A formal [3 + 3] cycloaddition reaction. 5. An enantioselective intramolecular formal aza-[3 + 3] cycloaddition reaction promoted by chiral amine salts

Article abstract of DOI:10.1021/jo050171s

A detailed account on chiral secondary amine salt promoted enantioselective intramolecular formal aza-[3 + 3] cycloadditions is described here for the first time. The dependence of enantioselectivity on the structural feature of these chiral amines is tho

Full text of DOI:10.1021/jo050171s

A Formal [3 + 3] Cycloaddition Reaction. 5. An Enantioselective
Intramolecular Formal Aza-[3 + 3] Cycloaddition Reaction
Promoted by Chiral Amine Salts
Aleksey I. Gerasyuto, Richard P. Hsung,* Nadiya Sydorenko, and Brian Slafer
Department of Chemistry, University of Minnesota, Minneapolis, Minnesota 55455
hsung@chem.umn.edu
Received January 27, 2005
A detailed account on chiral secondary amine salt promoted enantioselective intramolecular formal
aza-[3 + 3] cycloadditions is described here for the first time. The dependence of enantioselectivity
on the structural feature of these chiral amines is thoroughly investigated. This study also reveals
a very interesting reversal of the stereochemistry in the respective cycloadducts obtained using
C1- and C2-symmetric amine salts. In addition, the influence of solvents, counteranions, and
temperatures on the enantioselectivity is described, and a unified mechanistic model based on
experimental results as well as semiempirical calculations is proposed.
Introduction
Over the past several years we have been developing
an annulation reaction using vinylogous amides and R,â-
unsaturated iminium salts.1,2 Specifically, this annulation
provides a powerful synthetic access to nitrogen hetero-
cycles such as dihydropyridines 3 and quinolizidines 7
in intermolecular and intramolecular manifolds, respec-
tively [Figure 1].3-7 It represents a formal aza-[3 + 3]
cycloaddition8,9 with two of the five carbons along with
the nitrogen atom coming from the vinylogous amide [see
1 or 4] and the remaining three carbons coming from the
R,â-unsaturated iminium salt [see 2 or 5]. The net result
is the formation of two σ-bonds and a new stereocenter
adjacent to the nitrogen atom.
While both reaction manifolds provide a unique and
novel approach to the synthesis of alkaloids,7,10,11 they
proceed through very different mechanistic courses.5,6
The intermolecular formal aza-[3 + 3] cycloaddition
FIGURE 1. A formal aza-[3 + 3] cycloaddition.
(1) For reviews, see: (a) Hsung, R. P.; Kurdyumov, A. V.; Sydorenko,
N. Eur. J. Org. Chem. 2005, 1, 23-44. (b) Coverdale, H. A.; Hsung, R.
P. ChemTracts 2003, 16, 238. (c) Harrity, J. P. A.; Provoost, O. Org.
Biomol. Chem. 2005, 3, 1349-1358.
(2) For a review on vinylogous amide chemistry, see: Kucklander,
U. Enaminones as Synthons. In The Chemistry of Functional Groups:
The Chemistry of Enamines Part I; Rappoport, Z., Ed.; John Wiley &
Sons: New York, 1994; p 523.
reaction [1 + 2 f 3] mechanistically constitutes a tandem
Knoevenagel condensation-6π-electron electrocyclic ring-
closure involving a 1-azatriene intermediate.12-14 The
latter part of this tandem sequence, the pericyclic ring-
10.1021/jo050171s CCC: $30.25 © 2005 American Chemical Society
Published on Web 05/04/2005
4248
J. Org. Chem. 2005, 70, 4248-4256

A Formal [3 + 3] Cycloaddition Reaction. 5
closure, can be rendered highly stereoselective when
driven by a chiral auxiliary.5b,c,7b On the other hand, the
intramolecular process [4 f 7] proceeds through a
tandem sequence that features an N-1,4-addition via the
in situ generated R,â-unsaturated iminium salt 5, and a
C-1,2-addition through the iminium intermediate 6 [Fig-
ure 1].5b,6 Because it is possible to regenerate the
amine salt during the C-1,2-addition and subsequent
elimination, only a sub-stoichiometric amount of second-
ary amine salts was needed to afford 7.6
Consequently, this difference in mechanism allowed us
to envision the possibility of an asymmetric intramolecu-
lar formal aza-[3 + 3] cycloaddition using a catalytic
amount of chiral amine salt [Figure 1]. Given that
organocatalysis has attracted much attention recently,15-19
investigations here should represent a unique opportu-
nity to explore organocatalysis in a tandem process. In
the current paper, we reveal details on chiral amine salt-
promoted enantioselective intramolecular formal aza-[3
+ 3] cycloadditions.
(3) For recent studies in this area, see: (a) Goodenough, K. M.;
Moran, W. J.; Raubo, P.; Harrity, J. P. A. J. Org. Chem. 2005, 70, 207.
(b) Bose, G.; Nguyen, V. T. H.; Ullah, E.; Lahiri, S.; Go¨rls, H.; Langer,
P. J. Org. Chem. 2004, 69, 9128. (c) Abelman, M. M.; Curtis, J. K.;
James, D. R. Tetrahedron Lett. 2003, 44, 6527. (d) Hedley, S. J.; Moran,
W. J.; Price, D. A.; Harrity, J. P. A. J. Org. Chem. 2003, 68, 4286. (e)
Chang, M.-Y.; Lin, J. Y.-C.; Chen, S. T.; Chang, N.-C. J. Chin. Chem.
Soc. 2002, 49, 1079. (f) Hedley, S. J.; Moran, W. J.; Prenzel, A. H. G.
P.; Price, D. A.; Harrity, J. P. A. Synlett 2001, 1596. (g) Davies, I. W.;
Marcoux, J.-F.; Reider, P. J. Org. Lett. 2001, 3, 209. (h) Davies, I. W.;
Taylor, M.; Marcoux, J.-F.; Wu, J.; Dormer, P. G.; Hughes, D.; Reider,
P. J. J. Org. Chem., 2001, 66, 251. (i) Nemes, P.; Bala´zs, B.; To´th, G.;
Scheiber, P. Synlett 1999, 222. (j) Hua, D. H.; Chen, Y.; Sin, H.-S.;
Robinson, P. D.; Meyers, C. Y.; Perchellet, E. M.; Perchellet, J.-P.;
Chiang, P. K.; Biellmann, J.-P. Acta Crystallogr. 1999, C55, 1698.
(4) Also see: (a) P. Benovsky, P.; Stephenson, G. A.; Stille, J. R. J.
Am. Chem. Soc. 1998, 120, 2493. (b) Heber, D.; Berghaus, Th. J.
Heterocycl. Chem. 1994, 31, 1353. (c) Paulvannan, K.; Stille, J. R.
Tetrahedron Lett. 1993, 34, 215 and 6677. (d) Paulvannan, K.; Stille,
J. R. J. Org. Chem. 1992, 57, 5319. (e) Chaaban, J.; Greenhill, J. V.;
Rauli, M. J. Chem. Soc., Perkin Trans. 1 1981, 3120. (f) Hickmott, P.
W.; Sheppard, G. J. Chem. Soc. C 1971, 2112.
Results and Discussions
1. Feasibility and Absolute Stereochemistry. 1.1.
Synthetic Feasibility. To explore the feasibility of an
enantioselective formal aza-[3 + 3] cycloaddition, viny-
logous amide 8 was synthesized20 and its reaction was
examined using a series of L-proline based amine salts
9a-e, which are C1-symmetric, and various pyrrolidine-
based amine salts 10a-c, which are C2-symmetric. These
results are summarized in Table 1.
(5) For intermolecular formal aza-[3 + 3] cycloadditions, see: (a)
Sydorenko, N.; Hsung R. P.; Darwish, O. S.; Hahn, J. M.; Liu, J. J.
Org. Chem. 2004, 69, 6732. (b) Sklenicka, H. M.; Hsung, R. P.;
McLaughlin, M. J.; Wei, L.-L.; Gerasyuto, A. I.; Brennessel, W. W. J.
Am. Chem. Soc. 2002, 124, 10435. (c) Sklenicka, H. M.; Hsung, R. P.;
Wei, L.-L.; McLaughlin, M. J.; Gerasyuto, A. I.; Degen, S. J.; Mulder,
J. A. Org. Lett. 2000, 2, 1161. (d) Hsung, R. P.; Wei, L.-L.; Sklenicka,
H. M.; Douglas, C. J.; McLaughlin, M. J.; Mulder, J. A.; Yao, L. J. Org.
Lett. 1999, 1, 509.
It can be seen that in all cases the reaction proceeded
well to give the formal cycloadduct 11 in moderate to good
yields. The use of 30-40 mol % of the chiral catalyst was
sufficient to promote the cycloaddition [entries 1, 2, 6, 8,
and 10], while increasing the amount of the catalyst did
not necessarily lead to an improved enantioselectivity
[entries 3-5, 7, and 9].
(6) For intramolecular formal aza-[3 + 3] cycloaddition, see: Wei,
L.-L.; Sklenicka, H. M.; Gerasyuto, A. I.; Hsung, R. P. Angew. Chem.,
Int. Ed. 2001, 40, 1516.
More importantly, three features of the cycloaddition
captured our attention. First, C2-symmetric pyrrolidine-
(7) For our applications in natural product syntheses, see: (a) Luo,
S.; Zificsak, C. Z.; Hsung, R. P. Org. Lett. 2003, 5, 4709. (b) McLaughlin,
M. J.; Hsung, R. P.; Cole, K. C.; Hahn, J. M.; Wang, J. Org. Lett. 2002,
4, 2017.
(8) The term “formal [3 + 3]” was used to describe [3 + 3] carbo-
cycloadditions. See: (a) Seebach, D.; Missbach, M.; Calderari, G.;
Eberle, M. J. Am. Chem. Soc. 1990, 112, 7625. For earlier studies on
[3 + 3] carbo-cycloadditions, see: (b) Landesman, H. K.; Stork, G. J.
Am. Chem. Soc. 1956, 78, 5129.
(9) For recent reviews on formal carbo-[3 + 3] cycloadditions using
enamines, enol ethers, or â-ketoesters see: (a) Filippini, M.-H.;
Rodriguez, J. Chem. Rev. 1999, 99, 27. (b) Filippini, M.-H.; Faure, R.;
Rodriguez, J. J. Org. Chem. 1995, 60, 6872, and see refs 21-33 cited
within. (c) For reviews on metal-mediated stepwise [3 + 3] cycloaddi-
tion reactions, see: (d) Fru¨hauf, H.-W. Chem. Rev. 1997, 97, 523. (e)
Lautens, M.; Klute, W.; Tam, W. Chem. Rev. 1996, 96, 49.
(10) (a) Grundon, M. F. In The Alkaloids: Quinoline Alkaloids
Related to Anthranilic Acids; Academic Press: London, UK, 1988; Vol.
32, p 341. (b) Daly, J. W.: Garraffo, H. M.; Spande, T. F. The Alkaloids;
Cordell, G. A., Ed.; Academic Press: New York, 1993; Vol. 43, p 185.
(c) Michael, J. P. Nat. Prod. Rep. 1999, 16, 675 and 697. (d) Jones, T.
H.; Gorman, J. S. T.; Snelling, R. R.; Delabie, J. H. C.; Blum, M. S.;
Garraffo, H. M.; Jain, P.; Daly, J. W.; Spande, T. F. J. Chem. Ecol.
1999, 25, 1179. (e) Michael, J. P. Nat. Prod. Rep. 2000, 17, 579. (f)
Lewis, J. R. Nat. Prod. Rep. 2001, 18, 95.
(11) For a recent review on preparations of piperidines, see: Laschat,
S.; Dickner, T. Synthesis 2000, 1781.
(12) For leading references on electrocyclic ring-closures involving
1-azatrienes, see: (a) Maynard, D. F.; Okamura, W. H. J. Org. Chem
1995, 60, 1763. (b) de Lera, A. R.; Reischl, W.; Okamura, W. H. J. Am.
Chem. Soc. 1989, 111, 4051. For an earlier account, see: (c) Oppolzer,
V. W. Angew. Chem. 1972, 22, 1108.
(13) For recent elegant accounts on stereoselective ring-closure of
1-azatrienes, see: (a) Tanaka, K.; Katsumura, S. J. Am. Chem. Soc.
2002, 124, 9660. (b) Tanaka, K.; Mori, H.; Yamamoto, M.; Katsumura,
S. J. Org. Chem. 2001, 66, 3099. (c) Tanaka, K.; Kobayashi, T.; Mori,
H.; Katsumura S. J. Org. Chem. 2004, 69, 5906.
(14) For a review on rotational preferences leading to diastereomeric
induction during a 6π-electron electrocyclic ring closure see: Okamura,
W. H.; de Lera, A. R. Comprehensive Organic Synthesis; Trost, B. M.,
Fleming, I., Eds.; Paquette, L. A., volume Ed.; Pergamon Press: New
York, 1991; Vol. 5, pp 699-750.
(15) For a recent review on proline-catalyzed asymmetric reactions
see: List, B. Tetrahedron 2002, 58, 5573. For a comprehensive set of
reviews on asymmetric organocatalysis see: Acc. Chem. Res. 2004, 37,
487-631.
(16) (a) Ouellet, S. G.; Tuttle, J. B.; MacMillan, D. W. C. J. Am.
Chem. Soc. 2005, 127, 32. (b) Northrup, A. B.; MacMillan, D. W. C.
Science 2004, 305, 1752. (c) Brown, S. P.; Goodwin, N. C.; MacMillan,
D. W. C. J. Am. Chem. Soc. 2003, 125, 1192. (d) Paras, N. A.;
MacMillan, D. W. C. J. Am. Chem. Soc. 2002, 124, 7894. (e) Northrup,
A. B.; MacMillan, D. W. C. J. Am. Chem. Soc. 2002, 124, 2458. (f)
Austin, J. F.; MacMillan, D. W. C. J. Am. Chem. Soc. 2002, 124, 1172.
(g) Paras, N. A.; MacMillan, D. W. C. J. Am. Chem. Soc. 2001, 123,
4370. (h) Jen, W. S.; Wiener, J. J. M.; MacMillan, D. W. C. J. Am.
Chem. Soc. 2000, 122, 9874. (i) Ahrendt, K. A.; Borths, C. J.;
MacMillan, D. W. C. J. Am. Chem. Soc. 2000, 122, 4243.
(17) (a) Hechavarria Fonseca, M. T.; List, B. Angew. Chem., Int. Ed.
2004, 43, 3958. (b) List, B.; Hoang, L.; Martin, H. J. PNAS 2004, 101,
5839. (c) Vignola, N.; List, B. J. Am. Chem. Soc. 2004, 126, 450. (d)
Martin, H. J.; List, B. Synlett 2003, 1901. (e) Pidathala, C.; Hoang, L.;
Vignola, N.; List, B. Angew. Chem., Int. Ed. 2003, 42, 2785. (f) Hoang,
L.; Bahmanyar, S.; Houk, K. N.; List, B. J. Am. Chem. Soc. 2003, 125,
16. (g) List, B. J. Am. Chem. Soc. 2002, 124, 5656.
(18) (a) Marigo, M.; Bachmann, S.; Halland, N.; Braunton, A.;
Jorgensen, K. A. Angew. Chem., Int. Ed. 2004, 43, 5507. (b) Bella, M.;
Jorgensen, K. A. J. Am. Chem. Soc. 2004, 126, 5672. (c) Halland, N.;
Aburel, P. S.; Jorgensen, K. A. Angew. Chem., Int. Ed. 2004, 43, 1272
(d) Juhl, K.; Jorgensen, K. A. Angew. Chem., Int. Ed. 2003, 42, 1498.
(e) Melchiorre, P.; Jorgensen, K. A. J. Org. Chem. 2003, 68, 4151. (f)
Halland, N.; Aburel, P. S.; Jorgensen, K. A. Angew. Chem., Int. Ed.
2003, 42, 661. (g) Halland, N.; Hazell, R. G.; Jorgensen, K. A. S. J.
Org. Chem. 2002, 67, 8331. (h) Kumaragurubaran, N.; Juhl, K.;
Zhuang, W.; Bogevig, A.; Jorgensen, K. A. J. Am. Chem. Soc. 2002,
124, 6254.
(19) (a) Yamamoto, Y.; Momiyama, N.; Yamamoto, H. J. Am. Chem.
Soc. 2004, 126, 5962. (b) Hayashi, Y.; Yamaguchi, J.; Sumiya T.; Shoji,
M. Angew. Chem., Int. Ed. 2004, 43, 1112. (c) Alexakis, A.; Andrey, O.
Org. Lett. 2002, 4, 3611. (d) Betancort, J. M.; Barbas, C. F., III Org.
Lett. 2001, 3, 3737.
(20) See the Supporting Information for details of all preparations,
characterizations, and relevant Spartan′02 models.
J. Org. Chem, Vol. 70, No. 11, 2005 4249

Gerasyuto et al.
SCHEME 1. Independent Synthesis of S-15
TABLE 1. Chiral Pyrrolidine-Catalyzed Cycloaddition
entry
catalyst
mol %
yielda (%)
er (R:S)b
enantiomerically pure S-15 [Figure 2]. Interestingly,
when the S-enantiomer of C1-symmetric catalyst such as
9e was used, the cycloaddition favored the R-enantiomer,
while the 2S,5S-enantiomer of C2-symmetric catalyst 10c
gave predominantly S-enantiomer.
1
2
9ac
40
30
100
100
100
40
100
40
100
40
26
68
68
52
56
60
63
57
67
46
56:44
52:48
52:48
50:50
55:45
81:19
80:20
50:50
60:40
27:73
9bc
3
9b
4
9cc,d
9dd
5
6
9ee
7
9e
8
10af
ent-10bg
10cd
9
10
a Isolated yields. b Determined by CSP-HPLC. c Commercially
available. d The acetate salt was generated in situ by addition of
1.0 equiv of AcOH to the reaction mixture. e The acetate salt was
synthesized from commercially available free amine (see ref 20).
f For the synthesis, see ref 21. g The 2R,5R-isomer was used for
the reaction.
based catalysts gave higher enantiomeric ratios than
those obtained from using L-proline-based derivatives
[i.e., entries 2 and 5 versus 9 and 10, respectively].
Second, the enantioselectivity could be modestly en-
hanced by using sterically more bulky catalysts such as
9e [entry 6] and 10c [entry 10] but appears to have
reached the maximum. Third, it was found that C1- and
C2-symmetric catalysts provided cycloadduct 11 with
enrichment for the opposite enantiomer [entry 6 versus
10].
FIGURE 2. Determination of absolute configuration.
1.2. Absolute Configuration Assignment. To assess
the origin of observed enantioselectivity, the absolute
configuration of the major enantiomer was first deter-
mined. After numerous failed attempts to derivatize the
cycloadduct 11,22 S-15 was synthesized independently as
shown in Scheme 1,23 featuring a resolution of racemic
piperidine 12 with (+)-CSA using a procedure developed
by Toy.24
2. Catalyst Screening and Synthetic Scope. 2.1.
Catalyst Screening. In the search for an amine catalyst
that can provide improved asymmetric induction, chiral
piperidines 16a-f were screened [Table 2]. The first
observation was that the nature of the counteranion in
the catalyst did not appear to have any impact on the
enantioselectivity. For instance, using the L-tartrate salt
of pipecolic acid (16b) did not lead to any improvement
in the enantiomeric ratio when compared to the acetate
salt 16a [entry 2 versus 1]. Most disappointingly, none
of the chiral piperidines provided any meaningful enan-
tioselectivity.
Moreover, treatment of vinylogous amide 8 with the
amine salt 16e bearing a diphenylhydroxymethyl sub-
stituent did not lead to the desired dihydropyridine 11
under standard reaction conditions [entry 5]. Instead, the
reaction had to be heated at an elevated temperature
(130 °C) for 14 h in a sealed tube in which severe
decomposition occurred with only ca. 50% conversion.
Although we were able to isolate some 11, it was racemic.
The reaction temperature could be lowered to 80 °C when
using the more active trifluoroacetate salt 16f [see
Section 3 below for more details]. In this case, the
The absolute configuration was determined based on
CSP-HPLC comparison between the scalemic 15 obtained
via hydrogenation of the formal cycloadduct 11 and the
(21) Yamamoto, Y.; Hoshino, J.; Fujimoto, Y.; Ohmoto, J.; Sawada,
S. Synthesis 1993, 298.
(22) Attempts to dihydroxylate 11 only led to the isolation of the
osmate ester A, but we were not successful in securing an X-ray
structure of A.
(23) Nelson, N. A.; Tamura, Y. Can. J. Chem. 1965, 43, 1323.
(24) Toy, M. S.; Price, C. C. J. Am. Chem. Soc. 1960, 82, 2613.
4250 J. Org. Chem., Vol. 70, No. 11, 2005

A Formal [3 + 3] Cycloaddition Reaction. 5
TABLE 2. Chiral Piperidines as Catalysts
TABLE 4. Synthetic Scope
entry
1
piperidine catalyst
16ac,d R ) COOH
X ) OAc
mol % yielda (%) er (R:S)b
40
45
48:52
47:53
48:52
47:53
49:51
48:52
2
3
4
5
6
16bc
16ce
R ) COOH
X ) L-tartrate
R ) COOMe
X ) Cl
40
53
50
65
16dd R ) CH2OTBDPS
X ) OAc
100
40
42
16ee
R ) C(Ph)2OH
X ) Cl
R ) C(Ph)2OH
X ) O2CCF3
∼15f
62
16f
40
a Isolated yields. b Determined by CSP HPLC. c For its synthe-
sis, see ref 25a. d The acetate salt was generated in situ via
addition of 1.0 equiv of AcOH to the reaction mixture. e See ref
25a,b. f The reaction was carried out at 130 °C for 14 h. The
product 11 was isolated at ca. 50% conversion.
TABLE 3. MacMillan’s and Jørgensen’s Amine Salts
a Isolated yields. b Determined by CSP-HPLC. c E/Z mixtures
were used. d Determined by LC/MS (see ref 20).
[entry 5]. In addition, for reactions catalyzed with 17a-
c, a significant amount of decomposition occurred, thereby
leading to lower yields [entries 1-3].
entry
catalyst
mol %
yielda (%)
er (R:S)b
2.2. Synthetic Scope. On the basis of these results,
the acetate salt of R,R-diphenyl-2-pyrrolidinemethanol 9e
appears to be the most efficient catalyst to promote a
modestly enantioselective intramolecular formal aza-[3
+ 3] cycloaddition reaction of vinylogous amide 8. Thus,
various other vinylogous amides 20-24 were prepared
and employed in the formal aza-[3 + 3] cycloaddition
catalyzed by 9e. As illustrated in Table 4, all reactions
proceeded to give the respective cycloadducts 25-29 in
modest to good yields and comparable enantioselectivity
with the exception of vinylogous amide 24, bearing a
3-carbon tether, which gave only 20% ee [entry 5].
3. Solvent, Counteranion, and Temperature Ef-
fects. 3.1. Solvent Effect. We subsequently pursued a
series of studies to investigate what other factors could
effect or improve the enantioselectivity of this formal
cycloaddition. Toward this goal, the reaction of vinylogous
amide 8 was carried out in different solvents in the
presence of 1.0 equiv of the amine salt 9e [Table 5]. The
reaction was kept at room temperature for 1 h, and was
then heated to 80 °C for 2 h. At this point complete
consumption of 8 was observed by TLC. Only in the case
of acetic acid [entry 6], no formation of the desired
cycloadduct 11 was detected besides decomposition of
starting material. Under these conditions, there appears
to be no significant dependence of the enantioselectivity
1
2
3
4
5
6
7
17ac,d
50
50
16e
27
22
57
53
48
63
50:50
50:50
58:42
60:40
63:37
62:38
50:50
17bf
17cf
50
18g
100
100
100
100
19ag
19bg,h
19cg,h
a Isolated yields. b Determined by CSP-HPLC. c For its synthe-
sis, see ref 16i. f The reaction was carried out at room temperature
for 30 h. e Significant decomposition observed. f See ref 16d. g See
ref 18g. h The acetate salt was generated in situ by addition of
1.0 equiv of AcOH to the reaction mixture.
cycloaddition of 8 went to completion in 2 h and 11 was
isolated in good yield [entry 6]. However, the product
remained racemic. These results suggest that chiral
piperidines are even less useful as chiral catalysts than
chiral pyrrolidines for this particular reaction.
To continue our efforts, we turned to MacMillan’s
chiral amine salts 17a-c16d,i [Table 3] as well as those
developed by Jørgensen [18 and 19a-c].18g These amines
are known to catalyze a variety of organic transforma-
tions in a highly enantioselective manner.15-19 As sum-
marized in Table 3, the desired cycloadduct 11 was
isolated in all cases, but the highest enantiomeric ratio
was only 63:37 when using C2-symmetric catalyst 19a
J. Org. Chem, Vol. 70, No. 11, 2005 4251

Gerasyuto et al.
TABLE 5. Solvent Effect
TABLE 7. Temperature Effect
entry
solvent
yielda (%)
er (R:S)b
1
2
3
4
5
6
EtOAc
CHCl3
EtOH
acetone
THF
60
57
42
53
48
NA
81:19
76:24
81:19
84:16
83:17
decomp
entry cat. (mol %) temp (°C) time (h) yielda (%) er (R:S)b
1
2
3
4
9f (40)
rt
10
6
14
3
43
60
16
46
75:25
81:19
23:77
27:73
AcOH
9e (50)
10c (40)
10c (40)
80
rt
a Based on 1H NMR. b Determined by CSP-HPLC.
80
a Isolated yields. b Determined by CSP-HPLC.
TABLE 6. Counteranion Effect
of the amine salt to dissociate to its “free amine” and “free
acid” can exert an impact on the rate of the iminium salt
formation. The low reactivity of the chloride salt 9g can
be attributed to its higher resistance toward dissociation,
or 9g is simply a tighter ion-pair compared to the acetate
salt 9e and the trifluoroacetate 9f. At the same time, the
formation of R,â-unsaturated iminium salt from the
aldehyde and the “free amine” is also promoted by
protonation of the carbonyl group via the “free acid”.
Consequently, increased reactivity of the trifluoroacetate
salt 9f from the acetate salt 9e can be attributed to a
higher acidity of trifluoroacetic acid.
entry
catalyst
time (h)
yielda (%)
er (R:S)b
1
2
3
9e: X ) OAc
9f: X ) O2CCF3
9g: X ) Clc
1.5
1
12
61
58
66
80:20
79:21
76:24
a Isolated yield only. b Determined by CSP-HPLC. c For its
synthesis, see ref 26.
3.3. Temperature Effect. Finally, we examined the
temperature effect as shown in Table 7. Vinylogous amide
8 was reacted with chiral amine salt 10c and reactive
trifluoroacetate 9f at room temperature, and the result-
ing enantiomeric ratios were compared with reactions
carried out at 80 °C. The reaction took much longer at
room temperature and led to lower yields of 11 [entries
1 and 3], and no significant increase in enantioselectivity
was observed in either case.
4. A Proposed Mechanistic Model. 4.1. Racemiza-
tion Possibility and Enal Geometry. As indicated in
the Introduction, the intermolecular aza-[3 + 3] cycload-
dition undergoes a pericyclic ring-closure of 1-azatriene
intermediates en route to the formation of 1,2-di-
hydropyridines.5b Although intramolecular reactions pro-
ceed through a different mechanism altogether,5b,6 we still
questioned if the modest enantioselectivity obtained from
these chiral amine catalyzed cycloadditions could be a
result of racemization of the final cycloadduct via the
same 6π-electron electrocyclic ring-opening and ring-
closure through the 1-azatriene intermediate 30 [Table
8].
on solvent. The observed enantiomeric ratio stayed at
about 80:20, which was initially found for reactions
carried out in EtOAc [see also entry 1]. Only a slight
increase in selectivity was detected when acetone was
used [entry 4].
3.2. Counteranion Effect. We then examined the
possibility of a counteranion effect in the case of the most
effective amine, R,R-diphenyl-2-pyrrolidinemethanol. To-
ward that goal, vinylogous amide 21 was allowed to react
with three different amine salts 9e-g under standard
reaction conditions [Table 6].
In all cases, the desired cycloadduct 26 was isolated
in good yield, but there was again no significant depen-
dence on the nature of the counteranion. However, there
was a noticeable difference in reaction rate between the
three catalysts. It turned out that the trifluoroacetate salt
9f was the most active with complete conversion of 21 to
26 in 1 h [entry 2]. The reaction with the acetate salt 9e
took a longer time to complete [entry 1], while the
chloride salt 9g was found to be the least active [entry
3].
The observed difference in reactivity can be rational-
ized in two possibilities. First, the chloride salt 9g is not
as soluble in EtOAc as the acetate salt 9e or the
trifluoroacetate salt 9f. Second, a more important pos-
sibility would be that the difference in reactivity is
related to the dissociation capacity of the respective
amine salts.
To test this possibility, an enantiomerically enriched
sample of cycloadduct 11 was heated in toluene-d8. After
3 h at 85 °C, no loss of ee was detected [entry 2]. More
significantly, even after 11 was heated at 130 °C for 12
h, no racemization had occurred [entry 3]. Ab initio
calculations27 indicate 1-azatriene 30 to be destabilized
(25) (a) Shiraiwa, T.; Shinjo, K.; Kurokawa, H. Bull. Chem. Soc. Jpn.
1991, 64, 3251. (b) Portoghese, P. S.; Pazdernik, T. I.; Kuhn, W. L.;
Hite, G.; Shaf’ee, A. J. Med. Chem. 1968, 11, 12.
(26) Aggarwal, V. K.; Lopin, C.; Sandrinelli, F. J. Am. Chem. Soc.
2003, 125, 7596.
Since both “free amine” and “free acid” are needed in
a synergistic manner to generate the R,â-unsaturated
iminium salt from the corresponding aldehyde, the ability
4252 J. Org. Chem., Vol. 70, No. 11, 2005

A Formal [3 + 3] Cycloaddition Reaction. 5
TABLE 8. Racemization via Pericyclic Ring-Opening
SCHEME 2. Detection of the Key Intermediate
entry
time (h)
temp (°C)
er (R:S)b
1
2
3
0
3
12
rt
85
130
27:73
26:74
27:73
a Determined by CSP-HPLC.
TABLE 9. Impact of Enal Geometry
the enantiomeric induction is already determined. Pro-
tonation of these two enamines should give the respective
iminium salts R-34b and S-34b, and a subsequent C-1,2-
addition should furnish the second bond in this formal
cycloaddition sequence, leading to tricyclic intermediates
R-35 and S-35.
Finally, given the fact that we were able to detect the
diastereomeric mixture [see 32 in Scheme 2] related to
R-35 and S-35, regeneration of the chiral amine via
elimination is likely to be slow. This makes intuitive
sense, as generation of vinyl iminium salts R-36 and S-36
could be uphill energetically. Subsequent tautomerization
of R-36 and S-36 should lead to the enantiomeric 1,2-
dihydropyridines R-11 and S-11 concomitant with the
chiral amine salt turning over.
entry
E:Z
yielda (%)
er (R:S)b
1
2
100:0
27:73
60
58
81:19
77:23
a Isolated yields. b Determined by CSP-HPLC.
relative to 11 by ∼31 kcal/mol. Since the activation
barrier would be even higher, we believe that the 6π-
electrocyclic ring-opening process is not thermally acces-
sible under our standard reaction conditions.
To examine the influence of the enal geometry on
enantioselectivity, a mixture of E/Z-isomers of vinylogous
amide 8 was submitted to standard reaction conditions
using 9e [Table 9]. The product 11 was isolated in
comparable yield and its enantiomeric purity [entry 2]
was virtually the same as in the reaction of pure trans-
isomer [entry 1]. We believe that a facile isomerization
of Z-enal to the more stable E-isomer takes place under
the reaction conditions, thereby rendering the enal
geometry irrelevant to the asymmetric induction.
4.2. Detection of the Key Intermediate. When
vinylogous amide 8 was reacted with 1.0 equiv of the
amine salt 9e at room temperature in EtOH, the forma-
tion of the intermediate 32 was detected by 1H NMR and
mass spectroscopy [Scheme 2]. Interestingly, only two of
the four possible diastereomers were seen with a ratio
of 82:18, and heating of the intermediate 32 at 80 °C led
to the formation of 11 with an enantiomeric ratio of
81:19. This experimental outcome suggests that the
stereochemical induction likely occurs during the N-1,4-
addition step.
4.4. Computational Study. To further probe the
origin of the observed enantioselectivity as well as the
difference in the stereochemical outcomes from cycload-
dition reactions carried out with C1- and C2-symmetric
amine salts, computational studies were performed.
Foremost, when an iminium salt such as 33 [see Figure
3] is generated with a C1-symmetric catalyst such as 9e,
two geometric isomers can be formed. Ab initio calcula-
tions27 with a simplified model system (crotonaldehyde)
suggest that the formation of “cis”-isomer 37 is disfavored
by 3.5 kcal/mol relative to “trans”-isomer 38, presumably
due to allylic strain present in 37 [Figure 4]. Therefore,
to simplify the computational analysis, we assumed the
exclusive formation of the “trans”-isomer of the iminium
salt.
To understand the preferred formation of the R-
enantiomer of 11 when using the S-isomer of C1-sym-
metric amine salt 9e, we calculated the stabilities of the
two diastereomeric enamines 39 and 40 [Figure 5], which
are products after the initial N-1,4-addition step [see
Figure 3]. Using a Monte Carlo simulation in the
Spartan′02 program,28 we found that enamine 39, which
4.3. A Proposed Mechanism. With all of the above
experimental findings, we propose the following mecha-
nism for our enantioselective intramolecular aza-[3 + 3]
cycloaddition reaction [Figure 3]. The initial chiral imi-
nium salt 33 can undergo N-1,4-addition in either a Re-
or Si-facial approach toward the â-carbon, leading to two
diastereomeric enamines R-34a and S-34a. At this stage,
(28) Spartan′02, PC version; Wavefunction, Inc.: 18401 Von Karman
Avenue, Suite 370, Irvine, CA 92612; 2002.
(29) Klumpp, D. A.; Aguirre, S. L.; Sanchez, G. V.; de Leon, S. J.
Org. Lett. 2001, 3I, 2781.
(30) Aggarwal, V. K.; Sandrinelli, F.; Charmant, J. P. H. Tetra-
hedron: Asymmetry 2002, 13, 87.
(31) Willems, J. G. H.; Hersmis, M. C.; de Gelder, R.; Smits, J. M.
M.; Hammink, J. B.; Dommerholt, F. J.; Thijs, L.; Zwanenburg, B. J.
Chem. Soc., Perkin Trans. 1 1997, 963.
(27) Calculations were carried out in Spartan′02 software on a Dell
Precision 650 Dual Xeon (2.00 GHz) workstation.
J. Org. Chem, Vol. 70, No. 11, 2005 4253

Gerasyuto et al.
FIGURE 3. A proposed mechanism for the enantioselective formal aza-[3 + 3] cycloaddition.
FIGURE 4. Iminium salt stereoisomers.
FIGURE 6. C1-symmetric amine salts: relevant transition
states.
diphenylhydroxymethyl substituent on the pyrrolidine
ring and the cyclohexenone moiety in Pro-S TSq-42
[Figure 6].
To be cautious, we looked at all six other possible chair
transition states that differ in the relative localized
positions of the nitrogen substituent and the enal func-
tionality. Structures for four pairs of diastereomeric TSq-
[41-48] along with their energies are shown in Figure
7. None of the other transition states are lower in energy
than Pro-R TSq-41. In particular, TSq-43 and TSq-44
having the nitrogen substituent axial are destabilized in
comparison with Pro-R TSq-41 due to more severe 1,3-
diaxial interactions. A similar interpretation can be made
in the case of TSq-45 and TSq-46 where both side chains
are axial. Finally, placing both substituents equatorial
as shown in TSq-47 and TSq-48 leads to destabilization
of the transition state structures due to the gauche
interaction. On the basis of these computational results,
it can be postulated that the asymmetric intramolecular
formal aza-[3 + 3] cycloadditions promoted with C1-
symmetrical catalysts such as 9e proceed predominantly
through Pro-R TSq-41 giving rise to R-11 as the major
enantiomer.
FIGURE 5. C1-symmetric amine salts: enamine stability and
conformation.
would give R-11, is ∼1 kcal/mol more stable than the
corresponding diastereomer 40 leading to the S-enanti-
omer.
Based on this product stability analysis and the
preferred conformations shown for 39 and 40, the corre-
sponding transition states [TSq] 41 and 42 for the N-1,4-
addition step of the reaction were identified. It is
noteworthy that the localization of the transition state
structures was proven by the presence of one negative
eigenvalue and the corresponding imaginary frequency,
which reflect formation of the N-C bond. Models and
their respective energies are shown in Figure 6.
Both transition state structures represent a chair form
with the nitrogen substituent being equatorial and the
enal moiety being axial. Calculations indicate that Pro-R
TSq-41 is more stable relative to the corresponding
diastereomeric Pro-S TSq-42 by ∼1.41 kcal/mol. This is
consistent with calculations obtained for the respective
enamines 39 and 40. This energy difference can be
rationalized by the steric interaction between the bulky
A similar approach was used to rationalize the ob-
served formation of S-enantiomer of 11 when C2-sym-
metric amine salts are employed. To reduce the compu-
4254 J. Org. Chem., Vol. 70, No. 11, 2005

A Formal [3 + 3] Cycloaddition Reaction. 5
FIGURE 7. C1-symmetrical amine salts: all possible transition states.
SCHEME 3. The Effect of the r-Methyl Substituent
When the corresponding TSq-51 and TSq-52 were
localized, the same trend in relative stabilities was
observed. Pro-S TSq-52, which has the nitrogen substitu-
ent equatorial and the enal functionality axial, is lower
in energy than Pro-R TSq-51 where now both side chains
are axial [Figure 8]. Six other chairlike transition states
were also modeled but none of them show greater
stability than Pro-S TSq-52.
Most likely the steric interaction between one of the
ester groups on 10b and the piperidine ring is responsible
for destabilization of Pro-R TSq-51 in comparison with
Pro-S TSq-52 [Figure 8]. An increase in the substituent
size as shown in the amine salt 10c would further
disfavor Pro-R TSq-51, thereby leading to enhanced
enantioselectivity in favor of S-11. It can be suggested
that this particular steric interaction, which does not
exist in C1-symmetric amine salts, is responsible for the
opposite stereochemical outcome.
FIGURE 8. C2-symmetrical amine salts: enamine stability
and relevant transition states.
tational time, all calculations were performed with 10b
bearing two methyl ester substituents. It should be noted
that in this case only one geometric isomer of the iminium
salt exists due to the C2-symmetric nature of 10b.
Stabilities and conformations for the two diastereomeric
enamines 49 and 50 were again found through Monte
Carlo simulation in Spartan′02 [Figure 8].20 The relative
stability of conformers is now reversed compared to the
C1-symmetric case. Specifically, enamine 50, which would
yield the S-11, is ∼1 kcal/mol more stable than 49, which
would lead to the R-enantiomer.
J. Org. Chem, Vol. 70, No. 11, 2005 4255

Gerasyuto et al.
SCHEME 4. Tuning of the Amine Salts
SCHEME 5. “Rigid” Amine Salts
a For the synthesis see ref 31. b The acetate salt was generated
in situ by addition of 1.0 equiv of AcOH.
enantioselectivity. To ensure that the presence of the
tertiary hydroxyl group did not impede our efforts, the
amine salt 9i was employed but it also led to the racemic
11.
a Commercially available. b The acetate salt was generated in
situ by addition of 1.0 equiv of AcOH to the reaction mixture. c For
its synthesis see ref 29. d See ref 30.
We then examined the C2-symmetric 10d amine salt
bearing two bulky diphenylhydroxymethyl substituents,
but again, the desired cycloadduct 11 was isolated as a
racemate. These experimental results intuitively suggest
that more bulky catalysts could simply impede the
formation of the R,â-unsaturated iminium salt and
increase the opportunity for background reactions.
Finally, we tried to use more rigid chiral amine salts,
presuming that the conformational flexibility of pyrroli-
dine-based amine salts might have compromised the
enantioselectivity. To that end, amine salts 58 and 59
were tested in the cycloaddition of 8 but neither was
useful [Scheme 5]. It is nonetheless noteworthy that both
amine salts were efficient in promoting the reaction.
5. Final Attempts. One of the key assumptions in our
mechanistic model is the exclusive formation of the
“trans”-isomer of the iminium salt 38 [see the box in
Scheme 3] for the case of C1-symmetric amine salts. To
validate this assumption, vinylogous amide 53 containing
a methyl substituent at the R-position was synthesized
[Scheme 3]. Since the formation of the “cis”-isomer of
iminium salt 57 [see the box] intuitively should be further
disfavored in this case compared to 37, the enantio-
selectivity of the cycloaddition should stay the same if
not improve.
It was found that vinylogous amide 53 would undergo
formal cycloaddition only if the most active trifluoro-
acetate salt 9f was used [Scheme 3]. No product forma-
tion was detected after running the reaction overnight
with chloride salt 9g or acetate salt 9e. In the case of 9f,
only 40% conversion was observed after 14 h at 80 °C,
and after hydrogenation of initial cycloadduct 54, quino-
lizidine 55 was isolated as a single isomer. Relative
“trans”-stereochemistry was assigned based on the cou-
pling constants analysis as well as nOe data. But to our
surprise, CSP-HPLC revealed that quinolizidine 55 was
nearly racemic.
Due to the low-yielding and sluggish nature of this
reaction, it is not clear what caused the drop in the
enantioselectivity and we did not follow up in this
pursuit. It could be that the background reaction gained
momentum given that formation of the iminium salt is
now more difficult with severe allylic strain present in
both 56 and 57.
Conclusion
We have described here a detailed study using chiral
secondary amine salts to promote enantioselective in-
tramolecular formal aza-[3 + 3] cycloaddition. The de-
pendence of enantioselectivity on the structural feature
of these chiral amines has been thoroughly investigated,
and we found a very interesting reversal of the stereo-
chemistry in the respective formal cycloadducts obtained
using C1- and C2-symmetric amine salts. Effects of
solvents, counteranions, and temperatures on the enan-
tioselectivity were also investigated and appear to have
no real impact. On the basis of these experimental results
and some semiempirical calculations, a unified mecha-
nistic model is proposed.
Acknowledgment. R.P.H. thanks NIH [N.S.38049]
for funding.
We also attempted to modify the structures of the most
efficient catalysts 9e and 10c in a rational manner to
increase the enantioselectivity of the cycloaddition reac-
tion. As shown in Scheme 4, the use of the amine salt
9h, which has more extensive naphthyl-substituents
relative to 9e, did not lead to any improvement in the
Supporting Information Available: Experimental pro-
cedures, characterization data for all new compounds, 1H NMR
spectra, nOe data, and HPLC traces. This material is available
free of charge via the Internet at http://pubs.acs.org.
JO050171S
4256 J. Org. Chem., Vol. 70, No. 11, 2005