Routine use of natural abundance deuterium NMR in a polypeptidic chiral oriented solvent for the determination of the enantiomeric composition of chiral building blocks.

Article abstract of DOI:10.1021/ol020038b

[reaction: see text] Natural abundance deuterium 2D NMR spectroscopy in chiral liquid crystal was successfully used to efficiently analyze the enantiomeric composition of organic chiral building blocks involved in the syntheses of natural and synthetic bioactive products. The results reported here emphasize the high potential of this analytical strategy and prove its applicability for routinely determining enantiomeric excesses.

Full text of DOI:10.1021/ol020038b

ORGANIC
LETTERS
2002
Vol. 4, No. 10
1663-1666
Routine Use of Natural Abundance
Deuterium NMR in a Polypeptidic Chiral
Oriented Solvent for the Determination
of the Enantiomeric Composition of
Chiral Building Blocks
,†
Arnaud Parenty,† Jean-Marc Campagne,* Christie Aroulanda,‡ and
,‡
Philippe Lesot*
ICSN-CNRS, AVenue de la Terrasse, 91198 Gif sur YVette, France, and Laboratoire de
Chimie Structurale Organique, CNRS ESA 8074, ICMO, Baˆt. 410, UniVersite´ de
Paris-Sud, 91405 Orsay, France
philesot@icmo.u-psud.fr
Received February 13, 2002
ABSTRACT
Natural abundance deuterium 2D NMR spectroscopy in chiral liquid crystal was successfully used to efficiently analyze the enantiomeric
composition of organic chiral building blocks involved in the syntheses of natural and synthetic bioactive products. The results reported here
emphasize the high potential of this analytical strategy and prove its applicability for routinely determining enantiomeric excesses.
Organic chiral building blocks play an essential role in the
development of rapid and efficient syntheses of natural and
synthetic bioactive products. In connection with an ongoing
total synthesis project, we have identified 1 as a potential
chiral building block for the synthesis of several natural
products possessing interesting biological activities, such
as lasiodiplodin,1 dolatrienoic acid,2 dehydrocurcuvalin,3
macrolactines,4 leukotrienes,5 diplodialide A,6 brefeldin A,7
and mutolide8 (Scheme 1). Synthesis of enantiomerically pure
compound 1 from smaller chiral building blocks has been
previously described but requires lengthy routes (from
lactate2a or glutamate2b esters) or expensive starting material
(chiral nonracemic propylene oxide9). Alternatively, previ-
ously described stoichiometric10 or catalytic11 asymmetric
(4) Jaruchoktaweechai, C.; Suwanborirux, K.; Tanasupawatt, S.; Kitta-
koop, P.; Menasveta, P. J. Nat. Prod. 2000, 63, 984-986 and references
therein.
(5) Vidal, A.; Durand, T.; Vidal, J.-P.; Rossi, J.-C.; Ravasi, S.; Zarini,
S.; Sala, A. Bioorg. Med. Chem. Lett. 2000, 10, 665-668.
(6) Ireland, R. E.; Brown, F. R. J. Org. Chem. 1980, 45, 1868-1880.
(7) Fox, B. M.; Vroman, J. A.; Fanwick, P. E.; Cushman, M. J. Med.
Chem. 2001, 44, 3915-3924.
(8) Bode, H. B.; Walker, M.; Zeeck, A. Eur. J. Org. Chem. 2000, 8,
1451-1456.
(9) Millar, J. G.; Oehlschager, A. C. J. Org. Chem. 1984, 49, 2332-
2338.
* To whom correspondence should be addressed. (for the synthetic part
of this work: Email: Jean-Marc.Campagne@icsn.cnrs-gif.fr).
† ICSN-CNRS.
‡ Universite´ de Paris-Sud.
(1) Fuerstner, A.; Thiel, O. R.; Kindler, N.; Bartkowska, B. J. Org. Chem.
2000, 65, 7990-7995 and references therein.
(2) (a) Duffield, J. J.; Pettit, G. R. J. Nat. Prod. 2001, 64, 472-479. (b)
Moune´, S.; Niel, G.; Busquet, M.; Eggleston, I.; Jouin, P. J. Org. Chem.
1997, 62, 3332-3339.
(3) Liu, Y.; Li, Z.; Vederas, J. C. Tetrahedron 1998, 54, 15937-15958.
10.1021/ol020038b CCC: $22.00 © 2002 American Chemical Society
Published on Web 04/26/2002

efficiency of natural abundance deuterium NMR (NAD
NMR) in polymeric liquid-crystalline solvents made of poly-
γ-benzyl-L-glutamate (PBLG) dissolved in organic solvents
(CDCl3, DMF, etc.) and its application for routine analyses.14
This spectroscopic strategy provides an interesting and
original alternative when conventional NMR methods used
routinely in the laboratory (chiral solvating agents, chiral-
lanthanide shift reagents, etc.)15 fail or give rather poor
analytical results. In this technique, the spectroscopic enan-
tiodiscrimination principle is based on the fact that two
enantiomers embedded in a chiral oriented solvent are not
ordered in the same way, thus yielding the doubling of
spectra (one for each enantiomer).13 NAD NMR offers the
major advantage that neither chemical modification nor
isotopic labeling of solutes to be studied is required, while
all deuterated sites are simultaneously probed, thus increasing
the possibility to observe a chiral differentiation. Moreover,
previous work has shown that despite the low natural
abundance of deuterium nuclei (0.015%), routine spectrom-
eters (400 MHz) were able to record overnight 1D or 2D
NAD spectra with sufficient signal-to-noise (S/N) ratio to
afford workable analytical information for (bio)chemists.14
Scheme 1
reductions were moderately efficient in terms of enantio-
selectivity (78-83% enantiomeric excess). Consequently, we
attempted to devise an efficient, rapid, inexpensive, and
scalable route to enantiomerically pure 1 (Scheme 2).
The analysis of H-{1H} spectra acquired at natural
2
abundance level in the PBLG phase is rather simple. Indeed,
as a result of the absence of H-2H couplings, the spectra
2
consist of the superposition of independent quadrupolar
doublets (nuclei with spin I ) 1) corresponding to all
nonequivalent isotopomers in the mixture.14,15 The separation
between two components, centered on δ H
referred to as the quadrupolar splitting, denoted ∆νQ. When
Scheme 2
aniso
(≈ δ Hiso), is
2
1
the discrimination between enantiomers occurs, we observe
R
generally two doublets (∆νQS and ∆νQ ) centered at the same
frequency, one for each enantiomer.13,14
2
Figure 1a shows the 61.4-MHz natural abundance H-
{1H} signal of the methyl group 1 of 4 (enantioselective
route) dissolved in the PBLG/CHCl3 system at 300 K (see
numbering in Scheme 2).16 This NAD spectrum (in fact a
sum of columns) was extracted from a 2D autocorrelation
experiment named Q-COSY, which has been specially
designed to facilitate the analysis of overcrowded NAD
spectra,14 but in this example the analysis of the 1D NAD
spectrum would be possible without the help of a 2D
experiment. The NAD signals of the methyl group consist
of one intense quadrupolar doublet ascribed to the majority
isomer, R-(+)-4, while the minority isomer exhibits a
Following an asymmetric three-step route, access to 1 was
achieved as follows. Ketone 3 was obtained in 79% yield
by treating the lithium salt of commercial pentyne 2 with
acetic anhydride. The ketone 3 was then asymmetrically
reduced using 1% of Noyori’s hydrogen transfer catalyst12
to give alcohol 4 in 30-60% isolated yield.
At this stage of the synthesis, it was crucial to determine
accurately the enantiomeric excess (ee) of 4. For this purpose,
we turned our attention to NMR in chiral liquid crystals
(CLCs), which has been recently proposed in the field of
enantiomeric analysis.13 In particular we have explored the
(13) (a) Sarfati, M.; Lesot, P.; Merlet, D.; Courtieu, J. Chem. Commun.
2000, 2069-2081, and refs. cited. (b) Aroulanda, C.; Sarfati, M.; Courtieu,
J.; Lesot, P. Enantiomer 2001, 6, 281-287.
(14) (a) Lesot, P.; Merlet, D.; Loewenstein, A.; Courtieu, J. Tetrahe-
dron: Asym. 1998, 9, 1871-1881. (b) Merlet, D.; Ancian, B.; Courtieu, J.;
Lesot, P. J. Am. Chem. Soc. 1999, 121, 5249-5258.
(15) Parker, D. Chem. ReV. 1991, 91, 1441-1457.
(16) Sample Preparation and NMR. The CLC NMR samples of
R-(+)-4 and (()-4 were made from 100 mg of a chiral material, 100 mg of
PBLG (MW ≈ 120 000), and 350 mg of dry CHCl3. Hereafter we will use
the notation: amount (in mg) of “solute/polymer (DP)/cosolvent”. PBLG
is available from Sigma. The sample preparation is described in ref 13a.
NMR experiments were performed on a DRX-400 (using a 5 mm selective
probe for 2H and BBI probe for 13C). Broad-band 1H decoupling was applied
using the WALTZ sequence. All 2D Q-COSY spectra were zerofilled to 1
k (t1) × 2 k (t2) data points prior to the double FT. Exponential filtering in
both dimensions (LB1,2 ) 1 Hz) was applied.
(10) Marx, K.-H.; Raddatz, P.; Winterfeldt, E. Liebigs Ann. Chem. 1984,
3, 474-482.
(11) Almqvist, F.; Torstensson, L.; Gudmundsson, A.; Fredj, T. Angew.
Chem., Int. Ed. Engl. 1997, 36, 376-377.
(12) Matsumara, K.; Hashiguchi, S.; Ikariya, T.; Noyori, R. J. Am. Chem.
Soc. 1997, 119, 8738-8739.
1664
Org. Lett., Vol. 4, No. 10, 2002

This is because the parameters governing the strength of the
13C CSA mainly increase with the electronegativity of the
substituents as well as the hybridization state of the carbons
(|∆σ(sp)| > |∆σ(sp2)| > |∆σ(sp3)|).13 In practice the
spectroscopic enantiodiscrimination for a given carbon, i,
occurs when two independent resonances separated by the
quantity |∆σiS - ∆σiR| are observed.
Figure 1c and d shows the 100.6-MHz 13C-{1H} signals
of the acetylenic carbon (4) for R-(+)-4 and (()-4, respec-
tively. For this carbon, the difference of CSA causes a
chemical shift difference between the two enantiomers of
0.08 ppm (∼8 Hz). Note that the R-(+)-4 isomer exhibits
the most shielded signal, but this result is not general, because
there is no straightforward correlation between the absolute
configuration of enantiomers and their respective 13C chemi-
cal shifts. Spectral enantioseparations of 4 and 8 Hz were
also observed on C-1 and C-3 carbons, respectively (not
shown). As previously, the large difference in peak areas
for the enriched mixture of 4 indicates a strong ee, while no
difference was measured for (()-4 within the experimental
accuracy. This enables a simple and easy measurement of
the ee. Here again the ee value of the enriched mixture
obtained by simple peak integration of 13C-{1H} signals
(averaged value on a series of integrations) or deconvolution
is 95 ( 3%. The smaller error compared with the value
calculated from the NAD spectrum is due to a better S/N
ratio on the 13C-{1H} spectrum. This result confirms the
value determined using NAD NMR and therefore validates
the measurements performed by this technique.
Although already high, we have attempted to improve the
enantioselectivity of the asymmetric reduction. Indeed, by
carefully controlling the experimental conditions (ligand
purity, degazing process, etc.), only the signal of the enriched
enantiomer was spectroscopically detected using NAD NMR
in PBLG/CHCl3 (not shown). Results obtained by 13C-{1H}
NMR show a very weak signal for the minority S-(-)-isomer
that does not clearly emerge from noise. Both experimental
results imply therefore an ee g 97%. R-(+)-4 was then
submitted to an alkyne zipper reaction17 to give R-(-)-1 in
a 68% yield (Scheme 2). Although the alkyne zipper reaction
has been previously described as a racemization-free process,17b
we were keen to check it out and test again the potential of
NMR in CLC.
Figure 1. (a) Natural abundance 2H-{1H} signals associated with
deuterons of the methyl group 1 and (c) 13C-{1H} signals of C-4
recorded for R-(+)-4. The 2D Q-COSY spectrum (magnitude mode)
was acquired in 15 h, using 279 (t1) × 1760 (t2) data points with
400 scans per t1 increments. The recycle delay (relaxation +
acquisition) was ∼0.6 s. The 13C-{1H} spectrum was recorded in
2 h by adding 2500 scans of 32 k data points and zerofilled to 64
k points to increase the digital resolution. The interferogram was
acquired using a pulse angle of ∼60° and a recycle delay of ∼3 s.
1H decoupling was applied during the totality of the experiment in
order to benefit from the nuclear Overhauser effect. No digital
filtering was applied. (b and d) Same experimental conditions as
in a and c, but using (()-4.
quadrupolar doublet with very low intensity. To validate this
analysis, we have synthesized compound 4 using a racemic
route and then recorded the 2D NAD Q-COSY spectrum of
(()-4 with the same experimental conditions (NMR sample
and spectroscopic parameters). In contrast with the NAD
spectrum of R-(+)-4, we observe two intense doublets of
same area, indicating that the ee is null as expected for a
racemic mixture (see Figure 1b). Note that other chiral
deuterated isotopomers show a chiral discrimination, but the
methyl site is interesting for the ee measurement. Indeed it
possesses three magnetically equivalent deuterons that
contribute to the doublet intensity (optimal S/N ratio), and
in this example it exhibits a large difference of quadrupolar
splittings, allowing one to easily integrate the signal. The
ee value for R-(+)-4 calculated by simple peak integration
of methyl group signals (averaged value on a series of
integrations) or by a deconvolution process is 95 ( 5%.
One could argue that the accuracy of the ee is rather low
as a result of the extremely weak deuterium sensitivity at
In contrast with compound 4, the 13C-{1H} NMR for 1
gave poor spectral enantiodifferentiation (|∆σiS - ∆σiR| <
2 Hz) on all carbons of the molecule. This lack of separation
indicates that the differential ordering effect (DOE) is too
small to be clearly detected through a difference of 13C CSA
(at least at 100 MHz); this also includes the acetylenic
carbons C-6 and C-7. This situation probably arises because
the triple bond is now separated from the stererogenic center
by a chain of four σ bonds exhibiting complex conforma-
tional dynamics. Even if 13C-{1H} NMR is certainly
superior to the NAD NMR in terms of sensitivity and
experimental time, this tool is not convenient to determinate
1
natural abundance level (1.45 × 10-6 with respect to H).
To get clarification on this important point, proton-decoupled
carbon-13 (13C-{1H}) spectra for R-(+)-4 and (()-4 in the
PBLG system were recorded.15,20 In 13C-{1H} NMR, the
dominant order-dependent interaction is the 13C chemical
shift anisotropy (CSA) denoted ∆σ. 13C CSA is particularly
large for doubly and triply bonded and aromatic carbons.
(17) (a) Abrams, S. R.; Shaw, A. C. Organic Syntheses; Wiley: New
York, 1993; Collect. Vol. VIII, pp 146-152. (b) Midland, M. M.;
Halterman, R. L. Tetrahedron Lett. 1981, 22, 4171-4172.
Org. Lett., Vol. 4, No. 10, 2002
1665

In the NAD spectrum reported in Figure 2b, a single
quadrupolar doublet is now visible, implying a large ee. As
observed for compound 4, the outer quadrupolar doublet
corresponds to the R-(-)-1 isomer. This fact is, however,
fortuitous, because to date there is no simple correlation
between the R/S (or () descriptors, the type of molecules
studied, and the magnitude of quadrupolar splittings.19
Considering the results obtained with the derivative 4, the
absence of signal for the S-(+)-isomer enables an estimate
that the ee is > 97 %. This result is, therefore, fully consistent
with that published by Midland et al., who described a
racemization-free process for the alkyne zipper reaction.17b
Figure 2. 61.4-MHz natural abundance 2H-{1H} signals associated
with one of the diastereotopic deuterons of the methylene 3 for
(()-1 (a) and R-(-)-1. (b) Dissolved in the PBLG/CHCl3 system
at 300 K. Both 2D Q-COSY spectra (magnitude mode) were
recorded in 14 h using identical experimental conditions (211 (t1)
× 1600 (t2) data points, 320 scans per t1 increments, recycle delay
) 0.6 s). The small central peak (labeled by an asterisk) is an
artifact.
This second example illustrates that the NAD NMR in
PBLG can be considered as a highly valuable tool in the
field of the enantiomeric analysis when other methods fail.
The strong potential of this approach is mainly due to the
2
large sensitivity of H quadrupolar interaction to DOEs,
compared to other anisotropic interactions (dipolar coupling
or CSA). Thus even though the DOE between the S and R
isomers is weak, their residual quadrupolar splittings may
be sufficiently different to exhibit a spectroscopic enantio-
separation. When an unsuccessful result is obtained using
NAD NMR, the sample temperature, the composition of the
mixture, the organic solvent, or the polypeptide can be
changed.13
the ee of 1, and consequently NAD NMR appears to be a
very interesting alternative in this case.
2
Figure 2a and b reports the natural abundance H-{1H}
signal associated with one of the deuterons of the methylene
group 3 for (()-1 and R-(-)-1, respectively.18 The assign-
1
ment of this methylene was based on the analysis of H
In summary, we have described an efficient asymmetric
and multigram access to enantiomerically pure chiral building
block 1. Moreover, this study has clearly illustrated that NAD
NMR in CLC can be routinely used, because it provides an
efficient and practical solution for the enantiomeric analysis
of chiral building blocks in which the control of the
stereogenic carbons is crucial. Results reported here demon-
strate also the outstanding capacity of the NMR in weakly
oriented chiral medium in this respect, mainly due to the
important flexibility of this tool compared with other
isotropic NMR methods. The amount of solutes used in this
work (100 mg) for NAD experiments can appear discourag-
ing. That is why, using routine spectrometers (9.4 T), this
tool is highly adapted for the analysis of chiral building
blocks (MW ≈ 150), when a large amount of compound is
still available. No doubt the use of higher field NMR
spectrometers and/or cryogenic probe systems will overcome
this disadvantage and greatly improve the efficiency of this
technique for analyzing larger solutes. Finally, note that the
enantiomers can always be extracted from the NMR sample.
COSY and HMQC 2D spectra, safely assuming that the
δ H
aniso
is very close to δ Hiso. We first discuss the left trace
2
1
(a) in the Figure 2 ((()-1). Although various sites show a
discrimination, we have focused our analysis on this meth-
ylene because it offers a large spectral enantiodifferentiation
S
R
(|∆νQ - ∆νQ | ≈ 30 Hz) for one of its diastereotopic
deuteron. Even if a single deuteron participates in the signal’s
intensity, this situation remains more attractive compared to
that provided, for instance, by the methyl group’s deuterons,
R
in which case |∆νQS - ∆νQ | < 7 Hz. The two deuterons in
a methylene group are nonequivalent in chiral molecules.
This causes in principle two pairs of two quadrupolar
splittings (one for each enantiomer when the chiral discrimi-
nation occurs) centered on two distincts chemical shifts. In
aniso
2
this example, the difference of δ H
between the two
diastereotopic deuterons is around 0.1 ppm (∼6 Hz).
Although this difference is small, the Q-COSY spectrum
allows the selection, using adequate slices of the 2D data
set, of one of the diasterotopic deuterons. Figure 2a and b
presents only slices for the deuteron of the methylene 3
exhibiting a differentiation, the other one showing no
discrimination. This situation perfectly reflects the versatility
of chiral discrimination in CLC when two distinct chiral
isotopomers are considered. It also proves the undeniable
advantage of the NAD NMR that probes simultaneously all
possible deuterated sites of the chiral molecule.
Supporting Information Available: Experimental pro-
cedures for the synthesis of compounds 1 and 4. This material
is available free of charge via the Internet at http://pubs.acs.org.
OL020038B
(18) Sample composition of R-(-)-1 and (()-1: 80/100 (562)/350 and
100/100 (562)/380; see ref 16 for notation.
(19) Chalard, P.; Bertrand, M.; Canet, I.; The´ry, V.; Remuson, R. R.;
Jeminet, G. Org. Lett. 2000, 2, 2431-2434.
1666
Org. Lett., Vol. 4, No. 10, 2002