temperature. The borylation did not pro-
ceed at all in the absence ofRL(entry 5).
Unfavorable effects on the yield and enantio-
selectivity were apparent in the absence of
2,6-lutidine (entry 6) in accord with our prior
works ( 19 , 20 ). Increasing theRLloading
(6 mol %) led to diminished yield and en-
antioselectivity (entry 7). The absence ofLin
the presence ofRL(6 mol %, two equiv to
Ir) fully impeded the reaction (entry 8). The
catalytic reaction of1ausing the antipode
of the ligand, (S,S)-Lgave the identical re-
sult but with the opposite absolute configura-
tion of the product.
Effects of the structure of the urea-pyridine
ligands are also summarized in Fig. 2 (see fig.
S1 for the effects of other derivatives). The use
of a pyridine-urea ligand (U1)withamethyl-
ene linker at the pyridine 2-position instead of
RLredirected the site selectivity toward the
borylation of a proximalb-methylene C–H
bond (14%^1 H NMR yield for 4 ). Changing the
linker from the methylene group inU1back
to theortho-phenylene group inU2led again
to activation of theg-C–H bond, giving2a
(17%) but still with a trace amount of the
b-borylation product. The 4-pyridyl isomer
U3delivered onlyg-C–H borylation product
2abut with a low product yield (12%).
The enantioselective borylation also oc-
curred in high yields and excellent site- and
enantioselectivities with secondary carbox-
amides (1b–e,Fig.3A,seecaption)andesters
(1f–h,Fig.3B,seecaption).Inthesecases,the
secondary alkylboronate products (2b–h)were
sufficientlystabletobeisolatedbysilicagelcol-
umn chromatography.N-(t-Butyl)hexanamide
was borylated site selectively at theg-methylene
C–H bond of the aliphatic hydrocarbon back-
bone, giving2b(93% ee). TheN-t-butyl group,
which has terminal C(sp^3 )–Hbondsgto the
carbonyl group, remained intact. Hexanoic
acid anilide similarly gave2cwith excellent
enantioselectivity (98% ee). The C(sp^2 )–Hbonds
in theN-phenyl group, which are also located
gto the carbonyl group, stayed unaffected.
Likewise,N-benzyl andN-1-naphthylhexana-
mide derivatives delivered2d(98% ee) and
2e(95% ee) as the sole products, respectively.
Esters were equally amenable. Ethyl hexanoate
gave2f(90% ee) exclusively, resulting from
the activation of a methyleneg-C(sp^3 )–Hbond
in the aliphatic acyl group rather than the ter-
minalmethylC–Hbondsinthealkoxygroup.
Enhanced enantioselectivities were observed
inthecaseoft-butyl and benzyl hexanoic acid
esters, giving2g(93% ee) and2h(95% ee),
respectively.
Complementary to the reactivity of1a,var-
iousN,N-disubstituted hexanamides with
differentN-substituents were suitable sub-
strates (Fig. 3C). Thus, substrates with linear
or branchedN,N-dialkylamino groups, in-
cluding methyl (1i), ethyl (1j), hexyl (1k),
isopropyl (1l), and isobutyl (1m) groups, un-
derwent the borylation followed by the mild
stereoretentive oxidation with NaBO 3 with
exclusive site selectivity to theg-C–Hbond
(92 to 94% ee) of the hexanamide chain.N-
Cyclohexyl-N-methyl- andN,N-dicyclohexyl-
hexanamides were converted to3n(93% ee)
and3o(95% ee), respectively. Hexanamides
with cyclic amine moieties, including piperidine
(1p), morpholine (1q), andN-methylpiperazine
(1r), likewise participated in the reaction (92%
ee). Hexanamides bearingN-methylaniline
andN-methyl-N-benzylamine moieties sim-
ilarly gave3s(94% ee) and3t(96% ee) as the
sole products, respectively.
The applicability of theprotocoltowardvar-
ious carboxamides with different aliphatic
chains was investigated forN,N-dibenzyl-
substituted amide derivatives (Fig. 3D). The
valeric acid derivative, one carbon shorter than
1a,gave3uwith excellent enantioselectivity
(97% ee). The undecylic acid derivative under-
went the borylation at the C-4,g-C–H bond
to deliver product3v(96% ee). The borylation
of more congested substrates with cyclohexyl
(1w), benzyl (1x), or phenyl (1y)groupsatthe
g-methylene carbon also proceeded smoothly
(94 to 97% ee), demonstrating the substantial
tolerance of this protocol toward steric hin-
drance, whereas the sense of enantioselection
was reversed for the reaction of the phenyl-
substituted amide (3y).
The protocol was also applicable to unsat-
urated alkenoic acid derivatives (5a–d) (Fig.
3D). Thus, the benzyl ester (5a)of6-nonenoic
acid gave the homoallylic alcohol7a(93% ee)
upon oxidation of theg-boryl product (6a),
whereas the anilide (5b) delivered theg-boryl-
w-3-N-phenylnonenamide6b(94% ee). Fur-thermore, the anilide (5c) of 12-tridecenoic
acid with a distal terminal alkene moiety gave
6c(93% ee) as the only product. The reaction
of the anilide (5d) of linoleic acid gave6d
(95% ee) exclusively without any trace of
side reactions such as alkene migration in
either the product (85%) or the recovered
substrate (12%).
Figure 3E showcases thesynthetic utility of
the borylation protocol by the transformations
of theg-borylhexanoic acid anilide (R)-2c.A
lower loading (1 mol %) of the Ir-L*catalyst
enabled the gram-scale synthesis of (R)-2c
(1.32 g, 83%) from1cwithout erosion of the
enantioselectivity. Subsequently, mild oxi-
dation of (R)-2cprovided the corresponding
alcohol (R)- 8. The Rh-catalyzed stereoreten-
tive carboboration ( 19 )of(R)-2cto benzyl
isocyanate gave the enantioenricheda-alkyl-
glutaric acid diamide (R)- 9. The amination
of (R)-2callowed direct access to a pharma-
cologically interestingg-alkyl-g-aminobutyric
acid (GABA) derivative [(R)- 10 ]( 24 ). The transition-
metal free cross-coupling between (R)-2cand
3-bromoanisole ( 25 ) furnished theg-arylated
derivative (R)- 11 (98%ee).
To gain insight into the structural features
of the modular chiral catalyst as proposed
in Fig. 1B, we performed preliminary quan-
tum chemical calculations, focusing on C–H
bond cleavage by the catalyst, employingN,
N-dimethylpentanamide as a model substrate,
Ir(Bpin) 3 - L*as the catalyst backbone, andRL
as the receptor ligand. We used the GFN2-xTB
method ( 26 ) and the artificial force–induced
reaction (AFIR) method ( 27 , 28 ) to explore
various conformations ofRLand the amide
substrate associated with the cleavage of
the C–H bonds of the hydrogen-bondedReyeset al.,Science 369 , 970–974 (2020) 21 August 2020 4of5
Fig. 4. Three-dimensional representation of the calculated transition state (TS) leading to the major
enantiomer.(A) The geometrical features of TS-Rshowing C(sp^3 )–H···O andp/pnoncovalent interactions between
the substrate and the modular catalyst. (B) The space-filling model shows the reaction cavity and the efficient
fitting of the substrate within the catalytic pocket. The binaphthyl frameworks ofL*are shown in green, the
isopropyl (iPr) groups in the triisopropylsilyl (TIPS) moiety ofL*in blue, theRLin cyan, and the substrate in yellow.RESEARCH | REPORT