Molecular determinants for selective C25-hydroxylation of vitamins D2 and D3 by fungal peroxygenases

Article abstract of DOI:10.1039/c5cy00427f

Hydroxylation of vitamin D by Agrocybe aegerita and Coprinopsis cinerea peroxygenases was investigated in a combined experimental and computational study. 25-Monohydroxylated vitamins D₃ (cholecalciferol) and D₂ (ergocalciferol), compounds of high interest in human health and animal feeding, can be obtained through a reaction with both fungal enzymes. Differences in conversion rates, and especially in site selectivity, were observed. To rationalize the results, diffusion of D₂ and D₃ on the molecular structure of the two enzymes was performed using the PELE software. In good agreement with experimental conversion yields, simulations indicate more favorable energy profiles for the substrates' entrance in C. cinerea than for A. aegerita enzyme. On the other hand, GC-MS analyses show that while a full regioselective conversion of D₂ and D₃ into the active C₂₅ form is catalyzed by C. cinerea peroxygenase, A. aegerita yielded a mixture of the hydroxylated D₃ products. From the molecular simulations, relative distance distributions between the haem compound I oxygen atom and H₂₄/H₂₅ atoms (hydrogens on C₂₄ and C₂₅, respectively) were plotted. Results show large populations for O-H₂₅ distances below 3 ? for D₂ and D₃ in C. cinerea in accordance with the high reactivity observed for this enzyme. In A. aegerita, however, cholecalciferol has similar populations (below 3 ?) for O-H₂₅ and O-H₂₄, which can justify the hydroxylation observed in C₂₄. In the case of ergocalciferol, due to the bulky methyl group in position C₂₄, very few structures are found with O-H₂₄ distances below 3 ? and thus, as expected, the reaction was only observed at the C₂₅ position.

Full text of DOI:10.1039/c5cy00427f

Catalysis
Science &
Technology
View Article Online
View Journal | View Issue
PAPER
Molecular determinants for selective
C25-hydroxylation of vitamins D2 and D3 by fungal
peroxygenases†
Cite this: Catal. Sci. Technol., 2016,
6, 288
a
Fátima Lucas,‡ Esteban D. Babot,‡b Marina Cañellas,‡ac José C. del Río,b
*
Lisbeth Kalum,d René Ullrich,e Martin Hofrichter,e Victor Guallar,af
g
b
*
Angel T. Martínez and Ana Gutiérrez
Hydroxylation of vitamin D by Agrocybe aegerita and Coprinopsis cinerea peroxygenases was investigated
in a combined experimental and computational study. 25-Monohydroxylated vitamins D3 (cholecalciferol)
and D2 (ergocalciferol), compounds of high interest in human health and animal feeding, can be obtained
through a reaction with both fungal enzymes. Differences in conversion rates, and especially in site selectivity,
were observed. To rationalize the results, diffusion of D2 and D3 on the molecular structure of the two
enzymes was performed using the PELE software. In good agreement with experimental conversion yields,
simulations indicate more favorable energy profiles for the substrates' entrance in C. cinerea than for A.
aegerita enzyme. On the other hand, GC-MS analyses show that while a full regioselective conversion of
D2 and D3 into the active C25 form is catalyzed by C. cinerea peroxygenase, A. aegerita yielded a mixture of
the hydroxylated D3 products. From the molecular simulations, relative distance distributions between the
haem compound I oxygen atom and H24/H25 atoms (hydrogens on C24 and C25, respectively) were plotted.
Results show large populations for O–H25 distances below 3 Å for D2 and D3 in C. cinerea in accordance
with the high reactivity observed for this enzyme. In A. aegerita, however, cholecalciferol has similar
populations (below 3 Å) for O–H25 and O–H24, which can justify the hydroxylation observed in C24. In the
case of ergocalciferol, due to the bulky methyl group in position C24, very few structures are found with
O–H24 distances below 3 Å and thus, as expected, the reaction was only observed at the C25 position.
Received 20th March 2015,
Accepted 29th July 2015
DOI: 10.1039/c5cy00427f
www.rsc.org/catalysis
reactions include cytochrome P450, a family of haem proteins
playing a variety of physiological roles but often requiring an
Introduction
Selective oxygenations of aliphatic compounds are among the
most challenging reactions in organic chemistry for the regio-
and/or stereospecific synthesis of pharmaceuticals and fine
chemicals. Monooxygenases catalyzing such hydroxylation
auxiliary flavoenzyme (or a flavin-containing module) and a
source of reducing power to be activated by O2 – two facts
that limit their biotechnological applicability.
Recently, a new peroxidase type, which shares the active
site architecture and reaction mechanism of cytochrome
P450, but has the advantage of being activated directly by
H2O2, was isolated from Agrocybe aegerita,45 and later identi-
fied in a variety of sequenced basidiomycete genomes includ-
ing that of Coprinopsis cinerea.13 Due to the above character-
istics, these unspecific peroxygenases (EC 1.11.2.1) have huge
a Joint BSC-CRG-IRB Research Program in Computational Biology, Barcelona
Supercomputing Center, Jordi Girona 29, E-08034 Barcelona, Spain.
E-mail: fatima.lucas@bsc.es
b Instituto de Recursos Naturales y Agrobiología de Sevilla, CSIC, Reina Mercedes
10, E-41012 Seville, Spain. E-mail: anagu@irnase.csic.es; Fax: +32 954624002;
Tel: +32 954624711
c Anaxomics Biotech, Balmes 89, E-08008 Barcelona, Spain
d Novozymes A/S, Krogshoejvej 36, 2880 Bagsvaerd, Denmark
e TU Dresden, Department of Bio- and Environmental Sciences, Markt 23, 02763
Zittau, Germany
f ICREA, Passeig Lluís Companys 23, E-08010 Barcelona, Spain
g Centro de Investigaciones Biológicas, CSIC, Ramiro de Maeztu 9, E-28040
Madrid, Spain
biotechnological
potential
as
self-sufficient
mono-
oxygenases9,19 for hydroxylation of both aromatic2,3,24–27,42–44
and aliphatic compounds.4,15,33
The A. aegerita enzyme has been the most widely investi-
gated basidiomycete peroxygenase, but recent studies have
shown that the C. cinerea enzyme has comparative advantages
related to its high conversion yield/selectivity for some
hydroxylation reactions and its production as a recombinant
protein in an industrial expression host (by Novozymes,
Bagsvaerd, Denmark).5,6
† Electronic supplementary information (ESI) available: Molecular model for C.
cinerea peroxygenase obtained using A. aegerita's crystal structure as the
template. See DOI: 10.1039/c5cy00427f
‡ These three authors equally contributed to this work
288 | Catal. Sci. Technol., 2016, 6, 288–295
This journal is © The Royal Society of Chemistry 2016

View Article Online
Catalysis Science & Technology
Paper
Hydroxylation of vitamin D for the selective production of
its active C25-hydroxylated derivatives is one of the reactions
where the C. cinerea peroxygenase can be of biotechnological
use (Fig. 1).6 Supplementation with 25-hydroxyvitamin D has
a positive effect on different human diseases,11,20,21,28,41 and
also raises considerable interest for feeding broiler
chickens14,17,23,32 and other farm animals40 to reduce skele-
ton problems caused by rapid growth and reduced mobility.
Therefore, the use of peroxygenase in vitamin D hydroxyl-
ation represents an attractive alternative to the chemical
synthesis.
this compound was completely (100%) converted by the C.
cinerea enzyme within 60 min reaction (Fig. 2C) as compared
with the control reaction without peroxygenase (Fig. 2A). In
the A. aegerita peroxygenase reaction, up to 90% conversion
was observed (Fig. 2B). In the case of ergocalciferol (Fig. 2D–
F), a conversion of 85% was produced by C. cinerea while in
A. aegerita reaction, 81% conversion was observed. When
higher substrate concentrations were tested (e.g. 0.5 mM),
higher differences in the conversion rates by the two enzymes
were observed (40–50% and 20% conversions by the C.
cinerea and A. aegerita peroxygenases, respectively).
The goal of the present work is to rationalize the differ-
ences observed in cholecalciferol and ergocalciferol's (vita-
mins D3 and D2, respectively) conversion rates and site selec-
tivity by the A. aegerita and C. cinerea peroxygenases. The
work presented here consists of first, the enzymatic conver-
sion of these compounds by the two peroxygenases under the
same reaction conditions, and gas chromatography-mass
spectrometry (GC-MS) analyses to identify all reaction prod-
ucts. Then, the energy profiles and binding modes of vita-
mins D2 and D3 were determined by structure-based compu-
tational simulations, using the PELE software.10,12 Finally,
differences in site selectivity were investigated through the
analysis of the most favorable binding orientations in the
active site. Results show that molecular simulations can
effectively discriminate experimentally observed differences
in conversion rates and site selectivity.
Moreover, cholecalciferol and ergocalciferol conversion by
the C. cinerea peroxygenase showed a strict site selectivity
since it gave exclusively 25-hydroxycalciferol. Likewise, conver-
sion by the A. aegerita enzyme of ergocalciferol yielded exclu-
sively 25-hydroxyergocalciferol but for cholecalciferol, a mix-
ture of products was observed. The products include
25-hydroxycholecalciferol (64% of the initial substrate)
together with 24-hydroxycholecalciferol (21%) and 26/27-
hydroxycholecalciferol (7%), as confirmed by comparison with
true standards showing identical retention times and mass
spectra. The double peaks observed in the chromatograms
for both the substrate and the product (Fig. 2) correspond to
the isopyro (19β, 9β) and pyro (19α, 9α) isomers formed by a
Results and discussion
Experimental hydroxylation reactions
The conversion of cholecalciferol and ergocalciferol was
experimentally determined for both the A. aegerita and C.
cinerea peroxygenases. In the case of cholecalciferol (0.1
mM), GC-MS analyses of the reaction mixture revealed that
Fig.
(right) hydroxylation by the A. aegerita (B, E) and C. cinerea
peroxygenases (C, F), compared with control without enzyme
2 GC-MS analyses of cholecalciferol (left) and ergocalciferol
a
showing the substrate peaks (A, D), as TMS derivatives from 60 min
reactions. In all cases, the isopyro (left) and pyro (right) isomers from
the secosteroid thermal rearrangement are obtained (two small peaks
in A and D correspond to minor additional isomers).
Fig. 1 Enzymatic conversion of cholecalciferol (vitamin D3; 1) and
ergocalciferol (vitamin D2; 3) into their bioactive 25-hydroxylated
derivatives (2 and 4) by basidiomycete peroxygenases.
This journal is © The Royal Society of Chemistry 2016
Catal. Sci. Technol., 2016, 6, 288–295 | 289

View Article Online
Paper
Catalysis Science & Technology
thermal rearrangement involving ring-B closure that vitamin
D and its hydroxylated derivatives undergo due to the temper-
ature at which GC-MS EI(+) is carried out. Indeed, the pres-
ence of the two isomers during GC separation is a useful
indication that a secosteroid of the vitamin D type was
injected into the system.31 The position of the hydroxyl group
at the target C25 position was established by the MS of the
TMS derivative. The spectrum shows a prominent ion from
the C24–C25 bond cleavage, with characteristic fragment at m/z
131 and molecular ion at m/z 544. Additionally, characteris-
tic fragments at m/z 529 ([M-15]+), m/z 454 ([M-90]+), m/z 439
IJ[M-90-15]+), m/z 349 IJ[M-90-90-15]+) and m/z 413 were also
present. Therefore, both the chromatographic profiles and
the mass spectra correspond to the isomerized secosteroids.
as the fifth ligand of the haem iron (Cys36) and the distal
glutamic acid and arginine involved in haem activation by
H2O2 (Glu169 and Arg189),34 are conserved in the two
enzymes. Moreover, differences in the channel providing
access to the haem cofactor and the neighbor residues at the
channel entrance are shown in Fig. 4.
Computational modeling
Ligand diffusion energy profiles. For the simulations of
cholecalciferol and ergocalciferol access to the H2O2-activated
haem in the A. aegerita and C. cinerea peroxygenases, the
haem cofactor was modeled as compound I (an FeIVO· por-
phyrin cation radical complex). The substrate was placed close
to the entrance of the haem-access channel of the proteins
prepared at the optimal pH for peroxygenase activity (pH 7).
This initial location was identified using SiteMap,39 and from
there, the ligand was spawned inside the protein by PELE,10
following the distance between the reactive O atom in the
haem compound I and the cholecalciferol/ergocalciferol's C25
atom (Fig. 1).
PELE simulations were done in two stages: first, the sub-
strate is perturbed to reduce the C25–O distance and, when
this distance is below 5 Å, the substrate is free to explore the
active site cavity (at the haem distal side) with a 15 Å restrain.
In the first step, the ligand is perturbed with a combination
of large and small translations and rotations, ranging from
0.5 to 1.5 Å for translations, and 0.05 to 0.25 radians for rota-
tions. However, during the second stage, the translation
range is reduced (0.75–0.25 Å) to perform a finer active site
exploration. The plots shown in Fig. 5 correspond to three 48
h simulations, each with 80 processors, and show the sub-
strate-C25 to haem-O distance vs. the interaction energy
between the protein and the substrate at each of the different
poses explored.
Peroxygenase structure
A superimposition of the general structure and the haem
pocket residues in A. aegerita and C. cinerea peroxygenases is
shown in Fig. 3. The main differences are the two longer
loops (Fig. 3A, arrows) and the substitution of Phe69 by
Met69 (Fig. 3B, red label) in the C. cinerea enzyme. All other
haem pocket residues, including the proximal cysteine acting
Simulations show that for the A. aegerita peroxygenase,
the entrance of the substrates from the surface of the protein
is quite favorable but then, the access to the activated haem
is obtained against an uphill potential (Fig. 5A and B). For
the C. cinerea enzyme, the entrance is less open than in the
Fig. 3 Comparison of peroxygenase molecular structures. A) General
superimposition with the A. aegerita and C. cinerea proteins as pink
and yellow ribbons, respectively. The haem group (CPK-colored
spheres) and several haem pocket residues (pink- and CPK-colored
sticks, respectively) are shown, and two larger loops in the second
enzyme are indicated by arrows. B) Haem pocket residues in the A.
aegerita (pink sticks) and C. cinerea (CPK-colored sticks)
peroxygenases (red label indicates non-conserved methionine in posi-
tion 69 of the latter enzyme). A hypothetical product molecule identi-
fied as 4-hydroxymethylimidazole is shown in A (brown-colored vdW
spheres) and B (MZ0, brown-colored sticks).34 From A. aegerita PDB
2YOR, and C. cinerea homology model (provided in the ESI†).
Fig. 4 Solvent access surfaces showing differences in the size of the
haem (CPK-colored sticks) access channel in the peroxygenases of A.
aegerita (A) and C. cinerea (B). Several neighbor residues are shown as
CPK-colored vdW spheres, including Gly323 and Ser324 contributing
to occlude the haem channel in the C. cinerea enzyme.
290 | Catal. Sci. Technol., 2016, 6, 288–295
This journal is © The Royal Society of Chemistry 2016

View Article Online
Catalysis Science & Technology
Paper
Fig. 6 Superposition of cholecalciferol main active site positions (as
liquorice) obtained in PELE simulations on the A. aegerita (A) and C.
cinerea (B) peroxygenases. Haem cofactor also in liquorice and
selected residues as vdW spheres. The protein's active site access is
shown as a surface.
Fig. 5 Interaction energies vs. ligand distances from PELE simulations
for cholecalciferol (left) and ergocalciferol (right) entrance by the C25
end in the peroxygenases from A. aegerita (A, B) and C. cinerea (C, D).
The distances shown (Å) are between the reactive O atom in the haem
compound I and the calciferol C25 (for substrate numbering see Fig. 1).
Val77/Ala77, Phe121/Phe121, Thr192/Thre192, Ala195/Gly195,
Phe199/Phe199 and Glu196/Glu196 (Fig. S1†). It is worthy to
note that Phe69 also contributes to substrate positioning in
A. aegerita peroxygenase, but the homologous Met69 of the C.
cinerea enzyme (Fig. 3B) is placed away from cholecalciferol
and does not appear to have any effect on its position in the
active site. Ergocalciferol, however, with an extra methyl
group in position C24 and with a C22–C23 double bond (Fig.
1), is positioned in a slightly different manner in the active
site, and residues in position 69 (phenylalanine and methio-
nine) now interacts with the ligand (red arrows in Fig. S2†).
This extra constraint could be responsible for the lower reac-
tivity observed for D2 in both peroxygenases. The influence of
the side-chain structure (alkyl substituent and unsaturation
presence) on the activity of the A. aegerita and C. cinerea
peroxygenases was evidenced in the reaction of these
enzymes with different sterols (Table S1†).
Inspection of the haem entrance when both ligands are in
the active site reveals a better wrapping of the protein around
the ligands in C. cinerea peroxygenase, compared with the A.
aegerita enzyme. This is mainly due to the larger loop where
Gly323 is located, which is smaller in the A. aegerita
peroxygenase (Fig. 3A, black arrow). These are better illus-
trated in the haem-access channels of both peroxygenases in
Fig. 4, where the position of the above Gly323 is shown. The
narrower access to the haem in C. cinerea is also the result of
several hydrophobic amino acid substitutions. Replacements
to larger side chains in C. cinerea are dominant with: Ala73/
Ile73, Ala77/Val77, Gly195/Ala195, Gly241/Ala246 and Val244/
Leu249; and only one to a smaller amino acid, Ser240/Gly245.
Due to the larger entrance cavity of the A. aegerita
peroxygenase, both substrates remained solvent exposed at
the C2 end. In contrast, more favorable interactions are
established at the final substrate position inside the tighter
channel of the C. cinerea peroxygenase, which presents extra
interactions on the protein surface. The combination of the
surface interactions, along with a tighter haem cavity, result
in the improved interaction energies seen for D2 and D3 in
this protein.
A. aegerita peroxygenase and, for this reason, we observed a
constrained access to the protein (at around 12 Å). However,
the overall energy profile is more favorable, in particular for
cholecalciferol (Fig. 5C). Once inside the protein, the ligand
must surpass smaller barriers to reach the haem. From
Fig. 5, it is clearly seen that C. cinerea peroxygenase has the
most favorable energy profiles and well-defined minima in
the active site (with
C
25–O distance around
3
Å).
Fig. 5A and C show that the binding of cholecalciferol at the
haem site is more favorable in the C. cinerea peroxygenase
(−65 kcal mol−1) than in the A. aegerita enzyme (−40 kcal
mol−1). This difference comes from the fact that the first pro-
tein has a tighter binding pocket (see below). In the case of
ergocalciferol in Fig. 5B and D, the binding is also more
favorable in the C. cinerea (−42 kcal mol−1) than in the A.
aegerita enzyme (−30 kcal mol−1). These differences in the
energy profiles, which indicate a more favorable protein–
ligand interaction in C. cinerea, can explain the higher con-
version rates observed for the two compounds in this
peroxygenase.
Ligand binding. If we overlap cholecalciferol's positions
for the entire PELE simulations for both proteins, we find
two main orientations in the active site, as shown in
Fig. 6.
The two orientations in Fig. 6B correspond to the two min-
ima observed in cholecalciferol's diffusion in the C. cinerea
peroxygenase (Fig. 5C). One of the minima (in green) is
found at the binding distance to the haem O, and a second
(in blue), about 5 Å away from the haem oxygen (Fig. 5C).
The ligand's position observed in the C. cinerea peroxygenase
is also found in the A. aegerita enzyme (Fig. 6A), although
they do not correspond to the energy minima (Fig. 5A). Like-
wise, ergocalciferol can adopt two main positions in the bind-
ing pocket, although the minima (Fig. 5D) are less marked
than that found for cholecalciferol (Fig. 5C).
When cholecalciferol is in an optimal reacting position,
it is held in place by C. cinerea/A. aegerita peroxygenase
This journal is © The Royal Society of Chemistry 2016
Catal. Sci. Technol., 2016, 6, 288–295 | 291

View Article Online
Paper
Catalysis Science & Technology
Site selectivity. To investigate the different site selectivities
only 1.4 times superior to O–H24 while for the other systems,
it ranges between 10 to 20 times. This higher fraction of reac-
tive O–H24 distances can explain the observed formation of
C24 hydroxylated products in cholecalciferol. To sum up, and
in agreement with experimental results, from the relative fre-
quencies in C. cinerea peroxygenase, and also for
ergocalciferol in A. aegerita, we expect a completely regio-
selective reaction in C25. In contrast, for cholecalciferol in A.
aegerita enzyme, it seems that hydroxyl addition is possible
not only in C25 position, but also in other carbons (C24 and
C26/C27).
observed for D2 and D3 hydroxylation by A. aegerita and C.
cinerea peroxygenases, we have analysed the relative distance
distribution of the substrates' reactive hydrogen atoms in the
active site. We have considered as reactive those that can
approach the haem compound I oxygen atom close enough
to react. Thus, we have taken into account the hydrogen
atoms in positions C24, C25, C26 and C27 for both ligands, in
addition to C28 hydrogen atoms for ergocalciferol. We have
selected all structures (from the PELE simulations) where the
distance between H–O is below 5 Å and interaction energies
below −20 kcal mol−1 for D2 and −30 kcal mol−1 for D3. When
the relative frequency of these distances was computed
(Fig. 7), it can be seen that for C. cinerea, the O–H25 fre-
quency is dominant for both compounds. Similar conclu-
sions were obtained for both enzymes and substrates when
the distribution of angles of FeO–H36 was computed for
the different reactive positions in the cholecalciferol and
ergocalciferol side-chains (Fig. S3†).
Conclusions
Atomic level simulations have been used here to rationalize
the differences observed for ergocalciferol and cholecalcif-
erol's conversion rates and site selectivity in two
peroxygenases. The overall improved energy profiles in C.
cinerea, the presence of favorable minima, and a high frac-
tion of favorable O–H25 distances agrees well with experimen-
tally higher conversion rates.
The main structural differences between the A. aegerita
and C. cinerea peroxygenases that modify the access of chole-
calciferol to the activated haem are the changes Ala73/Ile73,
Ala77/Val77, Gly195/Ala195, Gly241/Ala246 and Val244/
The percentage of O–H25 distances below 3 Å in C. cinerea
is 54.5% for D3 and 36.2% for D2, whereas in A. aegerita it is
27.4% and 25.7%, respectively. Moreover, in the D3 simula-
tions in A. aegerita, the fraction of structures with O–H24
below 3 Å is 19.3%, which is quite high when compared to
the other cases. In fact, the fraction of O–H25 distances is
Fig. 7 Relative distance distributions of cholecalciferol (blue) and ergocalciferol (green) reactive hydrogen atoms to haem compound I oxygen
atom. Histograms for A. aegerita are in dark colours, and in light colours for C. cinerea.
292 | Catal. Sci. Technol., 2016, 6, 288–295
This journal is © The Royal Society of Chemistry 2016

View Article Online
Catalysis Science & Technology
Paper
Leu249, all to larger amino acids, with only Ser240/Gly245 to
a smaller one. These larger hydrophobic side chains augment
the interaction of D2 and D3 substrates with C. cinerea. The
better conversion rates observed for C. cinerea peroxygenase
do not originate in the active site itself (where Phe69 is
replaced by Met69), but in the ligand access to the active site
instead, especially in the entrance to the protein. In particu-
lar, the larger loop hosting Gly323 creates a barrier that
reduces the size of the entrance channel in the C. cinerea
peroxygenase. This, along with a tighter cavity, is reflected in
the more favorable interaction energies.
Although the improved reactivity of C. cinerea does not
appear to be affected by the larger Phe69 side chain, it does
hinder the entrance of ergocalciferol relative to cholecalcif-
erol. The presence of an extra methyl group that interacts
directly with position 69 in both proteins lowers the reactivity
of this compound. Finally, the computed relative frequency
of the distances between the activated haem oxygen and the
hydrogen atoms in C24 and C25 show that, despite the more
favourable hydroxylation on tertiary carbons, the reaction at
cholecalciferol C24 occurs when the ratio of favourable O–
H25/O–H24 distances decreases.
structure 4), was also used as standard for GC-MS analyses.
All compounds were from Sigma-Aldrich.
Enzymatic reactions
Reactions of cholecalciferol and ergocalciferol (0.1 and 0.5
mM) with the A. aegerita and C. cinerea peroxygenases (1 U)
were performed in 5 mL of 50 mM sodium phosphate, pH 7,
at 40 °C for 60 min, in the presence of 0.5 mM H2O2. The
substrates were previously dissolved in acetone, and added to
the buffer (the acetone concentration in the reaction was
40%). In control experiments, the substrates were treated
under the same conditions (including H2O2) but without
enzyme. After the enzymatic reactions, the products were
recovered by liquid–liquid extraction with methyl tert-butyl
ether, dried under N2, and redissolved in chloroform for GC-
MS analyses. BisIJtrimethylsilyl)trifluoroacetamide (Supelco) in
the presence of pyridine was used to prepare trimethylsilyl
(TMS) derivatives. An internal standard was added after the
enzymatic reactions to determine product yields.
GC-MS analyses
GC-MS analyses were performed with a Shimadzu QP2010
Ultra equipment, using a fused-silica DB-5HT capillary col-
umn (30 m × 0.25 mm internal diameter, 0.1 μm film thick-
ness) from J&W Scientific.16 The oven was heated from 120
°C (1 min) to 300 °C (15 min) at 5 °C min−1. The injection
was performed at 300 °C, the transfer line was kept at 300
°C, and helium was used as carrier gas. Compounds were
identified by mass fragmentography and by comparing their
mass spectra with standards, and quantitation was obtained
from total-ion peak area using molar response factors
obtained from cholecalciferol, ergocalciferol, 25-hydroxy-
cholecalciferol and 25-hydroxyergocalciferol standards. The
two latter compounds were also used as external standards
for calculation of product yields.
Materials and methods
Enzymes and chemicals
Peroxygenase (isoform II) was isolated from A. aegerita DSM
22459 grown in soybean medium using a combination of SP-
Sepharose chromatography and Mono-P chromatofocusing.45
A. aegerita DSM 22459 is deposited at the Deutsche
Stammsammlung für Mikroorganismen und Zellkulturen Braun-
schweig (Germany). C. cinerea peroxygenase was provided by
Novozymes A/S (Bagsvaerd, Denmark). The enzyme corre-
sponds to the protein model 7249 from the sequenced C.
cinerea genome available at the JGI (http://genome.jgi.doe.
gov/Copci1), which was expressed in Aspergillus oryzae and
purified using
a combination of S-Sepharose and SP-
PELE and other computational analyses
Sepharose ion-exchange chromatography (patent WO/2008/
119780). One peroxygenase unit is defined as the amount of
enzyme oxidizing 1 μmol of veratryl alcohol to veratraldehyde
(ε310 9300 M−1 cm−1) in 1 min at 24 °C, pH 7, after addition
of 0.5 mM H2O2 in the C. cinerea peroxygenase reactions and
2.5 mM H2O2 in those with the A. aegerita peroxygenase.
Vitamin D3, also known as calciol or cholecalciferol
The starting structures for PELE simulations were the A.
aegerita peroxygenase crystal (2YOR)34 and a homology model
for the C. cinerea peroxygenase structure obtained using
2YOR as the template.8 As the optimum pH for peroxygenase
activity is 7, the structures were prepared accordingly using
Schrodinger's Protein Preparation Wizard35 and H++ web
server.1 Histidines were δ-protonated, with the exception of
His82 (ε-protonated) and His118 and His251 (double proton-
ated). All acidic residues were deprotonated, except for Asp85
which was kept in its protonated state. The ergocalciferol and
cholecalciferol molecules were optimized with Jaguar37 at the
DFT/M06 level with the 6-31G** basis and a PBF implicit sol-
vent in order to obtain their electrostatic potential atomic
charges. Finally, the haem site was modeled as thiolate-
ligated compound I after being fully optimized in the protein
environment with quantum mechanics/molecular mechanics
(QM/MM) using QSite.38 The electronic calculations show
three unpaired electrons: two located on the oxoiron group
IJ(5Z,7E)-IJ3S)-9,10-seco-5,7,10IJ19)-cholestatrien-3-ol;
Fig.
1,
structure 1), was tested as substrate of the A. aegerita and C.
cinerea peroxygenases. 25-Hydroxyvitamin D3, also known as
calcidiol or 25-hydroxycholecalciferol IJ(5Z,7E)-IJ3S)-9,10-seco-
5,7,10IJ19)-cholestatriene-3,25-diol; Fig. 1 structure 2), was
used as standard for gas chromatography-mass spectrometry
(GC-MS) analyses. Vitamin D2, also known as ercalciol or
ergocalciferol IJ(5Z,7E,22E)-IJ3S)-9,10-seco-5,7,10IJ19),22-
ergostatetraen-3-ol; Fig. 1, structure 3), was also tested as sub-
strate of the two peroxygenases. 25-Hydroxyvitamin D2, also
known as ercalcidiol or 25-hydroxyergocalciferol IJ(5Z,7E,22E)-
IJ3S)-9,10-seco-5,7,10IJ19),22-ergostatetraene-3,25-diol; Fig. 1,
This journal is © The Royal Society of Chemistry 2016
Catal. Sci. Technol., 2016, 6, 288–295 | 293

View Article Online
Paper
Catalysis Science & Technology
and
a
third on the heme and less than 1% spin
8 L. Bordoli, F. Kiefer, K. Arnold, P. Benkert, J. Battey and T.
Schwede, Nat. Protoc., 2009, 4, 1–13.
contamination.
Once the initial protein structure was prepared and
ligands optimized, these were placed manually in identical
positions at the entrance of the haem-access channel and
PELE simulations were performed.10 PELE is a Monte Carlo-
based algorithm that produces new configurations through a
sequential ligand and protein perturbation, side chain pre-
diction and minimization steps, freely available at https://
pele.bsc.es. New configurations are then filtered with a
Metropolis acceptance test, where the energy is described
with an all-atom OPLS force field22 and a surface generalized
Born solvent.7 In this way, it is possible to locate and charac-
terize local and global minima structures for the most favor-
able protein–ligand interactions. PELE has been successfully
used in a number of ligand migration studies with both
small and large substrates.18,29,30
Distance and angle distributions were computed after
screening all structures from PELE simulations by using the
interaction energy and distance to the compound I oxygen
atom. The structures which showed interaction energies
below −20 kcal mol−1 for ergocalciferol and −30 kcal mol−1
for cholecalciferol were selected. Reactive hydrogen atoms
were considered to be below 5 Å from the compound I oxygen
atom for the distance analysis and 3Å for the angle study
(angles are much more sensitive and thus only near attack
conformations were selected). All the selected structures were
then used to compute the relative frequencies of the ligand's
hydrogen atoms at reactive distances to the haem oxygen and
FeO–H and O–H–C angles.
9 S. Bormann, A. G. Baraibar, Y. Ni, D. Holtmann and F.
Hollmann, Catal. Sci. Technol., 2015, 5, 2038–2052.
10 K. W. Borrelli, A. Vitalis, R. Alcantara and V. Guallar,
J. Chem. Theory Comput., 2005, 1, 1304–1311.
11 N. R. Buck, W. Claerhout, B. H. Leuenberger, E. Stoecklin,
K. Urban and S. Wolfram, US Pat., US20130210782 A1,
2013.
12 B. P. Cossins, A. Hosseini and V. Guallar, J. Chem. Theory
Comput., 2012, 8, 959–965.
13 D. Floudas, M. Binder, R. Riley, K. Barry, R. A. Blanchette, B.
Henrissat, A. T. Martínez, R. Otillar, J. W. Spatafora, J. S.
Yadav, A. Aerts, I. Benoit, A. Boyd, A. Carlson, A. Copeland,
P. M. Coutinho, R. P. de Vries, P. Ferreira, K. Findley, B.
Foster, J. Gaskell, D. Glotzer, P. Górecki, J. Heitman, C.
Hesse, C. Hori, K. Igarashi, J. A. Jurgens, N. Kallen, P.
Kersten, A. Kohler, U. Kües, T. K. A. Kumar, A. Kuo, K.
LaButti, L. F. Larrondo, E. Lindquist, A. Ling, V. Lombard, S.
Lucas, T. Lundell, R. Martin, D. J. McLaughlin, I.
Morgenstern, E. Morin, C. Murat, M. Nolan, R. A. Ohm, A.
Patyshakuliyeva, A. Rokas, F. J. Ruiz-Dueñas, G. Sabat, A.
Salamov, M. Samejima, J. Schmutz, J. C. Slot, F. St.John, J.
Stenlid, H. Sun, S. Sun, K. Syed, A. Tsang, A. Wiebenga, D.
Young, A. Pisabarro, D. C. Eastwood, F. Martin, D. Cullen,
I. V. Grigoriev and D. S. Hibbett, Science, 2012, 336,
1715–1719.
14 C. A. Fritts and P. W. Waldroup, J. Appl. Poult. Res., 2003, 12,
45–52.
15 A. Gutiérrez, E. D. Babot, R. Ullrich, M. Hofrichter, A. T.
Martínez and J. C. del Río, Arch. Biochem. Biophys.,
2011, 514, 33–43.
16 A. Gutiérrez, J. C. del Río, F. J. González-Vila and F. Martín,
J. Chromatogr., 1998, 823, 449–455.
17 J. M. Hernández, US Pat., US20130137662 A1, 2013.
18 A. Hernández-Ortega, F. Lucas, P. Ferreira, M. Medina, V.
Guallar and A. T. Martínez, J. Biol. Chem., 2011, 286,
41105–41114.
Acknowledgements
This work was supported by the INDOX IJKBBE-2013-7-613549)
and PELE (ERC-2009-Adg 25027) EU projects, and by the
BIO2011-26694 and CTQ2013-48287 projects of the Spanish
Ministry of Economy and Competitiveness.
19 M. Hofrichter and R. Ullrich, Curr. Opin. Chem. Biol.,
2014, 19, 116–125.
Notes and references
1 R. Anandakrishnan, B. Aguilar and A. V. Onufriev, Nucleic
Acids Res., 2012, 40, W537–W541.
2 E. Aranda, M. Kinne, M. Kluge, R. Ullrich and M. Hofrichter,
Appl. Microbiol. Biotechnol., 2009, 82, 1057–1066.
3 E. Aranda, R. Ullrich and M. Hofrichter, Biodegradation,
2010, 21, 267–281.
4 E. D. Babot, J. C. del Río, M. Cañellas, F. Sancho, F. Lucas,
V. Guallar, L. Kalum, H. Lund, G. Gröbe, K. Scheibner, R.
Ullrich, M. Hofrichter, A. T. Martínez and A. Gutiérrez, Appl.
Environ. Microbiol., 2015, 81, 4130–4142.
5 E. D. Babot, J. C. del Río, L. Kalum, A. T. Martínez and A.
Gutiérrez, Biotechnol. Bioeng., 2013, 110, 2332.
20 G. Jean, J. C. Terrat, T. Vanel, J. M. Hurot, C. Lorriaux, B.
Mayor and C. Chazot, Nephrol., Dial., Transplant., 2008, 23,
3670–3676.
21 G. Jones, Annu. Rev. Nutr., 2013, 33, 23–44.
22 G. A. Kaminski, R. A. Friesner, J. Tirado-Rives and W. L.
Jorgensen, J. Phys. Chem. B, 2001, 105, 6474–6487.
23 S. Käppeli, E. Frohlich, S. G. Gebhardt-Henrich, A. Pfulg, H.
Schäublin, R. Zweifel, H. Wiedmer and H. H. Stoffel, Arch.
Gefluegelkd., 2011, 75, 179–184.
24 A. Karich, M. Kluge, R. Ullrich and M. Hofrichter, AMB
Express, 2013, 3, 5.
25 M. Kinne, C. Zeisig, R. Ullrich, G. Kayser, K. E. Hammel and
M. Hofrichter, Biochem. Biophys. Res. Commun., 2010, 397,
18–21.
6 E. D. Babot, J. C. del Río, L. Kalum, A. T. Martínez and A.
Gutiérrez, ChemCatChem, 2015, 7, 283–290.
7 D. Bashford and D. A. Case, Annu. Rev. Phys. Chem.,
2000, 51, 129–152.
26 M. Kluge, R. Ullrich, C. Dolge, K. Scheibner and M.
Hofrichter, Appl. Microbiol. Biotechnol., 2009, 81, 1071–1076.
294 | Catal. Sci. Technol., 2016, 6, 288–295
This journal is © The Royal Society of Chemistry 2016

View Article Online
Catalysis Science & Technology
Paper
27 M. Kluge, R. Ullrich, K. Scheibner and M. Hofrichter, Green
Chem., 2012, 14, 440–446.
28 G. A. Leichtmann, J. M. Bengoa, M. J. G. Bolt and M. D.
Sitrin, Am. J. Clin. Nutr., 1991, 54, 548–552.
35 G. M. Sastry, M. Adzhigirey, T. Day, R. Annabhimoju and
W. Sherman, J. Comput.-Aided Mol. Des., 2013, 27, 221–234.
36 J. C. Schöneboom, S. Cohen, H. Lin, S. Shaik and W. Thiel,
J. Am. Chem. Soc., 2004, 126, 4017–4034.
29 D. Linde, R. Pogni, M. Cañellas, F. Lucas, V. Guallar, M. C.
Baratto, A. Sinicropi, V. Sáez-Jiménez, C. Coscolín, A.
Romero, F. J. Medrano, F. J. Ruiz-Dueñas and A. T. Martínez,
Biochem. J., 2015, 466, 253–262.
30 M. F. Lucas and V. Guallar, Biophys. J., 2012, 102, 887–896.
31 H. L. J. Makin and D. B. Gower, Steroid analysis, Springer,
NY, 2010.
37 Schrödinger. Jaguar 7.8, LCC, New York, 2011.
38 Schrödinger. QSite 5.7, LCC, New York, 2011.
39 Schrödinger. SiteMap 2.5, LCC, New York, 2011.
40 C. Simoes-Nunes and G. M. Weber, EP Pat., EP1516540 B1,
2004.
41 M. D. Sitrin and J. M. Bengoa, Am. J. Clin. Nutr., 1987, 46,
1011–1015.
32 M. Michalczuk, D. Pietrzak, J. Niemiec and J. Mroczk, Pol. J.
Food Nutr. Sci., 2010, 60, 121–126.
42 R. Ullrich, C. Dolge, M. Kluge and M. Hofrichter, FEBS Lett.,
2008, 582, 4100–4106.
33 S. Peter, M. Kinne, X. Wang, R. Ulrich, G. Kayser, J. T.
Groves and M. Hofrichter, FEBS J., 2011, 278, 3667–3675.
34 K. Piontek, E. Strittmatter, R. Ullrich, G. Grobe, M. J. Pecyna,
M. Kluge, K. Scheibner, M. Hofrichter and D. A. Plattner,
J. Biol. Chem., 2013, 288, 34767–34776.
43 R. Ullrich and M. Hofrichter, FEBS Lett., 2005, 579, 6247–6250.
44 R. Ullrich and M. Hofrichter, Cell. Mol. Life Sci., 2007, 64,
271–293.
45 R. Ullrich, J. Nuske, K. Scheibner, J. Spantzel and M.
Hofrichter, Appl. Environ. Microbiol., 2004, 70, 4575–4581.
This journal is © The Royal Society of Chemistry 2016
Catal. Sci. Technol., 2016, 6, 288–295 | 295