Communication
To assign the absolute configuration of the four stereoiso-
mers, we decided to follow Mosher's methodology. To this aim
starting from enantiomerically pure (–)-2a and (–)-2b, (R)- and
Table 1. Reduction of ethyl secodione 1 by different yeasts.
(
S)-MTPA esters (MTPA = α-methoxy-α-trifluoromethylphenyl-
acetic acid) of the alcohol at C-17 were prepared and the abso-
lute configuration at C-17 was unambiguously assigned as 17S
in both cases (see Supporting Information). Based upon the
already assigned relative configuration, the absolute configura-
tion of (–)-2a was established as 13R, 17S and the absolute
configuration of (–)-2b was established as 13S, 17S.
Yeasts[a]
Product
ee[b]
%]
m.c.
[%]
Time
[h]
[
Pichia anomala DBVPG 2873
P. capsulata CBS 1993
P. etchellsii CBS 2011
P. glucozyma CBS 5766
P. minuta CBS 1708
Rhodotorula glutinis NRRL 1587
R. rubra MIM146
S. cerevisiae CEN.PK113-7D
(13R,17S)-2a
(13R,17S)-2a
(13R,17S)-2a
(13S,17S)-2b
(13R,17S)-2a
(13S,17S)-2b
(13S,17S)-2b
(13R,17S)-2a
20
51
50
94
90
90
76
96
65
62
75
85
80
78
92
81
24
96
24
2
24
24
24
48
Corey and co-workers developed a method for the stereo-
selective reduction of methyl secodione (Torgov's diketone) by
using oxazaborolidine catalysis, which furnishes enantiopure
(
13R, 17R)-methyl secol in 86 % yield with 99 % ee after recrys-
[
5]
tallization, the preparation of which opened the route to a
stereoselective version of Torgov's synthesis of estrone. Our at-
tempts to apply Corey's method to the reduction of 1 gave
poor results, with formation of (13R, 17R)-2b in low yields (25–
[
a] To compare yields, all biotransformations were performed with the same
optical density of yeast culture (OD600/mL = 60). [b] Determined by analytical
chiral HPLC (details in the Supporting Information); in all the biotransforma-
tions de was > 98 %.
30 %) and poor enantioselectivity (75 % ee).
Biocatalytic reduction (either by using isolated ketoreduct-
ases or ketoreductases bound to whole microbial cells) is an
biotransformation with P. minuta CBS 1708 and S. cerevisiae
CEN.PK113-7D were further optimized by evaluating different
parameters (temperature, pH, substrate and cells concentration,
and co-substrate type and concentration).
effective alternative for obtaining enantiopure compounds.[6]
A
recombinant ketoreductase (KRED1-Pglu), recently reported as
an efficient stereoselective biocatalyst for the reduction of
[
7]
bulky diketones, was employed for the reduction of 1. KRED1-
Pglu was used in the presence of a catalytic amount of NADP+
and an enzyme-coupled system (glucose/glucose dehydrogen-
ase, GDH) for the regeneration of the cofactor. Different param-
eters of the biotransformation (substrate and enzyme concen-
trations, pH, type and co-substrate concentrations) were opti-
mized by using a Multisimplex, approach, previously used for
Table 2 reports the comparison between the results obtained
in the biotransformation of 1 accomplished on a 1 liter scale
with the best-performing biocatalysts (recombinant KR or
yeasts) used under optimized conditions. Although S. cerevisiae
and the recombinant ketoreductase furnished the highest
stereoselectivity, P. minuta resulted in the formation of the
highest amounts of enantiomerically pure (13R, 17S)-2a after a
single crystallization from diisopropyl ether (DIPE). Actually, P.
minuta allowed us to perform the biotransformation at sub-
[
8]
biotransformations; KRED1-Pglu under optimized conditions
0.05 Tris-HCl at pH 8.0, 30 °C, 6.5 m substrate concentration
(
M
M
in the presence of 3.2 % EtOH and 2.5 % glucose) gave the
desired stereoisomer (13R, 17S)-2a with outstanding stereose-
lectivity (> 98 % ee) and good reaction rates, but with moderate
yields (65 % after 6 h). Notably, no further reduction to diol was
observed even at prolonged times, indicating the total selectiv-
ity of KRED1-Pglu.
strate concentrations up to 4.7 g/L (15 mM) with > 95 % yield,
whereas S. cerevisiae and KRED1-Pglu provided high yields only
at much lower concentrations.
Table 2. Optimized reduction of 1 into (13R, 17S)-2a.
The relatively low yield observed with the recombinant
ketoreductases, led us to screen various yeasts to find more
productive biocatalysts. The screening was carried out with
yeasts previously known for their ability to reduce structurally
[
9–12]
different ketones.
Table 1 gives the results of the biotrans-
formation at the time of maximum production (only reactions
with yields above 50 % are reported).
Six strains (Pichia etchellsii CBS 2011, Pichia glucozyma CBS
m.c[a] [%] Isolated yield [%][b]
Biocatalyst
ee [%]
Time [h]
5
766, Pichia minuta CBS 1708, Rhodotorula glutinis NRRL-Y1587,
P. minuta[c]
S. cerevisiae
92
> 98
> 98
> 95
> 95
65
73
87
56
24
24
8
Rhodotorula rubra MIM 146, Saccharomyces cerevisiae
CEN.PK113-7D) gave monoreduction of 1 with yields higher
than 70 %. P. minuta CBS 1708 and S. cerevisiae CEN.PK113-7D
afforded the desired stereoisomer (13R, 17S)-2a with ee ≥ 90 %;
different stereoselectivity was detected with P. glucozyma CBS
[d]
e]
KRED1-Pglu[
[
a] Analytical yield determined by HPLC. [b] After crystallization from DIPE.
[c] Conditions: 0.1 phosphate buffer at pH 7.2, 30 °C, 15 m substrate
concentration in the presence of 3.2 % EtOH and 25 gdry weight/L of yeast cells.
d] Conditions: 0.1 phosphate buffer at pH 6.5, 30 °C, 10 m substrate
concentration in the presence of 3.5 % EtOH and 30 gdry weight/L of yeast
cells. [e] Conditions: 0.05 Tris-HCl buffer at pH 8.0, 28 °C, 6.5 m substrate
M
M
[
M
M
5766 and R. glutinis NRRL-Y1587, giving (13S, 17S)-2b with a
high ee value. Notably, all the biocatalysts employed (recombi-
nant KRED1-Pglu and whole cells) followed Prelog's rule, reduc-
ing the 17-carbonyl to give the corresponding (S)-alcohol. The
M
M
concentration in the presence of 3.2 % EtOH, 2.5 % glucose, KRED1-Pglu and
GDH.
Eur. J. Org. Chem. 2016, 1260–1263
www.eurjoc.org
1262
© 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim