that provide access to a variety of boronic
ester products ( 15 – 19 ). A distinct approach that
does not proceed via a boronate rearrange-
ment yet leverages a similar retrosynthetic
disconnection to the sequence proposed here
has been developed by Fu and co-workers, in-
volving the enantioconvergent nickel-catalyzed
cross-coupling of racemica-halo boronic ester
derivatives with alkylzinc nucleophiles ( 20 , 21 ).
In Fu’s method, both substituents geminal to
the boronic ester in the product are defined
in the enantiodetermining reaction; as a re-
sult, the products accessible by this strategy
are dictated by the intersecting scope of both
organozinc reagents anda-halo boronic esters
that can be effectively engaged by the catalyst.
In contrast, the sequence proposed here would
install only a single substituent in the enan-
tioselective catalytic reaction, followed by the
enantiospecific introduction of the second
substituent with a scope not dictated by the
chiral catalyst. The closest efforts towardan enantioselective, catalytic variant of the
Matteson homologation specifically are those
reported by Jadhav and Man in the chiral
Lewis acid–mediated rearrangement of a
dichloromethyl-substitutedn-butyl boronate,
wherein up to 88% enantiomeric excess (ee)
was obtained using superstoichiometric quan-
tities of a chiral bis(oxazoline) ligand ( 22 , 23 ).
We hypothesized initially that the rearrange-
ment of prochiral dichloromethyl boronates
might be enabled by anion-abstraction catal-
ysis. In particular, weakly Brønsted acidic dual–
hydrogen bond donor catalysts have been
applied broadly and successfully to promote
stereoselective nucleophilic substitution re-
actions via leaving group–abstraction path-
ways ( 24 – 26 ). The rearrangement of lithium
boronate substrate3a, prepared by addition
ofn-butyllithium to dichloromethyl boronic
acid pinacol ester ( 1 ), was selected as a model
reaction (Fig. 1D). The arylpyrrolidine-tert-
leucine–derived thiourea 4 was observed to
promote the formation of thea-chloro boronic
ester product5awith modest levels of enan-
tioselectivity (48% ee) when added to a pre-
formed solution of3a. However, the same
reaction proceeded with up to 92% ee when
thiourea 4 was present in the reaction mix-
ture during the generation of3afrom 1 and
n-butyllithium. Under these strongly basic
conditions, thiourea 4 was found to undergo
deprotonation followed byS-alkylation by dis-
placement of a chloride on the starting mate-
rial 1 or product5ato generate a mixture of
isolable isothiourea-boronate adducts (see
supplementary materials for additional de-
tails) ( 27 , 28 ).
The major adduct derived from 4 isolated
from the highly enantioselective“ 4 added be-
forenBuLi”reaction conditions was isothiourea-
boronate derivative6a. This species promoted
the rearrangement reaction in only 26% ee when
added to a preformed solution of boronate3a,
but in 94% ee when introduced to the reaction
mixture beforen-butyllithium. We hypothe-
sized that the active, highly enantioselective
catalyst might beLi-6a, generated cleanly
under the latter conditions by deprotonation
of6abyn-butyllithium but only in small quan-
tities in the presence of preformed boronate
3aalone. Indeed, the lithiated isothiourea-
boronate speciesLi-6agenerated quantitatively
by mixing6aand lithium hexamethyldisilazide
(LiHMDS) in a 1:1 ratio proved to be a highly
effective catalyst for the model reaction. Thus,
rearrangement of3aprepared either by ad-
dition ofn-butyllithium to 1 or by addition
of dichloromethyllithium [generated in situ
via deprotonation of dichloromethane by
lithium diisopropylamide (LDA)] ton-butyl
boronic ester2awas promoted byLi-6ain
94 to 95% ee.
Bench-stable isothiourea-boronate precat-
alyst6awas synthesized independently on754 5NOVEMBER2021•VOL 374 ISSUE 6568 science.orgSCIENCE
Fig. 2. Structural characterization of lithium-isothiourea-boronate catalysts.(A) ReactIR study of the
deprotonation and reprotonation of6a, consistent with the formation of an amide-lithium-isothiourea
chelate. Predicted IR stretching frequencies were calculated from structures optimized at the PCM(Et 2 O)-
M06-2X/6-31+G(d,p) level of theory, with explicit solvation of the lithium cation inLi-6aby dimethyl ether.
See supplementary materials for computational details. (B) X-ray crystal structure ofLi-6bconfirming
the presence of the proposed amide-lithium-isothiourea chelate in the solid state. Hydrogen atoms and a
molecule of pinacol coordinated to Li have been removed for presentation. (C) Assessment of the
performance of catalyst analogLi-6bin the rearrangement of lithium boronate3a. Yield value refers to
gas chromatography yield relative to internal standard.
RESEARCH | REPORTS