Journal of the American Chemical Society
Communication
(Scheme 5a).61 We also isolated cinnamyl tert-butyl ether 48,
which could arise from the decomposition of the IrIII(π-allyl)
intermediate derived from the less reactive enantiomer of the
carbonate.53,54 To further illustrate this mechanistic hypoth-
esis, the reaction was probed with enantioenriched carbonate
(S)-2a using either (S)-L1 or (R)-L1 (Scheme 5b). Ligand
(S)-L1 provided the desired product in 96% yield with
excellent stereoselectivity and no observable byproduct.
Alternatively, ligand (R)-L1 provided the three-component
coupling product in only 32% yield with greatly diminished
enantioselectivity. The linear cinnamyl ether 48 was obtained
in comparable yields. Lastly, enantioenriched carbonate (S)-2a
was coupled with boronate 1b in the presence of achiral ligand
L4 to obtain the product in quantitative yield with a slight drop
in the er compared to (S)-2a (Scheme 5b). From these results,
a proposed mechanism of the pathway is shown in Scheme 5c.
First, the racemic allylic carbonate undergoes a kinetic
resolution by the Ir(phosphoramidite) complex to provide
diastereomeric IrIII(π-allyl) intermediates 49 and 50, wherein
kS > kR. Loss of carbon dioxide renders the oxidative addition
step irreversible. The IrIII(π-allyl) intermediates can then either
react with alkenyl boronate to form the desired product or
react with tert-butoxide to form ether 48.
Quantitative assessment of the various reaction pathways
accessible to the IrIII(π-allyl) intermediates will require
additional investigation, but qualitative conclusions can be
drawn with existing data. The rate of product formation from
the fast-reacting enantiomer (kmat) is much faster than the rate
of ether formation (kether). This conclusion arises from the
observation that little ether was observed when optically active
substrate (S)-2a reacted in the presence of the matched ligand.
Alternatively, the rate of product formation from the slower-
reacting enantiomer (kmis) is similar to the rate of etherification
(k′ether) because the mismatched reaction formed ether 48 and
the coupled product in similar yields. Moreover, π−σ−π
isomerization of the IrIII(π-allyl) complex is slower than the
matched reaction (kmat > kisom) based on the modest drop in
optical activity observed using the enantioenriched substrate
and the achiral ligand L4 (Scheme 5b). However, the
mismatched reaction of (S)-2a in the presence of (R)-L1
proceeded with substantial epimerization to give the same
major enantiomer of product as from (R)-2a, indicating that
kisom ∼ kmis. Taken together, the results indicate the chiral
ligand (S)-L1 affects the enantioselectivity of the reaction
through effects on both the oxidative addition step and the
addition of the alkenyl boronate nucleophile.
a
Scheme 6. Setting Three Contiguous Stereocenters
a
The reaction was conducted on a 0.3 mmol scale using 2-bromo-2H-
chromene (0.3 mmol), s-BuLi (0.3 mmol), boronic ester (0.3 mmol),
[Ir(COD)Cl]2 (2.5 mol %), (S)-L1 (10 mol %), ( )-2a (0.63 mmol),
and LiCl (0.3 mmol). Yields of the isolated and purified material after
oxidation by treatment of boronic ester with NaOH and H2O2. The er
of the major diastereomer as determined by HPLC analysis. The dr
1
determined by H NMR.
alkyl-derived ate complexes. Primary (32, 33), secondary (34),
and tertiary (35) boronic esters all coupled cleanly and
selectively. Coupling products derived from tertiary boronic
esters often provided poor conversion to the alcohol (35)
during oxidation.
Diastereomeric products could be accessed by switching the
migrating group and the olefin substituent (Scheme 3b). For
example, addition of isopropenyllithium to phenyl boronic
ester provided 36 from R2 = Ph as the migrating organic
fragment. Alternatively, addition of methyllithium to 2-boryl
styrene provided the diastereomeric product 37 as a
consequence of R1 = Me now participating in the migration
step. We have demonstrated this flexibility with aryl−alkyl (36,
37), alkyl−alkyl (38, 39), and aryl−aryl (40, 3) derived
tertiary alcohols with similar yields and enantioselectivities of
the major diastereomer in each case.
To explore the synthetic utility of asymmetric allylation, the
coupling reaction was executed on a gram-scale to provide
tertiary boronic ester 41 in 88% yield and 8:1 dr and 95:5 er
(Scheme 4). Next, synthetic transformations were performed
on boronic ester 41 to demonstrate the compatibility of these
compounds for derivatizations. Oxidation of 41 followed by
iodoetherification of the tertiary alcohol provided crystalline
iodide 42, which established absolute and relative stereo-
chemistry. Other useful synthetic intermediates were prepared,
such as the trifluoroborate salt (43)58 and diene (44).59 The
boronic ester was arylated to provide both the furan (45)21
and pyridine (46)25 using transition-metal-free cross-coupling
chemistry. Finally, hydroboration/oxidation provided diol 47.
In all cases, good to excellent yields were obtained with
retention of stereochemistry.
Mechanistic Considerations. The Carreira group has
investigated the mechanisms of allylic substitution reactions
catalyzed by Ir[(S)-L1] complexes.60 However, the addition of
alkenyl boronates to Ir(π-allyl) intermediates and the
associated metalate shift gave rise to some interesting
mechanistic questions. To determine if the racemic carbonate
was undergoing a kinetic resolution, we conducted the
allylation using ate complex 1b and isolated both the product
36 and the unreacted starting material (Scheme 5). Recovered
carbonate (R)-2a was isolated in 97:3 enantioselectivity,
indicating that an efficient kinetic resolution had occurred
To gain insight into the diastereoselectivity of the reaction,
we investigated trisubstituted boronate 1d to form three
contiguous stereocenters (Scheme 6).35,62,63 Under optimized
conditions, product 51 (R = n-Bu) was isolated in 65% yield as
a 1:1 mixture of diastereomers, both of which display 99:1 er.
Likewise, 52 (R = Me) was obtained in similar yield and
stereoselectivity. In both diastereomers of both products,
irradiation of the benzopyran methine showed an NOE signal
corresponding to the OH hydrogen, while we observed no
correlation between the butyl or methyl group with the
adjacent sp3-hybridized methine. The NMR data is bolstered
by an X-ray crystal of 52a. This evidence suggests the
surprising conclusion that both diastereomers result from a cis
addition of the migrating R group and the allyl fragment across
the olefin, with the diastereomers arising from opposite facial
selectivity. Understanding the mode of addition for disub-
stituted alkenyl boronate complexes is an ongoing effort.
4925
J. Am. Chem. Soc. 2021, 143, 4921−4927