Bijvoet in solution reveals unexpected stereoselectivity in a michael addition

Article abstract of DOI:10.1002/chem.201002520

The absolute configuration of small crystallizable molecules can be determined with anomalous X-ray diffraction as shown by Bijvoet in 1951. For the majority of compounds that can neither be crystallized nor easily be converted into crystallizable derivat

Full text of DOI:10.1002/chem.201002520

FULL PAPER
DOI: 10.1002/chem.201002520
Bijvoet in Solution Reveals Unexpected Stereoselectivity in a Michael
Addition
Han Sun,[a] Edward J. dꢀAuvergne,[a] Uwe M. Reinscheid,*[a] Luiz Carlos Dias,*[b]
Carlos Kleber Z. Andrade,*[c] Rafael Oliveira Rocha,*[c] and Christian Griesinger*[a]
Abstract: The absolute configuration
of small crystallizable molecules can be
determined with anomalous X-ray dif-
fraction as shown by Bijvoet in 1951.
For the majority of compounds that
can neither be crystallized nor easily
be converted into crystallizable deriva-
tives, stereocontrolled organic synthesis
is still required to establish their abso-
lute configuration. In this contribution,
a new fundamental methodology for
resolving the absolute configuration
will be presented that does not require
crystallization. With residual dipolar
coupling enhanced NMR spectroscopy,
(ORD) spectra are predicted by DFT
calculations and compared to experi-
mental results. The combination of
these two steps reveals the absolute
configuration of a flexible molecule in
solution, which is a big challenge to
chiroptical methods and DFT in the
absence of NMR spectroscopy. Here
the absolute stereochemistry of the
product of a new Michael addition,
synthesized via a niobium(V) chiral
enolate, will be elucidated by using the
new methodology.
ensembles of
a limited number of
structures are created reflecting the
correct conformations and relative con-
figuration. Subsequently, from these
ensembles, optical rotation dispersion
Keywords: configuration determi-
nation · Michael addition · niobium ·
NMR spectroscopy · optical rota-
tion dispersion
Introduction
Although the determination of the relative configuration of
stereocenters was first established by Fischer in 1891
through organic synthesis,[1] the question of absolute config-
uration was not solved until the work of Bijvoet et al. in
1951 with anomalous X-ray diffraction.[2] For the majority of
compounds that can neither be crystallized nor easily be
converted into crystallizable derivatives, organic synthesis is
still required to establish their stereochemistry.[3,4] Over the
past two decades, the situation has been improved with the
development of chiroptical techniques, such as optical rota-
tion dispersion (ORD)[5] and vibrational circular dichroism
(VCD),[6] that allow the configuration of all stereogenic ele-
ments to be determined in solution. As this requires all con-
formations to be known, these techniques are limited to rel-
atively rigid compounds. For molecules with greater than
four rotatable bonds, the magnitude of the resulting confor-
mational space is immense rendering the required calcula-
tions computationally infeasible.[7–10] In this contribution, a
new fundamental methodology overcoming the flexibility
issue to resolve the absolute configuration will be presented.
Rather than sampling the entirety of the conformational
space, ensembles of ten structures derived by NMR spec-
troscopy, which reproduce the experimental NMR spectro-
[a] H. Sun, Dr. E. J. dꢀAuvergne, Dr. U. M. Reinscheid,
Prof. C. Griesinger
Department of NMR-Based Structural Biology
Max Planck Institute for Biophysical Chemistry
Am Fassberg 11, 37077 Gçttingen (Germany)
Fax : (+49)551-201-2202
E-mail: cigr@nmr.mpibpc.mpg.de
urei@nmr.mpibpc.mpg.de
[b] Prof. L. Carlos Dias
Laboratꢁrio de Sꢂntese Orgꢃnica, Instituto de Quꢂmica
Universidade Estadual de Campinas, C.P. 6154
Universidade Estadual de Campinas, 13084-971
Campinas, SP (Brazil)
Fax : (+55)19-3521-3023
E-mail: ldias@iqm.unicamp.br
[c] Prof. C. Kleber Z. Andrade, Prof. R. Oliveira Rocha
Laboratꢁrio de Quꢂmica Metodolꢁgica e Orgꢃnica Sintꢄtica
Instituto de Quꢂmica, Universidade de Brasꢂlia
Campus Universitꢅrio Darcy Ribeiro, C.P. 4478
70904-970, Brasꢂlia-DF (Brazil)
Fax : (+55)61-32734149
E-mail: rafaelrocha@unb.br
ckleber@unb.br
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.201002520.
Chem. Eur. J. 2011, 17, 1811 – 1817
ꢆ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1811

scopic data will be used to predict chiroptical properties of
the molecule in solution. The technique involves two steps:
residual dipolar coupling (RDC)-enhanced NMR spectros-
copy to enable the determination of the conformational en-
semble and the relative configuration;[11,12] and subsequent
prediction of ORD spectra by DFT calculations to differen-
tiate between enantiomers. Combined, the two steps reveal
the absolute configuration of a flexible molecule in solution.
We will show that the ORD values for the different confor-
mations in the ensemble vary tremendously and, therefore,
the determination of a faithful ensemble representing the
situation in solution is mandatory for calculating correct
ORD values with DFT. Here the absolute stereochemistry
of the product of a new Michael addition, synthesized via a
niobium(V) chiral enolate, will be elucidated using the new
methodology. The conjugated addition to a,b-unsaturated
compounds (Michael reaction) is one of the most important
Scheme 2. Possible diastereomers from the Michael addition of the nio-
bium(V) enolate of oxazolidinone (R)-1 to chalcone 2. Top left: (R,R,R)-
3, top right: (R,S,S)-3, bottom left: (R,R,S)-3, bottom right: (R,S,R)-3.
The Michael product (R)-3 (Scheme 1) was isolated with
a yield of 72%. The identity of this adduct was confirmed
by NMR spectroscopy and a high 95:5 ratio of two diaste-
reomers was determined by integration of the methyl signals
[13]
1
ꢀ
reactions in C C bond formation.
A few examples that
in the H NMR spectrum of the crude product. We are not
use metallic enolates prepared from chiral and achiral oxa-
zolidinones in conjugated addition reactions have been de-
scribed to prepare versatile intermediates in the synthesis of
compounds with interesting pharmacological activities.[14–19]
By using the new methodology we determine the relative
configuration of the reaction product, reveal with this an un-
expected stereoselectivity and provide an insight into mech-
anistic organic chemistry. In addition, we unequivocally de-
termine the absolute configuration of this Michael addition
product.
aware of any other example of an addition of a chiral oxazo-
lidinone enolate to chalcones. Furthermore, when TiCl4 was
used as the Lewis acid under similar conditions no reaction
was observed between these educts demonstrating the im-
portance of NbCl5 in this process.
Determination of the relative configuration of (R)-3: The
Michael adduct (R)-3 contains three stereocenters. The one
located in the oxazolidinone ring (C8) is known to have an
R configuration. To elucidate the relative configuration of
the two unknown stereocenters C2 and C3 in the major
product, J-coupling analysis was used to establish the anti
relative configuration in agreement with the chemical ap-
proach (see the Supporting Information). Therefore, the
possible diastereomers are reduced from four to just
(R,R,R)-3 and (R,S,S)-3, in which the stereocenters corre-
spond to C8, C2, and C3, respectively (Scheme 1). However,
the relative configuration of C2 and C3 to the known stereo-
center C8 in (R)-3 (Scheme 1) could not be defined by using
conventional NMR spectroscopic tools. To solve this prob-
lem we developed a new analysis using residual dipolar cou-
plings (RDCs) in addition to the conventional rotating-
frame Overhauser enhancement (ROE) distance restraints.
Since the two neighboring unknown stereocenters and the
known stereocenter in the oxazolidinone ring are separated
by one fully and two partially rotatable bonds, long-range
structure restraints were re-
Results
Studies for the niobium-based stereoselective Michael addi-
tion: The application of niobium(V) compounds as Lewis
acids has recently been demonstrated in a variety of organic
reactions.[20,21] Based on our previously established method-
ology for preparing niobium enolates of oxazolidinones,[22]
we decided to study the conjugated addition of these eno-
lates to a,b-unsaturated systems, especially chalcones. The
niobium(V)-catalyzed reaction is shown in Scheme 1.
This conjugate addition reaction involves the addition of
a niobium enolate of the chiral auxiliary N-propionyl oxazo-
lidinone (R)-1 to a chalcone 2. This can lead to the four dia-
stereomers shown in Scheme 2.
quired to find the correct con-
figuration. Such restraints were
provided by quantitative ROEs,
as well as by RDCs. A total
number of 40 ROEs were de-
rived from a ROESY experi-
ment with 400 ms mixing time.
A total of 36 local and remote
ROEs (see the Supporting In-
formation) were then used as
the restraints in the following
structure evaluation.
Scheme 1. Stereoselective conjugated addition of the niobium enolate of (R)-1 to chalcone (R)-3; d.r.=diaste-
reomeric ratio.
1812
www.chemeurj.org
ꢆ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2011, 17, 1811 – 1817

Bijvoet in Solution
FULL PAPER
Due to the flexibility of this molecule, conformational
averaging constitutes a methodological challenge. In this
study, to sample the conformational space we calculated a
molecular dynamics trajectory for each diastereomer in
vacuo. Then 10,000 ensembles, each of 10 conformations,
were generated for the diastereomers (R,R,R)-3 and (R,S,S)-
3. These 10 structures were randomly picked from the 1,000
snapshots of the molecular dynamics (MD) trajectory. The
ROEs and RDCs were then used to rank the ensembles of
both diastereomers according to the violation of the ROEs
and the quality of the fit to the RDCs (Q factor). The viola-
tion (U) for the ROE in ꢇ2 is defined by Equation (1):
in which j is the index over all RDCs. rcj alcd is the back-calcu-
lated ROE distance, averaged to the minus 6th power over
all members of the ensemble as shown in Equation (2):
Figure 1. The ten structures of the best ensemble of (R,R,R)-3. For all
ꢂ
ꢃ
1
ꢀ
ꢁꢀ6 ꢀ6
1
N
ꢀ
conformations, the C N bond is antiperiplanar.
1
N
calcd
j;i
rjcalcd
¼
S r
ð2Þ
i
To use the single alignment tensor approximation, the rel-
ative orientation of conformers in the ensemble is deter-
mined by superimposing each to the mean structure. For
each ensemble the ROE violation was averaged to the
minus 6th power over all members of the ensemble. The
dots in Figure 3 represent the ROE violation in ꢇ2, the Q
factor for the PH-gel data and the Q factor for the PAN-gel
data for each ensemble. From Figure 3 it is obvious that the
blue (R,R,R)-3 and the red (R,S,S)-3 datasets are well re-
solved. The ensembles of (R,R,R)-3 fulfill the ROEs and
in which i represents the members of the ensemble and N is
the total number of the structures in one ensemble. The Q
factor for the RDC is defined by Equation (3):
RDCs significantly better than those of (R,S,S)-3. For all
1
ꢀ
conformations and for both diastereomers, the C N bond is
antiperiplanar.
in which j is the index over all RDCs and the back-calculat-
ed RDC (Dcj alcd) is linearly averaged over all members of the
ensemble. The fitting for each ensemble was performed in-
dependently for the two sets of RDC data (from a poly-
acrylamide (PH) gel[23] and a poly(acrylonitrile) (PAN)
gel[24]) by calculating the alignment tensor using the Nelder–
Mead simplex optimization method.[25] The ten structures in
an ensemble were assumed to have equal populations. In
this study, we use the assumption that all the structures in
an ensemble have the same alignment tensor.[12,26] This as-
sumption is justified, since the overall shape of the confor-
mations in the ensembles is very similar (Figure 1 and
Figure 2).
Another, less robust, method is to compare the best en-
semble of (R,R,R)-3 to that of (R,S,S)-3. Here “best” is de-
fined as a combined Q factor by using quadratic averaging.
The Q factor for the ROE is defined in Equation (4):
vffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
u
U
S rjobs
u
QROE
¼
À
Á
t
ð4Þ
2
j
in which j is the index over all ROE distances and U is de-
fined in Equation (1). The combined Q factor is defined in
Equation (5):
Chem. Eur. J. 2011, 17, 1811 – 1817
ꢆ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
1813

C. Griesinger, U. M. Reinscheid, L. Carlos Dias, C. Kleber Z. Andrade, R. Oliveira Rocha et al.
(R,S,S)-3 has Q factors of 0.34 and 0.50, respectively. The
ROE violation of these two best ensembles is smaller than
3.5 ꢇ2, corresponding to a QROE factor of 0.10 [Eq. (4)].
Also the principal components of the alignment tensors Axx,
Ayy, and Azz are smaller for the best ensemble of (R,R,R)-3
than the (R,S,S)-3 ensemble (Table 1). The largest RDC
Table 1. The diagonalized alignment tensors and the Q factors of the
best ensembles of (R,R,R)-3 and (R,S,S)-3.
Configuration
(R,R,R)-3
AHCTUNGTRENN(GUN R,S,S)-3
Medium
Dxx [Hz]
Dyy [Hz]
Dzz [Hz]
Q factor
PAN gel
PH gel
PAN gel
PH gel
ꢀ37.64
ꢀ6.69
ꢀ52.05
ꢀ7.50
1.07
ꢀ0.72
2.71
ꢀ2.75
37.57
8.41
50.35
11.24
0.165
0.319
0.336
0.499
AHCTUNGTRENNUNG
value of (R)-3-anti in the PAN gel is 26.92 Hz, whereas that
of the PH gel is 7.12 Hz. In both alignment media, the Dzz
values of (R,R,R)-3 rather than (R,S,S)-3 are closer to the
maximum experimental dipolar couplings. A similar obser-
vation has been made during the analysis of previous config-
urational studies,[12] in which fitting the experimental dipolar
couplings with the wrong diastereomers yielded larger align-
ment tensors than fitting them with the correct diastereo-
mer.
1
ꢀ
Without the presence of an organometal, the C N amide
bond is in slow exchange on the H NMR spectroscopy time
1
scale at room temperature due to its double-bond character.
Since we see only one set of peaks, only one of the two pos-
Figure 2. The ten structures of the best ensemble of (R,S,S)-3. For all con-
1
ꢀ
formations, the C N bond is antiperiplanar.
sible configurations is populated. Above, we assumed that
1
ꢀ
the configuration is C N antiperiplanar. Here, we now
show that this is not only in agreement with the literature,
but is corroborated by the RDCs. We therefore repeated the
1
ꢀ
above described analysis with the C N bond in the synperi-
planar configuration and compared this with the results as-
suming the antiperiplanar configuration. From Figure 4 it is
1
ꢀ
obvious that the (R,R,R)-3 (C N antiperiplanar) configura-
Figure 3. Distinguishing diastereomers by using RDCs and ROEs. The
axes of the plot are the ROE violation in ꢇ2, the Q factor of the PH gel
RDC data and the Q factor of the PAN gel RDC data. The 10,000 en-
sembles of 2 diastereomers are symbolized in different colors: (R,R,R)-3
(blue) and (R,S,S)-3 (red). 2D projections are presented in the Support-
1
ꢀ
ing Information. For both diastereomers, the C N bond is antiperipla-
nar.
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
Q2PAN þ Q2PH þ Q2ROE
ð5Þ
Qtotal
¼
1
33
ꢀ
Figure 4. Comparison of the QPAN factors between the C
N antiperipla-
1
For (R,R,R)-3, the best ensemble has Q factors of 0.16
and 0.32 for PAN and PH gels, respectively, whereas
ꢀ
nar (green) and C N synperiplanar (blue) configuration of the diaste-
reomer (R,R,R)-3.
1814
www.chemeurj.org
ꢆ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2011, 17, 1811 – 1817

Bijvoet in Solution
FULL PAPER
tion fulfils the RDC data measured in the PAN gel much
1
ꢀ
better than the (R,R,R)-3 (C N synperiplanar) configura-
tion. This result is in agreement with the literature,[27,28]
1
ꢀ
which shows that the C N antiperiplanar configuration is
1
ꢀ
lower in energy than the C N synperiplanar configuration.
This analysis shows again the power of RDC, which could
be used to determine the relative orientation of different
moieties in one molecule. In contrast to the RDC, the NOE
is not very sensitive due to the flexibility of the molecule
and the minus 6th power dependence.
Determination of the absolute configuration: Extending the
methodology to determine the absolute configuration with-
out the knowledge that the stereochemistry of C8 is R, we
measured ORD values at 4 different wavelengths and com-
pared them with those calculated with DFT from the best
ensemble of structures as determined by NMR spectroscopy.
The MD snapshots were minimized before the DFT optimi-
zation. Table 2 and Figure 5 comprise the experimental and
Figure 6. Plot of the calculated ORD values for two different conforma-
tions, selected from the best ensemble of (R,R,R)-3.
Discussion
Table 2. Experimental and calculated ORD values at different wave-
lengths.
Determination of the absolute configuration of flexible mol-
ecules is currently a very difficult problem to resolve. As
mentioned in the introduction, chiroptical methods require
the knowledge of the conformational ensemble to be accu-
rate, which is very difficult to achieve in the absence of ex-
perimentally determined ensembles. Therefore, if the con-
formationally heterogeneous compound cannot be crystal-
lized, chemical synthesis is the last resort. Herein, we pres-
ent a new methodology using RDC-enhanced NMR spec-
troscopy to first determine the ensemble of conformations,
which reveals the relative configuration by the NMR spec-
troscopy data and gives a faithful description of the solution
ensemble. The chiroptical properties of the ensemble is then
predicted by DFT and compared to experimental data to es-
tablish the absolute configuration.
[a]DFT
l [nm]
(source)
T
[8C]
[a]exptl
(R,R,R)-3
(R,S,S)-3
E
ACHTUNGTRENN(UNG S,R,R)-3
N
589 (Na) 20.3 ꢀ35
578 (Hg) 20.6 ꢀ37
546 (Hg) 20.6 ꢀ42
436 (Hg) 20.6 ꢀ76
ꢀ59
ꢀ61
ꢀ70
82
87
104
252
59
62
70
ꢀ82
ꢀ87
ꢀ104
ꢀ252
ꢀ110
110
Our results show that for a highly flexible molecule, such
as compound (R,R,R)-3 that was synthesized by a novel
Nb(V)-catalyzed Michael addition, the relative configura-
tion can be determined by carefully measuring J-couplings,
NOE or ROE derived distances, and RDCs. Two different
alignment media were used to obtain the anisotropic RDC
data set. The distance and anisotropic data are then used to
select the best ensemble of 10 structures, whereby each
member is obtained randomly from free molecular dynamics
simulations, from a large pool of 10,000 different ensembles.
The fit of the RDCs used a single alignment tensor for all
different conformations in the ensemble: this assumption is
justified as illustrated in Figure 1 and Figure 2 from the sim-
ilar shape of all conformations. As opposed to previous ap-
proaches,[12,26] no weighting of the RDCs and ROEs is re-
quired, which establishes a more robust way of analyzing
the data. This procedure clearly shows that the ensembles
with (R,R,R)-3 configuration are in far better agreement
with the experimental data than the (R,S,S)-3 ensembles.
The stereochemical outcome ((R,R,R)-3) of the previously
unused niobium(V)-catalyzed version of the Michael addi-
Figure 5. Comparison of experimental ORD values with those calculated
by DFT for the four possible diastereomers of 3. Experimental ORD
&
curve (dashed line, ) and the calculated curves (solid lines) for the two
possible pairs of enantiomers with an anti relationship between C2 and
C3; +=(R,R,R)-3, ꢈ=(S,S,S)-3, =(R,S,S)-3, & =(S,R,R)-3.
*
calculated ORD data measured at different wavelengths.
The calculated ORD values for the different conformation
in one ensemble are highly divergent and even change sign
as demonstrated by Figure 6.
Chem. Eur. J. 2011, 17, 1811 – 1817
ꢆ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
1815

C. Griesinger, U. M. Reinscheid, L. Carlos Dias, C. Kleber Z. Andrade, R. Oliveira Rocha et al.
tion is quite surprising. Expected transition states of this re-
mined to be (R,R,R)-3. This answer to the hypothetical
question required no information about the configuration of
the auxiliary. Thus, the finding of the configuration of the
auxiliary used in the reaction, namely R, shows that the ap-
proach is useful for flexible molecules in solution. It should
be noted from Figure 6 that the calculated ORD values of
each member of the ensemble are highly divergent. Approx-
imately half of the conformations in the ensemble yield
single conformation ORD values that would suggest the
(S,S,S)-3 configuration. Only when we use the correct en-
semble of conformations and average over the individual
ORD values do we get the correct answer. Hence an accu-
rate ensemble approximating the full conformational space
of the molecule is essential for the calculation of the chirop-
tical properties. Clearly, ensembles without NMR spectros-
copy distance and RDC restraints would not be sufficient.
Thus, combining diastereomer differentiation by using iso-
tropic and anisotropic NMR spectroscopy data with enantio-
mer differentiation by experimental validation of the DFT-
calculated chiroptical properties of ensembles derived from
NMR spectroscopy, the hypothetical absolute stereochemis-
try question is unequivocally and indisputably answered,
and is (R,R,R)-3. This determination of absolute configura-
tion in solution establishes “Bijvoet in solution”. We coined
this term, since anomalous X-ray diffraction, the key to the
determination of absolute configuration from crystals, is a
chiroptical effect for X-ray radiation. This effect together
with the known conformation and relative configuration es-
tablished from conventional X-ray diffraction from crystals
allows the determination of the absolute configuration. With
Bijvoet in solution, we establish the conformational ensem-
ble in solution and the relative configuration by NMR spec-
troscopy. We then use chiroptical methods to establish the
absolute configuration.
action based on the literature[14–19] would lead to the (R,S,S)-
3 isomer (see the Supporting Information). Since the
(R,R,R)-3 isomer is formed instead, we suggest four new
transition states with antiperiplanar and synclinal attack on
the Si and Re faces of the chalcone (see the Supporting In-
formation). In all cases, niobium is complexed to the car-
bonyl group of the chalcone, which promotes its activation
and forms 8-membered cyclic transition states in that the
endo carbonyl group of the oxazolidinone is opposite to the
enolate to minimize electronic effects, such as the dipole
moment (Scheme 3). The lack of activity of TiCl4 in this re-
Scheme 3. Chair-like transition state (TS) for (R,R,R)-3.
action may be due to its inability to simultaneously complex
to both carbonyl groups of the oxazolidinone and the chal-
cone. Whether these transition states are indeed formed in
this reaction remains to be seen. They are, nevertheless,
chemically plausible and lead to the observed products.
Finally, we addressed the question of the absolute config-
uration of the Michael addition product formed in the reac-
tion presented in Scheme 1. Although C8 is known to be in
the R configuration, since the configuration of the chiral
auxiliary was known a priori, here we present the hypotheti-
cal that all stereocenters are unknown and attempt to
answer the absolute configuration question. This can be con-
sidered to be a test for the accuracy of the conformational
ensemble. In addition, it would establish the combination of
the ensembles determined by NMR spectroscopy and chi-
roptical measurement as a versatile method for the determi-
nation of the absolute configuration of flexible molecules.
Firstly, the J coupling shows that C2 and C3 must be
either R,R or S,S. Secondly, the combined ROE- and RDC-
based analysis provided the diastereomer (R,R,R)-3 over
(R,S,S)-3, but hypothetically, compound (R,R,R)-3 cannot be
distinguished from its (S,S,S)-3 enantiomer. To answer this
question, chiroptical measurements were employed. The chi-
roptical properties of the ensembles selected from the NMR
spectroscopic analysis were calculated by DFT and com-
pared to measured ORD values.
Conclusion
We have introduced a robust way of combining RDC-en-
hanced NMR spectroscopy for the determination of the rel-
ative configuration of flexible molecules with ORD meas-
urements in conjunction with predicted ORD values from
the ensemble derived by NMR spectroscopy. Since the
ORD values vary tremendously when the molecule adopts
different conformations, the determination of a faithful en-
semble representing the situation in solution is mandatory
(of course the ensemble could also be obtained by other
methods, not necessarily exactly in the way we propose).
This allowed us to determine the correct relative and abso-
lute configuration of the synthetic product (R,R,R)-3 in solu-
tion. Compound 3 was obtained in a new niobium(V)-based
The ORD values are shown in Figure 5 and summarized
in Table 2. Taking large errors in the calculated values into
account, the measured negative values are compatible with
both (R,R,R)-3 and (S,R,R)-3 configurations. Since the rela-
tive configuration of (S,R,R)-3 is incompatible with the
NMR spectroscopic analysis, and could therefore be elimi-
nated, the absolute configuration is unambiguously deter-
ꢀ
Michael-type C C bond-forming reaction, which requires a
b-substituted Michael acceptor. The reaction was highly ste-
reoselective and gave an unexpected stereochemical result
that could be rationalized by a plausible transition state.
1816
www.chemeurj.org
ꢆ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2011, 17, 1811 – 1817

Bijvoet in Solution
FULL PAPER
[9] E. Giorgio, M. Roje, K. Tanaka, Z. Hamersak, V. Sunjic, K. Naka-
nishi, C. Rosini, N. Berova, J. Org. Chem. 2005, 70, 6557.
[10] P. J. Stephens, F. J. Devlin, J. R. Cheeseman, M. J. Frisch, J. Phys.
Chem. A 2001, 105, 5356.
[11] C. M. Thiele, A. Marx, R. Berger, J. Fischer, M. Biel, A. Giannis,
Angew. Chem. 2006, 118, 4566; Angew. Chem. Int. Ed. 2006, 45,
4455.
[12] A. Schuetz, J. Junker, A. Leonov, O. Lange, T. Molinski, C. Grie-
singer, J. Am. Chem. Soc. 2007, 129, 15114.
[13] M. Christoffers, Synlett 2001, 0723.
[14] T. Yura, N. Iwasawa, T. Mukaiyama, Chem. Lett. 1987, 791.
[15] T. Mukaiyama, N. Iwasawa, T. Yura, R. S. J. Clark, Tetrahedron
1987, 43, 5003.
[16] T. Yura, N. Iwasawa, T. Mukaiyama, Chem. Lett. 1988, 1021.
[17] D. A. Evans, M. T. Bilodeau, T. C. Somers, J. Clardy, D. Cherry, Y.
Kato, J. Org. Chem. 1991, 56, 5750.
Experimental Section
General procedure: Oxazolidinone (R)-1 (0.23 g, 1.0 mmol) diluted in
dry dichloromethane (1 mL) was added to a suspension of NbCl5 (0.35 g,
1.3 mmol) in dry dichloromethane (1.0 mL) at 08C. The reaction was vig-
orously stirred for 5 min. Then triethylamine (0.35 mL, 2.5 mmol) was
added dropwise, resulting in a dark brown solution. After stirring for
5 min at 08C, a solution of chalcone 2 (1.1 mmol) in dry dichloromethane
(1.0 mL) was added and the reaction was monitored by TLC. The reac-
tion was stirred for 3 h at 08C and 72 h at room temperature and then sa-
turated ammonium chloride solution (NH4Cl, 10 mL) was added. The re-
action mixture was extracted with CH2Cl2 (3ꢈ15 mL). The organic layers
were washed with brine (10 mL), dried over Na2SO4 and concentrated in
vacuum. Flash chromatography (hexanes/EtOAc=60:40) afforded pure
Michael product (R)-3 in 72% yield as a yellow oil.
MD simulations: To sample the conformational space of this flexible
molecule, we calculated the molecular dynamics trajectory of each diaste-
reomer with the DISCOVER software (Biosym Technologies, San Diego,
CA, USA) using the consistent valence force field (CVFF),[29,30] a stan-
dard force field for small molecules. The MD-simulation was carried out
at 298 K for 1 ns. A single structure was logged every 1 ps so that a tra-
jectory with 1,000 structures was obtained.
[18] D. A. Evans, T. C. Somers, M. T. Bilodeau, F. Urpi, J. S. Clark, J.
Am. Chem. Soc. 1990, 112, 8215.
[19] J. Bach, R. Berenguer, J. Garcia, J. Vilarrasa, Tetrahedron Lett.
1995, 36, 3425.
[20] C. K. Z. Andrade, Curr. Org. Synth. 2004, 1, 333.
[21] C. K. Z. Andrade, R. O. Rocha, Mini Rev. Org. Synth. 2006, 3, 271.
[22] C. K. Z. Andrade, R. O. Rocha, L. M. Alves, P. P. Kallil, C. M. A.
Panisset, Lett. Org. Chem. 2004, 1, 109.
[23] P. Haberz, J. Farjon, C. Griesinger, Angew. Chem. 2005, 117, 431;
Angew. Chem. Int. Ed. 2005, 44, 427.
RDC and ROE analysis: Superimposing each conformer to the mean
structure was performed in the program MOLMOL2.1.[31] The full ROE
and RDC analysis was implemented in the program relax.[32,33]
[24] G. Kummerlçwe, J. Auernheimer, A. Leidlein, B. Luy, J. Am. Chem.
Soc. 2007, 129, 6080.
[25] A. Schuetz, T. Murakami, N. Takada, J. Junker, M. Hashimoto, C.
Griesinger, Angew. Chem. 2008, 120, 2062; Angew. Chem. Int. Ed.
2008, 47, 2032.
[26] J. A. Nelder, R. Mead, Comp. J. 1965, 7, 308.
[27] B. S. Lou, G. Li, F. D. Lung, V. J. Hruby, J. Org. Chem. 1995, 50,
5509.
[28] G. Li, D. Patel, V. J. Hruby, Tetrahedron Lett. 1994, 35, 2301.
[29] A. T. Hagler, E. Huler, S. Lifson, J. Am. Chem. Soc. 1974, 96, 5319.
[30] A. T. Hagler, S. Lifson, P. Dauber, J. Am. Chem. Soc. 1979, 101,
5122.
ORD analysis: The ten structures of the best ensemble of (R,R,R)-3 and
(R,S,S)-3 were used as starting structures for DFT geometry optimization
using Gaussian03 Revision C.02.[34] The optimizations were performed at
the B3LYP/6–31G(d) level of theory. The optical rotation dispersion cal-
culations at the four wavelengths 436, 546, 578, and 589 nm were per-
formed with the optimized structures as input coordinates with the same
basis set as the optimizations using the integral equation formalism var-
iant polarizable continuum model (IEFPCM) solvent continuum model
as implemented in Gaussian03 with DMSO as the solvent.
Acknowledgements
[31] R. Koradi, M. Billeter K. Wꢉthrich, J. Mol. Graphics 1996, 14, 51.
[32] E. J. dꢀAuvergne, P. R. Gooley, J. Biomol. NMR 2008, 40, 107.
[33] E. J. dꢀAuvergne, P. R. Gooley, J. Biomol. NMR 2008, 40, 121.
[34] Gaussian 03, Revision C.02, M. J. Frisch, G. W. Trucks, H. B. Schle-
gel, G. E. Scuseria, M. A. Robb, J. R. Cheeseman, J. A. Montgom-
ery, Jr., T. Vreven, K. N. Kudin, J. C. Burant, J. M. Millam, S. S.
Iyengar, J. Tomasi, V. Barone, B. Mennucci, M. Cossi, G. Scalmani,
N. Rega, G. A. Petersson, H. Nakatsuji, M. Hada, M. Ehara, K.
Toyota, R. Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda,
O. Kitao, H. Nakai, M. Klene, X. Li, J. E. Knox, H. P. Hratchian,
J. B. Cross, V. Bakken, C. Adamo, J. Jaramillo, R. Gomperts, R. E.
Stratmann, O. Yazyev, A. J. Austin, R. Cammi, C. Pomelli, J. W.
Ochterski, P. Y. Ayala, K. Morokuma, G. A. Voth, P. Salvador, J. J.
Dannenberg, V. G. Zakrzewski, S. Dapprich, A. D. Daniels, M. C.
Strain, O. Farkas, D. K. Malick, A. D. Rabuck, K. Raghavachari,
J. B. Foresman, J. V. Ortiz, Q. Cui, A. G. Baboul, S. Clifford, J. Cio-
slowski, B. B. Stefanov, G. Liu, A. Liashenko, P. Piskorz, I. Komaro-
mi, R. L. Martin, D. J. Fox, T. Keith, M. A. Al-Laham, C. Y. Peng,
A. Nanayakkara, M. Challacombe, P. M. W. Gill, B. Johnson, W.
Chen, M. W. Wong, C. Gonzalez, J. A. Pople, Gaussian, Inc., Wall-
ingford CT, 2004.
R.O.R., C.K.Z.A., and L.C.D. thank IQ-UnB, IQ-UNICAMP, CAPES,
CNPq, FINATEC, CT-Infra 0970/01 and C.G. thanks the Max Planck So-
ciety, the Deutsche Forschungsgemeinschaft (GRK 782 and FOR 934)
and the Fonds der Chemischen Industrie for financial support. The au-
thors thank Prof. Michael Reggelin for the ORD measurement and Dr.
Burkhard Luy for providing the PAN gel.
[1] E. Fischer, Ber. Dtsch. Chem. Ges. 1891, 24, 1836.
[2] J. M. Bijvoet, A. F. Peerdeman, A. J. Van Bommel, Nature 1951, 168,
271.
[3] W. Youngsaye, J. T. Lowe, F. Pohlki, P. Ralifo, J. S. Panek, Angew.
Chem. 2007, 119, 9371–9374; Angew. Chem. Int. Ed. 2007, 46, 9211.
[4] E. L. Eliel, S. H. Wilen, Stereochemistry of Organic Compounds,
Wiley, New York, 1994.
[5] L. D. Barron, Molecular Light Scattering and Optical Activity, Cam-
bridge University Press, Cambridge, 2004.
[6] P. L. Polavarapu, Vibrational Spectra: Principles and Applications
with Emphasis on Optical Activity (Studies in Physical and Theoreti-
cal Chemistry, Vol. 85), Elsevier, Amsterdam, 1998.
[7] T. D. Crawford, Theor. Chem. Acc. 2006, 115, 227.
[8] P. J. Stephens, D. M. McCann, J. R. Cheeseman, M. J. Frisch, Chirali-
ty 2005, 17, S52.
Received: August 31, 2010
Published online: January 12, 2011
Chem. Eur. J. 2011, 17, 1811 – 1817
ꢆ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
1817