Characterization of two steroid hydroxylases from different Streptomyces spp. and their ligand-bound and -unbound crystal structures

Article abstract of DOI:10.1111/febs.14729

Bacterial cytochrome P450 (CYP) enzymes are involved in the hydroxylation of various endogenous substrates while using a heme molecule as a cofactor. CYPs have gained biotechnological interest as useful biocatalysts capable of altering chemical structures by adding a hydroxyl group in a regiospecific manner. Here, we identified, purified, and characterized two CYP154C4 proteins from Streptomyces sp. W2061 (StCYP154C4-1) and Streptomyces sp. ATCC 11861 (StCYP154C4-2). Activity assays showed that both StCYP154C4-1 and StCYP154C4-2 can produce 2′-hydroxylated testosterone, which differs from the activity of a previously described NfCYP154C5 from Nocardia farcinica in terms of its 16α-hydroxylation of testosterone. To better understand the molecular basis of the regioselectivity of these two CYP154C4 proteins, crystal structures of the ligand-unbound form of StCYP154C4-1 and the testosterone-bound form of StCYP154C4-2 were determined. Comparison with the previously determined NfCYP154C5 structure revealed differences in the substrate-binding residues, suggesting a likely explanation for the different patterns of testosterone hydroxylation, despite the high sequence similarities between the enzymes (54% identity). These findings provide valuable insights that will enable protein engineering for the development of artificial steroid-related CYPs exhibiting different regiospecificity.

Full text of DOI:10.1111/febs.14729

Received Date : 09-Jul-2018
Revised Date : 20-Nov-2018
Accepted Date : 09-Dec-2018
Characterization of two steroid hydroxylases from different Streptomyces
spp. and their ligand-bound and -unbound crystal structures
Bikash Dangi1†, Chang Woo Lee2,3†, Ki-Hwa Kim1, Sun-Ha Park2, Eun-Ji Yu1, Chang-Sook Jeong2,3,
Hyun Park2,3, Jun Hyuck Lee2,3* and Tae-Jin Oh1,4,5
*
1Department of Life Science and Biochemical Engineering, Sunmoon University, Asansi 31460,
Republic of Korea
2Unit of Polar Genomics, Korea Polar Research Institute, Incheon 21990, Republic of Korea
3Department of Polar Sciences, University of Science and Technology, Incheon 21990, Republic of
Korea
4Genome-based BioIT Convergence Institute, Asansi 31460, Republic of Korea
5Department of Pharmaceutical Engineering and Biotechnology, Sunmoon University, Asansi 31460,
Republic of Korea
†These authors contributed equally to this work.
* To whom correspondence should be addressed:
Dr. Jun Hyuck Lee, Unit of Polar Genomics, Korea Polar Research Institute, Incheon 21990, Republic
of Korea, Tel: +82-32-760-5555; Fax: +82-32-760-5509; E-mail: junhyucklee@kopri.re.kr.
Prof. Tae-Jin Oh, Department of Pharmaceutical Engineering and Biotechnology, Sunmoon
University, Asansi 31460, Republic of Korea, Tel: +82-41-530-2677; Fax: +82-41-530-2279; E-mail:
tjoh3782@sunmoon.ac.kr.
Running title: Structures of Streptomyces steroid hydroxylases
This article has been accepted for publication and undergone full peer review but has not
been through the copyediting, typesetting, pagination and proofreading process, which may
lead to differences between this version and the Version of Record. Please cite this article as
doi: 10.1111/febs.14729
This article is protected by copyright. All rights reserved.

Article type
: Original Article
Keywords: crystal structure; cytochrome P450; Streptomyces; steroid hydroxylase; X-ray
crystallography
Conflicts of interest
The authors declare no conflicts of interest.
Abbreviations
Amp, ampicillin; CO, carbon monoxide; CYP, cytochrome P450; DMSO, dimethyl sulfoxide; DTT,
1,4-dithiothreitol; HPLC, high-performance liquid chromatography; IPTG, isopropyl-1-thio-β-D-
galactopyranoside; LC-MS, liquid chromatography mass spectrometry; NMR, nuclear magnetic
resonance; PIDA, diacetoxyiodobenzene; SDS-PAGE, sodium dodecyl sulfate polyacrylamide gel
electrophoresis
Abstract
Bacterial cytochrome P450 (CYP) enzymes are involved in the hydroxylation of various endogenous
substrates while using a heme molecule as a cofactor. CYPs have gained biotechnological interest as
useful biocatalysts capable of altering chemical structures by adding a hydroxyl group in a
regiospecific manner. Here, we identified, purified, and characterized two CYP154C4 proteins from
Streptomyces sp. W2061 (StCYP154C4-1) and Streptomyces sp. ATCC 11861 (StCYP154C4-2).
Activity assays showed that both StCYP154C4-1 and StCYP154C4-2 can produce 2-hydroxylated
testosterone, which differs from the activity of a previously described NfCYP154C5 from Nocardia
farcinica in terms of its 16α-hydroxylation of testosterone. To better understand the molecular basis
of the regioselectivity of these two CYP154C4 proteins, two crystal structures of the ligand-unbound
form of StCYP154C4-1 and the testosterone-bound form of StCYP154C4-2 were determined.
Comparison with the previously determined NfCYP154C5 structure revealed differences in the
substrate-binding residues, suggesting a likely explanation for the different patterns of testosterone
hydroxylation, despite the high sequence similarities between the enzymes (54% identity). These
findings provide valuable insights that will enable protein engineering for the development of
artificial steroid-related CYPs exhibiting different regiospecificity.
Introduction
Heme-containing cytochrome P450 enzymes (CYPs) are found in all kingdoms. CYPs are usually
termed P450, because they have an absorption-maximum peak at ~450 nm following their reaction
with carbon monoxide (CO). CYPs have been frequently studied since their discovery based on their
ability to synthesize and metabolize various cellular molecules. Additionally, CYPs are attractive due
to their capabilities of hydroxylation, epoxidation, decarboxylation, and dealkylation of endogenous
This article is protected by copyright. All rights reserved.

and xenobiotic compounds, with these reactions targeting various molecules, including cholesterol,
sterols, alkanes, and fatty acids [1-4]. In humans, CYPs are key enzymes involved in drug metabolism
and are targets for drug design; however, investigations tend to focus on bacterial rather than
mammalian CYPs due to the insolubility of membrane-bound eukaryotic CYPs, which limits their
study. By contrast, soluble bacterial CYPs are easily overexpressed and engineered, and CYP
mutagenesis is highly effective at increasing product yield. Therefore, bacterial CYPs have been the
basis for studies involving the drug-development industry [5-9]. Cytochrome P450 enzymes are
classified into two major groups based on their localization and redox partners. Class I includes
mitochondrial and bacterial P450s that use two separate redox partners: an iron-sulfur protein
(ferredoxin and adrenodoxin) and a flavin-containing reductase (ferredoxin reductase and
adrenodoxin reductase). Class II enzymes localize to microsomes and receive electrons from
NADPH-cytochrome P450 reductase (CPR) [10]. There are also other minor classes of P450
enzymes, such as P450BM3 (CYP102A1), which does not require separate redox partners due to the
fact that a diflavin reductase partner is contained within the polypeptide chain [11]. CYP154 family
members have been identified in several Actinobacteria, such as Streptomyces coelicolor,
Streptomyces griseus, and Nocardia farcinica. They oxidatively hydroxylate steroid molecules regio-
and stereoselectively [12-14].
Steroids are widespread organic compounds that are essential for multiple functions in plants and
animals [15]. Steroids are widely used as “anti-agents’’ for their anti-inflammatory, antimicrobial,
antifungal, and antiviral effects [16]. Modification by hydroxylation can strongly affect the biological
activity, polarity, and toxicity of steroids. The position at which a steroid is hydroxylated is of great
importance; in particular, the presence of hydroxyl or ketone groups at carbons 3, 11, 17, and 21
correlates with biological activity. Commercially, 11α-hydroxylation of progesterone to yield
cortisone and 11β-hydroxylation of 11-deoxycortisol to form hydrocortisone have been established at
an industrial scale [17,18]. Steroids are used as a source of nutrients by bacteria, in which steroid
hydroxylase is involved in some steroid catabolic pathway reactions for the degradation of an
aliphatic side chain or opening of a sterane ring [19]. A number of P450s from diverse bacterial
sources are known to be involved in steroid hydroxylation, including CYP106A family (CYP106A1
and CYP106A2), CYP154C family (CYP154C3, CYP154C5, and CYP154C8), CYP260 family
(CYP260A1 and CYP260B1), and CYP109 family (CYP109B1 and CYP109E1) members [20]. The
mycobacterium P450s CYP125A1 and CYP142A1 are involved in cholesterol degradation via
hydroxylation and its sequential oxidation to aldehyde and carboxylic acid [21]. In one particularly
interesting study, the screening of 250 bacterial P450s expressed in Escherichia coli identified 24
P450s that were able to hydroxylate testosterone at different positions in a regio- and stereoselective
manner [22].
In the CYP154 family, two crystal structures (CYP154A1 and CYP154C1) have been reported from
Streptomyces coelicolor A3(2) [12,23]. Additionally, various steroid (pregnenolone, progesterone,
androstenedione, and testosterone)-bound forms of CYP154C5 from N. farcinica have been solved,
with these proteins sharing 93% amino acid sequence identity [24]. Interestingly, CYP154A1
structure analysis revealed that the heme orientation was 180° opposite that of most other CYP
structures, including CYP154C1; however, the reason for this and its biological meaning remain
unclear. In steroid-bound CYP154C5 structures, the C16 position of bound steroid molecules faces
This article is protected by copyright. All rights reserved.

the heme iron. This structural feature possibly explains the regio- and stereoselectivity of CYP154C5
in the conversion of steroid molecules to yield 16α-hydroxylated steroid products [24,25].
In this study, we report two CYP154C4 crystal structures from Streptomyces spp. that include the
ligand-unbound form from Streptomyces sp. W2061 (StCYP154C4-1) and the testosterone-bound
form from Streptomyces sp. ATCC 11861 (StCYP154C4-2). Notably, crystals of the testosterone-
bound form of StCYP154C4-1 and apo-StCYP154C4-2 could not be obtained, despite their high
sequence identity (93%). Structural comparison of the two obtained structures revealed different
structural rearrangement of the α13-helix region according to testosterone binding. Additionally,
biochemical activity assays showed that StCYP154C4-1 and StCYP154C4-2 yielded a 2-
hydroxylated testosterone product through a mechanism that differs from that of other CYP154
proteins. Furthermore, these structures show different active site residue compositions that might
explain variations in the regiospecificity of the testosterone product among different CYP154 family
members. Our results provide useful insights for future protein engineering and applications of this
enzyme.
Results and Discussion
Expression, purification, and spectrophotometric characterization of StCYP154C4s
Both CYPs, StCYP154C4-1 and StCYP154C4-2, were expressed in their soluble form in Escherichia
coli BL21(DE3) cells. Each fraction eluted using different concentrations of imidazole-containing
potassium phosphate buffer was confirmed by sodium dodecyl sulfate polyacrylamide gel
electrophoresis (SDS-PAGE) analysis, which showed a band at ~65 kDa (Fig. 1A), although the
theoretical molecular weight calculated for both CYPs was ~45 kDa. The difference in molecular
weights was due to the Trx-His-s-enterokinase fusion sequence in the pET32a(+) vector transcribed
and translated along with the StCYP154C4s sequence. StCYP154C4s purity was determined by Rz
value, which was calculated as the ratio of absorbance at ʎmax of the soret band to the absorbance at
280 nm, with those for StCYP154C4-1 and StCYP154C4-2 at 1.32 and 1.28, respectively, indicating a
high degree of purity for both CYPs [26]. UV-visible absorption spectroscopy of the oxidized forms
of StCYP154C4-1 and StCYP154C4-2 showed a soret band at 417 nm. Similarly, in their reduced
state and CO-bound form, both CYPs showed a peak at 447 nm (Fig. 1B and 1C), which represents
the characteristic signature of CYP heme in its Fe2+CO complex form.
Substrate-binding assay
Optical-absorption spectroscopy is a major advantage for the physiochemical characterization of CYP
and has been extensively utilized to obtain concentration measurements, monitor substrate and
inhibitor binding, and find different kinetics parameters. Substrate binding by CYPs causes the
displacement of water as a sixth ligand, which subsequently causes a spin shift of the heme iron from
low (415–418 nm) to high (390–394 nm) [27]. StCYP154C4s showed this phenomenon upon binding
to the testosterone substrate, with soret absorption at 418 nm shifting to 392 nm upon addition of
testosterone (Fig. 1D) and indicating that both CYPs were capable of binding testosterone. Therefore,
testosterone was selected as a substrate for in vitro reactions catalyzed by StCYP154C4s. Similarly,
This article is protected by copyright. All rights reserved.

based on this, the dissociation constant (Kd) values for CYP154C4s were determined by titrating
different concentrations of substrate (testosterone) until saturation (Fig. 2). The Kd value for
StCYP154C4-1 was found to be less than 10 μM, indicating tight binding. StCYP154C4-2 had a
higher Kd value than CYP154C4-1 (Table 1).
Enzyme activity and product identification
In the reaction mixture (reaction system I) containing 3 μM CYP, 0.025 U Fdr, 6 μM Fdx, and 1 U
glucose-6-phosphate dehydrogenase, the conversion of all three substrates (progesterone, testosterone,
and androstenedione) was less than 10%. HPLC analysis of the progesterone reaction mixture showed
a single product peak (Fig. 3A). LC-MS analysis of the reaction mixture showed the exact mass of a
hydroxylated progesterone (data not shown). The progesterone product was identified as 16α-
hydroxyprogesterone by co-elution with an authentic standard (Fig. 3A, inset III). The testosterone
reaction mixture showed the formation of at least two products by both CYPs (Fig. 3B). Among them,
one of the products (P1) was identified as 16α-hydroxytestosterone based on co-elution with a
standard (Fig. 3B, inset III). Based on NMR structural elucidation, the other product (P2) observed in
the testosterone reaction mixture was identified as a mixture of 2α and 2β-hydroxytestosterone (Fig.
3C).
Having found a very low conversion rate with reaction system I, in another reaction (reaction system
II), increases in the concentrations of reaction constituents to 15 μM CYP, 15 µM Fdx, 0.1 U Fdr, and
5 U glucose-6-phosphate dehydrogenase resulted in not only high conversion (>70%) rates for all
three substrates but also variation in product distributions. Among the three products found in the
progesterone reaction mixture (Fig. 4A), P2 was identified as monohydroxylated progesterone with an
exact mass of [M+H]+ 331.2291. Products P3 and P1 had exact masses of [M+H]+ 329.2115 and
[M+H]+ 347.2222, respectively. Both of these products (P1 and P3) may have been obtained from the
monohydroxylated product (P2), which was further supported by the time-dependent reaction results,
in which the concentration of P2 decreased with an increase in the time interval, while the
concentrations of the other two products (P1 and P3) increased with an increase in the time interval
(Fig. 4B). P1 may have been derived from sequential hydroxylation of P2, while P3 may have resulted
from dehydrogenation of the hydroxyl group, yielding a keto group. In another reaction (reaction
system III), we reduced the concentrations used in reaction system II by half, and carried out the
reaction catalyzed by StCYP154C4-1. This resulted in mostly a single monohydroxylated product
(P2) that was found in reaction system II, while the other two products P1 and P3 exhibited very low
conversion rates (Fig. 4C). This clearly indicates that P1 and P3 must have resulted from sequential
oxidation of P2 by StCYP154C4-1. Surprisingly, the progesterone exclusively hydroxylated at the
C16α-position (P) found in reaction system I was not detected in reaction system II, based on HPLC
comparisons with the standard (16α-hydroxyprogesterone) (Fig. 4D). Another substrate, testosterone,
was converted into a number of products using reaction system II. P3 was identified as a major
product of catalysis by StCYP154C4-1 (Fig. 5A, inset II); however, in a reaction catalyzed by
StCYP154C4-2, P2 was observed as a major product (Fig. 5A, inset I). The P2 produced by reaction
system II was identified as a mixture of 2α and 2β testosterone based on HPLC retention time
comparisons (Fig. 5A, inset III). The possible 16α-hydroxytestosterone (P1) exhibited a very low rate
of conversion compared to the other products. Androstenedione was converted into one major
monohydroxylated product (P3) and two other minor products (P1 and P2) by StCYP154C4-1 (Fig.
5B, inset I). The LC-MS analysis of P1 and P2 showed the exact masses of dihydroxy- and
monohydroxyandrostenedione, respectively. In addition to three different androstenedione products
formed by catalysis with StCYP154C4-1, StCYP154C4-2 also resulted in the formation of another
monohydroxylated product (P4) (Fig. 5B, inset II).
This article is protected by copyright. All rights reserved.

We also carried out an in vitro reaction supported by oxygen surrogates diacetoxyiodobenzene
(PIDA) and hydrogen peroxide (H2O2). Iodosobenzene is a PhIO-activated compound similar to PIDA
that has been used in many cases for P450-catalyzed in vitro reactions [33-36]. A PIDA-supported
reaction catalyzed by both StCYP154C4s with a progesterone substrate resulted in major conversion
(P) into 16α-hydroxyprogesterone, which was identified based on co-elution with the standard (Fig.
5C). The StCYP154C4-2-catalyzed reaction with testosterone showed the formation of a single
product (P4), which was identified as a mixture of 2α and 2β-hydroxytestosterone (Fig. 5D, inset I). In
a reaction catalyzed by StCYP154C4-1, minor conversion into 16α-hydroxytestosterone (P1) and
major conversion into 2α and 2β-hydroxytestosterone were observed (Fig. 5D, inset II). PhIO- and
NADPH-supported reactions by P450 can be mechanistically distinct, leading to differences in
product profiles [36]. A similar difference in product profiles was recently reported by us with
CYP154C8-catalyzed reactions supported by NADPH or PhIO [37]. In addition to 16α-hydroxylation,
a diacetoxyiodobenzene (PhIO system)-supported reaction by CYP154C8 also catalyzed the 6β-
hydroxylation of progesterone and 11-ketoprogesterone, neither of which were observed when the
reactions were supported by NADPH or NADH systems. In another report, CYP17A1 produced
different products in the presence of iodosobenzene and NADPH-P450 reductase to form compound I
[35]. MycG, a P450 enzyme, was also reported to display altered catalytic activity when it was
partnered with an alternative surrogate redox partner [38]. Recently, CYP106A2, when partnered with
11 different redox partners in separate reactions, exhibited differences in product patterns caused by
differences in the binding mode of the redox partners [39]. The reaction in the presence of a high
concentration (40 mM) of H2O2 showed low conversion rates (~5%) of testosterone and progesterone
(data not shown). The major conversion via 2-hydroxylation was observed with testosterone, while
the progesterone reaction mixture showed a single product peak corresponding to 16α-hydroxylation.
Androstenedione was not converted to any product following catalysis with both CYPs in the
presence of H2O2.
To the best of our knowledge, StCYP154C4 is the first reported wild-type CYP enzyme that
hydroxylates testosterone to yield isomers of hydroxytestosterone. A similar pattern of product
formation was observed with CYP106A2, with dehydroepiandrosterone hydroxylated at the 7α and 7β
positions to form the stereoisomer [40]. In a similar manner, CYP106A2 was found to hydroxylate
abietic acid, a diterpene, at the 12α and 12β positions [41]. Not many wild-type P450s are known to
hydroxylate the C2 positions, although the best-characterized P450BM3 (CYP102A1) has been
engineered to obtain 2-hydroxylated testosterone [42-44]. StCYP154C4s may represent important
biocatalysts in the future, with the aim of developing steroids with different regiospecificities by
protein engineering. Recently, a steroidogenic P450 CYP260A1 from Sorangium cellulosum strain So
ce56 was engineered for stereo- and regioselective hydroxylation of progesterone at very important 1α
and 17α positions [45]. In another study involving the protein engineering of P450BM3 variant M01
A82W S72I, which is capable of producing 16α-progesterone, researchers inverted the
stereoselectivity to 16β using a single active site mutation [46]. Other P450s of bacterial origin have
been demonstrated to hydroxylate testosterone at different positions [13,22,25]. However, there have
been no reports of any wild-type CYP that hydroxylates at both positions 2α and 2β.
StCYP154C4s represent a suitable model for understanding the function of CYP154C sub-class P450
enzymes. Our results show that not only does the choice of redox partners affect the type and
selectivity of P450 reactions, but also the varied concentrations of P450s and auxiliary proteins (redox
partners) can alter the reaction type catalyzed by P450 enzymes. Such results provide a unique
opportunity to further exploit the chemistries behind altered activity and product distribution. The lack
of suitable redox partners for support for in vivo conversion of the substrates catalyzed by both
StCYP154C4s was a major problem for identifying the products by NMR. The in vivo bioconversion
carried out in E. coli cells, which contained a plasmid harboring the genes encoding StCYP154C4s,
putidaredoxin reductase (Pdr) and putidaredoxin (Pdx), did not result in the conversion of the three
substrates. However, traces of products were detected when the reaction was supported by Pdx and
This article is protected by copyright. All rights reserved.

Pdr in vitro. Considering the interesting functions of both P450s, further research is needed to find
suitable redox partners to support in vivo conversion, leading to the elucidation of the structures of the
modified products. Mutagenesis of both enzymes should also be carried out to identify the specific
residues involved in these functions.
Overall structure
The crystal structures of CYP154C4 from Streptomyces sp. W2061 (StCYP154C4-1) and
testosterone-bound CYP154C4 from Streptomyces sp. ATCC 11861 (StCYP154C4-2) were
determined at 2.2 Å and 2.7 Å, respectively (Fig. 6A and 6B). Those two CYP154C4s share 93%
sequence identity, with 97% positives and only 29 different residues. Both crystals belonged to space
group I212121 and contained a single molecule in each asymmetric unit. StCYP154C4-1 was solved by
molecular replacement using CYP154C1 from S. coelicolor (PDB code 1GWI; 73% sequence
identity) as a search model [12]. The StCYP154C4-1 structure was subsequently used for molecular
replacement to solve the StCYP154C4-2 structure. Both crystal structures showed typical
characteristics of a CYP, with the heme-binding site located at the center of an α-helix-rich region.
StCYP154C4-1 comprises 21 α-helices and 6 β-strands (Fig. 6A), whereas testosterone-bound
StCYP154C4-2 comprises 19 α-helices and 6 β-strands, with two small α-helices missing as compared
with the StCYP154C4-1 structure (Fig. 6B). Structural alignments using the Dali server
(http://ekhidna2.biocenter.helsinki.fi/dali/) confirmed that StCYP154C4 has the highest structural
similarity with CYP154C5 from N. farcinca (PDB code 4J6B), with a Z-score of 58.7 [24].
Additionally, CYP154C1 from S. coelicolor (PDB code 1GWI) and CYP154A1 from S. coelicolor
(PDB code 1ODO) also showed high levels of structural similarity [12,23].
The active sites of the StCYP154C4s
The cofactor heme is tightly bound in the center of StCYP154C4 enzymes. Carboxyl groups of heme
form salt bridges with the conserved residues Lys57, His94, Arg98, Arg291, and His348. The side
chain of Cys350 coordinates the central iron of heme on the proximal side. The nonpolar region of the
heme molecule is surrounded by hydrophobic residues (Fig. 7). Above the bound heme, there is a
hydrophobic substrate-binding site, with Pro82, Leu87, Phe173, Phe174, Leu234, Ala237, Ala238,
and Val288 residues comprising the nonpolar region of the substrate-binding pocket. Additionally,
several polar residues, including Gln233, Thr242, Thr285, and Gln391, are also located in the
substrate-binding pocket (Fig. 8). They may be responsible for hydrogen bonding with the substrate.
All of these residues are shared between StCYP154C4-1 and StCYP154C4-2 and suggest that both
enzymes likely share similar substrate specificity, as well as regiospecificity (Fig. 8A and 8B). As
expected, this prediction was confirmed by the results of biochemical assays. Structural comparison of
StCYP154C4-1 and StCYP154C4-2 showed a conformational change in the α13 helix following
testosterone binding (Fig. 8C). The Ala238 located on the α13-helix moved to near the testosterone
substrate and formed a hydrophobic interaction. This movement induces rearrangement of the α13
helix. The orientations of the other residues forming the substrate-binding pocket are also slightly
altered by testosterone binding. Specifically, interactions between StCYP154C4-2 and testosterone
involve the O3 and O17 atoms of testosterone interacting with the OG1 of Thr285 and the OE1 of
Gln233, respectively. It is possible that these two hydrogen bonds are important in determining the
orientation of testosterone binding. Moreover, the C18 and C19 atoms of testosterone are oriented on
the opposite side of the α13 helix and form nonpolar interactions with Leu79, Pro82, and Leu87. The
This article is protected by copyright. All rights reserved.

C2 atom of testosterone faces the Fe of the heme molecule and lies within 4.5 Å in a position that
enables 2′-hydroxylation. Activity assays confirmed that StCYP154C4 is capable of catalyzing both
2α-hydroxylation and 2β-hydroxylation of testosterone.
The testosterone-binding mode of the StCYP154C4s
Structural superposition of testosterone-bound StCYP154C4-2 and testosterone-bound NfCYP154C5
(PDB code 4J6D; 52% sequence identity) revealed that the two enzymes have different testosterone-
binding modes, as well as different residue compositions in the substrate-binding pocket (Fig. 9) [24].
The hydrophobic residues Leu79, Pro82, Leu87, and Leu234, which comprise the nonpolar region of
the binding pocket in StCYP154C4-2, are represented by Met84, Val88, Phe92, and Ala240,
respectively, in NfCYP154C5. These differences form a unique cavity shape in StCYP154C4 that
might influence ligand binding and regiospecificity. Specifically, Thr285 in StCYP154C4 interacts
with the O3 atom of testosterone, but is represented by Val291 at the same positon in NfCYP154C5.
By contrast, testosterone binding in the NfCYP154C5 structure shows that Gln398 interacts with the
O3 atom of testosterone. In the StCYP154C4-2 structure, the O3 atom of testosterone forms a
hydrogen bond with Thr285 located at a point deeper within the binding pocket than NfCYP154C5
Gln398. These differences in the substrate-binding site and individual side chain interactions likely
enable testosterone to bind in the opposite direction in StCYP154C4. Moreover, structural information
verified the results obtained from activity assays, which showed that StCYP154C4 exhibits both 2-
hydroxylation and 16α-hydroxylation activity on testosterone, whereas NfCYP154C5 exhibits only
16α-hydroxylation activity.
In conclusion, we identified broad and flexible substrate-hydroxylation activity on the part of both
StCYP154C4-1 and StCYP154C4-2 enzymes as compared with the previously characterized
NfCYP154C5 enzyme. Our findings suggest that differences in residue positioning in the active site
might account for the altered substrate position and catalytic activity observed in the CYP154C4
enzyme. Collectively, these results provide useful insight for subsequent protein engineering of
CYP154C4 family members in order to allow production of modified steroid compounds.
Materials and Methods
Materials
Testosterone was purchased from Tokyo Chemical Industry Co., Ltd. (Seoul, Korea). NADPH,
glucose-6-phosphate, glucose-6-phosphate dehydrogenase, spinach ferredoxin, spinach ferredoxin
reductase, hydrogen peroxide, PIDA, catalase, ampicillin (Amp), chloramphenicol, and α-
aminolevulinic acid were purchased from Sigma-Aldrich (Yongin, Korea). DNA polymerase, T4
DNA ligase, and dNTPs were obtained from Takara Bio (Shiga, Japan). Isopropyl-1-thio-β-D-
galactopyranoside (IPTG) and 1,4-dithiothreitol (DTT) were obtained from Duchefa Bohemie (Seoul,
Korea). All other chemicals were high-grade products obtained from commercially available sources.
Restriction enzymes (EcoRI and HindIII) were obtained from Takara Clontech (Seoul, Korea).
This article is protected by copyright. All rights reserved.

Sequence accession numbers
The gene responsible for encoding the CYP enzyme was searched based on the heme-binding domain
signature [FxxGx(H/R)xCxG] from the sequences of Streptomyces sp. W2061 and Streptomyces sp.
ATCC11861. The name of the CYP enzyme (CYP154C4) was assigned by Dr. David Nelson
(http://drnelson.uthsc.edu/cytochromeP450.html). The nucleotide sequences of StCYP154C4-1 and
StCYP154C4-2 have been deposited in GenBank under accession numbers MH444361 and
MH444362, respectively.
Cloning, expression, and purification of the StCYP154C4s
The StCYP154C4-1 and StCYP154C4-2 genes were amplified from genomic DNA of Streptomyces
sp. W2061 and Streptomyces sp. ATCC 11861, respectively. The StCYP154C-1 sequence was
amplified using 5′-GAA TTC ATG ACC CGT ATC GCG CTC-3′ (EcoRI site is underlined) as a
forward primer and 5′-AAG CTT TCA GGC GTC GCC GCC G-3′ (HindIII site is underlined) as a
reverse primer. Similarly, amplification of StCYP154C4-2 was performed using specific forward 5′-
GAA TTC ATG ACC CGT ATC GCG CTC G-3′ (EcoRI site is underlined) and reverse 5′-AAG CTT
TCA GGC GTC GTG GCC GAA-3′ (HindIII site is underlined) primers. The target genes were
cloned into the pET32a(+) vector and introduced into E. coli BL21(DE3) cells. Transformed cells
were grown overnight at 37°C for seed culture and then inoculated into LB media supplemented with
100 µg/mL Amp, followed by incubation (180 rpm) at 37°C until cell density ranged from 0.6 to 0.8
according to the optical density at 600 nm. Cell cultures were induced with 0.3 mM IPTG and
supplemented with 100 µg/mL 5-aminolevulinic acid hydrochloride and 0.5 mM FeCl3. Protein
expression was performed at 20°C for 48 h, followed harvesting of cells by centrifugation. The cell
pellets were further washed and resuspended in 50 mM Tris-HCl buffer (pH 7.4) containing 10%
glycerol, 100 mM NaCl, and 1 mM DTT. For purification, the cells were lysed by ultrasonication,
followed by centrifugation at 15,000 rpm for 30 min at 4°C. The soluble fraction of the cell lysate was
mixed and purified by Ni2+-affinity chromatography using a TALON His-tag. Protein-bound resins
were pre-equilibrated with two-column volumes of equilibration buffer (potassium phosphate, pH
7.4), and bound proteins were eluted with elution buffer [potassium phosphate (pH 7.4), 10%
glycerol, and 100 mM NaCl] containing 20 mM, 100 mM, and 300 mM imidazole, respectively.
Fractions of eluted proteins were analyzed for purity by SDS-PAGE, and pure fractions were
concentrated by ultrafiltration using Amicon centrifugal filters (Millipore, Billerica, MA, USA) with a
molecular weight cut-off of 30 kDa.
UV–visible absorption spectroscopy
UV-visible spectra of StCYP154C4s were measured at room temperature on a Biochrome Libra
S35PC UV/visible spectrophotometer (Biochrome Libra, Cambridge, UK). Absorption spectroscopy
was used to determine CYP concentration, purity, and substrate spin shift. Concentration of both
CYPs were determined based on CO-difference spectra using the extinction coefficient Ɛ450–490
=
91,000 mM−1 cm−1 [4]. The proteins were diluted to 2 mL with potassium phosphate buffer (pH 7.4)
and separated into two cuvettes (sample and reference). The sample was bubbled with CO for 1 min,
followed by addition of sodium dithionite in both cuvettes. The cuvettes were further scanned using
the spectrophotometer at between 400 nm and 500 nm repeatedly until the reading at ~450 nm
deceased relative to the highest value of the readings. To determine the purity (RZ value) of both
CYPs, the sample containing the CYP enzyme was scanned in the range of 200 nm to 500 nm,
This article is protected by copyright. All rights reserved.

whereas the reference contained only buffer. Substrate-binding shift was observed in a sample
containing 1 μM CYP and 20 μM substrate in potassium phosphate buffer (pH 7.4), with the reference
containing only and equal concentration of substrate. The sample and reference were scanned at
between 350 nm and 500 nm. Dissociation constant (Kd) values for StCYP154C4s (1 µM) were
determined by titrating the substrate (testosterone) until the CYPs were saturated. Kd values were
determined as described previously [47].
Enzyme activity assay
In vitro activity of both StCYP154C4s was determined in the presence of NADPH, PIDA, and H2O2,
separately. The substrate (testosterone) and PIDA were dissolved in dimethyl sulfoxide (DMSO) and
100% ethanol, making the stock concentrations 20 mM and 50 mM, respectively. Reaction system I
contained 3 μM StCYP154C4s, spinach ferredoxin (6 μM), spinach ferredoxin reductase (0.025 U),
glucose-6-phosphate dehydrogenase (1 U), glucose-6-phosphate (10 mM), MgCl2 (5 mM), catalase
(100 μg/mL), and substrate (500 μM) in 500 μL potassium phosphate buffer (pH 7.4). The reactions
were initiated with 500 μM NADPH. The reaction mixtures were incubated at 30°C for 2 h, followed
by extraction with ethyl acetate (1,000 μL), which was dried and analyzed by HPLC and LC-MS.
Reaction system II was carried out in the presence of StCYP154C4s (15 µM), Fdx (15 µM), Fdr (0.1
U), glucose-6-phosphate (100 mM), glucose-6-phosphate dehydrogenase (5 U), magnesium chloride
(10 mM), catalase (100 μg/mL), substrate (500 μM), and 1 mM NADPH. The reactions mixture were
incubated at 30°C for different time intervals ranging from 2–6 h. The reactions mixtures were
extracted and analyzed as described for reaction system I. The other two reactions, supported with
PIDA (2 μM) and H2O2 (40 mM), contained 3 μM StCYP154C4, with the two StCYP154C4s being
analyzed separately. All reactions mixtures were incubated at 30°C for 2 h and extracted and analyzed
as described for reaction system I.
Product purification and characterization
The product (P2) submitted for NMR analysis was prepared by an up-scaled (300 mL) in vitro
reaction catalyzed by StCYP154C-1. The reactions were performed separately in a 10-mL volume in
potassium phosphate buffer (pH 7.4) containing 250 μM substrate. The reactions were initiated with
the addition of 2 mM PIDA and incubated at 30°C for 2 h. Each reaction was extracted with 10 mL
ethyl acetate, dried in vacuo, and dissolved in HPLC-grade methanol. The sample was filtered through
a 0.45-µm pore polytetrafluoroethylene filter and subjected to preparative HPLC (Shimadzu, Kyoto,
Japan) with a C18 column (Mightysil RP-18 GP; 150 × 4.6 mm, 5 µm; Kanto Chemical, Tokyo,
Japan). The purity of the product (P2) was reconfirmed by HPLC. Fractions containing the product
(P2) were dried and dissolved in DMSO-d6, and the sample was further subjected to NMR analyses
(1H-NMR, 13C-NMR, ROSEY, COSY, HMBC, and HSQC) at 800 MHz using a Varain Unity
INOVA spectrometer (Varian, Palo Alto, CA, USA).
2α-Hydroxytestosterone:1H NMR (800 MHz, DMSO-d6) δ 5.63 (d, J = 1.6 Hz, 1H), 5.02 (d, J = 4.4
Hz, 1H), 4.48 (d, J = 4.8 Hz, 1H),4.14 (dt, J = 13.7, 4.9 Hz, 2H), 3.44 (dt, J = 8.5, 5.0 Hz, 1H), 2.35
(tdd, J = 14.0, 5.3, 1.8 Hz, 1H), 2.26 (ddd,J = 14.4, 4.3, 2.4 Hz, 1H), 2.12 (dd, J = 12.6, 5.5 Hz, 1H),
1.87 – 1.80 (m, 1H), 1.79 – 1.73 (m, 2H), 1.52 (qd,J = 11.0, 3.6 Hz, 1H), 1.53 – 1.45 (m, 1H), 1.52 –
1.44 (m, 1H), 1.44 (t, J = 13.3 Hz, 1H), 1.36 (ddd, J = 25.9,13.1, 3.9 Hz, 1H), 1.39 – 1.31 (m, 1H),
1.23 (s, 3H), 1.20 (dp, J = 18.1, 5.8 Hz, 1H), 0.98 (td, J = 12.9, 4.2Hz, 1H), 0.95 – 0.87 (m, 1H), 0.89
(s, 1H), 0.86 (td, J = 11.5, 11.1, 4.2 Hz, 1H), 0.68 (s, 3H). 13C NMR (201 MHz, DMSO-d6) δ 199.98
(C3), 171.15 (C5), 121.45 (C4), 80.29 (C17), 69.25 (C2), 54.45 (C9), 50.37 (C14), 44.96 (C1), 42.84
This article is protected by copyright. All rights reserved.

(C13), 40.61 (C10), 36.67 (C12), 35.04 (C8), 31.98 (C6), 31.75 (C7), 30.27 (C16), 23.43 (C15), 20.59
(C11), 18.06 (C19), 11.67 (C18).
2β-Hydroxytestosterone: 1H NMR (800 MHz, DMSO-d6) δ 5.65 (d, J = 1.3 Hz, 1H), 5.15 (d, J =
4.2 Hz, 1H), 4.48(d, J = 4.8 Hz, 1H), 3.99 (dt, J = 10.9, 4.5 Hz, 1H), 3.43 (dt, J = 8.6, 4.9, 4.8 Hz,
1H), 2.45(tdd, J = 13.0, 5.0, 1.0 Hz, 1H), 2.20 (ddd, J = 12.6, 4.3, 2.5 Hz, 1H), 2.07 (dd, J = 13.8,4.8
Hz, 1H), 1.88 – 1.83 (m, 1H), 1.87 – 1.80 (m, 1H), 1.79 – 1.73 (m, 1H), 1.63 – 1.61 (m,1H), 1.61 (t, J
= 5.2 Hz, 1H), 1.63 – 1.59 (m, 1H), 1.53 – 1.45 (m, 1H), 1.41 (dd, J = 13.2,3.8 Hz, 1H), 1.39 – 1.31
(m, 1H), 1.26 – 1.21 (m, 1H), 1.20 (dp, J = 18.1, 5.8 Hz, 1H),1.15 (s, 3H), 1.04 (td, J = 12.9, 4.2 Hz,
1H), 0.95 – 0.87 (m, 1H), 0.95 – 0.84 (m, 1H), 0.68(s, 3H). 13C NMR (201 MHz, DMSO-d6) δ
199.41 (C3), 172.90 (C5), 120.29 (C4), 80.26 (C17), 68.61 (C2), 51.21 (C9), 50.39 (C14), 43.30
(C13), 41.28 (C1), 40.69 (C10), 36.70 (C12), 35.62 (C8), 34.11 (C7), 32.50 (C6), 30.33 (C16), 23.48
(C15), 22.44 (C19), 22.14 (C11), 11.81 (C18).
Analytical methods
The reaction mixtures were analyzed by ultra-HPLC (UHPLC) using an SYNAPT G2-S/ACUITY
UPLC quadrupole time-of-light/electrospray ionization mass spectrometry (Waters, Milford, MA,
USA) in positive ion mode. The sample was injected into the UHPLC system and separated using a
Mightysil reversed phase C18 column (4.6 × 250 mm, 5 µm; Kanto Chemical). Acetonitrile (B) and
water (A) were used as mobile phases in a gradient system of B (15% for 0–10 min, 50% for 10–20
min, 70% for 20–25 min, and 15% for 25–40 min) at a flow rate of 1 mL/min. Detection of substrates
and their product was performed by UV absorbance at 242 nm.
Crystallization and data collection
Initial crystallization screenings of both StCYP154C4-1 and testosterone-bound StCYP154C4-2 were
performed using a crystallization robot (Mosquito; TTP Labtech, Cambridge, MA, USA) with the
sitting-drop vapor-diffusion method at 293 K in 96-well crystallization plates (Emerald Bio,
Bainbridge Island, WA, USA). Commercially available screening kits were used for screening,
including the MCSG suite I-IV (Microlytic, Burlington, MA, USA) and SaltRx (Hampton Research,
Aliso Viejo, CA, USA). Co-crystallization of testosterone-bound StCYP154C4-2 was prepared by
adding 10 mM testosterone. The drops were mixed with a 0.3 μL of protein solution and 0.3 μL of
reservoir solution, and equilibrated against 80 μL reservoir solution at 293 K. StCYP154C4-1 crystals
were grown under conditions of 25% (w/v) PEG3350, 0.2 M lithium sulfate, and 0.1 M Bis-Tris:HCl
(pH 6.5) (MCSG2 #B11). Crystals of testosterone-bound StCYP154C4-2 were obtained from
conditions of 1 M sodium citrate tribasic and 0.1 M sodium cacodylate:HCl (pH 6.5) (MCSG3 #A1).
Before mounting the crystals, those of testosterone-bound StCYP154C4-2 were briefly soaked in
paratone-N oil used as a cryoprotectant, whereas StCYP154C4-1 crystals were directly mounted. X-
ray diffraction data for both crystals were collected at the BL-5C beam line of the Pohang Accelerator
Laboratory (PAL; Pohang, Korea) at 100 K. The 2.2-Å resolution of the diffracted data for
StCYP154C4-1 contained 180 images, and the 2.7-Å resolution of the diffracted data for testosterone-
bound StCYP154C4-2 contained 150 images. Image data were collected at a 1° oscillation and an
exposure time of 1 s per image. All diffraction data were indexed, processed, and scaled with HKL-
2000 [48]. The statistics associated with the X-ray diffraction data are presented in Table 2.
This article is protected by copyright. All rights reserved.

Structure determination and refinement
Molecular replacement of StCYP154C4-1 was successfully performed using the MOLREP program
from the CCP4i package [49,50]. The crystal structure of CYP154C1 from S. coelicolor (PDB code
1GWI; 73% sequence identity) was used as the template model [12]. One molecule was predicted in
an asymmetric unit with a Matthews coefficient of 3.45 Å3Da−1, corresponding to 64.50% solvent
[51]. The resulting structure was manually rebuilt and refined using COOT and REFMAC5 [52-54].
The final model of StCYP154C4-1 showed Rwork and Rfree values of 17.24% and 22.17%, respectively.
Testosterone-bound StCYP154C4-2 was solved by molecular replacement using StCYP154C4-1 as a
template model. Rebuilding and refinement of the structure were performed using same methods as
those used for StCYP154C4-1. The final model of testosterone-bound StCYP154C4 had Rwork and Rfree
values of 20.31% and 25.10%, respectively. The qualities of both final models were checked using
Molprobity and PROCHECK [55,56]. Structure coordinates and structure-factor data for
StCYP154C4-1 and testosterone-bound StCYP154C4-2 have been deposited in the Protein Data Bank
with accession codes 6A7I and 6A7J, respectively.
Acknowledgments
We thank the staff at the X-ray core facility of the Korea Basic Science Institute (Ochang, Korea) and
BL-5C of the Pohang Accelerator Laboratory (Pohang, Korea) for their assistance with data
collection. This work was supported by the Korea Polar Research Institute (grant no. PE18210).
Additionally, we would like to thank the Division of Magnetic Resonance, Korea Basic Science
Institute, Ochang, Chungbuk, Korea, for their help with NMR analysis.
Author contributions
H.P., J.H.L., and T.-J.O. designed and supervised the project. B.D., K.-H.K., and E.-J.Y. performed
the cloning, expression, and protein purification experiments. C.W.L. and S.-H.P. performed
crystallization, collected X-ray data, and solved the protein structures. B.D., K.-H.K., C.-S.J., and E.-
J.Y. performed the enzyme activity experiment. B.D. and K.-H.K. performed the HPLC and NMR
analyses. B.D., C.W.L., J.H.L., and T.-J.O. wrote the initial manuscript, and H.P. contributed to
revisions. All authors discussed the results, commented on the manuscript, and approved the
manuscript.
References
1. Nebert DW & Gonzalez FJ (1987) P450 genes: structure, evolution, and regulation. Annu Rev
Biochem 56, 945-993.
2. Guengerich FP (2002) Cytochrome P450 enzymes in the generation of commercial products.
Nat Rev Drug Discov 1, 359-366.
3. Ortiz de Montellano PR (2005) Cytochrome P450: Structure, Mechanism, and Biochemistry,
3rd edn. Kluwer Academic/Plenum Publishers, New York.
4. Omura T & Sato R (1964) The carbon monoxide-binding pigment of liver microsomes. I.
Evidence for its hemoprotein nature. J Biol Chem 239, 2370-2378.
5. Bentley SD, Chater KF, Cerdeño-Tárraga AM, Challis GL, Thomson NR, James KD, Harris
DE, Quail MA, Kieser H, Harper D, Bateman A, Brown S, Chandra G, Chen CW, Collins M,
This article is protected by copyright. All rights reserved.

Cronin A, Fraser A, Goble A, Hidalgo J, Hornsby T, Howarth S, Huang CH, Kieser T, Larke
L, Murphy L, Oliver K, O'Neil S, Rabbinowitsch E, Rajandream MA, Rutherford K, Rutter
S, Seeger K, Saunders D, Sharp S, Squares R, Squares S, Taylor K, Warren T, Wietzorrek A,
Woodward J, Barrell BG, Parkhill J & Hopwood DA (2002) Complete genome sequence of
the model actinomycete Streptomyces coelicolor A3(2). Nature 417, 141-147.
6. Lamb DC, Guengerich FP, Kelly SL & Waterman MR (2006) Exploiting Streptomyces
coelicolor A3(2) P450s as a model for application in drug discovery. Expert Opin Drug
Metab Toxicol 2, 27-40.
7. Lamb DC, Waterman MR, Kelly SL & Guengerich FP (2007) Cytochromes P450 and drug
discovery. Curr Opin Biotechnol 18, 504-512.
8. Purnapatre K, Khattar SK & Saini KS (2008) Cytochrome P450s in the development of
target-based anticancer drugs. Cancer Lett 259, 1-15.
9. Di Nardo G & Gilardi G (2012) Optimization of the bacterial cytochrome P450 BM3 system
for the production of human drug metabolites. Int J Mol Sci 13, 15901-15924.
10. Waskell L & Kim JJ (2015) Electron transfer partners of cytochrome P450. In Cytochrome
P450: Structure, Mechanism, and Biochemistry (Ortiz de Montellano PR, ed), pp. 33-68.
Springer International Publishing, Cham, Switzerland.
11. Hannemann F, Bichet A, Even KM & Bernhardt R (2007) Cytochrome P450 systems–
biological variations of electron trasport chains. Biochim Biophys Acta 1770, 330-334.
12. Podust LM, Kim Y, Arase M, Neely BA, Beck BJ, Bach H, Sherman DH, Lamb DC, Kelly
SL & Waterman MR (2003) The 1.92-Å structure of Streptomyces coelicolor A3(2)
CYP154C1. A new monooxygenase that functionalizes macrolide ring systems. J Biol Chem
278, 12214-12221.
13. Makino T, Katsuyama Y, Otomatsu T, Misawa N & Ohnishi Y (2014) Regio- and
stereospecific hydroxylation of various steroids at the 16alpha position of the D ring by the
Streptomyces griseus cytochrome P450 CYP154C3. Appl Environ Microbiol 80, 1371-1379.
14. Choi KY, Park HY & Kim BG (2010) Characterization of bi-functional CYP154 from
Nocardia farcinica IFM10152 in the O-dealkylation and ortho-hydroxylation of
formononetin. Enzyme Microb Technol 47, 327-334.
15. Baker ME (2011) Origin and diversification of steroids: co-evolution of enzymes and nuclear
receptors. Mol Cell Endocrinol 334, 14-20.
16. Donova MV & Egorova OV (2012) Microbial steroid trasformations: current state and
prospects. Appl Microbiol Biotechnol 94, 1423-1447.
17. Peterson DH, Murray HC, Eppstein SH, Reineke LM, Weintraub A, Meister PD & Leigh
HM (1952) Microbiological trasfromations of steroids. 1. Introduction of oxygen at carbon-
11 of progesterone. J Am Chem Soc 74, 5933-5936.
18. Petzoldt K, Annen K, Laurent H & Wiechert R (1982) Process for the preparation of 11-beta-
hydroxy steroids. Patent 4,353,985, Schering Aktiengesellschaft, Berlin, Germany.
This article is protected by copyright. All rights reserved.

19. Galán B, Garcia-Fernández J, Felpeto-Santero C, Fernández-Cabezón L & Garcia JL (2017)
Bacterial metabolism of steroids. In Aerobic Utilization of Hydrocarbons, Oils and Lipids.
Handbook of Hydrocarbon and Lipid Microbiology (Rojo F, ed) Springer, Cham,
Switzerland.
20. Szaleniec M, Wojtkiewicz AM, Bernhard R, Borowski T & Donova M (2018) Bacterial
steroid hydroxylases: enzyme classes, their functions and comparison of their catalytic
mechanisms. Appl Microbiol Biotechnol 102, 8153-8171.
21. Ouellet H, Johnston JB & Ortiz de Montellano PR (2010) The Mycobacterium tuberculosis
cytochrome P450 system. Arch Biochem Biophys 493, 82-95.
22. Agematu H, Matsumoto N, Fujii Y, Kabumoto H, Doi S, Machida K, Ishikawa J & Arisawa
A (2006) Hydroxylation of testosterone by bacterial cytochromes P450 using the Escherichia
coli expression system. Biosci Biotechnol Biochem 70, 307-311.
23. Podust LM, Bach H, Kim Y, Lamb DC, Arase M, Sherman DH, Kelly SL & Waterman MR
(2004) Comparison of the 1.85 Å structure of CYP154A1 from Streptomyces coelicolor
A3(2) with the closely related CYP154C1 and CYPs from antibiotic biosynthetic pathways.
Protein Sci 13, 255-268.
24. Herzog K, Bracco P, Onoda A, Hayashi T, Hoffmann K & Schallmey A (2014) Enzyme-
substrate complex structures of CYP154C5 shed light on its mode of highly selective steroid
hydroxylation. Acta Crystallogr D Biol Crystallogr 70, 2875-2889.
25. Bracco P, Janssen DB & Schallmey A (2013) Selective steroid oxyfunctionalisation by
CYP154C5, a bacterial cytochrome P450. Microb Cell Fact 12, 95.
26. Su Z, Horner JH & Newcomb M (2012) Cytochrome P450 119 compounds I formed by
chemical oxidation and photooxidation are the same species. Chemistry doi:
10.1002/chem.201202254.
27. Luthra A, Denisov IG & Sligar SG (2011) Spectroscopic features of cytochrome P450
reaction intermediates. Arch Biochem Biophys 507, 26-35.
28. Kuriyama K, Kondo E & Tori K (1963) Conformation of ring A in some 2β-hydroxy- and
2β-acetoxy-Δ4-3-ketosteroids. Tetrahedron Lett 4, 1485-1491.
29. Duax WL, Eger C, Pokrywiecki S & Osawa Y (1971) Crystalline molecular structures of two
derivatives of 2β-hydroxytestosterone having an unusual A-ring conformation. J Med Chem
14, 295-300.
30. Osawa Y & Gardner J (1971) Synthesis and conformation of 2β-hydroxytestosterone. J Org
Chem 36, 3246-3247.
31. Burnett RD & Kirk DN (1974) An improved synthesis for 2β-hydroxytestosterone. J Chem
Soc Perkin Trans 1 2, 284-285.
32. Duax WL, Cody V, Griffin J, Hazel J & Weeks CM (1978) Steroid structure and function -
II. Conformational transmission and receptor binding of medroxyprogesterone acetate. J
Steroid Biochem 9, 901-907.
This article is protected by copyright. All rights reserved.

33. Hrycay EG & Bandiera SM (2012) The monooxygenase, peroxidase, and peroxygenase
properties of cytochrome P450. Arch Biochem Biophys 522, 71-89.
34. Gelb MH, Heimbrook DC, Mälkönen P & Sligar SG (1982) Stereochemistry and deuterium
isotope effects in camphor hydroxylation by the cytochrome P450cam monoxygenase
system. Biochemistry 21, 370-377.
35. Yoshimoto FK, Gonzalez E, Auchus RJ & Guengerich FP (2016) Mechanism of 17α,20-
lyase and new hydroxylation reactions of human cytochrome P450 17A1: 18O labeling and
oxygen surrogate evidence for a role of a perferryl oxygen. J Biol Chem 291, 17143-17164.
36. Bhakta MN, Hollenberg PF & Wimalasena K (2005) P450/NADPH/O2- and P450/PhIO-
catalyzed N-dealkylations are mechanistically distinct. J Am Chem Soc 127, 1376-1377.
37. Dangi B, Park H & Oh TJ (2018) Effect of alternative redox partners and oxidizing agents in
CYP154C8 catalytic activity and product distribution. ChemBioChem, in press, doi:
10.1002/cbic.201800284.
38. Zhang W, Liu Y, Yan J, Cao S, Bai F, Yang Y, Huang S, Yao L, Anzai Y, Kato F, Podust
LM, Sherman DH & Li S (2014) New reactions and products resulting from alternative
interactions between the P450 enzyme and redox partners. J Am Chem Soc 136, 3640-3646.
39. Sagadin T, Riehm JL, Milhim M, Hutter MC & Bernhard R (2018) Binding modes of
CYP106A2 redox partners determine differences in progesterone hydroxylation product
patterns. Commun Biol 1, 99.
40. Schmitz D, Zapp J & Bernhardt R (2014) Steroid conversion with CYP106A2 - production of
pharmaceutically interesting DHEA metabolites. Microb Cell Fact 13, 81.
41. Bleif S, Hannemann F, Lisurek M, von Kries JP, Zapp J, Dietzen M, Antes I & Bernhardt R
(2011) Identification of CYP106A2 as a regioselective allylic bacterial diterpene
hydroxylase. ChemBioChem 12, 576-582.
42. van Vugt-Lussenburg BM, Damsten MC, Maasdijk DM, Vermeulen NP & Commandeur JN
(2006) Heterotropic and homotropic cooperativity by a drug-metabolising mutant of
cytochrome P450 BM3. Biochem Biophys Res Commun 346, 810-818.
43. Vottero E, Rea V, Lastdrager J, Honing M, Vermeulen NP & Commandeur JN (2011) Role
of residue 87 in substrate selectivity and regioselectivity of drug-metabolizing cytochrome
P450 CYP102A1 M11. J Biol Inorg Chem 16, 899-912.
44. Kille S, Zilly FE, Acevedo JP & Reetz MT (2011) Regio- and stereoselectivity of P450-
catalysed hydroxylation of steroids controlled by laboratory evolution. Nat Chem 3, 738-743.
45. Khatri Y, Jóźwik IK, Ringle M, Ionescu IA, Litzenburger M, Hutter MC, Thunnissen AWH
& Bernhardt R (2018) Structure-based engineering of steroidogenic CYP260A1 for stereo-
and regioselective hydroxylation of progesterone. ACS Chem Biol 13, 1021-1028.
46. Venkataraman H, Beer SB, Bergen LA, Essen Nv, Geerke DP, Vermeulen NP &
Commandeur JN (2012) A single active site mutation inverts stereoselectivity of
16‐hydroxylation of testosterone catalyzed by engineered cytochrome P450ꢀBM3.
ChemBioChem 13, 520-523.
This article is protected by copyright. All rights reserved.

47. Dangi B, Kim KH, Kang SH & Oh TJ (2018) Tracking down a new steroid-hydroxylating
promiscuous cytochrome P450: CYP154C8 from Streptomyces sp. W2233-SM.
ChemBioChem 19, 1066-1077.
48. Otwinowski Z & Minor W (1997) Processing of X-ray diffraction data collected in
oscillation mode. Methods Enzymol 276, 307-326.
49. Vagin A & Teplyakov A (2010) Molecular replacement with MOLREP. Acta Crystallogr D
Biol Crystallogr 66, 22-25.
50. Winn MD, Ballard CC, Cowtan KD, Dodson EJ, Emsley P, Evans PR, Keegan RM, Krissinel
EB, Leslie AG, McCoy A, McNicholas SJ, Murshudov GN, Pannu NS, Potterton EA, Powell
HR, Read RJ, Vagin A & Wilson KS (2011) Overview of the CCP4 suite and current
developments. Acta Crystallogr D Biol Crystallogr 67, 235-242.
51. Kantardjieff KA & Rupp B (2003) Matthews coefficient probabilities: Improved estimates
for unit cell contents of proteins, DNA, and protein-nucleic acid complex crystals. Protein
Sci 12, 1865-1871.
52. Adams PD1, Afonine PV, Bunkóczi G, Chen VB, Davis IW, Echols N, Headd JJ, Hung LW,
Kapral GJ, Grosse-Kunstleve RW, McCoy AJ, Moriarty NW, Oeffner R, Read RJ,
Richardson DC, Richardson JS, Terwilliger TC & Zwart PH (2010) PHENIX: a
comprehensive python-based system for macromolecular structure solution. Acta Crystallogr
D Biol Crystallogr 66, 213-221.
53. Emsley P & Cowtan K (2004) Coot: model-building tools for molecular graphics. Acta
Crystallogr D Biol Crystallogr 60, 2126-2132.
54. Murshudov GN, Skubák P, Lebedev AA, Pannu NS, Steiner RA, Nicholls RA, Winn MD,
Long F & Vagin AA (2011) REFMAC5 for the refinement of macromolecular crystal
structures. Acta Crystallogr D Biol Crystallogr 67, 355-367.
55. Chen VB, Arendall WB 3rd, Headd JJ, Keedy DA, Immormino RM, Kapral GJ, Murray LW,
Richardson JS & Richardson DC (2010) MolProbity: all-atom structure validation for
macromolecular crystallography. Acta Crystallogr D Biol Crystallogr 66, 12-21.
56. Laskowski RA, Rullmannn JA, MacArthur MW, Kaptein R & Thornton JM (1996) AQUA
and PROCHECK-NMR: programs for checking the quality of protein structures solved by
NMR. J Biomol NMR 8, 477-486.
This article is protected by copyright. All rights reserved.

Table 1. Dissociation constant (Kd) values of StCYP154C4-1 and StCYP154C4-2 with steroid
substrates.
Substrate
Progesterone
Androstenedione
Testosterone
StCYP154C4-1 (Kd)
7.12 ± 1.08
StCYP154C4-C2 (Kd)
62.76 ± 3.35
11.35 ± 3.14
54.57 ± 6.33
9.11 ± 2.15
73.38 ± 5.21
This article is protected by copyright. All rights reserved.

Table 2. X-ray diffraction data collection and refinement statistics.
Testosterone-bound
StCYP154C4-2
Data set
StCYP154C4-1
X-ray source
Space group
PAL 5C beam line
PAL 5C beam line
I212121
I212121
a=83.755,
c=127.807
α=β=γ=90
b=115.032, a=90.381,
b=113.94,
Unit-cell parameters (Å, °)
c=128.361
α=β=γ=90
Wavelength (Å)
Resolution (Å)
Total reflections
Unique reflections
Average I/σ (I)
0.9795
0.9795
50.00–2.20 (2.24–2.20)
204234
50.00–2.70 (2.75–2.70)
95298
31349 (1506)
42.9 (7.9)
17614 (872)
31.9 (5.8)
a
Rmerge
0.087 (0.445)
6.5 (6.3)
0.080 (0.419)
5.4 (5.0)
Redundancy
Completeness (%)
Refinement
98.0 (94.7)
96.5 (96.8)
Resolution range (Å)
46.47-2.19 (2.25-2.19)
29794 (2081)
47.55-2.71 (2.78-2.71)
16737 (1059)
No. of reflections of working
set
No. of reflections of test set
No. of amino acid residues
No. of water molecules
1554 (108)
409
872 (47)
408
268
8
b
Rcryst
0.173 (0.200)
0.222 (0.296)
0.019
0.217 (0.561)
0.266 (0.594)
0.013
c
Rfree
R.m.s. bond length (Å)
R.m.s. bond angle (°)
2.002
1.766
Average
(protein)
Average
(solvent)
B
B
value (Å2)
value (Å2)
43.956
47.231
87.640
70.805
a Rmerge = ∑｜<I> − I｜/ ∑<I>.
b Rcryst = ∑｜|Fo| − |Fc|｜/ ∑|Fo|.
c Rfree was calculated using 5% of all reflections, which were excluded from the refinement stages
using high-resolution data. Values in parentheses refer to the highest resolution shells.
This article is protected by copyright. All rights reserved.

Figure 1. Spectral analysis of StCYP154C4s. (A) SDS-PAGE of StCYP154C4-1 and StCYP154C4-2.
The dithionite-reduced CO-bound form of StCYP154C4-1 (B) and StCYP154C4-2 (C). (D) Substrate-
binding shift in the spectra of StCYP154C4-1 (solid line) and StCYP154C4-2 (dotted line). The
maximum absorbance (soret peak) was observed at 392 nm and characterized by a high-spin shift
achieved due to substrate binding in the CYP active site.
This article is protected by copyright. All rights reserved.

Figure 2. Spectrometric absorbance of StCYP154C4-1 (A, B, and C) and StCYP154C4-2 (D, E, and
F) bound with different concentrations of testosterone substrate. Dissociation constant (Kd) values
were calculated by plotting the peak to trough (Abs389 – Abs419) absorbance against different
concentrations of the substrate. Upper panel shows the shift in spin of the heme induced by binding of
different concentrations of substrate, and lower panel shows the plot of peak-to-trough absorbance
against the corresponding concentration of substrate. A (progesterone), B (testosterone), and C
(androstenedione) correspond to StCYP154C4-1, and D (progesterone), E (testosterone), and F
(androstenedione) correspond to StCYP154C4-2.
This article is protected by copyright. All rights reserved.

Figure 3. HPLC chromatogram of the progesterone (A) and testosterone (B) reaction mixture
catalyzed by StCYP154C4-1 (inset I) and StCYP154C4-2 (inset II) supported by reaction system I. P
indicate the hydroxylated product of progesterone which was identified as 16α-hydroxyprogesterone.
(C) The ROSEY of the mixture of 2α-OH testosterone and 2β-OH testosterone. The distance between
a 2α hydrogen and each hydrogen of 2α- and 2β-OH testosterone by force field MMFF94
optimization [28-32].
This article is protected by copyright. All rights reserved.

Figure 4. (A) Catalytic oxidation of progesterone supported by reaction system II, and catalyzed by
StCYP154C4-1 (inset I) and StCYP154C4-2 (Inset II). (B) HPLC chromatograms of progesterone
reaction mixture catalyzed by StCYP154C4-1 for different time intervals supported by reaction
system II. The reaction interval is indicated (2 h, 4 h, and 6 h). (C) HPLC chromatogram comparison
of progesterone reaction mixture catalyzed by StCYP154C4-1 supported by reaction system II (inset
I) and reaction system III (inset II). (D) HPLC chromatogram of 16α-OH progesterone (authentic
standard; indicated by blue color) in the progesterone reaction mixture catalyzed by StCYP154C4-1
(indicated by black color) and StCYP154C4-2 (indicated by red color) supported by reaction system
II. P1, P2, P3, and 16α-OH P indicate product 1, product 2, product 3, and16α-OH progesterone,
respectively. LC-MS analysis of the products showed P1 as a dihydroxylated product ([M+H]+
347.2222), P2 as a monhydroxylated ([M+H]+ 331.2291), and P3 as a possible oxidation product of
the monohydroxylated product ([M+H]+ 329.2115).
This article is protected by copyright. All rights reserved.

Figure 5. (A) Catalytic oxidation of testosterone supported by reaction system II and catalyzed by
StCYP154C4-1 (inset III) and StCYP154C4-2 (inset I). Insets II shows the chromatograms of the
authentic standard (16α-hydroxytestosterone). P1, P2, P3, and P4 indicate product 1, product 2,
product 3, and product 4 of testosterone, respectively. (B) HPLC chromatogram of androstenedione
reaction mixture catalyzed by StCYP154C4-1 (inset I) and StCYP154C4-2 (inset II) supported by
reaction system II. Inset III shows the chromatogram of the control reaction, carried out in the absence
of NADPH. P1, P2, P3, and P4 indicate product 1, product 2, product 3, and product 4 of
androstenedione, respectively. (C) HPLC chromatogram of the progesterone reaction mixture
catalyzed by StCYP154C4-1 (inset I) and StCYP154C4-2 (inset II) supported by
diacetoxyiodobenzene (PIDA). Inset III shows the chromatogram of the authentic standard 16α-OH
progesterone (16α-OH P). P indicates product which was identified as16α-OH progesterone. (D)
HPLC chromatogram of the testosterone reaction mixture catalyzed by StCYP154C4-1 (inset I) and
StCYP154C4-2 (inset II) supported by diacetoxyiodobenzene (PIDA). Inset III shows the
chromatogram of the authentic standard 16α-OH progesterone (16α-OH P). P2 and P4 are two
different product peaks of testosterone.
This article is protected by copyright. All rights reserved.

Figure 6. Crystal structures and multiple sequence alignment of StCYP154C4-1 and testosterone-
bound StCYP154C4-2. (A) The crystal structure of StCYP154C4-1. The α-helices and β-strands are
represented as ribbons colored in salmon and lime, respectively. The bound heme molecule is shown
as a magenta stick model. (B) The overall structure of testosterone-bound StCYP154C4-2 is
represented as a ribbon, with the α-helices and β-strands colored in olive and slate, respectively. The
bound testosterone is shown as a cyan stick model with a blue 2Fo-Fc electron-density map
(contoured at 1δ). (C) Structure-based multiple sequence alignment of StCYP154C4s with CYP154C1
from Streptomyces coelicolor A3(2) (PDB code 1GWI; UniProtKB code: Q9L142), CYP154C5 from
Nocardia farcinica (PDB code 4J6D; UniProtKB code: Q5YNS8), CYP154A1 from Streptomyces
coelicolor (PDB code 1ODO; UniProtKB code: Q9KZR7), and P450 mono-oxygenase PikC from
Streptomyces venezuelae (PDB code 2WHW; UniProtKB code: O87605). The secondary structures of
testosterone-bound StCYP154C4-2 are shown above the multiple aligned sequences. The Ala238,
which altered its position following testosterone binding, is indicated with a black triangle. Residues
making specific interactions with testosterone are denoted by a black circle above the residues.
This article is protected by copyright. All rights reserved.

Figure 7. Heme binding modes of StCYP154C4-1 and StCYP154C4-2. Stereo view of superposed
StCYP154C4-1 (salmon) and StCYP154C4-2 (olive) shows that the iron of the heme is coordinated
with the side chain of Cys350. Carboxyl groups of the heme interact with Lys57, His94, Arg98, and
Arg291, while the nonpolar region of the heme forms hydrophobic interactions with Met86, Val101,
Leu105, Ile235, Leu246, Leu279, Pro284, Val288, Phe314, Ala342, and Phe343.
This article is protected by copyright. All rights reserved.

Figure 8. Substrate-binding site of the StCYP154C4s. The substrate-binding site of (A) StCYP154C4-
1 and (B) testosterone-bound StCYP154C4-2 are colored in salmon and olive, respectively. The
residues in the substrate-binding site are shown as stick models, and the bound heme molecule
(magenta) and testosterone (cyan) also shown as stick models. Hydrogen bonds formed between the
bound testosterone and StCYP154C4-2 are shown with red dashed lines. (C) Stereo view of
StCYP154C4-1 superposed with testosterone-bound StCYP154C4-2. The α13 helix undergoes
conformational changes during testosterone binding.
This article is protected by copyright. All rights reserved.

Figure 9. Stereo view of the superimposed structures of testosterone(cyan)-bound StCYP154C4-2
(olive) and testosterone(green)-bound NfCYP154C5 (PDB code 4J6D; slate). The residues in the
vicinity of the bound testosterone and testosterone molecules are indicated by sticks. Structural
comparisons of the two enzymes show different testosterone-binding modes.
This article is protected by copyright. All rights reserved.