Highly enantioselective aldol reaction with 2-trimethylsilyloxyfuran: The first catalytic asymmetric autoinductive aldol reaction

Full text of DOI:10.1002/(SICI)1521-3773(20000515)39:10<1799::AID-ANIE1799>3.0.CO;2-Z

COMMUNICATIONS
for 1 hour. Then 1 equivalent of 2a (0.5 mmol) followed by
1.5 equivalents of 1a were added. Butenolides 3a so obtained
were analyzed by H NMR spectroscopy in the presence of
Highly Enantioselective Aldol Reaction with
2-Trimethylsilyloxyfuran: The First Catalytic
Asymmetric Autoinductive Aldol Reaction**
1
chiral europium salt (Eu(hfc)3). It should be noted that, in the
presence of only Ti(OiPr)4, 1a does not add to 2a.[11] The first
experiment shows that 1a added to 2a, with( R)-BINOL2Ti as
the catalyst gave, after 2 hours, 3a withlow d.r. (53:47), a good
ee value (87%) for the major syn product, and moderate 50%
yield (this is quite often encountered in such reactions[13±15]).
Under the same reaction conditions, but at lower temperature
(À788C), the ee value of the syn butenolide 3a was
significantly increased (96%) but to the detriment of the
chemical yield (10%). When (À)-TADDOL was added to the
Ti(OiPr)4 and (R)-BINOL (1:1:1 ratio), a similar yield and
only a slight decrease in the ee value (70 relative to 87%) were
observed. We thought that these conditions may correspond
to a mismatchcase and so performed the same reaction with
(‡)-TADDOL and observed a quantitative reaction with
formation of the same major diastereomer (syn) 3a (d.r. ˆ
70:30) with82% ee. In all cases the absolute configuration of
the major isomer 3a was S,S as determined by comparison
withrelated products. [11] The last two experiments gave good
evidence that the (R)-BINOLTi(À)-TADDOL is less active
than the (R)-BINOLTi(‡)-TADDOL complex which is
probably formed,[13] and that the chiral induction is controlled
by (R)-BINOL. Again a temperature effect was observed and,
in the presence of a 1:1:1 mixture of Ti(OiPr)4, (R)-BINOL,
and (‡)-TADDOL at À788C, 1a added to 2a to afford the
major syn butenolide 3a witha better ee value (90 relative to
82%) but lower chemical yield (40 relative to 99%).
Á
Magali Szlosek and Bruno Figadere*
We report herein the first quantitative and highly enantio-
selective autoinductive aldol reaction (up to 96% ee) between
2-trimethylsilyloxyfuran (TMSOF) 1a and an achiral alde-
hyde 2. The origin of homochirality of biomolecules (such as
a-amino acids and sugars) remains a puzzling question.[1] It is
now well accepted that physical factors like circulary polar-
ized light can induce chirality in molecules through photo-
synthetic procedures.[2, 3] However the ee values so obtained
are very low and cannot be directly correlated withteh
enantiomeric purity of the known biomolecules. Thus, Soai
et al. have shown that amplification of a slight enantiomeric
imbalance in molecules could be due to an asymmetric
autocatalytic system, and this was proved to be the case for
the 1,2-addition of organozinc reagents to 2-methylpyrimi-
dine-5-carbaldehyde.[4] Since then, only catalytic asymmetric
1,2-addition of nucleophiles on aldehydes has been shown to
be either autocatalyzed[4±6] or autoinduced[7±9] by the chiral
product of the reaction, apart from several Diels ± Alder
reactions which were reported to be autoinduced by the cyclic
adduct.[10]
We recently disclosed that 1a adds to aldehyde 2a to afford
the corresponding butenolides 3a withgood selectivity (76 to
90% ee) when the (R)-(1,1'-bi-2-naphthol)2Ti complex ((R)-
BINOL2Ti) was used (Scheme 1).[11] In order to improve both
It was now possible to perform the aldol reaction with a
ªreluctantº aldehyde, heptadecanal 2b, in 55% yield witha
75:25 d.r. and withan ee value of 75% for the major syn
isomer 3b (see supporting information). So far, it is clear that
BINOL is necessary for the aldol reaction to occur, and that
addition of a second aliphatic alcohol (TADDOL for in-
stance) could form an even more reactive complex withboth
titanium(iv) and BINOL. It was therefore important to check
if the product of the reaction could interact with the catalytic
species. This investigation led us to describe the first auto-
inductive aldol reaction (Scheme 1 and Table 1).
Scheme 1. Activators: I ˆ (R)-BINOL, II ˆ (À)-TADDOL, III ˆ (‡)-
TADDOL; 1a: R1 ˆ R ˆ Me; 1b: R1 ˆ Me and R ˆ tBu; 1c: R1 ˆ Phand
R ˆ tBu; 2a ± g: see Table 2.
the chemical yield and the ee value of this reaction, we studied
the use of a second chiral ligand, as an activator, based on the
chiral ligand acceleration concept.[12±15] The catalysts were
prepared in situ at 208C in Et2O by mixing 20 mol% of
Ti(OiPr)4, (R)-BINOL, and the activator I ± III in a 1:1:1 ratio
The first experiment shows that, in the presence of
20 mol% of the catalyst (R)-BINOL2Ti and 5 mol % of
Table 1. 2-TMSOF (1a) addition to octanal (2a) in the presence of the
BINOL2Ti complex at À208C in Et2O.
Entry
activator[a]
yield of 3a [%]
syn:anti ratio
ee of syn
Á
[*] Priv.-Doz. Dr. B. Figadere, Dr. M. Szlosek
(abs. conf.)
Â
Laboratoire de Pharmacognosie, associe au CNRS (BIOCIS)
1
2
3
4
5
(S,S)-3a[b]
(R,R)-3a[c]
(RS,SR)-3a[d]
none[e]
99
99
57
99
90
70:30
70:30
60:40
70:30
60:40[g]
> 96 (S,S)[b]
40 (S,S)
92 (S,S)
70 (S,S)
> 96 (S,S)
Â
Â
Universite Paris-Sud, Faculte de Pharmacie
Â
Ã
rue Jean-Baptiste Clement, 92296 Chatenay-Malabry (France)
Fax : (‡33)1-46-83-53-99
E-mail: Bruno.Figadere@cep.u-psud.fr
none[f]
[**] The authors wish to thank J. C. Jullian for his help in NMR
experiments, Dr. X. Franck for stimulating discussions, and Prof. R.
Hocquemiller for his continuous interest in this study. Thanks are due
to M.J. Vega Capela for the experiments with heptadecanal. L. Tache
and B. Mbeninack are acknowledged for the preparation of the (R,R)-
2 isomer.
[a] 5 mol% solution; [b] (S,S)-3a with82% ee, and thus, the ee value for
the butenolide 3a is 96.88%, as calculated in ref. [8]; [c] (R,R)-3a with
>96% ee; [d] racemic anti butenolide 3a; [e] after 24 hours; [f] with
addition of a mixture of 2-TMSOF (1a) and octanal (2a) in four portions
(ratios of 2a to 1a ˆ 0.032:0.049, 0.093:0.139, 0.125:0.188, and
0.250:0.375 mmol with15, 15, 30, and 60 min. of delay, respectively);
[g] the anti product (4R,5S)-3a withan ee value of 90%.
Supporting information for this article is available on the WWW under
http://www.wiley-vch.de/home/angewandte/ or from the author.
Angew. Chem. Int. Ed. 2000, 39, No. 10
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2000
0570-0833/00/3910-1799 $ 17.50+.50/0
1799

COMMUNICATIONS
(S,S)-3a (82% ee), the reaction was quantitative with an ee
value of 96% determined for all syn butenolide 3a (which
corresponds to 96.9% ee for the newly formed butenolide, see
reference [8] for calculations). In the presence of the minor
enantiomer (R,R)-3a, a muchlower ee value was observed for
3a, as expected (40% ee, entry 2, Table 1). These results
suggested that the major chiral products formed in this
reaction are incorporated into the active chiral catalyst to
form new catalysts which are more active due to a ligand
accelerating effect (leading to higher chemical yields) and
which have a better asymmetric inductive effect (higher ee
values observed). However, when 5 mol% of racemic anti
isomer 3a was introduced into a 20 mol% (R)-BINOL2Ti
catalyst solution followed by 2a and then 1a, the expected
butenolides 3a were obtained in 57% yield as a 60:40 syn:anti
mixture, withan ee value of 92% for the major syn product.
This important result shows that there is not a significant
increase in the chemical yield (57 relative to 50%) or the ee
value (92 in comparison with 87%). This is, furthermore, in
accordance with the result obtained in the absence of added
aldol, therefore, we assume that the anti isomer has a
negligible influence on the course of the reaction.
We then examined the amplification of the ee values when
the reaction is performed in a stepwise manner. Indeed, by
adding the mixture of reagents 1a and 2a in four portions
(ratios of 2a:1a were 0.032:0.049, 0.093:0.139, 0.125:0.188,
and 0.250:0.375 mmol with15, 15, 30, and 60 minutes of delay,
respectively) to the reaction medium containing 20 mol% of
the (R)-BINOL2Ti catalyst, we observed the formation of 3a
withexcellent yield (90 relative to 50%) withan ee value of
96% after only 2 hours at À208C! This result may be
rationalized as follows: After the first addition of 6.5 mol%
of the mixed 1a and 2a, the major syn butenolide 3a so
formed can be incorporated into the new catalyst and gives
identical results to those observed when syn 3a is first added
to the reaction mixture. The amplification can be shown by
comparing the ee value obtained in the one addition protocol
(87%) with the value of 96% obtained after the ªstepwiseº
Several other aldehydes (2c ± g) were added with 1a in four
portions as above (Scheme 1 and Table 2). It is noteworthy
that aldehydes possessing either isolated double bonds, such
as 2c, or a long saturated aliphatic chain, like 2 f, gave the
corresponding aldols withexcellent yields and high ee values
(3c and 3 f, 94 and 96% ee, respectively). In the case of a,b-
unsaturated aldehyde 2e (Entry 3), we were delighted to
observe that the aldols 3e, which resulted from 1,2-addition,
were the major compounds formed (ratio of 1,2- to 1,4-
addition[19] ˆ 70:30). However, in this latter case the major
product is now the anti isomer, and the syn isomer only had an
ee value of 52%; the absolute configuration of the major
enantiomer was determined as (4S,5S). In the case of the
aromatic aldehyde 2d, the reaction was again quantitative
(Entry 2) and the major anti aldol 3d had an ee value of 90%,
whereas the syn product only had a 60% ee value (absolute
configurations have not yet been determined). The reverse
diastereoselectivity, as well as a lower enantioselectivity in
these two cases may be rationalized by including electronic
factors. Finally, when a bulky aldehyde, such as cyclohex-
ylcarbaldehyde 2g, is involved (Entry 5) a 70:30 mixture of
the syn:anti aldols 3g was obtained in 70% yield. The major
syn product had an ee value of 71%, whereas the anti isomer
showed a 79% ee value. Thus, as far as chemical yields and ee
values are concerned, the reaction seems sensitive to steric
hindrance, which is often the case with aldol reactions.[20]
In an attempt to improve the d.r., we turned our attention
to the use of different silyloxyfurans, 2-tert-butyldimethylsil-
yloxyfuran (TBDMSOF) 1b[21] and 2-tert-butyldiphenylsilyl-
oxyfuran (TBDPSOF) 1c. When these silyloxyfurans were
separately mixed with 2a in the presence of a (R)-BINOL2Ti
complex solution (20 mol%) in Et2O cooled to À208C, no
reaction occurred (after 12 hours). However, when the
reaction was performed witheither 1b or 1c in a stepwise
fashion at À208C (ratios of 2a to 1b or 1c ˆ 0.032:0.049,
0.093:0.139, 0.125:0.188, and 0.250:0.375 mmol with30, 60, 90,
and 180 min delay, respectively), the aldols 3a were obtained
in 60% yield as a 70:30 syn:anti mixture with ee values of 97%
for the (4S,5S) isomer and 86% for the (4R,5S) isomer for the
reaction involving 1b, and 45% yield as a 86:14 syn:anti
mixture witha value of 82% ee for the (4S,5S) isomer, for the
reaction with 1c (results not optimized). It is worthnoting that
the d.r. was improved from 60:40 to 70:30 and finally 86:14 by
using the more bulky reagents.
experiment. Indeed, these results are consistent with an
[16]
autoinduced process withamplification
of the ee value of
the product. In this last experiment, the ee value of the anti
aldol 3a was measured (90% ee), and its absolute config-
uration was assumed to be (4R,5S) by comparison withthe
sign of the specific rotation of the dihydrogenated compound
[17]
withepimuricatacin.
Thus, epimerization at the allylic
In conclusion, the reaction of 1a with achiral aldehydes in
the presence of BINOL/titanium(iv) complexes is the first
catalytic asymmetric autoinductive aldol reaction reported.[22]
Bothyields and enantiomeric excess values are good to
position of the anti aldol 3a (for example, through Et3N
treatment[18]) would eventually lead to the formation of the
major syn aldol compound 3a.
Table 2. Additions of various couples of 2-TMSOF (1a) and aldehyde (2c ± g) in the presence of the BINOL2Ti complex at À208C in Et2O.
Entry
aldehyde
time [h]
yield [%]
syn:anti ratio
ee of syn
ee of anti
3c ± g
(abs. conf.)
(abs. conf.)
1
2
3
4
5
E-decen-4-al (2c)
2
6
2
6
2
99
99
70[b]
80[d]
70
53:47
24:76
30:70
60:40
65:35
94 (S,S)
60 (ND)[a]
52 (S,S)
> 96 (S,S)
71 (S,S)
90 (R,S)
90 (ND)[a]
24 (R,S)[c]
90 (R,S)
79 (R,S)
benzaldehyde (2d)
E-nonen-2-al (2e)
tridecanal (2 f)
cyclohexylcarbaldehyde (2g)
[a] ND ˆ not determined; [b] 1,4-addition ˆ 30%; [c] measured on the hydrogenated product by comparison with the specific rotation of an authentic
compound; [d] when addition is performed in one portion: yield ˆ 20%, d.r. ˆ 60:40, ee ˆ 90% (syn S,S).
1800
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2000
0570-0833/00/3910-1800 $ 17.50+.50/0
Angew. Chem. Int. Ed. 2000, 39, No. 10

COMMUNICATIONS
Interconversion between m-h2,h2-C60 and
m3-h2,h2,h2-C60 on a Carbido Pentaosmium
Cluster Framework**
excellent, which is unusual for such reactions. The stepwise
procedure allows the highest ee values to be obtained together
with high chemical yields.[23] These results show the wide
scope of suchaldol reactions, withcertain limitations for
bulky and/or conjugated aldehydes (results not optimized).
The use of bulkier trialkylsilyloxyfurans slightly improved the
d.r.; further optimization studies are in progress. Furthermore
these findings may provide some evidence for amplification of
ee values of biomolecules (by chemical reactions) in the
chemical origins of life. Indeed, conditions for aldol reactions
may have been found in prebiotic systems. More experiments
are nevertheless needed to set a mathematical model for such
a mechanism.[24] Sucha process could also be involved in other
asymmetric reactions.[23, 25, 26]
Kwangyeol Lee, Chang Hoon Lee, Hyunjoon Song,
Joon T. Park,* Hong Young Chang, and
Moon-Gun Choi
Exohedral metallofullerenes have recently attracted much
attention concerning the effects of metal coordination on the
chemical and physical properties of C60.[1] Most approaches to
forming metal complexes have been based on metal ± C60 p-
complex chemistry, which has resulted in h2-C60, m-h2,h2-C60,
and m3-h2,h2,h2-C60 ligands in monometallic (for most met-
als),[2] bimetallic (Re2, Ru2, Ir2),[3] and metal cluster com-
plexes (Ru3, Os3, Ru5C, Ru6C, PtRu5C),[4, 5] respectively.
Metal clusters can potentially accommodate all these C60
bonding modes, but the interaction of C60 withcluster
frameworks has been, thus far, dominated by the face-capping
cyclohexatriene-like bonding mode, m3-h2,h2,h2-C60. Th em-
h2,h2-C60 bonding mode has never been observed on a cluster
framework, although it has been postulated as an intermedi-
ate for the transformation of [Os3(CO)11(h2-C60)] to [Os3-
(CO)9(m3-h2,h2,h2-C60)] by loss of carbonyl ligands.[5c] The
interconversion among the three kinds of the C60 ligands
remains to be established in the area of C60 ± metal cluster
chemistry. We have recently observed the elusive m-h2,h2-C60
bonding mode on an Os5C cluster framework, and further-
more demonstrated that the two C60 bonding modes m-h2,h2
and m3-h2,h2,h2 are interconvertible.
Received: November 12, 1999 [Z14266]
[1] K. Soai, T. Shibata, H. Morioka, K. Choji, Nature 1995, 378, 767.
[2] a) N. P. M. Huck, W. F. Jager, B. de Lange, B. L. Feringa, Science 1996,
273, 1686; b) Y. Zhang, G. B. Schuster, J. Org. Chem. 1995, 60, 7192.
[3] O. Buchardt, Angew. Chem. 1974, 86, 222; Angew. Chem. Int. Ed.
Engl. 1974, 136, 179.
[4] T. Shibata, J. Yamamoto, N. Matsumoto, S. Yonekubo, S. Osanai, K.
Soai, J. Am. Chem. Soc. 1998, 120, 12157.
[5] T. Shibata, H. Morioka, T. Hayase, K. Choji, K. Soai, J. Am. Chem.
Soc. 1996, 118, 471.
[6] T. Shibata, T. Hayase, J. Yamamoto, K. Soai, Tetrahedron: Asymmetry
1997, 8, 1717.
[7] A. H. Alberts, H. Wynberg, J. Am. Chem. Soc. 1989, 111, 7265.
[8] L. Shengjian, J. Yaozhong, M. Aiqiao, Y. Guishu, J. Chem. Soc. Perkin
Trans 1 1993, 885.
[9] H. Danda, H. Nishikawa, K. Otaka, J. Org. Chem. 1991, 56, 6740.
[10] a) F. Rebiere, O. Riant, H. B. Kagan, Tetrahedron: Asymmetry 1990, 1,
199; b) D. P. Heller, D. R. Goldberg, W. D. Wulff, J. Am. Chem. Soc.
1997, 119, 10551.
Reaction of [Os5C(CO)12(PPh3)(NCMe)2] withC
refluxing ClC6H5 produced a mixture of 1 and 2 (see the
in
60
Á
Â
[11] M. Szlosek, X. Franck, B. Figadere, A. Cave, J. Org. Chem. 1998, 63,
5169.
[Os5C(CO)11(PPh3)(m3-h2,h2,h2-C60)]
[Os5C(CO)12(PPh3)(m-h2,h2-C60)]
1
[12] D. J. Berrisford, C. Bolm, K. B. Sharpless, Angew. Chem. 1995, 107,
1159; Angew. Chem. Int. Ed. Engl. 1995, 34, 1059.
2
[13] K. Mikami, S. Matsukawa, T. Volk, M. Terada, Angew. Chem. 1997,
109, 2936; Angew. Chem. Int. Ed. Engl. 1997, 36, 2768.
[14] K. Mikami, S. Matsukawa, Nature 1997, 385, 613.
Experimental Section). The conversion of 1 into 2 could be
effected by heating a solution of 1 in ClC6H5 at 808C under
1 atm of carbon monoxide. Upon thermolysis at 1328C, 2 was
cleanly reconverted into 1 by loss of a carbonyl ligand
(Scheme 1).
The solid-state structure of 1 is isomorphous to that of the
ruthenium analogue [Ru5C(CO)11(PPh3)(m3-h2,h2,h2-C60)].[4b]
The structure of 2 (Figure 1) reveals a very intriguing feature
Â
[15] M. Chavarot, J. J. Byrne, P. Y. Chavant, J. Pardillos-Guindet, Y. Vallee,
Tetrahedron: Asymmetry 1998, 9, 3889.
[16] a) C. Puchot, O. Samuel, E. Dunach, S. Zhao, C. Agami, H. B. Kagan,
J. Am. Chem. Soc. 1986, 108, 2353; b) N. Oguni, Y. Matsuda, T.
Kaneko, J. Am. Chem. Soc. 1988, 110, 7877; c) M. Kitamura, S. Okada,
S. Suga, R. Noyori, J. Am. Chem. Soc. 1989, 111, 4028; d) C. Girard,
H. B. Kagan, Angew. Chem. 1998, 110, 3088; Angew. Chem. Int. Ed.
1998, 37, 2922; e) B. L. Feringa, R. A. van Delden, Angew. Chem.
1999, 111, 3624; Angew. Chem. Int. Ed. 1999, 38, 3418.
Á
Â
[17] B. Figadere, J.-C. Harmange, A. Laurens, A. Cave, Tetrahedron Lett.
[*] Prof. J. T. Park, Dr. K. Lee, C. H. Lee, H. Song
Department of Chemistry & Center for Molecular Science
Korea Advanced Institute of Science and Technonogy
Taejon, 305-701 (Korea)
1991, 32, 7539.
Á
Â
[18] B. Figadere, C. Chaboche, J. F. Peyrat, A. Cave, Tetrahedron Lett.
1993, 34, 8093.
[19] H. Nishikori, K. Ito, T. Katsuki, Tetrahedron: Asymmetry 1998, 9,
1165.
[20] S.-I. Kiyooka, H. Maeda, M. Abu Hena, M. Uchida, C.-S. Kim, M.
Horiike, Tetrahedron Lett. 1998, 39, 8287.
[21] G. Rassu, F. Zanardi, L. Battistini, E. Gaetani, G. Casiraghi, J. Med.
Chem. 1997, 40, 168.
Fax : (‡82)42-869-2810
E-mail: jtpark@kaist.sorak.ac.kr
H. Y. Chang, Prof. M.-G. Choi
Department of Chemistry, Yonsei University
Seoul, 120-749 (Korea)
[**] We are grateful to the Korea Science Engineering Foundation
(KOSEF) for financial support (project no. 1999-1-122-001-5) of this
research and a postdoctoral fellowship to K.L. The X-ray diffraction
studies were carried out at the X-ray Crystallographic Laboratory of
Yonsei University, which was supported in part by KOSEF.
[22] A. H. Alberts, H. Wynberg, J. Chem. Soc. Chem. Commun. 1990, 453.
[23] S. Kobayashi, Y. Fujishita, T. Mukaiyama, Chem. Lett. 1990, 1455.
[24] E. F. Kogut, J. C. Thoen, M. A. Lipton, J. Org. Chem. 1998, 63, 4604.
Á
Â
[25] M. Pichon, J. C. Jullian, B. Figadere, A. Cave, Tetrahedron Lett. 1998,
39, 1755.
Supporting information for this article is available on the WWW under
http://www.wiley-vch.de/home/angewandte/ or from the author.
[26] D. A. Evans, M. C. Kozlowski, J. A. Murry, C. S. Burgey, K. R.
Campos, B. T. Connell, R. J. Staples, J. Am. Chem. Soc. 1999, 121, 669.
Angew. Chem. Int. Ed. 2000, 39, No. 10
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 2000
0570-0833/00/3910-1801 $ 17.50+.50/0
1801