Origin of diastereocontrol in the oxy-michael reactions of δ-lactol anions: A computational and experimental study

Article abstract of DOI:10.1002/chem.200801368

The diastereoselectivity in the alkylation and Michael addition of naked 6-substituted δ-lactolates has been studied by density functional (B3LYP) calculations with ab initio (MP2) energy refinements. The resulting proposed model for the origins of stereocontrol in this reaction has been tested by experiment. The reactions lead to a high cis diastereoselectivity across the THP ring due to the preference for both the alkoxide and the 6-substituent to sit equatorial in the alkylation transition structure. In the oxyMichael addition of these lactolates to β-substituted nitroolefins, we propose that the high diastereoselectivity β- to the nitro group is a result of a combination of steric, stereoelectronic and solvation factors.

Full text of DOI:10.1002/chem.200801368

FULL PAPER
DOI: 10.1002/chem.200801368
Origin of Diastereocontrol in the Oxy-Michael Reactions of d-Lactol Anions:
A Computational and Experimental Study
Robert D. Richardson,*[b] Felix A. Hernandez-Juan,[a] John W. Ward,[a] and
Darren J. Dixon*[a]
Abstract: The diastereoselectivity in
the alkylation and Michael addition of
“naked” 6-substituted d-lactolates has
been studied by density functional
(B3LYP) calculations with ab initio
(MP2) energy refinements.The result-
ing proposed model for the origins of
stereocontrol in this reaction has been
tested by experiment.The reactions
lead to a high cis diastereoselectivity
across the THP ring due to the prefer-
ence for both the alkoxide and the 6-
substituent to sit equatorial in the alky-
lation transition structure.In the oxy-
Michael addition of these lactolates to
b-substituted nitroolefins, we propose
that the high diastereoselectivity b- to
the nitro group is a result of a combi-
nation of steric, stereoelectronic and
solvation factors.
Keywords:
computational
chemistry
·
diastereoselectivity
Michael addition
·
·
lactols
·
stereochemical models
Introduction
ported by Evans,[3] whereas Enders reported a range of
highly diastereoselective reagent-controlled oxy-Michael ad-
ditions of N-formylnorephedrine to 2-alkyl-1-nitroethenes
(however, the stereoselectivity is much reduced when 2-aryl-
1-nitroethenes are employed).[4] An alternative diastereose-
lective approach employing intramolecular delivery of a
hemiacetal nucleophile by using a glucose-derived chiral
auxillary was reported by Watanabe to give high yields and
diastereoselectivities.[5]
More recently, enantioselective catalytic oxy-Michael re-
actions have been reported that use several modes of cataly-
sis.Jacobsen demonstrated that the addition of aldoximes to
a,b-unsaturated imides proceeds with excellent enantiocon-
trol when catalysed by (salen)aluminium complex Lewis
acid catalysts,[6] whereas Maruoka reported a Brønsted acid-
catalysed addition of simple alcohols that proceeded with
low to moderate stereoselectivity.[7] Other organocatalytic
approaches have also been successful, including the secon-
dary amine-catalysed addition of benzaldoxime to a,b-unsa-
turated aldehydes developed by Jørgensen;[8] and the bifunc-
tional Brønsted acid-base catalytic intramolecular oxy-Mi-
chael reactions for boronic hemiesters developed by Falck.[9]
Organocatalysed intramolecular oxy-Michael reactions have
also been reported by Scheidt[10] and separately by Hinter-
mann[11] in the enantioselective syntheses of flavanones.
Many of these reactions have been found to be of use in the
synthesis of natural products.[2]
Despite being reported by Loydl[1] as early as 1878 in his
synthesis of malic acid, the full synthetic utility of the oxy-
Michael reaction was not realised until recent years.[2] These
developments have included many examples of stereoselec-
tive oxy-Michael additions that compliment the stereoselec-
tive aldol reactions in the synthesis of (protected) b-hydroxy
carbonyl (or carboxyl) compounds.It is worth mentioning
that many examples of oxy-Michael reactions require either
the reaction to be intramolecular or the use of oxygen nu-
cleophiles that are more complex than simple alcohols to
overcome the thermodynamic preference of the retro-Mi-
chael reaction.[2] A highly diastereoselective, intramolecular
oxy-Michael reaction under substrate-control has been re-
[a] Dr.F.A.Hernandez-Juan, J.W.Ward, Dr.D.J.Dixon
School of Chemistry, The University of Manchester
Oxford Road, Manchester, M139PL (UK)
Fax : (+44)161-275-4939
E-mail: darren.dixon@manchester.ac.uk
[b] Dr.R.D.Richardson
Cambridge University Chemical Laboratory, Lensfield Road
Cambridge, CB2 1EW (UK)
Current Address: School of Chemistry, Cardiff University
Park Place, Cardiff, CF103AT (UK)
Fax : (+44)2920874030
E-mail: richardsonr@cf.ac.uk
Recently, our group reported that the oxy-Michael addi-
tion of the “naked” anion of enantiopure d-lactol 1 to nitro-
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.200801368.
Chem. Eur. J. 2008, 14, 9607 – 9612
ꢀ 2008 Wiley-VCH Verlag GmbH & Co.KGaA, Weinheim
9607

olefins proceeds with excellent diastereoselectivity both
across the tetrahydropyranyl (THP) ring and at the newly
formed stereocentre b to the nitro group in adduct 2
(Scheme 1).[12] This reaction has been extended to a range of
Michael acceptors[13] and has found use in the syntheses of
biologically active b-amino alcohols.[14]
gate the cis selectivity across the THP ring in the reaction
between the anion derived from (R)-6-methyl-d-lactol 1
with chloromethane as a simple model system.As stated
earlier, the selectivity in such alkylation reactions is high.[15]
The use of a crown ether in the Michael addition reaction
means that the “naked” anion assumption is valid.[25]
Deprotonation of an anomeric mixture of lactol 1 with
potassium hexamethyldisilazide and addition of crown
ether[12] sets up an equilibrium of “naked” anions 3–6 and
their open-chain forms that can then be reacted with an
electrophile (Scheme 2).Assuming that the alkylation is ir-
Scheme 1.Diastereoselective oxy-Michael reactions of
KHMDS: potassium hexamethyldisilazide.
d-lactols.
The high C6–C2 stereocontrol has also been observed in
the reactions of the corresponding potassium salts with alkyl
and acyl halides and with acetic anhydride.[15] The reaction
of the anomeric hydroxyl of protected glucosides with alkyl
halides[16] or trichloroacetonitrile[17] in the presence of
sodium hydride in polar, aprotic solvents also occurs with
high kinetic preference for the formation of the b-anomer.
It is reasonable to suggest that the origin of the cis-selectivi-
ty in all these lactolate reactions is similar but, to date, no
computational study of these systems in solution has been
reported.Herein, we report our work leading to a model to
explain the stereochemical outcome of these reactions.
Computational Methods
All calculations were performed by using Jaguar.[18] Energy
minimizations and transition-state searches were performed
by using a continuum solvent model for THF at the B3LYP/
6-31+G(d) level[19] previously used by Salpin to study the
gas phase acidity of d-glucose.[20] Energies were refined at
Scheme 2.The anionic equilibrium. Energies are quoted as Gibbs free
energies in kcalmolÀ1 above anion 3.Numbers in square brackets are
imaginary frequencies.
the MP2/6-31++G
G
reversible and, therefore, under kinetic control, a full analy-
sis of the stereoselectivity requires determination of the
equilibrium ratios of anions 3–6,[26] the activation energy for
their interconversion and the activation energy for the reac-
tions of lactolates 3–6 with an electrophile.The calculated
relative energies of anions 3–6 and the lowest-energy open-
chain form 7 (found by a molecular mechanics conformation
search[27] followed by DFT re-optimisation and single-point
calculation on the ten lowest energy conformers) are shown
in Scheme 2.
Diequatorial anion 3 dominates the equilibrium, whereas
species 6 with the axial alkoxide is more prevalent than the
other conformer of the trans isomer 4 with the axial methyl
group.The concentrations of diaxial anion 5 and open-chain
forms are low.All located twist-boat conformations were
over 9 kcalmolÀ1 higher in energy than 3.The calculated cis/
trans ratio at 195 K is 180:1.
to have exactly one imaginary frequency that was followed
by Goodmanꢁs Quick Reaction Coordinate.[21] Frequencies
were determined at the B3LYP/6-31+G(d) level and scaled
by 0.98.[22] Energies are quoted as free energies at 195 K
(À788C) at the MP2/6-31++G
G(d)
ACHTREUNG
tions determined by vibrational analysis for the species in
THF solution.Solvation effects were modelled by using the
Jaguar Poisson-Boltzmann/SCRF continuum approach with
a variable cavity.[24] Tight SCF and geometry convergence
criteria were applied throughout.
Results and Discussion
Locating the transition states for the oxy-Michael reaction
of lactol 1 with any nitroolefin proved very computationally
demanding.Owing to this, it was decided to initially investi-
The solvation energies of the equatorial anions 3 and 4
are approximately 5 kcalmolÀ1 greater than those for the
9608
www.chemeurj.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co.KGaA, Weinheim
Chem. Eur. J. 2008, 14, 9607 – 9612

Oxy-Michael Reactions of d-Lactol Anions
FULL PAPER
axial anions 5 and 6 showing that the preference for the alk-
oxide to be equatorial is due to this larger solvation energy
that more than cancels out classical anomeric effects ob-
served in previous gas-phase calculations.[20] This greater sol-
vation energy arises from the larger dipole moment of the
equatorial anions (11.8 for equatorial alkoxide 3 compared
with 8.7 D for the equivalent axial alkoxide 6),[28] which
leads to it being stabilised by the polar solvent.This is con-
sistent with the effect reported by Wiberg for the conforma-
tional preference of trans-1,2-difluorocyclohexane in the gas
and solution phase.[29]
The interconversion pathways of anions 3–6 were studied
À
by driving the O1 C2 bond length (allowing minimisation
of all other variables) and re-optimisation at the saddle
point.Whilst pathways could be found connecting anions
3
and 6 with equatorial methyl groups to open forms via tran-
sition-states 8 and 9, respectively, no transition state con-
necting anions 4 and 5 with the open chain could be found.
However, it is reasonable to assume that rapid ring-flipping
pathways[31] allow equilibration of 3Q5 and 4Q6, hence, that
the activation barrier to anion equilibration is 14.1 kcal
molÀ1 (Scheme 2).
À
Anion 3 has a long O1 C2 bond (1.501 compared to
1.429 Š in the protonated form minimised under the same
conditions and 1.418 Š found in a crystal structure of the
When studying the reaction of anions 3–6 with chlorome-
thane, two pathways were found for the reaction of the
equatorial anions 3–4 with the electrophile approaching
such that the dihedral (for anion 3) y(C3-C2-O7-C(H3))=
À178.7 (10) or 58. 38 (11—disfavoured by a steric clash with
the axial proton on C3).Only one of these pathways could
be found for the reaction of axial alkoxides (for anion 6
y(C3-C2-O7-C(H3))=À155.98, 15).The other approach
would place the electrophile over the THP ring resulting in
substantial steric compression with the axial groups on C4
and C6 (Scheme 3).No pathways were found in which the
oxy-Michael adduct of 1 with b-methyl-b-nitrostyrene[2b]
and a short O7 C2 bond (1.321 Š), and the geometry at C2
is distorted considerably from tetrahedral (bond angles: f-
)
À
G
G
tent with donation of the alkoxide electron density to the
À
C2 O1 s*.The energy difference between anions 3 and 4
(4.3 kcalmolÀ1) is much greater than the A value of a
methyl group on a THP ring (2.3 kcalmolÀ1).[30] This was un-
À
À
expected because the elongated O1 C2 bond is expected to
electrophile approaches antiperiplanar to the C2 O1 bond,
which is consistent with this pathway being deactivated by
donation of alkoxide electron density in this region to the
decrease the 1,3-diaxial interactions.However, on closer in-
spection of anion 3, the geometrical distortion at C2 pushes
the axial proton closer to that on C6 than it is in lactol 1
(H–H distances: 2.286 and 2.384 Š, respectively) leading to
increased steric compression.This distortion is also observed
in the anion 4 with the axial methyl group and this is be-
lieved to be the cause of the increased A value calculated
here.No such distortion is seen in axial alkoxides 5 and 6.
À
C2 O1 s* (Scheme 3).
The activation energy for alkylation (24.2 kcalmolÀ1) is
much greater than that for anion equilibration (14.1 kcal
molÀ1).The consequence of this is a Curtin–Hammett
system in which the diastereoselectivity across the THP ring
in the alkylation is determined only by the relative energies
À1
Scheme 3.Alkylation of lactolate anions by chloromethane.Energies are quoted as Gibbs free energies in kcalmol
above anion 3 (+MeCl).Numbers
in parentheses are free energies relative to transition-state 10.Numbers in square brackets are imaginary frequencies.
Chem. Eur. J. 2008, 14, 9607 – 9612
ꢀ 2008 Wiley-VCH Verlag GmbH & Co.KGaA, Weinheim
www.chemeurj.org
9609

R.D.Richardson, D.J.Dixon et al.
of transition-states 10–15.The energies in Scheme 3 predict
a kinetic cis/trans ratio (i.e., of products cis-16/trans-16) at
195 K of 70:1.In previous experimental studies, the trans
isomer is hardly ever observed in these reactions by
1H NMR spectroscopic analysis,[12,15] consistent with the re-
action producing this high diastereoselectivity.The differ-
ence in solvation energies between transition-states 10 and
15 accounts for the majority of the calculated energy differ-
ence between them (3.1 of the 3.6 kcalmolÀ1).The energy
difference between transition-states 10 and 12 (1.6 kcal
molÀ1) is much closer to the A value of a methyl group on a
cyclohexane ring (1.7 kcalmolÀ1)[30] than it was in the anions
3 and 4.The geometry at C2 in transition-state 10 is closer
to tetrahedral than that in anion 3, making the distance be-
tween the axial protons on C2 and C6 larger (2.328 Š) and
reducing the steric 1,3-diaxial compression discussed earlier.
Our calculations suggest that the cis diastereoselectivity in
the alkylation of the anion of lactol 1 is a result of the pref-
erence for both the C6 substituent and the alkoxide to
occupy equatorial positions in the transition state for steric
and solvation reasons, respectively.With these structures lo-
cated, attention was turned to the reaction with nitroolefins.
The high diastereoselectivity at the b-centre in adduct 2 is
consistent across a wide range of alkyl and aryl-substituted
nitroethenes,[12] so the reaction between 1-nitropropene and
the anion of lactol 1 was studied to reduce computational
demand.Despite this, location of the oxy-Michael transition
states proved difficult.A full study would require at least 12
such transition states to be located (corresponding to the six
transition-states 10–15 with approaches leading to both the
R and S configurations at Cb).The reaction is, however,
highly cis selective so the major transition state must corre-
spond to one of the cis structures 10, 11 and 14 shown in
Scheme 3.Additionally, the diequatorial structures 3, 10 and
11 are always much lower in energy than the corresponding
diaxial species 5 and 14, so it is reasonable to assume that
the reaction proceeds through the addition of diequatorial
anion 3 to the electrophile.Taking this into account, the ste-
reoselectivity in the oxy-Michael reaction can be investigat-
ed by studying the reaction of nitropropene with anion 3
alone.
Scheme 4.Oxy-Michael addition transition states.Energies are quoted as
Gibbs free energies in kcalmolÀ1 above anion 3+nitropropene.Numbers
in parentheses are free energies relative to transition-state 17.Numbers
in square brackets are imaginary frequencies.
Figure 1.Second-order saddle point located with electrophile APP to
À1
C2H.Energies are quoted as Gibbs free energies in kcalmol
above
anion 3+nitropropene.Numbers in parentheses are free energies relative
to transition-state 17.Numbers in square brackets are imaginary frequen-
cies.
stabilised by a steric interaction with the axial proton on C3.
This suggests that the reaction proceeds in such a way that
À
the electrophile approaches antiperiplanar to the C2 C3
bond.The dihedral driving studies also show that there is a
strong preference for the nitropropene to approach with the
À
C=C double bond antiperiplanar to the exocyclic C2 O7
bond—a result of electrostatic repulsion between the residu-
al negative charge on the lactolate oxygen atoms and the de-
veloping nitronate anion.
The calculated energy difference between transition-states
17 and 18 suggests a kinetic diastereoselectivity at Cb of
10:1 in favour of (R)-19—much lower than the experimental
results for all nitroalkenes, but the computation does predict
the correct facial selectivity.Attempts to extend this compu-
tational study to any of the electrophiles that have been pre-
viously used experimentally[12] led to geometry convergence
failure.The distance between the methyl group on C b and
the axial proton on C2 in major transition-state 17 (3.025 Š)
is longer than that in the minor transition-state 18 (2.843 Š),
which suggests that the stereoselectivity may increase as this
methyl group is changed to a larger function.
During optimisation of the reaction, it was noticed that
the use of sodium bis(trimethylsilyl)amide and [15]crown-5
in place of the optimal potassium bis(trimethylsilyl)amide
and [18]crown-6 did lead to a reduced stereoselectivity at
Cb,[1] which showed that the counter-ion–crown ether com-
plex can still influence the course of the reaction.For the
factors discussed here, the conclusions drawn from these cal-
Transition-state searches on the reaction between nitro-
propene and lactolate 3 identified only transition-state 17
leading to (R)-19 and transition-state 18 leading to (S)-19
(Scheme 4).
Both these transition states involve the approach of the
electrophile approximately antiperiplanar to the endocyclic
À
C2 C3 bond.Extensive dihedral driving about the exocyclic
À
À
C2 O7 bond and the newly forming O7 Cb bond and re-
searching (see the Supporting Information for details) did
locate species 20, but vibrational analysis identified this as a
second-order saddle point (Figure 1).Following the second
imaginary frequency in either direction by QRC led back to
the previously located transition-state 17 without passing
through any local minima.No corresponding saddle point of
any order could be located that would lead to (S)-19.Spe-
cies 20 is 4.4 kcalmolÀ1 higher in energy than 17 and is de-
9610
www.chemeurj.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co.KGaA, Weinheim
Chem. Eur. J. 2008, 14, 9607 – 9612

Oxy-Michael Reactions of d-Lactol Anions
FULL PAPER
Table 1.Lactol modification and stereoselectivity.
culations on the oxy-Michael reaction are more tentative
than those drawn from the alkylation reaction above.
The solvation energy of major transition-state 17 is
5.4 kcalmolÀ1 greater than that for the minor transition-state
18 and this more than accounts for the energy difference be-
tween them.This suggests that the stereocontrol in this reac-
tion is controlled once again by preferential solvation of the
transition state leading to the R diastereomer of adduct
19—a result of the greater dipole moment on 17 (14.3 com-
pared to 13.1 D). The methyl group on C6 is distant to the
Cb substituent (4.765 Š in major transition-state 17 and
5.048 Š in minor transition-state 18).This suggests that the
role of this substituent is solely to control the conformation
of the THP ring and it does not interact with the electro-
phile.
Lactol
R1
R2
Product
Yield [%]
d.r. at Cb[a]
1
Me
H
2
99
60
55
62
>98:2
>98:2
>98:2
6:1
23a
23b
23c
Me
n-Hex
tBu
Me
H
H
24a
24b
24c
[a] Determined by 1H NMR spectroscopic analysis on the crude reaction
product.The cis/trans ratio across the THP ring was >98:2 in all cases.
In the original work,[1] nitropropene was not tested as an
electrophile.In light of the differences in the predicted dia-
stereomeric ratio from the current calculations (10:1) and
that obtained with the larger electrophiles used in the previ-
ous work (for example, b-nitrostyrene giving 99:1), we revis-
ited the reaction with nitropropene 21 with the 6R enantio-
mer of lactol 1 (in the computational studies, the 6S enantio-
mer was used in keeping with previous experimental
work[12]).The reaction was much less clean than that with
other electrophiles (probably due to reaction pathways in-
volving deprotonation of the g-carbon atom), but we ob-
served a 32:1 ratio of diastereomers in favour of (bS)-22
(Scheme 5).This corresponds to a difference in transition-
state energies of 1.3 kcalmolÀ1 compared to the calculated
0.9 kcalmolÀ1.The discrepancy between these activation en-
ergies is within the expected error of the techniques used
here.
calculated transition states in which the electrophile ap-
À
proached antiperiplanar to the C2 C3 bond, distant from
these groups.
The stereochemical outcome is an indirect result of the
preference for the nitro group to be as far from the endocy-
clic oxygen atom as possible in the transition state.When
the reaction is performed in the absence of crown ether, the
Lewis acidic counterion would be expected to give rise to a
cyclic transition state.[25,32] This would lead to a change in
stereoselectivity, as observed experimentally, but a full com-
putational study of this would be far too demanding on cur-
rent computational resources.
Conclusion
The stereoselectivity in the oxy-Michael reaction of d-lactol
anions can be explained as follows.The reaction proceeds
through a cis diequatorial transition state due to the prefer-
ence for the methyl group on C6 and the reacting anomeric
alkoxide to be equatorial for steric and solvation reasons.
The nitroolefin approaches the lactolate antiperiplanar to
À
the C2 C3 bond to minimise steric compression and with
À
the C=C double bond antiperiplanar to the exocyclic C2
O7 bond to minimise electrostatic repulsions.Finally, with
the (S)-lactol 1 the reaction proceeds with R selectivity
owing to preferential solvation of transition-state 17.The cis
selectivity across that THP ring is predicted accurately by
the computational study, and the predicted selectivity in the
oxy-Michael reaction also agrees with experiment when the
reaction is conducted with nitropropene.Further predictions
of this model about the effects of the lactol structure on the
stereoselectivity of this reaction have all been realised ex-
perimentally.
Previous explanations for the increased reactivity of equa-
torial alkoxides invoke lone-pair interactions present in the
b-anomer of the anions of glucosides making equatorial alk-
oxides more reactive.[17] Whilst there is evidence for the
equatorial alkoxides being more reactive (alkylation transi-
tion-state 15 lies 25.7 kcalmolÀ1 above axial anion 6, where-
as the corresponding transition-state 10 lies 24.1 kcalmolÀ1
above equatorial anion 3), the reaction is a Curtin–Hammett
Scheme 5.Stereoselective oxy-Michael reaction with nitropropene.dr..
diastereomeric ratio.
=
Further experimental studies into this stereochemical
model were also undertaken by modifying lactol 1 (Table 1).
The use of lactols 1 and 23b shows that increasing the size
of the C6 substituent from methyl to hexyl leads to no meas-
urable change in the stereoselectivity at Cb, which is consis-
tent with this group serving only as a conformational lock.
However, when a tert-butyl group was placed here (as in
lactol 23c) the selectivity was diminished, which is consis-
tent with the fact that this group lies closer to the substitu-
ent on Cb on the major transition-state 17 than it does in
the minor one, 18.The installation of geminal dimethyl
groups on C3 in lactol 23a has no measurable effect on the
stereoselectivity of the reaction, which is consistent with the
Chem. Eur. J. 2008, 14, 9607 – 9612
ꢀ 2008 Wiley-VCH Verlag GmbH & Co.KGaA, Weinheim
www.chemeurj.org
9611

R.D.Richardson, D.J.Dixon et al.
[14] D.J.Buchanan, D.J.Dixon, M.S.Scott, D.I.LainØ,
Tetrahedron:
system in which the relative energies of the alkylation tran-
sition state alone decided the stereochemical outcome of the
reaction, so it is the differences observed there that must be
the origin of stereocontrol in this reaction.
Asymm. 2004, 15, 195–197.
[15] D.J. Dixon, S.V. Ley, E.W. Tate,
J. Chem. Soc. Perkin Trans. 1
1999, 2665–2667; D.J.Dixon, S.V.Ley, E.W.Tate,
Perkin Trans. 1 1998, 3125–3126.
J. Chem. Soc.
[16] J.J. Oltvoort, M. Kloosterman, C.A.A. van Boeckel, J.H. van
Boom, Carbohydr. Res. 1984, 130, 147–163.
[17] R.R.Schmidt, J.Michel, Tetrahedron Lett. 1984, 25, 821–823.
[18] Jaguar version 4.2, Schrodinger, Inc., Portland, Oregon, 2000 with
Maestro version 5.0.019 interface.
[19] For Beckeꢁs 3-parameter non-local hybrid exchange potential, see:
A.D.Becke, J. Chem. Phys. 1993, 98, 5648–5652; for the Lee, Yang
and Parr non-local correlation function, see: C.Lee, W.Yang, R.G.
Parr, Phys. Rev. B 1988, 37, 785–789.
Acknowledgements
We thank Dr.J.Goodman, Dr.C.Bolton and the Chemical Information
Laboratory, Unilever Centre, Cambridge for computational facilities; the
EPSRC for a studentship (to RDR) and the EPSRC National Mass Spec-
trometry Service Centre, Swansea for mass spectrometric data.
[20] J.-Y. Salpin, J. Tortajada, J. Mass. Spectrom. 2004, 39, 930–941.
[21] M.A.Silva, J.M.Goodman, Tetrahedron Lett. 2003, 44, 8233–8236.
[22] A.P.Scott, L.Radom, J. Phys. Chem. 1996, 100, 16502–16513.
[23] The same single-point calculations performed at the MP2/6-31++
G**//B3LYP/6-31+G* level gave the same relative energies in di-
[1] F.Loydl, Justus Liebigs Ann. Chem. 1878, 192, 80.
[2] For reviews of recent developments and applications of the oxy-Mi-
chael reaction, see: a) C.F.Nising, S.Bräse,
Chem. Soc. Rev. 2008,
37, 1218–1228; b) A.Berkessel, Houben-Weyl Methods of Organic
Synthesis, Vol. E21e (Ed:s. G.Helmchen, RW. .Hoffmann, J.
Mulzer, E.Schaumann), Thieme, Stuttgart, 1996, pp.4818–4856.
[3] D.A. Evans, J.A. Gauchet-Prunet, J. Org. Chem. 1993, 58, 2446–
2453.
rectly comparable species to within 0.2 kcalmolÀ1
.
[24] This has been used to calculate solution phase pKa values to
Æ0.4 pKa units hence was deemed a suitably accurate model for the
solvation of “naked” anions: J.J.Klicic, R.A.Friesner, S-.Y.Liu,
W.C.Guida, J. Phys. Chem. A 2002, 106, 1327–1335; the following
parameters were used for THF: e=7.52, FW =72.11, 1=0.886 gmLÀ1
taken from CRC Handbook of Chemistry and Physics (Ed.: D. R.
Lide), 84th ed., CRC Press, Boca Raton, Florida, 2003.
[4] D.Enders, A.Haertwig, G.Raabe, J.Runsink,
Eur. J. Org. Chem.
1998, 1771–1792.
[5] H.Watanabe, K.Machida, D.Itoh, H.Nagatsuka, T.Kitahara,
Chir-
ality 2001, 13, 379–385.
[25] M.A. Beswick, D.S. Wright in
Comprehensive Organometallic
[6] CD. .Vanderwal, EN. .Jacobsen,
14724–14725.
J. Am. Chem. Soc. 2004, 126,
Chemistry (Eds.E.W. Abel, F.G.A. Stone, G. Wilkinson), Perga-
mon Press, Oxford, 1995, Chapter 1.
[7] T.Kano, Y.Tanaka, K.Maruoka, Tetrahedron 2007, 63, 8658–8664.
[26] Attempts to study the anionic equilibrium by NMR spectroscopy
gave no meaningful results.
[8] S. Bertelsen, P. DinØr, R.L. Johansen, K.A. Jørgensen,
J. Am.
Chem. Soc. 2007, 129, 1536–1537.
[9] D.R.Li, A.Murugan, J.R.Falck,
[27] Monte Carlo conformational search performed MM2 forcefield as
implemented in Macromodel version 8.0, W. C. Still, Columbia Uni-
versity, New York.
[28] Dipole moments were calculated by using ESP fitting to reaction
centres at the B3LYP/6-31+G* level.
J. Am. Chem. Soc. 2008, 130, 46–
J. Am. Chem. Soc. 2007, 129,
48.
[10] M.M.Biddle, M.Lin, K.A.Scheidt,
3830–3831.
[11] C.Dittmer, G.Raabe, L.Hintermann,
5886–5898.
[12] N.J. Adderley, D.J. Buchanan, D.J. Dixon, D.I. LainØ,
Eur. J. Org. Chem. 2007,
Angew.
[29] K.B.Wiberg, W.Hinz, R.M.Jarret, K.B.Aubrecht,
J. Org. Chem.
2005, 70, 8381–8384.
[30] For A values see: F.A. Carey, R.J. Sundberg, Advanced Organic
Chemistry, 3rd ed., Plenum Press, NY, 1990.
Chem. 2003, 115, 4373–4376; Angew. Chem. Int. Ed. 2003, 42, 4241–
4244.
[31] The activation barrier for chair-to-chair flipping in cyclohexane is
+10.2 kcalmolÀ1 at 206 K: F.A.Bovey, F.P.Hood, III, E.W.Ander-
son, R.L.Kornegay, J. Chem. Phys. 1964, 41, 2041–2044.
[32] A.Mengel, O.Reiser, Chem. Rev. 1999, 99, 1191–1223.
[13] a) D.J. Buchanan, D.J. Dixon, F.A. Hernandez-Juan,
Org. Lett.
2004, 6, 1357–1360; b) D.J. Buchanan, D.J. Dixon, F.A. Hernan-
dez-Juan, Org. Biomol. Chem. 2004, 2, 2932–2934; c) F.A.Hernan-
dez-Juan, X. Xiong S.E. Brewer, D.J. Buchanan, D.J. Dixon,
thesis 2005, 3283–3286; d) Xiong, C.Ovens, AW. .Pilling, JW. .
Ward, D.J.Dixon, Org. Lett. 2008, 10, 565–567.
Syn-
Received: July 6, 2008
Published online: September 26, 2008
9612
www.chemeurj.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co.KGaA, Weinheim
Chem. Eur. J. 2008, 14, 9607 – 9612