Asymmetric binary acid catalysis: A regioselectivity switch between enantioselective 1,2- and 1,4-addition through different counteranions of InIII

Article abstract of DOI:10.1002/anie.201101254

You can count on the anion: Simply swapping the anions of an indium Lewis acid leads to a remarkable regioselectivity switch between asymmetric 1,2- and 1,4-addition reactions. N-protected indoles and β,γ-unsaturated α-keto esters gave adducts with excell

Full text of DOI:10.1002/anie.201101254

Communications
DOI: 10.1002/anie.201101254
Asymmetric Catalysis
Asymmetric Binary Acid Catalysis: A Regioselectivity Switch between
Enantioselective 1,2- and 1,4-Addition through Different
Counteranions of InIII**
Jian Lv, Long Zhang, Yueming Zhou, Zongxiu Nie, Sanzhong Luo,* and Jin-Pei Cheng
The enantioselective Friedel–Crafts alkylation of indoles with
a,b-unsaturated aldehydes and ketones is one of the most
important approaches for the synthesis of biologically active
indole alkaloids.[1] Hence, a number of asymmetric cata-
lysts[2–4] have been developed for this type of reactions.
Though amenable to both 1,2- and 1,4-addition pathways,
most of these reactions afford 1,4-addition products and there
Scheme 1. Regioselective and enantioselective 1,2- and 1,4-addition
reactions.
are surprisingly few precedents for 1,2-addition in the
reactions.[5] This disparity may stem from the difficulties in
re-adjusting the innate reactivity of enals and enones, and the
instability and tendency of the 1,2-adducts to form bisindole
compounds.[4,6] For example, a recent report indicated that the
1,2-adduct from indole and a b,g-unsaturated a-keto esters
easily reacts with another indole to form bisindole in the
catalysis mediated by chiral N-triflylphosphoramides.[4] Given
the challenges associated with the control of 1,2- and 1,4-
additions in both a regio- and stereoselective manner, it is not
surprising that enantioselective additions of indoles to a,b-
unsaturated aldehydes and ketones with tunable regioselec-
tivity remain unknown. Herein, we present an asymmetric
binary acid catalyst that synergistically combines a chiral
phosphoric acid[7] and an indium halide salt to achieve both
1,2- and 1,4-selective Friedel–Crafts alkylation of indoles with
b,g-unsaturated a-keto esters (Scheme 1). In this catalytic
system, a simple swap of counteranions of indium(III) from
fluoride to bromide or chloride switches the regioselectivity
from 1,2- to 1,4-addition with excellent reactivity and
enantioselectivity in both cases.
between the Brønsted and Lewis acids would lead to mutually
enhanced acidity with concomitant generation of multiacidic
centers for synergistic catalysis. Interestingly, when N-methyl
indole (2a) was employed instead of free indole, the binary
acid 1a/MgF2 catalyzed reactions with b,g-unsaturated a-keto
esters afforded predominantly 1,2-adducts (e.g. 5a) with
exceedingly high regioselectivity (> 30:1) and excellent
enantioselectivity (Table 1, entry 2). This unexpected finding
promoted us to further explore the regioselectivity control in
this type of reaction.
Different Lewis acids, chiral phosphoric acids, and combi-
nations thereof were then examined in the model reactions of
indole 2a and keto ester 3a (see the Supporting Information
for details). Several general trends became evident from these
experiments:
(1) Besides their critical role in stereocontrol, the judicious
selection of chiral phosphoric acid, Lewis acid, and
combinations thereof was also essential to attain high
catalytic activity. This trend is highlighted by the observed
low reactivity or total lack of reactivity when phosphoric
acid (e.g. 1a and 1e; Table 1, entries 1 and 15) or Lewis
acid (e.g. InX3; Table 1, entry 9) was applied individually.
(2) The use of different Lewis acids have a dramatic impact
on the regioselectivity (Table 1, entries 2–8). Most inter-
estingly, in the cases with indium(III) Lewis acids, the
simple swap of counteranions of indium(III) was found to
switch the regioselectivity between 1,2- and 1,4-addition.
For example, InF3 in combination with phosphoric acid
1a is as effective as MgF2/1a in promoting 1,2-addition
with excellent regioselectivity and enantioselectivity
(87% yield, 93% ee; Table 1, entry 3). Meanwhile, the
use of either InBr3, InCl3, or In(OTf)3 gave exclusively
1,4-adduct (1,4-adduct/1,2-adduct > 30:1) with varied
enantioselectivity (Table 1, entries 4–6). The fact that
the 1,4-additions are generally much faster than 1,2-
additions (Table 1, entries 1–3 vs. 4–8) also pinpoints the
difficulties in tuning the innate activity toward 1,2-
In our earlier studies, it was found that a highly efficient
binary acid catalyst could be generated from two individually
inert species, phosphoric acid 1a and MgF2, for the asym-
metric Friedel–Crafts reactions of phenols and indoles.[8,9]
These findings validated our postulate that the complexation
[*] J. Lv, Dr. L. Zhang, Prof. S. Luo, Prof. J.-P. Cheng
Beijing National Laboratory for Molecular Sciences (BNLMS)
CAS Key of Molecular Recognition and Function
Institute of Chemistry, Chinese Academy of Sciences
Beijing 100190 (China)
Fax: (+86)10-6255-4449
E-mail: luosz@iccas.ac.cn
Y. Zhou, Prof. Z. Nie
BNLMS, CAS Key Laboratory of Analytical Chemistry for
Living Biosystems, Beijing 100190 (China)
[**] The project was supported by the Natural Science Foundation
(NSFC; 20972163 and 21025208) and MOST (2011CB808600). J.L.
thanks the China Postdoctoral Science Foundation for support.
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201101254.
6610
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2011, 50, 6610 –6614

Table 1: Screen of different Lewis acid in the 1,4- or 1,2-addition of
N-methyl indole 2a to a-keto ester 3a.[a]
1,4-addition reactions. These combinations gave excellent
regio- and stereoselective control.
Next, the scope of these catalytic systems was explored. In
the presence of 1a/InF3 (10 mol%), a variety of b,g-unsatu-
rated a-keto esters were subjected to the reaction with N-
methyl indole 1a and gave the desired 1,2-addition products
5a–h in good yields and with up to 99% ee (Table 2, entries 1–
8). The reactions accommodated N-alkyl-substituted indoles
including those bearing a methyl or a benzyl substituent,
Table 2: Enantioselective 1,2-addition of N-protected indoles 2 with
various a-keto esters 3.[a]
Entry 1/Lewis acid Yield [%][b] 4a/5a[c] ee of 4a [%][d] ee of 5a [%][d]
1
2
3
4
1a/none
1a/MgF2
1a/InF3
1a/InCl3
1a/InBr3
75
85
87
78
93
1:30
1:30
1:30
31:1
31:1
–
–
–
61
69
90
90
93
–
Entry Product
Entry Product
5
–
6
7
8
9
1a/In(OTf)3 70
1a/Sc(OTf)3 70
42:1 À15
–
–
–
–
45:1
6:1
0
68
–
1a/FeCl3
95
1
7
InF3 or InBr3 trace
–
10
11
12
13
14
15
1b/InBr3
1c/InBr3
1d/InBr3
1e/InBr3
1e/InF3
1e
trace
90
99
98
50
–
–
–
–
–
–
91
88
>30:1
>30:1
>30:1
95
88
97
–
1:30
1:30
2
3
4
5
6
8
28
–
[a] General conditions: 2a (0.10 mmol), 3a (0.12 mmol), 1a (10 mol%),
Lewis acid (10 mol%), and molecular sieves (M.S., 4 ꢀ, 20 mg) at
À708C, for entries 1–8, 14, and 15, 24 h; for entries 9–13, 12 h.
[b] Combined yield of isolated products. [c] Determined by NMR
analysis. [d] Determined by HPLC on a chiral stationary phase. CPA=
chiral phosphoric acid, Py=pyridyl, Tf=trifluoromethanesulfonyl.
9
addition. Furthermore, the regioselectivity switch seems
to be independent of the chiral phosphoric acid
employed, and a similar counteranion effect can be
observed with other phosphoric acids such as 1e (Table 1,
entries 13 and 14). To the best of our knowledge, this
represents a rare example on switchable regioselectivity
by tuning of the counteranion in enantioselective Friedel–
Crafts reactions.
10
11
12
(3) In addition, the binary acid system allows for combina-
torial flexibility in identifying an optimal catalytic system.
Accordingly, significantly improved enantioselectivity for
the 1,4-addition reaction can be achieved by varying the
phosphoric acid component. In this regard, two phospho-
ric acids 1c and 1e, generated by straightforward cation
metathesis of 1a with Li+ or remote substitute tuning
respectively, appeared as the optimal partners with InBr3
and gave > 95% ee in the 1,4-addition reaction (Table 1,
entries 11 and 13).
[a] General conditions: 2a (0.10 mmol), 3a (0.12 mmol), 1a (10 mol%),
InF3 (10 mol%), and M.S. (4 ꢀ, 20 mg) at À708C, for entry 1: 24 h; for
entries 3–6: 48 h; entries 2 and 7–12: 72 h. [b] In the absence of InF3.
Bn=benzyl.
Eventually, InF3 and InBr3, coupled with chiral phospho-
ric acid 1a and its simple derivatives (1c and 1e), were
identified as the optimal catalysts for the respective 1,2- and
Angew. Chem. Int. Ed. 2011, 50, 6610 –6614
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
6611

Communications
Table 3: Enantioselective 1,4-addition of N-protected indoles 2 with
various a-keto esters 3.[a]
however, no activity was observed with N-Boc-protected
indoles (data not shown). a-Keto ester 3b bearing an
isopropyl ester substituent was also an inert substrate, and is
probably due to the steric hindrance of the bulky ester moiety
(Table 2, entry 2). Aliphatic b,g-unsaturated a-keto esters
were also examined in the current catalysis and gave excellent
enantioselectivity but low reactivity (Table 2, entry 11).[10]
The use of an ethyl ester substituent led to only trace
amounts of product (Table 2, entry 12) and this outcome is
consistent with the result shown in Table 2, entry 2. In
addition, the reaction did not work with g-heteroatom-
substituted keto esters such as (E)-methyl-4-ethoxy-2-
oxobut-3-enoate, and is probably due to the electron-donat-
ing ethoxyl substituent (data not shown).
Entry Product
Entry Product
1
7
Overall, the 1,2-addition reactions proceeded cleanly with
only trace amounts of 1,4-adducts being detected in all cases.
In addition, no bisindole products were observed in our
studies and most of the indole starting material was easily
recovered. The use of only 1a in the absence of InF3 generally
gave lower activity and enantisoelectivity (Table 2, entries 3
and 5), thus highlighting again the synergistic combination of
1a and InF3. The absolute configuration of 5c was assigned as
S by X-ray diffraction analysis and other 1,2-adduct can
therefore be assigned by analogy (see the Supporting
Information).[11]
The scope of the 1,4-addition reactions was then examined
with 1c/InBr3 or 1e/InBr3 binary systems (Table 3). The
reaction worked well with a variety of b,g-unsaturated a-keto
esters bearing either g-aromatic or aliphatic substitutents and
N-protected indoles 2. These reactions gave the desired 1,4-
addition products 4a–l in good yields and with excellent
regioselectivity and enantioselectivity (Table 3, entries 1–12).
The reaction did not work with g-heteroatom-substituted keto
esters such as (E)-methyl 4-ethoxy-2-oxobut-3-enoate (data
not shown). In the cases examined, 1c/InBr3 and 1e/InBr3
performed equally well with the latter showing slightly better
enantioselectivity in several cases (Table 3, entries 1, 3, 4, 6,
and 8). Interestingly, the use of 1c/InBr3 enabled a successful
reaction (87% yield and 93% ee; (Table 3, entry 2) even
when 1e/InBr3 failed. The 1c/InBr3 system also showed
improved enantioselectivity over 1e/InBr3 with aliphatic b,g-
unsaturated a-keto esters (Table 3, entries 11 and 12).
Though the definite role of the lithium cation remains
unclear, these results further highlight the combinatorial
power of the binary acid catalysis. The absolute configuration
of 4i was assigned as S by X-ray diffraction analysis and other
1,4-adduct can therefore be assigned by analogy (see the
Supporting Information).[11]
2
3
4
8
9
10
5
6
11
12
[a] General conditions: 2a (0.10 mmol), 3a (0.12 mmol), 1e or 1c
(10 mol%), InBr3 (10 mol%), and M.S. (4 ꢀ, 20 mg) at À708C, for
entries 1–2: 12 h; for entries 3–6, 11 and 12: 24 h; for entries 7–10: 48 h.
Preliminary experiments were conducted to elucidate the
mechanism of the current binary acid catalysis and its
dramatic counteranion effect. An ESI-MS experiment
(anionic mode) of a solution containing 1:1 1e and InBr3
gave three prominent new peaks m/z 1186, 1663, 1936,
corresponding to 1e·InBr3, (1e)2, and (1e)2·InBr2. No other
peaks with a higher m/z ratio were observed, thus suggesting
that phosphoric acid would predominantly form a complex
with indium bromide in 1:1 and 2:1 ratio in the solution phase
(see the Supporting Information). The ESI-MS studies of 1a
and InBr3 gave similar results.[12] A Job plot by fluorescence
titration indicates a 1:1 complexation between 1e and InBr3
under dilute conditions (see the Supporting Information).
These results, together with the observation of a linear
correlation between the catalyst ee value and product
ee value, provide strong evidences for a catalytic active
complex involving a 1:1 ratio of phosphoric acid and indi-
um(III). In addition, the dramatic counteranion effect on
6612
www.angewandte.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2011, 50, 6610 –6614

Zhou, Y. Tang, J. Am. Chem. Soc. 2002, 124, 9030 – 9033; c) K. A.
Jørgensen, Synthesis 2003, 1117 – 1125; d) M. Bandini, M.
Fagioli, M. Garavelli, A. Melloni, V. Trigari, A. Umani-
Ronchi, J. Org. Chem. 2004, 69, 7511 – 7518; e) C. Palomo, M.
Oiarbide, B. G. Kardak, J. M. Garcia, A. Linden, J. Am. Chem.
Soc. 2005, 127, 4154 – 4155; f) Y. X. Jia, S. F. Zhu, Y. Yang, Q. L.
Zhou, J. Org. Chem. 2006, 71, 75 – 80; g) S. Yamazaki, Y. Iwata, J.
Org. Chem. 2006, 71, 739 – 743; h) R. Rasappan, M. Hager, A.
Gissibl, O. Reiser, Org. Lett. 2006, 8, 6099 – 6102; scandium
catalyzed: i) D. A. Evans, K. A. Scheidt, K. R. Fandrick, H. W.
Lam, J. Wu, J. Am. Chem. Soc. 2003, 125, 10780 – 10781; j) D. A.
Evans, K. R. Fandrick, H.-J. Song, J. Am. Chem. Soc. 2005, 127,
8942 – 8943; k) D. A. Evans, K. R. Fandrick, H.-J. Song, K. A.
Scheidt, R. Xu, J. Am. Chem. Soc. 2007, 129, 10029 – 10041; l) Y.
Liu, D. Shang, X. Zhou, Y. Zhu, L. Lin, X. Liu, X. Feng, Org.
Lett. 2010, 12, 180 – 183; iron catalyzed: m) L. Yang, Q. Zhu, S.
Guo, B. Qian, C. Xia, H. Huang, Chem. Eur. J. 2010, 16, 1638 –
1645.
regio- and stereoselectivity as well as the observation that
In(OTf)3 is almost nonselective compared with indium halide
(Table 1, entry 6) strongly suggest a unique catalytically
active complex should be invoked in the transition state in
which the halides remain as an integrate component rather
than simply as balancing anions. This is a scenario frequently
encountered but largely overlooked in the typical asymmetric
Lewis acid catalysis.[13]
Accordingly, the halides may serve as ligands that fine
tune the indium(III) metal center in our catalytic systems to
afford specific electronic and steric properties.[14–16] The high
level of control of both regio- and stereoselectivity could be
presumably understood as a result of a combination of innate
electronic tuning[15] and steric effects[14,16] upon bidendate
activation of the a-keto ester substrate with the indium
complex. In both aspects, halide anions would be the delicate,
but critical adjustor of the reaction outcome. Unfortunately,
the structures of the active catalytic complex remain unde-
termined so far, and thus limits clear-cut assignments of the
functioning modes of halides in this complicated catalytic
system (see the Supporting Information for tentative tran-
sition states and discussions). Further mechanistic and
structural studies are therefore highly warranted.
In summary, a regioselectivity switch and highly enantio-
selective Friedel–Crafts reaction of N-protected indoles with
b,g-unsaturated a-keto esters has been realized with the aid of
an asymmetric binary acid catalyst that synergistically com-
bines chiral phosphoric acid and indium halide. The essential
roles of the indium halides (InF3 and InBr3) lies in their
capacity to significantly enhance both the reactivity and
stereoselectivity in 1,2- and 1,4-addition as well as the
unexpected finding that a simple swap of the counteranions
of indium from fluoride to bromide or chloride leads to a
regioselective shift between 1,2- and 1,4-additions with
excellent regioselectivity and enantioselectivity in both
cases. These results underline the potential importance of
counteranions in tuning both chemoselectivity and stereose-
lectivity. Detailed mechanistic studies are ongoing and will be
reported in due course.
[3] For leading examples on immnium-based catalysis of asymmet-
ric Friedel–Crafts alkylations, see: a) N. A. Paras, D. W. C.
MacMillan, J. Am. Chem. Soc. 2001, 123, 4370 – 4371; b) J. F.
Austin, D. W. C. MacMillan, J. Am. Chem. Soc. 2002, 124, 1172 –
1173; c) N. A. Paras, D. W. C. MacMillan, J. Am. Chem. Soc.
2002, 124, 7894 – 7895; d) N. Halland, T. Hansen, K. A. Jørgen-
sen, Angew. Chem. 2003, 115, 5105 – 5107; Angew. Chem. Int. Ed.
2003, 42, 4955 – 4957; e) G. Bartoli, M. Bosco, A. Carlone, F.
Pesciaioli, L. Sambri, P. Melchiorre, Org. Lett. 2007, 9, 1403 –
1405; f) L. Hong, L. Wang, C. Chen, B. Zhang, R. Wang, Adv.
Synth. Catal. 2009, 351, 772 – 778.
[4] For a leading example of asymmetric Friedel–Crafts alkylation
catalyzed by chiral phosphoric acid, see: M. Rueping, B. J.
Nachtsheim, S. A. Moreth, M. Bolte, Angew. Chem. 2008, 120,
603 – 606; Angew. Chem. Int. Ed. 2008, 47, 593 – 596.
[5] For examples of asymmetric 1,2-addition of a,b-unsaturated
carbonyl compounds: aldol reaction: a) C. Zheng, Y. Wu, X.
Wang, G. Zhao, Adv. Synth. Catal. 2008, 350, 2690 – 2694; b) P.
Li, J. Zhao, F. Li, A. S. C. Chan, F. Y. Kwong, Org. Lett. 2010, 12,
5616 – 5619; Henry reaction: c) C. Christensen, K. Juhl, R. G.
Hazell, K. A. Jørgensen, J. Org. Chem. 2002, 67, 4875 – 4881;
rhodium-catalyzed addition of arylboronic acids to a-keto
esters: d) H.-F. Duan, J.-H. Xie, X.-C. Qiao, L.-X. Wang, Q.-L.
Zhou, Angew. Chem. 2008, 120, 4423 – 4425; Angew. Chem. Int.
Ed. 2008, 47, 4351 – 4353.
[6] H. Li, Y.-Q. Wang, L. Deng, Org. Lett. 2006, 8, 4063 – 4065.
[7] For pioneer contributions on chiral phosphoric acid catalysis,
see: a) M. Terada, Chem. Commun. 2008, 4097 – 4112; b) T.
Akiyama, Chem. Rev. 2007, 107, 5744 – 5758.
[8] J. Lv, X. Li, L. Zhang, S. Luo, J.-P. Cheng, Org. Lett. 2010, 12,
1096 – 1099.
Received: February 19, 2011
Revised: April 4, 2011
Published online: June 6, 2011
Keywords: asymmetric catalysis · counterion effect · Friedel–
Crafts alkylation · organocatalysis · switchable regioselectivity
[9] For a review on combined phosphoric acid and metal catalysis,
see: M. Rueping, R. M. Koenigs, I. Atodiresei, Chem. Eur. J.
2010, 16, 9350 – 9365.
[10] The exploration of aliphatic b,g-unsaturated a-keto esters has
been limited by the difficulities encountered in synthesizing
these substrates. For a recent example, see: D. Belmessieri, L. C.
Morrill, C. Simal, A. M. Z. Slawin, A. D. Smith, J. Am. Chem.
Soc. 2011, 133, 2714 – 2720; for the only available synthetic
method, see: H. Sugimura, Y. Kazuhiro, Bull. Chem. Soc. Jpn.
1992, 65, 3209 – 3211.
.
[1] For reviews on asymmetric Friedel–Crafts reactions, see: a) S.-L.
You, Q. Cai, M. Zeng, Chem. Soc. Rev. 2009, 38, 2190 – 2201;
b) T. B. Poulsen, K. A. Jørgensen, Chem. Rev. 2008, 108, 2903 –
2915; c) M. Bandini, A. Melloni, A. Umani-Ronchi, Angew.
Chem. 2004, 116, 560 – 566; Angew. Chem. Int. Ed. 2004, 43, 550 –
556; d) M. Bandini, A. Eichholzer, Angew. Chem. 2009, 121,
9786 – 9824; Angew. Chem. Int. Ed. 2009, 48, 9608 – 9644; e) V.
Terrasson, R. M. de Figueiredo, J. M. Campagne, Eur. J. Org.
Chem. 2010, 2635 – 2655.
[2] For leading examples on metal-catalyzed asymmetric Friedel–
Crafts reactions, see: copper catalyzed: a) K. B. Jensen, J.
Thorhauge, R. G. Hazell, K. A. Jørgensen, Angew. Chem. 2001,
113, 164 – 167; Angew. Chem. Int. Ed. 2001, 40, 160 – 163; b) J.
[11] CCDC 810832 (5c) and 813841 (4i) contain the supplementary
crystallographic data for this paper. These data can be obtained
free of charge from The Cambridge Crystallographic Data
Centre via www.ccdc.cam.ac.uk/data_request/cif.
[12] ESI-MS and 19F NMR analyses also indicate effective complex-
ation of InF3 with 1a or 1e in 1:1 or 1:2 ratio, see the Supporting
Information for details.
Angew. Chem. Int. Ed. 2011, 50, 6610 –6614
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
6613

Communications
[13] a) Acid Catalysis in Modern Organic Sythesis (Ed.: H. Yama-
keto ester 3a) stretching frequencies, with Dn = 18 cmÀ1 in the
presence 1a/InBr3 and Dn = 5 cmÀ1 in the case of 1a/InF3. This
trend is consistent with the observed regioselective reaction at
moto, K. Ishihara), Wiley-VCH, Weinheim, 2008; b) H. Yama-
moto, K. Futatsugi, Angew. Chem. 2005, 117, 1958 – 1977;
Angew. Chem. Int. Ed. 2005, 44, 1924 – 1942.
=
the C C bond with 1a/InBr3 (see the Supporting Information for
[14] B. L. Pagenkopf, E. M. Carreira, Chem. Eur. J. 1999, 5, 3437 –
3442.
[15] FT-IR characterizations of a-keto ester 3a treated separately
the IR spectra).
[16] S. E. Denmark, G. L. Beutner, T. Wynn, M. D. Eastgate, J. Am.
Chem. Soc. 2005, 127, 3774 – 3789.
=
with 1a/InF3 and 1a/InBr3 exhibit marked difference in C C (of
6614 www.angewandte.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2011, 50, 6610 –6614