The efficient Δ1-dehydrogenation of a wide spectrum of 3-ketosteroids in a broad pH range by 3-ketosteroid dehydrogenase from Sterolibacterium denitrificans

Article abstract of DOI:10.1016/j.jsbmb.2020.105731

Cholest-4-en-3-one Δ¹-dehydrogenase (AcmB) from Sterolibacterium denitrificans, a key enzyme of the central degradation pathway of cholesterol, is a protein catalyzing Δ¹-dehydrogenation of a wide range of 3-ketosteroids. In this study, we demonstrate the application of AcmB in the synthesis of 1-dehydro-3-ketosteroids and investigate the influence of reaction conditions on the catalytic performance of the enzyme. The recombinant AcmB expressed in E. coli BL21(DE3)Magic exhibits a broad pH optimum and pH stability in the range of 6.5 to 9.0. The activity-based pH optimum of AcmB reaction depends on the type of electron acceptor (2,6-dichloroindophenol - DCPIP, phenazine methosulfate - PMS or potassium hexacyanoferrate - K₃[Fe(CN)₆]) used in the biocatalytic process yielding the best kinetic properties for the reaction with a DCPIP/PMS mixture (k_cat/K_m = 1.4·10⁵ s^-1·M^-1 at pH 9.0) followed by DCPIP (k_cat/K_m = 1.0·10⁵ s^-1·M^-1 at pH = 6.5) and K₃[Fe(CN)₆] (k_cat/K_m = 0.5·10² s^-1·M^-1 at pH = 8.0). The unique feature of AcmB is its capability to convert both testosterone derivatives (C20–C22) as well as steroids substituted at C17 (C27-C30) such as cholest-4-en-3-one or (25R)-spirost-4-en-3-one (diosgenone). Apparent steady-state kinetic parameters were determined for both groups of AcmB substrates. In a batch reactor synthesis, the solubility of water-insoluble steroids was facilitated by the addition of a solubilizer, 2-hydroxypropyl-β-cyclodextrin, and organic co-solvent, 2-methoxyethanol. Catalytic properties characterization of AcmB was tested in fed-batch reactor set-ups, using 0.81 μM of isolated enzyme, PMS and aerobic atmosphere resulting in >99% conversion of the C17–C20 3-ketosteroids within 2 h. Finally, the whole cell E. coli system with recombinant enzyme was demonstrated as an efficient biocatalyst in the synthesis of 1-dehydro-3-ketosteroids.

Full text of DOI:10.1016/j.jsbmb.2020.105731

Journal Pre-proof
The efficient ꢀ1-dehydrogenation of a wide spectrum of 3-ketosteroids in
a broad pH range by 3-ketosteroid dehydrogenase from Sterolibacterium
denitrificans
Agnieszka M. Wojtkiewicz (Writing - original draft) (Investigation)
(Formal analysis) (Supervision)<ce:contributor-role>Writing - revew
´
and editing), Patrycja Wojcik (Writing - original draft) (Investigation)
(Validation) (Writing - review and editing) (Visualization), Magdalena
Procner (Investigation) (Formal analysis), Monika Flejszar
(Investigation), Maria Oszajca (Resources) (Supervision) (Writing -
review and editing), Mateusz Hochołowski (Investigation), Mateusz
Tataruch (Investigation), Beata Mrugała (Investigation)
(Methodology) (Formal analysis), Tomasz Janeczko (Supervision)
(Investigation) (Resources) (Writing - review and editing), Maciej
Szaleniec (Conceptualization) (Supervision) (Visualization) (Project
administration) (Writing - original draft) (Writing - review and editing)
(Funding acquisition)
PII:
S0960-0760(20)30256-9
DOI:
https://doi.org/10.1016/j.jsbmb.2020.105731
SBMB 105731
Reference:
To appear in:
Journal of Steroid Biochemistry and Molecular Biology
Received Date:
Revised Date:
Accepted Date:
20 April 2020
20 July 2020
27 July 2020
´
Please cite this article as: Wojtkiewicz AM, Wojcik P, Procner M, Flejszar M, Oszajca M,
Hochołowski M, Tataruch M, Mrugała B, Janeczko T, Szaleniec M, The efficient
ꢀ1-dehydrogenation of a wide spectrum of 3-ketosteroids in a broad pH range by
3-ketosteroid dehydrogenase from Sterolibacterium denitrificans, Journal of Steroid

Biochemistry and Molecular Biology (2020), doi: https://doi.org/10.1016/j.jsbmb.2020.105731
This is a PDF file of an article that has undergone enhancements after acceptance, such as
the addition of a cover page and metadata, and formatting for readability, but it is not yet the
definitive version of record. This version will undergo additional copyediting, typesetting and
review before it is published in its final form, but we are providing this version to give early
visibility of the article. 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.
© 2020 Published by Elsevier.

The efficient Δ1-dehydrogenation of a wide spectrum of 3-ketosteroids in a broad pH range
by 3-ketosteroid dehydrogenase from Sterolibacterium denitrificans
Agnieszka M. Wojtkiewicz1,#, Patrycja Wójcik1,#, Magdalena Procner1, Monika Flejszar2,1,3
,
Maria Oszajca3, Mateusz Hochołowski1, Mateusz Tataruch1, Beata Mrugała1, Tomasz Janeczko4,
Maciej Szaleniec1,*
1 Jerzy Haber Institute of Catalysis and Surface Chemistry, Polish Academy of Sciences,
Niezapominajek 8, PL30239, Krakow, Poland
2 Department of Physical Chemistry, Faculty of Chemistry, Rzeszow University of Technology,
Al. Powstańców Warszawy 6, PL35959 Rzeszów, Poland
3 Faculty of Chemistry, Jagiellonian University, Gronostajowa 2, PL30387 Kraków
4 Department of Chemistry, Wrocław University of Environmental and Life Sciences, Norwida
25, PL50375 Wrocław, Poland
# These authors contributed equally to this work
* Corresponding author: ncszalen@cyfronet.pl; tel: +48 12 6395 101, fax: +48 12 425 19 23
GRAPHICAL ABSTRACT
1

HIGHLIGHTS
 AcmB exhibits two pH optima: at pH 6.5 and 8-9
 pH optimum depends on electron acceptor used in the
reaction
 The most active enzyme is a crude extract treated with
FAD and Triton X-100
 Reaction with PMS at pH 8.0 yields highest conversions
 AcmB dehydrogenates 3-ketosaponins, e.g. (25R)-
spirost-4-en-3-one (diosgenone)
Abstract
Cholest-4-en-3-one Δ1-dehydrogenase (AcmB) from Sterolibacterium denitrificans, a key
enzyme of the central degradation pathway of cholesterol, is a protein catalyzing Δ1-
dehydrogenation of a wide range of 3-ketosteroids. In this study, we demonstrate the application
of AcmB in the synthesis of 1-dehydro-3-ketosteroids and investigate the influence of reaction
conditions on the catalytic performance of the enzyme. The recombinant AcmB expressed in E.
2

coli BL21(DE3)Magic exhibits a broad pH optimum and pH stability in the range of 6.5 to 9.0.
The activity-based pH optimum of AcmB reaction depends on the type of electron acceptor (2,6-
dichloroindophenol - DCPIP, phenazine methosulfate - PMS or potassium hexacyanoferrate -
K3[Fe(CN)6]) used in the biocatalytic process yielding the best kinetic properties for the reaction
with a DCPIP/PMS mixture (kcat/Km =1.4·105 s-1·M-1 at pH 9.0) followed by DCPIP
(kcat/Km=1.0·105 s-1·M-1 pH=6.5) and K3[Fe(CN)6] (kcat/Km =0.5·102 s-1·M-1 pH=8.0). The unique
feature of AcmB is its capability to convert both testosterone derivatives (C20-C22) as well as
steroids substituted at C17 (C27-C30) such as cholest-4-en-3-one or (25R)-spirost-4-en-3-one
(diosgenone). Apparent steady-state kinetic parameters were determined for both groups of
AcmB substrates. In a batch reactor synthesis, the solubility of water-insoluble steroids was
facilitated by the addition of a solubilizer, 2-hydroxypropyl-β-cyclodextrin, and organic co-
solvent, 2-methoxyethanol. Catalytic properties characterization of AcmB was tested in fed-batch
reactor set-ups, using 0.81 µM of isolated enzyme, PMS and anaerobic atmosphere resulting in
>99% conversion of the C17-C20 3-ketosteroids within 2 h. Finally, the whole-cell E. coli system
with recombinant enzyme was demonstrated as an efficient biocatalyst in the synthesis of 1-
dehydro-3-ketosteroids.
Keywords: 3-ketosteroid Δ1-dehydrogenase, Δ1-dehydrogenation, 1,2-dehydrogenation, steroid
biotransformation, phenazine methosulfate, diosgenone
1. Introduction
Steroids as ubiquitous contamination of water and soil environments serve as a carbon
and energy source for many microorganisms. Steroids degradation is postulated for many species
from phyla Actinobacteria and Proteobacteria from soil, eukaryotic and aquatic environments
3

[1]. For all steroid degraders one of the key enzymes in the catabolic pathway is flavin adenine
dinucleotide (FAD)-dependent 3-ketosteroid Δ1-dehydrogenase (KstD), which catalyzes 1,2-
dehydrogenation of androst-4-en-3,17-dione (AD) to androst-1,4-dien-3,17-dione (ADD). ADD
is further hydroxylated, which leads to the ring A opening. Depending on the microorganism
environment, either aerobic or anaerobic, the ring A opening proceeds via 9,10-seco- or 2,3-seco-
degradation pathway, respectively [2]. The first scenario was well characterized for cholesterol,
testosterone or cholate degradation pathway for Rhodococcus jostii RHA1, Mycobacterium
tuberculosis H37Rv, Comamonas testosteroni TA441 and Pseudomonas sp. strain Chol1 [1,3,4]
while the second scenario was confirmed for denitrifying Sterolibacterium denitrifcans Chol-1S
[5,6].
Several biotechnological approaches for steroid modifications have been developed in
recent decades. The first and already commercialized approach uses wild-type strains (of genus
Mycobacterium) to conduct the biotransformation of cholesterol or phytosterol to desired steroid
synthons [7,8]. The next-generation approach employed genetically improved strains with either
increased expression of a desired enzyme or knockout of the gene(s) responsible for the
degradation of the desired product [9–11]. Such an approach is characterized by increased
selectivity and productivity toward the desired product. Finally, novel systems are being
developed that focus on specific functionalization of steroids, such as regioselective
hydroxylation, Baeyer-Villiger oxidations or 1-dehydrogenation. As such modifications rely on
a singular specific catalyst it is convenient to overexpress the selected enzyme in strains that do
not contain enzymes for steroid catabolism. In the context of 1-dehydrogenation, the last
approach was recently intensely studied. For example, KstDs from Mycobacterium neoaurum JC-
12 [12,13], Mycobacterium neoaurum DSM 1381 [14] or Mycobacterium smegmatis mc2155 and
4

Rhodococcus erythropolis WY 1406 [15] overexpressed in other microorganisms were
characterized in the conversion of AD to ADD or hydrocortisone acetate to prednisolone acetate
[15]. It has been reported that different KstD isoforms present in the genome of the same
microorganism can differ in substrate preference [15–17], stability or pH optimum. Moreover, in
the last couple of years, KstDs have been engineered to improve their catalytic properties like
KstD3 from Arthrobacter simplex for AD to ADD conversion with increased tolerance to organic
solvents [18] or KstD from Mycobacterium neoaurum JC-12 that was co-expressed with catalase
to avoid inactivation by H2O2, a byproduct of a catalytic O2 reduction by flavin cofactor of the
enzyme [13].
In this work we concentrate on biocatalytic 1-dehydrogenation of various biologically relevant
3-ketosteroids by KstD from Sterolibacterium denitrificans Chol-1 named cholest-4-en-3-one Δ1-
dehydrogenase (AcmB, Anaerobic cholesterol metabolism enzyme B). Several previous reports
on FAD-dependent AcmB concentrated on its physiological role in the cholesterol degradation in
S. denitrificans [5,19] as well as its propensity for aggregation and determination of its functional
tertiary organization [20]. AcmB in physiological conditions is a monomer with a mass of 61.4 or
62 kDa for an overexpressed protein with and without His-tag [20]. Here, we demonstrate the
application potential of recombinant AcmB, overexpressed in E.coli BL21(DE)Magic, in the
synthesis of various 1-dehydro-3-ketosteroids and systematic analysis of its catalytic properties
and for the first time we demonstrate that pH optimum of KstD may depend on the type of
reoxidant used in the assay. Also for the first time we demonstrate that KstD can be applied to
dehydrogenation of a 3-ketosaponin, (25R)-spirost-4-en-3-one.
5

2. Materials and methods
2.1. Materials
All chemicals of analytical grade were purchased from Sigma-Aldrich (Germany), Avantor
Performance Materials (Poland), GE Healthcare (USA) or Carl Roth (Germany). Progesterone,
17-methyltestosterone and testosterone propionate were purchased from Sigma-Aldrich
(Germany); 6-dehydrotestosterone acetate was obtained from the resources of the Department of
Chemistry, the Wrocław University of Environmental and Life Sciences, Poland. Lupeol was a
gift of Dr. M. Malinowska, and T. Więcaszek, while diosgenone was a gift of Professor J.
Morzycki. Other steroids were purchased from Tokyo Chemical Industry (Japan) and Sigma-
Aldrich (Germany).
2.2. Protein expression and purification
A gene encoding AcmB was cloned into a pMCSG7 vector as described previously by Sofinska
et al work [20] and transformed into calcium chloride chemically competent E. coli
BL21(DE3)Magic. An overnight culture of E. coli BL21(DE3)Magic in 2% Lennox Broth (LB)
supplemented with 100 μg/mL ampicillin and 50 g/mL kanamycin was grown at 37°C, 180 rpm.
The overnight culture was diluted a hundred times in 1 L of ampicillin and kanamycin
supplemented LB and grown in the same conditions as the overnight culture until the OD600
reached 0.6. Then the temperature was reduced to 16°C and the culture was induced with 250 μM
isopropyl β-D-1-thiogalactopyranoside (IPTG). After 24 h cells were harvested by centrifugation
at 4500 g for 1 h at 4 °C. The cell pellets were resuspended 1:5 in the buffer 50 mM Tris-HCl pH
8.5, 150 mM NaCl, 5 mM imidazole, 10 % glycerol and 100 μM phenylmethylsulfonyl fluoride
(PMSF). The cell suspension was lysed by sonication (Sonics Vibra-Cell VCX500, 3 s on, 5 s
6

off, 5 min, 40% amplitude, 150 000 J), supplemented with 0.5% Triton-X 100 and 100 μM FAD
and incubated overnight at 4 °C. Cell debris were removed by centrifugation at 40 000 g for 1 h
at 4 °C. Afterward, the supernatant was applied onto a 5 mL HisTrap HP (GE Healthcare)
column. To remove impurities the column was washed with 8 column volumes of 50 mM Tris-
HCl pH 8.5, 150 mM NaCl, 15 mM imidazole, 10 % glycerol and 0.5% Triton-X 100. The target
protein was eluted with the analogous buffer with 300 mM imidazole concentration. The yellow-
colored AcmB fraction was applied onto a desalting column Econo-Pac 10DG (BioRad), eluted
with buffer 50 mM Tris-HCl pH 8.5, 150 mM NaCl, 10 % glycerol, 0.05% Triton-X 100 and 5
mM 2-mercaptoethanol (BME) and stored at -20 °C. A final enzyme concentration was
determined according to the Bradford method while FAD content (i.e., the concentration of
holoenzyme) was determined spectrophotometrically at 450 nm using a free FAD extinction
coefficient (ε450 = 13 100 M-1cm-1).
2.3. Thermostability assay
The enzyme thermostability in the function of pH was investigated using the ThermoFAD
method. This approach allowed us to determine the protein unfolding temperature by monitoring
changes in fluorescence signal upon the flavin cofactor release [21]. The assay was performed
using 5 μL of the purified and concentrated enzyme and 20 μL of the following buffers: 0.1 M
citrate buffer (pH 4.0–5.0), 0.1 M K2HPO4/KH2PO4 buffer (pH 6.0–8.0) and 0.1 M glycine-
NaOH buffer (pH 9.0–10). The thermal denaturation curves were collected by measuring the
fluorescence intensities on the FRET channel (450 to 490 nm) and the FAD fluorescence was
monitored at 470 nm. The fluorescence emission was followed at the temperature range from 4
°C to 95 °C with the 1°C increment per 15 s.
7

2.4. LC/MS methods
The LC/MS analyses were carried out on Agilent 1100 VL LC/DAD/MSD using Advanced
Materials Technology HALO 90 Å RP-Amide column (2.7 μm, 2.1×75 mm), thermostated at 40
oC, in acetonitrile/water gradient method described in Table S1. The MS detection was conducted
using ESI (compounds 1–10, 12–15) or APCI (compound 11) ion source in a positive ion scan
mode.
2.5. Enzyme activity assay
2.5.1. Steady-state spectrophotometric assay
The AcmB activity was measured using a UV-2700 spectrophotometer (Shimadzu) in 0.5 mL
quartz cuvettes with a 10 mm path length. The spectrophotometric assay was carried out at 30°C
in 0.1 M K2HPO4/KH2PO4 buffer pH 6.5 with 0.2 mM 2,6-dichloroindophenol (DCPIP) and 0.1
M progesterone in isopropanol (IPA) (final IPA concentration 2%). The reaction was started by
the addition of 0.15–15.2 μM AcmB. The reduction of DCPIP was followed at 700 nm (ε700[pH 6.5]
= 4576 M-1 cm-1). Each experiment was conducted in triplicate. The linear function was fitted to
the initial part (10 s) of the kinetic curves. Obtained data were fitted to the Michaelis-Menten
model with non-linear regression in OriginPro 2019 software.
2.5.2. Steady-state kinetics for cholest-4-en-3-one and cholest-4,6-dien-3-one
Apparent Km and Vmax values for cholest-4-en-3-one (7) and cholest-4,6-dien-3-one (8) were
determined in the steroid concentration range of 10 to 80 μM. Steroids were dissolved in 1,4-
8

dioxane (2%). Due to their low solubility, reactions were performed in the presence of 2% (w/v)
(2-hydroxypropyl)-β-cyclodextrin (HBC).
2.5.3. Steady-state kinetics for the artificial electron acceptors
Apparent steady-state kinetics parameters for the artificial electron acceptors were determined in
the concentration range of 20–500 μM for DCPIP, 0.1 –2 mM for phenazine methosulfate (N-
methylphenazinium methyl sulfate, PMS) with 200 μM DCPIP and 0.25–4 mM for K3[Fe(CN)6]
using 100 M progesterone (4) as a steroid substrate.
2.5.4. Temperature optimum and stability
The temperature optimum was determined according to the steady-state spectrophotometric
assay. The reactions were carried out at 20 °C, 30 °C, 35 °C, 40 °C and 50 °C. To evaluate the
AcmB stability protein was incubated on ice or in a water bath at 30 °C or 50 °C for 4 h. The
AcmB activity was measured every 30 minutes.
2.5.5. pH optimum and stability
The pH optimum was determined for the different artificial electron acceptors: DCPIP,
PMS/DCPIP mixture, PMS and K3[Fe(CN)6]. The reactions were conducted in the buffers pH
range from 4 to 10 specified in Thermostability assay. The dehydrogenation reaction was carried
out with 200 μM DCPIP, 200 μM PMS/200 μM DCPIP (ε700 dependent on pH, Fig. S1) or 2.5
mM K3[Fe(CN)6] (ε440 = 583 M-1 cm-1) and 100 M progesterone.
To establish the enzyme stability in buffers pH 6.5, 8.0 and 9.0, the protein was incubated at 4 °C
in the appropriate buffer for 2 h. The relative activity was evaluated as defined previously.
As DCPIP exhibits change in spectrum with the change of pH we determined molar extinction
coefficients in the pH range 4–10 using 0.2 mM DCPIP. The obtained values of ε700 were used in
the activity assays (Fig. S1).
9

2.5.6. Ionic strength
The influence of the ionic strength was examined by measuring the activity in the standard 0.1 M
K2HPO4/KH2PO4 buffer pH 6.5 with 50–700 mM NaCl added. The final ionic strength of the
prepared buffers was in the range of 168 to 877 mM. The activity was measured in triplicate.
2.5.7. Stopped-flow steady-state spectrophotometric assay
Apparent Km and Vmax values for steroids 1-6 (Fig. 4) were determined using a SX20 stopped-
flow apparatus (Applied Photophysics). The data were collected and processed using the Pro-
Data software. All the kinetic traces were registered at 700 nm. The temperature in the stopped-
flow experiments was controlled by a Labo Plus (Polyscience) thermostat bath. All experiments
were performed under anaerobic conditions obtained with the application of argon. The reactions
kinetics were followed at pH 6.5 in 0.1 M K2HPO4/KH2PO4 buffer. In a typical experiment
reaction mixture contained: 100 μM DCPIP, varying concentrations of the steroids (10 to 200
μM) in isopropanol (1%) and 6.13 nM of the enzyme. Solutions of steroids in K2HPO4/KH2PO4
buffer at a given concentration were prepared immediately before use. The pH optimum for PMS
was determined using buffers specified in Thermostability assay. The reaction was carried out
with 200 μM PMS (ε400 = 4805 M-1 cm-1) and 100 μM progesterone. Obtained data were analyzed
with the application of OriginPro 2019 software.
2.6. Reactor tests
Optimization of the reactor conditions for progesterone was carried out in 12 reactions in
duplicates with 2, 4 or 8% of 2-hydroxypropyl-β-cyclodextrin (HBC) (w/v) and 1, 2, 5 or 10% of
2-metroxyethanol (EGME) (v/v). Reaction mixtures (2 mL) consisted of 0.1 M K2HPO4/KH2PO4
10

buffer pH 6.7, 0.6 mM progesterone, 12.5 mM K3[Fe(CN)6] and 0.2 μM AcmB. Reactions were
carried out in a thermoblock at 30 oC and 750 rpm for 80 min.
The selection of the optimal reoxidant was carried out in 2 mL volume reactions in triplicates that
consisted of 0.1 M K2HPO4/KH2PO4 buffer or glycine-NaOH buffer with the optimal pH, 0.2
mM progesterone in EGME (4%), 0.8 mM artificial electron acceptor and 2.8 nM AcmB. The
reaction mixtures were incubated at 30°C, in the thermoblock at 800 rpm for 1 h. The tests were
conducted under either aerobic or anaerobic atmosphere.
The study of oxygen influence on the reaction was carried out in 10 ml volume reactors in
duplicates that consisted of 0.1 M K2HPO4/KH2PO4 buffer pH 6.5, 0.56 mM of DCPIP, 4%
HBC, 2% EGME, 0.12 μM AcmB and excess of 2 mM progesterone. The reactors were run
under aerobic conditions, under anaerobic conditions (98:2 N2:H2 v/v) and the continuous flash of
the air or N2 (gas flow rate: 24 mL/min). The reaction proceeded at 30 °C for 24 h.
The scaled-up synthesis was performed in 10 mL reactors in duplicates containing 0.1 M
K2HPO4/KH2PO4 pH 8.0, 5% HBC, 4.6–6.6 mM steroid (1, 3-8) in EGME (2%), 15 mM PMS
and 0.81 μM AcmB. The reaction proceeded at 30 °C, 250 rpm for 2 h.
The synthesis of (25R)-spirost-1,4-dien-3-one (10a) was performed in the fed-batch reactor
system. The reactor (80 mL) contained 0.1 M K2HPO4/KH2PO4 buffer pH 8.0, 3.2% (w/v) HBC,
0.6 mM diosgenon, 2% EGME (v/v), 4 mM PMS and 0.81 μM AcmB. The reaction proceeded at
30 °C, 250 rpm for 15 min, after which 1.6 mL of 30 mM diosgenon in EGME was added to the
reaction mixture (the final concentration: 1.18 mM diosgenon, 3.9% EGME). The reaction was
carried out in the same conditions for the next 4 h 15 min.
To determine the conversion of a substrate in the reactor tests, the samples of the reaction mixture
were diluted in acetonitrile (1:10), centrifuged at 14 000 g for 5 min, and analyzed on HPLC with
DAD. The analyses were carried out on Agilent 1100 VL LC/MSD using Advanced Materials
11

Technology HALO 90 Å RP-Amide column (2.7 μm, 2.1×75 mm) in acetonitrile/water gradient
method (Table S1).
2.7. Product isolation, purification and NMR analysis
All reaction mixtures were extracted with CHCl3 (3×300 ml), dried (MgSO4) and concentrated in
vacuo. Products were separated using preparative TLC plates (Silica Gel GF, 20×20 cm, 500 μm,
Analtech) and a hexane/acetone mixture (3:1, v/v) as an eluent. The structure of reaction products
was determined using 1HNMR, 13CNMR and correlation spectroscopy.
(25R)-spirosta-1,4-dien-3-on (10a) was purified by solid-phase extraction using C18 PolarPlus®
column (J.M. Baker®). After column activation, the reaction mixture was applied onto the
column and afterward, the column was washed with 3 column volumes of 40% IPA to avoid
PMS contamination. The product was eluted with 4 column volumes of pure IPA.
3. Results
3.1. Enzyme isolation and characterization
The gene encoding AcmB was expressed in E. coli BL21(DE3)Magic. The isolation
procedure yielded the protein with a specific activity of 8.9 ± 0.8 U/mg (29% yield and 16.4-fold
purification) (Table 1). Notably, a significant increase in total activity was observed upon
treatment of the crude cell extract with Triton X-100 and FAD, most probably due to enzyme de-
aggregation and FAD reconstitution. The loss of enzyme activity during affinity chromatography
can be most probably attributed to the loss of FAD, as the final FAD content was around 30%.
The molecular mass of 62.2 kDa was confirmed by SDS-PAGE (Fig. S2) while a bright yellow
color confirmed the presence of the FAD cofactor.
12

In order to establish optimal reaction conditions for the study of AcmB activity following
parameters were tested in a standard spectrophotometric assay: i) temperature, ii) ionic strength,
and iii) pH of the buffer system.
The initial activity of the enzyme was assayed in the range of 20–60 °C and the highest
activity was observed at 50 °C (Fig. 1A). Further increase of temperature above 50 °C resulted in
a dramatic loss of AcmB activity indicating a thermal inactivation of the enzyme. The prolonged
stability of the enzyme was determined for AcmB incubated at 4, 30 or 50 °C (Fig. 1B). the
enzyme turned out to be stable at 4 °C and 30 °C, while at 50 °C a loss of 70% of activity was
observed after 30 min and almost complete deactivation after 2 hours of the experiment. Based
on the obtained results, 30°C was chosen as optimal operational conditions.
The influence of ionic strength was estimated by the addition of the sodium chloride in
concentrations between 0 to 0.8 M to 0.1 M K2HPO4/KH2PO4 buffer. We observed a gradual
decrease of AcmB activity with increased NaCl content, i.e., approximately 30% decrease of
AcmB activity when the concentration of NaCl reached 300 mM or more (Fig. S3). This result
suggested that a relatively poor affinity of K3[Fe(CN)6] toward the enzyme (see 3.2.1) cannot be
compensated with its high concentration without decreasing the enzyme activity.
The influence of pH on the enzyme stability was assayed using the ThermoFAD approach
[21], which yielded values of melting temperatures (Tm) for the enzyme at pH 4–10 (Fig. S5).
The obtained Tm values demonstrated the stability of the enzyme at the pH range of 6.0 to 9.0
(Fig. 1C). This result was supplemented by a standard spectrophotometric assay for the enzyme
(activity assayed at pH 6.5), which was incubated for 2 h at 4 oC at the studied pH. Although
there were slight differences in the specific activities of enzyme samples incubated at different
13

pH values, we did not observe a significant decrease in activity during 2 h of experiment (Fig.
1D). These results showed that AcmB can be catalytically active in a wide 6.5–9 range of pH
without a negative impact on its overall performance. This result was of special importance, as
previously reported AcmB pH optimum of 6.5 [5] differed from other KstDs with pH optima in
the range of 7 [13] to 10 [22].
3.2. Optimization of steroid 1,2-dehydrogenation process
Based on optimized assay conditions concerning enzyme stability and activity, we have
proceeded to establish the best setup for efficient 1,2-dehydrogenation using progesterone as a
representative substrate. This process encompassed a study of: i) different electron acceptors, ii)
the influence of O2, iii) the effect of organic solvent together with steroid solubilizer, and iv) the
form of catalyst.
3.2.1. 3-ketosteroid dehydrogenase electron acceptors
The oxidation of ketosteroids by KstDs results in a reduction of the FAD cofactor. To
close the catalytic cycle the KstDs have to be re-oxidized by an external oxidant. In most studies
2,6-dichloroindophenol (DCPIP), phenazine methosulfate (PMS) or PMS with DCPIP are used
for such purposes [4,12,23]. As KstDs conduct catalytic reactions according to the bi-bi ping-
pong mechanism, the oxidant has to bind to the same active site as the substrate to access the
FADH2 cofactor. As a result, the type of acceptor used, its binding affinity and rate of FADH2 re-
oxidation have a direct influence on the observed rate of the catalytic reaction. In some studies
molecular oxygen was used as FAD re-oxidant [22,24]. This, however, generates H2O2 in enzyme
active site which may result in KstD inactivation [13].
14

We conducted experiments with i) 0.2 mM DCPIP (E’o = +317 mV for reduction to
semiquinone AH2 form and Eo = +230 mV for reduction from semiquinone to the reduced AH3
form, both at pH 7.0 [25]), ii) 0.2 mM PMS (E’o = +155 mV, +126 mV and +100 mV at pH 7, 8
and 9, respectively [25], two-electron reduction from A+ to AH form), iii) 2.5 mM potassium
hexacyanoferrate (III) (E’o = +426 mV at pH 7.8 in 0.1 M phosphate buffer [26], one-electron
reduction) or iv) a mixture of 0.2 mM DCPIP and 0.2 mM PMS. All these reoxidants were tested
at a broad pH spectrum, from 4 to 10 (Fig. 2). Experiments with DCPIP, K3[Fe(CN)6] and
DCPIP/PMS mixture were conducted using a spectrophotometric assay under aerobic conditions
while experiments with PMS were conducted using a stopped-flow approach under anaerobic
conditions. The stopped-flow technique was used in the case of PMS to limit the influence of its
photolysis on the determination of the initial activity [27]. In experiments with K3[Fe(CN)6], due
to the low activity of the AcmB with this acceptor, the amount of enzyme had to be increased 10-
fold from that used in experiments with DCPIP or PMS. Unexpectedly, the results indicated that
depending on the electron acceptor, the pH optimum differs between 6.5 for DCPIP, 8.0 for PMS
and K3[Fe(CN)6] and 9.0 for the DCPIP/PMS mixture (Fig. 2 A and B). For the last case, two pH
optima could be detected, one at pH 6.5, which is characteristic for DCPIP and the other, at pH 9,
closer to the pH optimum observed for PMS. Although DCPIP, K3[Fe(CN)6] and PMS were
tested before in the reaction with AcmB, all experiments were conducted at pH 6.5, which was
then considered as optimal conditions [5]. Our results indicate that the pH of the reaction mixture
should be dictated by the selected electron acceptor. Also our preliminary stopped-flow kinetic
studies of reoxidation rate indicate, that reoxidation may be controlling the observed steady-state
kinetics (data not shown).
15

To compare the robustness of each electron acceptor under its optimal conditions we have
determined the apparent kinetic parameters for DCPIP, K3[Fe(CN)6] and DCPIP/PMS mixture.
We were unable to study in detail the reaction kinetics for PMS due to its light-triggered
photoreaction, as an aqueous solution of PMS exposed to light is photolyzed and oxidized to
blue-colored pyocyanin and other yellow- and orange-colored side products [27].
The highest apparent turnover number of 35.9 s-1 was observed for DCPIP/PMS mixture
at pH 9.0 with app. Km of 0.25 mM followed by DCPIP at pH 6.5 with app. kcat of 12.2 s-1 and
app. Km of 0.12 mM, while K3[Fe(CN)6] exhibited the highest app. Km and lowest kcat. (Table 2,
Fig. S6). Interestingly, for the DCPIP/PMS mixture we observed enzyme inhibition by PMS with
inhibitor constant Ki of 2.2 mM (Fig. S6). Further evaluation of all four types of AcmB
reoxidants was prepared in batch reactions under aerobic atmosphere measuring the progesterone
conversion rate using HPLC as a detection method. This approach was designed to validate if the
observed high pH optimum for PMS and PMS/DCPIP is not caused by previously reported
instability of PMS at higher pH and non-enzymatic reduction of DCPIP [28]. The experiment
showed that the highest substrate to product conversion in time was observed with PMS (>99%
after 30 min), followed by PMS/DCPIP mixture (60% after 45 min), DCPIP (40% after 60 min)
and K3[Fe(CN)6] (0.1% after 21 h), respectively (Fig. 2 C). Similar tendency was observed in the
experiment conducted under anaerobic atmosphre (Fig 10S), although relative initial activity of
the enzyme in the reactor with DCPIP was approximatelly 25-30% lower with respect to the
reactor with PMS.
16

3.2.2. The oxygen influence on the reaction
It has been demonstrated previously that reduced FAD in KstD can be re-oxidized by molecular
oxygen producing H2O2 [13,22,24]. Although AcmB in the oxidized state remains stable during
the incubation under the anaerobic atmosphere (Fig. S9), during the reaction the reduced FAD,
even in the presence of other reoxidant, can activate molecular oxygen, which is dissolved in the
reaction media. This may result in the production of H2O2, which in turn may have a detrimental
effect on the enzyme activity. However, we also observed that the presence of O2 enables the
synthesis of higher yields of a product then would be allowed by the stoichiometry of the added
reoxidant. It appears that H2O2 formed in the system reacts with DCPIPH2 reoxidizing it to
DCPIP. We have confirmed in the independent tests that O2 without enzyme is not able to oxidize
DCPIPH2. Meanwhile, DCPIPH2 undergoes oxidation if treated with H2O2 even without the
presence of the enzyme. Therefore, O2 could be potentially used to regenerate enzyme by direct
oxidation of FADred and a non-enzymatic reaction of H2O2 with reduced DCPIPH2.
To investigate the influence of O2 on the 1,2-dehydrogenation we run four types of reactors with
0.5 mM DCPIP in pH 6.5 and excess of progesterone: i) under anaerobic conditions (98:2 N2:H2
v/v), ii) in reactor flushed with a constant flow of N2, iii) under aerobic conditions and finally, iv)
in reactor flushed with a constant flow of air. The second type of reactor was used as a control to
test for potential protein denaturation, caused by protein foaming in reactor ii (Fig. 3). Under
anaerobic conditions (i), the reaction almost stops after 60 min due to depletion of DCPIP (98%
conversion of DCPIP, 0.55 mM of 1-dehydroprogesterone) and after 24 h the 100% of DCPIP
conversion was reached (28.4% conversion of 1-dehydroprogesterone corresponding to 0.56
mM). Running reaction flushed with a constant flow of N2 (ii) resulted in slightly lower
performance and after 24 h the final conversion of DCPIP of 92% and product concentration of
0.52 mM, which indicates slight enzyme deactivation by flushing gas. Under aerobic conditions
17

(iii), the reaction proceeded slower in comparison to the anaerobic reactor (i) but upon almost
complete depletion of DCPIP at 60 min (86% of DCPIP conversion) the reaction progressed
steadily at a constant rate to reach 0.66 mM of 1-dehydroprogesterone concentration after 24 h ,
which corresponded to 118% of initial DCPIP concentration. Finally, the reactor flushed with air
(iv) performed in the most efficient way reaching 0.62 mM of the product at 60 min of reaction
(110% of the initial DCPIP concentration) and almost 1 mM of the product after 24 h (171% of
initial DCPIP concentration). Also in the latter case (iv) after 60 min the reaction rate was almost
constant but faster than the reference reaction under the aerobic conditions (iii) by approx. 20%.
These experiments demonstrated that the presence of O2 in the reaction system indeed allows
obtaining product yields significantly exceeding the concentration of the reoxidant and potential
enzyme inactivation by H2O2 is outweighed by an increase of the reaction rate due to reoxidation
of enzyme and reoxidant.
3.2.3. The effect of organic solvent together with steroid solubilizer
The observed reaction rate and an overall reaction yield strongly depend on the
availability of hydrophobic steroids in the enzyme-friendly water environment [29]. A high
concentration of the substrate was ensured by the presence of solubilizer, 2-hydroxypropyl-β-
cyclodextrin (HBC), and addition of organic co-solvent, 2-methoxyethanol (EGME) [30]. The
optimal content of HBC and EGME to solubilize C17–C20 3-ketosteroids was determined for
progesterone (Fig. 4, 4), a representative AcmB substrate, in the set of reactor tests with different
content of HBC and EGME and constant concentration of progesterone (63.8 μM, 0.2 g/L). The
highest enzyme activity was detected for 4–6% (w/v) HBC and 1–2% (v/v) EGME, while the
lowest enzyme activity was observed for 10% EGME and reactors with 2 or 8% HBC (Fig. S8A).
However, based on the conversion rate after 20 h (α20h), 99% conversion was achieved for
18

reactors with lower HBC concentration i.e., 1–2% (Fig. S8B). Therefore, assuming a longer time
of synthesis, low HBC concentration is also eligible for the total substrate conversion. Similar
experiments were conducted with a substrate concentration of 63.8, 127.6 and 319 μM (0.2, 0.4
and 1 g/L, respectively). The highest enzyme activity was observed for 3–5% HBC (w/v), 1–2%
(v/v) EGME and initial progesterone concentration in the range of 0.2–0.4 g/L (Fig. S8C-D).
Thus, 5% HBC and 2% EGME were chosen as optimal conditions for batch reactions with
progesterone and its derivatives.
3.2.4. The form of the catalyst in a 1,2-dehydrogenation reaction
We studied the influence of the type of AcmB catalyst on the final productivity of the
dehydrogenation process. In the batch system, we compared the catalyst with different degrees of
purity, i.e., resting cells suspended in a purification buffer, the cell-free extract (crude cell extract
after centrifugation, see 3.1) and enzyme after one-step affinity purification (Fig. 2D) using the
same amount and batch of the cells the catalyst originated from as a reference. The experiments
showed that whole-cell biocatalyst can be efficiently used in the dehydrogenation of 3-
ketosteroids. However, it is possible to obtain an even more efficient catalyst if the crude cell
extract (i.e., the closest approximation of enzyme activity inside the E. coli cells) is treated with
detergent and free FAD. We observe approximately a 2-fold increase of the enzymatic activity of
such crude cell extract after overnight Triton X-100 treatment in comparison with the crude cell
extract without treatment (Table 1). Apparently, the addition of the FAD with Triton X-100
results with FAD reconstitution of the enzyme which increases its activity. As a result, in the
reactor with the FAD-reconstituted cell-free extract catalyst we were able to convert 99% of
substrate in 20 min while in the reactor with whole cells system 93% of conversion was observed
after 2 h. Due to the loss of enzyme activity during purification in the case of the isolated enzyme
only 58% conversion was reached after 20 min, and 87% after 2 h.
19

3.3. AcmB substrate spectrum and reaction upscaling
Up to now, thirteen compounds were tested as AcmB substrates, six of them exhibited
activity as substrates and their kinetic parameters were determined [5]. We decided to expand the
library of AcmB substrates by testing C20–C22 testosterone derivatives (Fig. 4, 3–6, 13), C27
cholesterol derivatives (7–9), as well as C30 steroid derivatives with additional substituents or
ring systems (10–12) or C11, reduced (14, 15) steroid and steroid-like compounds (Fig. 4). C30
steroids derivatives, dehydroepiandrosterone acetate (13) and methyl-1(9)-octal-2-ones (14, 15)
were inactive, while eight of tested compounds turned out to be AcmB substrates. The identity of
the products was confirmed by LC/MS, which enabled detection of [M+H]+ or [M+K]+ signals
consistent with dehydrogenation of the substrate (Table S2) as well as 1HNMR and 13CNMR
spectra (see “Spectral data of isolated products” and Fig. S11–S28 of the Supplementary data).
To investigate AcmB substrate specificity, C20–C22 testosterone derivatives and
cholesterol derivatives were characterized by a steady-state kinetic approach yielding apparent
kinetic parameters (Table 3, Fig. S7). Measurements were performed using a stopped-flow
technique for the first group of steroids and a standard spectrophotometer for the latter. The
highest AcmB affinity to the substrate was observed for progesterone (4) (Km = 3.1 ± 0.2 µM).
The highest catalytic rate constant (kcat) was observed for a native AcmB substrate, androst-4-en-
3,17-dione (AD) (1) (kcat = 144.8 ± 3.9 s-1). The highest catalytic efficiency was observed for
progesterone (4) (kcat/Km= 1.4∙107 s-1M-1), while the catalytic efficiency of other C20-C22
ketosteroids was of the same order of magnitude. Surprisingly, the Km value of androstenedione
(1) (59.6 µM) turned out to be the highest among the studied C20–C22 ketosteroids.
Additionally, the apparent kinetic parameters of the cholest-4-en-3-one (7) and cholest-4,6-dien-
20

3-one (8) indicate lower enzyme affinity toward both substrates compared to progesterone (4) and
significantly lower kcat values comparable to ketosteroids with degraded aliphatic side chain (i.e.
1–6). It should be stressed here, that the Km values of these two compounds were obtained in the
presence of HBC, which influences the thermodynamics of enzyme-substrate formation resulting
in an apparent shift of Km toward higher values. Furthermore, due to limitations in the solubility
of 7 and 8, it was not possible to properly characterize the enzyme saturation region of the
Michaelis-Menten curve, which influenced the final values of apparent kcat (Fig. S7).
As a part of this study, we also tested the enzyme potential in the synthesis of the 1-
dehydrogenated 3-ketosteroids. The batch reactor tests in a 10 ml scale were run aerobically with
substrates concentrations in the range of 4.5–6 mM except for diosgenone (10), which had a
concentration of 1.2 mM, due to the low solubility. After two hours of reaction under optimal
conditions, the reactions were stopped and the conversions evaluated by HPLC (see Table S3).
Substrates (1-6) reached >99% conversion and the isolation yields were in the range of 60–86%.
The introduction of the aliphatic side chain to the steroid core at C17, like in the substrate (7) or
(8), resulted in a conversion decrease to 70–78% and isolation yield in the range of 50–59%. The
lower conversion of (7) and (8) compared to (1-6) is consistent with the results of steady-state
kinetics. Finally, for diosgenone (10) we observed 87% conversion at 1.2 mM concentration,
which could be further increased to 100% by addition of a fresh amount of enzyme after the
reaction reached apparent equilibrium (data not shown). The isolation yield of the product, (25R)-
spirosta-1,4-dien-3-on (10a), was 53%. It is the first report on the enzymatic dehydrogenation of
this saponin derivate, which was reported to exhibit biocidal, anti-inflammatory and anticancer
properties [31,32].
21

4. Discussion
The detailed catalytic characterization of AcmB showed that it can convert a wide range
of 3-ketosteroid compounds, ranging from usual for KstD C20–C22 3-ketosteroids through
substrates with C17 aliphatic side chain such as cholest-4-en-3-one derivatives, to previously
never reported saponin derivatives such as (25R)-spirost-4-en-3-one (diosgenone). Wide substrate
spectrum combined with the availability of the overexpression system, cheap reoxidation system,
enzyme stability in the wide range of pH, and possibility of applying whole-cells or crude
enzyme as a biocatalyst make the system under study very attractive for industrial application.
Our detailed kinetic studies revealed, that the enzyme exhibits various pH-optima
depending on the type of reoxidation agent used in the reaction. As the measured pH optima are
activity-dependent, this result indicates that the reoxidation of FAD may be the reaction limiting
step controlling the observed kinetics, as our preliminary pre-steady-state stopped-flow tests
suggest that FAD reoxidation may indeed be rate-limiting under the studied conditions. The fact
that the enzyme can efficiently catalyze dehydrogenation under mildly acidic conditions seems to
be a unique feature of AcmB and has intriguing mechanistic implications. The 1,2-
dehydrogenation is proposed to be initiated by substrate deprotonation by tyrosyl anion, the
formation of which was previously associated with neutral or basic pH optima of known KstDs
[13,15]. Our results clearly show that AcmB can catalyse dehydrogenation both at basic pH such
as 8 and 9 (i.e., standard for KstDs) and slightly acidic such as 6.5. This interesting fact is
currently under further investigation.
The artificial electron acceptors used in this study differ in structure and chemistry. As a
consequence, they differ in the affinity to the active site of the enzyme, as demonstrated by
22

different values of Km, and in their redox potential, which plays an important role in the
efficiency of FAD reoxidation. The interesting fact of two pH optima, one seemingly associated
with enzyme reoxidation by DCPIP (pH opt = 6.5) and the other with reoxidation by PMS (pH
opt = 8–9) may be associated with electrostatic interactions between reoxidants and FAD
cofactor. According to [33] majority of the reduced flavohydroquinone (FADred) at pH 6.5 is in a
-
charge-neutral FADredH3 form (pKa = 6.7) and in the anionic FADredH2 form at pH 8–9.
Meanwhile, PMS is positively charged in the pH range of 0–11 and DCPIP is a charge-neutral
state between pH 6 and 7.5 (pKa1 = 3.95, pKa2 = 9.4) [25]. Therefore, it can be speculated, that
PMS may be preferentially bound to the reduced active site that hosts negatively charged flavin
-
cofactor (FADredH2 ), while DCPIP is preferentially bound when FADred is in the charge-neutral
state (FADredH3). Spectrophotometric measurements indicate that in the mixture of PMS/DCPIP
the DCPIP is a final electron acceptor regardless of the pH reaction in both acidic and basic
conditions. This result is consistent with previous reports and with a relative redox potential of
both reoxidation agents (e.g., [34]). However, the steady-state kinetic measurements at pH 6.5
showed the same relative activity for reaction either with PMS or DCPIP (Fig. 2 A and B).
Therefore, at pH 6.5 for the reaction with a mixture of PMS/DCPIP we may observe a
competition of these two oxidants for the oxidation of the reduced flavin (FADredH3). The same
situation can still take place at pH 8–9 but with PMS being preferential reoxidant due to
electrostatic attraction with FAD. In the whole studied pH range, we expect a non-enzymatic
reaction of PMSH2 with DCPIP in the bulk. This hypothesis is to some extent consistent with the
overall poor performance of negatively charged [Fe(CN)6]3- which, despite the highest redox
potential, is the worst reoxidant of AcmB. It is not clear why the optimum of the reaction with
ferrocyanide occurs at pH 8.0. The PMS seems to be reduced in the two-electron process as we
have not observed accumulation of the characteristic green semiquinone form of PMS [34] during
23

reactor tests, nor we were able to detect it during HPLC analysis (only PMSH2 or PMS form).
Furthermore, we have not observed the emergence of the semiquinone form of the flavin during
its reoxidation by DCPIP (preliminary pre-steady state kinetics, data not shown), which suggests
that AcmB active site does not stabilize semiquinoid form of FAD, even though it can be
reoxidized in two one-electron processes by K3[Fe(CN)6]. As the observed reaction rates are
obtained for the steady-state conditions it is now apparent that to determine a true pH optimum of
1,2-dehydrogenation process one has to measure kobs of the half-reduction cycle in the function of
pH using the stopped-flow technique. Such studies are currently underway.
We have also demonstrated that the synthesis of 1-dehydrosteroids can be run under
aerobic conditions. We confirmed that the presence of O2 enables reoxidation of the enzyme
(preliminary stopped-flow tests, data not shown) and results with reoxidation of DCPIPH2. As a
result, one can almost double the theoretical yield of the reaction with respect to available
reoxidant, especially when the system is flushed with air (thus eliminating O2 diffusion
limitation). Although the detrimental influence of H2O2 on AcmB activity seems to be present, it
is outweighed by a higher yield of the reaction. We intend to further explore the influence of
H2O2 on AcmB activity in future studies.
The steady-state kinetics shows that androst-4-en-3,17-dione (1) is converted with the
highest apparent rate but it is a progesterone (4) which, under experimental conditions, is
characterized by the highest affinity to the enzyme with Km as low as 3.1 µM. We also have not
observed significant influence of small modification of the substituent at C17 in the case of 4–6
(Km in the range of 3–7 µM).
The extension to the undegraded isooctyl substituent of cholest-4-en-3-one (7) derivate
results with a 10-fold increase of the apparent Km. On the other hand, the kinetics of these
24

compounds had to be collected in the presence of HBC, which shifts apparent Km toward higher
values due to the substrate sequestration effect. The non-linear influence of cyclodextrins on
apparent kinetic parameters was previously reported in the case of soybean lipoxygenase and
tyrosinase from Streptomyces antibioticus [35,36]. However, usually in case of dehydrogenation
of steroids, the addition of hydrophobic cyclodextrins (i.e., modified with alkyl substituents) has
a beneficial effect on the biotransformation yields as their low solubility in water is the main
limiting factor [37–39]. As a result, it is very difficult to compare kinetic parameters obtained for
these two classes of substrates without detailed knowledge on the equilibrium constants of
substrate-HBC complex formation.
Despite slightly different conditions of the kinetic assays, our results are generally in line
with previous reports on AcmB kinetics [5], where progesterone (4) was also the best substrate
with the lowest Km while androstenedione’s (1) was approx. 20–25 times higher. Surprisingly, in
[5] the kinetic parameters for cholest-4-en-3-one (7) or cholest-5-en-3-one were established
without the help of HBC using 1–500 µM substrate concentration range. In our tests, we were
unable to achieve such high concentration without the addition of HBC, which in turn influences
the observed kinetics. However, in studies presented by Chiang et al. [5], such tests might have
been possible due to the elevated temperature of the assays (37 °C), although the presence of
HBC in assays was reported in the follow-up study [40]. In our case, the kinetic parameters of 7
and 8 were much more consistent (higher Km values in the range of 55–100 M, lower kcat values
in the range of 9–13 s-1) than those reported by Chiang et al. for 7 (Km 42 M, kcat 69 s-1) and
cholest-5-en-3-one (Km 9 M, kcat 14 s-1). We intend to further explore this issue and determine
the equilibrium constants of the HBC-steroid complex formation in order to get better insight into
the kinetic behaviour of HBC-3-ketosteroid-KStD system.
25

5. Summary
We demonstrated that AcmB can be efficiently applied as a biocatalyst in 1-
dehydrogenation of a wide range of 3-ketosteroids, including non-standard compounds with
isooctyl chain or even additional spiro-ring characteristic for saponins. The enzyme is stable in a
wide range of pH and, depending on the reoxidant used, exhibits multiple pH optima and can
utilize O2 as a supplementary reoxidant. The presence of pH optimum at acidic pH to some extent
challenges the currently accepted mechanistic hypothesis of KstD and is an interesting topic of
further mechanistic studies.
CRediT author statement
Agnieszka M. Wojtkiewicz: Roles/Writing - original draft;Investigation, Formal analysis, Supervision,
Writing/editing; Patrycja Wójcik: Roles/Writing - original draft; Investigation, Validation, Writing/editing,
Visualization; Magdalena Procner, Investigation, Formal analysis; Monika Flejszar: Investigation; Maria
Oszajca: Resources, Supervision, Writing/review; Mateusz Hochołowski: Investigation; Mateusz Tataruch:
Investigation; Beata Mrugała: Investigation, Methodology, Formal analysis; Tomasz Janeczko: Supervision,
Investigation, Resources, Writing/review ; Maciej Szaleniec: Conceptualization, Supervision, Visualization, Project
administration, Writing - original draft, Writing/editing, Funding acquisition.
Conflict of interest
The authors declare no conflict of interest.
Acknowledgments
The authors acknowledge partial financial support from the National Center of Science
under the OPUS grant number UMO-2016/21/B/ST4/03798. P.W. acknowledges the fellowship
with the project no. POWR.03.02.00-00-I013/16. M. T. acknowledges ICSC PAS Statutory Fund.
26

Authors acknowledge Dr. Magdalena Malinowska from Cracow University of Technology and
Tomasz Więcaszek from Natchem S.C for lupeol samples and professor Jacek Morzycki from
University Białystok for diosgenone.
27

References
[1] L.H. Bergstrand, E. Cardenas, J. Holert, J.D. van Hamme, W.W. Mohn, Delineation of
steroid-degrading microorganisms through comparative genomic analysis, MBio. 7 (2016)
1–14. https://doi.org/10.1128/mBio.00166-16.
[2] A. Rohman, B.W. Dijkstra, The role and mechanism of microbial 3-ketosteroid Delta(1)-
dehydrogenases in steroid breakdown., J. Steroid Biochem. Mol. Biol. 191 (2019) 105366.
https://doi.org/10.1016/j.jsbmb.2019.04.015.
[3] J. Knol, K. Bodewits, G.I. Hessels, L. Dijkhuizen, R. Van Der Geize, 3-Keto-5α-steroid
Δ1-dehydrogenase from Rhodococcus erythropolis SQ1 and its orthologue in
Mycobacterium tuberculosis H37Rv are highly specific enzymes that function in
cholesterol catabolism, Biochem. J. 410 (2008) 339–346.
https://doi.org/10.1042/BJ20071130.
[4] N. Van Oosterwijk, J. Knol, L. Dijkhuizen, R. Van Der Geize, B.W. Dijkstra, Structure
and catalytic mechanism of 3-ketosteroid-Δ4-(5α)- dehydrogenase from Rhodococcus
jostii RHA1 genome, J. Biol. Chem. 287 (2012) 30975–30983.
https://doi.org/10.1074/jbc.M112.374306.
[5] Y.R. Chiang, W. Ismail, S. Gallien, D. Heintz, A. Van Dorsselaer, G. Fuchs, Cholest-4-en-
3-one-Δ1-dehydrogenase, a flavoprotein catalyzing the second step in anoxic cholesterol
metabolism, Appl. Environ. Microbiol. 74 (2008) 107–113.
https://doi.org/10.1128/AEM.01968-07.
[6] P.H. Wang, T.H. Lee, W. Ismail, C.Y. Tsai, C.W. Lin, Y.W. Tsai, Y.R. Chiang, An
Oxygenase-Independent Cholesterol Catabolic Pathway Operates under Oxic Conditions,
PLoS One. 8 (2013). https://doi.org/10.1371/journal.pone.0066675.
[7] W.J. Marsheck, S. Kraychy, R.D. Muir, Microbial degradation of sterols., Appl. Microbiol.
23 (1972) 72–77.
[8] B. Galan, I. Uhia, E. Garcia-Fernandez, I. Martinez, E. Bahillo, J.L. de la Fuente, J.L.
Barredo, L. Fernandez-Cabezon, J.L. Garcia, Mycobacterium smegmatis is a suitable cell
factory for the production of steroidic synthons, Microb Biotechnol. 10 (2017) 138–150.
https://doi.org/10.1111/1751-7915.12429.
[9] L. Fernández-Cabezón, B. Galán, J.L. García, Unravelling a new catabolic pathway of C-
19 steroids in Mycobacterium smegmatis, Environ. Microbiol. 20 (2018) 1815–1827.
https://doi.org/10.1111/1462-2920.14114.
[10] M. V Donova, Transformation of steroids by actinobacteria: A review, Appl. Biochem.
Microbiol. 43 (2007) 1–14. https://doi.org/10.1134/S0003683807010012.
[11] W. Zhang, M. Shao, Z. Rao, M. Xu, X. Zhang, T. Yang, H. Li, Z. Xu, Bioconversion of 4-
androstene-3,17-dione to androst-1,4-diene-3,17-dione by recombinant Bacillus subtilis
expressing ksdd gene encoding 3-ketosteroid-Delta1-dehydrogenase from Mycobacterium
neoaurum JC-12., J. Steroid Biochem. Mol. Biol. 135 (2013) 36–42.
https://doi.org/10.1016/j.jsbmb.2012.12.016.
[12] X. Zhang, D. Wu, T. Yang, M. Xu, Z. Rao, Over-expression of Mycobacterium neoaurum
3-ketosteroid-Δ1-dehydrogenase in Corynebacterium crenatum for efficient bioconversion
of 4-androstene-3,17-dione to androst-1,4-diene-3,17-dione, Electron. J. Biotechnol. 24
28