Journal of the American Chemical Society
Article
often much more stable and reluctant to dissociate into
undergo ring-opening to generate the corresponding homo-
enolate C. Association of the homoenolate C with the
aminoalkoxide A and the enone 2 would generate a bimetallic
monomers, which is responsible for NLE in Et Zn addition to
2
aldehyde. A question relevant to the present reaction system is
what happens to such ethylzinc aminoalkoxide dimers upon
exposure to cyclopropanol. In attempts to address this
question, we measured 1H NMR spectra of zinc species
16
species D, followed by delivery of the homoenolate to give an
17
enolate species E. Proton transfer from the amino alcohol to
the enolate would furnish the conjugate adduct 3 while
formed by mixing L6 and Et Zn in a 1:1 ratio followed by
regenerating the aminoalkoxide dimer [A] . The heterochiral
2
2
dimer [A·ent-A], preferentially formed when using racemic
ligand, would be reluctant to dissociate into the monomer [A·
1
1
]. When there is no enone or ECA is sluggish, the
Except for the case using MeOH, the H NMR spectra showed
homoenolate C could be protonated, possibly via internal
proton transfer from the amino alcohol, to give the
corresponding ketone 1′, which could further participate in
enolate Michael addition under zinc catalysis. On the basis of
the solvent effect (Figure 1d), the ability of DMPU as a Lewis
base to Zn, rather than its bulk polarity as a solvent, appears to
be critical for the preferential CA of the homoenolate over its
protonation. Thus, we speculate that coordination of DMPU
to zinc would assist in dissociation of the dimeric species and/
or increasing the nucleophilicity of the homoenolate.
clearly resolved quartet (∼0.6 ppm) and triplet (∼1.6 ppm)
signals that could be assigned to an ethyl group on Zn. This
observation indicates that ethylzinc aminoalkoxide [(L6-
H)ZnEt] resists protonation of the remaining ethyl group by
a bulky alcohol such as tBuOH and iPrOH, which should also
be the case for the tertiary cyclopropanol substrates used in the
present reaction. Similar observations have been made for the
reluctance of [LZnEt] species toward protonation of the Zn−
15
Et bond by alcohol. Note also that, a sample prepared by
mixing Et Zn and tBuOH first, followed by L6, gave rise to a
2
1
Selected transformations of the enantioenriched cyclo-
pentene product 4aa are shown in Scheme 7. Conversion of
H NMR spectrum close to that of the above-mentioned
experiment. This suggests that the initially formed (tBuO)-
ZnEt underwent deprotonation of L6 with tBuO rather than
Et, thus generating [(L6-H)ZnEt] and tBuOH.
Given the reluctance of bulky alcohols to protonate the Zn−
Et bond of [(L6-H)ZnEt], we speculate that the formation of a
mixed zinc bisalkoxide species such as [(L6-H)(1-H)Zn] in
the present reaction system is less likely and that a zinc
cyclopropoxide species would form as a transient rather than
resting species via reversible protonation of the aminoalkoxide
ligand. With this conjecture, and on the basis of the above
NLE and other experiments as well as the literature knowledge
3a,13b
on chiral amino alcohol-catalyzed organozinc addition,
we
suggest a tentative catalytic cycle as illustrated in Scheme 6.
Coordination of the cyclopropanol 1 would break down the
homochiral ethylzinc aminoalkoxide dimer [A]2 into a
monomeric species [A·1]. Deprotonation of the cyclopropyl
OH with the internal aminoalkoxide base would reversibly
generate a cyclopropoxide species B, which would further
the benzoyl group to an anilide moiety via Beckmann
rearrangement was achieved through oxime formation and
subsequent treatment with Tf O, affording the product 5 in
2
good yield. Epoxidation with mCPBA provided the tetrasub-
stituted epoxide 6 in a diastereoselective manner, albeit in a
modest yield. Generation of a dienolate species from 4aa and
LiHMDS was followed by trapping with benzyl bromide,
which took place regioselectively at the α position to generate
a quaternary stereocenter in the product 7 in a diastereose-
lective manner.
Scheme 6. Possible Catalytic Cycle
CONCLUSION
■
In summary, we have developed an enantioselective ring-
opening conjugate addition of cyclopropanols to α,β-
unsaturated ketones via catalytic generation of zinc homo-
enolate. A chiral β-amino alkoxide ligand on zinc accelerates
the ring-opening of a cyclopropanol and aids in enantiose-
lective addition of the resulting homoenolate to chalcones and
related enones without assistance of a transition-metal catalyst.
The ensuing intramolecular aldol condensation of the 1,6-
diketone products allows for facile preparation of multi-
substituted cyclopentene derivatives with good to high
enantioselectivity. The reaction represents the first example
of enantioselective transformation of zinc homoenolate and
also a rare example of asymmetric transformations of metal
homoenolates in general. Further development and mecha-
4
779
J. Am. Chem. Soc. 2021, 143, 4775−4781