Enantioselective cycloadditions with α,β-disubstituted acrylimides

Article abstract of DOI:10.1021/ol050599d

(Chemical Equation Presented) The use of N-H imide templates provides a solution to the problem of rotamer control in Lewis acid catalyzed reactions of α,β-disubstituted acryloyl imides. Reactions proceed through the s-cis rotamer and with improved reactivity because A^1,3 strain is avoided. Enantioselective nitrone, nitrile oxide, and Diels-Alder cycloadditions demonstrate the principle.

Full text of DOI:10.1021/ol050599d

ORGANIC
LETTERS
2005
Vol. 7, No. 12
2349-2352
Enantioselective Cycloadditions with
-Disubstituted Acrylimides
r,â
Mukund P. Sibi,* Zhihua Ma, Kennosuke Itoh, Narayanasamy Prabagaran, and
Craig P. Jasperse
Department of Chemistry, North Dakota State UniVersity,
Fargo, North Dakota 58105-5516
mukund.sibi@ndsu.edu
Received March 18, 2005
ABSTRACT
The use of N-H imide templates provides a solution to the problem of rotamer control in Lewis acid catalyzed reactions of
acryloyl imides. Reactions proceed through the s-cis rotamer and with improved reactivity because A1,3 strain is avoided. Enantioselective
nitrone, nitrile oxide, and Diels Alder cycloadditions demonstrate the principle.
r,â-disubstituted
−
Rotamer control is a central requirement for enantioselective
addition reactions involving R,â-unsaturated carbonyl com-
pounds.1 Many successful reactions using chiral Lewis acid
catalysis have been developed for situations when the
R-carbon of the substrate is unsubstituted (e.g., crotonate 1).
Traditional templates such as oxazolidinone are well-suited
for crotonates because they provide organized chelation such
that substrates 1 react exclusively via the s-cis rotamer.2
However, no general solution for R,â-disubstituted substrates
such as tiglate 2 has been developed.3,4 In substrate 2, the
use of traditional auxiliaries results in problematic A1,3
interactions for both the s-cis and the s-trans rotamers. To
relieve strain, the C-C bond of the enoyl group twists. This
twisting not only leads to ill-defined conformations unsuited
for enantioselective reactions but also breaks conjugation,
contributing to greatly diminished reactivity at the â-carbon.
Thus although A1,3 interactions may be desirable with
R-unsubstituted substrates 1, they are undesirable when
R-substituted substrates 2 are used.
We reasoned that the simple use of N-H imide templates5
(3) would relieve A1,3 strain, reduce twisting, restore
conjugation, and improve reactivity. However, whether there
would be useful levels of s-cis/s-trans rotamer control in the
(1) For a general discussion, see: Principles of Asymmetric Synthesis;
Gawley, R. E.; Aube, J. Pergamon: Oxford, 1996. Also see: (a) Nishimura,
K.; Tomioka, K. J. Org. Chem. 2002, 67, 431. (b) Hawkins, J. M.; Nambu,
M.; Loren, S. Org. Lett. 2003, 5, 4293. (c) Birney, D. M.; Houk, K. N. J.
Am. Chem. Soc. 1990, 112, 4127. (d) Sibi, M. P.; Sausker, J. B. J. Am.
Chem. Soc. 2002, 124, 984. (e) Avalos, M.; Babiano, R.; Brao, J. L.; Cintas,
P.; Jimenez, J. L.; Palacios, J. C.; Silva, M. A. J. Org. Chem. 2000, 65,
6613.
(2) (a) Evans, D. A.; Miller, S. J.; Lectka, T.; von Matt, P. J. Am. Chem.
Soc. 1999, 121, 7559. (b) Evans, D. A.; Barnes, D. M.; Johnson, J. S.;
Lectka, T.; von Matt, P.; Miller, S. J.; Murry, J. A.; Norcross, R. D.;
Shaughnessy, E. A.; Campos, K. R. J. Am. Chem. Soc. 1999, 121, 7582.
(c) Castellino, S. J. Org. Chem. 1990, 55, 5197.
(3) For diastereoselective reactions with Oppolzer’s sultam as chiral
auxiliary and related work on tiglates and cycloalkenes, see: (a) Oppolzer,
W.; Kingma, A. J. HelV. Chim. Acta 1989, 72, 1337. (b) Oppolzer, W.;
Kingma, A. J.; Poli, G. Tetrahedron 1989, 45, 479. (c) Niu, D.; Zhao, K.
J. Am. Chem. Soc. 1999, 121, 2456. (d) Karlsson, S.; Hoegberg, H.-E. J.
Chem. Soc., Perkin Trans. 1 2002, 1076. (e) Davies, S. G.; Epstein, S. W.;
Garner, A. C.; Ichihara, O.; Smith, A. D. Tetrahedron: Asymmetry 2002,
13, 1555. (f) Adam, W.; Degen, H.-G.; Krebs, O.; Saha-Mller, C. J. Am.
Chem. Soc. 2002, 124, 12938. (g) Lee, W.-D.; Chiu, C.-C.; Hsu, H.-L.;
Chen, K. Tetrahedron 2004, 60, 6657. (h) Dieter, R. K.; Lu, K.; Velu, S.
E. J. Org. Chem. 2000, 65, 8715.
10.1021/ol050599d CCC: $30.25
© 2005 American Chemical Society
Published on Web 05/17/2005

absence of A1,3 interactions remained unclear. Our initial
study involving hydroxylamine additions to R,â-disubstituted
N-H acrylimides 3 was reported recently (eq 1).6 The
additions were highly enantio- and diastereoselective. The
level of selectivity suggested that high rotamer control was
involved, and the absolute configuration of the products 5
suggested that the reaction proceeded via the s-cis rotamer
of 3. Whether reaction involved the s-cis rotamer because
the population of the s-trans complex was minimal or because
both rotamers were in equilibrium but the s-cis rotamer
reacted selectively (Curtin-Hammett principle) was unclear.
As expected, the reactivity with N-H imide templates 3 was
dramatically higher than with A1,3 templates such as oxazo-
lidinone or N-methylbenzimide. Intermolecular radical ad-
ditions to substrates 3 mediated by MgI2-4 also proceed with
good yield and high diastereo- and enantioselectivity (maxi-
mal dr 99:1, 94% ee) and proceed via the s-cis rotamer.7
A number of catalysts have been reported for enantiose-
lective nitrone additions to R-unsubstituted acrylate and
crotonate type substrates with varying templates.8 However,
enantioselective nitrone additions to R,â-disubstituted sub-
strates are rare, presumably because of low reactivity.9 We
recently reported a chiral Cu(OTf)2-ligand 4 catalyst that
gives highly exo- and enantioselective nitrone additions to
pyrazolidinone crotonates and cinnamates.10 As expected,
however, when oxazolidinone tiglate 2 was tested under
analogous reaction conditions, conversion was negligible.
Our results using the same Cu(OTf)2-4 Lewis acid with
R,â-disubstituted N-H imide substrates 3 are shown in Table
1. Entries 1-5 show additions of N-methylnitrone 6a to
Table 1. Nitrone Cycloadditions
entry SM nitrone time (days) % yield (SM)a,b ee (%)cb 7/8
Following the success of our initial studies, we have now
investigated additional types of addition reactions in order
to assess whether useful s-cis rotamer control is general in
Lewis acid activated reactions of substrates 3. In this paper
we report our results involving nitrone, nitrile oxide, and
Diels-Alder cycloadditions to R,â-disubstituted acrylimides
3. The results show improved reactivity and good to high
enantioselectivity in each case, suggesting that the N-H imide
solution to the R,â-disubstitution problem may have con-
siderable generality.
1
2
3
4
5
6
7
8
9
3b
3c
3d
3e
3f
3f
3f
3f
3b
6a
6a
6a
6a
6a
6b
6c
6d
6d
6b
7
7
10
10
7
7
10
5
60(30)
57
63(25)
<15
82
89
89
77
50
94
89
91
99/1
81/19
99/1
92
97
94
89
86
98
99/1
99/1
99/1
94/6
85/15
99/1
7
7
10 3b
62(34)
a For reaction conditions see Supporting Information. b Recovered starting
material following chromatography in parentheses. c Chiral HPLC analysis.
(4) (a) For enantioselective conjugate addition to tiglates see: Doi, H.;
Sakai, T.; Iguchi, M.; Yamada, K.-i.; Tomioka, K. J. Am. Chem. Soc. 2003,
125, 2886-2887. Doi, H.; Sakai, T.; Iguchi, M.; Yamada, K.-i.; Tomioka,
K. Chem. Commun. 2004, 1850. (b) For an example of organocatalysis in
nitrone cycloaddition to cyclopentene carboxylic acid esters, see: Karlsson,
S.; Hgberg, H.-E. Eur. J. Org. Chem. 2003, 2782.
(5) For pioneering work on the use of imide-unsubstituted substrates in
conjugate additions, see: (a) Myers, J. K.; Jacobsen, E. N. J. Am. Chem.
Soc. 1999, 121, 8959 and (b) Sammis, G. M.; Jacobsen, E. N. J. Am. Chem.
Soc. 2003, 125, 4442 and references therein.
(6) Sibi, M. P.; Prabagaran, N.; Ghorpade, S. G.; Jasperse, C. P. J. Am.
Chem. Soc. 2003, 125, 11796.
(7) Sibi, M. P.; Petrovic, G.; Zimmerman, J. J. Am. Chem. Soc. 2005,
127, 2390.
(8) For recent reviews, see: (a) Synthetic Applications of 1,3-Dipolar
Cycloaddition Chemistry toward Heterocycles and Natural Products; Padwa,
A., Pearson, W. H., Eds.; John Wiley and Sons: Hoboken, NJ, 2003. (b)
Gothelf, K. V.; Jrgensen K. A. Chem Commun. 2000, 1449. (c) Gothelf,
K. V.; Jrgensen, K. A. Chem. Rev. 1998, 98, 863. Also see: (d) Gothelf,
K. V.; Thomsen, I.; Jrgensen, K. A. J. Am. Chem. Soc. 1996, 118, 59. (e)
Kobayashi, S.; Kawamura, M. J. Am. Chem. Soc. 1998, 120, 5840. (f)
Kanemasa, S.; Oderaotoshi, Y.; Tanaka, J.; Wada, E. J. Am. Chem. Soc.
1998, 120, 12355. (g) Desimoni, G.; Faita, G.; Mortoni, A.; Righetti, P.
Tetrahedron Lett. 1999, 40, 2001. (h) Iwasa, S.; Tsushima, S.; Shimada,
T.; Nishiyama, H. Tetrahedron 2002, 58, 227. (i) Suga, H.; Kakehi, A.;
Ito, S.; Sugimoto, H. Bull. Chem. Soc. Jpn. 2003, 76, 327.
imides 3 with varying R,â-substituents. In each case the
regioselectivity is very high, with only products forming in
which the oxygen end of the dipole adds to the â-carbon of
the acceptor. With the exception of â-unsubstituted acryla-
mide 3c, the diastereoselectivity is outstanding, strongly
favoring the 7-exo products over the 8-endo products.11 Most
importantly, enantioselectivity is consistently excellent,
around 90% or above. The one limitation evident in entries
1-5 is that of reactivity. Even with long reaction times,
(9) Mita, T.; Ohtsuki, N.; Ikeno, T.; Yamada, T. Org. Lett. 2002, 4, 2457.
Also see ref 4b.
(10) Sibi, M. P.; Ma, Z.; Jasperse, C. P. J. Am. Chem. Soc. 2004, 126,
718-719.
(11) Use of the same Cu(OTf)2-4 Lewis acid in our previous study (ref
10) also gave strong exo selectivity for crotonate and cinnamate substrates,
but the exo/endo selectivity with R,â-disubstituted substrates 3 is signifi-
cantly higher.
2350
Org. Lett., Vol. 7, No. 12, 2005

yields are generally modest and some starting material is
often recovered. The relative reactivities of cyclopentene and
cyclohexene substrates 3f and 3e are interesting. Cyclopen-
tene substrate 3f showed unusually good reactivity (entry
5), whereas cyclohexene analogue 3e had unusually poor
reactivity (entry 4). Entries 5-8 screen the influence of the
nitrone substituents. Results are consistently excellent with
N-alkyl nitrones (entries 5-7), including the N-benzyl nitrone
6c (89% yield, 94% ee, 99:1 exo/endo). The N-phenyl nitrone
6d shows somewhat reduced enantio- and diastereoselectivity
with both cyclopentene substrate 3f (entry 8) and tiglate 3b
(entry 9).
One other set of results not shown in the Table 1 also
bears comment. A brief screening of the impact of the imide
R group, using tert-butyl imide 3b, isopropyl imide 3a, and
benzimide 3g, showed significant dependence on the imide
R group. The alkyl imides 3a and 3b showed much better
reactivity than the benzimide 3g, with the bulkier tert-butyl
imide 3b providing optimal yield, enantioselectivity, and
diastereoselectivity. That the nature of the R group can be
adjusted to optimize chemistry when N-H imides are used
is an important variable to keep in mind.12
The absolute and relative configuration of the exo cy-
cloadduct from cyclopentene 3f and nitrone 6b (entry 6) was
established by X-ray crystallography. The absolute config-
uration suggests reaction via the s-cis rotamer, since the
configuration of the â-carbon in the product is the same as
when nitrones or radicals add to R-unsubstituted oxazolidi-
none and pyrazolidinone crotonates and cinnamates under
the influence of the same Cu(OTf)2-4 Lewis acid.13 The
regioselectivity and relative configuration of the racemic
cycloadduct from tiglate 3b and nitrone 6b (entry 10) was
also established by X-ray crystallography.
The nitrone cycloaddition results in Table 1 confirm two
important points. First, rotamer control for R,â-disubstituted
acrylamides with N-H imide templates is again very high.
Enantioselectivities as high as 97% (entry 6) and 98% (entry
10) would be impossible without strong s-cis/s-trans rotamer
control. Second, the yields with the N-H imide template as
compared to the A1,3 oxazolidinone template (negligible
conversion for substrate 1) confirm that use of N-H imides
greatly improves reactivity for R,â-disubstituted substrates.
We have recently reported that both MgX2-4 and NiX2-4
complexes catalyze highly enantioselective and regioselective
nitrile oxide additions to pyrazolidinone crotonates.14 We
have now screened these same Lewis acids for their ability
to catalyze nitrile oxide additions to R,â-disubstituted
acrylimides 3 (Table 2). As shown in entries 1-5, Ni(ClO4)2,
Mg(ClO4)2, and MgI2 all provide good catalysis, resulting
in yields of 79-99% (entries 1, 2, and 4). Mg(ClO4)2 gave
Table 2. Nitrile Oxide Cycloadditions
entry
SM
LA
MgI2
time (days)
% yielda
ee (%)b
1
2
3
4
5
6
7c
8d
3g
3g
3g
3g
3g
3a
3b
12
2
2
2
4
4
7
7
7
82
79
51
99
74
56
c
65
81
82
77
37
83
Mg(ClO4)2
Mg(NTf2)2
Ni(ClO4)2
Ni(SbF6)2
Mg(ClO4)2
Mg(ClO4)2
Mg(ClO4)2
12
95
a For reaction conditions see Supporting Information. Isolated yield after
column purification. b Chiral HPLC analysis. c A complex mixture formed.
d A complex mixture of side products formed.
slightly higher ee’s than Ni(ClO4)2 did (81% versus 77%,
entries 2 and 4). All of these reactions were completely
regioselective for the C-adduct 10, in which the carbon
terminus of the dipole adds to the â-carbon of the acceptor,
to the exclusion of O-adduct 11. The reaction using Mg-
(ClO4)2 shows a surprising sensitivity to the imide R group,
with imides 3g and 3a reacting much more cleanly than imide
3b (compare entries 2, 6, with 7). This is in curious contrast
to the nitrone reactions, in which benzimide 3g showed
inferior reactivity.12
Entry 8 shows a comparison reaction involving a substrate
12 with a pyrazolidinone template rather than an N-H imide
template.15 The reactivity is very low. This again demon-
strates the superior reactivity with N-H imide templates as
compared to A1,3 templates, and once again the reasonably
good enantioselectivities (77-83% ee, entries 2-4 and 6)
indicate a reasonable level of rotamer control.
A host of catalysts have been developed for enantiose-
lective Diels-Alder additions, such that the ability to catalyze
cyclopentadiene addition to crotonate acceptors is a standard
benchmark for new chiral Lewis acids. However, enantiose-
lective Diels-Alder addition to R,â-disubstituted acrylimides
remains an unsolved problem, particularly when neither
substituent has electron-withdrawing character to improve
reactivity.16 Our results on Diels-Alder reaction of R,â-
alkyldisubstituted N-H acrylimides 3 are shown in Table 3.
The Lewis acids MgI2-4 and Cu(OTf)2-4 that catalyzed the
(12) Sensitivity to the imide R group was also observed in amine and
radical additions, see refs 6 and 7.
(13) (a) Addition of nitrones to pyrazolidinone crotonates/cinnamates:
ref 10. (b) Addition of radicals to oxazolidinone crotonates/cinnamates: Sibi,
M. P.; Chen, J. J. Am. Chem. Soc. 2001, 123, 9472-9473. (c) Addition of
radicals to pyrazolidinone crotonates/cinnamates: Sibi, M. P.; Prabagaran,
N. Synlett 2004, 13, 2421-2424. Each of these reactions occurs via s-cis
rotamers.
(15) This pyrazolidinone template was very effective in our previous
nitrile oxide additions to crotonates and cinnamates; see ref 14.
(16) For the use of tiglates as dienophiles, see: (a) de Miranda, D. S.;
da Conceicao, G. J. A.; Zukerman-Schpector, J.; Guerrero, M. C.;
Schuchardt, U.; Pinto, A. C.; Rezende, C.; Marsaioli, A. J. J. Braz. Chem.
Soc. 2001, 12, 391. (b) Dogan, O.; Oppolzer, W. Turk. J. Chem. 2001, 25,
273. (c) Paczkowski, R.; Maichle-Mossmer, C.; Maier, M. E. Org. Lett.
2000, 2, 3967. (d) Liu, W.-C.; Liao, C.-C. Synlett 1998, 912.
(14) Sibi, M. P.; Itoh, K.; Jasperse, C. P. J. Am. Chem. Soc. 2004, 126,
5366-5367.
Org. Lett., Vol. 7, No. 12, 2005
2351

stituted systems as compared to R-unsubstituted analogues
is not eliminated entirely. Even in the absence of A1,3 strain,
electron-donating R-alkyl substituents still have a rate-
retarding electronic effect associated with a higher LUMO,
and in the case of concerted cycloadditions also have a steric
effect. Nevertheless, while the use of N-H imide templates
does not bring the reactivity of R,â-dialkylated acrylamides
all the way back up to the level of R-unsubstituted analogues,
it appears general that the use of these templates goes a long
way toward lessening reactivity problems.17 Given that both
nitrone and Diels-Alder cycloadditions are relatively slow
in general, the fact that these cycloadditions proceed in useful
yields using R,â-dialkylated substrates 3 is impressive.
The enantioselectivities in this paper, combined with our
previously reported hydroxylamine6 and radical7 additions,
clearly demonstrate that useful levels of rotamer control exist
when R,â-disubstituted N-H acrylimides are activated by
Lewis acids. Calculations by Houk on electrophilic reactions
of R,â-unsaturated carbonyls predict an inherent electronic
reactivity preference for s-cis conformations.1c In our reac-
tions it is unclear whether the s-cis rotamer of 3 reacts
preferentially because the ground-state population of s-trans
rotamer is minimal, or because both rotamers are in equi-
librium with the s-cis rotamer reacting selectively (Curtin-
Hammett principle), or both. Nevertheless, rotamer prefer-
ence appears to be usefully high for five different types of
addition reactions (hydroxylamine,6 radical,7 nitrone, nitrile
oxide, and Diels-Alder). The rotamer preference also
appears to be high using four metal cations (CuX2, MgX2,
NiX2, ScX3) with widely differing Lewis acidities and
coordination environments. Whatever the origins, our results
suggests that useful levels of rotamer control is likely to be
fairly general when R,â-disubstituted N-H acrylimides are
activated by Lewis acids.
Table 3. Diels-Alder Cycloadditions
entry
SM
temp, time (h)
% yielda,b
ee (%)c
drd
1
2
3
4
5
3b
3d
3f
3a
3h
-40°, 24
-20°, 24
-20°, 24
-40°, 24
-40°, 24
71
53
70
67
78
80
80
65
65
70
84:16
71:29
73:27
60:40
81:19
a For reaction conditions see Supporting Information. b Isolated yield after
column purification. c Chiral HPLC analysis. d NMR analysis.
nitrone and nitrile oxide cycloadditions were too weak to
catalyze cyclopentadiene addition. The stronger 3+ Lewis
acid Sc(OTf)3 in conjunction with chiral pybox ligand 13
did provide enough reactivity to give modest yields. Entries
1-3, in which the R- and â-alkyl substituents vary, show
modest yields, enantioselectivity, and diastereoselectivity. Of
several imide R groups tested (entries 1, 4, and 5), the larger
tert-butyl group gives optimal enantio- and diastereoselec-
tivity (80% ee, 84:16 dr, entry 1). Enantioselectivity drops
sharply with less than 50% Lewis acid (70%, 60%, and 12%
ee with 100%, 50%, and 30% catalyst, respectively, in
additions to substrate 3h). Most likely the enantioselectivity
in these reactions is limited not by a lack of rotamer control
but rather by the Sc(OTf)3-13 Lewis acid.
The results in this paper demonstrate that nitrones, nitrile
oxides, and dienes can all add to R,â-disubstituted acrylim-
ides in good-to-excellent yields, with good-to-excellent regio-
and diastereoselectivities, and with good-to-excellent enan-
tioselectivities. This is true even when neither the R- nor
â-substituents are electron-withdrawing. Two new stereo-
centers including chiral quaternary carbons are formed in
each case.
Acknowledgment. We thank the National Science Foun-
dation (NSF-CHE-0316203) for financial support.
Supporting Information Available: Characterization
data for compounds and experimental procedures; crystal-
lographic data in CIF format. This material is available free
of charge via the Internet at http://pubs.acs.org.
OL050599D
Reactivities of R,â-dialkylated substrates in these reactions
are consistently much higher with N-H imide templates than
when traditional oxazolidinone templates or other A1,3
templates are used, just as was true in our hydroxylamine
and radical additions.6,7 The diminished reactivity of R-sub-
(17) The enhanced reactivity with N-H imide templates is not unique to
R,â-dialkylated substrates. We have also observed in several reactions that
crotonates and cinnamates with N-H imide templates are more reactive than
with oxazolidinone templates. Presumably the use of N-H imides provides
an electronic advantage apart from the differing sensitivity to A1, 3 strain.
2352
Org. Lett., Vol. 7, No. 12, 2005