Steroid hydroxylations with Botryodiplodia malorum and Colletotrichum lini

Article abstract of DOI:10.1016/j.steroids.2006.01.014

An improved procedure for the microbial hydroxylations of dehydroepiandrosterone (DHEA, 1) and 15β,16β-methylene-dehydroepiandrosterone (2) was studied using whole cells of Botryodiplodia malorum and Colletotrichum lini. C. lini catalyzed 7α- and 15α-hydroxylation of 1 and 7α-hydroxylation of 2, while B. malorum gave 7β-hydroxylation of both the substrates. The stability of the enzymatic activity was higher in the presence of co-substrates (i.e., glucose or mannitol) allowing for repeated batches of the biotransformations. The yields of 7α,15α-dihydroxy-1 production were improved obtaining 5.8 g l^-1 (recovered product) from 7.0 g l^-1 of substrate. The structures of the hydroxylated products were assigned by a combination of two-dimensional NMR proton-proton and proton-carbon correlation techniques.

Full text of DOI:10.1016/j.steroids.2006.01.014

steroids 7 1 ( 2 0 0 6 ) 429–434
available at www.sciencedirect.com
journal homepage: www.elsevier.com/locate/steroids
Steroid hydroxylations with Botryodiplodia malorum and
Colletotrichum lini
a
a
b
c
c
Andrea Romano , Diego Romano , Enzio Ragg , Francesca Costantino , Roberto Lenna ,
d
a,∗
Raffaella Gandolfi , Francesco Molinari
a
Dipartimento di Scienze e Tecnologie Alimentari e Microbiologiche, Universit a` degli Studi di Milano, via Celoria 2,
0133 Milano, Italy
Dipartimento di Scienze Molecolari Agroalimentari, Universit a` degli Studi di Milano, via Celoria 2, 20133 Milano, Italy
Industriale Chimica, via Grieg 13, 21047 Saronno (VA), Italy
2
b
c
d
Istituto di Chimica Organica “Alessandro Marchesini”, Universit a` degli Studi di Milano, via Venezian 21, 20133 Milano, Italy
a r t i c l e i n f o
a b s t r a c t
Article history:
An improved procedure for the microbial hydroxylations of dehydroepiandrosterone (DHEA,
1) and 15␤,16␤-methylene-dehydroepiandrosterone (2) was studied using whole cells of
Botryodiplodia malorum and Colletotrichum lini. C. lini catalyzed 7␣- and 15␣-hydroxylation of
1 and 7␣-hydroxylation of 2, while B. malorum gave 7␤-hydroxylation of both the substrates.
The stability of the enzymatic activity was higher in the presence of co-substrates (i.e., glu-
cose or mannitol) allowing for repeated batches of the biotransformations. The yields of
Received 4 April 2005
Received in revised form 28
December 2005
Accepted 11 January 2006
Published on line 31 March 2006
−
1
7
␣,15␣-dihydroxy-1 production were improved obtaining 5.8 g l (recovered product) from
7.0 g l 1 of substrate. The structures of the hydroxylated products were assigned by a combi-
nation of two-dimensional NMR proton–proton and proton–carbon correlation techniques.
© 2006 Elsevier Inc. All rights reserved.
−
Keywords:
Hydroxylation
Botryodiplodia malorum
Colletotrichum lini
Dehydroepiandrosterone (DHEA)
15␤,16␤-Methylene-3-␤-hydroxy-5-
androsten-17-one
1
.
Introduction
production of important pharmaceutical intermediates [2].
Hydroxylation at positions 7 and 15 of dehydroepiandros-
Microbial modification of steroids is
method for the preparation of a number of hydroxysteroids
1–3], but large scale transformations are sometimes limited
by low yields due to low solubility of the substrates and by
inhibition effects [4]. Microbial hydroxylations are among
the most studied and useful transformations since they
can be achieved under mild conditions with high chemo-,
regio- and stereoselectivity [5]. Hydroxylation at positions
a
long-established
terone (3␤-hydroxy-5-androsten-17-one, DHEA, 1) has been
reported by using Colletotrichum lini allowing for the prepa-
[
ration of 3␤,7␣,15␣-trihydroxy-5-androsten-17-one, while
5
a
process for preparing 3␤,7␤-dihydroxy-ꢀ -steroids is
based on the hydroxylation of the position 7 of 15␤,6␤-
methylene-3-␤-hydroxy-5-androsten-17-one (2) catalyzed
by Botryodiplodia malorum [7]. The products of these bio-
transformations (3␤,7␣,15␣-trihydroxy-5-androsten-17-one
5
7
and 15 with controlled stereoselectivity allows for the
and 3␤,7␤-dihydroxy-ꢀ -steroids) are key-intermediates for
∗
Corresponding author. Tel.: +39 0250316695; fax: +39 0250316694.
E-mail address: francesco.molinari@unimi.it (F. Molinari).
0039-128X/$ – see front matter © 2006 Elsevier Inc. All rights reserved.
doi:10.1016/j.steroids.2006.01.014

4
30
steroids 7 1 ( 2 0 0 6 ) 429–434
Fig. 1 – Hydroxylations of dehydroepiandrosterone (DHEA, 1) and 15␤,16␤-methylen-dehydroepiandrosterone (2) with
Botryodiplodia malorum and Colletotrichum lini.
◦
the synthesis of pharmacologically active steroids, such as
␤,7␤,15␤,16␤-dimethylene-3-oxo-17␣-pregn-4-ene-21,17-
carbolactone, an aldosterone antagonist [6,7].
at 110 C for 24 h. Dry weight biomass was in the range of
6
1.5–1.6 g l 1. C. lini was grown using two cultivation steps; a
−
−
vegetative phase using a liquid medium with glucose (30 g l 1)
−
1
◦
The importance of these molecules led us to study methods
for improving the performances of the microbial transforma-
tion, such as manipulation of the reaction medium or immobi-
lization of the mycelium. We have also studied the possibility
of expanding the applicability of these biotransformations for
obtaining both the 7-OH epimers of 1 and 2. The complete
pattern of the modifications studied in this work is reported
in Fig. 1.
corn steep liquor (10 g l ) and tap water at pH 7.0, 25 C, recip-
rocal shaking (120 spm). The next phase was a fermentative
phase where a richer liquid medium was employed (glucose
30 g l 1, corn steep liquor 10 g l , soy meal 10 g l , NaNO3
−
−1
−1
−
1
−1
−1
−1
2 g l , KH2PO4 1 g l , K2HPO4 2 g l , MgSO4·7H2O 0.5 g l
,
−
1
−1
KCl 0.5 g l , FeSO4·7H2O 20 mg l , pH 7.0). The liquid cultures
were carried out in 2.0 l Erlenmeyer flasks containing 200 ml of
◦
medium at 25 C for 72 h under reciprocal shaking (120 spm).
Dry weight biomass was in the range of 8.9–9.1 g l 1.
−
2
.
Experimental
2
.2.
Biotransformations
2
.1.
Microorganisms and microbial growth
Substrates were kindly furnished by Industriale Chimica
Saronno, Italy). Biotransformations with submerged cultures
(
B. malorum CBS 134.50 and C. lini CBS 112.21 were from Cen-
were carried out in 2.0 l Erlenmeyer flasks containing 200 ml
of the biotransformation suspension. Biotransformations with
resting mycelia were carried out by harvesting the mycelium
from submerged cultures by filtration on paper filters and by
suspending the mycelium in aqueous buffers (200 ml). Reac-
tions were started by addition of the substrates, followed
by addition of co-solvents and/or co-substrates; the incuba-
tion was performed under reciprocal shaking (120 spm) and
the reaction was followed by HPLC. The biotransformation
traalbureau voor Schimmelcultures (Baarn, Holland) and were
routinely maintained on PDA agar (potato infusion 200 g l 1,
−
glucose 20 g l 1, agar 15 g l , pH 5.6). B. malorum was cultured
−
−1
in 2.0 l Erlenemeyer flasks containing 200 ml of liquid medium
(
soy meal 10 g l 1, glucose 10 g l , pH 6.2) at 25 C for 72 h
− −1 ◦
under reciprocal shaking (120 strokes per minute, spm). The
dry weights were determined after centrifugation of 100 ml of
cultures. Mycelia were washed with distilled water and dried

steroids 7 1 ( 2 0 0 6 ) 429–434
431
medium was then filtered and extracted with methyl isobutyl
ketone. The organic extracts were dried over Na2SO4 and the
solvent was removed; the crude product was purified by flash
chromatography (hexane/acetone, 1/1, vanillin stained).
polynomial. Chemical shifts were measured in ı (ppm), using
the acetone residual signal as a reference and setting the
methyl proton and carbon resonances at 2.1 and 29.8 ppm,
respectively.
The [˛]D were measured in methanol (c 1.0) using a Perkin-
Elmer Polarimeter (Model 343 Plus). Melting points were mea-
sured by differential scanning calorimetry analysis (Perkin-
Elmer DSC 6).
2
.3.
Immobilization
The mycelia obtained by filtration of the culture broths (200 ml)
were suspended in 60 ml of sodium alginate (Sigma Chem-
icals) (3.0%, w/w) dissolved in distilled water. The resulting
suspension was degassed and extruded through a thin injec-
tion needle into 50 ml of an iced calcium chloride solution
3.
Results
(
0.1 M). The bead size had an average diameter in the range
3.1.
Biotransformations with C. lini
◦
of 2.5–2.8 mm. After hardening for 3 h at 4 C, the beads were
filtered under vacuum and used for the biotransformation or
stored at 4 C in fresh calcium chloride (0.01 M) solution. The
amount of mycelium in the beads was estimated by dissolv-
ing 10 g of beads in 2% sodium poliphosphate solution; the
suspension was then filtered through a pre-dried weighted fil-
C. lini showed no extracellular activity in the hydroxylation
of 1; therefore, the optimization of the biotransformation was
performed using mycelium suspended in aqueous media. The
optimization was carried out by simultaneously evaluating
different parameters of the biotransformation (temperature,
pH, type of buffer, substrate and mycelium concentration)
using the Multisimplex experimental design [8]. The profile
◦
◦
ter paper (Whatman No. 1) and dried at 100 C to a constant
weight.
◦
of the hydroxylation of 1 under optimized conditions (30 C,
−
concentration of the mycelium = 10 g l 1, concentration of the
2
.4.
Repeated use of mycelium
−
1
substrate = 1 g l in phosphate buffer at pH 6.4) is reported in
Fig. 2.
Mycelium was reused in successive biotransformation cycles.
At the end of 24 h, the mycelium was paper filtered and resus-
pended in fresh reaction mixture, and substrate was added.
The reaction was discontinued when the yield was less than
The reaction occurred stepwise with formation of the 7␣-
◦
hydroxy derivative (3, mp = 216–218 C, [˛]D = −56.86) before
formation of the known compound 4.
8
0% after 24 h.
One of the main drawbacks in steroid biotransformation is
the low solubility of the substrates in water, which diminishes
reaction rates and overall productivity. Different methods for
improving the solubility of the substrate were investigated,
such as the use of co-solvents (acetone, dimethylformamide,
dimethylsulfoxide, ethanol and mixtures of these solvents),
and addition of different cyclodextrins. The use of dimethyl-
formamide (2%, v/v) gave the highest conversion rates and
yields, while cyclodextrins did not significantly improve the
performance of the bioconversion. Table 1 shows the results
2
.5. Analytical methods
Samples (0.5 ml) were taken at intervals and extracted with an
equal volume of methyl isobutyl ketone; substrate and product
concentrations were determined by HPLC using a Lichrospher
6
0 Select B column (Merck, Darmstadt, Germany), UV detec-
tion at 210 nm with a Merck-Hitachi 655-22 detector and ace-
tonitrile as eluent with a flow rate of 1.0 ml min 1. The mobility
of substrates and products was: 1 = 4.1 min; 2 = 4.2 min; 3 and
−
◦
of the biotransformation in phosphate buffer (pH 6.0) at 30 C
5
= 5.7 min; 4 = 6.4 min; 6 and 7 = 5.6 min.
NMR spectroscopy. NMR spectra were performed at room
performed with different amounts of substrate dissolved in
dimethylformamide.
temperature on
a Bruker AMX-600 (Bruker Spectrospin,
Rheinstetten, Germany), operating at 600.1 MHz frequency
for the proton nucleus, using a 5 mm broad-band reverse
probe equipped with gradients. Suitable amounts of sam-
ple (between 3 and 10 mg) were dissolved in acetone-d6
(
ISOTEC, USA) and transferred into 5 mm NMR tubes (type 535-
PP, ALDRICH). Two-dimensional proton–proton (TOCSY and
COSY-DQF) and gradient based carbon–proton chemical shift
correlation experiments were performed using standard pulse
sequences, present in the spectrometer library. One-bond
(
HMQC) and multiple-bond (HMBC) carbon–proton shift corre-
lation experiments were performed assuming carbon–proton
coupling constant values of 145 and 8 Hz, respectively. HMQC
spectra were acquired with Broad Band-proton decoupling
using a WALTZ-16 pulse sequence. Spectra were processed
using XWINNMR software (Bruker) on a Silicon Graphics INDY
workstation. Raw data were transformed to 2K × 2K real data
Fig. 2 – Hydroxylation of dehydroepiandrosterone (1, ᭹)
into 3␤,7␣-dihydroxy-5-androsten-17-one (3, ꢀ) and
◦
points with a 90 shifted-sine bell squared weighting func-
3␤,7␣,15␣-trihydroxy-5-androsten-17-one (4, ꢁ) with
tion and zero-filling. Baseline was corrected with a fifth degree
Colletotrichum lini.

4
32
steroids 7 1 ( 2 0 0 6 ) 429–434
Table 1 – Production of
3
3
␤,7␣-dihydroxy-5-androsten-17-one (3) and
␤,7␣,15␣-trihydroxy-5-androsten-17-one (4) by
biotransformation using different concentrations of
dehydroepiandrosterone (DHEA, 1) with Colletotrichum
lini
−1
−1
−1
Time (h)
1 g l
2.5 g l
5 g l
3
4
3
4
3
4
4
0.55
0
0.52
1.05
0
0
0
0
0
0
0
2
4
7
4
8
2
0
0
0
0.98
1.02
0.85
0.97
1.94
1.80
0.50
1.50
2.75
0
Fig. 4 – Repeated batches of hydroxylation of
dehydroepiandrosterone (DHEA, 1) into
3
␤,7␣-dihydroxy-5-androsten-17-one (3) with
−1
Starting from 2.0 g l of substrate, complete conversion of
into 4 was obtained with little degradation of the product at
Colletotrichum lini.
1
prolonged times. The reaction was very slow with a substrate
concentration above 3 g l 1, and predominantly compound 3
−
1
added. The substrate was added in a dimethylformamide solu-
tion but was also finely ground and dispersed in a Tween 80
solution (0.1%, v/v); results are reported in Fig. 4. Glucose con-
centration was kept around 10 g l by means of continuous
addition.
−
was formed. When 5 g l of substrate were employed, 3 was
the only product observed even at prolonged times. The re-
usability of the mycelium was tested, but the stability was
quite poor since after the third addition, the activity was very
slow. Two strategies were attempted for improving the oper-
ational stability of the enzymes involved: immobilization of
the mycelium in Ca-alginate and addition of co-substrates
−
1
Repeated-batch operation was maintained for seven cycles,
before decrease of the enzymatic activity could be observed.
The use of Tween 80 guaranteed higher operational stability
(
10 g l 1) for maintaining the enzymes in a metabolically active
−
of the catalyst. The overall product recovery was 5.8 g l of 4
−1
−
1
state (Fig. 3).
starting from 7 g l of substrate.
The immobilization resulted in a lower production of the
desired molecule and did not significantly improve the stabil-
ity of the enzymatic activity, while the use of co-substrates,
such as glucose and mannitol, proved effective as stabilizing
agents.
Mycelium of C. lini was also used for the hydroxylation of
2 under the optimized conditions found for the biotransfor-
mation of 1. The only product produced was identified as the
7␣-derivative 7 (mp not determined because the compound
decomposed before reaching melting point, [˛]D = −67.30) with
complete molar conversion within 72 h.
The mycelium was also used in repeated batch reactions:
after 24 h of reaction, the mycelium was filtered, the product
was recovered from the aqueous solution and the mycelium
was resuspended in fresh reaction buffer; at the beginning of
3.2.
Biotransformations with B. malorum
−1
−1
each reaction cycle, 1 g l of 1 and 10 g l of glucose were
The biotransformation of 1 was also studied with B. malo-
rum. It was followed the same approach used for optimising
the bioconversion with C. lini (influence of pre-treatment of
the substrate, addition of co-solvents and co-substrates, effect
of substrate concentration, immobilization of the mycelium).
The best results were obtained by using free mycelium in the
presence of dimethylsulfoxide (2%) starting from a substrate
concentration of 1.0 g l 1. The employment of these conditions
−
yielded 70% molar conversion within 72 h, furnishing com-
◦
pound 5 (mp = 188–200 C, [˛]D = −6.60), as the only detectable
product. No traces of hydroxylation in other position could be
detected after 5 days.
The hydroxylation of 2 with B. malorum was carried out
under the optimal conditions found for the biotransformation
of 1; the reaction gave the known 7␤-hydroxy derivative 6 with
−
1
complete conversion starting from 1.0 g l of substrate in 72 h.
3
.3. Determination of the hydroxylation positions and
1
13
relative configurations by H and C NMR
Fig. 3 – Re-usability of Colletotrichum lini as free and
immobilized cells. Yields of
The structures of the hydroxylation products of 1 and 2 were
determined by a combination of two-dimensional NMR tech-
niques. Fig. 5 shows the expansions of the COSY-DQF and
3
␤,7␣,15␣-trihydroxy-5-androsten-17-one (4) through three
reaction cycles (72 h).

steroids 7 1 ( 2 0 0 6 ) 429–434
433
Fig. 6 – Selected traces of the 13C/ H HMBC spectrum
600 MHz) performed on a D2O-exchanged sample of
␤,7␣,15␣-trihydroxy-5-androsten-17-one (4), dissolved in
acetone-d6. Trace a: 1D spectrum.
1
(
3
Fig. 5 – (A) Expansion of a COSY-DQF spectrum (600 MHz,
performed on D2O-exchanged) of
3
␤,7␣,15␣-trihydroxy-5-androsten-17-one (4) dissolved in
acetone-d6. The mono-dimensional spectrum is shown on
top. (B) Expansion of a C/ H HMQC spectrum for the same
compound.
CH-7, whose resonance was assigned at 4.2 ppm, and with
CH3-19 (0.95 ppm). The hydroxylation at position 7 was also
proved by a TOCSY connectivity (not shown) between CH-7
and the olefinic CH-6 (5.5 ppm). The last proton at 4.4 ppm
should, therefore, be assigned to CH-15. This position was con-
firmed by the connectivity observed in the COSY-DQF experi-
ment between CH-15 and the methylene protons CH2-16 (2.05
and 2.90 ppm, Fig. 5) and by the two-bond carbon–proton
interactions involving the same methylene protons and C-
15. The assignment of the geminal protons at 2.05 and
2.9 ppm to the CH2-16 group was confirmed by the two-bond
carbon–proton connectivity with C-17 (198 ppm, Fig. 6d). The
same trace showed the correlation of C-17 with protons CH3-
18 (0.85 ppm), providing a cross-check for the assignments.
The relative configurations of the hydroxylated carbons were
determined by means of the proton–proton vicinal coupling
constant values. The 3JHH value between protons CH-7 and
CH-8 was measured less than 4 Hz (confirmed by the weak
interaction displayed in the COSY-DQF spectrum), and such
1
3
1
HMQC spectra for 7␣,15␣-dihydroxy 1. For the same molecule,
traces of the HMBC spectrum are reported in Fig. 6 for the most
significant carbon resonances. The HMQC experiment clearly
identifies, in the 60–70 ppm 13C chemical shift region, all pro-
tonated carbons directly attached to the hydroxyl groups. The
hydroxylated carbon at position 3 was already present in 1
and other related steroids, such as cholesterol; thus, its reso-
nance was readily assigned by analogy with literature data [9].
An independent assignment was made by observing the long-
2
range connectivity between the unique quaternary sp carbon
at position C-5 (144 ppm) and the methylene protons CH2-
4
(2.2–2.3 ppm, Fig. 6b), which, in turn, correlated by vicinal
couplings with the proton CH-3 in the COSY-DQF experiment
Fig. 5). C-5 had another long-range interaction with proton
(

4
34
steroids 7 1 ( 2 0 0 6 ) 429–434
−
1
a low value indicated that the relative orientation must have
been sin. Since CH-8 was known to be ␤, CH-7 must also have
been ␤ and, consequently, OH-7 was ␣. In the same compound,
OH-15 was found to be ␣ because of the large vicinal cou-
pling constant value with H-14 ( J14,15 = 9.7 Hz, J15,16␣ = 7.9 Hz,
3J14,15␤ = 6.2 Hz). Such values were measured from the analysis
of the H-15 multiplet in a resolution-enhanced 1D spectrum
and the corresponding COSY-DQF cross-peak.
the recovery of 5.7–5.8 g l
conversion.
of product with 75–80% molar
r e f e r e n c e s
3
3
[
1] Murray HC, Peterson DH. Upjohn Co., USA. Oxygenation of
steroids by Mucorales fungi. U.S. Patent 2602769; 1952.
With the above described reasoning, the relative configu-
rations of the hydroxylation sites in the isolated derivatives of
[2] Mahato B, Mazumder I. Current trends in microbial steroid
transformations. Phytochemistry 1989;34:883–98.
[
3] Fernandes P, Cruz A, Angelova B, Pinheiro HM, Cabral JMS.
Microbial conversion of steroid compounds: recent
developments. Enzyme Microb Technol 2003;32:688–705.
4] Mahato BM, Garai S. Advances in microbial steroid
biotransformation. Steroids 1997;62:332–45.
5] Holland HL. Recent advances in applied and mechanistic
aspects of the enzymatic hydroxylation of steroids by
whole-cell biocatalyst. Steroids 1999;64:178–86.
2
were similarly determined.
[
[
4
.
Conclusions
The hydroxylation of dehydroepiandrosterone 1 and 15␤,16␤-
methylene-dehydroepiandrosterone 2 by microbial means has
been optimized in terms of yields and selectivity. A suited
combination of substrates and whole-cell biocatalysts allowed
for the production of different hydroxylated compounds (3–7),
as optically pure epimers. The structures and relative con-
figurations of the hydroxylated products were assigned by
a combination of two-dimensional NMR proton–proton and
proton–carbon correlation techniques.
[
[
[
6] Petzoldt K, Laurent H, Wiechert R. Schering, Germany.
3␤,7␤,15␣-Trihydroxy-5-androsten-17-one, its 3,15-dipivalate,
and their preparation. U.S. Patent 4435327; 1984.
7] Petzoldt K, Wiechert R, Laurent H, Nicklasch K, Bittler D.
Schering, Germany. Process for preparing
5
3
␤,7␤-dihydroxy-ꢀ -steroids. U.S. Patent 4416985, in press.
8] Walters FH, Parker LR, Morgan SL, Deming SN. Sequential
simplex optimization. Boca Raton: CRC Press; 1991.
The production of 3␤,7␣,15␣-trihydroxy-5-androsten-17-
one was performed under repeated batch mode allowing for
[9] Breitmaier E, Voelter W. 13C NMR spectroscopy. Weinheim,
New York: Verlag Chemie; 1978.