Asymmetric synthesis of febrifugine and isofebrifugine using yeast reduction

Article abstract of DOI:10.1039/b005028h

The antimalarial agents febrifugine ((+)-1) and isofebrifugine ((+)-2) were asymmetrically synthesized from chiral piperidin-3-ol ((+)-4), which was prepared by the reductive dynamic optical resolution of the 3-piperidone derivatives ((±)-3) using baker's

Full text of DOI:10.1039/b005028h

Asymmetric synthesis of febrifugine and isofebrifugine using yeast reduction
Yasuo Takeuchi,* Kumiko Azuma, Kentaro Takakura, Hitoshi Abe and Takashi Harayama
Faculty of Pharmaceutical Sciences, Okayama University, Tsushima-naka 1-1-1, Okayama 700-8530, Japan.
E-mail: take@pharm.okayama-u.ac.jp
Received (in Cambridge, UK) 22nd June 2000, Accepted 21st July 2000
Published on the Web 11th August 2000
The antimalarial agents febrifugine ((+)-1) and isofebrifu-
gine ((+)-2) were asymmetrically synthesized from chiral
piperidin-3-ol ((+)-4), which was prepared by the reductive
dynamic optical resolution of the 3-piperidone derivatives
(( )-3) using baker’s yeast.
((2)-3, 34%) also had high enantiomeric purity (90% ee). Since
epimerization of (2)-3 with K2CO3 for 4 h gave (±)-3 in 67%
yield, we could be sure of the possibility of reductive dynamic
optical resolution. Epimerization that maintained the reductive
activity of the yeast was examined under a variety of
concentrations of K2CO3 and EtOH in water. Stirring a mixture
of (±)-3, baker’s yeast, sucrose, and K2CO3 in EtOH and water
at rt for 90 h afforded (+)-4 in 62% yield (97% ee) along with
(2)-3 in 14% yield (14% ee).6
Febrifugine ((+)-1) is an antimalarial agent that was isolated
from Dichroa febrifuga and Hydrangea umbellata along with
isofebrifugine ((+)-2).1a–c The plane structure2a of (+)-1 and
(+)-2 was first proposed in 1950. Subsequently, the relative2b
and absolute2c structures were proposed on the basis of Baker’s
synthetic work.3a–c The relative configuration2d of (+)-1 was
corrected in 1973 and then the absolute structures2e,f of (+)-1
and (+)-2 were corrected in 1999 (Fig. 1). We previously
synthesized the racemates of (+)-1 and (+)-2 via the unusual
Claisen rearrangement and highly diastereoselective reduc-
tion.4a,b In this paper, we describe the asymmetric synthesis of
(+)-1 and (+)-2 based on our synthetic method.
Isofebrifugine ((+)-2) was prepared by improving our
previous method (Scheme 2). The intramolecular bromoether-
ification of (+)-4 using NBS afforded octahydrofuro[3,2-b]-
pyridine (5) in 87% yield. The HPLC data indicated that this
was a 3:1 mixture of the diastereomeric isomers. To improve the
yield and reproductivity of preparing Z-protected isofebrifugine
(7) from 5, we designed a 2-methoxy derivative 6. The new
intermediate 6 could be prepared by a 2-step reaction in high
yield (90%) as a 4+1 mixture of the diastereomeric isomers. The
steps were dehydrobromination using potassium tert-butoxide
and bromoetherification using NBS and MeOH. Deacetaliza-
tion of 6 followed by a coupling reaction with 4(3H)-
quinazolinone afforded 7 in 69% yield. The hydrogenolysis of
7 gave isofebrifugine ((+)-2) in 62% yield as a crystalline solid.
The melting point,1b 1H-NMR data,7a,b and optical rotation1b
for (+)-2 agreed with reported values for the natural product.8 It
was clear that the yeast reduction of (±)-3 followed the Prelog
rule9 with high enantio- and diastereoselectivity to give (+)-4
having the absolute configuration of 2S,3S.
In our previous method for synthesizing (±)-febrifugine
((±)-1), the large differences4b in the melting point and
solubility of (±)-1 and (±)-2 played a convenient role for the
isolation of (±)-1. However, we could not isolate pure (+)-1 by
refluxing (+)-2 in EtOH. Although it is known that isomeriza-
tion of the congeners of (+)-2 run under the reversible Michael
reactions,7a,10 there are no reports of (+)-2 itself. We examined
the isomerization of (+)-2 to (+)-1 in various solvents, including
toluene, DMF, pyridine, EtOH, water, and 10% HCl aq. Heating
(+)-2 at 80 °C for 30 min in water resulted in the largest ratio
(2+1) of (+)-1 to (+)-2 among these solvents. On the other hand,
the epimerization of (+)-2 in 10% HCl aq. was not observed
under the same conditions. Based on these results, we isolated
pure (+)-111 as the hydrochloride salt and its physicochemical
Fig. 1 Structures of febrifugine and isofebrifugine
As a key asymmetric reaction, we selected the yeast reduction
of 3-piperidone derivatives ((±)-3)4a,b protected by a benzyl-
oxycarbonyl (Z) group (Scheme 1). The reaction of (±)-3 with
baker’s yeast and sucrose in EtOH and water at rt for 24 h
1
afforded chiral piperidin-3-ol ((±)-4), for which the H-NMR
data agreed with reported values4b for (±)4, in high yield (40%)
and enantiomeric purity (98% ee).5 The remaining piperidone
Scheme 2 Reagents and conditions: a, NBS, MeCN, rt, 0.5 h, 87%; b, (i)
tBuOK, THF, 0 °C, 0.5 h; (ii) NBS, MeOH, rt, 1 h, 90%; c, (i) H+, MeCN,
rt, 1 h; (ii) 4(3H)-quinazolinone, K2CO3, DMF, rt, 2 h, 69%; d, H2,
20%-Pd(OH)2/C, MeOH, rt, 3.5 h, 62%; e, (i) H2O, 80 °C, 15 min; (ii) H+,
73%.
Scheme 1
DOI: 10.1039/b005028h
Chem. Commun., 2000, 1643–1644
This journal is © The Royal Society of Chemistry 2000
1643

properties1c and spectral data7a,b were identical with those
reported for the natural product.
(31000 g, 10 min) was dried, filtered, and concentrated. MeCN was
added to the residue and the precipitates were filtered through a
membrane filter (0.5 mm). The filtrate was concentrated and the residue
was subjected to column chromatography (silica gel). The first eluant
(AcOEt–hexane, 1+7) gave (2)-3 (1.02 g, 90% ee based on HPLC with
a chiral column) as a colorless oil, [a]2D9 245.5 (c 1.00, EtOH). HPLC
conditions: column, Chiralcel OJ (Daicel); column temperature, rt;
eluant, hexane–IPA, 37+3; flow rate = 1.5 ml min21; wavelength, 254
nm; tR = 8.54 and 11.17 min. The second eluant (AcOEt–hexane, 1+3)
gave (+)-4 (1.20 g, 98% ee based on HPLC with a chiral column) as a
viscous colorless oil. HPLC conditions: column, Chiralcel OJ (Daicel);
column temperature, rt; eluent, hexane–IPA, 37+3; flow rate = 1.5 ml
min21; wavelength, 254 nm; tR = 6.43 and 7.51 min.
We were able to prepare febrifugine ((+)-1) in 15.2% yield
and 6 isolated steps from (±)-3 without using compounds that
were very expensive, toxic, or dangerous. We think that our
method is widely applicable to the synthesis of the derivatives
needed to study the structure–activity relationship of febrifu-
gine.
We are grateful to the SC-NMR Laboratory of Okayama
University for the use of the facilities.
6 Yeast Reduction (Method B); A mixture of (±)-3 (1.00 g), K2CO3
(1.00 g), baker’s yeast (10 g), and sucrose (30 g) in EtOH (30 ml) and
water (300 ml) was stirred at rt for 90 h. The reaction mixture was
treated in the same way as described above. The eluant (AcOEt–hexane,
1+7) gave (2)-3 (0.14 g, 14% ee based on HPLC with a chiral column)
as a colorless oil. The second eluant (AcOEt–hexane, 1+3) gave (+)-4
(0.62 g, 97% ee based on HPLC with a chiral column) as a viscous
colorless oil, [a]2D4 +76.2 (c 1.00, EtOH).
Notes and references
1 (a) J. B. Koepfli, J. F. Mead and J. A. Brockman Jr., J. Am. Chem. Soc.,
1947, 69, 1836; (b) J. B. Koepfli, J. F. Mead and J. A. Brockman Jr.,
J. Am. Chem. Soc., 1947, 69, 1048; (c) F. Ablondi, S. Gordon, J. Morton
II and J. H. Williams, J. Org. Chem., 1952, 17, 14.
2 (a) J. B. Koepfli, J. A. Brockman and J. Moffat, J. Am. Chem. Soc.,
1950, 72, 3323; (b) B. R. Baker, F. J. McEvoy, R. E. Schaub, J. P. Joseph
and J. H. Williams, J. Org. Chem., 1953, 18, 178; (c) R. K. Hill and
A. G. Edwards, Chem. Ind., 1962, 858; (d) D. F. Barringer, G.
Berkelhammer and R. S. Wayne, J. Org. Chem., 1973, 38, 1937; (e) S.
Kobayashi, M. Ueno, R. Suzuki and H. Ishitani, Tetrahedron Lett.,
1999, 40, 2175; (f) S. Kobayashi, M. Ueno, R. Suzuki, H. Ishitani, H.-S.
Kim and Y. Wataya, J. Org. Chem., 1999, 64, 6833.
7 (a) S. Uesato, Y. Kuroda, M. Kato, Y. Fujiwara, Y. Hase and T. Fujita,
Chem. Pharm. Bull, 1998, 46, 1; (b) K. Murata, F. Takano, S. Fushiya
and Y. Oshima, J. Nat. Prod., 1998, 61, 729.
8 Isofebrifugine; mp 130–131 °C (lit.1b 129–130 °C). [a]D22 +124.3 (c
0.50, CHCl3) {lit.1b [a]D25 +131 (c 0.35, CHCl3)}. The 1H NMR
spectrum agreed with that reported in the literature.7a,b
9 V. Prelog, Pure Appl. Chem., 1964, 9, 119.
3 (a) B. R. Baker, R. E. Schaub, F. J. McEvoy and J. H. Williams, J. Org.
Chem., 1952, 17, 132; (b) B. R. Baker, F. J. McEvoy, R. E. Schaub, J. P.
Joseph and J. H. Williams, J. Org. Chem., 1953, 18, 153; (c) B. R. Baker
and F. J. McEvoy, J. Org. Chem., 1955, 20, 136.
4 (a) Y. Takeuchi, H. Abe and T. Harayama, Chem. Pharm. Bull., 1999,
47, 905; (b) Y. Takeuchi, M. Hattori, H. Abe and T. Harayama,
Synthesis, 1999, 1814.
5 Yeast Reduction (Method A); A mixture of (±)-3 (3.00 g), baker’s
yeast (30 g), and sucrose (30 g) in EtOH (30 ml) and water (300 ml) was
stirred at rt for 24 h. AcOEt (900 ml) was added to the mixture and
stirred at rt for 10 min. The AcOEt layer separated by centrifugation
10 D. F. Barringer, G. Berkelhammer, S. D. Carter, L. Goldman and A. E
Lanzilotti, J. Org. Chem., 1973, 38, 1933.
11 Febrifugine dihydrochloride; Isofebrifugine ((+)-2, 0.34 g) in water
(10 ml) was heated at 80 °C for 15 min. To the mixture, 10% HCl aq. (2
ml) was added and the solvent was evaporated off. The residue was
washed with hot EtOH to give almost pure febrifugine dihydrochloride
((+)-1•2HCl, 0.31 g, 73%), which was recrystallized from a mixture of
EtOH and water (9+1), mp 218–219 °C (decomp.) (lit.1c 223–225 °C
(decomp.)). [a]D29 +13.3 (c 1.01, H2O) {lit.1c [a]D31 +12.8 (c 0.85, H2O)}.
The 1H NMR spectrum of free base ((+)-1) agreed with that reported in
the literature.7a,b
1644
Chem. Commun., 2000, 1643–1644