CYP260B1 acts as 9α-hydroxylase for 11-deoxycorticosterone

Article abstract of DOI:10.1016/j.steroids.2017.08.006

Steroids and their oxyfunctionalized counterparts are valuable compounds for the pharmaceutical industry; however, the regio- and stereoselective introduction of oxygen is a challenging task for the synthetic chemistry. Thus, cytochromes P450 play an important role for the functionalization of steroidal compounds. In this study, we elucidated the main product of 11-deoxycorticosterone conversion formed by CYP260B1 from Sorangium cellulosum So ce56 as 9α-OH 11-deoxycorticosterone by NMR spectroscopy. This is, to the best of our knowledge, the first identification of a 9α-hydroxylase for this substrate. In addition, the major side product was identified as 21-OH pregna-1,4-diene-3,20-dione. Studies using 1α-OH 11-deoxycorticosterone as substrate suggested that the major side product is formed via dehydrogenation reaction. This side reaction was considerably decreased by employing the CYP260B1-T224A mutant, which showed an increased selectivity of about 75% compared to the 60% of the wild type for the 9α-hydroxylation. To scale up the production, an E. coli based whole-cell system harboring the CYP260B1-T224A variant as well as two heterologous redox partners was used. Employing growing cells in minimal medium led to a productivity of about 0.25 g/l/d at a 50 ml scale showing the biotechnological potential of this system.

Full text of DOI:10.1016/j.steroids.2017.08.006

Accepted Manuscript
CYP260B1 acts as 9α-hydroxylase for 11-deoxycorticosterone
Martin Litzenburger, Rita Bernhardt
PII:
S0039-128X(17)30138-1
http://dx.doi.org/10.1016/j.steroids.2017.08.006
STE 8140
DOI:
Reference:
Steroids
To appear in:
Received Date:
Accepted Date:
10 July 2017
15 August 2017
Please cite this article as: Litzenburger, M., Bernhardt, R., CYP260B1 acts as 9α-hydroxylase for 11-
deoxycorticosterone, Steroids (2017), doi: http://dx.doi.org/10.1016/j.steroids.2017.08.006
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers
we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and
review of the resulting proof before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

CYP260B1 acts as 9α-hydroxylase for 11-deoxycorticosterone
Martin Litzenburger and Rita Bernhardt
Saarland University, Institute of Biochemistry, Campus B.2.2, 66123 Saarbruecken, Germany
Corresponding author:
Rita Bernhardt
Saarland University, Institute of Biochemistry, Campus B2.2, 66123 Saarbruecken, Germany
E-mail: ritabern@mx.uni-saarland.de
Tel:
+49 (0)681 302 4241
Fax: +49 (0)681 302 4739
Abstract:
Steroids and their oxyfunctionalized counterparts are valuable compounds for the pharmaceutical
industry; however, the regio- and stereoselective introduction of oxygen is a challenging task for the
synthetic chemistry. Thus, cytochromes P450 play an important role for the functionalization of
steroidal compounds. In this study, we elucidated the main product of 11-deoxycorticosterone
conversion formed by CYP260B1 from Sorangium cellulosum So ce56 as 9α-OH 11-
deoxycorticosterone by NMR spectroscopy. This is, to the best of our knowledge, the first
identification of a 9α-hydroxylase for this substrate. In addition, the major side product was identified
as 21-OH pregna-1,4-diene-3,20-dione. Studies using 1α-OH 11-deoxycorticosterone as substrate
suggested that the major side product is formed via dehydrogenation reaction. This side reaction was
considerably decreased by employing the CYP260B1-T224A mutant, which showed an increased
selectivity of about 75% compared to the 60% of the wild type for the 9α-hydroxylation. To scale up
the production, an E. coli based whole-cell system harboring the CYP260B1-T224A variant as well as
two heterologous redox partners was used. Employing growing cells in minimal medium led to a
productivity of about 0.25 g/l/d at a 50 ml scale showing the biotechnological potential of this system.
1

Keywords:
Steroids, 11-deoxycorticosterone, CYP260B1, Sorangium cellulosum So ce56, cytochromes P450, 9α-
hydroxylase.
Abbreviations:
AdR, adrenodoxin reductase; Adx4-108, adrenodoxin (truncated form); P450, cytochrome P450; DOC,
11-deoxycorticosterone; HP-β-CD, 2-hydroxypropyl-β-cyclodextrin; T224A mutant, CYP260B1-
T224A.
2

Introduction:
Steroids are sterol-derived compounds of specific structure that contain 3 cyclohexane rings and a
cyclopentane ring. They have been identified in many living systems such as mammalians, insects,
plants, yeasts or fungi, in which they play important roles like acting as hormones [1]. In mammalian
organisms, steroid hormones are capable of controlling the salt and water balance
(mineralocorticoids), the glucose metabolism (glucocorticoids) or the sexual differentiation and
reproduction (sexual hormones). Their specific roles are thereby dependent on functional groups
attached to the steroidal scaffold like the hydroxyl group at position C11β that is crucial for the anti-
inflammatory activity of steroids [2]. Such oxyfunctionalizations are mainly performed by
cytochromes P450, a superfamily of heme-thiolate proteins named due to the spectral absorbance peak
of their carbon-monoxide-bound species at 450 nm. As steroids and their functionalized counterparts
are of high pharmaceutical interest representing the second largest group of drugs next to antibiotics
[3], the investigation of new P450s with high activity and selectivity as well as the discovery of
enzymes with novel functionalities are important issues for the white biotechnology [4, 5]. Besides
mammalian P450s involved in steroid biosynthesis, several bacterial P450s were already identified to
act as steroid hydroxylases [6]. For instance, the selective hydroxylation of testosterone at diverse
positions by bacterial P450s using an E. coli expression system was previously demonstrated [7]. In
addition, CYP106A1 and CYP106A2 from Bacillus megaterium strain ATCC 13368 were identified
as versatile steroid hydroxylases with a high biotechnological potential [8-11]. Improvements of
CYP106A2 by different mutagenesis approaches to alter the selectivity and enhance the activity were
performed, too [12, 13]. Aside from these examples, P450s from Sorangium cellulosum So ce56 are
also able to functionalize steroids as shown for CYP260A1 and its novel 1α-hydroxylase activity
towards androgens and derivatives of them as well as CYP260B1 and its 6β-hydroxylase activity
towards 11-deoxycortisol [14, 15]. The selectivity of steroid conversions using bacterial P450s is often
low; however, mutagenesis approaches as demonstrated for CYP260A1 as well as CYP106A2 led to
an improved or even altered selectivity of hydroxylation [13, 16].
CYP260B1 was previously identified as steroid hydroxylase with a broad substrate range, in which
DOC was converted with a high activity. However, the main product was not identified due to the low
selectivity of hydroxylation [15]. In this study, we elucidated the structure of this main product as 9α-
OH DOC by NMR spectroscopy. To reach a higher selectivity, the previously created T224A mutant
of CYP260B1 that already caused an altered selectivity for progesterone hydroxylation was tested
[17]. With that variant, the major side product formation was considerably decreased leading to an
increased selectivity of 75% towards 9α-OH DOC during in vitro and in vivo experiments. A whole-
cell system harboring the T224A variant as well as two heterologous redox partners was used to scale
up the production. Utilizing M9CA medium and EtOH as solvent for DOC led to a production of
about 0.25 g/l/d in shaking flasks.
3

Experimental:
Chemicals and strains:
Isopropyl β-D-1-thiogalactopyranoside and 5-aminolevulinic acid were purchased from Carbolution
Chemicals (Saarbruecken, Germany). Bacterial media were obtained from Becton Dickinson
(Heidelberg, Germany). The E. coli strain C43(DE3) was obtained from Novagen (Darmstadt,
Germany). Standards of 9α-OH DOC and 11α-OH DOC were a kind gift from Schering AG (Berlin,
Germany). 1α-OH DOC was produced and purified as described previously [18]. All other chemicals
were purchased from standard sources in the highest purity available.
Expression and purification of the enzymes:
The expression and purification of CYP260B1 and its T224A variant was performed as described
previously [15, 17]. The truncated bovine adrenodoxin (Adx4-108) and adrenodoxin reductase were
expressed and purified as described elsewhere [19, 20].
Determination of the binding constants:
Dissociation constants were determined by difference spectroscopy using tandem cuvettes as described
elsewhere [21]. CYP260B1 and the T224A variant (5 µM) were diluted in 50 mM potassium
phosphate buffer (pH 7.4) and titrated with increasing concentrations of DOC from a 0.5 mM stock
solution in dimethyl sulfoxide. Difference spectra were recorded from 350 to 500 nm. To determine
the binding dissociation constant (Kd), the averaged peak-to-trough absorbance differences (ΔAmax
)
were plotted against the substrate concentration (see Supplemental Figures 1a and b). Plots were fitted
with Origin 8.6 software (OriginLab Corporation, Northampton, MA) by tight binding quadratic
equation [ΔA =(Amax/2[E]){(Kd+[E]+[S])-{(Kd+[E]+[S])2– 4[E][S]}1/2}], whereby ΔA represents the
peak-to-trough absorbance difference at every substrate concentration, A is the maximum absorbance
difference at saturation, [E] is the enzyme concentration, and [S] is the substrate concentration. All
titrations were performed in triplicate.
In vitro conversions:
A reconstituted in vitro system containing CYP260B1 or its T224A variant (0.5 µM), AdR (1.5 µM),
Adx4-108 (10 µM), MgCl2 (1 mM), glucose-6-phosphate (5 mM) and glucose-6-phosphate
dehydrogenase (1U) in a final volume of 250 µl of potassium phosphate buffer (20 mM, pH 7.4) was
used. The steroids (10 mM stock solution in EtOH) were added to a final concentration of 200 µM.
The reactions were initiated by adding 500 µM NADPH. After 30 min at 30°C the reactions were
quenched with 500 µl EtOAc and extracted twice. The combined organic phases were evaporated in
vacuum and prepared for HPLC analyses.
4

Steady state kinetic measurements were performed by using in vitro conversions under the same
conditions as described above except that substrate concentrations were varied from 10 to 200 µM and
that reactions were stopped after 1 to 10 minutes by flash freezing the samples in liquid nitrogen
followed by extraction with EtOAc. Data were fitted to the Michaelis–Menten equation by non-linear
regression using the software Origin 8.6 to obtain Km- and kcat-values. To determine the total turnover
number, reactions were performed with a substrate concentration of 1 mM and carried out for 20 h. All
other parameters were equal to those described above. Mean values and standard deviations were
calculated from three experiments.
Whole cell conversions:
C43(DE3) cells were transformed with two plasmids, one encoding CYP260B1 or its T224A variant
(pET17b) and the other one encoding Adx4-108 and FpR [22]. An overnight culture was prepared in
NBI medium containing 100 µg/ml ampicillin and 50 µg/ml streptomycin. The main culture was
inoculated with the overnight culture (1:100) and grown in M9CA medium also containing ampicillin
(100 µg/ml) and streptomycin (50 µg/ml). 300 ml baffled flasks each filled with 50 ml of the main
culture were used for the experiments. After 3.5 h at 37°C the cultures were induced with 1 mM
isopropyl β-D-1thiogalactopyranoside and 0.5 mM 5-aminolevulinic acid and further grown for 20.5 h
at 28°C followed by substrate addition (from 100 mM stock) to a final concentration of 1.5 mM DOC
unless stated otherwise. Reactions were performed for 24 h and 0.5 ml samples were taken and
extracted twice with ethyl acetate. The organic layers were pooled, evaporated to dryness and prepared
for HPLC analyses.
Likewise, the conversion in TB medium was performed, except that the induction was done after 2.5 h
and that proteins were expressed for 24 h before adding the substrate.
HPLC analysis:
HPLC analyses were performed on a system consisting of a PU-2080 HPLC pump, an AS-2059-SF
autosampler, and a MD-2010 multi wavelength detector (Jasco, Gross-Umstadt, Germany). A
Nucleodur 100–5 C18 column (125 x 4 mm, Macherey–Nagel, Düren, Germany) was used at 40°C.
10% (v/v) acetonitrile in water (A) and pure acetonitrile (B) were used as mobile phases. A gradient
from 10% to 60% B over 25 min followed by 3 min of 60% B at a flow rate of 1 ml/min was used for
separating the analytes. The relative amounts of the products were estimated using the relative peak
area of the product compared to the combined peak areas of substrate and products at 240 nm.
Purification of the products:
For product elucidation, the reactions were performed with wild type CYP260B1 in 4x 2 l flasks each
filled with 250 ml M9CA medium for 6 h with a substrate concentration of 0.5 mM. All other
conditions were the same as described above. The extract of this whole cell conversion was dissolved
5

in ethyl acetate and loaded on a silica column. An ethyl acetate – hexane mixture of 8:2 was used as
mobile phase for product separation. Fractions were collected and monitored with p-anisaldehyde
staining. Product containing fractions were pooled, evaporated to dryness and prepared for NMR
analysis.
NMR analysis:
NMR spectra were recorded on a Bruker (Rheinstetten, Germany) DRX 500 NMR spectrometer. A
1
1
combination of H, 13C, H,1H-COSY, HSQC, HMBC and NOESY was used to elucidate the
1
13
structures. All chemical shifts are relative to CHCl3 (δ=7.24 for H NMR) or CDCl3 (δ=77.00 for C
NMR) using the standard δ notion in parts per million (ppm).
NMR data of 21-OH pregna-1,4-diene-3,20-dione (1):
1H NMR (500 MHz, CDCl3): 7.01 (d, 1H, J=10.1 Hz, H1), 6.21 (dd, 1H, J= 1.9, 10.1 Hz, H2), 6.04 (s,
1H, H4), 4.15 (dd, 2H, J= 4.6, 5.9 Hz, H21), 3.23 (t, 1H, J= 4.5 Hz, OH), 2.49-2.40 (m, 2H, H17,
H6a), 2.37-2.31 (m, 1H, H6b), 2.23-2.15 (m, 1H, H11a), 1.97-1.90 (m, 2H, H7a, H12a), 1.80-1.67 (m,
4H, H11b, H15a, H16), 1.67-1.56 (m, 1H, H8), 1.38-1.29 (m, 2H, H15b, H12b), 1.20 (s, 3H, H19),
1.16-1.08 (m, 2H, H14), 1.07-1.00 (m, 2H, H11b), 0.69 (s, 3H, H18); 13C NMR (125 MHz, CDCl3):
209.97 (C20), 186.22 (C3), 168.59 (C5), 155.43 (C1), 127.61 (C2) 123.97 (C4), 69.32 (C21), 58.87
(C17), 55.60 (C14), 52.05 (C9), 44.76 (C13), 43.37 (C10), 38.17 (C12), 35.42 (C8), 33.43 (C7), 32.66
(C6), 24.60 (C15), 22.86 (C16), 22.64 (C11), 18.64 (C19), 13.51 (C18).
NMR data of 9α-OH 11-deoxycorticosterone (3):
1H NMR (500 MHz, CDCl3): 5.86 (d, 1H, J= 1.9 Hz, H4), 4.17 (ddd, 2H, J= 4.9, 7.8, 19.7 Hz, H21),
3.24 (t, 1H, J= 4.8 Hz, OH), 2.51 (t, 1H, J= 9.3 Hz, H17), 2.45-2.41 (m, 1H, H6a), 2.41-2.37 (m, 3H,
H1a H2a, H2b), 2.31-2.16 (m, 2H, H6b, H16a), 1.92 (ddd, 1H, J= 4.1, 12.1 (x2), H8), 1.72-1.75 (m,
1H, H16b), 1.75-1.66 (m, 5H, H1b, H11a, H12a, H14, H15a), 1.66-1.54 (m, 3H, H7a, H11b, H12b),
1.43 (ddd, 1H, J= 4.9, 12.6 (x2), H7b), 1.33-1.29 (m, 1H, H15b), 1.30 (s, 3H, H19), 0.67 (s, 3H, H18);
13C NMR (125 MHz, CDCl3): 210.06 (C20), 198.83 (C3), 167.76 (C5), 127.23 (C4), 76.27 (C9), 69.37
(C21), 58.72 (C17), 49.46 (C14), 44.43 (C10, C13), 37.33 (C8), 33.97 (C2), 33.84 (C12), 31.59 (C6),
28.45 (C1), 26.67 (C11), 25.32 (C7), 24.26 (C15), 22.90 (C16), 19.85 (C19), 12.59 (C18).
Results:
Elucidation of the products:
Previous studies by our group showed that wild type CYP260B1 is capable of an efficient conversion
of diverse steroidal compounds, in which DOC was converted with a high yield [15]. However, the
6

main product was not elucidated due to the low selectivity. As shown in Figure 1A, the main product
(3) and a major side product (1) were formed in higher amounts compared to the other side products.
Both products were purified and analyzed by NMR spectroscopy. The main product (3) showed a
tertiary hydroxyl group and was identified as 9α-OH DOC. Its retention time and absorption maximum
matches those of the authentic standard. The major side product showed an additional double bond
between C1 and C2 and was therefore identified as 21-OH pregna-1,4-diene-3,20-dione. The minor
side products were formed in insufficient amounts for structure elucidation via NMR spectroscopy.
However, two of these products were compared with authentic standards and identified as 1α-OH
DOC and 11α-OH DOC.
DOC conversion by the T224A mutant:
As previous docking studies suggested that threonine 224 forms a hydrogen bond with the keto group
of the A ring of DOC [15], we tested the T224A mutant lacking this interaction towards its selectivity
for DOC conversion. This variant showed an increased selectivity of 75% compared to the wild type
showing a selectivity of 60% concerning the main product. The formation of the major side product
was considerably decreased by this mutant (see Figure 1). To compare the wild type with its variant in
more detail, kinetic and binding studies with DOC as substrate were performed. The results are
summarized in Table 1 and showed that the wild type has a kcat value of 23.5 min-1 compared with 12.3
min-1 for the mutant. Binding of DOC to the mutant was also weaker with a Kd of about 210 nM
compared with 21 nM for the wild type. Moreover, the ability of the mutant to further convert the
main product was investigated. The mutant T224A showed a two times lower activity than the wild
type (see Supplemental Figure 2).
Investigation of the side product formation:
The formation of the major side product (1) was further investigated by in vitro conversions of 1α-OH
DOC (2) by both enzymes (see Figure 2). A previous study demonstrated that CYP260A1 from the
same strain was capable of forming the same side product out of 1α-OH DOC [18]. However, neither
the wild type nor the T224A variant were able to form 21-OH pregna-1,4-diene-3,20-dione out of 1α-
OH DOC (see Supplemental Figure 4). The reaction seems, therefore, to be catalyzed by a
dehydrogenation reaction of the P450 itself as described for several other P450s [23].
Screening of wild type CYP260B1 and its T224A mutant towards a steroid library:
To check whether the selectivity of other steroidal compounds is also influenced by the T224A
variant, a library comprising the following steroids was screened with both enzymes: androstenedione,
aldosterone, corticosterone, cortisol, 11-deoxycorticosterone, 11-deoxycortisol, estradiol, estrone, 17α-
OH progesterone, progesterone and testosterone. All tested steroids were converted by both enzymes,
in which only progesterone and DOC were converted with an increased selectivity by the T224A
7

variant. The largest effect was achieved for the conversion of DOC with a selectivity of 75%, whereas
the conversion of progesterone was shifted towards another main product with a selectivity of about
35%.
Optimization approaches for the whole-cell conversion:
The T224A mutant was chosen for the whole-cell experiments due to the higher selectivity for the 9α-
OH DOC production and less activity to further convert this desired product. All whole-cell
conversions were performed with growing cells in medium. TB medium and the minimal medium
M9CA were investigated to test their effects on the bioconversion of DOC by T224A mutant. As
shown in Figure 3A, the conversion in M9CA was significantly higher than that in TB medium.
Therefore, all following experiments were performed in M9CA medium. In addition, the substrate
concentration was varied from 0.5 to 2.0 mM showing that 1.5 mM was consumed to about 70% after
24h, whereas 0.5 mM, 1.0 mM and 2.0 mM were consumed completely, to ~90% or to ~50%,
respectively (see Supplemental Figure 3). To avoid a P450 based consecutive reaction of the main
product, a substrate concentration of 1.5 mM was used for the experiments. Solvents such as EtOH
and DMSO to dissolve DOC or 45% HP-β-CD complexing this steroid were used to investigate their
influence on the bioconversion. As shown in Figure 3B, EtOH was the most suitable solvent for the
T224A dependent conversion of DOC. The biotransformation with DMSO as solvent for DOC was
not as efficient as EtOH and HP-β-CD as complexing agent showed the lowest product formation.
Thus, all whole-cell experiments were performed with EtOH as solvent for DOC.
As the substrate availability inside the cells is a common problem for the bioconversion of
hydrophobic substrates [24, 25], the effect of diverse cell permeabilizing agents was tested. EDTA (20
mM), polymyxin B (30 µg/ml) and toluene (1%) were used for that purpose. However, none of these
additives was able to increase the 9α-OH DOC production. Utilizing EDTA halved the productivity
and the addition of polymyxin B and toluene even dropped the productivity to less than 10% compared
with the control not containing any cell permeabilizing additive (see Figure 3C). Thus, the substrate
availability inside the cells might not be limiting the bioconversion of DOC.
Time dependent bioconversion of DOC by T224A mutant:
A time dependent conversion of DOC was performed to check if the productivity is decreasing after a
certain time. The reaction was performed for 48 h with a substrate concentration of 1.5 mM which was
almost completely consumed after that time. The reaction process precedes a hyperbolic curve as
shown in Figure 4. In the first 24 h the product formation showed no considerably decrease. Then the
reaction was slowed down as only lower amounts of substrate were available. These results
8

demonstrate the efficiency of this enzyme towards the production of 9α-OH DOC. A product yield of
about 0.25 g/l/d was obtained after 24h.
Discussion:
Steroids are an important group of pharmaceuticals primarily used in birth control, hormone-
replacement therapy, inflammatory conditions, and cancer treatment [26]. Their functionalization is of
high interest, whereby hydroxylation at positions C11 and C16 is so far mainly produced by microbial
transformations in steroid industry revealing the lack of efficient chemical techniques for the selective
introduction of oxygen into steroids. During the past years, however, a novel and challenging
procedure using a recombinant yeast strain has been developed for the production of hydrocortisone
[27]. As other positions like C9 and C14 are also of industrial interest, the discovery of alternative and
more effective methods is an important issue [28]. For that reason, bacterial P450s might be used to
perform such reactions due to their high product spectrum, which can be further improved towards
activity and selectivity with molecular biological methods as shown for P450BM3(F87A). The
directed evolution of this enzyme led to either 2β-OH testosterone, 15β-OH testosterone, 2β-OH
progesterone or 16β-OH progesterone with almost complete regio- and stereoselectivity [29].
Here we elucidated the main product of DOC conversion by CYP260B1 from So ce56 as 9α-OH DOC
by NMR spectroscopy. To best of our knowledge, this product was so far only achieved by
biotransformation with whole organisms and not by a specified P450. Fungi like Circinella strains or
Ascochyta linicola and bacteria such as Nocardia strains were previously reported to metabolize DOC
to its 9α-hydroxy derivative [30-32]. Another approach to obtain 9α-OH DOC was performed by
hydroxylating androstenedione at position C9 by Mycobacterium species followed by synthetic routes
of C17 side chain modification [33]. However, the utilization of a whole-cell organism lacks the
information on which specific enzyme is capable of performing the desired reaction. As a result,
strategies for enzyme engineering via molecular biological methods are not feasible hampering the
optimization of such a biotechnological process. In contrast, we identified here a specific P450, whose
crystal structure has been recently solved. Previous docking studies of steroidal compounds into the
substrate-free crystal suggested that threonine 224 forms a hydrogen bond with the C3 keto-group of
C21-steroids [15]. For that reason, we tested the previously created T224A mutant of CYP260B1 that
already showed an altered selectivity for progesterone conversion [17] towards the selectivity of DOC
conversion. This variant showed an increased selectivity of 75% compared with the 60 % of the wild
type concerning the 9α-hydroxylation of DOC. Albeit the activity and binding strength of the mutant
were lower than those of the wild type, we have chosen the T224A mutant for whole-cell conversions
due to its higher selectivity and less activity to further convert the desired 9α-OH DOC. The increased
selectivity for 9α-OH DOC was thereby mainly caused by the lower 21-OH pregna-1,4-diene-3,20-
9

dione formation. This side product formation was further investigated as the family member
CYP260A1 from the same strain was able to form this product out of 1α-OH DOC [18]. However,
neither the wild type nor the T224A variant catalyzed this reaction leading to the assumption that the
side product was formed via dehydrogenation catalyzed by the P450 itself.
With the biocatalytic system harboring the T224A mutant and employing growing cells in M9CA
medium, a production of about 0.25 g/l/d was achieved. Cell permeabilizing agents or expensive
substrate complexing compounds like 2-HP-β-CD were not necessary to reach a high productivity.
The time dependent biotransformation showed a continuous conversion for 48h until nearly all
substrate was consumed demonstrating the durability of the system. Nevertheless, this process might
be further improved by site directed mutagenesis of additional amino acids that are important for the
binding and orientation of DOC to reach a higher selectivity and activity. Such an approach has been
successfully used to change the regioselectivity of CYP106A2 and improve the obtained mutants
concerning their activities [13]. Moreover, performing the reactions in a bioreactor instead of shaking
flasks could increase the space-time yield of this bioconversion [25]. To improve the selectivity of this
enzyme towards the other screened steroidal compounds, saturation mutagenesis [34] of supposed
amino acids binding and interacting with functional groups of the C- or D-ring such as tyrosine 63,
serine 274 or alanine 275 might be performed [15].
The resulting 9α-OH DOC can be used as intermediate in the synthesis of 9α-halo-11β-
hydroxysteroids that show a high physiological activity [35]. In the 1950s, Fried and Sabo discovered
that the anti-inflammatory bioactivity of steroids correlates with the size of the halogen substituent
attached at position C9, in which fluorine showed a stronger effect as chlorine, bromine or iodine. The
bioactivity of the 9α-fluoro steroids was 10 times higher than that of the parent hormones [36, 37].
Since that time, many steroidal drugs were marketed and the most potent corticosteroid hormones are
9α-fluoro derivatives [35, 38]. These 9α-halo-11β-hydroxysteroids diminish the oxidation of the 11-
hydroxyl group providing an increased duration of activity on the receptor which leads to their high
anti-inflammatory effect [39]. Synthetic routes for the production of 9α-fluoro derivatives often begin
with 9(11)-dehydro steroids followed by an epoxidation step and subsequent ring opening [36, 37, 40].
Therefore, the 9α-OH DOC product can serve as precursor for the 9(11)-dehydro compound that can
be further derivatized to the corresponding 9α-fluoro-11β-hydroxysteroid [41]. Alternatively, the
hydroxyl group might be replaced by fluorine via nucleophilic fluorination [42].
In conclusion, we identified 9α-OH DOC as major product for DOC conversion by CYP260B1. The
previously created T224A variant showed an increased selectivity of 75% compared to 60% of the
wild type. With our whole-cell system harboring this mutant, a productivity of about 0.25 g/l/d was
achieved at a 50 ml scale demonstrating the biotechnological potential of this system.
10

Declaration of interest:
Conflict of interest: none
Acknowledgements:
The authors thank Birgit Heider-Lips for the purification of AdR and Adx and Dr. Josef Zapp for
measuring the NMR samples. This work was supported by a grant of the Deutsche
Forschungsgemeinschaft (Be1343/23-2) to R.B.
11

References:
[1] M.V. Donova, O.V. Egorova, Microbial steroid transformations: current state and prospects, Appl
Microbiol Biotechnol 94(6) (2012) 1423-47.
[2] R. Bernhardt, M.R. Waterman, Cytochrome P450 and Steroid Hormone Biosynthesis, The
Ubiquitous Roles of Cytochrome P450 Proteins, John Wiley & Sons, Ltd2007, pp. 361-396.
[3] W.Y. Tong, X. Dong, Microbial biotransformation: recent developments on steroid drugs, Recent
Pat Biotechnol 3(2) (2009) 141-53.
[4] R. Bernhardt, V.B. Urlacher, Cytochromes P450 as promising catalysts for biotechnological
application: chances and limitations, Appl Microbiol Biotechnol 98(14) (2014) 6185-203.
[5] R. Bernhardt, Cytochromes P450 as versatile biocatalysts, J Biotechnol 124(1) (2006) 128-45.
[6] V.B. Urlacher, M. Girhard, Cytochrome P450 monooxygenases: an update on perspectives for
synthetic application, Trends Biotechnol 30(1) (2012) 26-36.
[7] H. Agematu, N. Matsumoto, Y. Fujii, H. Kabumoto, S. Doi, K. Machida, J. Ishikawa, A. Arisawa,
Hydroxylation of testosterone by bacterial cytochromes P450 using the Escherichia coli expression
system, Biosci Biotechnol Biochem 70(1) (2006) 307-11.
[8] A. Berg, J.-A. Gustafsson, M. Ingelman-Sundberg, Characterization of a cytochrome P-450-
dependent steroid hydroxylase system present in Bacillus megaterium, J Biol Chem 251(9) (1976)
2831-2838.
[9] F.M. Kiss, Y. Khatri, J. Zapp, R. Bernhardt, Identification of new substrates for the CYP106A1-
mediated 11-oxidation and investigation of the reaction mechanism, FEBS lett 589(18) (2015) 2320-6.
[10] F.M. Kiss, D. Schmitz, J. Zapp, T.K. Dier, D.A. Volmer, R. Bernhardt, Comparison of
CYP106A1 and CYP106A2 from Bacillus megaterium - identification of a novel 11-oxidase activity,
Appl Microbiol Biotechnol 99(20) (2015) 8495-514.
[11] M. Lisurek, M.J. Kang, R.W. Hartmann, R. Bernhardt, Identification of monohydroxy
progesterones produced by CYP106A2 using comparative HPLC and electrospray ionisation collision-
induced dissociation mass spectrometry, Biochem Biophys Res Commun 319(2) (2004) 677-82.
[12] C. Virus, M. Lisurek, B. Simgen, F. Hannemann, R. Bernhardt, Function and engineering of the
15beta-hydroxylase CYP106A2, Biochem Soc Trans 34(6) (2006) 1215-8.
[13] K.T. Nguyen, C. Virus, N. Gunnewich, F. Hannemann, R. Bernhardt, Changing the
regioselectivity of a P450 from C15 to C11 hydroxylation of progesterone, Chembiochem 13(8)
(2012) 1161-6.
[14] Y. Khatri, M. Ringle, M. Lisurek, J.P. von Kries, J. Zapp, R. Bernhardt, Substrate Hunting for the
Myxobacterial CYP260A1 Revealed New 1alpha-Hydroxylated Products from C-19 Steroids,
Chembiochem 17(1) (2016) 90-101.
[15] S.G. Salamanca-Pinzon, Y. Khatri, Y. Carius, L. Keller, R. Muller, C.R. Lancaster, R. Bernhardt,
Structure-function analysis for the hydroxylation of Delta4 C21-steroids by the myxobacterial
CYP260B1, FEBS lett 590(12) (2016) 1838-51.
12

[16] Y. Khatri, Y. Carius, M. Ringle, C.R. Lancaster, R. Bernhardt, Structural characterization of
CYP260A1 from Sorangium cellulosum to investigate the 1alpha-hydroxylation of
a
mineralocorticoid, FEBS lett 590(24) (2017) 4638-4648.
[17] M. Litzenburger, R. Lo Izzo, R. Bernhardt, Y. Khatri, Investigating the roles of T224 and T232 in
the oxidation of cinnamaldehyde catalyzed by myxobacterial CYP260B1, FEBS lett 591(1) (2017) 39-
46.
[18] Y. Khatri, A. Schifrin, R. Bernhardt, Investigating the effect of available redox protein ratios for
the conversion of a steroid by a myxobacterial CYP260A1, FEBS Lett. 2017 Apr;591(8):1126-1140.
doi: 10.1002/1873-3468.12619. Epub 2017 Apr 6.2017.
[19] H. Uhlmann, R. Kraft, R. Bernhardt, C-terminal region of adrenodoxin affects its structural
integrity and determines differences in its electron transfer function to cytochrome P-450, J Biol Chem
269(36) (1994) 22557-64.
[20] Y. Sagara, A. Wada, Y. Takata, M.R. Waterman, K. Sekimizu, T. Horiuchi, Direct expression of
adrenodoxin reductase in Escherichia coli and the functional characterization, Biol Pharm Bull 16(7)
(1993) 627-30.
[21] J.B. Schenkman, Studies on the nature of the type I and type II spectral changes in liver
microsomes, Biochem 9(10) (1970) 2081-91.
[22] M. Ringle, Y. Khatri, J. Zapp, F. Hannemann, R. Bernhardt, Application of a new versatile
electron transfer system for cytochrome P450-based Escherichia coli whole-cell bioconversions, Appl
Microbiol Biotechnol 97(17) (2013) 7741-54.
[23] F.P. Guengerich, Common and uncommon cytochrome P450 reactions related to metabolism and
chemical toxicity, Chem Res Toxicol 14(6) (2001) 611-50.
[24] S. Janocha, R. Bernhardt, Design and characterization of an efficient CYP105A1-based whole-
cell biocatalyst for the conversion of resin acid diterpenoids in permeabilized Escherichia coli, Appl
Microbiol Biotechnol 97(17) (2013) 7639-49.
[25] M.T. Lundemo, J.M. Woodley, Guidelines for development and implementation of biocatalytic
P450 processes, Appl Microbiol Biotechnol 99(6) (2015) 2465-83.
[26] J.M. Beale, J. Block, R. Hill, Organic medicinal and pharmaceutical chemistry, Lippincott
Williams & Wilkins, Philadelphia, 2010.
[27] F.M. Szczebara, C. Chandelier, C. Villeret, A. Masurel, S. Bourot, C. Duport, S. Blanchard, A.
Groisillier, E. Testet, P. Costaglioli, G. Cauet, E. Degryse, D. Balbuena, J. Winter, T. Achstetter, R.
Spagnoli, D. Pompon, B. Dumas, Total biosynthesis of hydrocortisone from a simple carbon source in
yeast, Nat Biotechnol 21(2) (2003) 143-9.
[28] S.B. Mahato, S. Garai, Advances in microbial steroid biotransformation, Steroids 62(4) (1997)
332-45.
[29] S. Kille, F.E. Zilly, J.P. Acevedo, M.T. Reetz, Regio- and stereoselectivity of P450-catalysed
hydroxylation of steroids controlled by laboratory evolution, Nat Chem 3(9) (2011) 738-43.
13

[30] A. Schubert, R. Siebert, Microbiological 9-hydroxylation of steroids, DD 20643, 1961.
[31] G.D. Searle, Company, 9α-Hydroxy steroids by microbiological oxidation, GB 862701, 1961.
[32] O.K. Sebek, 9α-Hydroxylation of steroids with Ascochyta, US 3116220, 1963.
[33] J.N.M.M. Batist, Arthur Friedrich Preparation of 9α-hydroxy-17-methylene steroids as
intermediates for corticosteroids, EP 336521, 1989.
[34] M.T. Reetz, Laboratory Evolution of Stereoselective Enzymes: A Prolific Source of Catalysts for
Asymmetric Reactions, Angew Chem Int Ed Engl 50(1) (2011) 138-174.
[35] J. Del Rosso, S.F. Friedlander, Corticosteroids: options in the era of steroid-sparing therapy, J Am
Acad Dermatol 53(1) (2005) 50-58.
[36] J. Fried, E.F. Sabo, Synthesis of 17α-hydroxycorticosterone and its 9α-halo derivatives from 11-
epi-17α-hydroxycorticosterone, J Am Chem Soc 75(9) (1953) 2273-2274.
[37] J. Fried, E.F. Sabo, 9α-Fluoro derivatives of cortisone and hydrocortisone, J Am Chem Soc 76(5)
(1954) 1455-1456.
[38] J.A. Hogg, Steroids, the steroid community, and Upjohn in perspective: a profile of innovation,
Steroids 57(12) (1992) 593-616.
[39] S. Diederich, E. Eigendorff, P. Burkhardt, M. Quinkler, C. Bumke-Vogt, M. Rochel, D.
Seidelmann, P. Esperling, W. Oelkers, V. Bahr, 11β-Hydroxysteroid dehydrogenase types 1 and 2: an
important pharmacokinetic determinant for the activity of synthetic mineralo-and glucocorticoids, J
Clin Endocrinol Metab 87(12) (2002) 5695-5701.
[40] C.J. Sih, The Synthesis of 9α-Hydroxy Steroids, J Org Chem 26(11) (1961) 4716-4718.
[41] L. Sedlaczek, L.L. Smith, Biotransformations of steroids, Crit Rev Biotechnol 7(3) (1988) 187-
236.
[42] T. Liang, C.N. Neumann, T. Ritter, Introduction of fluorine and fluorine-containing functional
groups, Angew Chem Int Ed Engl 52(32) (2013) 8214-64.
14

Table 1: Dissociation constants, selectivity and kinetic parameters of DOC oxidation catalyzed by
wild-type CYP260B1 and its T224A variant.
Enzyme
Wild type
T224A
Km [µM]*
7.94 ± 2.99
3.67 ± 2.01
kcat [min-1]*
23.52 ± 0.96
12.33 ± 0.39
TTN*
Kd [nM]
21 ± 10
213 ± 70
Selectivity
960
60%
610
75%
* Values were determined by measuring the total amount of products.
Results are obtained from triplicate experiments, ± standard deviation.
15

Figure captions:
Figure 1: Chromatograms of in vitro reactions of DOC by purified wild type (A) and its T224A variant
(B) after almost complete conversion. Marked peaks 1-3 were elucidated by NMR spectroscopy (1 and
3) or by comparison with an authentic standard of 1α-OH DOC (2).
Figure 2: Reactions catalyzed by CYP260B1 and its T224A variant.
Figure 3: Influences of the media, solvents and cell permeabilizing agents on the whole-cell
conversion of DOC by T224A mutant. The comparison of M9CA and TB medium (A), different
solvents and complexing agents to dissolve DOC (B) and diverse cell permeabilizing agents (C) are
shown. Results are obtained from triplicate experiments, ± standard deviation.
Figure 4: Time dependent production of 9α-OH DOC by the T224A mutant. Results are obtained from
triplicate experiments, ± standard deviation.
16

Figure 1A

Figure 1B

Figure 2

Figure 3A

Figure 3B

Figure 3C

Figure 4

Highlights:



CYP260B1 acts as 9α-hydroxylase for 11-deoxycorticosterone
The T224A mutant of CYP260B1 increases the selectivity
A productivity of about 0.25 g/l/d can be achieved
17