Utaka 1989
Utaka 1989
Utaka 1989
Masanori Utaka,* Satoshi Konishi, Ami Mizuoka, Toshiyasu Ohkubo, Takashi Sakai,
Sadao Tsuboi, and Akira Takeda
Department of Synthetic Chemistry, Faculty of Engineering, Okayama University, Tsushima, Okayama 700,
Japan
Received M a y 1 , 1989
In a previous paper,’ we have reported that (E)-and interest in view of the mechanism of bakers’ yeast re-
(2)-methyl 2,4,4-trichloro-2-butenoate (le) were reduced duction and the methodology of organic synthesis.
to (R)- and (S)-2,4,4-trichlorobutanoic acids (2e), respec- Although many optically active a-halo acids or esters
tively, by use of fermenting bakers’ yeast (Saccharomyces have been prepared by using naturally occurring chiral
cereuisiae) (eq 1). The R and S acids 2e were effectively acids such as a-amino acids3 or lactic acid4 as starting
materials, the asymmetric syntheses have been rather rare.
Reported syntheses include a microbial hydrogenation of
the carbon-carbon double bond of a-halo substituted
1
L enoate anions by using Clostridium sp. La l6and an
2
asymmetric halogenation of camphor-10-sulfonic acid de-
R
rived esters with N-bromo- or N-chlorosuccinimide.6 The
a CH3CHl d CHzCI present method reduces the carbon-carbon double bond
b (CH3)zCH e CHCli of a-chloro alkenoate by use of easily available bakers’
c CHg(CH2)3 f CCI, yeast. The scope and limitation of the method and its
application to the preparation of a-amino acids are fully
converted to optically pure L- and D-armentomycin (5e),2 described.
both enantiomers of a naturally occurring antibiotic agent
(eq 2). The remarkable feature of this reduction is the Results and Discussion
Preparation of the Substrate. The substrates 1 were
readily prepared by starting from aldehydes and methyl
a-chloroacetoacetate according to the method r e p ~ r t e d . ~
The products obtained were a mixture of E and 2 isomers
except for the case of R = CC13, where only the 2 isomer
was produced. The separation was carried out by using
a silica gel column. Determination of the E and 2 geom-
etry for la-c was made by using the chemical shift of the
highly effective stereochemical control for each geometrical olefinic hydrogen on the basis of the downfield shift for
isomer to produce the (R)- and (S)-a-chloro acids 2e se-
lectively. The E isomer [(E)-le] was reduced to (R)-2e in (3) Barrett, G. C. In Chemistry and Biochemistry of the Amino Acids;
92% ee (60% yield), while the 2 isomer [(Z)-le] was re- Barrett, G. C., Ed.; Chapman and Hall: London, 1985; Chapter 11.
(4) Uenakai, H. Japanese Patent S 61-68445, 1986.
duced to (S)-2e in 98% ee (65% yield). Thii result attracts (5) Hashimoto, H.; Simon, H. Angew. Chem., Znt. Ed. Engl. 1975,14,
106. Simon, H.; Gunther, H.; Bader, J.; Jischer, W. Angew. Chem., Znt.
Ed. Engl. 1981,20,861. Simon, H.; Bader, J.; Gunther, H.; Neumann,
(1) Utaka, M.; Konishi, S.: Okubo, T.: Tsuboi, S.: Takeda, A. Tetra- S.; Thanos, J. Ann. N.Y. Acad. Sci. 1984, 434, 171. Simon, H.;Bader,
hedron Lett. 1987,28,1447. J.; Gunther, H.; Neumann, S.; Thanos, J. Angew. Chem., Znt. Ed. Engl.
(2) (a) Argoudelis, A. D.; Herr, R. R.; Mason, D. J.; Pyke, T. R.; Zieserl, 1985, 24, 539.
J. F. Biochemistry 1967.6.165. (b) Arzoudelis. A. D.: Herr, R. R.: Mason. (6) Oppolzer, W.;Dudfield, P. Tetrahedron Lett. 1985,26, 5037.
D. J. Japanese Patent S 46-16676, 1$7L ( 7 ) Tsuboi, S.; Uno, T.; Takeda, A. Chem. Lett. 1978, 1325.
Table I. Asymmetric Reduction of Methyl 2-Chloro-2-alkenoates (1) to Methyl 2-Chloroalkanoates (2) with Fermenting
Bakers' Yeast
RCH=CClCOZCHB (1) RCHZCHClCOzCH3 (2)
R EIZ time, h % yield' % ee [ ( Y ] ~deg
, (c, CHC13)b confign
a CH3CHp E 40-44 23-28 (0) 47 +10.6 (0.62) R
Z 30-53 23-35 (2-3) >98 -24.7 (1.03) S
b (CH3)pCH E 41-47 6-10 (0) 68 +24.1 (1.26) R
z 48-77 16-19 (5) >98 -38.1 (0.74) S
C CHS(CH& E 30-60 30-40 (0) 25 +5.4 (1.63) R
z 76-96 32-40 (4-8) >98 -23.9 (1.19) S
d CHpCl z 30-68 0 (20-66) - - -
e CHClp E 28-30 54-65 (0) 92 +42.4 (1.32) R
Z 30-32 58-71 (0) 98 -44.0 (1.69) S
f cc1, z 11-23 41-69 (10) >98 -10.5 (1.35) S
'Recoveries of 1 are given in parentheses. bAt temperatures of 22-26 "C.
the 2 acids in comparison with the upfield shift for the those of perfluoroalkylated vinyl compo~nds.'~a,p-Un-
corresponding E acids as reported for (E)- and (2)-2- saturated carboxylic acids or esters also seem to be good
chloro-2-pentenoic acidsas For Id-f, in which the group substrates for the bakers' yeast reduction, but they need
R contains one to three chlorine atoms, the determination a close examination. They have been reported to be good
was made by conversion of Id-f to methyl cr-chloro- substrates in cases of ethyl 3-(perfluoroalky1)-2-a.kenoate~~~
c r ~ t o n a t eusing
~ , ~ tributyltin hydride. Details of the work but not good in cases of ethyl 4-hydroxy-2-alkenoate~.~~~.~
will be reported in due course. We have found that, in the reduction of methyl 2-
Conditions for Fermentation and Reduction. Uses chloro-Zalkenoates (I),the products were obtained as free
of bakers' yeast in three modes (fresh, dry, and immobi- acids. However, for example, when ethyl 2,4,4-trichloro-
lized) were examined to improve the yield and the optical butenoate was used as the substrate, the ethyl ester was
purity. The immobilization10was done by using sodium recovered intact in 62% yield, being accompanied by a
alginate or K-carrageenan mixed with dry bakers' yeast in trace of the acidic reduction product. Therefore, it can
ratios of 0.25 and 0.5 per 1 g of the yeast. We found no be said that the ester cannot be the substrate, and that
appreciable improvement in the optical purity by all yeasts it is quite probable that methyl 2-chloro-2-alkenoates (1)
mentioned above, but attained higher yields with an easy were reduced after hydrolysis. The hydrolysis is likely
extraction by use of the low-ratio (0.25:l) immobilized catalyzed by hydrolytic enzymes in yeast cells,I5 and the
yeast. For example, the yield of 2e was 45-50% and methyl ester (1)is hydrolyzed much faster than the ethyl
54-65% in the cases of dry and alginate-immobilized ester.I6 The choice of the methyl esters (1) as substrates
bakers' yeast, respectively. A clear contrast was also ob- rather than the free acids was made because of the ease
served in the 0% and 23-35% yields of 2a, where the of preparation and purification.
immobilization may prevent product losses due to con- We have also found that (E)-2-heptenoic acid and its
tamination with intractable impurities. Since alginate is methyl ester were not reduced to heptanoic acid by bakers'
more easily handled than K-carrageenan, we used the al- yeast. The acid was recovered almost quantitatively after
ginate-immobilized yeast generally. incubation for 4 days, and the methyl ester was hydrolyzed
The concentration of substrate was 2 mmol per 500 mL to the acid in 82% yield, leaving the rest in 14% recovery.
of culture broth with 2.5 g of dry bakers' yeast. We did The fact that methyl 2-chloro-2-heptenoate (IC) was re-
not attempt to maximize the substrate concentration, but duced to optically active 2-chloroheptanoic acid (2c) in-
higher concentrations showed a tendency to suppress the dicates that the electronegative 2-chloro ~ubstituent'~ is
fermentation. essential for the reduction, the conjugate carboxyl group
Bakers' Yeast Reduction of Methyl 2-Chloro-2- being insufficient to activate the carbon-carbon double
alkenoates. The reduction of ketones to optically active bonds.
secondary alcohols by using bakers' yeast has been studied As shown in Table I, the yields of reduction products
extensively." In contrast, fewer studies have been re- 2a-c are rather low. This may be attributable partly to
ported for the reduction of carbon-carbon double bonds losses in the extraction process, especially for 2a, and partly
to optically active saturated carbon chains. The double to the metabolism of yeast cells, especially for 2b.
bonds are limited to those conjugated with ketone or al-
dehyde carbonyl groups,l2 those of allylic alcohols,13 and ~ ~ ~ ~~~~~ ~ ~ ~
The rate-enhancing effect of the chlorine-containing good agreement with those of 2-chloro esters prepared from
substituent is also observed in Table I. For substrates l b L-amino acids. Therefore, 2a-c from (2)-la-c were as-
and le, the dichloromethyl group enhances the rate, about signed the S configuration, and those from (E)-la-c were
twice as fast as the isopropyl group. Since both the groups assigned the R configuration.
have comparable bulkiness, the difference in the rates can The configuration of 2e was determined by conversion
be attributable to the electron-withdrawing effect of two to 2-amino-4,4-dichlorobutanoic acid (5e) according to the
chlorine atoms. The trichloromethyl group for If further method reported (eq 2).21 Thus naturally occurring (S)-5e
enhances the rate, twice as fast as the dichloromethyl was obtained from (R)-2e, which was prepared by the
group. The monochloromethyl group for Id, however, bakers' yeast reduction of (E)-le. Similarly, (R)-5ewas
completely prevented reduction. When Id was subjected obtained from (S)-2e which was prepared from (2)-le.
to the reduction conditions for 30 h, Id was recovered The configuration of 2f was determined by conversion
intact in 42% yield accompanied by hydrolysis to the acid to 2e with tributyltin hydride. The specific rotation of 2e
in 24% yield. After 68 h, the recovery was decreased to thus obtained was found to be comparable to that of (S)-2e
23% and nothing else was obtained. It is quite puzzling from (2)-le,indicating that 2f has the S configuration.
what has happened to Id. Conversion of 2b and 2e to a-Amino Acids. The
Reactivity and Stereochemistry in the Reduction conversion of (R)- and (S)-2eto (S)- and (R)-2-amino-
of E and 2 Isomers. As shown in Table I, the reaction 4,4-dichlorobutanoic acids [ (S)-and (R)-5e],respectively,
time for la-c indicates that the E isomers are reduced was carried out according to the method reported for ra-
faster than the corresponding 2 isomers. The 2 isomers cemic ethyl 2,4,4-trich1orobutanoate2l with minor modi-
were recovered in small amounts. Interestingly, the E fications (eq 2).
isomers with higher reactivity afforded the reduced As expected, 2e was converted to 5e without any ap-
products with lower % ee. preciable loss of % ee. We treated hydrochloride 5'e with
The stereochemical course of the reduction is highly diazomethane in ether to obtain the free amino ester of
dependent on the geometry of the double bond to be re- 4e. The 'H NMR spectra of this free amino ester mea-
duced. All of the 2 isomers led to the S configuration with sured in the presence of (+)-Eu(hfc), indicated that the
198% ee while the E isomers afforded the R configuration % ee of 2e was still retained after the sequence of sub-
with 25-92% ee, depending on the R group. Why were stitution, hydrogenation, and hydrolysis. Thus both op-
the E isomers reduced with lower ee values? This may tically pure enantiomers of 5e (armentomycin) were ob-
arise from the isomerization of the E isomers to the 2 tained by a single crystallization in the last step. This is
isomers and/or from the lower enantioselectivity of the the Fist preparation of both enantiomers of armentomycin.
yeast reductase toward the E isomers. We have observed In a similar way, (S)-2b was converted to (R)-5b (D-
that both the E and 2 isomers are stable and can be stored leucine). However, conversion of (S)-2fto (R)-5fhas not
without isomerization, and that the E isomer is not isom- been successful because of the failure to reduce (R)-3fto
erized to the 2 isomer in the culture solution. Therefore, (R)-4f by several means used, such as 5% Pd-BaSO, in
it is concluded that the ee values reflect the enantiose- EtOH-HC1, NaBH,-THF-MeOH,= SnC12in MeOH,= and
lectivity of the yeast reductase toward the E and 2 isomers. NaBH3CN-THF-MeOH for 3f and for the free acid of 3f.
Interestingly the isopropyl group directs the stereo- The preparation of 3f from 2f using NaN, in DMF was
chemistry more effectively than the ethyl or butyl, but less better carried out at a lower, ice-bath temperature to se-
effectively than the highly polar dichloromethyl group of cure a 64% yield compared with an 18% yield at room
comparable bulkiness. Thus, the polarity of the group has temperature.21
again played an important role in the stereochemistry as
well as in the rate enhancement discussed above. Experimental Section
Determination of Optical Purity. We have deter- All boiling and melting points are uncorrected. IR spectra were
mined the % ee by measuring 100-MHz lH NMR spectra recorded with a JASCO A-102 spectrometer. 'H NMR spectra
in the presence of (+)-Eu(hfc), a t 27 or 45 OC by using a at 60 MHz were obtained on a JEOL PMX 60SI spectrometer.
1:3 mixture of CDC13 and CCl, as solvent. The tempera- 'H NMR spectra at 100 MHz were obtained on a JEOL FX 100
ture was varied to obtain well-resolved sharp NMR signals spectrometer. TMS in organic solvents and sodium 4,4-di-
methyl-4-sila-1-pentanesulfonate(DSS)in D20 were used as in-
and to shift an overlapping signal of moisture. Separation ternal standards. Bulb-to-bulb distillation was performed by using
of the ester methyl signals for the R and S enantiomers a Shibata GTO-250 glass tube oven. HPLC was performed by
2 was ascertained by measuring the spectra of the following using a Hitachi chromatograph equipped with an RI detector.
racemic methyl esters: (f)-2a,18120 mol % (+)-Eu(hfc),, Elemental analyses were performed by E. h a n o of this laboratory
45 "C, 6 5.78 and 5.84; (*)-2b,18100 mol %, 27 "C, 6 4.34 on a Yanagimoto CHN analyzer MT-3.
and 4.36; (f)-2e,19 170 mol 70,27 "C, 6 4.16 and 4.22; Materials. Dry bakers' yeast was purchased from Oriental
(f:)-2f,l9150 mol %, 27 OC, 6 4.81 and 4.88. Yeast, Tokyo. Sodium alginate was from Nakarai Chemicals,
Determination of the Absolute Configuration. Kyoto, and K-carrageenan was from Sansho, Osaka. Tris[3-
Commercially available (S)-2-aminopentanoic acid (L- [ (heptafluoropropyl)hydroxymethylene]- (+)-camphorat01euro-
norvaline), (S)-2-amino-4-methylpentanoic acid (L-leucine), pium(II1) [Eu(hfc),]was obtained from Aldrich.
Immobilization of Bakers' Yeast.24 Immobilization of
and (S)-2-aminohexanoic acid (L-norleucine) were con- bakers' yeast in calcium alginate gel beads was carried out by a
verted to the corresponding (S)-methyl2-~hloroalkanoates modification of the method of Kientan and Bucke.= A 6-g sample
with retention of configuration according to the method of dry bakers' yeast was suspended in 120 mL of water containing
reported.20 Their specific rotations were measured and 1.5 g of sodium alginate. This suspension was transferred to a
used as references. In a chiral sense, 2-chloro esters 2a-c
obtained from (2)-la-c showed their specific rotations in
(21)Heinzer, F.;Martin, P. Helu. Chin. Acta 1981, 64, 1379.
(22)Sakai, K.; Yokoyama, S.; Ookawa, A. Synthesis 1987, 48.
(18)Prepared by chlorination of the corresponding acid chloride fol- (23)Maiti, S. N.;Singh, M. P.; Micetich, R. G. Tetrahedron Lett.
lowed by methanolysis. 1986,27, 1423.
(19)Prepared by radical addition of CHC1, or CCl, to methyl acrylate. (24)Bucke, C. and Chibata, I.; Tosa, T.; Sato, T.; Takata, I. In
See: Martin, P.; Steiner, E.; Bellui, D. Helu. Chim. Acta 1980, 63, 1947. Methods in Enzymology; Mosbach, K., Ed.; Academic: Orlando, 1987;
(20)Aketa, K.;Terashima, S.; Yamada, S. Chem. Pharm. Bull. 1976, Vol. 135,Section 11.
24,621. (25)Kierstan, M.; Bucke, C. Biotechnol. Bioeng. 1977, 29, 387.
4992 J. Org. Chem., Vol. 54, No. 21, 1989 Utaka et al.
(26) Takamatsu, S.; Yamamoto, K.; Tosa, T.; Chibata, I. J.Ferment. (27) The Merck Index, 9th ed.;Merck & Co., Inc.: Rahway, NJ, 1976;
TechnoL. 1981, 59, 489. p 713.