Enantioselective production of 3-hydroxy metabolites of tibolone by yeast reduction

Article abstract of DOI:10.1016/j.steroids.2007.09.008

The enantioselective reduction of tibolone into the corresponding 3α-hydroxy or 3β-hydroxy metabolite can be controlled by choosing suited strains of yeasts and biotransformation conditions. A restricted screening performed among 52 yeasts showed that the 3α-epimer was preferentially obtained with high epimeric purity with various strains (i.e. with Kluyveromyces lactis CBS 2359), while only Saccharomyces cerevisiae CBS 3093 gave the 3β-epimer as major product. The reduction of tibolone with K. lactis CBS 2359 and S. cerevisiae CBS 3093 was optimised. S. cerevisiae CBS 3093 furnished a 96:4 ratio of 3β/3α with complete molar conversion within 72 h when the initial concentration of substrate was below 2.5 g/L. K. lactis CBS 2359 gave a 99:1 ratio of 3α/3β with complete conversion in 64 h.

Full text of DOI:10.1016/j.steroids.2007.09.008

steroids 7 3 ( 2 0 0 8 ) 112–115
available at www.sciencedirect.com
journal homepage: www.elsevier.com/locate/steroids
Enantioselective production of 3-hydroxy
metabolites of tibolone by yeast reduction
Diego Romanoa, Valerio Ferrarioa, Diego Moraa,
Roberto Lennab, Francesco Molinaria,∗
a
`
Dipartimento di Scienze e Tecnologie Alimentari e Microbiologiche, Universita degli Studi di Milano,
via Celoria 2, 20133 Milano, Italy
b
Industriale Chimica, via Grieg 13, 21047 Saronno (VA), Italy
a r t i c l e i n f o
a b s t r a c t
Article history:
The enantioselective reduction of tibolone into the corresponding 3␣-hydroxy or 3␤-hydroxy
metabolite can be controlled by choosing suited strains of yeasts and biotransformation
conditions. A restricted screening performed among 52 yeasts showed that the 3␣-epimer
was preferentially obtained with high epimeric purity with various strains (i.e. with
Kluyveromyces lactis CBS 2359), while only Saccharomyces cerevisiae CBS 3093 gave the 3␤-
epimer as major product. The reduction of tibolone with K. lactis CBS 2359 and S. cerevisiae
CBS 3093 was optimised. S. cerevisiae CBS 3093 furnished a 96:4 ratio of 3␤/3␣ with complete
molar conversion within 72 h when the initial concentration of substrate was below 2.5 g/L.
K. lactis CBS 2359 gave a 99:1 ratio of 3␣/3␤ with complete conversion in 64 h.
Received 14 June 2007
Received in revised form
26 July 2007
Accepted 16 September 2007
Published on line 2 October 2007
Keywords:
Enantioselective reduction
Saccharomyces cerevisiae
Kluyveromyces lactis
© 2007 Elsevier Inc. All rights reserved.
Tibolone
17␣-Hydroxy-7␣-methyl-19-
norpregn-5(10)-en-20-in-3-one
15␤,7␣-Methyl-19-norpregn-5(10)-
en-20-in-3␤,17␤-diol
7␣-Methyl-19-norpregn-5(10)-en-20-
in-3␤,17␤-diol
1.
Introduction
density [3]. Tibolone is quickly metabolised into its 3␣-hydroxy
metabolite (2, 7␣-methyl-19-norpregn-5(10)-en-20-in-3␣,17␤-
Tibolone (1, 17␣-hydroxy-7␣-methyl-19-norpregn-5(10)-en-
20-in-3-one) is a synthetic pro-hormone with estrogenic, pro-
gestagenic and androgenic actions [1]. Results of a randomised
controlled trial of tibolone versus continuous combined
estrogen-only HRT (Hormone Replacement Therapy) suggest
both options are equally effective at relieving vasomotor
symptoms of the menopause [2]. Moreover, all existing stud-
ies have shown a positive action of tibolone on bone mineral
diol) and 3␤-hydroxy metabolite (3, 7␣-methyl-19-norpregn-
5(10)-en-20-in-3␤,17␤-diol) by 3␣/3␤ hydroxysteroid dehydro-
genases (HSDH) (Fig. 1) in the intestine and the liver; the
3-OH-epimers can be interconverted, but only through their
previous conversion into mono- or disulphated compounds
[4]. Only 3␤-OH-tibolone can be converted into the ꢀ4 isomer
(4, Fig. 1) which shows a clear binding to the progesterone
receptor, which is comparable to that of natural progesterone;
∗
Corresponding author. Tel.: +39 0250319148; fax: +39 0250319191.
E-mail address: francesco.molinari@unimi.it (F. Molinari).
0039-128X/$ – see front matter © 2007 Elsevier Inc. All rights reserved.
doi:10.1016/j.steroids.2007.09.008

steroids 7 3 ( 2 0 0 8 ) 112–115
113
Fig. 1 – Metabolism of tibolone and its active metabolites.
the ꢀ4 isomer activates both the progesterone receptor and
the androgen receptor but not the estrogen receptor [5].
The pharmacological importance of these metabolites has
stimulated the search for synthetic methods able to pro-
vide both the 3-OH-epimers of tibolone. Reduction of tibolone
with lithium tri-tert-butoxyaluminum hydride (LTTBA) pre-
dominantly furnishes the 3␣-OH-tibolone in a ratio ␣/␤ of
96:4; inversion of the configuration can be achieved by a
Mitsunobu reaction [6]. Stereoselective preparation of the 3-
hydroxy metabolites of tibolone can be also accomplished by
a chemoenzymatic approach where tibolone is reduced with
diisobutylaluminum hydride (DIBAL) giving a 7/3 mixture of
3␣/3␤-OH-tibolone; Candida antarctica lipase B (CALB) catal-
ysed only the acetylation of the 3␣-OH functionality, therefore
allowing for the recovery of 3␤-OH-tibolone [7]. These meth-
ods are laborious and characterised by low overall yields,
especially for the production of the 3␤-OH-epimer. Microbial
reduction of the 3-keto group of tibolone is a more straight-
forward method for obtaining both 3␣- and 3␤-OH-tibolone.
In this paper, the enantioselective microbial reduction of
tibolone has been developed using yeasts belonging to differ-
ent genera.
solutions with a Bruker AMX-600 (Bruker Spectrospin, Rhein-
stetten, Germany), operating at 600.1 MHz frequency for the
proton nucleus. Chemical shifts were measured in ı (ppm),
using the acetone residual signal as a reference and setting
the methyl proton resonance at 2.1. The [˛]D were measured in
methanol (ca. 1.0) using a Perkin-Elmer Polarimeter (Model 343
Plus). Melting points were measured by differential scanning
calorimetry analysis (Perkin-Elmer DSC 6).
2.1.
Microorganisms and microbial growth
The microorganisms were from CBS (Centraal Bureau voor
Schimmelcultures, Baarn, The Netherlands). They were cul-
tured in 750 mL Erlenmeyer flasks with 100 mL of malt broth
pH 6.0 for 48 h, at 27 ◦C on a rotary shaker set at 200 rpm. Cells
from 48 h submerged cultures were centrifuged and washed
with 0.1 M phosphate buffer, pH 7.0. Saccharomyces cerevisiae
CBS 3093 and Kluyveromyces lactis CBS 2359 were employed in
optimization studies accomplished using cells from a 8 L fer-
menter containing 4.0 L of liquid medium for 48 h, pH 6.0, at
27 ◦C and agitation speed 100 rpm. The dry weights were deter-
mined after centrifugation of 100 mL of cultures, cells were
washed with distilled water and dried at 110 ◦C for 24 h.
2.
Experimental
2.2.
Biotransformations
Solvents and reagents were purchased from Sigma–Aldrich.
TLC was performed on silica gel 60 F254 plates (Merck,
Darmstadt, Germany). NMR spectra were recorded in CDCl3
The substrate was kindly furnished by Industriale Chimica
(Saronno, Italy). Washed cells were suspended in aqueous
buffers (200 mL) to reach the same biomass concentration

114
steroids 7 3 ( 2 0 0 8 ) 112–115
obtained at the end of the growth. Reactions were started
by addition of the substrate (either as micronized powder or
dissolved in co-solvents) and/or co-substrates; the incubation
was performed under reciprocal shaking (120 spm) and the
reaction was followed by HPLC.
The biotransformation with S. cerevisiae CBS 3093 under
optimised conditions was carried out following this procedure:
tibolone (2 g) was added to 63 mL of ethanol (95%) and the solu-
tion heated at 65 ◦C until complete dissolution of tibolone; the
solution at 65 ◦C was added to 1 L of phosphate buffer 0.1 M at
pH 6.0 and 28 ◦C, containing 20 g (dry weight) of yeasts. The
biotransformation mixture was maintained under agitation
(100 rpm).
When the reaction was over, the biotransformation
medium was paper-filtered and the aqueous phase extracted
with ethyl acetate. The organic extracts were dried over
Na2SO4 and the solvent removed; the crude product was
purified by flash chromatography (hexane/ethyl acetate, 7/3,
vanillin stained). Products were assigned on the basis of
their chemico-physical data which were in agreement with
those reported in Ref. [6]: 7␣-methyl-19-norpregn-5(10)-en-20-
in-3␣,17␤-diol (2): [˛]D + 68 (ca. 1, ethyl acetate); 1H NMR d
0.76 (d, J = 12 Hz, 3H); 0.84 (s, 3H); 2.56 (s, 1H); 3.80 (m, 1H).
MS (EI): m/z 314 (M+); 7␣-methyl-19-norpregn-5(10)-en-20-in-
3␤,17␤-diol (2): [˛]D + 16 (ca. 1, ethyl acetate); 1H NMR d 0.76 (d,
J = 12 Hz, 3H); 0.84 (s, 3H); 2.56 (s, 1H); 4.04 (m, 1H). MS (EI): m/z
314 (M+)
Fig. 2 – Influence of co-substrate on the reduction of
tibolone with S. cerevisiae CBS 3093.
tion followed by TLC and HPLC. Table 1 shows the strains able
to reduce tibolone into its 3-hydroxy derivatives with molar
conversions above 10% after 48 h.
The biotransformation with Kluyveromyces and Pichia
strains furnished 3␣-OH-tibolone 2 with high purity, while
only S. cerevisiae CBS 3093 gave predominantly 3␤-OH-tibolone
3, although with a slow reaction rate.
Since the synthetic difficulty in obtaining the epimer 3,
further experiments were aimed at the optimization of the
biotransformation with S. cerevisiae CBS 3093. The optimiza-
tion was performed by simultaneously evaluating parameters
affecting both dehydrogenase expression during the growth
(carbon source supplied, pH, aeration) and the biotransforma-
tion (temperature, type of buffer and cell concentration) using
the Multisimplex experimental design [8]. A medium based on
malt extract added with 0.5% of yeast extract and kept at pH
5.8 with an aeration rate of 1.2 vvm gave 9.3 g/L of dry S. cere-
visiae CBS 3093 with the highest 3␤-HSDH activity. These cells
were employed under optimised conditions of biotransforma-
tion (28 ◦C, concentration of the cells = 20 g L−1 in phosphate
buffer 0.1 M at pH 6.0) for testing the influence of different
co-substrates (25 g/L) on the transformation. Fig. 2 shows the
molar conversion and the epimeric ratio obtained after 24 h.
Ethanol and isopropanol proved to be the best co-substrate
and it was assumed that they might contribute to the solubi-
lization of the steroid substrate as well. Biotransformations
2.3.
Analytical methods
Samples (0.5 mL) were taken at intervals and extracted with
an equal volume of ethyl acetate; substrate and product con-
centrations were determined by HPLC using a Hypersil ODS
(250 mm × 4.6 mm) column (Supelco Inc., Bellefonte, PA, USA),
UV detection at 210 nm with a Merck-Hitachi 655-22 detector;
a mixture of acetonitrile/water (1/1) was used as eluent with a
flow rate of 0.8 mL min−1. The mobility of substrate and prod-
ucts was: 1 = 12.5 min; 2 = 9.0 min; 3 = 9.9 min, allowing for the
independent measurement of 3␣- and 3␤-HSDH activity.
3.
Results
A preliminary screening was performed among 52 yeasts
belonging to different genera grown for 48 h on medium con-
taining malt extract as carbon source. Neat substrate (1 g/L)
and glucose (25 g/L) were added to whole cultures and the reac-
Table 1 – Reduction of tibolone with different yeasts
3␣–OH (2)/3␤-OH (3)
Molar conversion (%)
Time (h)
Kluyveromyces lactis CBS 2359
Kluyveromyces marxianus CBS 397
Kluyveromyces marxianus CBS 607
Kluyveromyces marxianus CBS 1553
Kluyveromyces marxianus CBS 2231
Pichia anomala CBS 110
>99/1
>99/1
98/2
98/2
99/1
89/11
95/5
12/88
62
33
26
12
35
26
64
34
48
48
48
48
120
48
Pichia pastoris CBS 2612
Saccharomyces cerevisiae CBS 3093
168
168

steroids 7 3 ( 2 0 0 8 ) 112–115
115
ethanol at 65 ◦C before biotransformation showed that this
pre-treatment did not affect the biotransformation.
Fig. 3 shows the results of the biotransformation performed
with different amounts of substrate dissolved in hot ethanol
(65 ◦C) and then added to the suspension of cells.
Almost complete molar conversion was achieved with
initial substrate concentrations below 2 g/L, while marked
inhibition was encountered above 8 g/L. The time-course of
the epimeric ratio of the reaction performed with 2 g/L of sub-
strate under optimised conditions is reported in Fig. 4.
The epimeric ratio ␤/␣ changed along the transformation,
since the ␣-epimer was formed only during the first hours; the
final ␤/␣ ratio was 96/4.
K. lactis CBS 2359 was used for the optimization of the bio-
transformation aimed at the obtainment of 3␣-OH-tibolone
2. The same strategy for the optimization of the biotransfor-
mation with S. cerevisiae was followed. The best results were
obtained by adding micronized substrate to a yeast suspension
in the presence of glucose (50 g L−1) starting from a substrate
concentration of 1.5 g L−1. The employment of these condi-
tions yielded 95% molar conversion within 60 h, furnishing an
epimeric ratio ␣/␤ of 99/1.
Fig. 3 – Influence of initial substrate concentration on the
reduction of tibolone with S. cerevisiae CBS 3093.
r e f e r e n c e s
[1] Albertazzi P, Di Micco R, Zanardi E. Tibolone: a review.
Maturitas 1998;30:295–305.
[2] Hammar M, Christau S, Nathorst-Boos J, Rud T, Garre K. A
doubleblind, randomised trial comparing the effects of
tibolone and continuous combined hormone replacement
therapy in postmenopausal women with menopausal
symptoms. BJOG 1998;105:904–11.
[3] Devogelaer JP. A review of the effects of tibolone on the
skeleton. Expert Opin Pharmacothr 2004;5:941–9.
[4] Kloosterboer HJ. Tibolone: a steroid with a tissue-specific
mode of action. J Steroid Biochem Mol Biol 2001;76:231–8.
[5] Schoonen WGEJ, Deckers G, de Gooyer ME, de Ries R,
Kloosterboer HJ. Hormonal properties of norethisterone,
7-methyl-norethisterone and their derivates. J Steroid
Biochem Mol Biol 2000;74:213–22.
Fig. 4 – Time-course of the formation of 3␣-hydroxy and
3␤-hydroxy tibolone during the reduction of tibolone with
S. cerevisiae CBS 3093.
[6] Colombo D, Ferraboschi P, Ronchetti F, Toma L. Magn Reson
Chem 2002;40:581–8.
[7] Ferraboschi P, Colombo D, -Elahi SR. A practical
chemoenzymatic approach to the synthesis of 3-hydroxy
metabolites of tibolone. Tetrahedron: Asymmetry
2002;13:2583–6.
were, therefore, carried out by adding the substrate as an
ethanol solution at different temperatures and ethanol con-
centrations. The addition of the substrate in ethanol (5% v/v
final concentration of ethanol in the biotransformation) at
65 ◦C strongly improved rate and molar conversion (96% after
24 h); an experiment where the whole cells were treated with
[8] Walters FH, Parker LR, Morgan SL, Deming SN. Sequential
simplex optimization. Boca Raton: CRC Press; 1991.