Iridium-catalyzed C-H borylation of heteroarenes: Scope, regioselectivity, application to late-stage functionalization, and mechanism

Article abstract of DOI:10.1021/ja412563e

A study on the iridium-catalyzed C-H borylation of heteroarenes is reported. Several heteroarenes containing multiple heteroatoms were found to be amenable to C-H borylation catalyzed by the combination of an iridium(I) precursor and tetramethylphenanthroline. The investigations of the scope of the reaction led to the development of powerful rules for predicting the regioselectivity of borylation, foremost of which is that borylation occurs distal to nitrogen atoms. One-pot functionalizations are reported of the heteroaryl boronate esters formed in situ, demonstrating the usefulness of the reported methodology for the synthesis of complex heteroaryl structures. Application of this methodology to the synthesis and late-stage functionalization of biologically active compounds is also demonstrated. Mechanistic studies show that basic heteroarenes can bind to the catalyst and alter the resting state from the olefin-bound complex observed during arene borylation to a species containing a bound heteroarene, leading to catalyst deactivation. Studies on the origins of the observed regioselectivity show that borylation occurs distal to N-H bonds due to rapid N-H borylation, creating an unfavorable steric environment for borylation adjacent to these bonds. Computational studies and mechanistic studies show that the lack of observable borylation of C-H bonds adjacent to basic nitrogen is not the result of coordination to a bulky Lewis acid prior to C-H activation, but the combination of a higher-energy pathway for the borylation of these bonds relative to other C-H bonds and the instability of the products formed from borylation adjacent to basic nitrogen.

Full text of DOI:10.1021/ja412563e

Article
pubs.acs.org/JACS
Iridium-Catalyzed C−H Borylation of Heteroarenes: Scope,
Regioselectivity, Application to Late-Stage Functionalization, and
Mechanism
Matthew A. Larsen and John F. Hartwig*
Department of Chemistry, University of California, Berkeley, California 94720, United States
S
* Supporting Information
ABSTRACT: A study on the iridium-catalyzed C−H
borylation of heteroarenes is reported. Several heteroarenes
containing multiple heteroatoms were found to be amenable
to C−H borylation catalyzed by the combination of an
iridium(I) precursor and tetramethylphenanthroline. The
investigations of the scope of the reaction led to the
development of powerful rules for predicting the regioselec-
tivity of borylation, foremost of which is that borylation occurs
distal to nitrogen atoms. One-pot functionalizations are reported of the heteroaryl boronate esters formed in situ, demonstrating
the usefulness of the reported methodology for the synthesis of complex heteroaryl structures. Application of this methodology
to the synthesis and late-stage functionalization of biologically active compounds is also demonstrated. Mechanistic studies show
that basic heteroarenes can bind to the catalyst and alter the resting state from the olefin-bound complex observed during arene
borylation to a species containing a bound heteroarene, leading to catalyst deactivation. Studies on the origins of the observed
regioselectivity show that borylation occurs distal to N−H bonds due to rapid N−H borylation, creating an unfavorable steric
environment for borylation adjacent to these bonds. Computational studies and mechanistic studies show that the lack of
observable borylation of C−H bonds adjacent to basic nitrogen is not the result of coordination to a bulky Lewis acid prior to
C−H activation, but the combination of a higher-energy pathway for the borylation of these bonds relative to other C−H bonds
and the instability of the products formed from borylation adjacent to basic nitrogen.
Scheme 1. Examples of the Borylation of Heteroarenes
1. INTRODUCTION
The C−H bond functionalization of heteroarenes is an
important synthetic method because heteroarenes are common
units in the structures of pharmaceuticals, agrochemical
products, and electronic materials.1 Although transition-metal
catalyzed reactions of arenes, such as cross-coupling reactions,
typically occur with some heteroarenes,2−4 the more basic the
heteroarene and the more nitrogen atoms contained in the ring
system of the heteroarene, the more challenging the translation
of the catalytic reaction of an arene is to the catalytic reaction of
the heteroarene.
In the past decade, the iridium-catalyzed C−H borylation of
arenes has become a widely used method for the functionaliza-
tion of arenes because of its ability to produce highly versatile
aryl organoboronate ester intermediates from arenes without
the need for reactive groups, such as halides or sulfonates.5,6
C−H borylation occurs with some heteroarenes, and the
reactions of benzo-fused or non-benzo-fused five-membered
ring heteroarenes containing one heteroatom are some of the
most reactive substrates for C−H borylation.7−9 A few
examples of the borylation of heteroarenes containing multiple
heteroatoms and high nitrogen content also have been reported
(Scheme 1), but the scope of these reactions has not been
explored.10−14
of heteroarenes could allow for rapid derivatization of more
complex medicinally important compounds and natural
In addition to improving methods for the modification of
heteroarene building blocks, an understanding of the borylation
Received: December 15, 2013
Published: February 7, 2014
© 2014 American Chemical Society
4287
dx.doi.org/10.1021/ja412563e | J. Am. Chem. Soc. 2014, 136, 4287−4299

Journal of the American Chemical Society
Article
products. The combination of C−H borylation and elaboration
of the heteroarylboronate ester by methods such as Suzuki
coupling,15 Chan−Lam−Evans coupling,16 trifluoromethyla-
tion,17 and oxidation18 can lead to many derivatives after a
single C−H bond functionalization reaction.
state. The binding constant of the heteroarene to the catalyst
could also be sufficiently high to form 3 irreversibly, preventing
any appreciable formation of borylated product.
With these challenges associated with the borylation of
heteroarenes in mind, we sought to explore the boundaries of
the reaction scope, assess the applicability of this reaction to
late-stage functionalization, reveal differences between the
mechanism of the borylation of heteroarenes and the borylation
of arenes, and investigate the origins of observed regioselectiv-
ities.
Here, we report a detailed study on the borylation of
heteroarenes. The scope of the reactions we studied
encompasses benzoxazoles, pyrimidines, as well as unprotected
benzimidazoles, pyrazoles, and azaindoles. These heteroarenes
are common heteroaryl motifs in medicinal chemistry22−31 and
are present in several currently marketed drugs (Figure 1).
One challenge facing the application of catalysis to
heteroarenes is the diversity of electronic properties of
heteroarenes. The electron density at different positions of
the heteroarene varies with the distribution of heteroatoms.
Moreover, binding of a basic atom to the catalyst can poison its
activity. The propensity of different heteroarenes to poison the
catalyst varies because the basicity of the nonbonded electron
pairs on the nitrogen atoms vary.19 Thus, it is important to
understand the trends in rate and selectivity for the borylation
of these substrates and the mechanistic origin of these trends.
There are many five- and six-membered ring heteroarenes, and
all of these rings can contain substituents that have the
potential to affect the rate and selectivity of the reaction. Thus,
we sought to understand how the nitrogen atoms in the
heteroaryl rings affect the reaction rates and the site selectivity.
The regioselectivity for the borylation of heteroarenes differs
substantially from the regioselectivity for the borylation of
arenes, and the regioselectivity for the borylation of various
heteroarenes differs from each other. For the borylation of
arenes, the regioselectivity is controlled almost exclusively by
steric effects,5 and this sterically controlled selectivity is
complementary to the origin of selectivity during classical
arene functionalization reactions, such as directed ortho
metalation or Friedel−Crafts reactions. In contrast, the
regioselectivity for the borylation of heteroarenes is controlled
by a combination of electronic and steric effects. The
regioselectivity for the borylation of pyrrole, indole, furan,
and thiophene tends to occur at the most acidic C−H bond.20
However, the regioselectivity for the borylation of heteroarenes
containing a basic nitrogen, such as pyridine, quinoline, and
other azines, undergoes borylation at the C−H bond located
beta or gamma to the basic nitrogen.9
Figure 1. Pharmaceuticals containing benzoxazole, benzimidazole,
pyrimidine, pyrazole, or azaindole moieties.
Basic heteroaryl substrates, such as pyridine, could occupy
the vacant coordination site necessary for C−H activation in
the 16-electron, five-coordinate Ir(III) species 1 (Scheme 2).21
These reactions are amenable to one-pot sequences involving
borylation and functionalization for the late-stage modification
and the synthesis of biologically active heteroarenes. Finally, we
report findings from mechanistic and computational studies
that elucidate the origin of the regioselectivity. These studies
show why borylation tends to occur at positions distal to
nitrogen more rapidly than it occurs at positions adjacent to the
nitrogen of azines and the nitrogen of azoles containing
multiple heteroatoms.
Scheme 2. Coordination of Heteroarenes to the Active
Catalyst
2. RESULTS AND DISCUSSION
2.1. Method Development. The most common reaction
system for the borylation of arenes is B2pin2 (bis(pinacolato)-
diboron) as the boron source and a combination of
[Ir(COD)OMe]2 (bis(1,5-cyclooctadiene)di-iridium(I)-
dimethoxide) and dtbpy (4,4′-di-tert-butyl bipyridine) as
catalyst. Recently, the borylations of octane,32 cyclic ethers,32
chloromethylsilanes,33 benzylic C−H bonds,34 and arenes35
have been shown to occur in higher yield when catalyzed by the
combination of [Ir(COD)OMe]2 and Me4phen (3,4,7,8-
tetramethyl-1,10-phenanthroline) than when catalyzed by the
combination of [Ir(COD)OMe]2 and dtbpy. The increased
reactivity of the Me4phen-Ir system over the dtbpy-Ir system is
presumably due to the greater electron-donating ability and
backbone rigidity of Me4phen compared to those of dtbpy.
When the heteroarene binds more strongly than the olefin
bound to the precatalysts or when the concentration of the
heteroarene is sufficiently high, relative to the olefin, the resting
state of the catalyst could shift from the alkene complex 2 to
heteroarene complex 3. In such a scenario, the reaction will be
slower, relative to the analogous reaction with 2 as the resting
4288
dx.doi.org/10.1021/ja412563e | J. Am. Chem. Soc. 2014, 136, 4287−4299

Journal of the American Chemical Society
Article
We began our studies on the borylation of basic heteroarenes
containing multiple heteroatoms by investigating both of the
catalytic systems for the borylation of benzo-fused azoles
(Table 1). Initially we observed low conversion for the
Table 1. Effect of Ligand on the Borylation of Selected
a
Azoles
Figure 2. (a) Reaction profile for the borylation of 5-bromo-2-
methylbenzoxazole with B2pin2 (1.0 equiv), 0.25 mol % [Ir(COD)-
OMe]2 and 0.5 mol % ligand in THF at RT. (b) Reaction profile for
the borylation of 4-chloro-2-methyl-7-azaindole with B2pin2 (1.5
equiv), 1.0 mol % [Ir(COD)OMe]2 and 2.0 mol % ligand. Yields were
determined by GC analysis.
a
Table 2. Effect of B2pin2 Stoichiometry on Product Yield
a
b
Reactions were conducted on a 0.25 mmol scale. Determined by 1H
c
NMR spectroscopy. Reaction conducted at room temperature.
d
Reaction conducted at 80 °C.
reactions of 2-methylbenzoxazole and 5-bromo-2-methylben-
zoxazole (entries 1 and 2) catalyzed by the combination of
Me4phen and [Ir(COD)OMe]2 at temperatures below 50 °C.
This lack of reactivity was most likely due to the low solubility
of Me4phen in THF. However, heating a mixture of B2pin2,
Me4phen, and [Ir(COD)OMe]2 in THF at 80 °C and
subsequent addition of this mixture to the heteroarene led to
high conversions, even at room temperature.
The yields from the borylation reactions of 5-bromo-2-
methylbenzoxazole conducted with dtbpy-Ir and Me4phen-Ir as
catalyst were similar to each other (Table 1, entry 2). However,
the difference in yield between the two catalyst systems was
significant for the borylations of azoles containing free N−H
bonds, such as 2-methylbenzimidazole and 4-chloro-2-methyl-
7-azaindole (entries 3 and 4). In these cases, the combination
of iridium and Me4phen was more active than the combination
of iridium and dtbpy.
Reaction progress for the borylation of 5-bromo-2-methyl-
benzoxazole (Figure 2a) and 4-chloro-2-methyl-7-azaindole
(Figure 2b) conducted with dtbpy-Ir or Me4phen-Ir as catalyst
was followed by gas chromatography (GC). Initial rates for
formation of the product from the two substrates were similar
with both catalyst systems, and the reaction profiles for the
borylation of 5-bromo-2-methyl benzoxazole were essentially
identical. However, the borylation of 4-chloro-2-methyl-7-
azaindole catalyzed by the dtbpy-Ir system stalled at 78 min
with a product yield of 27% and only 70% conversion of
starting material. Borylation of this substrate under identical
conditions, but with Me4phen-Ir as catalyst, formed product in
72% yield with a standard exponential appearance of product.
Further reaction development with Me4phen-Ir complexes as
catalyst revealed that the stoichiometry of B2pin2 relative to
heteroarene significantly influenced the yields for reactions of
a
b
Reactions were conducted on a 0.25 mmol scale. Determined by GC
c
d
analysis. Reaction conducted at RT. Reaction conducted at 80 °C.
heteroarenes containing N−H bonds, but had little effect on
reactions of heteroarenes lacking N−H bonds (Table 2). The
borylation of substrates containing free N−H bonds (Table 2,
entry 2) required 1.5 equiv of B2pin2 to form the product in
high yield. The requirement for additional B2pin2 to obtain a
high yield was found to result from the formation of an N-Bpin
species, a finding that was also reported by Smith, Maleczka,
and Krska during the preparation of this document. A detailed
investigation of the N−H borylation phenomena will be
presented later.
2.2. Scope of the Borylation of Heteroarenes. With
Me4phen-Ir as catalyst, we found that benzoxazoles and
pyrimidines, as well as unprotected benzimidazoles, pyrazoles,
and azaindoles, underwent the borylation process in good yield.
Benzoxazoles underwent borylation at room temperature
(Chart 1), provided that a substituent was present at the 2-
position, leading to heteroaryl boronate esters 4a−f in good to
excellent yields. Borylation occurred almost exclusively at the 7-
position, ortho to oxygen (confirmed via independent synthesis
of 4a). Smith, Maleczka, and Singleton proposed that the
regioselectivity of the borylation of benzodioxole ortho to
oxygen resulted from the greater acidity of the corresponding
C−H bond at this position.20 The same effect might be leading
to borylation adjacent to oxygen in benzoxazoles. The
borylation of 2-methylbenzothiazole led to a mixture of
products 4g within a substantially lower yield than was
observed for borylation of the oxygen analogue.
4289
dx.doi.org/10.1021/ja412563e | J. Am. Chem. Soc. 2014, 136, 4287−4299

Journal of the American Chemical Society
Article
a
a
Chart 1. Borylation of Benzoxazoles and Benzothiazoles
Chart 3. Borylation of Benzimidazoles
a
Reactions were conducted on a 0.25 mmol scale. Yields were
determined by 1H NMR spectroscopy, and isolated yields are reported
in parentheses for reactions conducted on a 0.50 mmol scale.
resonances were observed for these products due to the rapid
tautomerism of benzimidazoles. A slightly lower yield was
obtained for 4-methyl 6b than for the less electron-rich 6a, 6c,
and 6d, but the yield for the reaction of 6b was still at a useful
level.
Smith, Maleczka, and Krska reported, during our current
study, the borylation of unsubstituted pyrazoles and azaindoles.
Here we present a more complete scope of these classes of
substrates, particularly those containing varying substitution
patterns. In analogy to the regioselectivity observed for
benzimidazoles, the borylation of pyrazoles (Chart 4) occurred
a
Reactions were conducted on a 0.25 mmol scale. Yields were
determined by 1H NMR spectroscopy, and isolated yields are reported
in parentheses for reactions conducted on a 0.50 mmol scale.
b
c
Reaction conducted with 0.75 equiv of B2pin2. Reaction conducted
at 50 °C with 1.0 mol % [Ir(COD)OMe]2 and 2.0 mol % Me4phen.
Yields were determined by GC chromatography.
a
Chart 2. Borylation of Pyrimidines
a
Chart 4. Borylation of Pyrazoles
a
Reactions were conducted on a 0.25 mmol scale. Yields were
determined by 1H NMR spectroscopy, and isolated yields are reported
in parentheses for reactions conducted on a 0.50 mmol scale.
a
Reactions were conducted on a 0.25 mmol scale. Yields were
b
determined by 1H NMR spectroscopy, and isolated yields are reported
in parentheses for reactions conducted on a 0.50 mmol scale.
Reaction was conducted with 2.5 mol % [Ir(COD)OMe]2 and 5.0
c
mol % Me4phen. Reaction was conducted with 0.75 equiv of B2pin2.
The borylation of pyrimidines led to formation of heteroaryl
boronate esters 5a−c (Chart 2), bearing the boryl group
exclusively at the 5-position. This selectivity is complementary
to the selectivity of radical C−H functionalizations of
pyrimidines recently reported by the Baran group.36 Once
again, the heteroarene containing a substituent at the 2-position
was more reactive than the analogue lacking a substituent at
this position. For example, the reaction of the parent
pyrimidine 5a occurred in 14% yield, whereas the reaction of
the 2-methyl analogue 5b occurred in 70% yield. Heteroaryl
boronates 5a−c were not stable to silica gel chromatography;
therefore, we purified these materials by Kugelrohr distillation.
However, heteroaryl boronate esters also can be elaborated in
situ, as will be shown later.
beta to nitrogen in most cases. This site is the one at which
electrophilic aromatic substitution typically occurs in these
heterocycles.37 The exception to this trend of reaction at the
position beta to nitrogen was the reaction of trifluoromethyl-
substituted 7d, which underwent borylation adjacent to
nitrogen. Borylation beta to nitrogen was also observed by
Smith and Maleczka for Boc-protected pyrazole.10 In contrast,
the borylation of N-methylpyrazole has been reported to occur
adjacent to the N-methyl moiety.38 The borylation of
unprotected pyrazole to generate 7a, as has been previously
reported,11 occurs with selectivity that is similar to the
selectivity for the borylation of the bulky Boc-protected
pyrazole, due to rapid formation of a bulky N-Bpin species,
which will be discussed later in this report.
Benzimidazoles substituted at the 2-position led to the
formation of products 6a−d (Chart 3) from borylation distal to
Reactions of unprotected azaindoles formed boryl-azaindoles
8a−e (Chart 5) in moderate to excellent yields. Products 8a−c
all resulted from borylation at the 3-position. Like the
1
nitrogen at the 5- or 6-positions. Only one set of H NMR
4290
dx.doi.org/10.1021/ja412563e | J. Am. Chem. Soc. 2014, 136, 4287−4299

Journal of the American Chemical Society
Article
a
Chart 5. Borylation of Azaindoles
Figure 3. Rules for predicting regioselectivity.
a
Reactions were conducted on a 0.25 mmol scale. Yields were
determined by 1H NMR spectroscopy, and isolated yields are reported
in parentheses for reactions conducted on a 0.50 mmol scale.
b
aryl hydrogen with the furthest downfield chemical shift was
Reaction conducted on a 0.50 mmol scale at RT with 1.0 equiv of
1
typically the site of reactivity. The H NMR chemical shifts of
B2pin2.
aryl hydrogens adjacent to basic nitrogen atoms and free N−H
bonds are typically downfield of those for other aryl hydrogens.
Thus, the lack of borylation adjacent to nitrogen is a deviation
from Steel’s observation, and the correlation between chemical
shift and site selectivity does not extend in a general fashion to
the borylation of basic heteroarenes. Factors beyond the
electronic properties of the C−H bond control the position of
the borylation of these heteroarenes.
Second, substrates containing sulfur or oxygen in their
aromatic systems undergo borylation ortho to these heter-
oatoms with moderate to excellent selectivity. Our scope
demonstrates this trend in the borylation of benzothiazoles and
benzoxazoles, with the borylation of benzothiazoles being
significantly less selective than the borylation of the analogous
benzoxazoles. This selectivity does correlate with the predicted
site selectivity for C−H borylation postulated by Steel et al.
because its origin is likely electronic.44
Third, the borylation of basic heteroarenes has a large steric
effect on regioselectivity. For example, no borylation is
observed on the tolyl ring of 2-(p-tolyl)benzoxazole (Chart 1,
4e). The main difference between the regioselectivity for the
borylation of heteroarenes and the borylation of arenes is that
the electronic properties of the heteroarenes can sometimes
override the steric preference (benzoxazoles undergo ortho-
borylation, rather than meta-borylation).
2.4. Tandem Reactions. In several cases, we experienced a
significant difference between spectroscopic yields and isolated
yields of the heteroaryl boronate esters formed from C−H
borylation. This difference results from several of the products
being unstable or immobile on silica. The difference in the two
yields was as large as 32% (product 5b, Chart 2). However, one
valuable aspect of the borylation of C−H bonds is the ability to
use the arylboronate esters in situ because the reaction gives
only volatile HBpin or H2 as side product and occurs with low
loadings of iridium.45−49
borylation of pyrazoles, the site for the borylation of azaindoles
is the site at which electrophilic aromatic substitution typically
occurs.39 This selectivity contrasts that for the borylation of
indole; the reaction of indole occurs at the 2-position.40 This
regioselectivity was also observed for the borylation of N-Boc-
7-azaindole, which yielded compound 8f.41 However, when a
substituent was present at the 3-position, borylation of the
azaindole occurred at the 5-position. For example, the
borylation of 3-ethyl azaindole occurred selectively at the 5-
position to form boryl-azaindole 8d. In addition, the borylation
of 4-chloro-2-methyl-7-azaindole led to functionalization at the
5-position; this selectivity is analogous to that reported for the
ortho metalation of N-TIPS-4-chloro-7-azaindole.42
2.3. Guidelines for Predicting the Regioselectivity of
the Borylation of Heteroarenes. Our results show clearly
that the regioselectivity for borylation of heteroarenes
containing two heteroatoms is high in many cases. However,
these selectivities might seem difficult to predict at first glance.
Thus, it is important to provide some guidelines that allow one
to predict the regioselectivity of the borylation of heteroarenes.
Just a few trends allow one to make these predictions (Figure
3).
First, the borylation rarely occurs adjacent to nitrogen,
whether the nitrogen atom is contained in the ring (e.g.,
pyrazoles) or is external to the ring (e.g., benzimidazoles). This
result is consistent with the selectivity for the borylation of N−
H bearing heterocycles reported by Smith and Maleczka.11
Only one exception to this trend was observed, and this case
(the borylation of trifluoromethyl-substituted pyrazole 7d,
Chart 4) involved the reaction of a heteroarene containing a
strong electron-withdrawing group that presumably perturbed
the electronic properties of the ring.43 This trend also was
followed whether the nitrogen was contained in an N−H bond
or was unsaturated and basic. Although a simple trend, the
origins of this trend are due to multiple factors, and these
factors will be presented later in this paper.
Thus, to avoid isolation of the heteroaryl boronate ester
intermediates, one-pot tandem reaction sequences are used,
and several reports describe tandem borylation/functionaliza-
tion sequences.15,17,47,49,50 Here, we report several new
applications of this methodology to the functionalization of
heteroarylboronate esters generated by C−H bond borylation.
Steel and Marder recently discussed the relationship between
1
the site of borylation of disubstituted arenes and the H NMR
chemical shift of the aryl hydrogens.44 They showed that the
4291
dx.doi.org/10.1021/ja412563e | J. Am. Chem. Soc. 2014, 136, 4287−4299

Journal of the American Chemical Society
Article
Related sequences formed heteroaryl alcohols. Borylation
and oxidation of the intermediate with sodium perborate gave
9e in high yield and with minimal side products.
Scheme 3. C−H Functionalization via Tandem Borylation/
a
Functionalization Sequences
We also showed that the products could be converted to
trifluoroborates, which are stable, crystalline reagents for a
range of transformations. Treatment of the crude product from
the borylation of 4-chloro-2-methyl-7-azaindole with KHF2 led
to formation of the trifluoroborate 9f in 50% yield. These
tandem reaction sequences demonstrate the versatility of
boronate ester intermediates for the synthesis of complex
heteroaryl compounds.
2.5. Applications to Synthesis and Late-Stage
Functionalization. We next sought to assess the suitability
of the combination of C−H borylation and functionalization
for the late-stage functionalization of biologically active
compounds. Compound 11 (Scheme 4) is an intermediate in
Scheme 4. Tandem C−H Borylation and Oxidation as a
Route to a Potential Diabetes Therapeutic
a
Conditions: (a) B2pin2 (1−1.5 equiv), [Ir(COD)OMe]2 0.25−1.0
mol %, Me4phen 0.5−2.0 mol %, THF, RT−80 °C, 16−48 h, removal
of solvents. (b) 4-Bromotoluene (1.3 equiv), Pd(dba)2 5 mol %,
QPhos 5 mol %, CsF (3 equiv), dioxane, H2O, 100 °C, 21 h. (c) 2-
Bromopyridine (1.3 equiv), complex 10.5 mol %, K2CO3 (3 equiv),
dioxane, H2O, 100 °C, 2 h. (d) 4-Bromobenzonitrile (1.3 equiv),
complex 10.5 mol %, K2CO3 (3 equiv), dioxane, H2O, 100 °C, 2 h. (d)
[(phen)CuCF3] (1.2 equiv), KF (1 equiv), DMF, 50 °C, 21 h. (e)
NaBO3·4 H2O (3 equiv), THF, H2O, RT, 15 min. (f) KHF2 (4.6
equiv), MeOH, H2O, 5 h.
Scheme 3 shows several different methods for the
functionalization of the crude products of heteroaryl borylation,
including Suzuki coupling, copper-mediated functionalization,
and oxidation. For example, biaryl 9a formed in good yield by a
Suzuki coupling catalyzed by the combination of QPhos and
Pd(dba)2. Application of this catalytic system to the synthesis of
biaryl 9b containing multiple nitrogen atoms gave no product.
However, conducting the reaction with the single-component
precatalyst 10, which is similar to a precatalyst shown to be
highly active for Suzuki couplings,51 led to the formation of 9b
in 65% yield, based on the starting heteroarene.
We were unable to isolate the heteroaryl boronate ester
formed by the borylation of 3-dimethylamino azaindole.
However, the product of C−H borylation of this substituted
heteroarene generated in situ underwent Suzuki coupling with
10 as catalyst to form the 5-aryl derivative 9c in 67% yield
based on the starting heteroarene. The reactions that yielded 9b
and 9c are rare examples of Suzuki couplings of basic heteroaryl
boronate esters containing free N−H moieties.
The trifluoromethyl-substituted heteroarene 9d was synthe-
sized by a protocol developed by our group for the oxidative
trifluoromethylation of arylboronate esters.17 Borylation of the
benoxazole, followed by treatment of the heteroarylboronate
ester with (phen)Cu(CF3) gave the 7-trifluoromethyl-sub-
stituted benzoxazole 9d. A similar sequence52 conducted with
Togni’s reagent as the source of the CF3 group also generated
9d, but the product was not isolable from the side products.
a synthesis, developed by AstraZeneca, of the potential type II
diabetes therapeutic 12.53 As described in AstraZeneca’s report,
the synthesis of 11 was accomplished by a tandem sequence of
Miyaura borylation and oxidation to give a 23% yield of the
desired product. In contrast, the tandem sequence of C−H
borylation and oxidation occurred to form 11 in a 91% yield
and with a less expensive starting material.
A benzoxazole-containing c-Met kinase inhibitor has been
identified as a potential cancer therapeutic.54 The penultimate
product 13 (Scheme 5) in the synthesis is converted to the
inhibitor by removal of the Boc and tert-butyl groups.
Structure−activity relationship (SAR) studies varying the
substitution pattern on the benzoxazole moiety led to the
discovery of several species with high c-Met kinase inhibition.
The substitution patterns on the benzoxazole moieties were
established in the second step of the reported synthesis (see
Supporting Information [SI] for synthetic scheme). Thus, SAR
studies could be greatly accelerated if a structure formed later in
the synthetic sequence could be modified.
Using our rules for prediciting the regioselectivity, we
predicted that borylation would occur selectively ortho to
oxygen. Indeed, the heteroarylboronate ester intermediate 14
formed in 51% yield (as determined by 1H NMR spectroscopy)
from the borylation, and the Suzuki coupling of 2-
4292
dx.doi.org/10.1021/ja412563e | J. Am. Chem. Soc. 2014, 136, 4287−4299

Journal of the American Chemical Society
Article
17 was chosen as catalyst because thermal dissociation of COE
leads to the catalytically active 16-electron species and because
17 can be synthesized in gram quantities.21 The synthesis of the
Me4phen-ligated analogue of 17 has been reported;11 however,
we were unable to synthesize this complex in an acceptable
amount and purity for these studies.
Scheme 5. Late-Stage Functionalization of a c-Met Kinase
Inhibitor
We measured initial rates for the borylation of 16 with varied
concentrations of 16, B2pin2, and 17 (Table 3). We observed a
first-order dependence of the rate on the concentration of the
catalyst (Figure 4), zero-order dependence of the rate on the
bromopyridine with 14 led to 15 (a highly potent c-Met kinase
inhibitor) in 20% isolated yield from the parent arene. Although
15 was obtained in low yield, the amount of material would be
sufficient for initial testing. Thus, this sequence further
demonstrates the utility of C−H borylation for rapid
derivatization of biologically active compounds.
2.6. Mechanistic Studies of Heteroarene Borylation.
Our studies of the scope and applications of heteroaryl C−H
borylation demonstrate that the iridium-catalyzed C−H
borylation is a powerful tool for the functionalization of
heteroarenes, even those containing multiple nitrogen atoms.
The scope of the reaction encompasses several heteroaryl
motifs that are prevalent in medicinal chemistry. However, to
predict the regioselectivity and to understand the factors
controlling the rate and scope, an understanding of the
mechanism for the borylation of these heteroarenes is needed.
Thus, we conducted experimental mechanistic studies and
computational studies on the reported reactions.
Figure 4. Dependence of the initial rate for the formation of 18 on the
concentration of catalyst 17.
concentration of B2pin2, and zero-order dependence of the rate
on the concentration of the heteroarene. The zero-order
dependence in substrate is distinct from the first-order
dependence in substrate of the rate of arene borylation
measured previously.21
This result is consistent with the mechanism shown in
Scheme 6. In this scheme, the resting state of the catalyst
contains the heteroarene, instead of the olefin of the precursor.
Dissociation of the heteroarene, followed by reaction of the
released heteroarene at the C−H bond, or simple rearrange-
ment of the heteroarene from a nitrogen ligand to a π-arene or
C−H σ complex would then lead to the transition state for C−
H bond cleavage.
2.6.1. Kinetic Studies of the Borylation of Represen-
tative Heteroarenes. We began our mechanistic investiga-
tions by conducting kinetic studies on the borylation of 3-
picoline 16 catalyzed by complex 17 (Table 3). We chose 16 as
a model basic heteroarene because of its simple structure and
because the borylation of 16 yields a single product.50 Complex
Scheme 6. Proposed Catalytic Cycle for the Borylation of 3-
Picoline
a
Table 3. Kinetic Data for the Borylation of 3-picoline
entry [B2pin2] (M) 3-picoline (M) [catalyst] (M) initial rate (M·s−1
)
1
2
3
4
5
0.50
0.50
0.50
1.5
0.50
0.50
0.50
0.50
1.5
0.0025
0.0050
0.010
0.010
0.010
1.6 0.1 × 10−6
3.6 0.1 × 10−6
7.6 0.4 × 10−6
7.9 0.3 × 10−6
8.2 0.3 × 10−6
0.50
a
Reactions were conducted on a 0.25 mmol scale. The formation of
product 18 was observed by gas chromatography.
4293
dx.doi.org/10.1021/ja412563e | J. Am. Chem. Soc. 2014, 136, 4287−4299

Journal of the American Chemical Society
Article
To assess the generality of the observed zero-order
dependence in heteroarene, we conducted kinetic studies on
the borylation of benzoxazole 19 (Table 4) by 1H NMR
corresponding to 3-picoline complex 20 and free COE were
observed over the course of the reaction, allowing us to assign
20 as the resting state.
We determined the constant K1 for the equilibrium depicted
in eq 1. However, due to the fast exchange that occurred at RT,
we conducted the H NMR spectroscopic measurements at
Table 4. Kinetic Data for the Borylation of 5-Bromo-2-
methylbenzoxazole 19
a
1
−30 °C, so that the free (δ = 5.60 ppm) and bound (δ = 3.93
ppm) olefinic proton resonances would be independently
discernible. The equilibrium constant K1 was determined to be
1.0
0.6. Under the conditions of our kinetic studies, 3-
picoline is present in a 50−200-fold excess over iridium
complex 17. The equilibrium constant indicates that picoline
adduct 20 will be present in a similar excess over olefin adduct
17 during the catalytic reaction.
Unlike the borylation of 3-picoline, the rate of the borylation
of benzoxazole 19 was found to have a partial order
dependence on the concentration of the heteroarene (vide
supra). To determine if this difference in order was due to
poorer binding of 19 to iridium than of 3-picoline to iridium,
we determined the equilibrium constant for the exchange of
COE with the heteroarene 19. This equilibrium constant was
determined to be 3
equilibrium constant indicates that the resting state of the
catalyst would likely consist mostly of the alkene complex.
a
Reactions were conducted on a 0.25 mmol scale. The formation of
1
product 4b was observed by H NMR spectroscopy.
spectroscopy. Benzoxazole 19 contains more steric bulk around
the basic nitrogen than does 16 (Table 3). In this case,
benzoxazole 19 would bind more weakly to the catalyst than
the parent oxazoles, and the resting state of the catalyst could
be the alkene complex. However, when we measured the initial
rates for the formation of 4b with varying concentrations of 19
we observed a weak dependence of the rate on the
concentration of 19. On the basis of the initial rates, we
approximated the order of the reaction in 19 to be ∼0.4 under
the conditions detailed in Table 4.
1 × 10−3 (Chart 6). This small
Chart 6. Equilibrium Between Heteroarene-Bound and
COE-Bound Forms of the Trisboryl Complex
2.6.2. Relative Binding Affinities of the Heteroarenes
and Cyclooctene. NMR spectroscopic studies were con-
ducted to determine the relative binding constants of the
heteroarenes and COE to iridium, as well as to determine the
resting state of the catalyst. We conducted studies with 3-
picoline 16 and 5-bromo-2-methyl-benzoxazole 19, so that we
could compare the trends in binding of these substrates to
iridium with the kinetic data for the borylation of these
substrates.
We first mixed 3-picoline 16 and COE-ligated iridium
trisboryl complex 17 in THF-d8 at room temperature and
1
analyzed the composition of the iridium complex by H NMR
However, when we monitored by 1H NMR spectroscopy the
borylation of 19 at 40 °C catalyzed by 1 mol % of complex 17
(Table 4), we observed a broad COE resonance at 5.40 ppm
and a broad resonance corresponding to the 6,6′-proton of
dtbpy at 9.82 ppm. The chemical shift of the COE under these
conditions is closer to that of free COE than to that of bound
COE, and the dtbpy resonance is closer to that of the
heteroarene-bound complex (9.98 ppm) than the COE-bound
complex (9.41 ppm), implying that most of the catalyst was
bound by the heteroarene instead of COE.
spectroscopy. Significant amounts of free COE were generated,
as determined by the observation of a broad olefinic proton
resonance for COE at 5.20 ppm. COE bound to this Ir
fragment resonates at 3.93 ppm, and free COE resonates at
5.60 ppm. Thus, the observed signal indicates rapid exchange of
free and bound COE on the NMR time scale, with
approximately 75% of the COE in the free state. The resonance
for the 6- and 6′- protons of bound dtbpy was also shifted
downfield from those of complex 17. We assigned this
resonance to the picoline adduct 20 (eq 1).
From the 1H NMR spectrum we approximated the
equilibrium constant for the displacement of COE by 19 to
be 2 × 10−2 at 40 °C, which is ∼6 times larger than the
equilibrium constant at −30 °C. This apparently greater
preference for the heteroarene-complex at 40 °C than at −30
°C is likely due to the endothermicity of the displacement of
the olefin by 19. Indeed, a van’t Hoff analysis of this process,
revealed that ΔH = +4.7 0.2 kcal/mol and ΔS = +3.8 0.4
eu. With these thermodynamic data, we calculated the
theoretical equilibrium constant at 40 °C to be 2.5 × 10−2
which is consistent with our experimental value. Finally, on the
basis of the observed chemical shift of the dtbpy resonance, the
1
The borylation of 3-picoline was monitored by H NMR
spectroscopy with 2 mol % of 17 as catalyst to determine the
resting state of the catalyst for this process. Sharp resonances
4294
dx.doi.org/10.1021/ja412563e | J. Am. Chem. Soc. 2014, 136, 4287−4299

Journal of the American Chemical Society
Article
ratio of olefin-bound 17 to benzoxazole bound complex is 0.4,
which is roughly consistent with our kinetic data in Table 4.
Chart 6 shows the equilibrium constants for the displace-
ment of COE by substrates 16 and 19 as well as by the
products of borylation 18 and 4b. 3-Picoline 16 binds much
more strongly to iridium than does benzoxazole 19. This
observation is consistent with our initial rate data for the
borylation of these two substrates, since the borylation of 19
(Table 4) is much more rapid than the borylation of 16 (Table
3). Substrate 16 undergoes stronger binding to iridium than the
corresponding product of borylation 18, which shows why the
borylation of 16 occurs without significant product inhibition.
However, some product inhibition is expected for the
borylation of 19, since the product of borylation 4b has
essentially the same ability to bind to iridium as the starting
material.
Figure 5. (a) Reaction of free N−H containing 23 with HBpin. (b)
Reaction of 23 with HBpin catalyzed by 17, which rapidly forms 24.
2.7. Investigations of the Origin of the Regioselectiv-
ity. Several studies were conducted to reveal the origin of the
regioselectivity of the borylation of basic heteroarenes. These
studies included analyzing if the borylation of N−H bonds
would create a large substituent on nitrogen and which
heteroarenes would undergo this process, as well as
competition studies with pyridines to address the question of
whether binding of nitrogen to iridium or boron would be
preventing the borylation at the C−H bonds adjacent to basic
nitrogen atoms. We also concluded DFT calculations of the
reaction pathways for the borylation of pyiridine adjacent to
basic nitrogen.
2.7.1. N−H Borylation. To determine whether the
borylation of N−H bonds affects regioselectivty, we inves-
tigated the origin of the lack of borylation α to free N−H
moieties of basic heteroarenes. The lack of borylation adjacent
to the N−H bond of the heteroarenes in this study was
unexpected because indole undergoes borylation at the 2-
position.40
Figure 6. (a) Reaction of 23 with trisboryl complex 17 to form 24 and
1
Ir−H species 25 as observed by H NMR spectroscopy, followed by
preferential reaction of 25 with B2pin2 to form 17. (b) Preferential
reaction of 23 with HBpin catalyzed by 17 to form 24. Footnote a:
Corrected for the generation of HBpin.
We hypothesized that borylation occurs beta to free N−H
moieties in the heteroarenes in the current study because of the
steric bulk of an N-Bpin species resulting from N−H
borylation. The following observations led to this hypothesis:
(1) more than one equivalent of B2pin2 was required to achieve
high yields of heteroaryl boronate ester, (2) borylation beta to
nitrogen is the same regioselectivity observed for the borylation
of N-Boc-protected heteroarenes, and (3) evolution of gas
(presumed to be H2) was observed upon addition of a mixture
of B2pin2, ligand, and Ir precatalyst to the substrates containing
N−H bonds.
and Maleczka that the N-borylation of 7-azaindole occurs in the
presence of an iridium catalyst and B2pin2.11
To investigate the mechanism of the formation of N-boryl
24, azaindole 23 was mixed with equimolar amounts of
1
trisboryliridium complex 17 (Figure 6a) in THF-d8. The H
NMR spectrum of this reaction contained an iridium hydride
signal (δ −4.72 ppm) as well as complete conversion of 23 to
24. The observed iridium hydride 25 appeared to be identical
to the bisboryl hydride observed previously.21 The formation of
25 suggests an N−H activation/N−B reductive elimination
pathway for the formation of 24.
If N−H borylation proceeds by a mechanism analogous to
that of C−H borylation with B2pin2 as the boron source, one
would expect HBpin to be the byproduct, not H2. However, the
HBpin generated by this pathway could react rapidly with the
free N−H bond without catalysis, generating H2 as a byproduct.
To test whether HBpin reacts with the N−H bonds of the
azoles, 4-chloro-2-methyl-7-azaindole 23 (Figure 5a) was mixed
To determine if the iridium hydride would be recycled to the
trisboryl 17, we investigated the reactivity of iridium hydride 25
with HBpin and B2pin2. Addition of one equivalent of both
HBpin and B2pin2 to the mixture containing 25 resulted in the
formation of 17 and full conversion of B2pin2, suggesting that
25 reacts selectively with B2pin2 over HBpin. However, the N−
H borylation of 23 with one equivalent of both HBpin and
B2pin2 catalyzed by 17 (Figure 6b) results in the quantitative
formation of 24 with only 28% conversion of B2pin2 and 71%
conversion of HBpin, correcting for the generation of HBpin
from the reaction of B2pin2 with 23. These experiments
demonstrate that HBpin must react with 23 via a pathway that
does not involve the N−B reductive elimination to form 25.
On the basis of these results we propose that N−H
borylation under our reaction conditions occurs first by N−H
1
with HBpin in THF. No gas evolution was observed, and H
NMR spectroscopy confirmed that the N−H proton was still
intact. However, when the same experiment was conducted in
the presence of 0.5 mol % of iridium complex 17 (Figure 5b),
vigorous gas evolution was observed, and a new species that
lacked an N−H 1H NMR resonance was formed. Also, the 11B
NMR spectrum of this species contained a new signal at 22.8
ppm. We assigned this new species as N-borylated azaindole 24.
This result is consistent with the recent findings from Smith
4295
dx.doi.org/10.1021/ja412563e | J. Am. Chem. Soc. 2014, 136, 4287−4299

Journal of the American Chemical Society
Article
activation and N−B reductive elimination (red pathway,
The observed regioselectivity could originate from coordi-
nation of pyridine to B2pin2 in either of two ways: (1) the two
species could form a stable adduct quantitatively, or (2) the
transition state for the C−H activation of a transiently formed
adduct could be much lower in energy than the transition state
for C−H activation of free pyridine.
Scheme 7). This pathway produces HBpin, which can then
Scheme 7. Proposed Mechanism for the N−H Borylation of
Basic Heteroarenes
The first proposal was shown to be invalid by measuring the
11B NMR spectrum of a 1:1 mixture of pyridine and B2pin2 in
THF. Only B2pin2 was observed; no upfield shift that would
correspond to a Lewis acid−base N−B adduct was observed.
The second proposal was assessed by conducting competi-
tion studies and by analysis of our kinetic data. We conducted a
competition experiment involving the borylation of one
equivalent of 2,6-lutidine 26 and 3-picoline 16 with one
equivalent of B2pin2 catalyzed by 17 (eq 2). If the transition
state for the C−H activation of a transiently formed N−B
adduct is much lower in energy than that of the free
heteroarene, then one would expect the borylation depicted
in eq 2 to be much more selective for picoline 16 over lutidine
26 because 26 would bind much more weakly to a Lewis acid
than would 16 due to the steric properties of 26. In the event,
the heteroaryl boronate esters 27 and 18 were formed in equal
react with 23 to produce 24 and H2 gas by a different pathway,
perhaps via coordination of 23 to the trisboryl complex,
enhancing the N−H acidity (blue pathway, Scheme 7).
In summary, the mechanism proposed in Scheme 7 accounts
for the rapid N−H borylation of basic heteroarenes and the
formation of H2 gas observed under our conditions. The ability
of basic heteroarenes to coordinate to iridium, which may
enhance the acidity of the N−H bond, explains why N−H
borylation of basic heteroarenes occurs faster than the catalytic
C−H borylation, whereas the N−H borylation of indoles and
pyrroles does not occur on the time scale of C−H borylation.55
We also considered the possibility that a catalytic amount of N-
borylated product formed from the N−B reductive elimination
pathway could act as a Lewis base to activate HBpin, in analogy
to the activation of HBpin by triethylamine reported by Smith
and Maleczka for the N-borylation of pyrrole and indole.11
However, when azaindole 23 was treated with HBpin and 10
mol % of N-borylated 24 in THF, no appreciable reaction was
observed over the course of 30 min. In contrast, the N-
borylation of azaindole 23 in the presence of iridium occurs in
less than 5 min. Therefore, the N-borylation of azaindole 23 in
the absence of iridium is not competitive with the Ir-catalyzed
reaction.
2.7.2. Origin of the Lack of Borylation Ortho to Basic
Nitrogen. 2.7.2.1. Experimental Studies. Finally, we inves-
tigated the origin of the absence of products from the
borylation of C−H bonds located alpha to basic nitrogen
atoms. This regioselectivity was first observed in a study on the
borylation of pyridine.40 Pyridine was found to form products
from borylation at only the 3- and 4-positions in a statistical
ratio without formation of product from borylation at the more
kinetically acidic 2-position. This regioselectivity is general
among the reactions of basic heteroarenes, except for a few
examples involving the borylation of pyrazines and dtbpy
adjacent to nitrogen.13,14 At the time, it was proposed that
precoordination of the basic nitrogen to a Lewis acidic species,
creating steric bulk around nitrogen, would disfavor borylation
at the 2-position. Under the reported conditions for the
borylation of pyridine, there are only two Lewis acidic species
at the start of the reaction: B2pin2 and iridium.
1
amounts, as determined by H NMR spectroscopy: 42% and
41% yield, respectively.
Finally, a reaction of the catalyst with an adduct of the
heteroarene with B2pin2 generated in small quantities by an
unfavorable equilibrium would be first order in B2pin2. This
prediction is inconsistent with the zero-order dependence of
the reaction of 3-picoline on the concentration of B2pin2. Thus,
our experimental data strongly argue against binding of the
heteroarene to the diboron reagent as the origin of the
regioselectivity for the borylation of basic heteroarenes.
Having ruled out the binding of pyridines to B2pin2 during
the reaction mechanism, we considered the potentiality that the
heteroaryl substrate would bind to one equivalent of iridium
while a second equivalent of iridium would react with the C−H
bond of the bound pyridine. Such a mechanism would be
second order in the concentration of catalyst if C−H activation
is turnover limiting. Inconsistent with this pathway, the
borylation of 3-picoline is first order in catalyst (Table 3).
However, if dissociation of heteroarene from a complex
analogous to picoline adduct 20 (eq 1) were rate-limiting, then
the reaction would be first order in catalyst and zero order in
substrate and B2pin2. These are the kinetic orders we observed
for the borylation of 3-picoline (Table 3). Yet rate-limiting
dissociation of heteroarene can be discounted because the
heteroarene and COE undergo exchange on the NMR time
scale (vide supra). These data are inconsistent with pre-
coordination of heteroarene to a Lewis acidic species as the
origin of the observed selectivity.
2.7.2.2. Computational Studies. Having concluded that the
origin of the lack of reactivity adjacent to a basic nitrogen atom
was likely to result from electronic effects, we conducted DFT
calculations of the reaction pathway for the borylation of
pyridine at the 2-, 3-, and 4-positions. We calculated the
pathway for the reaction of pyridine at these positions with the
4296
dx.doi.org/10.1021/ja412563e | J. Am. Chem. Soc. 2014, 136, 4287−4299

Journal of the American Chemical Society
Article
Figure 7. Reaction pathways for the borylation of pyridine at the 2-, 3-, and 4-positions calculated at the M06L level with lanl2dz and 6-31g(d,p)
basis sets and IEFPCM THF solvation model.
16-electron complex generated from thermal dissociation of
COE from 17. The calculations were conducted with M06L
functionals, the lanl2dz basis set for iridium, and the 6-31g(d,p)
basis set for all other atoms. Solvation in THF was modeled
using the IEFPCM formalism. The computed pathways
describing the borylation of pyridine at the 3- and the 4-
postions (red and blue pathways, respectively, Figure 7) predict
turnover-limiting C−H activation. This computational result is
consistent with experimentally determined side-by-side kinetic
isotope effects of 2.9 and 2.4 for reaction at the 3- and the 4-
positions, respectively (see SI for details). However, the
pathways depicted in Figure 7 do not predict the high
Figure 8. (a) Decomposition of 2-borylpyridine 28 during the
borylation of pyridine-d5. (b) Decomposition of 2-borylpyridine 28
during the borylation of benzene-d6.
selectivity for borylation at the 3- and 4-positions over the 2-
position. The highest energy along the pathway for reaction at
the 2-position is only about 1 kcal/mol higher than that along
the pathway for reaction at the 3- or 4-position.
2.7.2.3. Assessment of the Stability of the 2-Borylpyridine
Products under the Reaction Conditions and Conclusion on
the Origin of Regioselectivity for the Borylation of Pyridines
and Diazines. Considering the results of the computational
studies, we considered that the lack of observable products
from borylation alpha to the basic nitrogen atom in pyridines
and diazines might originate from the combination of the
higher barrier of reaction at this site over other sites and the
instability of the products formed from borylation alpha to
basic nitrogen under the reaction conditions.
To test the second part of this hypothesis, we conducted the
borylation of pyridine-d5 in the presence of 5 mol % of
unlabeled 2-borylpyridine 28 and monitored the fate of the 2-
borylpyridine (Figure 8a). Heating of this reaction mixture led
to the decomposition of 2-borylpyridine to multiple unidentifi-
able products. This decomposition was faster than the
borylation of pyridine-d5; only about 5% yield (based on
B2pin2) of the pyridine-d5 borylation products was observed at
the time that the decomposition of the 2-borylpyridine was
complete. Thus, the more rapid decomposition of 2-
borylpyridyl products than the formation of borylpyridine
products is likely the origin of the lack of observation of such
products during the borylation of pyridines.
We also conducted the borylation of benzene-d6 in the
presence of 5 mol % of 2-borylpyridine 28 (Figure 8b). Under
these conditions only a 24% yield of PhBpin was observed at
the time of complete consumption of the 2-borylpyridine.
Thus, even in the presence of benzene-d6, a noncoordinating
substrate containing six reactive C−D bonds, the decom-
position of the 2-borylpyridine was faster than the C−H
borylation process.
We also assessed the mass balance of the reaction of 4-
picoline (29) with B2pin2. Because of the substituent at the 4-
position, 4-picoline should not form significant amounts of the
borylation product. However, it has not been reported whether
pyridines lacking sterically accessible C−H bonds are stable to
the reaction conditions or undergo decomposition, perhaps
through borylation at the 2-position. If 4-picoline undergoes
4297
dx.doi.org/10.1021/ja412563e | J. Am. Chem. Soc. 2014, 136, 4287−4299

Journal of the American Chemical Society
Article
slow reaction at the 2-position, followed by more rapid
decomposition, we should observe (1) conversion of 4-picoline
to a set of decomposition products, (2) conversion of B2pin2 to
HBpin, and (3) catalyst decomposition.
Indeed, heating of 4-picoline (Figure 9) with B2pin2 at 80 °C
in the presence of 10 mol % of complex 17, led to the
(4) No borylation occurred alpha or ortho to free N−H
moieties, due to the rapid formation of a bulky N-Bpin species.
(5) Our studies also showed that borylation did not occur
alpha or ortho to basic nitrogen moieties. This selectivity was
not due to binding of the heteroarene to a Lewis acidic species
through the basic nitrogen, as we proposed earlier, but this
selectivity may be due to the combination of a higher-energy
pathway for the borylation adjacent to basic nitrogen relative to
the borylation at other sites and the instability of the products
from borylation adjacent to basic nitrogen under the reaction
conditions
(6) Heteroarenes were shown to bind to the catalyst, and the
strength of this binding determines the resting state of the
catalyst and ultimately the dependence of the rate of the
reaction on the concentration of substrate.
We expect that our studies will greatly broaden the
applications of iridium-catalyzed borylation in the field of
medicinal chemistry due to our demonstration of the high and
predictable selectivity of the reaction as well as the tolerance of
the reaction to high nitrogen content of the substrates.
ASSOCIATED CONTENT
■
S
* Supporting Information
Experimental procedures and spectra for all new compounds.
Cartesian coordinates and energies of calculated ground states
and transition states. This material is available free of charge via
the Internet at http://pubs.acs.org.
Figure 9. Decomposition of 4-picoline 29, B2pin2, and catalyst as
determined by H NMR spectroscopy.
1
AUTHOR INFORMATION
■
Corresponding Author
jhartwig@berkeley.edu
formation of several unidentifiable aromatic compounds,
according to 1H NMR spectroscopy, a small amount of
heteroaryl boronate 30, the distinctive broad quartet of HBpin,
and complete decomposition of the trisboryl complex after 7 h.
At this time, the conversion of 4-picoline was 40%, but the yield
of 3-boryl-4-picoline 30 was only ∼10%, implying that 4-
picoline was converted to a series of products resulting from
decomposition of the product of borylation at the 2-position.
The conversion of B2pin2 paralleled the conversion of 4-
picoline, which suggests that C−H borylation is leading to the
formation of the multiple aromatic products. At this time we do
not speculate on the mechanism of decomposition of the
proposed 2-borylpyridyl products or the catalyst, but the
instability of 2-borylpyridine derivatives is well docu-
mented.56−58
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
We thank the NSF (CHE-01213409 to J.F.H. and CHE-
0840505 to the Molecular Graphics and Computation Facility
at UC Berkeley) for support of this work. We thank Johnson
Matthey for a gracious donation of [Ir(COD)OMe]2. M.A.L.
thanks Amgen for a predoctoral fellowship and Dr. Carl Liskey,
Dr. Timothy Boller, and Christo Sevov for helpful con-
versations.
REFERENCES
■
(1) Welsch, M. E.; Snyder, S. A.; Stockwell, B. R. Curr. Opin. Chem.
Biol. 2010, 14, 347.
3. CONCLUSIONS
(2) Ge, S.; Hartwig, J. F. Angew. Chem., Int. Ed. 2012, 51, 12837.
(3) Ye, M.; Gao, G.-L.; Edmunds, A. J. F.; Worthington, P. A.;
Morris, J. A.; Yu, J.-Q. J. Am. Chem. Soc. 2011, 133, 19090.
(4) Meng, L.; Kamada, Y.; Muto, K.; Yamaguchi, J.; Itami, K. Angew.
Chem., Int. Ed. 2013, 52, 10048.
(5) Mkhalid, I. A. I.; Barnard, J. H.; Marder, T. B.; Murphy, J. M.;
Hartwig, J. F. Chem. Rev. 2009, 110, 890.
(6) Hartwig, J. F. Acc. Chem. Res. 2011, 45, 864.
(7) Ishiyama, T.; Takagi, J.; Yonekawa, Y.; Hartwig, J. F.; Miyaura, N.
Adv. Synth. Catal. 2003, 345, 1103.
(8) Ishiyama, T.; Nobuta, Y.; Hartwig, J. F.; Miyaura, N. Chem.
Commun. 2003, 0, 2924.
(9) Ishiyama, T.; Miyaura, N. Pure Appl. Chem. 2006, 78, 1369.
(10) Kallepalli, V. A.; Shi, F.; Paul, S.; Onyeozili, E. N.; Maleczka, R.
E.; Smith, M. R. J. Org. Chem. 2009, 74, 9199.
Our studies have demonstrated the potential of the iridium-
catalyzed borylation to functionalize heteroarenes with
predictable selectivities. Several important conclusions about
the origin of the reactivity and selectivity can be drawn from
our data.
(1) Investigations of the borylation of basic heteroarenes
catalyzed by the combination of [Ir(COD)OMe]2 and either
Me4phen or dtbpy revealed that the Me4phen-Ir catalyst is
more active for this transformation than the dtbpy-Ir catalyst.
(2) With the Me4phen-Ir catalyst, the borylation of
benzoxazoles, pyrimidines, and unprotected benzimidazoles,
pyrazoles, and azaindoles occurred in good yield and with high
selectivity.
(3) Much of the high selectivity was derived from the effect
of free N−H and basic nitrogens on selectivity.
(11) Preshlock, S. M.; Plattner, D. L.; Maligres, P. E.; Krska, S. W.;
Maleczka, R. E.; Smith, M. R. Angew. Chem., Int. Ed. 2013, 52, 12915.
4298
dx.doi.org/10.1021/ja412563e | J. Am. Chem. Soc. 2014, 136, 4287−4299

Journal of the American Chemical Society
Article
(12) Klecka, M.; Pohl, R.; Klepetarova, B.; Hocek, M. Org. Biomol.
Chem. 2009, 7, 866.
(13) Mkhalid, I. A. I.; Coventry, D. N.; Albesa-Jove, D.; Batsanov, A.
S.; Howard, J. A. K.; Perutz, R. N.; Marder, T. B. Angew. Chem., Int. Ed.
2006, 45, 489.
(14) Harrisson, P.; Morris, J.; Marder, T. B.; Steel, P. G. Org. Lett.
2009, 11, 3586.
(15) Robbins, D. W.; Hartwig, J. F. Org. Lett. 2012, 14, 4266.
(16) Tzschucke, C. C.; Murphy, J. M.; Hartwig, J. F. Org. Lett. 2007,
9, 761.
Shukla, L.; Morris, J.; Lin, Z.; Marder, T. B.; Steel, P. G. Chem. Sci.
2012, 3, 3505.
(45) Maleczka, R. E.; Shi, F.; Holmes, D.; Smith, M. R. J. Am. Chem.
Soc. 2003, 125, 7792.
(46) Kikuchi, T.; Nobuta, Y.; Umeda, J.; Yamamoto, Y.; Ishiyama, T.;
Miyaura, N. Tetrahedron 2008, 64, 4967.
(47) Murphy, J. M.; Liao, X.; Hartwig, J. F. J. Am. Chem. Soc. 2007,
129, 15434.
(48) Beck, E. M.; Hatley, R.; Gaunt, M. J. Angew. Chem., Int. Ed.
2008, 47, 3004.
(49) Liskey, C. W.; Liao, X.; Hartwig, J. F. J. Am. Chem. Soc. 2010,
132, 11389.
(17) Litvinas, N. D.; Fier, P. S.; Hartwig, J. F. Angew. Chem., Int. Ed.
2012, 51, 536.
(50) Murphy, J. M.; Tzschucke, C. C.; Hartwig, J. F. Org. Lett. 2007,
9, 757.
(51) Kinzel, T.; Zhang, Y.; Buchwald, S. L. J. Am. Chem. Soc. 2010,
132, 14073.
(52) Liu, T.; Shao, X.; Wu, Y.; Shen, Q. Angew. Chem., Int. Ed. 2012,
51, 540.
(18) Webb, K. S.; Levy, D. Tetrahedron Lett. 1995, 36, 5117.
(19) Revuelta, J.; Machetti, F.; Cicchi, S. In Modern Heterocyclic
Chemistry; Wiley-VCH Verlag GmbH & Co. KGaA: Weinheim,
Germany, 2011.
(20) Vanchura, I. I. B. A.; Preshlock, S. M.; Roosen, P. C.; Kallepalli,
V. A.; Staples, R. J.; Maleczka, J. R. E.; Singleton, D. A.; Smith, M. R.
Chem. Commun. 2010, 46, 7724.
(21) Boller, T. M.; Murphy, J. M.; Hapke, M.; Ishiyama, T.; Miyaura,
N.; Hartwig, J. F. J. Am. Chem. Soc. 2005, 127, 14263.
(22) Ertan, T.; Yildiz, I.; Tekiner-Gulbas, B.; Bolelli, K.; Temiz-
Arpaci, O.; Ozkan, S.; Kaynak, F.; Yalcin, I.; Aki, E. Eur. J. Med. Chem.
2009, 44, 501.
(53) Scott, J. S.; Birch, A. M.; Brocklehurst, K. J.; Broo, A.; Brown, H.
̈
S.; Butlin, R. J.; Clarke, D. S.; Davidsson, O.; Ertan, A.; Goldberg, K.;
Groombridge, S. D.; Hudson, J. A.; Laber, D.; Leach, A. G.; MacFaul,
P. A.; McKerrecher, D.; Pickup, A.; Schofield, P.; Svensson, P. H.;
Sorme, P.; Teague, J. J. Med. Chem. 2012, 55, 5361.
̈
(54) Lee, J.; Han, S.-Y.; Jung, H.; Yang, J.; Choi, J.-W.; Chae, C. H.;
Park, C. H.; Choi, S. U.; Lee, K.; Ha, J. D.; Lee, C. O.; Ryu, J. W.; Kim,
H. R.; Koh, J. S.; Cho, S. Y. Bioorg. Med. Chem. Lett. 2012, 22, 4044.
(55) Paul, S.; Chotana, G. A.; Holmes, D.; Reichle, R. C.; Maleczka,
R. E.; Smith, M. R. J. Am. Chem. Soc. 2006, 128, 15552.
(23) Kumar, D.; Jacob, M. R.; Reynolds, M. B.; Kerwin, S. M. Bioorg.
Med. Chem. 2002, 10, 3997.
(24) Potashman, M. H.; Bready, J.; Coxon, A.; DeMelfi, T. M.;
DiPietro, L.; Doerr, N.; Elbaum, D.; Estrada, J.; Gallant, P.; Germain,
J.; Gu, Y.; Harmange, J.-C.; Kaufman, S. A.; Kendall, R.; Kim, J. L.;
Kumar, G. N.; Long, A. M.; Neervannan, S.; Patel, V. F.; Polverino, A.;
Rose, P.; van der Plas, S.; Whittington, D.; Zanon, R.; Zhao, H. J. Med.
Chem. 2007, 50, 4351.
(56) Abraham, M. H.; Grellier, P. L. In The Metal−Carbon Bond
(1985); John Wiley & Sons, Ltd.: New York, 2010; p 25.
(57) Campeau, L.-C.; Fagnou, K. Chem. Soc. Rev. 2007, 36, 1058.
(58) Liu, Y.; Milo, L. J., Jr.; Lai, J. H. ARKIVOC 2013, 135.
(25) Bansal, Y.; Silakari, O. Bioorg. Med. Chem. 2012, 20, 6208.
(26) Adler, T. K.; Albert, A. J. Med. Chem. 1963, 6, 480.
(27) Hong, S.; Kim, J.; Seo, J. H.; Jung, K. H.; Hong, S.-S.; Hong, S. J.
Med. Chem. 2012, 55, 5337.
(28) Vangveravong, S.; Taylor, M.; Xu, J.; Cui, J.; Calvin, W.; Babic,
S.; Luedtke, R. R.; Mach, R. H. Bioorg. Med. Chem. 2010, 18, 5291.
(29) El-Deeb, I. M.; Lee, S. H. Bioorg. Med. Chem. 2010, 18, 3961.
(30) Orjales, A.; Mosquera, R.; Lop
́
ez, B.; Olivera, R.; Labeaga, L.;
Nunez, M. T. Bioorg. Med. Chem. 2008, 16, 2183.
́
̃
(31) Shao, H.; Shi, S.; Huang, S.; Hole, A. J.; Abbas, A. Y.; Baumli, S.;
Liu, X.; Lam, F.; Foley, D. W.; Fischer, P. M.; Noble, M.; Endicott, J.
A.; Pepper, C.; Wang, S. J. Med. Chem. 2013, 56, 640.
(32) Liskey, C. W.; Hartwig, J. F. J. Am. Chem. Soc. 2012, 134, 12422.
(33) Ohmura, T.; Torigoe, T.; Suginome, M. J. Am. Chem. Soc. 2012,
134, 17416.
(34) Cho, S. H.; Hartwig, J. F. J. Am. Chem. Soc. 2013, 135, 8157.
(35) Preshlock, S. M.; Ghaffari, B.; Maligres, P. E.; Krska, S. W.;
Maleczka, R. E.; Smith, M. R. J. Am. Chem. Soc. 2013, 135, 7572.
(36) O’Hara, F.; Blackmond, D. G.; Baran, P. S. J. Am. Chem. Soc.
2013, 135, 12122.
(37) Ichikawa, H.; Nishioka, M.; Arimoto, M.; Usami, Y. Heterocycles
2010, 81, 1509.
(38) Kallepalli, V. A., .Maleczka R. E., Jr., Onyeozili E., Smith M. R.,
III U.S. Patent 7,709,654, 2007.
(39) Schneider, C.; David, E.; Toutov, A. A.; Snieckus, V. Angew.
Chem., Int. Ed. 2012, 51, 2722.
(40) Takagi, J.; Sato, K.; Hartwig, J. F.; Ishiyama, T.; Miyaura, N.
Tetrahedron Lett. 2002, 43, 5649.
(41) For a prior report of this reaction see: Kallepalli, V. A.; Shi, F.;
Paul, S.; Onyeozili, E. N.; Maleczka, R. E.; Smith, M. R. J. Org. Chem.
2009, 74, 9199.
(42) L’Heureux, A.; Thibault, C.; Ruel, R. Tetrahedron Lett. 2004, 45,
2317.
(43) The observed regioselectivity also could originate from the
larger size of the CF3 group relative to the Bpin group.
(44) Tajuddin, H.; Harrisson, P.; Bitterlich, B.; Collings, J. C.; Sim,
N.; Batsanov, A. S.; Cheung, M. S.; Kawamorita, S.; Maxwell, A. C.;
4299
dx.doi.org/10.1021/ja412563e | J. Am. Chem. Soc. 2014, 136, 4287−4299