Boron-catalyzed silylative reduction of quinolines: Selective sp3 C-Si bond formation

Article abstract of DOI:10.1021/ja510674u

A silylative reduction of quinolines to synthetically versatile tetrahydroquinoline molecules involving the formation of a C(sp³)-Si bond exclusively β to nitrogen is described. Triarylborane is a highly efficient catalyst (up to 1000 turnovers), and silanes serve as both a silyl source and a reducing reagent. The present procedure is convenient to perform even on a large scale with excellent stereoselectivity. Mechanistic studies revealed that the formation of a 1,4-addition adduct is rate-limiting while the subsequent C(sp³)-Si bond-forming step from the 1,4-adduct is facile.

Full text of DOI:10.1021/ja510674u

Communication
pubs.acs.org/JACS
3
Boron-Catalyzed Silylative Reduction of Quinolines: Selective sp C−
Si Bond Formation
†
,‡,§
Seewon Joung,†,‡,§ Sung-Woo Park,†,‡ Sehoon Park,*,†,‡
Narasimhulu Gandhamsetty,
,†,‡
and Sukbok Chang*
†
Center for Catalytic Hydrocarbon Functionalizations, Institute for Basic Science (IBS), Daejeon 305-701, South Korea
Department of Chemistry, Korea Advanced Institute of Science and Technology (KAIST), Daejeon 305-701, South Korea
‡
*
S Supporting Information
3
heterocycles accompanied by the generation of a C(sp )−Si
ABSTRACT: A silylative reduction of quinolines to
bond in the reduced products would be plausible via consecutive
hydrosilylations under Lewis acidic catalytic conditions.
Herein we report the development of a boron-catalyzed
silylative reduction of quinolines (Scheme 1) in which the N-
synthetically versatile tetrahydroquinoline molecules in-
3
volving the formation of a C(sp )−Si bond exclusively β to
nitrogen is described. Triarylborane is a highly efficient
catalyst (up to 1000 turnovers), and silanes serve as both a
silyl source and a reducing reagent. The present procedure
is convenient to perform even on a large scale with
excellent stereoselectivity. Mechanistic studies revealed
that the formation of a 1,4-addition adduct is rate-limiting
Scheme 1. Catalytic Silylative Reduction of Quinolines
3
while the subsequent C(sp )−Si bond-forming step from
the 1,4-adduct is facile.
hemical reduction of multiple bonds of readily available
C
chemical feedstocks is one of the most fundamental
1
transformations in organic chemistry. In particular, the
reductive transformation of N-heteroaromatics such as quino-
lines is of special interest since the reduced products
aromatic is converted to a piperidinyl unit with the incorporation
3
of a C(sp )−Si bond exclusively β to the nitrogen atom. The
process is highly diastereoselective in that the pre-existing
substituents control the syn or anti relationship in a predictable
manner. In this procedure, the silane plays a dual role as an
incorporating silyl source as well as a reducing agent.
Considering the fact that the C−Si bond is synthetically
equivalent to a C−OH bond, our procedure creates a catalytic
route to versatile synthetic building units of functionalized
alkaloids. The present catalytic system is highly efficient,
achieving turnover numbers (TONs) of up to ∼1000, and is
workable on a large scale.
(
tetrahydroquinolines) are important building blocks as well as
versatile intermediates in the synthesis of alkaloids, pharmaceut-
icals, and agrochemicals. Although stoichiometric metal hydrides
(
been employed for the reduction of N-heteroaromatics, these
methods suffer from lack of chemo- and regioselectivity, limited
substrate scope, and generation of copious amounts of waste. In
this regard, hydrogen gas is undoubtedly one of the most
straightforward and atom-efficient means to effect reduction,
thereby leading to a number of catalytic hydrogenation
procedures of aromatic N-heterocycles. However, these
processes are typically operated under high pressure (H ) and/
2
including NaBH and LiAlH ) or reactive metals (e.g., Na) have
4 4
3
Inspired by recent reports of metal-catalyzed hydrosilylation of
4
5
b,c
N-aromatic compounds,
we initially chose Et SiH as a
2 2
2
reducing agent for the reductive hydrosilylation of quinoline (1a)
or at elevated temperature, giving rise to exhaustive reduction
products. From a synthetic point of view, such a tendency toward
using 0.01 equiv of B(C F ) as a catalyst in CDCl (eq 1).
6
5 3
3
complete reduction of N-heterocycles with H limits oppor-
2
tunities for the construction of functionalized alkaloids.
On the other hand, hydrosilanes have been used as a
convenient and cheap alternative to H in the metal-mediated
2
reductive conversion of N-heteroaromatics, often leading to
partially reduced products. Recently, tris(pentafluorophenyl)-
1
5
Monitoring of this reaction by H NMR spectroscopy revealed
that it proceeded smoothly to give quantitative conversion within
6 h at 65 °C, affording 1,3-bis-silylated tetrahydroquinoline as a
major product; a minor product that does not bear a C−Si bond
was also detected. The N-silylated species were readily converted
borane [B(C F ) ] and related Lewis acid analogues have been
6
5 3
6
shown to be viable catalysts for the hydrogenation and
hydrosilylation of unsaturated compounds such as imines,
7
olefins, and/or N-heteroaromatics. Being aware of the
advantages of adopting metal-free protocols and also utilizing
the excellent catalytic activity of B(C F ) in the hydrosilylation
Received: October 17, 2014
6
5 3
of imines, we envisaged that the silylative dearomatization of N-
©
XXXX American Chemical Society
A
dx.doi.org/10.1021/ja510674u | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Communication
a
Table 1. Substrate Scope in the Silylative Reduction
a
b
Reactions were conducted on a 0.5 mmol scale (substrate) in CHCl (0.5 mL), and yields are for isolated compounds unless otherwise stated. 8
3
c
d
equiv of Et SiH . Obtained as a single diastereomer. Isolated as the silanol after hydrolytic oxidation using [Ru(p-cymene)Cl ] as a catalyst (ref 8).
2
2
2 2
e
f
5
mol % B(C F ) catalyst. Crude yields of initially formed products and isolated yields of N-protected derivatives are shown in parentheses.
6 5 3
to the corresponding isolable N−H products 1b and 1c in 86%
and 5% yield, respectively, upon silica gel chromatography.
Encouraged by this finding of an unusual silylative reduction of
quinoline, we next investigated the optimization of the reaction
conditions (scope of silanes and effects of solvent) in the reaction
of 1a using 1 mol % B(C F ) at 65 °C [see the Supporting
70%, and 61% yield, respectively. 3-Methylquinoline was
efficiently converted to the corresponding product 8b with the
3
formation of a quaternary C(sp )−Si bond. The reaction of 4-
methylquinoline was highly selective, leading exclusively to the
10
trans product 9b; longer reaction times (24 h) were required,
however. We were pleased to observe that substrates bearing
chloro, bromo, or fluoro groups at C5−C8 were selectively
reduced to their corresponding products (10b−16b) in good to
excellent yields. It was notable that the reductions of 7-chloro-, 8-
chloro-, and 8-bromoquinoline (14a−16a) occurred with high
efficiency, even at room temperature, in short reaction times (10
min for 16b).
6
5 3
Information (SI)]. This revealed that the use of sterically
analogous silanes such as Ph SiH , MePhSiH , and Me PhSiH
2
2
2
2
also resulted in the formation of the silylated product in yields
and selectivities comparable to those obtained from Et SiH .
2
2
Along this line, the reactions with bulkier silanes such as
t
Bu SiH , Et SiH, and Ph SiH became sluggish, leading to 1b in
2
2
3
3
low yields under the same conditions. In addition, the reduction
efficiency was moderate when 1,1,3,3-tetramethyldisiloxane
We next investigated quinolines having multiple substituents.
2-Methylquinolines bearing halides (17a−19a) or an additional
methyl group (20a) were readily converted to the corresponding
products (17b−20b) in moderate to good yields. The relative
stereochemistry between the pre-existing methyl group and the
newly generated silyl group was determined to be cis, which can
be rationalized by an anti addition of Si−H to a double bond by
(
TMDS) was employed to afford 1b in 78%, whereas
polymethylhydrosiloxane (PMHS) did not give 1b. Among
various solvents screened, CHCl was the most effective and thus
3
was chosen for the subsequent studies, although the solvent
effects were not substantial.
7c
With the optimized conditions [1 mol % B(C F ) , 4 equiv of
virtue of the B(C F ) catalyst. A substrates having an aryloxy
6
5 3
6 5 3
9
Et SiH , CHCl , 65 °C, 6−12 h], the reactivity scope of
group at C6 (21a) was converted to the desired product (21b) in
88% yield without unwanted cleavage of the Caryl−O bond, thus
proving the tolerance of those building units. Moreover,
quinolines multisubstituted at various positions were also facile
for the silylative reduction, leading to the desired products 22b−
27b in good yields. We also explored the reactivity of
2
2
3
quinolines was explored (Table 1). Quinolines bearing a methyl
group at C6, C7, or C8 underwent the reaction efficiently to give
the silylated products (2b−4b) in high yields (81−96%). 8-
Isopropylquinoline and 5- and 7-phenylquinoline were also
viable under the optimal conditions, giving 5b, 6b, and 7b in 90%,
B
dx.doi.org/10.1021/ja510674u | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Communication
polyaromatics and isoquinolines. Benzoquinolines were highly
reactive, affording products 28b and 29b in excellent yields. As
expected, both N-aromatic rings of 1,7-phenanthroline were
the SI). The same reaction at higher initial concentration (2.5 M
Et SiH ) displayed faster and quantitative formation of 1b′
2
2
−1
(green squares) in 1 h (v = 0.01044 M min ), and then rapid
i
11
reduced, leading to bis-silylated product 30b in good yield.
consumption of the accumulated 1b′ furnished 1b as the final
Although isoquinoline reacted rather slowly to give a moderate
yield of product 31b under more forcing conditions, the silyl
group was still installed at the position β to nitrogen. On the
other hand, having halide substituents on isoquinolines increased
the reactivity while maintaining high regioselectivity, giving
moderate to good yields of N-sulfonylated products (32b−34b).
In order to gain insights into the reaction pathway, a series of
mechanistic experiments were performed (Scheme 2). When the
product (red triangles) within 5 min (Scheme 2B). In addition,
12
an adduct of 1a with borane, (C H N)B(C F ) (1a′), was
9
7
6 5 3
clearly observed by NMR spectroscopy during the conversion of
1a to 1b′. As a result, this study led us to propose that (i) the final
product 1b is formed via a 1,4-addition adduct 1b′; (ii) the
formation of 1b′ from quinoline is the rate-limiting step; (iii) the
formation of 1b′ is first-order-dependent on the silane
concentration; (iv) the conversion of the 1,4-addition inter-
mediate 1b′ to the final product 1b begins almost at the peak
concentration of 1b′; and (v) a quinoline−borane adduct (1a′) is
a resting species.
Scheme 2. Preliminary Mechanistic Studies
On the basis of the above observations, we propose the
reaction pathway for the B(C F ) -catalyzed silylative reduction
6
5 3
of 1a shown in Scheme 2C. In parallel, an energy profile for this
catalysis was computed by carrying out density functional theory
(
DFT) calculations using Me SiH as a model silane (Scheme
2 2
2
D). The reaction is proposed to work largely via two stages:
formation of the 1,4-addition adduct 1b′ (rate-limiting) and
subsequent hydrosilylation. Initially, B(C F ) rapidly forms a
6
5 3
stable adduct with 1a, leading to 1a′ as a resting species. The 1,4-
adduct formation takes place with the active species, the borane−
silane complex (C F ) B·H SiMe (A), by transfer of its silylium
6
5 3
2
2
cation to 1a to form quinolinium salt B possessing a borohydride
anion, in which the hydride is then delivered to C4 of B to release
5
b,c,6e,13,14
1
,4-adduct 1b′.
Interestingly, rapid buildup of the
1
5
silylating reagent A occurs only after nearly complete
conversion of 1a to 1b′, and thus, completion of the first
reduction is required before the second reduction cycle can be
initiated. Most importantly, following quantitative conversion of
1
a to 1b′, the C2−C3 π electrons (1b′) interact with the
electron-deficient silicon of a boron-coordinated hydride to form
prereaction complex C. An electrophilic silyl group of C is then
transferred selectively to C3, forming a new C(sp )−Si bond to
give intermediate D via transition state TS-I involving a small
barrier of 8.0 kcal mol− (calculated relative to 1b′ + A). The final
product 1b is eventually liberated by hydride attack at C2 of
intermediate D with the regeneration of active species A via
transition state TS-II. The energy barrier for this hydride transfer
16
3
7b
1
−1
to cationic C2 of D was calculated to be 7.8 kcal mol . Overall,
the pathway furnishing the substitution into 1b from 1b′ via two
−1
transition states TS-I and TS-II requires only ∼8.0 kcal mol of
free energy. Such low energy barriers are consistent with the
kinetic behavior of a rapid conversion of 1b′ to 1b (Scheme 2B).
Furthermore, we calculated the Fukui function f − of 1b′
(
shown in the dotted rectangle at the left in Scheme 2C) in an
17
attempt to rationalize the observed β-silylation (C3) selectivity.
reduction of 1a was conducted with Ph SiD , the silylated
This revealed that electron density at C3 of 1b′ is most viable for
2
2
+
product 1b-d was obtained with complete incorporation of
the external electrophile transfer (Me HSi ) relative to those at
2
3
3
deuterium at C2 and C4 but without deuteration at C3 (Scheme
other positions, thereby resulting in a C(sp )−Si bond β to the
2A), indicating that a highly regioselective pathway is operative
nitrogen atom (see the SI).
under the present reduction conditions. Similarly, when
The synthetic utility of the present silylative reduction was also
explored (Scheme 3). The reaction of benzoquinoline (28a) was
performed successfully on a gram scale using 0.1 mol % B(C F )
quinoline-d7 was treated with Et SiH , the proton was
2
2
incorporated only at C2 and C4 (see the SI).
Next, we monitored the reaction progress by H NMR
6
5 3
1
to give a 95% yield of product 28b (TON ≈ 1000), and
18
spectroscopy with two different initial concentrations of silane
subsequent N-benzylation and Tamao oxidation of 28b
(
Scheme 2B). Linear consumption of 1a (blue circles) was
delivered N-benzyl-β-hydroxytetrahydrobenzo[h]quinoline
(36) in high yield (Scheme 3A). Similarly, reactions of various
quinoline derivatives (1a, 8a, 9a, and 10a) worked well on a large
scale, leading to the corresponding products 1b, 8b (quaternary
observed at the lower initial concentration (0.9 M Et SiH ) to
2
2
afford the 1,4-addition adduct 1b′, giving rise to about 50%
−1
conversion in 2 h with an initial rate (v ) of 0.00248 M min (see
i
C
dx.doi.org/10.1021/ja510674u | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Communication
Methods; Wiley: New York, 2008. (d) Denmark, S. E. Organic Reactions,
Vol. 71; Wiley: New York, 2008.
Scheme 3. Synthetic Applications
(
2) (a) Baliah, V.; Jeyaraman, R.; Chandrasekaran, L. Chem. Rev. 1983,
3, 379. (b) Kallstrom, S.; Leino, R. Bioorg. Med. Chem. 2008, 16, 601.
c) Bentley, K. W. Nat. Prod. Rep. 2006, 23, 444. (d) Sridharan, V.;
Suryavanshi, P. A.; Menendez, J. C. Chem. Rev. 2011, 111, 7157.
3) (a) Gribble, G. W. Chem. Soc. Rev. 1998, 27, 395. (b) Pitts, M. R.;
Harrison, J. R.; Moody, C. J. J. Chem. Soc., Perkin Trans. 1 2001, 955.
c) Forrest, T. P.; Dauphinee, D. A.; Deraniyagala, S. A. Can. J. Chem.
985, 63, 412.
4) (a) Bianchini, C.; Barbaro, P.; Macchi, M.; Meli, A.; Vizza, F. Helv.
Chim. Acta 2001, 84, 2895. (b) Glorius, F. Org. Biomol. Chem. 2005, 3,
171. (c) Kuwano, R. Heterocycles 2008, 76, 909. (d) Wang, D.-S.; Chen,
Q.-A.; Lu, S.-M.; Zhou, Y.-G. Chem. Rev. 2012, 112, 2557.
5) (a) Hao, L.; Harrod, J. F.; Lebuis, A.-M.; Mu, Y.; Shu, R.; Samuel,
8
(
̈
̈
́
(
(
1
(
4
(
E.; Woo, H.-G. Angew. Chem., Int. Ed. 1998, 37, 3126. (b) Gutsulyak, D.
V.; van der Est, A.; Nikonov, G. I. Angew. Chem., Int. Ed. 2011, 50, 1384.
(
2
c) Ko
013, 52, 10076.
6) (a) Chase, P. A.; Welch, G. C.; Jurca, T.; Stephan, D. W. Angew.
Chem., Int. Ed. 2007, 46, 8050. (b) Spies, P.; Schwendemann, S.; Lange,
S.; Kehr, G.; Frohlich, R.; Erker, G. Angew. Chem., Int. Ed. 2008, 47,
7543. (c) Eros, G.; Mehdi, H.; Papai, I.; Rokob, T. A.; Kiraly, P.;
Tarkanyi, G.; Soos, T. Angew. Chem., Int. Ed. 2010, 49, 6559. (d) Farrell,
J. M.; Hatnean, J. A.; Stephan, D. W. J. Am. Chem. Soc. 2012, 134, 15728.
e) Geier, S. J.; Chase, P. A.; Stephan, D. W. Chem. Commun. 2010, 46,
̈
nigs, C. D. F.; Klare, H. F. T.; Oestreich, M. Angew. Chem., Int. Ed.
(
C−Si bond), 9b (trans stereochemistry), and 10b (TON ≈
1
000), all in high yields.
To examine a plausible asymmetric induction, the optically
̈
̋
́
́
active quinoline (S)-37a was subjected to the silylative reduction.
Pleasingly, a single diastereomeric product 37b (>99% d.r.) was
found to form with the concomitant generation of two new
stereogenic centers (Scheme 3B), the stereochemistry of which
was determined after conversion of 37b to 39 (crystal structure is
shown) in two steps via 38 (>98% ee).
́
́
́
(
4
2
884. (f) Wang, H.; Fro
008, 5966. (g) Greb, L.; On
̈
hlich, R.; Kehr, G.; Erker, G. Chem. Commun.
a-Burgos, P.; Schirmer, B.; Grimme, S.;
̃
Stephan, D. W.; Paradies, J. Angew. Chem., Int. Ed. 2012, 51, 10164.
(h) Reddy, J. S.; Xu, B.-H.; Mahdi, T.; Frohlich, R.; Kehr, G.; Stephan, D.
̈
In summary, we have developed a silylative reduction of
W.; Erker, G. Organometallics 2012, 31, 5638. (i) Liu, Y.; Du, H. J. Am.
Chem. Soc. 2013, 135, 12968.
3
quinolines catalyzed by B(C F ) in which a new C(sp )−Si
6
5 3
bond is generated exclusively β to nitrogen. The reaction scope is
broad, including quinolines, benzoquinolines, and isoquinolines,
and the stereochemistry of the silylative reduction products was
controlled by the position (C2 and C4) of substituents of the
substrates. The rate-limiting step was found to be the formation
of the initial 1,4-addition adduct, while the subsequent silylation
is rather facile. This procedure is convenient and scalable, giving
high turnover numbers (up to 1000). Asymmetric induction was
also realized (>99% d.r.), offering a new route to functionalized
alkaloids that can be used in synthetic and medicinal chemistry.
(
7) (a) Lambert, J. B.; Zhao, Y.; Wu, H. J. Org. Chem. 1999, 64, 2729.
(b) Blackwell, J. M.; Sonmor, E. R.; Scoccitti, T.; Piers, W. E. Org. Lett.
2000, 2, 3921. (c) Rubin, M.; Schwier, T.; Gevorgyan, V. J. Org. Chem.
2002, 67, 1936. (d) Tan, M.; Zhang, Y. Tetrahedron Lett. 2009, 50, 4912.
(
e) Blackwell, J. M.; Morrion, D. J.; Piers, W. E. Tetrahedron 2002, 58,
247. (f) Bezier, D.; Park, S.; Brookhart, M. Org. Lett. 2013, 15, 496.
g) Simonneau, A.; Oestreich, M. Angew. Chem., Int. Ed. 2013, 52,
1905.
8
́
(
1
(8) Lee, M.; Ko, S.; Chang, S. J. Am. Chem. Soc. 2000, 122, 12011.
(9) Initial mixing of B(C F ) and Et SiH in CHCl followed by the
6
5
3
2
2
3
addition of 1a resulted in a higher yield of 1b (86%) than when the silane
was added last into a solution of B(C F ) and 1a in CHCl (79%) (see
the SI). Thus, we applied the former protocol to the other substrates in
6
5
3
3
ASSOCIATED CONTENT
■
*
S
Supporting Information
this study unless otherwise specified.
(10) The resulting anti (trans) relationship between the two groups at
Procedures and additional data. This material is available free of
charge via the Internet at http://pubs.acs.org.
C3 and C4 in 9b is especially noteworthy in view of the fact that the
B(C F ) -catalyzed hydrosilylation of 1-methyl-1-cyclohexene resulted
6
5 3
in the corresponding cis product (see ref 7c.).
11) The relative stereochemistry of the two newly generated C−Si
bonds was not determined.
12) Geier, S. J.; Gille, A. L.; Gilbert, T. M.; Stephan, D. W. Inorg. Chem.
2009, 48, 10466.
13) Hermeke, J.; Mewald, M.; Oestreich, M. J. Am. Chem. Soc. 2013,
35, 17537.
AUTHOR INFORMATION
■
(
Corresponding Authors
sehoonp@kaist.ac.kr
sbchang@kaist.ac.kr
Author Contributions
N.G. and S.J. contributed equally.
(
(
1
§
(14) Poddubnyi, I. S. Chem. Heterocycl. Compd. 1995, 31, 682.
Notes
(15) (a) Parks, D. J.; Piers, W. E. J. Am. Chem. Soc. 1996, 118, 9440.
The authors declare no competing financial interest.
(
b) Parks, D. J.; Blackwell, J. M.; Piers, W. E. J. Org. Chem. 2000, 65,
090. (c) Houghton, A. Y.; Hurmalainen, J.; Mansikkamaki, A.; Piers, W.
E. Nat. Chem. 2014, 6, 983.
16) (a) Nikonov, G. I.; Vyboishchikov, S. F.; Shirobokov, O. G. J. Am.
Chem. Soc. 2012, 134, 5488. (b) Sakata, K.; Fujimoto, H. J. Org. Chem.
013, 78, 12505.
3
̈
ACKNOWLEDGMENTS
This research was supported by the Institute for Basic Science
IBS-R010-D1) in Korea.
■
(
(
2
(
17) Yang, W.; Mortier, W. J. J. Am. Chem. Soc. 1986, 108, 5708.
REFERENCES
■
(18) Tamao, K.; Ishida, N.; Tanaka, T.; Kumada, M. Organometallics
(
1) (a) Rylander, P. N. Catalytic Hydrogenation in Organic Syntheses;
1
983, 2, 1694.
Academic Press: New York, 1979. (b) De Vries, J. G.; Elsevier, C. J.
Handbook of Homogeneous Hydrogenation; Wiley-VCH: Weinheim,
Germany, 2007. (c) Andersson, P. G.; Munslow, I. J. Modern Reduction
D
dx.doi.org/10.1021/ja510674u | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX