Chiral silylation reagents: Determining configuration via NMR-spectroscopic coanalysis

Article abstract of DOI:10.1021/ol049233b

(Chemical Equation Presented) Derivatization with (+)- and (-)-chloromenthoxydiphenylsilane was used to determine the absolute configuration of the insect defensive agent pinoresinol (1). Although the ¹H NMR chemical shift differences of the re

Full text of DOI:10.1021/ol049233b

ORGANIC
LETTERS
2004
Vol. 6, No. 18
3019-3022
Chiral Silylation Reagents: Determining
Configuration via NMR-Spectroscopic
Coanalysis
Frank C. Schroeder, Douglas B. Weibel, and Jerrold Meinwald*
Department of Chemistry and Chemical Biology, Cornell UniVersity,
Ithaca, New York 14853-1301
circe@cornell.edu
Received April 27, 2004 (Revised Manuscript Received June 30, 2004)
ABSTRACT
Derivatization with (+)- and (−)-chloromenthoxydiphenylsilane was used to determine the absolute configuration of the insect defensive agent
pinoresinol (1). Although the 1H NMR chemical shift differences of the resulting two diastereomers are small, 1H NMR spectroscopy provided
for the unambiguous assignment of the natural product’s configuration. For this purpose, a new approach involving NMR spectra of mixtures
of diastereomers was used. Our method resembles coinjecting mixtures of samples of known and unknown configuration in GC and HPLC.
One of the most widely used techniques for determining the
absolute stereochemistry of natural products involves the
incorporation of a chiral derivatizing agent (CDA) and
subsequent analysis via NMR spectroscopy.1 These methods
frequently involve the introduction of a chiral substituent
directly at or in close proximity to the chiral center of interest
to achieve complete separation of corresponding NMR
signals in the two complementary diastereomers. While
complete separation is desirable, it is often not possible to
introduce a chiral substituent sufficiently close to the chiral
center to achieve such a high degree of chemical shift
separation, especially if the chiral center is not at or next to
a hydroxy or amino group. Using the comparatively much
smaller chemical shift differences induced by derivatizing a
chiral, conformationally flexible molecule at a location far
away from the chiral centers of interest seems less obvious.2
In this situation, corresponding signals in the 1H NMR spectra
of the resulting complementary diastereomers would usually
not be well separated, although the respective chemical shift
values might be somewhat different.3 We here report the
determination of the absolute configuration of a natural
product, pinoresinol (1), via derivatization with a chiral
silylating agent at the periphery of the molecule, using the
resulting minute H NMR chemical shift differences for
unambiguous stereochemical assignment.
We recently identified pinoresinol (1) as a minor com-
ponent in the defensive secretion obtained from larvae of
the European cabbage butterfly, Pieris rapae.4 This finding
was somewhat surprising because pinoresinol, though widely
1
(1) (a) Dale, J. A.; Mosher, H. S. J. Am. Chem. Soc. 1973, 95, 512-
519. (b) Seco, J. M.; Quin˜oa´, E.; Riguera, R, J. Org. Chem. 1999, 64, 4669-
4675. (c) Kato, N. J. Am. Chem. Soc. 1990, 112, 254-257. (d) Anderson,
R. C.; Shapiro, M. J. J. Org. Chem. 1984, 49, 1304-1305. (e) Doolittle, R.
E.; Heath, R. R. J. Org. Chem. 1984, 49, 5041-5050. (f) Ohtani, I.; Kusumi,
T.; Kashman, Y.; Kakisawa, H. J. Am. Chem. Soc. 1991, 113, 1092-4096.
For a recent review, see: Seco, J. M.; Quinoa, E.; Riguera, R. Chem. ReV.
2004, 104, 17-117.
(2) Seco, J. M.; Latypov, Sh.; Quin˜oa´, E.; Riguera, R. Tetrahedron 1997,
53, 8541-8564.
(3) For an example for the use of small chemical shift differences to
establish remote stereochemical relationships, see: Boyle, C. D.; Kishi, Y.
Tetrahedron Lett. 1995, 36, 5695-5698.
(4) Schroeder, F. C.; Grant, J. B.; Weibel, D. B.; Smedley, S. R.; Bolton,
K. L.; Meinwald, J.; Eisner, T. Proc. Natl. Acad. Sci. U.S.A., submitted.
10.1021/ol049233b CCC: $27.50 © 2004 American Chemical Society
Published on Web 08/13/2004

occurring in plants,5 had never been reported as a defensive
agent in insects. Accordingly, we found ourselves con-
fronted with the need to determine the absolute configuration
of the material isolated from the caterpillar’s secretion. The
amount of pinoresinol that can be isolated from the secretion
is quite small (∼100 µg isolated from >2000 larvae) and
does not suffice for determining its absolute configuration
via chiroptical methods. Furthermore, with only a small
amount of material available, we were hesitant to use CDA-
based methods that would require multiple steps, such as a
partial degradation of pinoresinol (1) to expose hydroxy
groups closer to the stereocenters in 1 and their subsequent
derivatization. Instead, we considered directly derivatizing
the phenolic hydroxy groups in 1 with a CDA, hoping this
would allow us to differentiate the resulting diastereomers
by NMR.
in the menthoxy substituents appear at slightly different
chemical shift values (Figure 1). In terms of the size of the
Given the large intramolecular distance between the
phenolic hydroxy groups and the chiral centers in pinoresinol,
only very small differences in the chemical shift values for
any pair of diastereomers arising from such a derivatization
can be expected. It seemed highly unlikely that one could
achieve complete separation of corresponding NMR signals
in the spectra of complementary diastereomers with either
commonly available CDAs or our recently described chiral
silylation reagents.6 Nevertheless, since silylation can be
achieved under mild conditions and works well even for trace
amounts of material, we decided to derivatize a sample of
authentic (+)-pinoresinol, isolated from Forsythia suspensa
leaves,7 with the (+)- and (-)-isomers of chloromenthoxy-
diphenylsilane, 2 (Scheme 1). The phenolic hydroxy groups
reacted smoothly, and the desired diastereomeric derivatives
3 and 4 were produced in good yields.
1
Figure 1. 0.65-0.95 ppm section of H NMR spectra (benzene-
d6, 500 MHz): (A) (-)-menthoxydiphenylsilyl derivative of (+)-
pinoresinol (3); (B) (+)-menthoxydiphenylsilyl derivative of (+)-
pinoresinol (4); (C) (-)-menthoxydiphenyl-silyl derivative of
pinoresinol isolated from cabbage butterfly caterpillars (5).
chemical shift difference observed, using benzene-d6 as the
solvent gave the best results, as had been the case in earlier
applications of the silylation technique.6a Next, a sample of
pinoresinol (100 µg) isolated from cabbage butterfly caterpil-
lars was derivatized with (-)-menthoxydiphenylsilyl chlo-
1
ride. As shown in Figure 1, the H NMR spectrum of the
caterpillar derived silyl ether (5) resembled that of the (+)-
menthoxydiphenylsilyl derivative of (+)-pinoresinol (4) more
closely than that of its diastereomer 3, thus suggesting that
the caterpillar secretes (-)-pinoresinol. However, the chemi-
cal shift differences observed are extremely small, generally
of the order of 0.004 ppm and smaller (2 Hz or less at 500
MHz). Chemical shifts are often not reproducible at this level,
because many factors such as sample concentration, water
content, or pH can influence the exact values in a difficult
to predict manner. Therefore, a mere comparison of spectra
does not allow for unambiguous assignment of configuration
in cases such as the pinoresinol derivatives 3 and 4 where
spectra show only minute differences.
Scheme 1. Derivatization of (+)-Pinoresinol
To neutralize the variability of chemical shift resulting
from the above-mentioned factors, we introduced a simple
modification to our protocol for NMR-based stereochemical
assignment. Instead of relying on chemical shift values
relative to a reference such as the TMS or the solvent signal,
only chemical shift differences between the two diastereo-
mers observed in samples of mixtures of the two were
considered. This approach removes the effect of chemical
shift variability caused by sample-specific factors since not
only are both diastereomers affected in the same way, but
more importantly, mixing two diastereomers there are only
two possible outcomes: either there are two sets of signals
1
The H NMR spectra show minute differences in the
chemical shift values of the two diastereomers 3 and 4. Most
prominently, the doublets corresponding to the methyl groups
(5) Kasahara, H.; Miyazawa, M.; Kameoka, H. Phytochemistry 1997,
44, 1479-1482.
(6) (a) Weibel, D. B.; Walker, T. R.; Schroeder, F. C.; Meinwald, J.
Org. Lett. 2000, 2, 2381-2383. (b) Williamson, R. T.; Barrios Sosa, A.
C.; Mitra, A.; Seaton, P. J.; Weibel, D. B.; Schroeder, F. C.; Meinwald, J.;
Koehn, F. E. Org. Lett. 2003, 5, 1745-1748.
(7) Kitagawa, S.; Hisada, S.; Nishibe, S. Phytochemistry 1984, 23, 1635-
1636.
3020
Org. Lett., Vol. 6, No. 18, 2004

1
in the H NMR spectra or just one. In fact, the absolute
In all cases, reactions proceeded smoothly and produced
pairs of silylated derivatives in near 100% yield. For (S)-
citronellol, the chemical shift differences between its two
diastereomeric silylation products are quite large (Figure 3A).
magnitude of the chemical shift differences is of little
importance when following this approach, as long as there
are differences the spectrometer is able to resolve.
In preparation for the mixing experiments, NMR samples
of 3 and 4 were prepared with concentrations matching that
of the caterpillar-derived sample. Subsequently, spectra of
mixtures containing 3 and 4 in a ratio of 1:2 and 2:1 were
acquired. These spectra were then used as reference spectra
for the subsequent analysis of the caterpillar-derived material
(5), which was divided into two portions. Of these, one was
added to the sample of pure 3 and the other one was added
to the sample of pure 4. The results are shown in Figure 2.
1
Figure 2. 0.65-0.95 ppm section of H NMR spectra (benzene-
d6, 500 MHz): (A) and (B) mixtures of 3 and 4 in approximate
ratios of 1:2 and 2:1, respectively; (C) mixture of (+)-menthoxy-
diphenylsilyl derivative of (+)-pinoresinol (4) and 5; (D) mixture
of (-)-menthoxydiphenylsilyl derivative of (+)-pinoresinol (3) and
5.
While the mixture of caterpillar derived material (5) and 4
shows just one set of H NMR signals (Figure 2C), there
Figure 3. Sections of 1H NMR spectra (benzene-d6, 500 MHz (A
and B) and 600 MHz (C and D)): (A) 1:1 mixture of the (+)- and
(-)-2 derivatives of (S)-citronellol (7); (B) 2:1 mixture of the (+)-
and (-)-2 derivatives of N-acetyl-L-tyrosine ethyl ester (6); (C) 1:2
mixture of the (+)- and (-)-2 derivatives of (5S)-5,9-dimethyl-8-
decen-1-ol (8); (D) 1:1 mixture of the (+)- and (-)-2 derivatives
of 11-methyltridecanol (9).
1
are clearly two sets of signals in the spectrum of the mixture
containing 3 (Figure 2D). Since 5 was obtained through
derivatization with (-)-menthoxydiphenylsilyl chloride, it
must be the enantiomer of 4, thus unambiguously defining
the material isolated from the caterpillar secretion as (-)-
pinoresinol.
In this case, a straightforward comparison of 1H NMR spectra
would suffice in order to assign configuration. For the silyl
derivatives of N-acetyl-L-tyrosine methyl ester (6) and (5S)-
5,9-dimethyl-8-decen-1-ol (8), the chemical shift differences
In light of these results, we were interested in whether
our method is applicable to other natural products where the
site of derivatization is far removed from the chiral center(s).
For this purpose, we prepared the (+)- and (-)-isomers of
chloromenthoxydiphenylsilane derivatives of N-acetyl-L-
tyrosine ethyl ester (6), (S)-citronellol (7), (5S)-5,9-dimethyl-
8-decen-1-ol (8), which we prepared from (S)-citronellyl
bromide,8 and 11-methyltridecanol (9).
(8) S)-5,9-Dimethyl-9-decen-1-ol (8) was synthesized via a short sequence
starting with reacting diethyl malonate with (S)-citronellyl bromide. For
reaction conditions, see: Kletzke, P. G.; J. Org. Chem. 1964, 29, 1363-
1366.
Org. Lett., Vol. 6, No. 18, 2004
3021

are much smaller, as can be expected since the chiral centers
in these derivatives are further removed from each other
(Figure 3B,C). Also, it should be noted that in the case of
(8) the resulting derivatives are conformationally highly
very much the same manner as in the NMR-based procedure
described here. Of course, using NMR one does not depend
on the separation of just one peak. In most cases, several
signals in the NMR spectra of a pair of diastereomers will
show chemical shift differences that could be taken advantage
of.
1
flexible. In fact, differences in the H NMR spectra of
different samples of the (+)-2 derivative of 8 are of the same
order of magnitude as differences between the spectra
between the (+)- and (-)-derivatives. Therefore, a mere
In summary, we have shown that extremely small differ-
ences in the NMR spectra of pairs of diastereomers can be
used for unambiguous assignment of configuration. We
imagine that this method will be of great advantage especially
in situations where the stereocenter of interest is far removed
from the site of derivatization, and thus the differences in
1H chemical shift values are likely to be minute. In the
present case, the menthoxy methyl groups responsible for
the NMR signals used in this analysis are actually each
eleven bonds away from the nearest chiral center in the
pinoresinol substrate. In addition, the assignment of the
absolute configuration of pinoresinol isolated from Pieris
rapae represents the first application of our recently intro-
duced chiral silylation reagents to the derivatization of
phenols. This method should be particularly useful for the
stereochemical analysis of flavonoids and lignans, many of
which have important biological activity or are of medicinal
interest.11
1
comparison of the respective H NMR spectra would not
suffice to unambiguously determine absolute configuration
of 8 and might also not be satisfactory for the tyrosine
derivative 6. However, spectra of mixtures of the two
diastereomers of 6 and 8 immediately show that, in fact, there
are resolvable chemical shift differences (Figure 3B,C),
which could be used for configurational assignment as
described above for pinoresinol. Considering the conforma-
tional flexibility of 8, determination of its absolute config-
uration using other NMR-based methods or chromatographic
methods such as chiral GC or HPLC would seem very
difficult. However, the ultimate limitation for NMR-based
differentiation of diastereomers is determined by the resolu-
tion of available spectrometers. For example, for a 1:1
mixture of the two diasteromeric silylation products of 11-
methyltridecanol (9) even a well-resolved 1H NMR spectrum
acquired at 600 MHz shows only one set of signals (Figure
3D).9
Our simple version of NMR-based stereochemical analysis
resembles the common GC or HPLC method of co-injecting
an unknown as a mixture with a reference sample. For
example, to determine the absolute configuration of a volatile
compound by chiral GC,10 one would usually begin with
finding suitable conditions for separating the enantiomers
into two peaks. Co-injection of the sample of unknown
configuration with samples of either enantiomer would then
generate chromatograms showing either two peaks or one
peak, which then allows for assignment of configuration in
Acknowledgment. This work was supported in part by
the National Institutes of Health (GM 53830 and training
grant GM 08500). We thank Dr. Andrew E. Taggi for helpful
comments. The hospitality of the American Academy of Arts
& Sciences to J.M. during the preparation of this manuscript
is acknowledged with pleasure.
Supporting Information Available: Procedures for the
preparation of 2 and the derivatization of pinoresinol; full
spectra of mixtures of derivatives 3-5. This material is
available free of charge via the Internet at http://pubs.acs.org.
OL049233B
(9) In some cases, the use of 13C NMR spectroscopy might help to further
extend the scope of NMR-based assignment of absolute configuration.
However, the amounts of material needed would be much higher than when
(11) Lewis, N. G., Davin, L. B. In ComprehensiVe Natural Product
Chemistry; Barton, D. H. R., Nakanishi, K., Meth-Cohn, O., Eds.;
Elsevier: London, 1999; pp 639-712.
1
using H NMR.
(10) Schurig, V. TrAC, Trends Anal. Chem. 2002, 21, 647-661.
3022
Org. Lett., Vol. 6, No. 18, 2004