Enantioselective [1,2] Wittig rearrangement using an external chiral ligand

Article abstract of DOI:10.1002/(SICI)1521-3773(19991216)38:24<3741::AID-ANIE3741>3.0.CO;2-5

Even with only a catalytic amount of chiral bis(dihydrooxazole) 1 as external ligand, the rearrangement of benzyl ethers to alcohols [Eq. (1)] proceeds with high enantioselectivity (over 60 % ee). This reaction represents the first example of an enantiose

Full text of DOI:10.1002/(SICI)1521-3773(19991216)38:24<3741::AID-ANIE3741>3.0.CO;2-5

COMMUNICATIONS
[4] For discussions of the transition state geometry of the aldol reaction,
see a) H. E. Zimmerman, M. D. Traxler, J. Am. Chem. Soc. 1957, 79,
1920 ± ; b) S. E. Denmark, B. R. Henke, J. Am. Chem. Soc. 1991, 113,
2177 ± 2194, and references therein; c) C. Gennari, S. Vieth, A.
Comotti, A. Vulpetti, J. M. Goodman, I. Paterson, Tetrahedron 1992,
48, 4439 ± 4458.
approximately 1000-fold higher than that reported for any
other catalytic antibody.[3f, 10] The catalytic efficiency of
1
the antibody for this substrate, 3.3 Â 105 s 1 m , compares
favorably with the efficiency of natural muscle aldolase,
1
4.9 Â 104 s 1 m , in the retro-aldolization of its substrate
fructose-1,6-bisphosphate.[11]
[5] a) P. G. Schultz, R. A. Lerner, Science 1995, 269, 1835 ± 1842; b) N. R.
Thomas, Nat. Prod. Rep. 1996, 13, 479 ± 511.
In conclusion, we have demonstrated that combining
transition state analogy and reactive immunization design
into a single hapten can result in increases both in the output
of catalysts from the immune system as well as their efficiency.
This strategy resulted in the characterization of the most
proficient antibody catalysts prepared to date. Antibodies
93F3 and 84G3 catalyze a wide array of aldol reactions with ee
values exceeding 95% in most of the cases studied. A new
stereogenic center is formed when acetone is the aldol donor
substrate by attack on the re-face of the aldehyde, which
provides the antipodal complement of ab38C2 in aldol
reactions. Both aldol enantiomers may be accessed through
aldol and retro-aldol reactions. These catalysts should provide
access to a wide variety of enantiomerically enriched synthons
with application to natural product syntheses.
[6] B. List, C. F. Barbas III, R. A. Lerner, Proc. Natl. Acad. Sci. USA
1998, 95, 15351 ± 15355.
[7] I. Paterson, J. M. Goodman, M. A. Lister, R. C. Schumann, C. K.
McClure, R. D. Norcross, Tetrahedron 1990, 46, 4663 ± 4684.
[8] We have identified two catalysts with enantioselectivities similar to
ab38C2.
[9] A. R. Radzicka, R. A. Wolfenden, Science 1995, 267, 90 ± 93.
[10] N. R. Thomas, Appl. Biochem. Biotechnol. 1994, 47, 345 ± 372.
[11] Data for muscle aldolase was reported at 48C: A. J. Morris, D. R.
Tolan, Biochemistry 1994, 33, 12291 ± 12297.
Received: April 29, 1999
Revised version: August 4, 1999 [Z13339IE]
German version: Angew. Chem. 1999, 111, 3957 ± 3960
Enantioselective [1,2] Wittig Rearrangement
Using an External Chiral Ligand**
Keywords: aldol reactions ´ asymmetric synthesis ´ catalytic
antibodies ´ enantiomeric resolution ´ retro reactions
Katsuhiko Tomooka,* Kyoko Yamamoto, and
Takeshi Nakai*
[1] For reviews of the aldol reaction, see a) S. Masamune, W. Choy, J. S.
Peterson, L. R. Sita, Angew. Chem. 1985, 97, 1 ± 31; Angew. Chem. Int.
Ed. Engl. 1985, 24, 1 ± 30; b) C. H. Heathcock, Aldrichim. Acta 1990,
23, 99 ± 111; c) D. A. Evans, Science 1988, 240, 420 ± 426; d) C. J.
Cowden, I. Paterson, Org. React. 1997, 51, 1; e) A. S. Franklin, I.
Paterson, Contemp. Org. Synth. 1994, 1, 317.
[2] a) S. G. Nelson, Tetrahedron: Asymmetry 1998, 9, 357 ± 389; b) A.
Yanagisawa, Y. Matsumoto, H. Nakashima, K. Asakawa, H. Yama-
moto, J. Am. Chem. Soc. 1997, 119, 9319 ± 9320; c) E. M. Carreira, W.
Lee, R. A. Singer, J. Am. Chem. Soc. 1995, 117, 3649 ± 3650; d) D. A.
Evans, D. W. C. MacMillan, K. R. Campos, J. Am. Chem. Soc. 1997,
119, 10859 ± 10860; e) D. J. Ager, M. B. East, Asymmetric Synthetic
Methodology, CRC Press, Boca Raton, 1996; f) C. H. Wong, G. M.
Whitesides, Enzymes in Synthetic Organic Chemistry, Pergamon,
Oxford, 1994; g) C. H. Wong, R. L. Halcomb, Y. Ichikawa, T.
Kajimoto, Angew. Chem. 1995, 107, 453 ± 474; Angew. Chem. Int.
Ed. Engl. 1995, 34, 412 ± 432; h) W. D. Fessner, Curr. Opin. Chem.
Biol. 1998, 2, 85 ± 89.
[3] a) J. Wagner, R. A. Lerner, C. F. Barbas III, Science 1995, 270, 1797 ±
1880; b) R. Björnestedt, G. Zhong, R. A. Lerner, C. F. Barbas III, J.
Am. Chem. Soc. 1996, 118, 11720 ± 11724; c) G. Zhong, T. Hoffmann,
R. A. Lerner, S. Danishefsky, C. F. Barbas III, J. Am. Chem. Soc. 1997,
119, 8131 ± 8132; d) C. F. Barbas III, A. Heine, G. Zhong, T.
Hoffmann, S. Gramatikova, R. Björnestedt, B. List, J. Anderson,
E. A. Stura, E. A. Wilson, R. A. Lerner, Science 1997, 278, 2085 ±
2092; e) T. Hoffmann, G. Zhong, B. List, D. Shabat, J. Anderson, S.
Gramatikova, R. A. Lerner, C. F. Barbas III, J. Am. Chem. Soc. 1998,
120, 2768 ± 2779; f) G. Zhong, D. Shabat, B. List, J. Anderson, S. C.
Sinha, R. A. Lerner, C. F. Barbas III, Angew. Chem. 1998, 110, 2609 ±
2612; Angew. Chem. Int. Ed. 1998, 37, 2481 ± 2484; g) S. C. Sinha, J.
Sun, G. Miller, C. F. Barbas III, R. A. Lerner, Org. Lett. 1999, in press;
h) B. List, D. Shabat, G. Zhong, J. M. Turner, A. Li, T. Bui, J.
Anderson, R. A. Lerner, C. F. Barbas III, J. Am. Chem. Soc. 1999, 121,
7283 ± 7291; i) For an alternative aldolase antibody strategy see J. L.
Reymond, Angew. Chem. 1995, 107, 2471 ± 2473; Angew. Chem. Int.
Ed. Engl. 1995, 34, 2285 ± 2287; J. L. Reymond, Y. Chen, J. Org. Chem.
1995, 60, 6970 ± 6979.
Since its discovery by Wittig and Löhmann in 1942,[1] the
reaction of a-lithiated ethers, now known as the [1,2] Wittig
rearrangement, has attracted much interest from both mech-
anistic and synthetic points of view.[2] This type of carbanion
rearrangement is recognized to proceed by means of the
radical dissociation ± recombination mechanism [Eq. (1)].[2, 3]
Despite its long history, however, no enantioselective
versions of the Wittig rearrangement have been developed
yet. Clearly, the radical character provides a great challenge.
We now disclose the first enantioselective Wittig rearrange-
ment which relies upon an asymmetric lithiation protocol[4] in
which (S,S)-bis(dihydrooxazol) 3 serves as an external chiral
ligand[5, 6] [Eq. (2)]. The most striking feature is that the
[*] Prof. Dr. K. Tomooka, K. Yamamoto, Prof. Dr. T. Nakai
Department of Applied Chemistry
Graduate School of Science and Engineering
Tokyo Institute of Technology
Meguro-ku, Tokyo 152-8552 (Japan)
Fax: (‡81)3-5734-3931
E-mail: ktomooka@o.cc.titech.ac.jp
takeshi@o.cc.titech.ac.jp
[**] This work was supported by a Grant-in-Aid for Scientific Research
from the Ministry of Education, Science, Sports and Culture, Japan
and by the Research for the Future Program, administered by the
Japan Society of the Promotion of Science.
Angew. Chem. Int. Ed. 1999, 38, No. 24
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 1999
1433-7851/99/3824-3741 $ 17.50+.50/0
3741

COMMUNICATIONS
ment of [D3]1 (racemate) afforded (S)-[D3]2 (>90% D
content) in 87% yield and with essentially the same enantio-
meric excess as previously observed for 1 [Eq. (3)], indicating
rearrangement proceeds with essentially the same enantiose-
lectivity even when only a catalytic amount of chiral ligand 3 is
used.
that the lithiation stereochemistry has nothing to do with the
product stereochemistry. The most striking feature of this
asymmetric catalysis is that during the course of reaction the
chiral ligand 3 switches the lithium species partner at each
stage, thereby playing the dual role as a catalyst for the
lithiation step and a chiral auxiliary for the rearrangement
step. The exact mechanism of the asymmetric induction at the
rearrangement step is unclear at present; more detailed work
is needed.
Encouraged by this success, we also examined the enantio-
selective rearrangement of ethers 4 ± 6 (TMS ˆ trimethylsilyl,
TBDPSˆ tert-butyldiphenylsilyl, TIPSˆ triisopropylsilyl) using
tBuLi/3 (2.0 equiv each). The results thus obtained are
We found that the rearrangement of dibenzyl ether (1), a
historic Wittig substrate,[1] when induced with the premixed
complex tBuLi/3 (1.0 equiv each) in diethyl ether at 788C,
afforded alcohol (S)-2 in 55% yield and with 60%ee, along
with recovered 1 in 43% yield.[7, 8] This is the first example of
an enantioselective Wittig rearrangement, though the enan-
tioselectivity is moderate.[9] Significantly enough, when either
one more equivalent of tBuLi was added after the reaction or
the complex tBuLi/3 (2.0/1.0) was initially used, the yield of 2
increased to 90 ± 94% while the same level of enantioselec-
tivity (62%ee) was maintained. This means that two equiv-
alents of tBuLi are required for the complete reaction. At this
point we made two suppositions: 1) The initial lithiation
would be effected almost exclusively by the tBuLi/3 complex;
tBuLi itself, which is known to exist as a dimer in diethyl
ether,[10] could not function as an efficient lithiating agent
under these conditions.[11] 2) The tBuLi dimer could partic-
ipate in complexing with the resulting 3-bound lithium
alkoxide, thereby taking half out of action to form the
tBuLi/alkoxide complex[12, 13] while regenerating the active
complex tBuLi/3. These arguments explain how the enantio-
selective version could proceed even with a catalytic amount
of the chiral ligand 3.
shown below the formula drawings.[14, 15] Of particular interest
is the rearrangement of rac-6, which provides the ter-
tiary alcohol 7 with a relatively high enantioselecitivity
[Eq. (4)].
Therefore, the rearrangement of 1 was carried out using
10 mol% of 3 and two equivalents of tBuLi under the same
conditions. As anticipated, alcohol (S)-2 was obtained in
equally high enantioselectivity and chemical yield (60%ee,
86%). A similar use of 5 mol% of 3 resulted in 54%ee and
81% chemical yield. Scheme 1 depicts a plausible asymmetric
catalytic cycle in which the enantioselectivity might be
determined at the radical-recombination step involving the
chiral ligand-bound anion radical. Indeed, a similar rearrange-
We have described the first example of an enantioselective
[1,2] Wittig rearrangement which can proceed even in an
asymmetric, catalytic fashion. This is also the first example of
an enantioselective, catalytic carbanion-radical rearrange-
ment. Thus, this work opens a new chapter for classic Wittig
chemistry as well as for asymmetric reactions of radicals and
organolithium species.
Experimental Section
To a solution of (S,S)-3 (11.2 mg, 0.038 mmol) in diethyl ether (6 mL) at
788C was added a commercial solution of tBuLi in pentane (0.48 mL of a
1.57m solution, 0.75 mmol) by means of a syringe, and the mixture was
stirred for 15 min. The mixture was maintained at 788C, and a solution of
1 (75.9 mg, 0.38 mmol) in diethyl ether was added dropwise by means of a
syringe. The reaction mixture was stirred at 788C for 2 h and then treated
with a saturated aqueous solution of ammonium chloride. The resulting
mixture was extracted with diethyl ether three times, and the combined
organic layers were washed with brine, dried over anhydrous sodium
sulfate, and concentrated in vacuo. The crude product was purified by
Scheme 1. Plausible catalytic cycle of the enantioselective [1,2] Wittig
rearrangement.
3742
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 1999
1433-7851/99/3824-3742 $ 17.50+.50/0
Angew. Chem. Int. Ed. 1999, 38, No. 24

COMMUNICATIONS
chromatography on silica gel (hexane/diethyl ether 10/1) to give 65.4 mg of
(S)-alcohol 2 (86% yield, 60%ee).
The Design of Leadlike Combinatorial
Libraries
Received: May 3, 1999 [Z13365IE]
German version: Angew. Chem. 1999, 111, 3955 ± 3957
Simon J. Teague,* Andrew M. Davis, Paul D. Leeson,
and Tudor Oprea
Keywords: asymmetric catalysis ´ asymmetric synthesis ´
rearrangements ´ Wittig reactions
Combinatorial chemistry is now widely applied in the drug-
discovery process, for both the identification and optimization
of lead compounds (chemical starting points). Initially the key
factors in the design of libraries intended for finding lead
compounds were considered to be library size and diversity.[1]
More recently consideration has been given to designing
libraries in which the members have druglike physicochemical
properties.[2] Druglike properties are most commonly defined
using the ªrules of 5º: Mr is smaller than 500, the calculated
logarithm of the octanol ± water partition coefficient (clgP) is
less than 5, there are less than five hydrogen-bond donor
atoms, and the sum of the number of nitrogen and oxygen
atoms is less than 10.[3] These guidelines have achieved
widespread acceptance as defining the limiting properties of
most orally active drugs which are able to be absorbed by
passive mechanisms. However, it should be borne in mind that
these are empirical rules derived by examination of the
properties of existing drugs. Herein we propose that the
properties required of library compounds intended to provide
leads suitable for further optimization may be rather differ-
ent.
[1] G. Wittig, L. Löhmann, Liebigs Ann. Chem. 1942, 550, 260 ± 268.
[2] Reviews: a) U. Schöllkopf, Angew. Chem. 1970, 82, 795 ± 805; Angew.
Chem. Int. Ed. Engl. 1970, 9, 763 ± 773; b) R. W. Hoffmann, Angew.
Chem. 1979, 91, 625 ± 634; Angew. Chem. Int. Ed. Engl. 1979, 18, 563 ±
572; c) J. A. Marshall in Comprehensive Organic Synthesis, Vol. 3
(Eds.: B. M. Trost, I. Fleming, G. Pattenden), Pergamon, Oxford, 1991,
pp. 975 ± 1014; d) K. Tomooka, T. Nakai, J. Synth. Org. Chem. Jpn.
1996, 54, 1000 ± 1008; e) K. Tomooka, H. Yamamoto, T. Nakai, Liebigs
Ann. Chem. 1997, 1275 ± 1281.
[3] When R1 is an allylic moiety the rearrangement proceeds over-
whelmingly in the symmetry-allowed [2,3]-sigmatropic fashion. Re-
views on the [2,3] Wittig rearrangement: ref. [2b, c]; T. Nakai, M.
Mikami, Chem. Rev. 1986, 86, 885 ± 902; T. Nakai, M. Mikami, Org.
React. 1994, 46, 105 ± 209.
[4] Reviews on asymmetric lithiation protocols mostly with ( )-sparteine
as the external chiral ligand: a) P. Beak, A. Basu, D. J. Gallagher, Y. S.
Park, S. Thayumanavan, Acc. Chem. Res. 1996, 29, 552 ± 560; b) D.
Hoppe, T. Hense, Angew. Chem. 1997, 97, 2376 ± 2410; Angew. Chem.
Int. Ed. Engl. 1997, 36, 2283 ± 2316.
[5] Prepared by the reported method: S. E. Denmark, N. Nakajima, O. J.-
C. Nicaise, A. M. Faucher, J. P. Edwards, J. Org. Chem. 1995, 60,
4884 ± 4892.
[6] Reviews on asymmetric reactions mediated by chiral bis(dihydroox-
azol)s: a) C. Bolm, Angew. Chem. 1991, 103, 556 ± 558; Angew. Chem.
Int. Ed. Engl. 1991, 30, 542 ± 543; b) A. Pfaltz, Acc. Chem. Res. 1993,
26, 339 ± 345; c) O. Reiser, Angew. Chem. 1993, 105, 576 ± 578; Angew.
Chem. Int. Ed. Engl. 1993, 32, 547 ± 549; synthetic utilities of chiral
bis(dihydrooxazol)/alkyllithium complexes: d) asymmetric addition to
imines: S. E. Denmark, N. Nakajima, O. J.-C. Nicaise, J. Am. Chem.
Soc. 1994, 116, 8797 ± 8798; e) enantioselective rearrangement of
epoxides: D. M. Hodgson, G. P. Lee, Tetrahedron: Asymmetry 1997, 8,
2303 ± 2306; f) enantioselective [2,3] Wittig rearrangement: K. To-
mooka, N. Komine, T. Nakai, Tetrahedron Lett. 1998, 38, 5513 ± 5516.
[7] The enantioselectivity was determined by 1H NMR analysis of the (S)-
MTPA ester (MTPA ˆ a-methoxy-a-(trifluoromethyl)phenylacetic
acid). The absolute configuration was assigned by the specific
rotation: (S)-2 (50%ee) [a]D23 ˆ ‡28.9 (c ˆ 0.97 in EtOH); value of
[a]1D8 ˆ ‡55.9 (c ˆ 1.40 in EtOH) in the literature: G. Berti, F. Bottani,
P. L. Ferrarini, B. Macchia, J. Org. Chem. 1965, 30, 4091 ± 4096.
[8] When sBuLi/( )-sparteine was used instead of tBuLi/3, (S)-2 was
obtained in 36% yield and with 24%ee, and 1 was recovered in 66%
yield.
Our analysis starts from consideration of the common
sources of leads for drug discovery (Figure 1). These have
been divided broadly into three types. The first are leadlike,
Figure 1. Classification of leads by binding affinity. Lower potency leads
are subdivided into leadlike and druglike by reference to their Mr and clgP
values.
[9] When the reaction temperature was lowered to 1108C, 71%ee was
observed with 37% chemical yield.
[10] T. F. Bates, M. T. Clarke, R. D. Thomas, J. Am. Chem. Soc. 1988, 110,
5109 ± 5112.
[11] Treatment of 1 with tBuLi alone in diethyl ether under the same
conditions did not gave any trace of 2. In sharp contrast, the reaction
in THF gave 2 in 70% yield.
[12] For the chemistry of RLi/R'OLi complexes, see P. Caubetre, Chem.
Rev. 1993, 93, 2317 ± 2334.
[13] We found that the separately premixed tBuLi/lithium alkoxide of 2 is
not reactive enough to induce the rearrangement.
low-affinity (>0.1 mm)[4] compounds which have low molec-
ular weight and clgP, typified by some endogenous molecules,
for example histamine and GABA. These have been con-
verted into drugs, through the optimization of potency and
pharmacokinetic profile, by increasing molecular weight and
lipophilicity (Table 1, entries 1 ± 8). The second major source
of leads is typified by high affinity and molecular weight. It
[14] The rearrangement of 4 and 6b with 10 mol% of 3 resulted in 53%ee
(56%) and 54%ee (35%), respectively.
1
[*] Dr. S. J. Teague, A. M. Davis, P. D. Leeson, T. Oprea
Department of Medicinal Chemistry
AstraZeneca R&D Charnwood
[15] The enantiomeric excesses were determined by H NMR analysis of
the (S)-MTPA esters. The absolute configurations of the major
enantiomers obtained from 4 and 5 were not determined. The (R)
configuration of 7a was assigned by conversion into the known (R)-
Bakewell Road
methyl a-benzylmandelate: [a]D23
ee); value of [a]D26
Sullivan, J. R. Beck, A. Pohland, J. Org. Chem. 1963, 28, 2381 ± 2385.
ˆ
13.5 (c ˆ 1.20 in CHCl3) (65%
Loughborough, Leicestershire LE11 5RH (UK)
Fax: (‡44)1509-645-571
ˆ
32.4 (c ˆ 3.4 in CHCl3) in the literature: H. R.
E-mail: simon.teague@charnwood.gb.astra.com
Angew. Chem. Int. Ed. 1999, 38, No. 24
ꢀ WILEY-VCH Verlag GmbH, D-69451 Weinheim, 1999
1433-7851/99/3824-3743 $ 17.50+.50/0
3743