Side chain removal from corticosteroids by unspecific peroxygenase

Article abstract of DOI:10.1016/j.jinorgbio.2018.03.011

Two unspecific peroxygenases (UPO, EC 1.11.2.1) from the basidiomycetous fungi Marasmius rotula and Marasmius wettsteinii oxidized steroids with hydroxyacetyl and hydroxyl functionalities at C17 - such as cortisone, Reichstein's substance S and prednisone

Full text of DOI:10.1016/j.jinorgbio.2018.03.011

Accepted Manuscript
Side chain removal from corticosteroids by unspecific
peroxygenase
René Ullrich, Marzena Poraj-Kobielska, Steffi Scholze, Claire
Halbout, Martin Sandvoss, Marek J. Pecyna, Katrin Scheibner,
Martin Hofrichter
PII:
S0162-0134(18)30159-4
doi:10.1016/j.jinorgbio.2018.03.011
JIB 10462
DOI:
Reference:
To appear in:
Journal of Inorganic Biochemistry
Received date:
Accepted date:
13 March 2018
22 March 2018
Please cite this article as: René Ullrich, Marzena Poraj-Kobielska, Steffi Scholze, Claire
Halbout, Martin Sandvoss, Marek J. Pecyna, Katrin Scheibner, Martin Hofrichter , Side
chain removal from corticosteroids by unspecific peroxygenase. The address for the
corresponding author was captured as affiliation for all authors. Please check if
appropriate. Jib(2017), doi:10.1016/j.jinorgbio.2018.03.011
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.

ACCEPTED MANUSCRIPT
Side chain removal from corticosteroids by unspecific peroxygenase
1
#
1#
1
2
2
1,3
René Ullrich , Marzena Poraj-Kobielska , Steffi Scholze , Claire Halbout , Martin Sandvoss , Marek J. Pecyna ,
4
1
Katrin Scheibner and Martin Hofrichter *
1
TU Dresden, International Institute Zittau, Department of Bio- and Environmental Sciences, Markt 23, 02763
Zittau, Germany
2
Sanofi-Aventis Deutschland GmbH, R&D, Integrated Drug Discovery IDD, Isotope Chemistry & Metabolite
Synthesis ICMS, G876, 65926 Frankfurt/Hoechst, Germany
3
University of Applied Sciences Zittau-Görlitz, Faculty of Natural and Environmental Sciences, Theodor-Körner-
Allee 16, 02763 Zittau, Germany
4
BTU Cottbus-Senftenberg, Faculty of Environment and Natural Sciences, Universitätsplatz 1, 01968,
Senftenberg, Germany
*
3
Corresponding author (e-mail address: martin.hofrichter@tu-dresden.de, phone: +49 3583 612 4121, fax: +49
583 612 4100)
#
These authors contributed equally to this work
1

ACCEPTED MANUSCRIPT
Abstract
Two unspecific peroxygenases (UPO, EC 1.11.2.1) from the basidiomycetous fungi Marasmius rotula
and Marasmius wettsteinii oxidized steroids with hydroxyacetyl and hydroxyl functionalities at C17 -
such as cortisone, Reichstein’s substance S and prednisone - via stepwise oxygenation and final fission
of the side chain. The sequential oxidation started with the hydroxylation of the terminal carbon (C21)
leading to a stable geminal alcohol (e.g. cortisone 21-gem-diol) and proceeded via a second
oxygenation resulting in the corresponding -ketocarboxylic acid (e.g. cortisone 21-oic acid). The latter
decomposed under formation of adrenosterone (4-androstene-3,11,17-trione) as well as formic acid
and carbonic acid (that is in equilibrium with carbon dioxide); fission products comprising two carbon
atoms such as glycolic acid or glyoxylic acid were not detected. Protein models based on the crystal
structure data of MroUPO (Marasmius rotula unspecific peroxygenase) revealed that the bulky
cortisone molecule suitably fits into the enzyme’s access channel, which enables the heme iron to
come in close contact to the carbons (C21, C20) of the steroidal side chain. ICP-MS analysis of purified
MroUPO confirmed the presence of magnesium supposedly stabilizing the porphyrin ring system.
Keywords
Peroxygenation; peroxide shunt; P450; heme-thiolate; deacylation; geminal alcohol
1
. Introduction
Steroidal compounds influence animals and humans in various ways [1, 2]. They comprise several
biologically active classes of substances such as bile acids, cardio-active drugs (cardiotonics), steroid
sapogenins, steroid alkaloids as well as sexual and corticosteroidal hormones [3]. As hormones, they
regulate metabolic and ontogenetic processes in humans from fetal life to adulthood. Because of their
biological effects, steroids are not only of general scientific interest but also of medical relevance, not
least with respect to pharmaceutical applications. In animals including humans, steroids are made from
lanosterol that derives from cyclization of the triterpene squalene, which for its part is a product of the
mevalonate pathway [4]. One group of pharmaceutically relevant steroids are C21 hormones
(pregnane core), which include gluco- and mineralocorticoids produced in the adrenal cortex as well
as their precursor progesterone. The latter can be transformed via hydroxylation to 17-hydroxy
progesterone and subsequent side chain removal or C21 hydroxylation to androstenedione and 11-
deoxycortisol (Reichstein’s substance S), respectively; the former is the direct precursor of the male
sexual hormone testosterone and the latter of the glucocorticoid cortisol. Both Reichstein’s substance
S and androstenedione are important intermediates in industrial steroid syntheses.
Steroids used for therapeutic or other pharmaceutical purposes are largely products of partial
synthesis starting from animal cholesterol or herbal phytosterols or sapogenins. The addition, removal
2

ACCEPTED MANUSCRIPT
and modification of functional groups to, from and at these starting materials is of central importance
when producing the wide variety of steroidal drugs. These modifications are achieved using
conventional organic synthesis and/or microbial whole-cell biotransformations [5, 6]. Cytochrome
P450 enzymes (P450s, CYPs) are the key biocatalysts of steroid biotransformation and catalyze
hydroxylation, epoxidation, aromatization and side chain removal [2]. Such reactions are integral part
of steroid biosynthesis pathways in humans and animals (see above) and some of them can be also
found in microorganisms (fungi, bacteria); typically in form of single transformations with often
unknown function for the microbes (e.g. 11-, 11- and 16-hydroxylation by Bacillus,
Mycobacterium, Curvularia or Cunninghamella) [6, 7]. However, microbial systems do not mimic all
relevant human steroid transformations. Thus side-chain removal (C-C bond cleavage) from
progesterone or pregnenolone is exclusively accomplished by vertebrate CYP17A1 (also known as a
1
7-hydroxylase/17,20-lyase). It proceeds through two intrinsic reactions: in the first one, the enzyme
hydroxylates the 17-position and in the second reaction, it cleaves the C-C bond between C17 and
C20 in a lyase-like reaction resulting in the release of acetate [8]. To our best knowledge, however,
steroids hydroxylated in position C21 such as Reichstein’s substance S, cortisol or cortisone are not
subject of C-C cleavage by CYP17A1 or any microbial P450. Since, on the other hand, Reichstein’s
substance S is a key precursor in industrial steroid syntheses [9], it would be worthwhile to have a
catalyst that selectively removes the respective hydroxyacetyl moiety to form the corresponding keto-
derivatives (17-ones).
Unspecific peroxygenase (UPO, EC 1.11.2.1) is a secreted fungal enzyme type that efficiently transfers
oxygen from peroxides (R-OOH) to diverse substrates including unactivated hydrocarbons [10].
Literally, such enzymes strictly follow the peroxide ‘shunt’ pathway of P450s and in this context, they
4
6
can reach total turnovers numbers of 10 to almost 10 for oxygen transfer reactions [11-13]. As P450s,
UPO enzymes are heme-thiolate proteins but they do not share any sequence homology with them.
Highest sequence homology has been ascertained to fungal chloroperoxidase (CPO, EC 1.11.1.6) that
is actually a UPO with strong halide oxidation activity [14]. Among the UPO-catalyzed reactions are
hydroxylations, epoxidations, O- and N-dealkylations, oxidations of organic hetero atoms and bromide
as well as one-electron oxidations. Their catalytic cycle involves formation of a highly reactive oxoferryl
intermediate (compound I) and a protonated basic compound II (oxoferryl hydroxide) that are the
active oxygen species in hydroxylation reactions [15, 16]. Crystal structure of the UPO from Agrocybe
aegerita (AaeUPO) was solved and has given first insights into the molecular organization of this
exceptional peroxidase type [17, 18].
Here we report two UPOs from the mushrooms Marasmius rotula (MroUPO) [19] and Marasmius
wettsteinii (MweUPO, novel wild-type UPO first described herein), which catalyze the selective
3

ACCEPTED MANUSCRIPT
deacylation of 17,21-dihydroxycorticosteroids. To our best knowledge, this reaction has not been
reported so far for any microbial P450 and represents a new reaction type for UPOs.
2
. Materials and methods
2
.1. Chemicals
Formic acid (pro analysis) and acetonitrile (ACN, gradient grade for liquid chromatography) were
purchased from VWR (Darmstadt, Germany). All other chemicals including cortisone (S I), prednisone
(
S II), Reichstein’s substance S (S III), and adrenosterone (P III) were obtained from Sigma-Aldrich
GmbH (Schnelldorf, Germany) in the highest available purity.
2
.2. Enzyme preparation and characterization
The extracellular peroxygenases from the basidiomycetous fungi Agrocybe aegerita (AaeUPO; main
isoform, 46 kDa) and Marasmius rotula (MroUPO; main isoform, 32 kDa denaturated and 58 kDa
native) were produced and purified as described previously [19, 20]. The enzyme preparations were
homogeneous according to sodium dodecyl polyacrylamide gel electrophoresis (SDS-PAGE) and
analytical high performance liquid chromatography-size exclusion chromatography (HPLC-SEC). The
-1
-1
specific activities of the AaeUPO and MroUPO preparations were 99.6 U mg and 26.0 U mg ,
respectively. One unit (1 U) corresponds to the oxidation of 1 µmol of veratryl alcohol (3,4-
dimethoxybenzyl alcohol) to veratraldehyde (3,4-dimethoxybenzaldehyde) in 1 min at 23°C [20];
activity measurements were performed in 50 mM sodium phosphate buffer at pH 7.0 (AaeUPO) or 5.5
(
MroUPO and MweUPO).
A novel wild-type peroxygenase was produced with Marasmius wettsteinii (MweUPO) strain 582
deposited at the culture collection of the International Institute Zittau (Technical University of
Dresden, Germany) and at DSMZ Braunschweig (Deutsche Stammsammlung für Mikroorganismen und
Zellkulturen, Germany) under DSM 106021. The fungus was isolated from fruiting bodies growing on
spruce needle litter (Picea abies) near the Kottmar mountain in East Saxony (Germany) in September
2
015 (more information on the fungal species is given in the Supplemental Information section). Stock
cultures were maintained in culture slants on 2% malt extract agar at 4°C in the dark. For enzyme
production, the content of a completely overgrown agar-plate was homogenized in 80 mL sterile tap
water and used for the inoculation of liquid cultures (5% vol/vol of the suspension). Submerged
cultivation was performed in 500-mL Erlenmeyer flasks containing 200 mL of a complex liquid medium
according to Gröbe et al. [19] on a rotary shaker (120 rpm, at 23°C). The time course of MweUPO
production is shown in Fig. S1. Fungal cultures were harvested after three weeks when the UPO activity
-
1
had reached approximately 2,500 U L . Crude extracts of concentrated culture liquid contained at least
4

ACCEPTED MANUSCRIPT
five isoforms of MweUPO. The major isoform was isolated via different steps of fast protein liquid
-
1
chromatography (FPLC) and had a final specific activity of 37.5 U mg ; the corresponding purification
table is given in Table S1. Homogeneity of the final MweUPO preparation was proven by SDS-PAGE and
HPLC-SEC and compared to MroUPO and AaeUPO (Fig. S2 and S3). Molecular mass of the native
enzyme was 62 kDa (determined by HPLC-SEC; Fig. S2) and 32.5 kDa according to denaturing SDS-PAGE
(Fig. S3) indicating, as in the case of MroUPO, a physiological dimer [21]. HPLC-SEC of purified MweUPO
was carried out using an Agilent Series 1200 instrument equipped with a diode array detector (Agilent
Technologies Deutschland GmbH, Waldbronn, Germany) and a Yarra 3µ SEC-2000 column (300 x 7.8
mm, Phenomenex, Aschaffenburg, Germany). Separation of UPO proteins was achieved at 25°C under
-
1
isocratic conditions (flow rate 1 mL min ) with sodium acetate/sodium chloride (50 mM/100 mM, pH
.9) as eluent. Calibration occurred with a gel filtration low-molecular weight calibration kit from GE
Healthcare (Munich, Germany). SDS-PAGE was performed as described previously [22].
6
2
.3. Enzymatic reaction setup
-
1
Typical reaction set-ups (0.2-1.0 mL) contained UPO protein (1.0-2.0 U mL corresponding to 0.04-0.08
-
1
µM = 0.014-0.028 mg mL ), 0.5-2.0 mM steroidal substrate (stock solution in 50% ACN resulting in a
final ACN concentration of 5 vol%) and 50 mM potassium phosphate buffer (pH 7.0 or 5.5). The
reactions stirred at room temperature were started with H
2
O by either adding it directly (0.1 to 2 mM
2
-
1
2
H O
2
) or supplying it continuously with a syringe pump (0.5 mM h for 4 hours). For kinetic studies,
aliquots were taken from the reaction mixtures every 20 min and analyzed by high performance liquid
chromatography-mass spectrometry (HPLC-MS).
The reaction setup was scaled up to 50 mL to separate and purify the products of enzymatic steroid
-
1
conversion for subsequent NMR analysis. The reaction solution contained 20 U (5.6 µg mL ) of purified
MroUPO, 80 mg of the substrate (S I cortisone, S II prednisone or S III Reichstein’s substance S;
chemical structures are shown in Fig. 1), 2.5 mL of ACN (final concentration of 5% vo/vol) and
potassium phosphate buffer (50 mM, pH 5.5). The reaction was started by the addition of hydrogen
peroxide (H
2
O , 1 mM) and repeated every 6 min over 1 hour (final concentration 10 mM).
2
2
.4. Analytical methods
HPLC-MS method I (for standard reactions). Products of enzymatic syntheses were analyzed by high
performance liquid chromatography (HPLC) using an Agilent Series 1200 instrument equipped with a
diode array detector and an electrospray ionization mass spectrometer (ESI-MS) (Agilent Technologies
Deutschland GmbH, Waldbronn, Germany). Unless otherwise stated, reverse phased chromatography
on a Luna C18 column (2 mm diameter by 250 mm length, 3 µm particle size, Phenomenex,
-
1
Aschaffenburg, Germany) was used. Column was eluted at a flow rate of 0.2 mL min and 40°C with
5

ACCEPTED MANUSCRIPT
aqueous 0.01% vol/vol ammonium formate (pH 3.5)/ACN, 95:5 for 5 min, followed by a 15-min linear
gradient to 100% ACN. Mass spectrometric measurements were made in positive or negative ESI mode
in a mass range from 70 to 900, step size 0.1, drying gas temperature 350°C, capillary voltage 4000 V
for positive mode and 5500 V for negative mode. Reaction products were identified relative to
authentic standards, based on their retention times, UV absorption spectra, and mass spectral data
([M + H] or [M - H] ions).
HPLC-MS method II. Additional analyses were performed using a Kinetex C18 column (2 mm diameter
by 250 mm length, 5 µm particle size, Phenomenex), which was eluted with a flow rate of 1 mL min-1
at 40°C with aqueous 0.01% vol/vol ammonium formate (pH 3.5)/ACN, 95:5 for 2 min, followed by a
1
2-min linear gradient to 100% ACN. MS parameters were the same as described under method I.
HPLC-MS method III. Kinetic studies on the conversion of cortisone by MroUPO and MweUPO were
carried out by applying a more rapid HPLC method. Respective samples were analyzed using a
Phenomenex Kinetex Phenyl-Hexyl column (2.6 µm particle size, 150 x 2.1 mm). The flow rate was 0.4
-
1
mL min and the temperature set to 40°C. The following gradient of 0.01 vol% ammonium formate
pH 3.2) and ACN was performed: 0-0.2 min 20% ACN followed by a linear gradient to 65% ACN (0.2-
0 min) and to 85% (10-12 min). The MS parameters were the same as in method I.
(
1
Ion chromatography (IC) method for organic acids. Organic acids were analyzed by IC using a Dionex
Series ICS1100 system (Thermo Fisher GmbH, Idstein, Germany) fitted with an IonPac® ICE-AS1 column
(4 mm diameter by 250 mm length, Dionex, Germering, Germany). The column was eluted at 30ºC and
-
1
1
.6 mL min with an aqueous perfluorobutyric acid solution (1 mM) over 20 min. Products were
identified relative to the authentic standards by their retention times.
+
HPLC-UV method for organic acids. Analyses were performed using a Rezex™ ROA-Organic Acid H (8%)
column (7.8 mm diameter by 300 mm length, Phenomenex). The column was eluted at 50°C and 0.5
-
1
mL min with an aqueous sulfuric acid solution (0.005 N) over 25 min. Acids were recorded at 210 nm
and identified relative to authentic standards by their retention times.
2
.5. Product isolation
Products of enzymatic steroid oxidation were separated and purified by preparative HPLC using a 1260
Infinity Agilent system (Agilent Technologies Deutschland GmbH) equipped with a Kinetex C18 column
(Axia packed, 21.2 mm diameter by 150 mm length, 5 µm particle size, Phenomenex), which was eluted
-
1
at 20 mL min and room temperature with water/ACN 70:30 for 1 min under isocratic conditions,
followed by a 5-min linear gradient to 70% ACN and a 1.5-min gradient to 100% CAN, which was hold
for 0.5 min and then set to 30% ACN within 1 min. Eluting peaks (detected at 240 nm) were sampled
with an Infinity fraction collector (fraction collector; preparative scale, G1364B, Agilent Technologies
6

ACCEPTED MANUSCRIPT
Deutschland GmbH). The purity of isolated products was checked by analytical HPLC-MS (method I
above).
2
2
.6. NMR spectroscopy experimental conditions
.65 mg of P I and 0.5 mg of P II were transferred into capillary NMR tubes of 2.5 × 100 mm in 150 µL
dimethylsulfoxide-d6 (99.8%, Deutero, Kastellaun, Germany). Proton, COSY (correlated spectroscopy),
HSQC (heteronuclear single quantum coherence) and HMBC (heteronuclear multiple bond correlation)
spectra were recorded on a Bruker Avance 500 MHz spectrometer equipped with a 2.5 mm selective
inverse SEI probe (Bruker, Rheinstetten, Germany). For technical reasons, some HMBC spectra had to
be recorded on a Bruker Avance III 600 MHz instrument equipped with a 1.7 mm inverse SEI probe
(part of the sample was transferred in to 1.7 × 100 mm microtubes). For chemical shift prediction, the
ACD/C+H NMR predictors, release 2015, pack2, from ACDlabs (Toronto, Ontario, Canada) were used.
2
.7. Precipitation of carbon dioxide with barium hydroxide.
Carbon dioxide (CO ) formation in the course of side chain removal from cortisone was qualitatively
proved in a separate experiment. Two vials filled with 1 mL of saturated barium hydroxide [Ba(OH)
were connected via silicon tubes with two further vials containing the reaction mixture with MroUPO
active sample) or a control without enzyme (Fig. S7). The reaction occurred under helium atmosphere.
Before starting the reaction, the intrinsic CO was stripped by flushing with helium, over 10 min.
Afterwards the reaction was started as described above but under continuous bubbling of helium
through the vials. The stripped gaseous compounds were transferred into the Ba(OH) vials where the
CO was trapped as insoluble barium carbonate (BaCO ; compare Fig. S7).
2
2
]
(
2
2
2
3
2
.8. Determination of the metal content in MroUPO
Purified MroUPO was washed with deionized water resulting in a final protein concentration of 4 µM
determined by a standard Bradford assay. Inductively coupled plasma optical emission spectroscopy
(
ICP-OES) and ICP mass spectrometry (ICP-MS) were performed on an Optima 3000 system and an Elas
DRC-e device (both PerkinElmer Inc., MA, USA), respectively. The OES-system was calibrated with the
ICP multi-element standard solution IV’ (Mg, K, Ca, Mn, Fe, Cu) as well as the ‘phosphorus ICP
‘
-
1
-1
standard’ (both Merck, Darmstadt, Germany) in 1 mg L and 10 mg L concentrations. Deionized water
served as enzyme blank solution. Measurements were performed according to the manufacturers’
-
1
specifications. Different concentrations (2, 15, and 50 μg L ) of the ‘ICP multi-element standard
solution VI’ (V, Mn, Fe, Co, Cu; Merck, Darmstadt, Germany) were used to calibrate the MS unit.
2
.9. Crystal structure and 3D-modelling
7

ACCEPTED MANUSCRIPT
The crystal structure data of MroUPO have been made available recently (protein data bank entries
5
FUJ and 5FUK). Based on them, first structures were published indicating that MroUPO is a dimeric
protein with a relatively wide heme access channel [21]. Cortisone was docked into the MroUPO active
site cleft using the open-source program AutoDock Vina, version 1.1.2 [23]. The MroUPO structure
(PDB 5FUJ) and the cortisone molecule (3D conformation obtained from PubChem [24], CID 222786; 4
rotatable bonds) were prepared for docking in AutoDockTools [25], version 1.5.6. In order to define
the active site of MroUPO, cortisone as substrate was first placed manually at the entrance of the
protein’s binding pocket using PyMOL (The PyMOL Molecular Graphics System, Version 2.0
Schrödinger, LLC). The grid box was centered at x: −10.607, y: -17.81 and z: −28.584 with sizes of 20.25,
1
6.5, and 9.75ꢀÅ, respectively. Exhaustiveness was tested between 8 to 100. No more than twenty
−
1
output poses (energy range 5 kcal mol with a maximum number of 30 for binding modes) were
generated. In separated runs, the best three models occurred repeatedly. The docking model with
−
1
highest affinity (−8.3ꢀkcalꢀmol ) in all runs shows close proximity of cortisone C21 to heme iron (4.4 Å)
and is shown in Figure 6.
3
. Results
3
.1. Enzymatic conversion of cortisone (S I)
+
Both MroUPO and MweUPO were found to be capable of converting cortisone (S I, m/z M+H : 361) by
almost 100%, while AaeUPO did not accept this compound as substrate. For the first 20 min of the
-
1
-1
reaction, we calculated turnover numbers (TON) of 6.4 min and 6.1 min , respectively. These TON
were limited by the limited amount of H gently and continuously supplied via a syringe pump, which
2
O
2
-
1
would theoretically entail a maximum TON of 6.9 min under the conditions applied (co-substrate
limitation to prevent heme damage). The first product that appeared within 20 min had a molecular
+
mass of 376 (P I; m/z M+H : 377) indicating a hydroxylated metabolite (Fig. 2). 20 min later, a second
+
product was detectable with a molecular mass of 374 (P II; m/z M+H : 375) probably representing an
oxidation product of P I. P II showed an unusual peak shape (tailing) in the HPLC elution profile that is
characteristic for keto-enol tautomerism or the presence of an aldehyde hydrate (that is in equilibrium
with the corresponding free aldehyde). After 80 min, the formation of P I had reached its maximum,
while the peak of P II still increased. 160 min after starting the reaction, a third product (P III) emerged
+
with a smaller mass of 300 (m/z M+H : 301); it was identified by means of an authentic standard as
adrenosterone and simultaneously, P II started to decrease (Fig. 2). At the end of the experiment (after
2
20 min), only 20% of P II’s maximum peak area remained, while P III reached a maximum
concentration of 359 and 347 µM in the case of MroUPO and MweUPO, respectively. The respective
8

ACCEPTED MANUSCRIPT
results are summarized in Fig. 2, 3 and S5. Compared to the Marasmius UPOs, AaeUPO merely
converted S I to negligible extent (<1%), which resulted in traces of P II and P III (Fig. 3).
The observed HPLC peak pattern and the almost identical UV spectra of substrate and products, caused
by the unchanged chromophore (double bonds in the steroidal A ring), prompted us to analyze the MS
data more in detail to identify possibly hidden additional products. Indeed, the hint of an additional
mass eluting at the same time as the substrate (S I) was observed. Therefore, the HPLC method was
modified and improved (among others by changing the stationary phase). Then, we analyzed the
extracted ion counts of the respective MS data (given in Fig. S5 for MweUPO) again and found a fourth
product (P IV; most likely 17-hydroxy-3,11-dioxoandrost-4-ene-17-carboxylic acid) with a molecular
+
mass of 346 (m/z M+H : 347). Fig. S5 displays a further metabolite (P V) with a mass of 316 (m/z: 317
+
M+H ), probably a hydroxylation product of the fission product P III. Both P IV and P V were only
detectable in traces and could not be quantified because of lacking standards; thus changes in their
concentration were deduced from the corresponding peak areas (based on UV-absorption).
3
.2. NMR spectroscopic analysis of the products P I and P II
13
1
Particular efforts were made to unambitiously identify the products P I, P II and P III by C/ H-NMR
spectroscopy. The NMR assignment of P I was based on the analysis of the side chain resonances in
conjunction with chemical shift prediction and comparison to cortisone reference spectra (Table S2).
In the side chain, a CH signal at 84.1/5.39 ppm indicating an alcoholic substitution was observed along
with a carbonyl function at 207.0 ppm. Two possible keto-enol tautomeric structures were considered.
The COSY showed that the CH signal at 84.1/5.39 ppm couples to two OH doublets at 6.12 and 6.36
ppm, respectively. Moreover, we observed HMBC correlations between the carbonyl resonance at
2
07.0 ppm and the CH proton at 5.39 ppm and the OH protons at 6.12 and 6.36 ppm. Furthermore, we
observed HMBC couplings between the carbon at 84.05 ppm and two OH (6.12 and 6.36 ppm). This
clearly establishes P I to be structure 1 (cortisone 21-gem-diol) given in Fig. 4, along with an additional
minor species in the same sample (P Ia accounting for approximately 14%). The structure of P Ia being
the hydrate of an aldehyde prompted us to further inspect the heteronuclear spectra: Fig. 4 shows the
aldehyde section of the respective HMBC spectrum. A correlation between a CH at 9.66/195.2 ppm
1
2
(
spread by its J (C,H) coupling of 189 Hz) and a carbonyl C at 200.5 ppm (showing a J (C,H) coupling
of 27 Hz) indicates the presence of the free aldehyde P Ia (cortisone 21-al) in the sample (Fig. 4).
For the second metabolite (Table S3), P II, we suggest the structure of a carboxylic acid (cortisone 21-
oic acid) given in Fig. 5. Inspection of the proton, COSY and HSQC spectra showed that the sterane
scaffold resonances were unchanged, while side chain CH resonances could not be observed. An
additional OH proton resonance appeared at 7.09 ppm (integral one proton). Unfortunately, the
9

ACCEPTED MANUSCRIPT
13
13
sample’s concentration was too low to observe the carbonyl C signals, through neither direct C-
NMR nor HMBC.
P III was the major product at the end of the reaction (with about 70% yield) and identified as
adrenosterone (4-androst-ene-3,11,17-trione) by means of an authentic standard. As the
corresponding fission products, we found traces of formic acid (HCOOH) and carbonic acid (H
2
CO ) by
3
ion chromatography (IC; data not shown) and HPLC with UV detection (Fig. S6). In addition, the
formation of carbon dioxide (that is in equilibrium with carbonic acid and bicarbonate) in the reaction
solution was confirmed by the formation of a whitish smear of precipitating barium carbonate (BaCO
when gaseous substances formed were stripped from the reaction solution and flushed through a
barium hydroxide solution (Fig. S7). Unexpectedly, glycolic acid (HO-H C-COOH) and glyoxylic acid
OHC-COOH) – the supposed fission products – were not detectable in the reaction solutions. On the
other hand, glycolic acid was found to be a substrate of MroUPO that was rapidly oxidized into glyoxylic
acid, formic acid and carbon dioxide as proved by HPLC-UV, IC and BaCO formation (data not shown).
3
)
2
(
3
The residual UPO activity in the reaction solution followed an interesting biphasic time course (both
for MroUPO and MweUPO). As long as S I and P I were present in the reaction solution, the enzyme
activity remained rather stable, however, after their consumption (between 120 and 140 min), the
activity strongly decreased, probably caused by unfavorable catalysis during conversion of P II into the
final product P III (Fig. 2).
The conversion of prednisone (S II) and Reichstein’s substance S (S III) by MroUPO and MweUPO gave
similar results as for S I. Both compounds were converted in three ways following the mass changes
Δm/z= +16, Δm/z= +14 and Δm/z= -60. The latter mass indicates that the last step was again a side
chain removal resulting in the formation of 4-androsten-3,17-dione and 1,4-androstadien-3,11,17-
trione, respectively (data not shown). AaeUPO oxidized S II and S III to negligible extent (<1%) with
traces of S (-60) as sole product.
3
.3. Structure and metal content of MroUPO
Based on the recently published coordinates and first crystal structure of MroUPO [21], we performed
a docking analysis using cortisone (substrate S I) as ligand and MroUPO as receptor with the open-
source program AutoDock Vina. Corresponding Fig. 6 shows that the corticosteroid molecule almost
perfectly fits into the active site of MroUPO, while exposing the side chain to be removed towards the
reactive heme center. In the computed model, cortisone’s C21 appears in close proximity to the heme
iron (4 Å). The binding pocket virtually consists just of aliphatic residues (leucine, Leu; isoleucine, Ile)
and contains merely one aromatic residue (phenylalanine, Phe). One polar serine residue (Ser156) is
also found in this pocket; however, this residue does seemingly not form polar contacts to cortisone
in the best fitting docking models. The keto groups at C3 and C11 of cortisone lie outside the binding
1
0

ACCEPTED MANUSCRIPT
pocket. The hydroxyl group at C17 as well as the keto group at C20 do not have polar contacts to amino
acid residues of MroUPO. In the second best fitting model (with correct orientation of the
−
1
hydroxyacetyl group to heme but slightly less binding affinity of -8.2 kcalꢀmol ), these both oxygen
functionalities build intra-molecular contacts. The hydrogen of the hydroxyl group at C21 can build a
polar contact to the hydroxyl-oxygen of the carboxylic group of a glutamate (Glu157, the distal acid-
base catalyst) and thereby stabilizes correct substrate binding.
Since magnesium (Mg) had been found in the crystal structure of AaeUPO [18], we checked whether
MroUPO does also contain this probably stabilizing metal. ICP-MS and ICP-OES analyses in fact revealed
that Mg was along iron (Fe) the most abundant metal in purified MroUPO. Its amount corresponded
to nearly one mol Mg per UPO molecule (Table S4). In contrast, the concentration of other metals,
including manganese (Mn) and calcium (Ca) reported to be present in other heme peroxidases [26-28],
were noticeably lower.
4
. Discussion
Two UPOs from the mushroom genus Marasmius (MroUPO, MweUPO) were found to convert
corticosteroids with hydroxyacetyl (-CO-CH OH) and alcohol (-OH) functionalities at C17 via stepwise
2
oxygenation/oxidation and final removal of the side chain (C-C cleavage, deacylation). In contrast, the
model enzyme AaeUPO known to oxidize diverse other substrates, failed in catalyzing this reaction [10,
2
9]. In a previous study, three UPOs including AaeUPO and MroUPO were shown to hydroxylate
steroids with longer side chains (with 8-10 carbon atoms) such as cholesterol, ergosterol and sitosterol
preferably at tertiary C25, their scission, however, was not observed [30]. Another UPO from the
ascomycetous mold Chaetomium globosum (CglUPO) has recently been demonstrated to selectively
epoxidize (at 4,5-position) testosterone that is neither a substrate of MroUPO nor of AaeUPO [31].
The ability of MroUPO and MweUPO to convert bulky substrates such as cortisone (S I), prednisone
(
S II) and Reichstein’s substance S (S III) can be attributed to the size and molecular architecture of the
enzymes’ active site. Both Marasmius enzymes belong to the cluster (subfamily) of “short” UPOs
whereas AaeUPO is a “long” UPO that differs in several features from the former ones [29]. Thus, the
heme access channel of MroUPO is about 4 Å shorter and up to 5 Å wider than that of AaeUPO.
Furthermore, the hydrophobic amino acid residues flanking the heme access channel are different.
While AaeUPO contains nine rigid Phe and one Tyr forming a rather narrow ‘carafe-shaped’ heme
channel (Fig. 7 B), MroUPO’s heme channel comprises only one Phe but ten more flexible aliphatic
amino acids (1 Ala, 2 Leu, 7 Ile; Fig. 6) and thus forms a rather open and symmetric frustum (Fig. 7 A)
[18]. These structural differences may explain, which is why MroUPO (and also MweUPO that shares
1
1

ACCEPTED MANUSCRIPT
more than 90% sequence identity with MroUPO, unpublished result) harbor and convert bulky
substrates such as corticosteroids whereas AaeUPO cannot (Fig. 6).
A plausible reaction sequence involving the stepwise oxygenation and final deacylation of cortisone-
and related substrates (S I-III) by MroUPO (and respectively by MweUPO) can be described as follows.
In the first step, the enzymes hydroxylates the position C21 in the side chain by means of H
2
O
2
(peroxygenation) resulting in the formation of P I that is a geminal alcohol (21-gem-diol, aldehyde
hydrate) being in equilibrium with the corresponding free aldehyde (cortisone 21-al, P Ia) (Fig. 8). A
similar reaction was observed during the oxidation of 5-hydroxymethylfurfural (HMF) by AaeUPO (or
aryl alcohol oxidase) leading to the corresponding dialdehyde (diformylfurane), about 50% of which
was in equilibrium with its stable hydrate (gem-diol detectable by MS) [32, 33].
In the next step, the Marasmius UPOs catalyze a second hydroxylation leading to the corresponding
-ketocarboxylic acid P II (corstisone 21-oic acid). However, so evident the structure of P II is (compare
respective NMR data), so uncertain is the exact point of enzymatic attack at P I. This is due to the
different tautomeric and hydrated forms of P I. Theoretically – in addition to P I and P Ia (the latter
detected by NMR, Table S2) – two further tautomeric structures, a 21-endiol and an -hydroxy-21-oic
acid, are conceivable; all of them would be substrates of Marasmius UPOs and converted into P II. In
our opinion, the most probable route may proceed via the hydroxylation of the free aldehyde (P Ia;
Fig. 8). This reaction type via abstraction of hydrogen (C-H abstraction) and OH-rebound has been
reported for various UPO substrates including aromatic (e.g. benzaldehyde) and aliphatic (n-alkanals)
aldehydes, which are oxidized into the corresponding carboxylic acids (e.g. benzoic acid and n-
alkanoicfatty acids, respectively) [34, 35]. An alternative route may start from the 21-gem-diol and
proceed via an additional hydroxylation at C21 to form a putative 21-gem-triol that spontaneously
dehydrates under formation of the 21-oic acid, Figure 8). A third variant may follow the hydroxylation
of the -hydroxy-21-oic acid tautomer of P I to give the corresponding 20-gem-diol that dehydrates to
P II; eventually, the epoxidation of the tautomeric 21-enediol could contribute to P I conversion as well
(not shown). Corticosterone with the same hydroxyacetyl moiety as cortisone (but lacking OH at C17)
was reported to be analogously oxidized by a panel of liver CYPs (especially of the CYP3A and CYP4A
subfamilies), involving 21-gem-diol, 21-gem-triol and 21-al intermediates and resulting in the final
formation of the corresponding 21-oic acid [36].
The last step in the conversion of corticosteroids by Marasmius UPOs proceeds via the removal of the
side chain, which results, in the case of P II, in the appearance of adrenosterone (17-one, P III) (Figure
9
). Although the formation of the latter compound was unambiguously demonstrated (by an authentic
standard), it is the most difficult-to-explain step of the reaction sequence [37] (Figure 9). Taking into
consideration the existing knowledge on steroidal side-chain removal by 17-hydroxylase/C17,20-
lyase (CYP17A1) [8], recently described UPO-catalyzed fatty acid decarboxylation [21] and existing
1
2

ACCEPTED MANUSCRIPT
knowledge on chemical decarboxylation of -ketoacids by hydrogen peroxide [38, 39], it appears to
be plausible that P II undergoes both enzymatic and chemical oxidation in the course of side-chain
removal (Figure 9). Thus, a lyase-like reaction resulting in C-C bond scission between C17 and C20,
release of glyoxylic acid (that is immediately oxidized/cleaved by UPO to formic acid and CO /carbonic
2
acid) and the formation of a new double bond at C17 (carbonyl, P III) is as conceivable as the peroxide-
-
driven oxidation of P II (caused by the peroxo-anion/HOO at the pH used [38]), which results via a
peroxo-intermediate in the release of CO and a C17-ol/C17-oic acid derivative (P IV, the mass of which
2
+
fits to the detected m/z [H ] of 347); in a final spontaneous step, less stable P IV may decay to formic
acid and P III (17-one) (Figure 9). Whether a compound-0-like UPO-intermediate (i.e. a bound peroxo-
anion-heme species) is involved as proposed for the C17,20-lyase reaction of CYP17A1 remains unclear
[40]. The rapid inactivation of Marasmius UPOs during the obviously unfavorable conversion of P II
into P III may support this assumption (Figure 2).
2+
The presence of equimolar amounts of magnesium (Mg ) in purified MroUPO confirms the importance
of this metal as structural element in UPO proteins. Strong indication of its presence were also found
in the crystal structure of AaeUPO and it was detected by ICP-MS/OES analysis in highly pure
preparations of the enzyme [18]. It may stabilize the porphyrin ring system that is under permanent
2+
ring strain. Interestingly, the Mg ion sits at the same position as Mn in manganese peroxidase (MnP,
EC 1.11.13.1) [41]. In both cases, the metals are chelated by conserved acidic amino acid residues (Glu,
Asp) and the heme propionate [18, 41]. The presence of manganese was proposed for CPO based on
its crystal structure (but without appropriate metal analysis) [26]. However, when considering the
2+
function of Mn in MnP [42] and the almost identical covalent atomic radiuses of Mn (1.39 Å) and
2+
2+
Mg (1.41 Å) [43], it is more conceivable that redox-inactive Mg is present in CPO representing a
halide-oxidizing UPO [14].
5
. Conclusions
After recently reported fatty acid shortening by MroUPO [21], steroidal side-chain removal is the
second example of a C-C cleaving reaction (deacylation) catalyzed by this enzyme type. Thereby
another prototypical reaction of P450s (in this case lyase-activity of CYP17A1) with physiological and
biotechnological relevance could be ascertained for a secreted fungal peroxygenase. Future studies
will have to elucidate the role of UPO compound 0 and whether other fungal UPOs (of which more
than 4,000 sequences are known so far) may catalyze similar reactions without loosing their activity
during the scission process.
Abbreviations
1
3

ACCEPTED MANUSCRIPT
Aae, Agrocybe aegerita; ACN, acetonitrile; CPO, chloroperoxidase; CYP, cytochrome P450; ESI,
electrospray ionization; HPLC, high performance liquid chromatography; IC, ion chromatography; ICP-
OES, inductively coupled plasma optical emission spectrometry; Mro, Marasmius rotula; Mwe,
Marasmius wettsteinii; MS, mass spectrometry; NMR, nuclear magnetic resonance; SDS-PAGE, sodium
dodecyl sulfate polyacrylamide gel electrophoresis; SEC, size-exclusion chromatography; S I, cortisone;
S II, prednisone, S III, Reichstein’s substance S; TON, turnover number; UPO, unspecific peroxygenase
Conflicts of interest
The authors declare that they have no conflict of interest.
Acknowledgements
We thank U. Schneider for isolating Marasmius wettsteinii and maintaining the Zittau Fungal Culture
Collection and H. Heidenreich for performing ICP-OES and ICP-MS as well as A. Karich for the
photography of M. rotula (graphical abstract). This work was supported by the German Ministry of
Education and Research (Bundesministerium für Bildung und Forschung) within the program
‘
BioIndustrie 2021-Cluster Integrierte Biotechnololgie 2021’ (project 0315877) and in part by the
integrated European Union projects INDOX (KBBE 2013.3.3-04) and ENZOX2 (H2020-BBI-PPP-2015-2-1-
20297). Ongoing sequencing and annotation of the genome of M. wettsteinii is supported by the AiF
7
project PEROXYMEER (IGF 19636 BG/3).
Appendix A. Supplementary Information (SI)
References
[
[
[
1] K. Bloch, Steroids. 57 (1992) 378-383.
2] W.L. Miller, R.J. Auchus, Endocr. Rev. 32 (2011) 81-151.
3] D.E. Cane, D.A. Bochar, J. Freisen, C.V. Stauffacher, V.W. Rodwell, M. Rohmer, T. Koyama, K.
Ogura, M.L. Wise, R. Croteau, T.M. Hohn, Isoprenoids including cartenoids and steroids, first ed.,
Pergamon, Oxford, 1999.
[
4] R.A. Hill, H.L.J. Makin, D.N. Kirk, G.M. Murphy, Dictionary of steroids, first ed., Chapman and
Hall/CRC, London, 1991.
1
4

ACCEPTED MANUSCRIPT
[
[
[
[
[
[
[
[
5] J.A.R. Salvador, S.M. Silvestre, V.M. Moreira, Curr. Org. Chem. 10 (2006) 2227-2257.
6] S.B. Mahato, S. Garai, Steroids. 62 (1997) 332-345.
7] H.N. Bhatti, R.A. Khera, Steroids. 77 (2012) 1267-1290.
8] D. Porubek, Curr. Top. Med. Chem. 13 (2013) 1364-1384.
9] L. Sedlaczek, L.L. Smith, Crit. Rev. Biotechnol. 7 (1988) 187-236.
10] M. Hofrichter, R. Ullrich, Curr. Opin. Chem. Biol. 19 (2014) 116-125.
11] M. Kluge, R. Ullrich, K. Scheibner, M. Hofrichter, Green Chem. 14 (2012) 440-446.
12] Y. Ni, E. Fernández-Fueyo, A.G. Baraibar, R. Ullrich, M. Hofrichter, H. Yanase, M. Alcalde, W.J.H.
van Berkel, F. Hollmann, Angew. Chem. Int. Ed. Engl. 55 (2016) 798-801.
[
13] A.E.W. Horst, S. Bormann, J. Meyer, M. Steinhagen, R. Ludwig, A. Drews, M. Ansorge-
Schumacher, D. Holtmann, J. Mol. Catal. B: Enzym. 133 (2016) S137-S142.
[
14] H. Kellner, M.J. Pecyna, M. Buchhaupt, R. Ullrich, M. Hofrichter, Genome Announc. 4 (2016)
e00774-00716.
[
15] X. Wang, S. Peter, M. Kinne, M. Hofrichter, J.T. Groves, J. Am. Chem. Soc. 134 (2012) 12897-
2900.
1
[16] X. Wang, R. Ullrich, M. Hofrichter, J.T. Groves, Proc. Natl. Acad. Sci. U.S.A. 112 (2015) 3686-3691.
[
17] K. Piontek, R. Ullrich, C. Liers, K. Diederichs, D.A. Plattner, M. Hofrichter, Acta Crystallogr. F. 66
2010) 693-698.
(
[
18] K. Piontek, E. Strittmatter, R. Ullrich, G. Gröbe, M.J. Pecyna, M. Kluge, K. Scheibner, M.
Hofrichter, D.A. Plattner, J. Biol. Chem. 288 (2013) 34767-34776.
1
5

ACCEPTED MANUSCRIPT
[
19] G. Gröbe, R. Ullrich, M.J. Pecyna, D. Kapturska, S. Friedrich, M. Hofrichter, K. Scheibner, AMB
Express. 1 (2011) 31.
[
20] R. Ullrich, J. Nüske, K. Scheibner, J. Spantzel, M. Hofrichter, Appl. Environ. Microbiol. 70 (2004)
575-4581.
4
[
21] A. Olmedo, J.C. del Río, J. Kiebist, R. Ullrich, M. Hofrichter, K. Scheibner, A.T. Martínez, A.
Gutiérrez, Chem. Eur. J.2017) 16985-16989.
[
22] C. Liers, M.J. Pecyna, H. Kellner, A. Worrich, H. Zorn, K.T. Steffen, M. Hofrichter, R. Ullrich, Appl.
Environ. Microbiol. 97 (2013) 5839-5849.
[
[
23] O. Trott, A.J. Olson, J. Comput. Chem. 31 (2010) 455-461.
24] S. Kim, P.A. Thiessen, E.E. Bolton, J. Chen, G. Fu, A. Gindulyte, L. Han, J. He, S. He, B.A.
Shoemaker, J. Wang, B. Yu, J. Zhang, S.H. Bryant, Nucleic Acids Res. 44 (2016) D1202-D1213.
[
25] G.M. Morris, R. Huey, W. Lindstrom, M.F. Sanner, R.K. Belew, D.S. Goodsell, A.J. Olson, J.
Comput. Chem. 30 (2009) 2785-2791.
[
[
[
26] M. Sundaramoorthy, J. Terner, T.L. Poulos, Structure. 3 (1995) 1367-1378.
27] G. Nie, S.D. Aust, Arch. Biochem. Biophys. 337 (1997) 225-231.
28] M. Matsubara, J. Suzuki, T. Deguchi, M. Miura, Y. Kitaoka, Appl. Environ. Microbiol. 62 (1996)
4
066-4072.
[
29] M. Hofrichter, H. Kellner, M.J. Pecyna, R. Ullrich, Fungal unspecific peroxygenases: Heme-
thiolate proteins that combine peroxidase and cytochrome P450 properties, in: E.G. Hrycay, S.M.
Bandiera (Eds.), Monooxygenase, peroxidase and peroxygenase properties and mechanisms of
cytochrome P450, Springer International Publishing, Cham, 2015, pp. 341-368.
1
6

ACCEPTED MANUSCRIPT
[
30] E.D. Babot, J.C. del Río, M. Cañellas, F. Sancho, F. Lucas, V. Guallar, L. Kalum, H. Lund, G. Gröbe,
K. Scheibner, R. Ullrich, M. Hofrichter, A.T. Martínez, A. Gutiérrez, Appl. Environ. Microbiol. 81 (2015)
130-4142.
4
[
31] J. Kiebist, K.-U. Schmidtke, J. Zimmermann, H. Kellner, N. Jehmlich, R. Ullrich, D. Zänder, M.
Hofrichter, K. Scheibner, ChemBioChem. 18 (2017) 563-569.
[32] J. Carro, P. Ferreira, L. Rodríguez, A. Prieto, A. Serrano, B. Balcells, A. Ardá, J. Jiménez-Barbero, A.
Gutiérrez, R. Ullrich, M. Hofrichter, A.T. Martínez, FEBS J. 282 (2015) 3218-3229.
[
33] L. Kalum, M.D. Morant, H. Lund, J. Jensen, I. Lapainaite, S.N. H., S. Pedersen, L.H. Østergaard, F.
Xu, Enzymatic oxidation of 5-hydroxymethylfurfural and derivatives thereof, WO 2014/015256 A2,
Denmark, 2014.
[
34] M. Kinne, C. Zeisig, R. Ullrich, G. Kayser, K.E. Hammel, M. Hofrichter, Biochem. Biophys. Res.
Commun. 397 (2010) 18-21.
[
35] A. Olmedo, C. Aranda, J.C. del Río, J. Kiebist, K. Scheibner, A.T. Martínez, A. Gutiérrez, Angew.
Chem. Int. Ed. Engl. 55 (2016) 12248-12251.
[
36] T. Wang, Y.M. Shah, T. Matsubara, Y. Zhen, T. Tanabe, T. Nagano, S. Fotso, K.W. Krausz, T.M.
Zabriskie, J.R. Idle, F.J. Gonzalez, J. Biol. Chem. 285 (2010) 7670-7685.
[
[
[
[
37] S. Bonomo, F.S. Jørgensen, L. Olsen, J. Chem. Inf. Model. 57 (2017) 1123-1133.
38] E. Melzer, H.L. Schmidt, Biochem. J. 252 (1988) 913-915.
39] B. Siegel, J. Lanphear, J. Org. Chem. 44 (1979) 942-946.
40] M. Akhtar, J.N. Wright, Acyl-carbon bond cleaving cytochrome P450 enzymes: CYP17A1,
CYP19A1 and CYP51A1, in: E.G. Hrycay, S.M. Bandiera (Eds.), Monooxygenase, peroxidase and
peroxygenase properties and mechanisms of cytochrome P450, Springer International Publishing,
Cham, 2015, pp. 107-130.
1
7

ACCEPTED MANUSCRIPT
[
[
[
41] M. Sundaramoorthy, K. Kishi, M.H. Gold, T.L. Poulos, J. Biol. Chem. 269 (1994) 32759-32767.
42] H. Wariishi, K. Valli, M.H. Gold, J. Biol. Chem. 267 (1992) 23688-23695.
43] The periodic table of the elements, Manganese: radii of atoms and ions.
https://www.webelements.com/manganese/atom_sizes.html, 2018 (accessed 5 January 2018).
[
44] S. Nakajin, M. Shinoda, M. Haniu, J.E. Shively, P.F. Hall, J. Biol. Chem. 259 (1984) 3971-3976.
1
8

ACCEPTED MANUSCRIPT
Legends
Figure 1: Chemical structures of corticosteroids used in this study: S I cortisone (A), S II prednisone (B)
and S III Reichstein’s substance S (C).
Figure 2: Time courses of the conversion of cortisone by MroUPO and MweUPO. Analyses were carried
out by HPLC at 240 nm. Data are given in mAU*s of the respective peak areas (top) and in µM (bottom)
along with the residual UPO activities (bottom). Black circles - cortisone (S I), blue up-pointing triangles
-
product P I (cortisone 21-gem-diol, m/z 376), red diamonds - product P II (cortisone 21-oic acid, m/z
74), green squares - product P III (adrenosterone, m/z 300), brown crosses - sum of peak areas, black
3
asterisks - cortisone (S I) control, orange down- pointing triangles - residual UPO activity. Arrows mark
the point of complete substrate consumption (S I, cortisone). Reaction solutions contained 1 U mL-1
UPO, 0.5 mM cortisone (S I) in 10 mM potassium phosphate buffer (pH 5.5). H
2
O was supplied by
2
-1
syringe pump at 0.5 mM h .
-
1
Figure 3: HPLC elution profiles of reaction solutions containing cortisone (S I, 0.5 mM) and 1 U mL of
three different UPOs; they were recorded after different reaction times (0-220 min). H was supplied
2
O
2
-
1
by syringe pump at 0.5 mM h over the indicated time. A) MroUPO, B) MweUPO and C) AaeUPO.
Figure 4: Chemical structure of the cortisone metabolite P I with a molecular mass of 376.44 g mol-1
and the molecular formula C21
H
28
O confirmed by NMR (left) as well as the corresponding dehydrated
6
aldehyde P Ia being in equilibrium with P I. Bottom: HMBC spectrum of P I indicating the presence of
the corresponding free aldehyde (P Ia).
-
1
Figure 5: Proposed structure of metabolite P II with a molecular mass of 374.43 g mol and following
molecular formula: C21
26
H O
6
Figure 6: 3D-model of MroUPO when harboring cortisone (S I, in magenta) as substrate (MroUPO-
cortisone complex). Left side: intersection of MroUPO with bound cortisone; middle: detailed view of
the substrate channel (heme access channel) including catalytic residues Glu-157 (distal acid-base
catalyst) and His-86 (charge stabilizer). Polar contacts of the hydroxyl group at C21 of cortisone are
shown as yellow dashed lines; right side: top view of the substrate channel with and without bound
cortisone (all residues forming the substrate channel are depicted and labeled). For convenience of
the reader, the stabilizing magnesium ion (green ball) is shown in all views.
Figure 7: Tangential cross sections through crystal structure-based 3D-models of MroUPO and
AaeUPO; the former (A) has a rather open and symmetric frustum-like heme access channel whereas
that of the latter (B) is narrower and ‘carafe-shaped’.
1
9

ACCEPTED MANUSCRIPT
Figure 8: Postulated reactions involved in the side chain oxidation of S I (cortisone) by MroUPO and
MweUPO. First hydroxylation at C21 results in P I (cortisone 21-gem-diol, identified by NMR) that is in
equilibrium with the corresponding free aldehyde P Ia (cortisone 21-al, also identified by NMR). Both
P I and P Ia can be subject of a second hydroxylation at C21 resulting in the direct formation of P II
#
(
cortisone 21-oic acid, detected by NMR) or in putative cortisone gem-triol that is in equilibrium with
P II.
Figure 9: Possible reaction sequences explaining the side chain removal from P II (cortisone 21-oic acid)
by MroUPO and MweUPO. 1) A lyase-like activity of UPO similar to that described in [44] may cleave
the C-C bond between C17 and C20 resulting in P III (adrenosterone, identified by an authentic
standard) and putative glyoxylic acid that can be cleaved by UPO into formic acid and carbonic acid /
carbon dioxide (as demonstrated by authentic standards and barium carbonate precipitation). 2) P II
may also chemically react with H
2
O to form a tetrahedral intermediate [38] that decomposes to
2
carbon dioxide and P IV (17-hydroxy-3,11-dioxoandrost-4-ene-17-carboxylic acid, as indicated by mass
data).
2
0

ACCEPTED MANUSCRIPT
Highlights





peroxygenases from mushrooms of the genus Marasmius catalyze cortisone deacylation
oxidation starts with hydroxylation at C21 and formation of a stable 21-gem-diol
nd
alpha-keto-21-oic acid results from 2 oxygenation and decays under C1-compound release
adrenosterone, the final product of cortisone oxidation, is an anabolic drug precursor
peroxygenase of Marasmius rotula is a functional dimer with a stabilizing magnesium
2
1

Graphics Abstract

Figure 1

Figure 2

Figure 3

Figure 4

Figure 5

Figure 6

Figure 7