Chiral Lewis Acid Catalysis in Nitrile Oxide Cycloadditions

Article abstract of DOI:10.1021/ja0318636

We describe examples of highly regio- and enantioselective nitrile oxide cycloadditions to unsaturated alkenes using substoichiometric amounts of a chiral Lewis acid. Pyrazolidinones proved to be effective achiral templates in the cycloadditions providing

Full text of DOI:10.1021/ja0318636

Published on Web 04/13/2004
Chiral Lewis Acid Catalysis in Nitrile Oxide Cycloadditions
Mukund P. Sibi,* Kennosuke Itoh, and Craig P. Jasperse
Department of Chemistry, North Dakota State UniVersity, Fargo, North Dakota 58105
Received December 20, 2003; E-mail: Mukund.Sibi@ndsu.nodak.edu
Scheme 1
Nitrile oxide cycloaddition to olefins is an important synthetic
transformation.1,2 Diastereoselective nitrile oxide cycloadditions
have been investigated extensively, and successful methods are at
hand.3 However, the development of enantioselective variants using
chiral Lewis acids has been hampered by the use of coordinative
amine bases for the generation of the nitrile oxide, by the high
donor ability of the oxygen atom of the dipole,4 and by the
propensity of the dipole to dimerize.2 Ukaji and Inomata have
reported examples of enantioselective nitrile oxide cycloadditions
with allylic alcohols using chiral catalysts.5 In this contribution we
describe examples of highly regio- and enantioselective nitrile oxide
cycloadditions to electron-deficient alkenes using substoichiometric
amounts of a chiral Lewis acid.
Table 1. Evaluation of Templates in Nitrile Oxide Cycloadditionsa,b
We began our investigation with the goal of identifying achiral
templates “Z” that would provide optimal regioselectivity in nitrile
oxide cycloadditions to crotonates, since regioselectivity is often
difficult to control (Scheme 1).6 We reasoned that regioselectivity
might result if the pathway leading to the O-adduct 4 could be
suppressed by steric interactions of the R1 group with either bulky
templates and/or bulky Lewis acids. During template screening
(Table 1), we used a chiral Lewis acid prepared from magnesium
iodide and a bisoxazoline derived from amino indanol (6a) (30 mol
% catalytic loading). Preformed mesityl nitrile oxide (7a), a stable
dipole, was chosen as the reagent since this would obviate the use
of an amine base for its generation. To aid us in the assessment of
regio- and stereoselectivity, the product isoxazoline amides were
reduced to alcohols 10 and 11,7 such that the same products would
result regardless of initial template. Reaction with the oxazolidinone
crotonates 5a-5c proceeded with low regioselectivity and varying
enantioselectivity (entries 1-3). Use of the 3,5-dimethylpyrazole
template (5d) reversed the regioselectivity (entry 4). We have
recently reported on a novel class of achiral pyrazolidinone
templates, which contain a fluxional nitrogen atom and which have
proven to be very effective in several enantioselective transforma-
tions.8 Three such templates (5e-5g) differing in the size of the
N1 substituent were investigated (entries 5-7). Reactions using
each of the pyrazolidinone templates were both highly regio- and
enantioselective providing the C-adduct exclusively. The size of
the R group had minimal impact. These results demonstrate that
chiral Lewis acid-mediated nitrile oxide cycloadditions proceed with
high enantioselectiVity and high regioselectiVity.
yield,
%
10 (ee),
%
11 (ee),
%
c
e
e
ent
subs.
time, h
10:11 ratiod
1
2
3
4
5
6
7
5a
5b
5c
5d
5e
5f
67
41
69
61
88
84
98
24
24
24
96
12
2.5
24
2.2:1 (8a:9a)
1.6:1 (8b:9b)
3:1 (8c:9c)
1:5 (8d:9d)
99:1 (8e:9e)
99:1 (8f:9f)
99:1 (8g:9f)
82
28
86
37
95
99
99
14
00
08
03
-
-
-
5g
a For details of the reaction conditions see Supporting Information. b 30
mol % Lewis acid. c Isolated yield. d Regioisomer ratio determined by H
1
NMR (500 MHz). e Chiral HPLC.
high selectivity. The combination of ligand 6d and Cu(OTf)2, which
has provided excellent enantioselectivity for various transformations,
proved to be ineffective (entry 13). Lowering the catalytic loading
of 6a and MgI2 from 30 to 10 to 5 mol % reduced both
regioselectivity and enantioselectivity (compare entry 1 with entries
14 and 15).
We have carried out a short breadth and scope study by both
varying the enoyl portion of the substrate and varying the nitrile
oxide (Table 3). The N-benzyl pyrazolidinone template and the
chiral Lewis acid derived from MgI2 and 6a were held constant.
Four substrates were examined using 7a as the dipole (entries 1-4).
All of these reactions were highly efficient, providing the products
in good yields and high regio- and enantioselectivity. Even the less
reactive cinnamate acceptor 5i gave excellent results (entry 3). The
nitrile oxide was also varied (entries 1, 5-10). To avoid potential
problems involving coordination of the Lewis acid by amine bases,
we have devised a novel method for the generation of unstable
nitrile oxides from hydroximinoyl chlorides using Amberlyst 21
as the base.9 Cycloaddition with several aryl nitrile oxides gave
The effect of the chiral Lewis acid on the nitrile oxide
cycloadditions using template 5f was evaluated next (Table 2). A
remarkable impact of ligand is shown in entries 1-5. Reactions
using MgI2 and ligands 6b-6e in place of 6a gave adducts with
poor C/O selectivity and negligible enantioselectivity. The nature
of the magnesium counterion had limited influence, with magnesium
perchlorate and triflimide giving results almost as good as with
MgI2 (entries 1, 6-8). Nickel salts (entries 9 and 10) also gave the
C-adduct in high enantioselectivity, although with lower C/O
regioselectivity as compared to MgI2. These results show that
several chiral Lewis acids are effective in providing C-adducts with
9
5366
J. AM. CHEM. SOC. 2004, 126, 5366-5367
10.1021/ja0318636 CCC: $27.50 © 2004 American Chemical Society

C O M M U N I C A T I O N S
Table 2. Evaluation of Chiral Lewis Acids
Figure 1. Stereochemical model
yield,
%
8 ee,
%
9 ee,
%
a
c
c
ent
LA
MgI2
MgI2
MgI2
MgI2
MgI2
Mg(ClO4)2
Mg(NTf2)2
Mg(SbF6)2
Ni(ClO4)2
Ni(SbF6)2
Zn(OTf)2
Cu(OTf)2
Cu(OTf)2
MgI2
ligand
8:9b
6b-6e are apparently too small. While MgI2/6a provides good
enantioselectivity even with oxazolidinone templates 5a and 5c,
high regioselectivity requires the bulky pyrazolidinone templates
5e-5g. We believe these templates may clutter the rear quadrant
above the alkene such that the carbon end of the dipole prefers
approach from the front quadrant for steric reasons. In contrast to
some other reactions,8 the size of the R group on the pyrazolidiones
5e-5g had little observable influence on enantioselectivity.
1
2
3
4
5
6
7
8
6a
6b
6c
6d
6e
6a
6a
6a
6a
6a
6a
6a
6d
6a
6a
84
88
89
79
88
89
83
92
74
88
76
67
34
90
85
99:1
2:1
4:1
99
00
19
12
40
98
97
72
92
96
11
00
-07
97
16
-
00
04
17
14
-
33
57
20
-
04
29
27
59
02
7:1
11:1
32:1
21:1
17:1
15:1
10:1
3:1
3:1
2:1
13:1
4:1
9
10
11
12
13
14d
15e
Acknowledgment. This work was supported by NSF-CHE-
0316203 and NSF-EPS-0132289.
Supporting Information Available: Characterization data for
compounds 5-16 and experimental procedures. This material is
available free of charge via the Internet at http://pubs.acs.org.
MgI2
a Isolated yield. b Regioisomer ratio determined by 1H NMR (500 MHz).
c Chiral HPLC. d 10 mol % Lewis acid. e 5 mol % Lewis acid.
References
(1) (a) Karlsson, S.; Hogberg, H.-E. Org. Prep. Proced. Int. 2001, 33, 103.
(b) Gothelf, K. V.; Jørgensen, K. A. Chem. Commun. 2000, 1449. (c)
Kanemasa, S. Synlett 2002, 1371. (d) Ukaji, Y.; Inomata, K. Synlett 2003,
1075. (d) Cycloaddition Reactions in Organic Synthesis; Kobayashi, S.,
Jørgensen, K. A., Eds.; Wiley-VCH: Weinheim, 2002.
Table 3. Reactions with Various Dipolarophiles and Nitrile Oxides
(2) (a) Torsell, K. B. G. Nitrile Oxides, Nitrones, and Nitronates in Organic
Synthesis; VCH: Weinheim, 1988. (b) Synthetic Applications of 1,3-
Dipolar Cycloaddition Chemistry Toward Heterocycles and Natural
Products; Padwa, A., Pearson, W. H., Eds.; Wiley: Chichester, 2002.
(3) Gothelf, K. V.; Jørgensen, K. A. Chem. ReV. 1998, 98, 863.
(4) Coordination of metal to nitrile oxide: (a) Kanemasa, S.; Nishiuchi, M.;
Kamimura, A.; Hori, K. J. Am. Chem. Soc. 1994, 116, 2324. (b)
Yamamoto, Y.; Watanabe, S.; Kadotani, K.; Hasegawa, M.; Noguchi, M.;
Kanemasa, S. Tetrahedron Lett. 2000, 41, 3131. (c) Rasmussen, B. S.;
Elezcano, U.; Skrydstrup, T. J. Chem. Soc., Perkin Trans. 1 2002, 1723.
(d) Faita, G.; Paio, A.; Quadrelli, P.; Rancati, F.; Seneci, P. Tetrahedron
2001, 57, 8313.
(5) (a) Ukaji, Y.; Sada, K.; Inomata, K. Chem. Lett. 1993, 1847. (b) Shimizu,
M.; Ukaji, Y.; Inomata, K. Chem. Lett. 1996, 455. (c) Tsuji, M.; Ukaji,
Y.; Inomata, K. Chem. Lett. 2002, 1112. For antibody catalysis, see:
Toker, J. D.; Wentworth, P.; Hu, Y.; Houk, K. N.; Janda, K. D. J. Am.
Chem. Soc. 2000, 122, 3244.
yld,
%
8 ee,
%
a
c
ent
sub. R
nitrile oxide R
prod
8:9b
1
1
2
3
4
5
6
7
8
9
R ) Me 5f
R ) Et 5h
R ) Ph 5i
7a
7a
7a
8f, 9f
8h, 9h
8i, 9i
8j, 9j
8k, 9k
8l, 9l
84 99:1
86 99:1
85 99:1
75 99:1
75 99:1
78 99:1
99
99
99
99
99
86
96
99
92
79
R ) CO2Et 5j 7a
R ) Me 5f
R ) Me 5f
R ) Me 5f
R ) Me 5f
R ) Me 5f
R1 ) Ph 7b
R1 ) 2-Cl-Ph 7c
R1 ) 4-Cl-Ph 7d
8m, 9m 70 99:1
R1 ) 4-MeOPh 7e 8n, 9n
61 10:1
44 99:1
63 33:1
(6) Regioselectivity in nitrile oxide cycloadditions: (a) Toma, L.; Quadrelli,
P.; Perrini, G.; Gandolfi, R.; Valentin, C. D.; Corsaro, A.; Caramella, P.
Tetrahedron 2000, 56, 4299. (b) Caramella, P.; Reami, D.; Falzoni, M.;
Quadrelli, P. Tetrahedron 1999, 55, 7027. (c) Weidner-Wells, M. A.;
Fraga-Spano, S. A.; Turchi, I. J. J. Org. Chem. 1998, 63, 6319. (d)
Wallace, R. H.; Liu, J.; Zong, K. K.; Eddings, A. Tetrahedron Lett. 1997,
38, 6791.
R1 ) t-Bu 7f
R1 ) i-Bu 7g
8o, 9o
8p, 9p
10 R ) Me 5f
a Isolated yield. b Regioisomer ratio determined by 1H NMR (500 MHz).
c Chiral HPLC.
(7) See Supporting Information for experimental details.
(8) Achiral templates with fluxional groups in synthesis see: (a) Sibi, M. P.;
Venkatraman, L.; Liu, M.; Jasperse, C. P. J. Am. Chem. Soc. 2001, 123,
8444. (b) Corminboeuf, O.; Quaranta, L.; Renaud, P.; Liu, M.; Jasperse,
C. P.; Sibi, M. P. Chem. Eur. J. 2003, 9, 28. (c) Sibi, M. P.; Liu, M. Org.
Lett. 2001, 3, 4181. (d) Sibi, M. P.; Ma, Z.; Jasperse, C. P. J. Am. Chem.
Soc. 2004, 126, 718.
the C-adduct preferentially in high enantioselectivity and good
yields (entries 5-8). Aliphatic nitrile oxides also provided the
C-adducts with good selectivity, although the reactions were slower
and proceeded in lower yields (entries 9 and 10).
(9) The nitrile oxide can be generated using in situ Amberlyst 21, or by passing
the hydroximinoyl chlorides through an external bed of Amberlyst 21
immediately prior to injection into the reaction mixture. See Supporting
Information for details of the reaction setup.
(10) Yoshida, Y.; Ukaji, Y.; Fujinami, S.; Inomata, K. Chem. Lett. 1998, 1023.
(11) Competing background reaction does explain entry 15 of Table 2 (5 mol
% of MgI2 as catalyst). Copper triflate also appears to be ineffective as a
catalyst (Table 2, entries 12 and 13).
(12) MgX2/6a gives the same sense of facial selectivity in other conjugate
additions (amine, radical) with various templates (imide, pyrazolidione,
oxazolidinone), see: (a) Sibi, M. P.; Prabagaran, N.; Ghorpade, S. G.;
Jasperse, C. P. J. Am. Chem. Soc. 2003, 125, 11796. (b) Reference 8c.
(c) Sibi, M. P.; Chen, J. J. Am. Chem. Soc. 2001, 123, 9472.
The absolute stereochemistry of adduct 8k was determined to
be S,S by converting it to a known compound.10 In general, control
reactions in the absence of Lewis acid were slower than Lewis-
acid-catalyzed reactions.11 Thus, the superior results using the
combination of bulky ligand 6a and bulky templates 5e-5g reflect
shielding effects rather than superior rate acceleration. A tentative
model for the cycloaddition is shown in Figure 1. In our model, a
five- or six-coordinate magnesium is bound to the ligand and to
the bidentate substrate in an s-cis conformation. Shielding by the
ligand blocks the bottom face of the alkene.12 The high enantiose-
lectivity with templates 5e-5g requires the bulky ligand 6a; ligands
JA0318636
9
J. AM. CHEM. SOC. VOL. 126, NO. 17, 2004 5367