M.L. Mascotti et al. / Steroids 109 (2016) 44–49
45
spraying the p-anisaldehyde reagent, as described elsewhere. To
assess the biotransformation profile of each substrate, samples
were analyzed by GC-FID using a Clarus 500 instrument (Perkin
Elmer) using pure analytic standards. Conversion was determined
using a 007 (methyl 5% phenyl silicone) capillary column (Quad-
rex) and calculated using relative areas. Product purifications were
done by column chromatography on silica gel 60 (230-400 mesh,
Merck). Mixtures of Hx: EtOAc were initially used as mobile
phases, changing the composition from 100:0 to 30:70 by a 10%
per column volume (Vc). Finally, two Vc of pure EtOAc followed
by two Vc of EtOAc: MeOH (90:10) were employed. The identity
of the isolated compounds was determined by NMR, HRMS-ESI
and GC/LRMS-EI. The 1H and 13C NMR spectra were recorded on
a AC-200 or AMX-400 spectrometer (at 200.13 or 400.16 MHz,
and at 50.23 or 100.62 MHz, respectively) (Bruker) using CDCl3
as solvent with TMS as internal standard. HRMS-ESI measurements
were done on a microTOF-Q II (Bruker) coupled to a UPLC-DAD
1200 (Agilent) at a resolution of 5000 (5% valley definition), by
70 eV electron ionization, at an accelerating voltage of 8 kV. GC/
LRMS-EI were performed at 70 eV using an ion trap (GCQ Plus)
with MSn (Finnigan, Thermo-Quest), operated at a fundamental
rf-drive of 1.03 MHz. Helium was used as the damping gas at an
uncorrected gauge reading of 6Á10À5 Torr. An OV-5 (5% diphenyl
95% dimethylpolysiloxane) capillary column (OVS) was used.
Compounds were analyzed by comparing their mass spectra with
the NIST spectral library (Wiley). Structural data is given in the
Supplementary Information.
2. Experimental
2.1. Chemicals and microorganisms
All the chemicals were purchased in analytical grade from
Sigma Aldrich or Merck and were used without further purifica-
tion. Culture media components were obtained from Britania.
The Aspergillus strains used comprised A. flavus, A. fumigatus and
A. japonicus, which were acquired from the collections of Universi-
dad de Buenos Aires (Argentina), Universidad Nacional de Rio
Cuarto (Argentina) and Colección de Cultivos Fúngicos IIB-INTECH
(Argentina), respectively, as well as A. candidus and A. parasiticus,
which were both obtained from Universidad Nacional del Litoral
(Argentina). Microorganisms stored at 4 °C in Czapek Yeast (CY)
agar slopes were inoculated to erlenmeyers containing CY liquid
medium, which composition per 1 L distilled H2O is 5 g yeast
extract, 30 g sucrose, 1 g K2HPO3, 0.3 g NaNO3, 0.5 g KCl, 0.5 g
MgSO4Á7H2O, 0.01 g FeSO4Á7H2O, 0.005 g CuSO4Á5H2O and 0.01 g
ZnSO4Á7H2O. Cultures were incubated at 28 °C and 150 rpm.
2.2. Microbial screening
To test the capability of the Aspergillus strains to transform the
model BVMO substrate rac-bicyclo[3.2.0]hept-2-en-6-one (1), bio-
catalysts were prepared according to previous reports [31,34].
Then, 1 (2.5 mM final concentration) was added to 50 mL erlen-
meyer flasks containing 500 mg wet cells (2.5% (p/v)) of an individ-
ual strain each in 20 mL of CY medium. Biotransformations were
run by duplicate –mean values are given- and incubated at 28 °C
and 150 rpm during 72 h. Controls without 1 and without biocata-
lyst were also run. Samples were taken as described in Section 2.3.1
and analyzed by GC-FID using a Clarus 500 instrument (Perkin
Elmer) and pure analytic standards. Conversion was determined
employing a Rt-bDEXse (Restek) chiral capillary column and calcu-
lated using relative areas of the substrate and the lactone-based
products only. The enantiomeric excesses were calculated as done
by Chen et al. [35]. See Supplementary Information for further
details.
3. Results and discussion
3.1. Screening of BVMO activity in Aspergillus strains
Considering previous reports in literature describing the poten-
tial of the Aspergillus genus as a biocatalyst to perform Baeyer-
Villiger oxidations of the D-ring in several steroids [36–39], we
selected a small collection of strains from this genus to screen
for this enzymatic activity by using rac-bicyclo[3.2.0]hept-2-
en-6-one (1) as model substrate [40]. Since redox cofactors are
needed, we employed growing cell cultures in order to display
self-sufficient transformations. Although all the strains assayed
were capable to oxidize the model substrate 1 to the corresponding
normal and/or abnormal lactones (Table 1, Supplementary
Information), A. parasiticus exhibited the highest regio- and enan-
tioselectivity by yielding mainly the normal (1S, 5R) and the abnor-
mal (1R, 5S) lactones from the enantiomers (1R, 5S) and (1S, 5R) of
substrate 1, respectively (Fig. 1). Since dehydrogenase activity is
ubiquitous and constitutive in these fungi, carbonyl reduction
products were observed as well (data not shown). Because of these
results, A. parasiticus was selected as the biocatalyst to perform the
transformation of DHEA (2).
2.3. Biotransformation of steroids
Biocatalysts were prepared as referred above. In the analytical
scale, 50 mL erlenmeyer flasks containing 500 mg wet cells (2.5%
(p/v)) of A. parasiticus in 20 mL CY medium were used. To start
the biotransformations, each substrate was dissolved in DMSO or
DMF (1% (v/v) final concentration) and added to the erlenmeyer
flasks. Biotransformations were incubated at 28 °C and 150 rpm
and were run by duplicate, mean values are shown. Controls with-
out biocatalyst and without substrate were also included. To per-
form time-course experiments, multiples batches were prepared
and samples were taken by withdrawing batches and controls in
duplicate at a time. In the semipreparative scale, the same proce-
dure described for the analytical scale was followed, but using a
biotransformation volume of 200 mL in a 500 mL erlenmeyer flask.
A total of five batches were run and combined after 96 h.
3.2. Biotransformation of dehydro-epi-androsterone (DHEA)
When the biotransformation of 2 was assayed in analytical scale
(100 mg), five products were detected (Fig. 2) by GC/ LRMS-EI and
NMR analyses (see Supplementary Information). Among them,
three compounds are supposed to directly derive from the
substrate, being androstenediol (3) –a product of the reduction of
the carbonyl in the D-ring-, androstenedione (4) –obtained by
the oxidation of hydroxyl in position 3 and the subsequent isomer-
ization of the C@C from C5 to C4- and the homo-lactone 3b-
2.3.1. Work-up and analysis
Each batch was filtered in vacuo to remove cells and then reac-
tion mixtures were extracted three times with one volume of ethyl
acetate. Organic layers were combined, dried with sodium sulfate,
concentrated in vacuo and spotted on silica gel 60 F254 TLC plates
(Merck) to qualitatively assess the substrate consumption and the
appearance of biotransformation products. The chromatographic
analysis was done using the mixture n-hexane: ethyl acetate as
mobile phase and plates were revealed under UV light and by
hydroxy-17a-oxa-D-homoandrost-5-en-17-one (5), produced after
a D-ring lactonization reaction. Supported in literature reports, it
is possible to propose that these molecules were obtained by the
action of a dehydrogenase, a 3b-HSD/isomerase (E2) and a BVMO