ORGANIC CHEMISTRY
Remote steric control for undirectedmeta-selective
CÐH activation of arenes
Boobalan Ramadoss†, Yushu Jin†, Sobi Asako, Laurean Ilies
Regioselective functionalization of arenes remains a challenging problem in organic synthesis. Steric
interactions are often used to block sites adjacent to a given substituent, but they do not distinguish the
remaining remote sites. We report a strategy based on remote steric control, whereby a roof-like
ligand protects the distantparasite in addition to theorthosites, and thereby enables selective
activation ofmetacarbon-hydrogen (C–H) bonds in the absence oforthoorparasubstituents. We
demonstrate this concept for iridium-catalyzedmeta-selective borylation of various monosubstituted
arenes, including complex drug molecules. This strategy has the potential to expand the toolbox of
C–H bond functionalization to previously nondifferentiable reaction sites.
D
irect functionalization of arenes is inte-
gral to the synthesis of pharmaceuticals,
agrochemicals, and a wide variety of
fine chemicals and polymers ( 1 – 4 ).
However, controlling regioselectivity in
this process remains challenging, particularly
in the case of functionalizing electron-rich
monosubstituted arenes at themetaposition
( 5 , 6 ). Transition metal–catalyzed C–H activa-
tion methods have the potential to provide a
general solution to this problem ( 7 – 9 ), and to
date, regiocontrol has typically been achieved
by taking advantage of directing groups ( 10 ) or
by electronically or sterically biasing the sub-
strate. The steric control strategy is successful
in blocking proximal sites (Fig. 1A, a) ( 11 , 12 ),
but differentiation of remote positions such
asmetaandparais difficult. A very bulky
catalyst can block both theorthoandmeta
C–H sites and enableparafunctionalization
(Fig. 1A, b) ( 13 , 14 ), butmeta-selective activa-
tion of a simple monosubstituted arene in the
presence of reactiveparaC–H bonds (Fig. 1A,
c) has been elusive. Bifunctional ligands that
bind a donor or acceptor group on the sub-
strate through attractive interactions such as
hydrogen bonding, Lewis acid-base interac-
tions, or ion pairing have been used formeta
functionalization ( 15 – 17 ), but this strategy
requires specialized groups attached to the
substrate. Other strategies ( 18 – 26 ) that have
been used for remote functionalization also
require specialized directing groups installed
on the substrate. We present here a strategy
based on remote steric control, whereby a
roof-like ligand blocks aparasite (Fig. 1B) to
achievemeta-selective C–H activation in the
presence oforthoandparaC–H bonds. Thus,
a spirobipyridine ligand (L12) in combination
with an iridium catalyst can borylate mono-
substituted arenes selectively at the otherwise
featurelessmetaC–H sites, with regioselectivity
as high as >20:1 (Fig. 1C). Moreover, whereas
sterically demanding ligands typically lower
the activity of the catalyst, the roof design
induces remote steric bias without affecting
the coordination sphere of the metal catalyst;
as shown for the reaction of a protected phenol
(1o), the spirobipyridine ligand is more active
(84% yield) than the typically used bipyridine
ligand, dtbpy (4,4′-di-tert-butyl-2,2′-bipyridine,
24% yield). Under these reaction conditions,
compounds such as alkylbenzenes, anilines, pro-
tected phenol, aryl ethers, and arylsilanes were
selectivelymeta-borylated.
Bipyridine derivatives are privileged ligands
in transition-metal catalysis ( 27 ); for exam-
ple, they are the ligand of choice in iridium-
catalyzed C–H borylation ( 28 ). Commonly used
bipyridine compounds such as dtbpy enable
high reactivity ( 29 ), but control of regiose-
lectivity is poor. There have been efforts to
attach substrate recognition units to the bipyr-
idine core to achieve regioselectivity, but donor
or acceptor groups on the substrate, such as
carbonyl, pyridine, or ammonium, are re-
quired ( 15 – 17 ). The iridium catalyst is typically
sensitive to steric bias, and the steric control
strategiesaandbdepicted in Fig. 1A have
been previously applied ( 11 , 13 , 14 ). To achieve
the remote steric control strategyc,weenvi-
sioned the exploitation of the unexplored space
perpendicular to the coordination plane of the
bipyridine ligand, and we designed a spirobi-
pyridine compound that possesses a bridging
C(sp^3 ) carbon, which allows construction of a
“steric roof”that is expected to block thepara
site and achievemetaselectivity.
The key features of the ligand design are
described in Fig. 2 for the iridium-catalyzed
borylation oftert-butylbenzene (1a). We chose
this substrate because thetert-butyl group is a
standard steric marker in organic synthesis,
and it lacks functionality that may electron-
ically interact with the catalyst. The Ir catalyst
coordinated by the standard dtbpy ligand washighly reactive, and as expected, we obtained a
mixture ofmeta(2a, 40%),para(4a, 20%),
and diborylated (3a, 41%) products. Themeta
selectivity (4.0:1) was enhanced by diboryla-
tion because, for steric reasons, only themeta
product2acan further react to give3a.
Rigidifying the plane of bipyridine (L1) did
not affect the selectivity (3.1:1), whereas an
unsubstituted spirobipyridine compound (L2)
gave a larger amount of diborylated3a, result-
ing in a slightly higher selectivity (6.8:1). The
selectivity was largely unaffected when a phenyl
substituent (L3) was introduced into the fluorene
ring. A pyridyl group (L4) greatly decreased
the reactivity, possibly because of competi-
tive coordination or ligand borylation ( 30 ).
An improvement in selectivity was observed
when we increased the size of the substi-
tuents on the fluorene ring, as predicted by our
“steric roof”model. Thus, groups such as 3,5-
dimethylphenyl (L5), 3,5-di-tert-butylphenyl (L6),
2,6-dimethylphenyl (L7), 2,4,6-trimethylphenyl
(L8), and 9-anthracenyl (L9) gave 7.7:1 to 9.1:1
metaselectivity; notably, forL7,L8, andL9,
diborylation was greatly retarded (2%, 7%,
and 3%, respectively), which proves that the
metaselectivity is genuine and not artifi-
cially enhanced by diborylation. A spirobipyr-
idine ligand bearing a substitutedN-carbazolyl
group (L10) was less selective. The problem
of conformational flexibility for these groups
was solved by using a more rigid Bpin group
(L12), where the boronic ester group aligns in
a planar fashion with the fluorene group (see
fig. S3 for the x-ray structure of the ligand). By
usingL12, we achieved highmetaselectivity
(12:1) while keeping diborylation low (7%).
Changing the borylating reagent to pinacol-
borane (HBpin) (see table S2 for details on the
optimization of the reaction conditions), the
reaction using ligandL12gave the borylated
product with high yield (84%) and highmeta
selectivity (23:1) while diborylation was kept
low (9%). Recovery of the starting material ac-
counted for the remaining mass balance. Mod-
ification of the boronic ester group (L13) gave
a similar result. The use of a bulky triphenylsilyl
group (L11) gave moderate selectivity, prob-
ably because of its rotational flexibility.
Using ligandL12, we next explored the
iridium-catalyzed borylation of a variety of
monosubstituted arenes such as alkylbenzenes,
substituted anilines, protected phenol, and
ethers to achieve high to moderatemetaselec-
tivity, according to the size of the substituent
on the arene (Fig. 3). These electron-rich arene
substrates typically giveortho/paraselectiv-
ity under electrophilic aromatic substitution
conditions ( 31 ) andmeta/paraselectivity under
catalytic C–H bond activation conditions
( 7 – 9 , 28 ). Therefore,meta-selective func-
tionalization is of great synthetic interest,
especially because the boronic ester group can
be readily converted to various functionalities658 11 FEBRUARY 2022•VOL 375 ISSUE 6581 science.orgSCIENCE
RIKEN Center for Sustainable Resource Science, 2-1 Hirosawa,
Wako, Saitama 351-0198, Japan.
*Corresponding author. Email: [email protected] (S.A.);
[email protected] (L.I.)
These authors contributed equally to this work.
RESEARCH | REPORTS