ACS Catalysis
Research Article
borane 2b (20 mol %) and HBpin 6 (1.2 equiv) at 0 °C enabled
the asymmetric reduction of 4-phenyl-3-butyn-2-one 1a in 76%
yield and 89% e.e. This matched the yield and enantioselectivity
obtained using stoichiometric myrtanyl borane 2b (90% yield,
89% e.e.), giving 99% e.f. and establishing B−O transborylation
as a mechanism of turnover for asymmetric main-group catalysis.
The catalytic asymmetric reduction was further applied to other
substrate classes; however, this proved unsuccessful in the cases
of acetophenone and 4-phenyl-3-buten-2-one (no reaction) and
an α-keto ester and an α-keto thioester (poor e.e.) (see Table S1
in the Supporting Information).
enantiofidelity in the reduction of ethyl- (1s) and trifluor-
omethyl-substituted (1v) ketones, with the enantiofidelity
increasing from 63% to 82% e.f. and from 19% to 83% e.f.,
respectively. The enantiofidelity could also be improved by
reducing the H-B-9-BBN 2d loading (10 mol %) while
maintaining the β-pinene β-4 loading (20 mol %); the
enantiofidelity of ethyl ketone 1s increased from 63% to 85%
e.f. Reducing the reaction temperature to −20 °C improved the
enantiofidelity (to 85% e.f.), albeit with reduced yield (22%).
Applying B−O transborylation to substrates derived from
biologically active compounds proved successful. Asymmetric
reduction of the galactopyranose-derived substrate 1x gave high
diastereofidelity (95% d.f.). Ketone 1y, derived from propofol
(Diprivan), was reduced with excellent enantiofidelity (99%
e.f.). Gram-scale reduction of ketone 1a under the standard
conditions gave excellent enantiofidelity and yield (80% yield,
98% e.f.).
Two mechanisms of boron−boron exchange have been
proposed: ligand redistribution and transborylation (Scheme
3a).7,8,14 For ligand redistribution, the B−O bond of the borinic
ester 3a is maintained and the supporting ligands are exchanged
with HBpin. Transborylation breaks the B−O bond on the
borinic ester 3a by σ-bond metathesis, with the boron atom of
alkoxyboronic ester 7a originating from HBpin. The reaction of
H10Bpin with borinic ester 3a gave only the 10B-labeled
alkoxyboronic ester 10B-7a, as determined by 10B and 11B
NMR spectroscopy (Scheme 3a). Therefore, exchange
proceeded by B−O transborylation, not ligand redistribution.
The thermodynamic properties of the B−O transborylation
were determined using an Eyring plot constructed over the
temperature range 301−315 K (Scheme 3b; see section S9 in the
Supporting Information).15 This supported a highly ordered
transition-state structure for B−O transborylation with a large
negative entropy value (ΔS⧧ = −21.5 eu)16 and a Gibbs free
energy (ΔG⧧298 = 22.7 kcal mol−1) similar to those of B−C(sp2)
(ΔG⧧ = 20.3 kcal mol−1)7b and B−C(sp3) (ΔG⧧ = 28 kcal
mol−1)7a transborylation reactions.
When all mechanistic investigations were taken into account,
a catalytic cycle for the B−O transborylation-driven asymmetric
ketone reduction was proposed (Scheme 3c). Enantioselective
hydroboration of the ketone 1 by the borane catalyst 2b through
a Meerwein−Ponndorf−Verley-type transition state gives the
enantioenriched borinic ester 3 and releases β-pinene β-4
(enantioselective hydroboration).11,17 B−O/B−H transborylation
of borinic ester 3 with HBpin 6 gives the alkoxyboronic ester
product 7 and releases H-B-9-BBN 2d (transborylation). The
borane catalyst 2b is regenerated by highly chemo-, regio-, and
diastereoselective hydroboration of β-pinene β-4 by H-B-9-BBN
2d (alkene hydroboration).
In summary, B−O transborylation has been established and
applied as a turnover mechanism for asymmetric main-group
catalysis. A catalytic Midland reduction has been enabled, using
B−O/B−H transborylation and myrtanyl borane 2b as the
asymmetric catalyst, across a range of functionalized substrates
with excellent enantiofidelity. B−O transborylation was found to
proceed by a σ-bond metathesis mechanism. Modification of the
catalytic protocol to reduce racemic background reductions by
achiral boron reagents (H-B-9-BBN 2d and HBpin 6) ensured
high enantiofidelity for challenging substrates. This application
of B−O/B−H transborylation demonstrates the potential of
transborylation to be used as a general platform for main-group
catalysis.
The substrate scope of the catalytic asymmetric hydro-
boration was explored using myrtanyl borane 2b as the catalyst,
generated in situ by reaction of H-B-9-BBN 2d (20 mol %) and
β-pinene β-4 (20 mol %) (Table 1). 4-Phenyl-3-butyn-2-one 1a
underwent hydroboration with excellent yield (90%) and
enantiofidelity (99% e.f.). Substitution on the aromatic ring
was tolerated, with excellent enantiofidelity observed for 4-tert-
butyl (1b, 89% e.f.), 4-methyl (1c, 94% e.f.), 3-methyl (1d, >99%
e.f.), and 2-methyl (1e, 97% e.f.) groups. Use of the 4-fluoro
derivative 1f gave good enantiofidelity (88% e.f.) whereas
decreased enantiofidelity was observed for the 3-chloro
analogue 1g (66% e.f.). Lewis basic ether substituents 1n
(84% e.f.) and 1m (92% e.f.) and the thioether 1o (90% e.f.) gave
high enantiofidelity, although the 4-methoxy-substituted 1r gave
lower enantiofidelity (50% e.f.). Reduced enantiofidelity was
observed with the dimethylamino-bearing ketone 1q (60% e.f.).
Excellent chemoselectivity was observed, with groups expected
to react with boranes being tolerated. Nitrile (1w, 91% e.f.), ester
(1j, 74% e.f.), and amide substituents (1i, >99% e.f.) all gave
excellent enantiofidelity. Propargylic ketones bearing electron-
withdrawing substituents, such as 1f (73% e.e.), 1g (46% e.e.),
and 1j (67% e.e.), were reduced in moderate to good e.e.,
presumably due to a greater rate of background, unselective
reduction by HBpin. Propargylic ketones bearing electron-
donating substituents 1a−e consistently gave improved
enantioselectivities (89−77% e.e.). However, in contrast to
ketones bearng electron-donating groups about the arene,
substrates bearing a mesomeric donor in the para position, 1q
(52% e.e.) and 1r (44% e.e.), gave moderate to poor
enantioselectivty. The greater Lewis basicity of these substrates
may increase the rate of unselective reduction, by greater
coordination to the achiral boranes. Although a higher rate of
reaction was achieved at 18 °C, the enantioselectivity was
decreased (5a (92% yield, 35% e.e.), 5m (88% yield, 49% e.e.), 5s
(92% yield, 34% e.e.), and 5v (83% yield, 26% e.e.), presumably
as a result of the low temperature required for enantioselectivity
in the Midland reduction.
Sterically encumbered ketones 1s (63% e.f.) and 1t (39% e.f.)
gave poor to moderate enantiofidelity. Presumably, slow
hydroboration by the enantioenriched borane allowed signifi-
cant background reduction by the less sterically demanding,
achiral boranes H-B-9-BBN 2d and HBpin 6. The trideuter-
iomethyl-substituted ketone 1h was tolerated, but electron-
withdrawing groups such as monofluoromethyl (1u, 44% e.f.)
and trifluoromethyl (1v, 19% e.f.) gave reduced enantiofidelity.
The trifluoromethyl ketone 1v was reduced to the racemic
alcohol ( )-5v by HBpin 6 in 86% yield under the reaction
conditions, indicating that unselective hydroboration by HBpin
6 outcompetes the enantioselective reaction.
Controlling the concentration of achiral boranes (H-B-9-BBN
2d and HBpin 6) could suppress the rate of unselective
hydroboration. The slow addition of HBpin improved the
2038
ACS Catal. 2021, 11, 2034−2040