Modified bile acids and androstanes—Novel promising inhibitors of human cytochrome P450 17A1

Article abstract of DOI:10.1016/j.jsbmb.2020.105777

Cytochromes P450 are key enzymes for steroid hormone biosynthesis in human body. They are considered as targets for the screening of novel high efficient drugs. The results of screening of bile acids and androstane derivatives toward human recombinant steroid 17α-hydroxylase/17,20-lyase (CYP17A1) are presented in this paper. A group of steroids, binding with micromolar or submicromolar affinity (in a range from 9 μM – less than 0.1 μM), was identified. Results presented here showed that these steroidal compounds are able to decrease rate of hydroxylation of essential CYP17A1 substrate – progesterone, while some compounds completely inhibited enzyme activity. Structure-activity relationship (SAR) analysis based on in vitro and in silico studies showed that high affinity of the enzyme to bile acids derivatives is correlated with side chain hydrophobicity and presence of hydroxyl or keto group at C₃ position. From the other side, bile acid-derived compounds with more polar side chain or substituents at C₇ and C₁₂ positions possess higher K_d values. Among androstane-derived steroids couple of Δ⁵-steroids with hydroxyl group at C₃ position, as well as 16,17-secosteroids, were found to be high affinity ligands of this enzyme. The data obtained could be useful for the design of novel highly efficient inhibitors of CYP17A1, since the bile acids-derived compounds are for first time recognized as effective CYP17A1 inhibitors.

Full text of DOI:10.1016/j.jsbmb.2020.105777

Journal Pre-proof
Modified bile acids and androstanes - novel promising inhibitors of
human cytochrome P450 17A1
Yaraslau Dzichenka, Michail Shapira, Aliaksei Yantsevich, Tatsiana
´
´
Cherkesova, Ljubica Grbovic, Marina Savic, Sergey Usanov,
ˇ
´
Suzana Jovanovic-Santa
PII:
S0960-0760(20)30302-2
DOI:
https://doi.org/10.1016/j.jsbmb.2020.105777
SBMB 105777
Reference:
To appear in:
Journal of Steroid Biochemistry and Molecular Biology
Received Date:
Revised Date:
Accepted Date:
11 June 2020
31 October 2020
1 November 2020
´
Please cite this article as: Dzichenka Y, Shapira M, Yantsevich A, Cherkesova T, Grbovic L,
ˇ
´
´
Savic M, Usanov S, Jovanovic-Santa S, Modified bile acids and androstanes - novel
promising inhibitors of human cytochrome P450 17A1, Journal of Steroid Biochemistry and
Molecular Biology (2020), doi: https://doi.org/10.1016/j.jsbmb.2020.105777
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.

Modified bile acids and androstanes - novel promising inhibitors of human
cytochrome P450 17A1
Running title.
Novel promising ligands of human cytochrome P450 17A1.
Yaraslau Dzichenka1, Michail Shapira1, Aliaksei Yantsevich1, Tatsiana Cherkesova1,
Ljubica Grbović2, Marina Savić2, Sergey Usanov1, Suzana Jovanović-Šanta2
1Institute of Bioorganic Chemistry NAS of Belarus, Kuprevicha Street, 5/2 Minsk,
220141, Belarus.
2University of Novi Sad Faculty of Sciences, Department of Chemistry, Biochemistry
and Environmental Protection, Trg Dositeja Obradovića 3, Novi Sad, 21000, Serbia.
*Corresponding author.
Yaraslau Dzichenka, Institute of Bioorganic Chemistry NASB, Kuprevicha Street, 5/2
Minsk, 220141 Belarus, Tel/Fax 375-17-3698273; E-mail: dichenko@iboch.by
Highlights
 Modified bile acids and androstane derivatives binds with high affinity by human CYP17A1.
 Novel ligands decrease rate of hydroxylation of progesterone – natural CYP17A1 substrate.
 Bile acids derivatives can completely inhibit enzyme activity toward progesterone.
 In vitro results and SAR analysis allowed to recognize important pharmacophores.

Abstract.
Cytochromes P450 are key enzymes for steroid hormone biosynthesis in human
body. They are considered as targets for the screening of novel high efficient drugs.
The results of screening of bile acids and androstane derivatives toward human
recombinant steroid 17α-hydroxylase/17,20-lyase (CYP17A1) are presented in this
paper. A group of steroids, binding with micromolar or submicromolar affinity (in a
range from 9 µM – less than 0.1 µM), was identified. Results presented here showed
that these steroidal compounds are able to decrease rate of hydroxylation of essential
CYP17A1 substrate – progesterone, while some compounds completely inhibited
enzyme activity. Structure-activity relationship (SAR) analysis based on in vitro and
in silico studies showed that high affinity of the enzyme to bile acids derivatives is
correlated with side chain hydrophobicity and presence of hydroxyl or keto group at
C3 position. From the other side, bile acid-derived compounds with more polar side
chain or substituents at C7 and C12 positions possess higher Kd values. Among
androstane-derived steroids couple of Δ5-steroids with hydroxyl group at C3 position,
as well as 16,17-secosteroids, were found to be high affinity ligands of this enzyme.
The data obtained could be useful for the design of novel highly efficient inhibitors of
CYP17A1, since the bile acids-derived compounds are for first time recognized as
effective CYP17A1 inhibitors.
Abbreviations.
CYP – cytochrome P450; SAR – structure-activity relationship; RMSD – root-mean-
square deviation; DMSO – dimethylsulfoxide; CYP17A1 – steroid 17α-
hydroxylase/17,20-lyase; IPTG – isopropyl β-D-1-thiogalactopyranoside; PMSF —
phenylmethylsulfonyl fluoride; Em-913 – Emulgene-913; DTT — dithiothreitol;
CHAPS — 3-[(3-Cholamidopropyl)dimethylammonio]-1-propanesulfonate hydrate;
TLC — thin layer chromatography; HRMS — high-resolution mass spectrometry;
GAL – galeterone; P4 – progesterone; P5 – pregnenolone; 17OH-P4 – 17α-hydroxy
progesterone; 17OH-P5 – 17α-hydroxy pregnenolone; PAINS – pan-assay interference
compounds.
Keywords.
Cytochrome P450, androstane compounds, bile acids derivatives, CYP17A1 inhibitors.
1. Introduction
Cytochromes P450 (CYP) present a superfamily of heme-containing enzymes
existing in all life kingdoms and participating in many different physiological
processes. Human CYP enzymes catalyze biotransformation of xenobiotics (including

drugs), biosynthesis and metabolism of cholesterol, steroid hormones, bile acids, fat-
soluble vitamins and polyunsaturated fatty acids [1].
CYP17A1 (steroid 17α-hydroxylase/17,20-lyase or 17,20-desmolase) is an
important enzyme, playing crucial role in steroid hormone biosynthesis. This enzyme
is expressed in all steroidogenic and few non-steroidogenic tissues [2]. CYP17A1
catalyzes two independent reactions in one active site [3]: hydroxylation at C17 position
of natural substrates pregnenolone (P5) or progesterone (P4) and breaking of a bond
between C17 and C20 of their 17-hydroxy analogues. The ratio of these two activities is
physiologically important and may direct steroid hormone biosynthesis to the
production of either corticoid or sex hormones [4] (Figure 1).
Mutations in CYP17A1 gene or overexpression of the protein are associated with
different human endocrine diseases: congenital adrenal hyperplasia (due to 17α-
hydroxylase deficiency form) [5], isolated 17,20-lyase deficiency [6], polycystic ovary
syndrome [7] and Cushing’s syndrome [8]. It was also found that changes in the
enzyme activity, due to point mutations or other reasons, lead to the castration-resistant
prostate cancer, which imposes CYP17A1 as one of the key targets in the treatment of
this disease [9]. Having in mind great medical importance of CYP17A1 inhibition, in
the present study we tested a panel of modified bile acids and androstane derivatives
against human CYP17A1, in order to find novel enzyme ligands – from the one hand,
and to determine novel types of biological activity for a group of compounds which
suppress cancer cells proliferation. Such biology-driven medicinal chemistry of
anticancer drugs is very promising and can make significant progress in the
development of drugs for prostate cancer treatment.
2. Materials and methods.
2.1 Protein expression and purification.
Human CYP17A1 was expressed and purified according to known procedure [10]
with slight modifications. Protein was truncated by removing 27 amino acids from N-
termini and MAKKT sequence was added instead of truncated segment. CYP17A1 was
overexpressed in the E. coli BL-21 cells. Expression was induced by the addition of
0.5 mM isopropyl β-D-1-thiogalactopyranoside (IPTG), and 0.4 mM δ-aminolevulinic
acid was added as a precursor of heme biosynthesis. After induction cells were cultured
at 26 °С during 48 h. Harvested cells were resuspended in Buffer A (50 mM Tris-HCl
buffer, pH 7.4, containing 20% glycerol and 0.3 M NaCl), containing 0.5 mM
phenylmethylsulfonyl fluoride (PMSF). The cells suspension was frozen under –73 °C.
Cells were disrupted by Emulsiflex C3 homogenizer, followed by addition of
Emulgen 913 (Em-913) at final concentration 1%. The lysate was cleared by
centrifugation (100000 g) during 1 h. Supernatant was applied to a Ni2+-NTA-agarose
column, equilibrated with buffer A. The column was washed with 3 volumes of Buffer

Acontaining 0.2% Em-913 and then with 20 volumes of Buffer Acontaining 0.2% Em-
913 and 25 mM imidazole with 50 μM P4 (Buffer B). Protein was eluted from the
column with buffer A containing 250 mM imidazole and 0.2% Em-913. Eluted
fractions of CYP17A1 were applied to a column with hydroxyapatite equilibrated with
10 mM potassium-phosphate buffer, pH 7.4, containing 20% glycerol, 0.2% Em-913,
0.1 mM dithiothreitol (DTT) and 0.3 M NaCl. The column was washed with 10
volumes of 50 mM potassium-phosphate buffer, pH 7.4, containing 20% glycerol, 0.2%
Em-913, 0.1 mM DTT and 0.3 M NaCl. The protein was eluted by 0.6 M potassium-
phosphate buffer, pH 7.4, containing 20% glycerol, 0.2% Em-913, 0.1 mM DTT and
0.3 M NaCl. Protein preparations were kept under –73 °C.
2.2. Modified steroids used in enzyme binding tests.
Synthesis of tested analytically pure steroidal compounds was performed
according to known procedures, while characterization and purity of all substances
were tested and documented by thin layer chromatography (TLC), high-resolution
13
mass spectrometry (HRMS), 1H and C NMR [11-32]. Purity was also checked by
measuring melting points of the synthesized compounds; they were narrow range (1-2
degrees) and in agreement with literature data. Synthetic procedures with short
13
description and raw NMR spectra (1H and C) are presented in Supplemental
Materials. Stock and working solutions (10-2, 10-3, 10-4 M) for titration studies were
prepared in dimethylsulfoxide (DMSO).
2.3. In vitro screening of the potential ligands.
Initial screening was performed using high throughput screening (HTS) approach.
The 96-well plate was filled with compounds solutions in 50 mM potassium-phosphate
buffer (pH 7.4), containing 0.2% 3-[(3-Cholamidopropyl)dimethylammonio]-1-
propanesulfonate hydrate (CHAPS) and 0.3 M NaCl (final ligand concentration was
80 μM). After adding protein solution (final concentration 1 μM) difference spectrum
(350-500 nm) was recorded (protein+ligand vs. protein+DMSO) using SpectraMax i3
spectrophotofluorometer (Molecular Devices, USA). Compounds with typical spectral
response, corresponding to binding of substrate (weak field ligand, Type I spectral
response; λmax=393 nm, λmin=417 nm) or inhibitor (strong field ligand, Type II spectral
response; λmin=413 nm, λmax=432 nm) molecules were detected and picked up for
further experiments.
Affinity of the ligands was analyzed using spectrophotometric titration approach
in 50 mM potassium-phosphate buffer (pH 7.4), containing 0.2% CHAPS and 0.3 M
NaCl with final CYP17A1 concentration 1 µM. Ligand solution (stock solutions with
concentrations from 10-4 up to 10-2 M) was added to the experimental cuvette and equal
volume of the solvent (DMSO) to the control cuvette. To determine dissociation
constant of enzyme-ligand complex (Kd) equation for the tight binding was used
(titration data were approximated using of Levenberg-Marquardt algorithm):

[L]t [R]0  Kd  ([L]t [R]0  Kd )2 4[R]0[L]t
2[R]0
A  Amax 
where A – amplitude of the spectral change at [L] ligand concentration; Amax
–
amplitude of the spectral change at saturation concentration of the ligand; [L]t –
concentration of the ligand at saturation; [R]0 – total protein concentration.
2.4. Analysis of protein activity in reconstituted system.
Enzyme activity was reconstituted at 37°C in 25 mM HEPES buffer (pH 7.4),
containing 10 mM MgCl2, 0.1% CHAPS, 0.5 μM CYP17A1, 1 μM NADPH-
cytochrome P450 reductase (CPR) as a redox partner, 50 μM substrate, 8 mM sodium
isocitrate and 1 U/ml isocitrate dehydrogenase. Reaction was started by addition of
NADPH (final concentration 0.25 mM). Concentration of P4 and inhibitors in
competition test was 50 μM.
Aliquots were taken after 0, 2, 5, 10 and 30 min, and extracted by
dichloromethane. Organic phase was evaporated and dissolved in MeOH for HPLC-
MS analysis.
Reaction mixture was analyzed by HPLC-MS, using ZORBAX Extend-C18 (2.1
× 50 mm, 1.8 μm) column (Agilent, USA). Mobile phase A consisted of 0.2% formic
acid solution in water, while mobile phase B consisted of 0.2% formic acid solution in
acetonitrile. The volume of the sample applied to the column was 1.0 μL, and the shape
of the elution gradient was the following: 5% phase B for 2 minutes, 5% to 95% phase
B for 6 minutes and 95% for 7 minutes. The mobile phase flow rate was 250 μL/min,
and column temperature was 40°C. The quadrupole time-of-flight mass analyzer Q-
TOF 6550 (Agilent, USA), used for the detection of the products of enzymatic reaction,
was equipped with an electrospray ionization source (APESI), with the following
parameters: a carrier gas temperature of 400°C, carrier gas flow rate of 12 L/min,
capillary voltage of 4 kV, and fragmentor voltage of 90 V. The mass analyzer was
operated in range 200–1600 m/z.
2.5. In silico screening. Spatial structures of the modified steroids were built
from their 2D structures using MolView service. Before docking structures were
minimized in UCSF Chimera software (2000 steepest descent steps and 1000 conjugate
gradient steps with step size 0.02 Å) [33]. An automated Python-based (v. 3.6) protocol
was used for the in silico screening of CYP17A1 potential ligands. Docking and further
analysis were based on known crystal structures of CYP17A1 with different substrates
or inhibitors from Protein Data Bank. Docking was performed using AutoDock Vina
software (v.1.1.2) [34] with next parameters: 20 diverse positions of the ligand in the
active site of protein; exhaustiveness – 256; energy range – 4 kcal/mol. After docking
all hits were ranked according to their root-mean-square deviation (RMSD) towards

similar known ligands from PDB and according to AutoDock Vina scoring function
values.
3. Results and Discussion.
Purified human recombinant CYP17A1 possesses a typical P450 absorption
spectrum with the heme iron being in a low-spin state, with a Soret region maximum
at 418 nm (corresponding α, β-bands located at 567 and 535 nm, respectively) (Figure
2). Protein molecular mass corresponds to the expected value (Figure 2, insets A, B).
CYP17A1 binds its substrate P4 and inhibitor econazole (Figure 2, inset C). Carbonyl
complex of the protein has characteristic maximum in the absorption spectrum at 450
nm (Figure 2, inset D), indicating that the protein is stable and represents its active
form.
Such well characterized human recombinant CYP17A1 was used for the in vitro
screening of affinity of a series of modified bile acids and androstane derivatives for
binding to the active site of this enzyme. Before screening an in silico analysis proving
that these molecules are not pan-assay interference compounds (PAINS) [35] was
provided using functionality of Python3 RDKit library (version 2019.03.3). According
to our results, only compound L12 (Table 2) does not satisfy PAINS filter.
Among all compounds tested in vitro some synthetic steroids proved themselves
as CYP17A1 active site ligands (Table 1), while other did not bind to the enzyme
according to spectrophotometric titration results (Table 2).
Spectrophotometric titration experiments showed that human CYP17A1 binds
steroids from both classes: bile acids derivatives with long side chain at position 17, as
well as D-seco or D-homo androstane derivatives, with similar affinities in some cases.
Spectral changes for all compounds correspond to Type I spectral response (‘substrate-
like‘ molecules). According to the data presented previously by Auchus et al. [36], this
enzyme has relatively small active site cavity, so it should be difficult to bind
compounds with large differences in structure, in both dimension and substituents.
However, bile acid derivatives, having A/B-ring cis-conformation, bind to the enzyme
with high affinity. This is very important finding because steroidal ligands of
CYP17A1, different from androstane-type compounds, has for the first time been
indicated. Moreover, according to the literature, among known cytochromes P450 only
CYP3A4 is able to bind bile acids in the active site [37]. In other cases bile acids act
on the activity of CYP through regulation of the expression of corresponding genes via
binding to nuclear receptors [38].
In silico screening of binding affinity of the modified steroids toward CYP17A1
showed that interaction energy for all tested compounds (which was calculated using
AutoDock Vina scoring function) is comparable with consequent energy, calculated

for the ‘essential’ substrates of the enzyme (P4, P5, 17OH-P4, 17OH-P5). However,
no correlation between binding energy and the affinity of the ligand was found.
It was found that novel detected ligands can interact with conserved amino acids
A113, F114, N202, IE205, D298, G301, A302, T306, V366, A367, I371 and V483,
forming active site surface. Most of them are taking part in stabilization of natural
substrates (P4, P5, 17OH-P4, 17OH-P5) of CYP17A1. Especially N202 is important,
for which it was found that formation of the hydrogen bond between the C3 substituent
of the steroidal substrate and the amino acid residue is crucial for the first step of
activity of CYP17A1 – hydroxylation at C17 position [39]. Moreover, ‘tightness’ of this
bond, depending on the stereochemistry of the steroid A/B-rings, increases the
possibility of steroid hydroxylation at C16 position due to increased mobility of the
substrate molecule, as it was shown for the progesterone and pregnenolone: keto group
can only serve as hydrogen bond acceptor, but hydroxyl group can serve either
hydrogen bond acceptor or hydrogen bond donor [39].
According to our data, localization of the ligands in CYP17A1 active site is quite
similar to the position of the known substrates (RMSD is lower than 3 Å) except for
compounds 9 and 11, which have bulky substitution groups at C12 and C6, respectively.
In case of compound 11 steroidal plane is turned around the axis C3-C16 so C6-oximino
group of the ligand contacts with the conserved L214 (Figure 3, B). Docking of
compound 9 showed that A/B rings moiety of the steroidal core is located farther from
the I-helix, in comparison with its location in complex with P5 or P4. This bile acid
derivative forms bonds with A105, I205, V236, R239 (Figure 3, A). Relatively high Kd
values for these compounds (9 and 11), confirm that such localization is not optimal
for the ligands of CYP17A1.
Analysis of in vitro screening results for bile acids-derived compounds allowed
us to determine some crucial structural elements for ‘good ligands’.
Compounds with more hydrophobic side chain, including methyl- (compounds
1, 5, 6, 7 and 9) or butyl- (compound 3, Kd=0.17±0.04 uM) ester moiety are binding
with lower Kd values than those with hydrophilic carboxyl group (compounds 2, 4, 8).
According to in silico results, compounds with more hydrophobic side chain oriented
in such a way that these parts of the molecules occupy hydrophobic pocket formed by
V366, A367, I371 and V483 (Figure 3, C). This correlates with the binding mode of a
well-known CYP17A1 inhibitor – galeterone (GAL; PDB ID: 3SWZ) – it is known
that one of the key points of the binding of this compound is interaction of its
benzimidazole moiety with these amino acids [40].
Obtained data also showed that bile acid derivatives with any group (oxo- or
ester-derivatives) at C12 are binding with lower affinity (possess higher Kd values, as
compound 9), compared to the molecules without any substituent (compounds 7, which
is methyl-ester of lithocholic acid, and 6), or do not bind at all (as cholic and

deoxycholic acid derivatives L1, L6, L4 and L8). Analysis of docking results showed
that C12-group locates too close to conserved A302 and T306 from I-helix (Figure 3,
D). This interaction could decrease stability of the complex ‘CYP17A1-steroid
molecule’.
Among bile acid-derived compounds with oxo or hydroxyl group at C3, which
bind with high affinity to human CYP17A1 active site heme iron (compounds 3, 5 and
7), any correlations between nature of substituent at this position and affinity of the
tested molecules could not be found. Our data also showed that bile acid derivatives
with acetyl ester or other substitution group at C3 do not bind to human CYP17A1 (L1,
L4, L6 and L9). According to in silico results, it could be due to inappropriate
interactions of bulky C3-group of steroids with A/B cis-conformations with residues
I198, Y201, L243 and F300, forming pocket over the top of α-helix I and along the
underside of α-helices F and G (Figure 4, B). Another reason is failure to form
hydrogen bond network with conserved N202 and R239, and other polar residues from
α-helices F, G and I, which is necessary for ligand stabilization (Figure 4, A) [40].
Ligands with low Kd values both between bile acid derivatives with planar (Δ4-
steroids) and non-planar (A/B-ring cis-junction) steroidal core were found, but in case
of A/B-ring cis molecules binding affinity was higher (compounds 1, 3 and 7). This
could be because I-helix contains conserved G301 and A302, whose location permits
ligands to adopt right conformation inside the active site (Figure 3, D). Interestingly,
for abiraterone and GAL planar α-face of the molecule is one of the key features that
make them effective inhibitors of CYP17A1 [40].
There were no ligands found for CYP17A1 by in vitro HTS among C-homo
lactones derived from cholic acid (Table 2). It was also found that modified bile acids
with any group at C7 position, namely cholic acid derivatives, were not recognized as
ligands for the enzyme (no spectral response was detected), except compound 8, which
is binding relatively weak. Nevertheless, in the latter case addition of hydroxyl group
dramatically decreases affinity (Kd=108±6 μM for compound 8 comparing to
Kd=26.0±0.7 μM for compound 2). According to docking results this could be due to
unfavorable interactions of α-oriented functional group in this position with conserved
hydrophobic amino acids arising from I-helix (G301, A302).
In vitro analysis of androstane derivatives binding also allowed us to select
compounds that interact with CYP17A1 active site. In general, such androstane-
derived steroids were not so good ligands for the protein as bile acid derivatives: only
two of the tested compounds (12 and 13, Table 1) bind with micromolar or
submicromolar affinity. It was found that Δ5-steroids with hydroxyl group at C3
(compounds 10, 12, 13) interact with the enzyme more tightly than other androstane
derivatives. Addition of long 2-hydroxyethyl ether chain at C3 position significantly
decreases affinity of steroid (Kd =0.33±0.01 μM for compound 13, towards Kd

=42.8±2.9 μM for compound 14). According to docking results this could be due to
unfavorable interaction of C3-substituent group of compound 14 with surface of the
cavity, formed by residues arising from α-helices F and G (Figure 3, B). Similar
situation is for 3,6-dioxime 11. Results also showed that 16,17-secosteroids (12, 13)
are more preferable for binding than D-homo lactones (10, 15, Table 1).
In order to analyze possible products of the enzymatic reaction we performed
reconstruction of the enzyme activity, offering modified steroids (Table 1) as
substrates. According to the mass spectrometry analysis of the reaction mixture
extracts, any hydroxylated derivatives of steroidal compounds were not detected in
reconstituted system after 30 min of incubation. From the fact that some of the tested
molecules (compounds 1 (Kd=0.72±0.09 µM), 3 (Kd=0.17±0.04 µM), 7 (Kd=0.05±0.03
µM) and 13 (Kd=0.33±0.01 μM)) possess unusually high affinity or moderately high
affinity, compared to essential ligands of human CYP17A1 (P5 (Kd=0.86±0.12 µM)),
P4 (Kd=1.92±0.09 µM), 17OH-P4 (Kd=2.32±0.35 µM) and 17OH-P5(Kd=0.63±0.05
µM) [10]), they could be considered as competitive inhibitors of the enzyme, that is
very useful for the creation of novel drugs with high efficiency.
To check this hypothesis, we analyzed relative activity of CYP17A1 toward
progesterone in the presence of these modified steroids. It was found that modified
steroids are able to decrease rate of first phase catalytic reaction of CYP17A1, namely
P4 hydroxylation (Figure 5). The most prominent compounds among them were
compounds 1 and 6 (corresponding Kd values are 0.72±0.09 µM and 5.6±0.4 μM,
respectively): they completely blocked hydroxylation of P4. In the latter case
(compound 6) affinity was lower for modified steroid than for progesterone. From the
other hand, compounds with Kd values at submicromolar range (compounds 3, 5, 7 and
13) demonstrated good inhibiting potential.
Given data shed light on a novel type of biological activity of modified steroids,
some of which were recognized as perspective anticancer agents. Among all novel
ligands of CYP17A1, D-secoandrostane compound 13 stood out in that, since good
anti-proliferative activity of this compound against prostate cancer cells PC-3 line was
stated previously (IC50 was 6.3µM) [21]. Tight binding to human CYP17A1, one of the
key enzymes in biosynthesis of sex hormones and a key target in prostate cancer
treatment, followed with high percentage of the enzyme inhibition, allowed us to
present these novel ligands of the enzyme as lead compounds for the development of
high-efficient therapeutics for the treatment of hormone-responsive diseases,
particularly prostate cancer. This is especially emphasized for compounds expressing
anti-proliferative effect for prostate cancer cells. From the other side, the results
obtained can diminish tested compounds, which can interact with other targets than
CYP17A1, from further usage in therapy, or redirect design of novel drugs, as well as
their possible pharmacological usage.

Author statement
The authors are very grateful for reviewing our manuscript (SBMB-D-20-
00042). We corrected the manuscript according to the comments and suggestions. We
send the revised manuscript with corrections visible in ‘track changes’, while answers
are provided for each comment. We hope that corrected manuscript will suitable for
publication in Journal of Steroid Biochemistry and Molecular Biology.
Author Contributions.
Conceptualization and design: Y.D. and S. J-Š.; Synthesis and characterization of
modified steroids: Lj.G. and M.S.; Protein expression and purification: T.S.; In vitro
experiments and analysis of data: Y.D., M.Sh. and S. J-Š.; In silico experiments and
analysis of data: Y.D. and M.Sh.; HPLC analysis of CYP17A1 activity and
interpretation of the results: A.Y.; Drafting and revising the article: Y.D., S. J-Š., A.Y.
and S.U.
Conflicts of Interest.
No potential conflicts of interest.
Acknowledgements
Presented results are obtained in the frame of Belarus-Serbia bilateral projects (Projects
No X18SRBG-002 / 51-03-003036/2017-09/02 and X20SRBG-004 / 337-00-
00612/2019-09/04) which are being realized between Institute of Bioorganic
Chemistry of NAS of Belarus and University of Novi Sad, Faculty of Sciences. The
authors also acknowledge financial support of the Ministry of Education, Science and
Technological Development of the Republic of Serbia (Grant No. 451-03-68/2020-14/
200125).
References
1. Estabrook RW (2003) A passion for P450s (rememberances of the early history
of research on cytochrome P450). Drug Metab Dispos 31, 1461–1473.
2. Storbeck KH, Swart P, Africander D, Conradie R, Louw R, Swart AC (2011)
16α-hydroxyprogesterone: origin, biosynthesis and receptor interaction. Mol
Cell Endocrinol 336, 92–101.
3. Hall PF (1991) Cytochrome P-450 C21scc: one enzyme with two actions:
hydroxylase and lyase. J Steroid Biochem Mol Biol 40, 527–532.

4. Sushko TA, Gilep AA, Usanov SA (2012) Mechanism of intermolecular
interactions of microsomal cytochrome P450s CYP17 and CYP21 involved in
steroid hormone biosynthesis. Biochem (Moscow) 77, 585–592.
5. Maitra A, Shirwalkar H (2003) Congenital adrenal hyperplasia: biochemical
and molecular perspectives. Indian J Exp Biol 41, 701–709.
6. Miller WL (2012) The Syndrome of 17,20 Lyase Deficiency. J Clin Endocrinol
Metab 97, 59–67.
7. Qin KN, Rosenfield RL (1998) Role of cytochrome P450c17 in polycystic
ovary syndrome. Mol Cell Endocrinol 145, 111–121.
8. Ogo A, Haji M, Ohashi M, Nawata H (1991) Markedly increased expression of
cytochrome P-450 17 alpha-hydroxylase (P-450c17) mRNA in adrenocortical
adenomas from patients with Cushing's syndrome. Mol Cell Endocrinol 80, 83–
89.
9. Sushko TA, Gilep AA, Usanov SA (2014) Genetics, Structure, Function, Mode
of Actions and Role in Cancer Development of CYP17. Anti-cancer Agents Med
Chem 14(1), 66–76.
10.Sushko TA, Gilep AA, Yantsevich AV, Usanov SA (2013) Role of microsomal
steroid hydroxylases in Δ7-steroid biosynthesis. Biochem (Moscow) 78, 282–
289.
11.Popadyuk II, Markov AV, Morozova EA, Babich VO, Salomatina OV,
Logashenko EB, Zenkova MA, Tolstikova TG, Salakhutdinov NF (2017)
Synthesis and evaluation of antitumor, anti-inflammatory and analgesic activity
of novel deoxycholic acid derivatives bearing aryl- or hetarylsulfanyl moieties
at the C-3 position. Steroids 127, 1–12.
12.Miljković D, Kuhajda K, Hranisavljević J (1996) Selective C-12 oxidation of
cholic acid. J Chem Res, Synop 2, 106-107
13.Grbović Lj, unpublished results
14.Lin H-S, Rampersaud AA, Richett ME, Harper RW, Beavers LS, McClure DB,
Gardner AJ, Eacho PI, Foxworthy PS, Gadski RA (1999) Synthesis and in vitro
biological activity of 4α-(2-propenyl)-5α-cholest-24-en-3α-ol: A24,25-dehydro
analog of the hypocholesterolemic agent 4α-(2-propenyl)-5α-cholestan-3a-ol.
Steroids 64, 217–227.
15.Uddin MH, Roy MC, Tanaka J (2011) Cytotoxic cholic acid type sterones from
a marine soft coral Paraminabea Sp. Chem Nat Compd 47, 64–67.
16.Li Y, Dias JR (1996) Syntheses of α- and β-dimers of lithocholic acid esters.
Org Prep Proced Int 28(2), 203–209.
17.Tohma M, Mahara R, Takeshita H, Kurosawa T (1986) A convenient synthesis
of 3β,12α-, 3β,7α-, and 3β,7β-dihydroxy-5 cholen-24-oic acids: unusual bile
acids in human biological fluids. Steroids 48(5-6) 331–338.

18.Penov-Gaši K, Cvjetićanin S, Stojanović S, Kuhajda K, Stupavský L, Čanadi J,
Molnár-Gábor D, Mijačević L, Sakač M (2000) Chemical transformations of
3β,17β-dihydroxy-16-oximino-5-androstene. Acta Period Technol 31(B), 675.
19.Savić MP, Klisurić OR, Penov-Gaši KM, Jakimov DS, Sakač MN,
Djurendić EA (2016) Synthesis, structural analysis and cytotoxic activity of
novel A- and B-modified D-homo lactone androstane derivative. J Chem
Crystallogr 46, 84–92.
20.Miljković D, Petrović J, Stajić M, Miljković M (1973) The Beckmann
fragmentation reaction of some alpha-hydroxy ketoximes. J Org Chem 38,
3585–3588.
21.Đurendić EA, Zaviš (Savić) MP, Sakač MN, Kojić VV, Bogdanović GM, Penov-
Gaši KM (2008) Synthesis and biological evaluation of some A,D-ring modified
16,17-secoandrostane derivatives. Coll Czech Chem Comm 73(5), 627–636.
22.Savić MP, Ajduković JJ, Plavša JJ, Bekić SS, Ćelić AS, Klisurić OR, Jakimov
DS, Petri ET, Djurendić EA (2018) Evaluation of A-ring fused pyridine D-
modified androstane derivatives for antiproliferative and aldo-keto reductase
1C3 inhibitory activity. Med Chem Commun 9, 969–981.
23.Opsenica D, Pocsfalvi G, Juranić Z, Tinant B, Declercq J-P, Kyle DE, Milhous
WK, Šolaja BA (2000) Cholic acid derivatives as 1,2,4,5-Tetraoxane Carriers:
Structure and Antimalarial and Antiproliferative Activity. J Med Chem 43,
3274–3282.
24.do Nascimento PG, Lemos TL, Almeida MC, de Souza JM, Bizerra AM,
Santiago GM, da Costa JG, Coutinho HD (2015) Lithocholic acid and
derivatives: Antibacterial activity. Steroids 104, 8–15.
25.Abdu-Allah HHM, Chang TT, Li W-S (2016) Synthesis of B- and C-ring-
modified lithocholic acid analogues as potential sialyltransferase inhibitors.
Steroids 112, 54–61.
26.Tinajero-Delgado V, Romero-Ávila M, Flores-Álamo M, Arteaga MAI
Synthesis (2014) NMR characterization and crystal Structure of methyl 3a,7a-
dihydroxy-12-oxo-5β-cholanate. J Chem Crystallogr 44, 487–492.
27.Popadyuk II, Markov AV, Salomatina OV, Logashenko EB, Shernyukov AV,
Zenkova MA, Salakhutdinov NF (2015) Synthesis and biological activity of
novel deoxycholic acid derivatives. Bioorg Med Chem 23, 5022–5034.
28.Uekawa T, Ishigami K, Kitahara T (2004) Short-step synthesis of chenodiol
from stigmasterol. Biosci Biotechnol Biochem 68(6), 1332–1337.
29.Djurendić E, Sakač M, Zaviš M, Gaković A, Čanadi J, Andrić S, Klisurić O,
Kojić V, Bogdanović G, Penov Gaši K (2008) Synthesis and biological
evaluation of some new A,B-ring modified steroidal D-lactones. Steroids 73,
681–688.

30.Djurendić EA, Savić MP, Klisurić OR, Sakač MN, Bogdanović GM, Jakimov
DS, Penov Gaši KM (2012) Synthesis, X-ray, structural analysis, and cytotoxic
activity of some new androstane D-homo lactone derivatives. Struct Chem 23,
1761–1767.
31.Djurendić EA, Zaviš MP, Sakač MN, Čanadi JJ, Kojić VV, Bogdanović GM,
Penov Gaši KM (2009) Synthesis and antitumor activity of new D-seco and D-
homo androstane derivatives. Steroids 74, 983–988.
32.Zaviš MP, Djurendić EA, Petri ET, Ćelić A, Klisurić OR, Sakač MN, Jakimov
DS, Kojić VV, Penov Gaši KM (2013) Synthesis, structural analysis and
antiproliferative activity of some novel D-homo lactone androstane derivatives.
RSC Advs 3, 10385–10395.
33.Pettersen EF, Goddard TD, Huang CC, Couch GS, Greenblatt DM, Meng EC,
Ferrin TE (2004) UCSF Chimera – a visualization system for exploratory
research and analysis. J Comput Chem 25(13), 1605–1612.
34.Trott O, Olson AJ (2010) AutoDockVina: improving the speed and accuracy of
docking with a new scoring function, efficient optimization and multithreading.
J Comput Chem 31, 455–461.
35.Baell JB, Holloway GA (2010) New substructure filters for removal of pan
assay interference compounds (PAINS) from screening libraries and for their
exclusion in bioassays. J. Med. Chem. 53(7), 2719–2740.
36.Auchus RJ, Miller WL (1999) Molecular modeling of human P450c17
(17alpha-hydroxylase/17,20-lyase): insights into reaction mechanisms and
effects of mutations. Mol Endocrinol 13(7), 1169–1182.
37.Chen J, Zhao KN, Chen C (2014) The role of CYP3A4 in the biotransformation
of bile acids and therapeutic implication for cholestasis. Ann Transl Med 2(1),
7–16.
38.Juřica J, Dovrtělová G, Nosková K, Zendulka O (2016) Bile acids, nuclear
receptors and cytochrome P450. Physiol. Res. 65(4), 427–440.
39.Yadav R, Petrunak EM, Estrada DF, Scott EE (2017) Structural insights into
the function of steroidogenic cytochrome P450 17A1. Mol Cell Endocrinol 441,
68–75.
40.DeVore NM, Scott EE (2012) Structures of cytochrome P450 17A1 with
prostate cancer drugs abiraterone and TOK-001. Nature 482(7383), 116–119.
Figure 1. Steroidogenic reaction pathways catalyzed by human CYP17A1.

Figure 2. Absorbance spectrum of human recombinant CYP17A1 in ligand-free form.
Insets: MALDI-TOF mass-spectrum (A), SDS-PAGE (14% gel, Coomassie Blue G-
250 staining) of purified CYP17A1 (ST – protein ladder (Page RulerTM Plus Prestained
Protein Ladder, 10 to 250 kDa (Thermo Fisher Scientific)), 1 – purified protein) (B),
difference spectra of CYP17A1 in complex with P4 (substrate – dotted line) and
econazole (inhibitor – solid line) (C) and carbonyl complex spectrum (D)

1,2
1
A
B
0,8
0,6
0,4
0,2
0
0,04
0,03
0,02
0,01
0,00
0,04
0,03
0,02
0,01
0
C
D
350 370 390 410 430 450 470 490
400 420 440 460 480 500
-0,01
-0,02
-0,03
-0,04
-0,05
, nm
, nm
-0,01
-0,02
-0,03
-0,04
-0,05
250 300 350 400 450 500 550
, nm
600 650 700
Figure 3. Fragment of human CYP17A1 active site in complex with compound 9 and
P5 (A), compound 11 and P5 (B), compound 3 and GAL (C) and compound 1 (D). In
the stick and sphere representations, non-carbon atoms are indicated in blue (N) and
red (O), Fe atom is indicated as orange sphere. Molecules are colored by dark grey
(haem), green (compounds 1, 3, 9 and 11), and gold (P5 and GAL). Surface of G301,
A302, E305 and T306 is colored by grey.

A
B
C
D
Figure 4. CYP17A1 structure, representing amino acid residues forming active site
pocket, in complex with GAL (PDB ID: 3SWZ) (A), compounds L4 (Table 2) and 14
(B). In the stick and sphere representations, non-carbon atoms are indicated in blue (N)
and red (O), Fe atom is indicated as orange sphere. Molecules are colored by dark grey
(heme), green (compound L4), salmon (compound 14) and gold (GAL). Hydrogen
bonds are indicated as green lines. Surface is colored by cornflower blue.

A
B
Figure 5. Relative conversion rate of P4 in the presence of novel human CYP17A1
inhibitors. Conversion rate of P4 in the absence of modified steroids is taken as 100%.
Tick labels of X axis correspond the labels of compounds in Table 1.

Tables
Table 1. Novel steroidal ligands of human recombinant CYP17A1, found during
in vitro screening, and their binding parameters.
Titration
Spectral
Compound
Structure
Kd, μM
limits, μM
parameters
Type I
(max 390 nm,
min 420 nm)
ΔA = 0.050
P4 [10]
0.1-50
0.1-50
0.1-50
0.1-50
1.92±0.09
Type I
17OH-P4
[10]
(max 390 nm,
min 420 nm)
ΔA = 0.020
2.32±0.35
0.86±0.12
0.63±0.05
Type I
(max 390 nm,
min 420 nm)
ΔA = 0.070
P5 [10]
Type I
17OH-P5
[10]
(max 390 nm,
min 420 nm)
ΔA = 0.050
Type I
(max 387 nm,
min 423 nm)
ΔA = 0.046
1[11]
0.1-49
0.72±0.09

Type I
(max 387 nm,
min 424 nm)
ΔA = 0.034
2 [12]
0.1-97
0.1-10
26.0±0.7
Type I
(max 387 nm,
min 421 nm)
ΔA = 0.053
3 [13]
0.17±0.04
Type I
(max 388 nm,
min 418 nm)
ΔA = 0.017
4 [14]
5 [15]
6 [14]
0.1-59
0.1-20
0.1-10
13.2±0.4
0.46±0.04
5.6±0.4
Type I
(max 388 nm,
min 424 nm)
ΔA = 0.044
Type I
(max 386 nm,
min 419 nm)
ΔA = 0.038
Type I
(max 388 nm,
min 425 nm)
ΔA = 0.053
7 [16]
0.1-20
0.05±0.03

Type I
(max 385 nm,
min 419 nm)
ΔA = 0.022
8 [12]
0.1-97
0.1-89
108±6
Type I
(max 389 nm,
min 417 nm)
ΔA = 0.010
9 [17]
9.59±1.01
Type I
(max 388 nm,
min 424 nm)
ΔA = 0.030
10 [18]
11 [19]
0.1-97
18.5±0.4
Type I
(max 389 nm,
min 423 nm)
ΔA = 0.010
0.1-108
97.3±25.1
Type I
(max 383 nm,
min 419 nm)
ΔA = 0.040
12 [20]
13 [21]
0.1-97
0.1-20
6.4±0.5
Type I
(max 387 nm,
min 419 nm)
ΔA = 0.075
0.33±0.01
Type I
(max 389 nm,
min 427 nm)
ΔA = 0.019
14 [21]
0.1-59
42.8±2.9

Type I
(max 389 nm,
min 423 nm)
ΔA = 0.044
15 [22]
0.1-97
11.2±0.6
Table 2. Steroids that showed weak (low) or no binding to human recombinant CYP17A1.
Comp.
Structure
Comp.
Structure
L1
L10
[23]
[29]
L2
L11
[24]
[18]
L3
L12
[13]
[30]
L4
L13
[13]
[31]

L5
L14
[25]
[30]
L6
L15
[26]
[21]
L7
L16
[17]
[32]
L8
L17
[27]
[22]
L9
[28]