Evidence of asymmetric autocatalysis in organocatalytic reactions

Article abstract of DOI:10.1002/anie.200603517

(Chemical Equation Presented) When product and catalyst are the same: The chiral product 3 of an asymmetric Mannich reaction (see scheme), investigated under various experimental conditions and by calculations, is shown to act as a chiral catalyst for its

Full text of DOI:10.1002/anie.200603517

Angewandte
Chemie
DOI: 10.1002/anie.200603517
Asymmetric Autocatalysis
Evidence of Asymmetric Autocatalysis in Organocatalytic Reactions**
Michael Mauksch,* Svetlana B. Tsogoeva,* Irina M. Martynova, and Shengwei Wei
Dedicated to Professor LutzF. Tiet ze on the occasion of his 65th birthday
[
8]
Asymmetric autocatalysis is the process of automultiplication
of a chiral compound in which the chiral product acts as a
Mannich reaction [Eq. (1)] under various reaction condi-
tions. The choice fell on an example that allows for specific
product–substrate interactions through hydrogen-bonded
complexes of the product molecules with the prochiral
substrate.
[
1]
chiral catalyst for its own formation. Such reactions offer
striking advantages as the chiral product does not need to be
separated from the chiral catalyst and as no other chiral
catalyst than the product itself is involved. Several examples
are known, all of which involve organometallic reagents and
[
1–3]
are hence restricted to certain reaction types;
the most
prominent of these is the Soai reaction for its ability to
amplify a tiny initial enantiomeric excess of a chiral pyrimidyl
alkanol in the presence of iPr Zn to almost enantiomeric
2
[
3c]
purity. A catalytic asymmetric autoinductive aldol reaction
in the presence of Ti (binol) complexes was described by
Szlosek and Figad›re. Whereas a purely organic specimen of
product catalysis is known, such as the autocatalytic aldol
reaction, asymmetric examples of such reactions have not
yet been reported.
IV
First, we prepared the product catalyst with l-proline as
external catalyst in 98% ee (S) and with d-proline in 99% ee
[
4]
[
8,9]
(R) enantiomeric purity.
Experiments carried out under
[
5]
various conditions (different concentrations of educt; differ-
ent solvents; different ee values, absolute configuration, and
loadings of catalyst; variable temperatures and reaction
times; see Table 1) revealed to our surprise that the product
is formed in nearly the same enantiomeric purity as that of the
initially added product catalyst at 15 mol% (Table 1,
However, Sievers and von Kiedrowski described a tem-
plate-based mechanism of autocatalytic self-replication of
[
6]
(
naturally asymmetric) oligonucleotide strands. Moreover,
Bolm and co-workers cited a fully organocatalytic example of
enantioselective autoinduction in the production of a chiral
cyanhydrine, catalyzed by a cyclic dipeptide. The product and
the external catalyst together form here a more efficient
catalytic species, which resulted in higher ee values and
[
10]
entries 1–4, 9–12).
Higher effective ee values of up to
96% were obtained with very high catalyst loadings (Table 1,
[
7]
yields.
Table 1: Mannich reaction catalyzed by product 3 [Eq. (1)].
These results led us to pose the question of whether the
product alone could in principle act as an inductor of chirality
and in any asymmetric organic reaction. Reversible reactions
appear to be badly suited for such an undertaking, as
racemization might occur by the reverse reaction. Addition-
ally, if the reaction-accelerating feature of the product is
poorly developed, competition with the “uncatalyzed” reac-
tion further limits the achievable enantiomeric excesses.
Herein, we demonstrate for the first time that even in
quite ordinary (and reversible) asymmetric organic reactions
Entry
Catalyst 3
Loading ee [%]
mol%]
T
[8C]
Reaction
time
[days]
Product 3
Yield
ee [%]
(red. ee)
[
a]
[b]
[
(config.)
[%]
[
c,d]
1
2
3
4
5
6
7
8
9
0
1
2
3
15
15
15
15
5
98 (S)
98 (S)
98 (S)
98 (S)
98 (S)
98 (S)
98 (S)
98 (S)
29 (S)
55 (S)
75 (S)
99 (R)
99 (R)
99 (R)
98 (S)
98 (S)
25
25
25
25
25
25
25
25
0
2
4
6
4
4
6
4
6
4
4
4
4
4
4
4
4
20
48
45
53
38
32
33
32
25
27
27
42
40
38
39
41
87 (79, S)
85 (81, S)
91 (89, S)
86 (83, S)
28 (19, S)
23 (11, S)
29 (27, S)
23 (20, S)
25 (22, S)
56 (57, S)
69 (66, S)
94 (92, R)
96 (94, R)
95 (90, R)
94 (90, S)
95 (91, S)
[c,d,e]
[c,d]
[
[
[
[
[
[
c,f]
c,d]
c,d]
c,d]
c,d]
d,g]
5
1
1
(
as opposed to reactions involving organometallic species) the
15
15
15
15
30
50
30
50
chiral product alone could act as a catalyst with high
stereoselectivity. As an example, we chose the asymmetric
[d,g]
[d,g]
1
1
1
1
0
0
[
[
[
[
[
c,d]
c,d]
c,d]
c,d]
c,d]
25
25
25
25
25
[
*] Dr. M. Mauksch, Prof. Dr. S. B. Tsogoeva, Dr. I. M. Martynova, S. Wei
Institut für Organische und Biomolekulare Chemie
Universität Göttingen
14
15
16
Tammannstrasse 2, 37077 Göttingen (Germany)
Fax: (+49)551-39-9660
E-mail: stsogoe@gwdg.de
[
[
(
a] Yields after subtraction of the initially added product catalyst.
b] Enantioselectivities were determined by chiral HPLC analysis
Daicel Chiralpak OD) in comparison with authentic racemic material.
[
**] The authors gratefully acknowledge generous financial support from
Reduced ee values are given in parentheses (see Ref. [10]). [c] Educt
concentration: 0.25 molL . [d] Acetone as solvent. [e] A repetition of
the Deutsche Forschungsgemeinschaft.
À1
Supporting information for this article is available on the WWW
under http://www.angewandte.org or from the author.
experiment 2 with a different substrate batch produced a yield of 59%
À1
with 93% ee. [f] DMSO as solvent. [g] Educt concentration: 0.5 molL
.
Angew. Chem. Int. Ed. 2007, 46, 393 –396
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
393

Communications
entries 13–16). These results resemble those of Soai et al. for a
[
2a]
special metal-containing system.
The use of dimethyl
sulfoxide (DMSO) as solvent produces only slightly better
results in terms of both ee values and yields than acetone
[
11]
(
Table 1, entry 2 vs 4).
To see whether the product acts as a catalyst or is merely
stoichiometric, we carried out some experiments with lower
initial concentrations of product catalyst (Table 1, entries 5–
8
). A comparison of these results with varying catalyst
loadings (1–15 mol%) of a Mannich product catalyst with
8% ee (Table 1, entries 2–8) shows that the yield does not
depend very strongly on the amount of the catalyst added or
Scheme 1. a) Proposed general catalytic cycle and b–e) pre-complex
equilibria with educt (E), R, and S product dimers. Free enthalpies of
reaction for the Mannich reaction (DG [kcalmol ]) are based on
9
À1
[
12]
computations (see text).
the reaction time. However, the impact on enantioselectiv-
ities is high. We suppose therefore that asymmetric autoca-
talysis is in operation here, also because of the still sizeable ee
values observed even for very low initial concentrations of
product catalyst. Intriguingly, better yields were obtained
when the reaction was halted after 4 days (Table 1, compare
entries 2 vs 3; 5 vs 6; and 7 vs 8). A comparison of experiments
that are stopped after 6 days with those that are halted after
To elucidate further the existence and role of the species
involved and to possibly corroborate our intuitive mechanistic
proposal (Scheme 1), we carried out density functional theory
(DFT) calculations (Table 2) with the B3LYP functional and
4
days shows a tendency for decrease in product ee values for
Table 2: Computed species at B3LYP/6-31G level for the Mannich
reaction [Eq. (1)].
longer reaction times, indicating the involvement of racemi-
zation, either due to the reversibility of the reaction or to an
enantiomerization of the product (Table 1, compare entries 5
À1 [a]
Entry
Species
Eabs [Hartree]
Erel [kcalmol ]
[
b]
(NIMAG)
[
13]
vs 6; and 7 vs 8).
1
2
3
4
5
6
7
8
9
cis-2
trans-2
(S)-3
À707.18136014
À707.18968948
À900.31395423
À1800.64298375
À1800.64201430
À1607.50385835
À1607.50605890
À1607.51274783
À1607.51796910
À1800.58622694
À1800.59027031
À900.25703468
6.0 (0)
0.0 (0)
0.0 (0)
Next, we studied the influence of the reaction temper-
ature and the initial ee value of the catalyst on the investigated
reaction. Carrying out the reaction at 08C (Table 1, entries 9–
1
[c]
(S)-3·(R)-3
(S)-3·(S)-3
(S)-3·cis-2 (I)
(S)-3·cis-2 (II)
(S)-3·trans-2 (I)
(S)-3·trans-2 (II)
TS for (R)-3
TS for (S)-3
TS for (R,S)-3
3.0 (0)
3.0 (0)
À1
1) with an educt concentration of 0.5 molL in acetone and
[c]
with catalyst ee values of 29%, 55%, and 75% gave the
Mannich product in around 25–27% yield with 25%, 56%,
and 69% ee, respectively. Hence, yields are more strongly
influenced here by reaction temperature than by catalyst ee
values (Table 1, entry 2 vs 9–11). We expected from the
inspection of the Gibbs–Helmholtz equation that the result-
ing ee values would be higher at a lower temperature.
However, the values achieved relative to the initial values are
not higher than in some of the experiments carried out at
room temperature.
9.1 (0)
9.4 (0)
3.8 (0)
0.0 (0)
[
d]
[e]
1
1
1
0
1
2
31.3 (1)
29.5 (1)
32.2 (1)
[
d]
[f]
[g]
[h]
[a] Inclusive ZPEand thermal corrections. [b] Number of imaginary
frequencies. [c] With respect to two monomeric (S)-3 molecules. [d] For
catalyzed reaction with acetone based on the structure (S)-3·trans-2 (II)
(Figure 1). [e] Relative to entry 4. [f] Relative to entry 5. [g] For uncata-
The catalyst showed a high efficiency in transferring
chirality from itself onto the chemically identical product
lyzed reaction of acetone with trans-2 without product catalyst.
[h] Relative to entry 3.
(
Table 1). The transfer of chirality appears less than optimal,
apparently because of the competing uncatalyzed reaction
which can account for significant contributions to the
achievable yields. To explain the fairly good induction of
chirality, we propose a catalytic cycle (Scheme 1) which
involves hydrogen-bonded pre-complexes between the sub-
strate (E) and product and which could be attacked by the
enol tautomer of acetone.
[
16]
the 6-31G basis set with the Gaussian03 quantum chemis-
try package. Full optimizations were carried out for minima
and transition-state structures of the Mannich reaction
involving the S-configured product catalyst (Figure 1). Fre-
quency computations had been done to include zero-point
energy (ZPE) and thermal corrections.
In principle, this Mannich reaction might also proceed by
an aminocatalytic mechanism that involves the formation of
We found that both hetero- (R,S) and homochiral (S,S)
product dimers are less stable with respect to the monomers,
with almost equal energies (Table 2, entry 3 vs 4 and 5). The
formation of dimer competes with the complexation of the
product with the substrate, which is disfavored by only
[
14]
an enamine from the product and acetone. However, the
evidence from the ESI mass spectra was somewhat incon-
[
15]
clusive.
The observation that even the aldol reaction
À1
between acetone and p-nitrobenzaldehyde gave ee values of
the isolated product of 71–75% (3–27% yield, 6 days) when
the product with 76% ee was initially added in various
amounts led us to the conclusion that the enamine mechanism
probably does not explain the results reported here.
3.2 kcalmol with respect to the separate educt and product
À1
molecules. The latter energy lies a mere 0.3 kcalmol higher
than either dimerization energy.
It is implicated that the product is formed from the trans-
educt/S-product complex (Table 2, entry 9). The resulting
3
94
www.angewandte.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2007, 46, 393 –396

Angewandte
Chemie
entries 10 and 11; Figure 1), or, alternatively, that could
thermodynamically favored; Scheme 1d) “annihilate” to
(
release the substrate and give more inactive meso product
dimers.
The amount of enantiomeric excess is limited by the
stereospecificity of the product catalyst (i.e. the branching
ratio, Scheme 1) and the extent of the racemizing, slightly
disfavored “uncatalyzed” reaction (Table 2, entry 12). The
mechanism we anticipate here is fully general and, in our
opinion, not restricted to the organic reaction discussed
herein.
In conclusion, we have demonstrated here for the first
time that the product alone might play an important role for
yields and achievable enantiomeric excesses in asymmetric
organocatalysis. For the Mannich reaction under various
reaction conditions, we have shown that the product can be a
promising catalyst for its own formation.
Our results pose challenging questions, particularly for
organic chemists: What is the role of the product on the yields
and ee values in those enantioselective organocatalytic
reactions reported in the literature? Could the product
itself indeed be an effective catalyst in organocatalytic
reactions? If so, then it would remove the necessity to
separate the product from the catalyst, which could save costs
in commercial applications. In principle, it might involve the
condition that the catalyst could be self-multiplied to any
extent. What remains is to find the proper conditions for
higher ee values and yields. Extensions of the concept to other
reactions with active product catalysis are presently being
carried out in our laboratory.
Figure 1. Transition-state structures for the formation of S (TS-A) and
R (TS-B) enantiomers of the Mannich product.
product dimer is most probably homochiral (Scheme 1a, main
branch) because the transition-state structure for the forma-
tion of S-configured product (entry 11, Table 2) is 1.8 kcal
Received: August 28, 2006
Revised: October 16, 2006
Published online: December 5, 2006
À1
mol lower in energy with respect to the respective transition
state for the formation of R-configured product. Heterochiral
dimers might result from the reaction when the absolute
configurations of product and product catalyst are not
matching (i.e. less than 100% stereospecificity; side branch
not shown). The regeneration of educt–product complexes
from free substrate molecules and product dimers is disfa-
Keywords: asymmetric autocatalysis · density functional
.
calculations · Mannich reaction · organocatalysis ·
transition states
À1
vored by 3.4 kcalmol (Scheme 1b). Heterochiral complexes
are in equilibrium with a racemic mixture of educt–product
complexes (Scheme 1d). Homochiral dimers and monomers
of product are most likely not directly involved in the catalytic
step. Entropy-favored racemization can occur by the uncata-
lyzed back reaction (Scheme 1e).
[
[
1] K. Soai, T. Shibata, I. Sato, Acc. Chem. Res. 2000, 33, 382.
2] a) K. Soai, S. Niwa, H. Hori, J. Chem. Soc. Chem. Commun.
1990, 982; b) T. Shibata, K. Choji, H. Morioka, T. Hayase, K.
Soai, Chem. Commun. 1996, 751.
[3] a) S. Tanji, Y. Kodaka, A. Ohno, T. Shibata, I. Sato, K. Soai,
Tetrahedron: Asymmetry 2000, 11, 4249; b) I. Sato, D. Omiya, T.
Saito, K. Soai, J. Am. Chem. Soc. 2000, 122, 11739; c) I. Sato, H.
Urabe, S. Ishiguro, T. Shibata, K. Soai, Angew. Chem. 2003, 115,
Hydrogen transfer from the enol oxygen to the nitrogen of
the Schiff base initiates the approach of the methylene carbon
(
(
C ) from the enol towards the carbon (C ) of the aldimine
a b
Figure 1). The reaction coordinate is represented by a six-
329; Angew. Chem. Int. Ed. 2003, 42, 315; d) K. Soai, T. Shibata,
H. Morioka, K. Choji, Nature 1995, 378, 767; e) for a review, see:
K. Soai, I. Sato, Chirality 2002, 14, 548.
membered transition state with a C –C distance of 2.845 Š
a
b
(
TS-A, Figure 1) to 2.847 Š (TS-B, Figure 1). The respective
[
4] M. Szlosek, B. Figad›re, Angew. Chem. 2000, 112, 1869; Angew.
Chem. Int. Ed. 2000, 39, 1799.
computed dipole moments of transition-state structures for
the formation of S and R enantiomers (3.97 and 3.12 Db,
respectively) offer a possible explanation for the influence of
solvent polarity (Table 1, entry 2 vs 4).
Product dimers are in equilibrium with a small amount of
educt–product pre-complexes (Scheme 1b,c; Table 2,
entry 9), from which either more product could form, for
example, via transition states TS-A or TS-B (Table 2,
[5] H. Fujisawa, T. Nakagawa, T. Mukaiyama, Adv. Synth. Catal.
004, 346, 1241.
[
2
6] D. Sievers, G. von Kiedrowski, Nature 1994, 369, 221.
7] a) C. Bolm, F. Bienewald, A. Seger, Angew. Chem. 1996, 108,
[
1767; Angew. Chem. Int. Ed. Engl. 1996, 35, 1657; b) H. Danda,
H. Nishikawa, K. Otaka, J. Org. Chem. 1991, 56, 6740.
8] A. Córdova, W. Notz, G. Zhong, J. M. Betancort, C. F. Bar-
bas III, J. Am. Chem. Soc. 2002, 124, 1842.
[
Angew. Chem. Int. Ed. 2007, 46, 393 –396
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
395

Communications
[
9] We cannot preclude that the product may have played a role
during its formation in the presence of proline as well.
[14] For reviews on asymmetric aminocatalysis, see: a) B. List, Synlett
2001, 1675; b) B. List, Chem. Commun. 2006, 819.
[15] A small ion peak at m/z 329.3 [M+Na+H] could be interpreted
+
[
10] To assess the effect of asymmetric induction, we have also
included the “reduced” ee values of the product (ee ) after
as originating from the enamine, while another peak at m/z 473.5
[M+H] may indicate the formation of a product–substrate
complex.
p
+
deduction of the effect of the initial catalyst on the ee values of
the isolated product: ee = (n /(n Àn ))·(ee ’À(n ·ee /n )), with
p
1
1
o
p
o
o
1
[
16] Gaussian03, RevisionA.11.4, M. J. Frisch, G. W. Trucks, H. B.
Schlegel, G. E. Scuseria, M. A. Robb, J. R. Cheeseman, V. G.
Zakrzewski, J. A. Montgomery, Jr., R. E. Stratmann, J. C.
Burant, S. Dapprich, J. M. Millam, A. D. Daniels, K. N. Kudin,
M. C. Strain, O. Farkas, J. Tomasi, V. Barone, M. Cossi, R.
Cammi, B. Mennucci, C. Pomelli, C. Adamo, S. Clifford, J.
Ochterski, G. A. Petersson, P. Y. Ayala, Q. Cui, K. Morokuma,
N. Rega, P. Salvador, J. J. Dannenberg, D. K. Malick, A. D.
Rabuck, K. Raghavachari, J. B. Foresman, J. Cioslowski, J. V.
Ortiz, A. G. Baboul, B. B. Stefanov, G. Liu, A. Liashenko, P.
Piskorz, I. Komaromi, R. Gomperts, R. L. Martin, D. J. Fox, T.
Keith, M. A. Al-Laham, C. Y. Peng, A. Nanayakkara, M.
Challacombe, P. M. W. Gill, B. Johnson, W. Chen, M. W. Wong,
J. L. Andres, C. Gonzalez, M. Head-Gordon, E. S. Replogle,
J. A. Pople, Gaussian, Inc., Pittsburgh PA, 2002.
ee = ee value of the initially added product catalyst, ee ’ =
o
p
observed ee value of isolated product, n = product initially
o
added [mol], n = total isolated product [mol].
1
[
11] We employed also CH Cl and toluene as solvents. While the
2
2
reaction in toluene gave only traces of product, the yield in
methylene chloride was 4–6% with an ee value of the isolated
product close to that of the initially added product (98% ee).
12] Traces of p-methoxyaniline have been detected after workup.
Unused starting material could be retrieved in all experiments,
that is, the reaction does not proceed to completion.
13] After keeping the product 3 for one month at À218C, we
observed a decrease in enantiomeric purity from 98% ee to 53%
ee. However, no loss of enantiomeric purity was observed after
[
[
4
days at room temperature.
3
96
www.angewandte.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2007, 46, 393 –396