tajima, N.; Kuroda, R.; Matsubara, Y.; Imai, Y. Cryst.
Growth Des. 2010, 10, 1341−1345. (c) Hige, S.; Shiota, N.; Kobayashi,
Y.; Wakabayashi, T.; Sato, T.; Imai, Y. Tetrahedron 2012, 68, 9999−
10004.
AUTHOR INFORMATION
Corresponding Author
*E-mail: hirose@apc.saitama-u.ac.jp.
Notes
■
(10) Chen, Y.-C.; Wang, H.; Li, D.-M.; Zheng, Y.-S. Eur. J. Org.
Chem. 2013, 2013, 1521−1529.
The authors declare no competing financial interest.
(11) Kinbara, K.; Kobayashi, Y.; Saigo, K. J. Chem. Soc., Perkin Trans.
2 1998, 1767−1775.
ACKNOWLEDGMENTS
■
This study was supported by a program for the “Future
Scientist” program at Saitama University provided by Japan
Science and Technology Agency. This paper is dedicated to
Professor Teruaki Mukaiyama in celebration of the 40th
anniversary of the Mukaiyama aldol reaction.
(12) Saigo, K.; Ogawa, S.; Kikuchi, S.; Kasahara, A.; Nohira, H. Bull.
Chem. Soc. Jpn. 1982, 55, 1568−1573.
(13) TCI Fine Chemicals 2012-2013; Tokyo Chemical Industry Co.,
Ltd.: Tokyo, 2012; No. 41, p 1364.
(14) (a) Kato, Y.; Kitamoto, Y.; Morohashi, N.; Kuruma, Y.; Oi, S.;
Sakai, K.; Hattori, T. Tetrahedron Lett. 2009, 50, 1998−2002.
(b) Kitamoto, Y.; Suzuki, K.; Morohashi, N.; Sakai, K.; Hattori, T. J.
Org. Chem. 2013, 78, 597−605.
(15) Saigo, K.; Sugiura, I.; Shida, I.; Tachibana, K.; Hasegawa, M.
Bull. Chem. Soc. Jpn. 1986, 59, 2915−2916.
REFERENCES
■
(1) (a) Rouhi, A. M. Chem. Eng. News 2003, 81, May 5, 45−55.
(b) Murakami, H. From racemates to single enantiomers − Chiral
synthetic drugs over the last 20 years. In Novel Optical Resolution
Technologies; Sakai, K., Hirayama, N., Tamura, R., Eds.; Springer:
Berlin, Germany, 2007; p 273−299.
(2) (a) Jacques, J.; Collet, A.; Wilen, S. H. Enantiomers, Racemates,
and Resolutions; Wiley & Sons: New York, 1981. (b) Eliel, E. L.; Wilen,
S. H.; Mander, N. Stereochemistry of Organic Compounds; Wiley &
Sons: New York, 1994. (c) Kozma, D. Optical Resolutions via
Diastereomeric Salt Formation; CRC Press: London, 2002.
(16) Nohira, H.; Mizuguchi, K.; Murata, T.; Yazaki, Y.; Kanazawa,
M.; Aoki, Y.; Nohira, M. Heterocycles 2000, 52, 1359−1370.
(17) (a) Kobayashi, Y.; Kodama, K.; Saigo, K. Org. Lett. 2004, 6,
2941−2944. (b) Kodama, K.; Kobayashi, Y.; Saigo, K. Chem.Eur. J.
2007, 13, 2144−2152. (c) Kodama, K.; Kobayashi, Y.; Saigo, K. Cryst.
Growth Des. 2007, 7, 935−939. (d) Kobayashi, Y.; Soetrisno; Kodama,
K.; Saigo, K. Tetrahedron: Asymmetry 2008, 19, 295−301.
(18) Kinbara, K.; Kobayashi, Y.; Saigo, K. J. Chem. Soc., Perkin Trans.
2 2000, 111−119.
(3) (a) Dalmolen, J.; Tiemersma-Wegman, T. D.; Nieuwenhuijzen, J.
W.; van der Sluis, M.; Echten, E.; Vries, T. R.; Kaptein, B.; Broxterman,
Q. B.; Kellogg, R. M. Chem.Eur. J. 2005, 11, 5619−5624.
(b) Leeman, M.; Brasile, G.; Gelens, E.; Vries, T.; Kaptein, B.;
Kellogg, R. Angew. Chem., Int. Ed. 2008, 47, 1287−1290.
̌ ̌ ́
(19) (a) Friscic, T.; Jones, W. Cryst. Growth Des. 2009, 9, 1621−
1637. (b) Loots, L.; Wahl, H.; van der Westhuizen, L.; Haynes, D. A.;
le Roex, T. Chem. Commun. 2012, 48, 11507−11509.
(20) Altomare, A.; Burla, M. C.; Camalli, M.; Cascarano, G. L.;
Giacovazzo, C.; Guagliardi, A.; Moliterni, A. G. G.; Polidori, G.;
Spagna, R. J. Appl. Crystallogr. 1999, 32, 115−119.
(4) (a) Mills, W. H.; Elliott, K. A. C. J. Chem. Soc. 1928, 1291−1302.
(b) Sealock, R. R.; Speeter, M. E.; Schweet, R. S. J. Am. Chem. Soc.
1951, 73, 5386−5388. (c) Berson, J. A.; Ben-Efraim, D. A. J. Am.
(21) Sheldrick, G. M. SHELX-97, Programs for the Refinement of
̀
Chem. Soc. 1959, 81, 4083−4087. (d) Brienne, M. J.; Ouannes, C.;
Crystal Structures; University of Gottingen: Gottingen, Germany, 1997.
̈
̈
Jacques, J. Bull. Soc. Chim. Fr. 1967, 613. (e) Varech, D.; Jacques, J.
Tetrahedron 1972, 28, 5671−5679. (f) Nohira, H.; Tamura, H.;
Asakura, M.; Terunuma, D. Japan Pat. 1984-110656; Chem. Abstr.
1985, 102, 5896. (g) Nohira, H.; Nohira, M. Japan Pat. 1983-32872;
Chem. Abstr. 1983, 99, 70782. (h) Kanoh, S.; Muramoto, H.;
Kobayashi, N.; Motoi, M.; Suda, H. Bull. Chem. Soc. Jpn. 1987, 60,
3659−3662. (i) Yoshioka, R.; Okamura, K.; Yamada, S.; Aoe, K.; Da-
te, T. Bull. Chem. Soc. Jpn. 1998, 71, 1109−1116. (j) Bal
Kiss, V.; Gajary, A.; Juvancz, Z.; Fogassy, E. Tetrahedron: Asymmetry
2001, 12, 3435−3439.
(5) Faigl, F.; Fogassy, E.; Nog
Tetrahedron: Asymmetry 2008, 19, 519−536.
́
int, J.; Egri, G.;
́
́
́ ́
radi, M.; Palovics, E.; Schindler, J.
(6) (a) Sakai, K.; Sakurai, R.; Hirayama, H. Tetrahedron: Asymmetry
2004, 15, 1073−1076. (b) Sakai, K.; Sakurai, R.; Nohira, H.; Tanaka,
R.; Hirayama, N. Tetrahedron: Asymmetry 2004, 15, 3495−3500.
(c) Sakai, K.; Sakurai, R.; Hirayama, N. Tetrahedron: Asymmetry 2006,
17, 1812−1816. (d) Sakurai, R.; Yuzawa, A.; Sakai, K.; Hirayama, N.
Cryst. Growth Des. 2006, 6, 1606−1610. (e) Sakurai, R.; Sakai, K.;
Kodama, K.; Yamaura, M. Tetrahedron: Asymmetry 2012, 23, 221−224.
(7) (a) Taniguchi, K.; Aruga, M.; Yasutake, M.; Hirose, T. Org.
Biomol. Chem. 2008, 6, 458−463. (b) Hirose, T.; Begum, M.; Islam, M.
S.; Taniguchi, K.; Yasutake, M. Tetrahedron: Asymmetry 2008, 19,
H
dx.doi.org/10.1021/jo401517q | J. Org. Chem. XXXX, XXX, XXX−XXX