Rapid determination of enantiomeric excess of α-chiral cyclohexanones using circular dichroism spectroscopy

Article abstract of DOI:10.1021/ol2004885

Chemical equations presented. Ketone handedness was discriminated using circular dichroism (CD) spectroscopy by monitoring the metal-to-ligand charge transfer (MLCT) bands of complexes between [Cu^I((S)-1)(CH ₃CN)₂]PF₆ and derivatized α-chiral cyclohexanones (4). This method was able to quantify the enantiomeric excess of unknown samples using a calibration curve, giving an absolute error of ±7%. The analysis was fast, allowing potential application of this assay in high-throughput screening (HTS).

Full text of DOI:10.1021/ol2004885

ORGANIC
LETTERS
2011
Vol. 13, No. 9
2298–2301
Rapid Determination of Enantiomeric
Excess of r-Chiral Cyclohexanones Using
Circular Dichroism Spectroscopy
Diana Leung and Eric V. Anslyn*
Department of Chemistry and Biochemistry, The University of Texas at Austin, Austin,
Texas 78712, United States
anslyn@austin.utexas.edu
Received March 2, 2011
ABSTRACT
Ketone handedness was discriminated using circular dichroism (CD) spectroscopy by monitoring the metal-to-ligand charge transfer (MLCT)
bands of complexes between [CuI((S)-1)(CH3CN)2]PF6 and derivatized R-chiral cyclohexanones (4). This method was able to quantify the
enantiomeric excess of unknown samples using a calibration curve, giving an absolute error of (7%. The analysis was fast, allowing potential
application of this assay in high-throughput screening (HTS).
Asymmetric synthesis is well-recognized as a cost-effec-
tive method for the production of enantiomerically pure
products.1 Combinatorial methods combined with high-
throughput screening (HTS) have become a tool for find-
ing appropriate catalysts/auxiliaries for asymmetric synth-
esis due to the allowance for a large number of candidates
to be rapidly examined. However, the determination of the
enantiomeric excess (ee) of reaction products is a limiting
factor in catalyst discovery. Chromatographic techniques
are commonly used to determine ee due to their high
accuracy, but the intrinsically serial analysis restricts their
ultimate use in HTS.2
Optical signaling techniques are well adapted for HTS
because they use instruments such as circular dichroism
(CD) spectrophotometers, fluorimeters, and UVꢀvis spec-
trophotometers, which require less time for analysis com-
pared to chromatographic techniques. Typically, a single
measurement can be conducted in under a minute, allow-
ing for rapid analysis of samples for HTS of asymmetric
catalysts/auxiliaries. Due to the advantage of speed, nu-
meroustechniquesusing optical signaling approaches have
been investigated for the determination of ee.3ꢀ7 These
optical signaling techniques have the advantage that they
can be transitioned to a microwell plate format, allowing
(3) Molecular Sensors: (a) Nieto, S.; Dragna, J. M.; Anslyn, E. V.
Chem.;Eur. J. 2010, 16, 227. (b) Nieto, S.; Lynch, V. M.; Anslyn, E. V.;
Kim, H.; Chin, J. Org. Lett. 2008, 10, 5167. (c) Nieto, S.; Lynch, V. M.;
Anslyn, E. V.; Kim, H.; Chin, J. J. Am. Chem. Soc. 2008, 130, 9232. (d)
Leung, D.; Anslyn, E. V. J. Am. Chem. Soc. 2008, 130, 12328. (e) Leung,
D.; Folmer-Andersen, J. F.; Lynch, V. M.; Anslyn, E. V. J. Am. Chem.
Soc. 2008, 130, 12318. (f) Shabbir, S. H.; Regan, C. J.; Anslyn, E. V.
Proc. Natl. Acad. Sci. U.S.A., Early Edition 2009, 1. (g) Zhu, L.; Shabbir,
S. H.; Anslyn, E. V. Chem.;Eur. J. 2006, 13, 99. (h) Liu, S.; Pestano,
J. P. C.; Wolf, C. J. Org. Chem. 2008, 73, 4267. (i) Richard, G. I.;
Marwani, H. M.; Jiang, S.; Fakayode, S. O.; Lowry, M.; Strongin,
R. M.; Warner, I. M. Appl. Spectrosc. 2008, 62, 476. (j) Wolf, C.; Liu, S.;
Reinhardt, B. C. Chem. Commun. 2006, 4242. (k) Corradini, R.;
Paganuzzi, C.; Marchelli, R.; Pagliari, S.; Dossena, A.; Duchateau, A.
J. Inclusion Phenom. Macrocyclic Chem. 2007, 57, 625. (l) Corradini, R.;
Paganuzzi, C.; Marchelli, R.; Pagliari, S.; Sforza, S.; Dossena, A.;
Galaverna, G.; Duchateau, A. J. Mater. Chem. 2005, 15, 2741.
(4) IR thermography: (a) Reetz, M. T.; Becker, M. H.; Kuhling,
K. M.; Holzwarth, A. Angew. Chem., Int. Ed. 1998, 37, 2647. (b) Reetz,
M. T.; Becker, M. H.; Liebl, M.; Furstner, A. Angew. Chem., Int. Ed.
2000, 39, 1236. (c) Reetz, M. T.; Hermes, M.; Becker, M. H. Appl.
Microbiol. Biotechnol. 2001, 55, 531. (d) Millot, N.; Borman, P.; Anson,
M. S.; Campbell, I. B.; Macdonald, S. J. F.; Mahmoudian, M. Org.
Process Res. Dev. 2002, 6, 463.
(1) (a) Tsukamoto, M.; Kagan, H. B. Adv. Synth. Catal. 2002, 344,
453. (b) Charbonneau, V.; Ogilvie, W. W. Mini-Rev. Org. Chem. 2005, 2,
313.
(2) Traverse, J. F.; Snapper, M. L. Drug Discovery Today 2002, 7,
1002.
r
10.1021/ol2004885
Published on Web 04/12/2011
2011 American Chemical Society

for an even more rapid means of analysis compared to
single-cuvette analyses.
number of functional handles to complex to the Cu(I)
metal center. To create ligands, they were derivatized with
1-methyl-1-(2-pyridyl) hydrazine (3) to produce bidentate
hydrazones 4 (Scheme 1).
Here we demonstrate the use of CD spectroscopy to
determine ee values of R-chiral cyclohexanones. An ana-
lyte binding event that would induce CD signals in wave-
lengths longer than 320 nm was sought because the
analytes themselves, most organic functional groups, and
organic solvents are CD silent in this region. This would
potentially allow the technique to be used on crude reac-
tion mixtures without purification, making the analysis
step even more rapid. For these purposes, the chiral metal
complex used in this study was the 2,20-bis(diphenylphos-
phino)-1,10-binaphthyl (BINAP) copper(I) complex
([CuI(1)(CH3CN)2]PF6). Previous studies have shown that
the CD active MLCT bands (Figure 1) are modulated by
complexation of chiral analytes, allowing for enantioselec-
tive discrimination.3a,c
We postulated that enantioselective discrimination of 4
by the chiral BINAP complex could arise via different twist
angles that the two enantiomers of 4 impart on the naphthyl
rings in 1. As shown in Figure 2, (S)-4 is predicted to impart
a smaller change in the twist of the two naphthyl moieties of
[CuI((S)-1)(CH3CN)2]PF6. This is due to the avoidance of
steric interactions between the R-group and the phosphine
ligand. On the other hand, binding of the enantiomer (R)-4
induces a steric clash between the substituent in (R)-4 and
the phosphine ligand, potentially inducing a larger twist to
prevent such interactions. This postulate is illustrated by the
Newman projection of the two diastereomeric complexes,
[CuI((S)-1)((R)-4)]PF6 and [CuI((S)-1)((S)-4)]PF6 in Fig-
ure 2. As discussed below, such twists are reflected in the CD
spectra, where the predicted disfavored diastereomeric com-
plexes have a larger change in their CD signals from the
initial MLCT band of [CuI((S)-1)(CH3CN)2]PF6. By mon-
itoring this change, chiral analytes can be enantioselec-
tively discriminated.
Figure 1. CD spectra of [CuI((S)-1)(CH3CN)2]PF6 (400 μM) in
CH3CN.
Scheme 1. Derivatization of R-Chiral Cyclohexanones (2) with
1-Methyl-1-(2-pyridyl) Hydrazine (3) To Produce a
Bidentate Analyte (4), Followed by Complexation to
[CuI((S)-1)(CH3CN)2]PF6 To Produce [CuI((S)-1)(4)]PF6
Figure 2. Steric interactions allowing for enantioselective dis-
crimination of hydrazone 4 with [CuI((S)-1)(CH3CN)2]PF6.
Because there are only a few enantiomerically pure
commercially available R-chiral cyclohexanones, some of
our analytes needed to be synthesized. Oxidation with
pyridinium chlorochromate (PCC) of commercially
(5) Liquid Crystals: (a) Eelkema, R.; van Delden, R. A.; Feringa,
B. L. Angew. Chem., Int. Ed. 2004, 43, 5013. (b) van Delden, R. A.;
Feringa, B. L. Chem. Commun. (Cambridge, U.K.) 2002, 174. (c) Van
Delden, R. A.; Feringa, B. L. Angew. Chem., Int. Ed. 2001, 40, 3198. (d)
Walba, D. M.; Eshdat, L.; Korblova, E.; Shao, R.; Clark, N. A. Angew.
Chem., Int. Ed. 2007, 46, 1473.
(6) Molecular Imprinted Polymers: Shimizu, K. D.; Snapper, M. L.;
Hoveyda, A. H. Chem.;Eur. J. 1998, 4, 1885.
(7) Enzymes/Antibodies: (a) Abato, P.; Seto, C. T. J. Am. Chem. Soc.
2001, 123, 9206. (b) Onaran, M. B.; Seto, C. T. J. Org. Chem. 2003, 68,
8136. (c) Dey, S.; Karukurichi, K. R.; Shen, W.; Berkowitz, D. B. J. Am.
Chem. Soc. 2005, 127, 8610.
The analytes targeted in this study, R-chiral cyclohex-
anones (2), presented a challenge due to their limited
Org. Lett., Vol. 13, No. 9, 2011
2299

available enantiomerically pure alcohols 2-phenylcyclo-
hexanol and 2-(2-phenylpropan-2-yl)cyclohexanol pro-
duced enantiomerically pure ketones 2-phenylcyclohex-
anone and 2-(2-phenylpropan-2-yl)cyclohexanone, respec-
tively. Using a polarimeter, the optical rotations of each
ketone’s enantiomer pair were confirmed to be equal and
opposite. In addition, the molar optical rotation of
2-phenylcyclohexanone had been previously reported
in the literature, and our sample was found to be in
agreement.8
Derivatization of enantiomerically pure (R)- and (S)-
2-phenylcyclohexanone with hydrazine 3 produced hydra-
zone 5, while 2-(2-phenylpropan-2-yl)cyclohexanone did
not react with hydrazine 3. This was most likely due to the
bulky substituent at the R-position of 2-(2-phenylpropan-2-yl)-
cyclohexanone. Fortunately, derivatization of (R,S)- and
(S,R)-menthone with hydrazine 3 produced hydrazones 6
as a second analyte.
but this step is a single addition to the time of analysis
for all samples. The derivatization step can be carried out
in parallel for all the samples requiring analysis, especially
since HTS of catalysts/auxiliaries often uses microreactors.
Hence, this synthetic step (5ꢀ26 h) can potentially be
incorporated into the asymmetric reaction scheme.
Upon obtaining derivatized enantiomerically pure hy-
drazones, 5 and 6, titrations with [CuI((S)-1)(CH3CN)2]-
PF6 were conducted to determine the assay’s ability to
discriminate the enantiomers. There was no CD signal in
the visible region from either enantiomers of 6, nor from the
enantiomers of 6 with the Cu(I) source, [CuI(CH3CN)4]PF6
(see Supporting Information). A large CD signal was
observed for [CuI((S)-1)(CH3CN)2]PF6, [CuI((S)-1)((R,S)-
6)]PF6, and [CuI((S)-1)((S,R)-6)]PF6 (Figure 4). The MLCT
bands are different for each enantiomer of 6, allowing for
enantioselective discrimination.
Condensation of hydrazine 3 with R-chiral ketone 2
creates two isomers (E)- and (Z)-4 (Figure 3). Through
HR-MS analysis of the hydrazones 5 and 6, one peak was
found for each compound, correlating to their [M þ H]þ.
However, in the 13C NMR studies, it was observed that
duplicate peaks were present, whichdid notcorrelate tothe
starting material. One set of peaks had a significantly
smaller intensity and chemical shifts close to the major
peaks. In addition, the 1H NMR showed extra smaller sets
of peaks, whose chemical shifts were also similar to the
major peaks. This lead us to conclude that both (E)- and
(Z)-isomers have been formed. Although two isomers were
present, no purification was performed, since the relative
ratio of isomers was consistent each time the reaction was
conducted (see Supporting Information). Due to the steric
clash between the methyl group and the R-substituent on
the R-position of (Z)-4, the dominant product from the
condensation was postulated to be (E)-4.
Figure 4. CD spectra of [CuI((S)-1)(CH3CN)2]PF6 (401 μM) in
CH3CN (black), [CuI((S)-1)(CH3CN)2]PF6 (401 μM) mixed
with (R,S)-6 (4 mM) (red), and [CuI((S)-1)(CH3CN)2]PF6 (401
μM) mixed with (S,R)-6 (4 mM) (green).
The molar ellipticity of the MLCT band at 350 nm (λmax
of the MLCT band) was monitored during a titration, and
the enantiomers of 6 werediscriminatedasshown in Figure
5a. After being able to enantioselectively discriminate 6
with[CuI((S)-1)(CH3CN)2]PF6, the scope of thisstudywas
extended to target 5. As shown in Figure 5b, hydrazone 5
was also successfully enantioselectively discriminated,
although to a lesser degree.
The enantiomer that induced the largest signal change
was (R,S)-6, which correlates to the hypothesis that the
enantiomer with an (R)-R-stereocenter would lead to more
steric interactions with the BINAP moiety, forcing a larger
twist of the naphthyl rings, and a larger change in the
MLCT band. This was also true of analyte 5, as shown by
the largerchangeinthe MLCTband for (R)-5 comparedto
(S)-5.
Figure 3. (E)- and (Z)-hydrazone produced upon derivatization
of R-chiral ketone (2) with hydrazine 3.
To demonstrate the assay’s ability to determine en-
antiomeric excess values of unknown samples, a cali-
bration curve was generated by monitoring the molar
ellipticity at 350 nm with different ee solutions of 5
One might argue that a technique requiring derivatiza-
tion of the analyte for the determination of ee is not
amenable for HTS of asymmetric catalysts/auxiliaries,
(8) Johnson, C. R.; Zeller, J. R. Tetrahedron 1984, 40, 1225.
Org. Lett., Vol. 13, No. 9, 2011
2300

100%) with small signal change, the rest of the calibration
curve (ꢀ100% to 50%) shows significant curvature. The
use of the enantiomeric host [CuI((R)-1)(CH3CN)2]PF6
willproducea mirror imagecalibration curve whose region
of maximum signal change closely corresponds to that of
minimum signal change from the [CuI((S)-1)(CH3CN)2]-
PF6 host shown. Thanks to this property, using both
receptors separately should make it possible to differenti-
ate between even closely separated ee samples across the
whole ee range.
Figure 6. Molar ellipticity at 350 nm with (5% error bars as a
function of ee performed with the addition of 5 (10 mM) into a
solution of [CuI((S)-1)(CH3CN)2]PF6 (400 μM) in CH3CN.
Figure 5. CD isotherm at 350 nm with the addition of (a) (R,S)-
or (S,R)-6 (15.4 and 19.1 mM, respectively) into a solution of
[CuI((S)-1)(CH3CN)2]PF6 (398 μM) in CH3CN. (b) (R)- or (S)-5
(18.1 and 22.2 mM, respectively) into a solution of [CuI((S)-1)-
(CH3CN)2]PF6 (400 μM) in CH3CN.
In summary, using commercially available [CuI((S)-1)-
(CH3CN)2]PF6, R-chiral cyclohexanones were enantiose-
lectively discriminated using circular dichroism spectro-
scopy. The speed at which CD operates is advantageous
for high-throughput screening of asymmetric catalysts/
auxiliaries. We have demonstrated this system’s ability to
discriminate enantiomers of derivatized cyclohexanones
and determine the ee of unknown samples with approxi-
mately (7% absolute error.
titrated into enantiomerically pure [CuI((S)-1)(CH3CN)2]PF6
(Figure 6). The ee calibration curve was subjected to
linear and second-degree polynomial regression, and
the best-fit curve was used to determine the ee of five
independently prepared unknown samples. The cali-
bration curve had considerable scatter, and yet, as now
described, the determination of ee values for unknown
samples was quite accurate. Errors were calculated for
each sample by taking the absolute difference between
the actual ee and the experimentally determined ee. The
average absolute error for determining the ee of 5 for
the five unknown samples was (6.6%.
Acknowledgment. We gratefully acknowledge the fi-
nancial support from the National Institutes of Health
(GM77437) and Welch Foundation (F-1151).
Supporting Information Available. Experimental pro-
1
cedures, H and 13C NMR data, spectral data, and ee
Though the calibration curve using the [CuI((S)-1)-
(CH3CN)2]PF6 host (Figure 6) has a region (50%ꢀ
determinations. This information is available free of
charge via the Internet at http://pubs.acs.org.
Org. Lett., Vol. 13, No. 9, 2011
2301