Structure-Guided Evolution of Aryl Alcohol Oxidase from Pleurotus eryngii for the Selective Oxidation of Secondary Benzyl Alcohols

Article abstract of DOI:10.1002/adsc.201900134

Aryl alcohol oxidase (AAO) is a fungal flavoenzyme capable of oxidizing aromatic primary alcohols into their correspondent aldehydes through a stereoselective hydride abstraction. Unfortunately, this enzyme does not act on secondary benzyl alcohols in racemic mixtures due to the strict control of substrate diffusion and positioning at the active site restricted to primary benzyl alcohols. Here we describe the engineering of AAO from Pleurotus eryngii to oxidize chiral benzyl alcohols with high enantioselectivity. The secondary benzyl alcohol oxidase was remodeled at the active site through four cycles of structure-guided evolution, including a final step of in vivo site-directed recombination to address the positive epistatic interactions between mutations. The final variant, with five substitutions and a renovated active site, was characterized at biochemical and computational level. The mutational sculpting helped position the bulkier (S)-1-(p-methoxyphenyl)-ethanol, improving the mutant's catalytic efficiency by three orders of magnitude relative to the native enzyme while showing a high enantioselectivity (ee >99%). As a promising candidate for racemic resolution, this evolved secondary benzyl alcohol oxidase maintained its natural stereoselective mechanism while displaying activity on several secondary benzyl alcohols. (Figure presented.).

Full text of DOI:10.1002/adsc.201900134

Title: Structure-Guided Evolution of Aryl Alcohol Oxidase from
Pleurotus eryngii for the Selective Oxidation of Secondary Benzyl
Alcohols
Authors: Javier Viña-Gonzalez, Diego Jimenez-Lalana, Ferran Sancho,
Ana Serrano, Angel T Martinez, Victor Guallar, and Miguel
Alcalde
This manuscript has been accepted after peer review and appears as an
Accepted Article online prior to editing, proofing, and formal publication
of the final Version of Record (VoR). This work is currently citable by
using the Digital Object Identifier (DOI) given below. The VoR will be
published online in Early View as soon as possible and may be different
to this Accepted Article as a result of editing. Readers should obtain
the VoR from the journal website shown below when it is published
to ensure accuracy of information. The authors are responsible for the
content of this Accepted Article.
To be cited as: Adv. Synth. Catal. 10.1002/adsc.201900134
Link to VoR: http://dx.doi.org/10.1002/adsc.201900134

10.1002/adsc.201900134
Advanced Synthesis & Catalysis
FULL PAPER
DOI: 10.1002/adsc.201((will be filled in by the editorial staff))
Structure-Guided Evolution of Aryl Alcohol Oxidase from
Pleurotus eryngii for the Selective Oxidation of Secondary
Benzyl Alcohols
Javier Viña-Gonzaleza, Diego Jimenez-Lalanaa, Ferran Sanchob, Ana Serranoc, Angel T.
Martinezc, Victor Guallarb,d and Miguel Alcaldea*
a
Department of Biocatalysis, Institute of Catalysis, CSIC, Cantoblanco, 28049 Madrid, Spain.
Fax: (+31)- 91 5854760; phone: (+34)- 91 5854806; e-mail:malcalde@icp.csic.es
Barcelona Supercomputing Center, Jordi Girona 31, 08034 Barcelona, Spain.
Biological Research Center, CSIC, Ramiro de Maeztu 9, 28040 Madrid, Spain.
ICREA, Passeig Lluís Companys 23, 08010 Barcelona, Spain.
b
c
d
Received: ((will be filled in by the editorial staff))
Supporting information for this article is available on the WWW under http://dx.doi.org/10.1002/adsc.201######.((Please
delete if not appropriate))
Abstract. Aryl alcohol oxidase (AAO) is a fungal active site, was characterized at biochemical and
flavoenzyme capable of oxidizing aromatic primary alcohols computational level. The mutational sculpting helped
into their correspondent aldehydes through a stereoselective position the bulkier (S)-1-(p-methoxyphenyl)-ethanol,
hydride abstraction. Unfortunately, this enzyme does not act improving the mutant´s catalytic efficiency by three orders
on secondary benzyl alcohols in racemic mixtures due to the of magnitude relative to the native enzyme while showing a
strict control of substrate diffusion and positioning at the high enantioselectivity (ee > 99%). As a promising
active site restricted to primary benzyl alcohols. Here we candidate for racemic resolution, this evolved secondary
describe the engineering of AAO from Pleurotus eryngii to benzyl alcohol oxidase maintained its natural
oxidize chiral benzyl alcohols with high enantioselectivity. stereoselective mechanism while displaying activity on
The secondary benzyl alcohol oxidase was remodeled at the several secondary benzyl alcohols.
active site through four cycles of structure-guided evolution, Keywords: aryl alcohol oxidase; secondary benzyl
including a final step of in vivo site-directed recombination alcohols; Saccharomyces cerevisiae; directed evolution.
to address the positive epistatic interactions between
mutations. The final variant, with five substitutions and a
renovated
cyclic oxidation-reduction sequences, after the
selective oxidation of one of the enantiomers, the
Introduction
With global annual sales of $ 1 trillion, an
achiral intermediate (ketone or imine) is non-
increasingly important challenge in drug development
selectively reduced to the racemic initial material.
that the pharmaceutical sector must overcome is that
posed by chiral chemistry.[1] Indeed, enantiopure
building blocks are in strong demand to produce
drugs with particular biological activities, while they
are also paramount for the production of fine
chemicals.[2] Biocatalysis has found an important
niche in the field of chiral technology, with
enzymatic and whole-cell biotransformations offering
stereo-, regio- and chemio-selectivity under mild
reaction conditions. Classically, two fundamental
approaches are followed in the industrial production
of chiral molecules: enantioselective synthesis and
racemic mixture separation. While asymmetrical
synthesis involves complex and expensive
processes,[3] dynamic kinetic resolution (DKR) is
currently one of the most efficient sources of chiral
molecules, in which separation is coupled to the in
situ re-racemization of one of the enantiomers.
Another recently described transformative source of
enantiomers is cyclic de-racemization. Based on
After a certain number of cycles, a theoretical 100%
yield of the non-reactive enantiomer can be
accumulated.[4] The core step in racemic separation,
kinetic resolution, has been achieved for secondary
alcohols using either whole-cell systems[5] or isolated
enzymes, such as alcohol dehydrogenases,[6] lipases[7]
and particularly, oxidases.[8] This latter solution is the
simplest means to prepare optically pure secondary
benzyl alcohols, given their trivial requirements (only
needing
O2
from
the
air)
and
high
enantioselectivity.[8a,9]
Among these enzymes, aryl alcohol oxidase (AAO,
EC 1.1.3.7) is a promising candidate for the
enantioselective oxidation of chiral benzyl alcohols.
AAO is a flavoenzyme belonging to the GMC
(glucose-methanol-choline)
superfamily
of
oxidoreductases and it is naturally secreted as a part
of the fungal enzymatic consortium involved in lignin
degradation.[10] Accepting a variety of aromatic
1
This article is protected by copyright. All rights reserved.

10.1002/adsc.201900134
Advanced Synthesis & Catalysis
alcohols as substrates, the activity of AAO is initially
divided into two half-reactions. The first reductive
half-reaction involves highly enantioselective hydride
transfer from the alcohol’s α-C to the FAD co-factor.
This process yields the corresponding aldehyde, such
that the FAD can then be reoxidized by O2, releasing
H2O2 as a by-product of the second oxidative half of
the reaction.[11] Unfortunately, the active site of AAO
is buried under a hydrophobic constriction formed by
residues Tyr92, Phe397 and Phe501. As a result,
substrates bulkier than primary aromatic alcohols
cannot easily be accommodated, reducing the
enzyme’s activity on chiral molecules to a residual
trace. Until recently the failure to functionally
express AAO in a heterologous host had prevented its
directed evolution. However, fusing the enzyme to a
chimeric signal prepro-leader has enabled this protein
to be successfully evolved for secretion by yeast.[12]
Taking advantage of this expression system, here
we have combined different laboratory evolution
strategies to unlock the activity of AAO on secondary
aromatic alcohols. We first carried out a carefully
structure-guided campaign of evolution using chiral
1-(p-methoxyphenyl)-ethanol as the substrate,
thereby generating a palette of secondary benzyl
alcohol oxidase mutants. Employing in vivo site-
directed recombination approaches, mutations were
curated by comparing them with their correspondent
parental reversions. The differential oxidation of
secondary benzyl alcohols by the final benzyl alcohol
oxidase variant was characterized, while the rationale
behind these changes was analyzed computationally
at the atomic level.
As a substrate for the screening assay we chose 1-
(p-methoxyphenyl)-ethanol, a chiral molecule with
similar structure to the natural p-methoxybenzyl
alcohol substrate, the oxidation of which by AAO can
be rapidly assessed in a coupled Amplex Red/HRP
assay (see Experimental Section for details). As the
activity of AAO on secondary alcohols is irrelevant,
no response was detected with 1-(p-methoxyphenyl)-
ethanol when the parental FX9 was screened in
microtiter fermentations of the supernatant (i.e.
cultures in 96 well plates). As indicated previously,
AAO’s failure to oxidize secondary aromatic alcohols
is due to steric perturbation of the residues forming
the catalytic cavity when trying to accommodate
bulkier, chiral molecules. Specifically, Phe501 is
thought to be a steric liability at the active site,
particularly given the very weak but detectable
activity on 1-(p-methoxyphenyl)-ethanol of a F501A
variant expressed in E. coli after in vitro refolding.[21]
A more recent rational attempt to achieve secondary
alcohol oxidation based on PELE (Protein Energy
Landscape Exploration) suggested Ile500 was another
possible obstacle for ligand diffusion. Indeed, the
I500A substitution conferred better transit of the
substrate to the active site due to channel
broadening.[22]
To ensure that the activity on secondary benzyl
alcohol could be measured at the start of the
laboratory evolution campaign, we first prepared
three mutants from the FX9 secretion variant: I500A,
F501A and 500A-F501A (Figure S2 A, B). When
isolated from yeast supernatants these variants did not
appear to act on 1-(p-methoxyphenyl)-ethanol. Hence,
we performed combinatorial saturation mutagenesis
of the Ile500-Phe501 residues and we found several
clones with activity on 1-(p-methoxyphenyl)-ethanol,
which were scaled up to a 100 mL flask to estimate
the overall improvement in activity. Of these, the
I500Q-F501W, I500L-F501I, I500M-F501V and
I500M-F501W mutants presented a 5-, 15-, 30- and
160-fold enhancement in activity relative to the
parental FX9, respectively. The activity of the
I500M-F501W mutant (named 15G12) was further
tested against a panel of primary and secondary
alcohols (Table 1). Its specific activity on primary
alcohols was dramatically reduced depending on the
chemical nature of the molecule, with practically no
activity on p-methoxybenzyl alcohol (c.a. 0.6% of
that of the parental FX9 variant). By contrast, the
activity of the 15G12 variant on secondary aromatic
alcohols rose from undetectable levels, to weak yet
evident activity on 1-phenyl ethanol and 4-fluoro-α-
methylbenzyl alcohol, and up to 0.3 U/mg on 1-(p-
methoxyphenyl)-ethanol (Table 1). Given that
hydride abstraction of p-methoxybenzyl alcohol to
the flavin of AAO occurs exclusively from the pro-R
position, in deracemization reactions of secondary
Results and Discussion
Laboratory evolution
First generation: Unlocking activity for secondary
aromatic alcohols
The departure point of this study was a secretion
mutant of the AAO from Pleurotus eryngii named
FX9. This mutant is the product of several rounds of
directed evolution aimed to promote functional
expression in Saccharomyces cerevisiae.[12] In this
FX9 variant, the AAO is fused to a chimeric prepro-
leader (preα-factor-proKiller) that enhanced secretion
by introducing the F[3α]S, N[25proK]D, T[50proK]A
and F[52proK]L mutations into the pre-α and pro-
Killer leaders. In addition, two substitutions were
included in the mature AAO: L170M in an α-helix
situated at the protein surface, and the consensus,
ancestral mutation H91N in the FAD attachment loop.
These latter mutations enhanced stability and
improved production by S. cerevisiae to 4.5 mg/L and
by Pichia pastoris in bioreactor to 25.5 mg/L, while
conserving the general biochemical features of the
native AAO.[12]
2
This article is protected by copyright. All rights reserved.

10.1002/adsc.201900134
Advanced Synthesis & Catalysis
alcohols R-hydrogen abstraction should produce S
enantioselective oxidation of the alcohol to the
corresponding
ketone.
Table 1. Specific activity of AAO variants with primary and secondary alcohols.
Specific activities were estimated in 100 mM phosphate buffer pH 6.0 containing 5 mM of each alcohol. Each reaction was performed
by triplicate and substrate conversion was followed by measuring the absorption at 563nm (ε563 = 56000 M-1 cm-1) using the
HRP/Amplex red coupled assay as described in the experimental section.
To confirm this, reactions were performed with the
optically pure (R) and (S)-4-fluoro-α-methylbenzyl
alcohol enantiomers, and as expected, only activity
on the latter was detected.[23]
Paradoxically, while the initial search for a wider
space for secondary aromatic alcohol accommodation
focused on the I500A-F501A mutations, the bulky
alcohol would appear to be much better oxidized in a
narrower catalytic pocket following the substitution
of Ile500 and Phe501 by the more expansive Met and
Trp, respectively (Figure S2 C, D). Thus, these two
mutations at the active site reposition the secondary
alcohol, favouring catalysis, as confirmed by
computational analysis (see below).
Second and third generations: Searching for new
beneficial mutations
After unlocking the activity on secondary benzyl
alcohols, different protein segments of the 15G12
mutant were targeted for random mutagenesis and
DNA recombination by MORPHING, with a view to
further optimize its activity on secondary alcohols.
This focused structural evolution tool allows the
protein to be divided into defined mutagenic areas,
each of which can be interrogated in conjunction with
the in vivo recombination of the different DNA
fragments in S. cerevisiae.[14] This approach has
already been successfully applied during the
evolution of AAO towards functional expression in
yeast[12] and on this occasion, the design involved the
study of three protein blocks in independent libraries
(Figure 1A, B). The first of these was the MA block
at the N-terminus (Leu48-Thr100), which is
associated with the access channel and the FAD-
binding domain, and that contains important
determinants like Pro79, Asn91, Tyr92 and Val90.
Figure 1. MORPHING fragments for focused evolution.
(A) Front and back views of AAO, with the MA, MB and
MC MORPHING blocks shown in purple, red and blue,
respectively. (B) The dark grey box corresponds to the
signal peptide, and the three mutagenic fragments
considered for MORPHING are shown as purple, red and
blue boxes. For the first MORPHING library, mutagenic
block MA was in vivo assembled with high-fidelity
fragments HA1 and HA2. For the second library,
3
This article is protected by copyright. All rights reserved.

10.1002/adsc.201900134
Advanced Synthesis & Catalysis
mutagenic fragment MB was recombined from high-
fidelity fragments HB1 and HB2. The third library was
constructed with the assembly of mutagenic block MC and
high-fidelity fragment HC. (C) The catalytic pocket of
AAO with the contribution of the MORPHING blocks MA,
MB and MC (purple, red and blue, respectively). Model
prepared with the crystal structure of the AAO from P.
eryngii (PDB 3FIM).
4
This article is protected by copyright. All rights reserved.

10.1002/adsc.201900134
Advanced Synthesis & Catalysis
Figure 2. Laboratory evolution of AAO for the oxidation of secondary benzyl alcohols. New mutations are represented as
stars and accumulated mutations as squares. The chimeric prepro-leader is depicted in grey and the mature AAO in blue.
The TAI (total activity improvement) refers to the fold improvement of AAO activity with 1-(p-methoxyphenyl)-ethanol
as a substrate and it was estimated relative to the FX9 parental type from S. cerevisiae supernatants: *Mutants from the
MB library; **Mutants from the MA library; ***Mutant from the ep-PCR library.
5
This article is protected by copyright. All rights reserved.

10.1002/adsc.201900134
Advanced Synthesis & Catalysis
The second MB block (Leu310-Ile417) covers the
wall of the catalytic pocket and it contains the
aromatic Phe397, a residue implicated in substrate
positioning and product release. The third MC block
(Glu490-Gln566) is situated in the C-terminal region,
and it contains the catalytic His502 and several amino
acids related to substrate positioning (Met500,
Trp501 and His546). Together, these three blocks
encompass the complete active site and the aromatic
bottleneck formed by Tyr92, Phe397 and Trp501
(Figure 1C).[16,23,24]
We carried out the three independent libraries of
MORPHING; besides, we prepared a conventional
error-prone PCR (ep-PCR) mutagenic library that
targeted the whole AAO gene (in total over 4,000
clones were screened). From the pool of libraries,
seven mutants with stronger activity on the secondary
1-(p-methoxyphenyl)-ethanol were selected for
further analysis, ranging from a 1.4 to 2.2-fold
enhancement over 15G12 (Figure 2). The most
successful MORPHING corresponded to the MA
block, with five of the seven improved AAO variants
from this library. Interestingly, all the mutations
identified were located within a 14 amino acid span,
from Ile76 to Val90. The mutations carried by the
best variant (3F10) were I76V and M83I, which in
conjunction with the rest of the substitutions in this
stretch (A77V, R80C, I76V and V90A) highlight the
importance of the access channel in modulating
oxidative activity. Indeed, it should be noted that the
R80C substitution was also found in the
After careful evaluation of the mutations obtained in
the second and third generations (Table 2), we
decided to undertake a final round of evolution to
assess whether there were any positive epistatic
effects among the mutations. We constructed a
combinatorial library by in vivo site-directed
recombination, such that the 10 mutations and their
corresponding reversions could be rapidly combined
in an one-pot transformation reaction, evaluating the
library in order to obtain the optimal combination of
substitutions for the oxidation of chiral alcohols
(Figure S1). From this ensemble of mutations, the
three best variants identified shared the same
backbone of substitutions: A77V-R80C-V340A. The
third best variant was the 24H3 mutant that carried
the
A77V-R80C-Q174R-F332L-V340A-Q466R
mutations, displaying activity more than 420-fold
better than the FX9 parental type. In the case of 4D3,
the only difference from the 24H3 variant was the
absence of the Q174R mutation, which translated into
a 480-fold increase in activity, highlighting a
detrimental effect of Q174R within this mutational
context. The further purging of Q466R and F332L
gave rise to the LanDo variant that carried the A77V-
R80C-V340A mutations in conjunction with I500M-
F501W and the 6 secretion mutations of FX9, this
variant representing the best performer with a total
Mutation
Variant
Library
Secondary
motif
triple 6G3 mutant from the whole gene ep-PCR
library, together with two superficial substitutions:
E39G and Q466R (Figure 2). The MORPHING
method was also successful in fragment MB, where
the 12D12 variant presented two mutations at the
surface: F332L and V340A. Applying focused
mutagenesis to the area corresponding to the catalytic
pocket, the MC block, was ineffective since variants
with improvements were not detected.
To further enhance secondary alcohol oxidation,
the best variants from each library were submitted to
ep-PCR and in vivo shuffling in S. cerevisiae: 3F10
and 11H2 from the MA library, the 12D2 variant
from the MB library, and the 6G3 variant obtained
through whole-gene mutagenic amplification. From
this third generation, the 3C11 variant was seen to
enhance the activity roughly 1.2-fold compared to
3F10 (412-fold relative to the FX9 parental type),
retaining the I76V mutation from 3F10 and acquiring
the V90A change, also previously detected in the 9F2
variant, as well as incorporating the new Q174R
substitution (Figure 2).
GAGE39GGGG
ATTI76VGTT
6G3
ep-PCR
MA
Loop
Loop
Loop
3F10, 4D8
GCGA77VGTG
CGCR80CTGC
11H2
6G3, 7B4
3F10
MA
ep-PCR,
MA
Loop
Loop
ATGM83IATA
GTTV90AGCT
MA
MA,
mutagenic
shuffling
9F2
Loop
Mutagenic
Shuffling
CAAQ174RCGA
TTCF332LCTC
3C11
Alpha helix
Alpha helix
12D2
MB
GTTV340AGCT
CAAQ466RCGA
12D2
6G3
MB
Alpha helix
Loop
Fourth generation: Mutational polishing by in vivo
site-directed recombination
ep-PCR
6
This article is protected by copyright. All rights reserved.

10.1002/adsc.201900134
Advanced Synthesis & Catalysis
800-fold enhancement of activity relative to the
parental type (Figures 2, 3).
two hour reaction. The configuration of the remaining
alcohol in the reaction was confirmed by optical
rotation (Figure S3, Table S2), the positive rotation
corresponding to the R enantiomer meaning the
natural oxidation of the S enantiomer by AAO was
maintained after evolution.[23] Despite the remarkably
specific activity for the secondary 1-(p-
methoxyphenyl)-ethanol (2.9 U/mg), the five new
mutations carried by the LanDo variant did not
negatively affect its secretion (4.6 mg/L). The activity
of LanDo with secondary alcohols was tested against
available commercial secondary (aromatic) alcohols
that were representative of the structural scope of the
AAO. The initial turnover rates of the LanDo variant
relative to the wildtype AAO (wtAAO, heterologous
expressed in E. Coli after in vitro refolding) increased
30, 20 and 100-fold times for 1-(p-methoxyphenyl)-
ethanol, p-fluoro-α-methylbenzyl alcohol and 1-
phenylethanol, respectively. Indeed, the LanDo
variant even showed activity on 1-phenylpropanol, a
substrate not oxidized by wtAAO (Table 3). The
kinetic parameters were measured under air-saturated
conditions for 1-(p-methoxyphenyl)-ethanol, and for
the primary alcohols p-methoxybenzyl alcohol and
2,4-hexadien-1-ol. LanDo displayed an outstanding
Table 2. Selected mutations for site-directed
recombination.
increase
in
the
catalytic
efficiency
for
enantioselective oxidation of 1-(p-methoxyphenyl)-
ethanol, three orders of magnitude. Interestingly, the
activity on primary alcohols that was dramatically
reduced after inserting the I500M-F501W pair in the
first cycle of evolution in 15G12 variant (Table 1,
Figure 2), was recovered to a considerable extent for
p-methoxybenzyl alcohol and the aliphatic 2,4-
hexadien-1-ol. This result indicates the beneficial
Figure 3. Location of the mutations in the evolved
secondary benzyl alcohol oxidase. (A) FX9 parental type
and (B) LanDo mutant. The FAD molecule is represented
in yellow, the catalytic base His502 and His546 is depicted
in red, and the consensus ancestral mutation Asn91 is in
light pink. A77V, R80C, V340A, I500M and F501W are
represented following a color code (before and after
mutation). The mutation L170M at the surface of the
enzyme is not present in the fragment represented. The
inset shows a detail of the protein surface at the access
channel. The model was prepared with a crystal structure
of the AAO from P. eryngii (PDB 3FIM).
Biochemical characterization
To characterize the LanDo mutant and the FX9
parental variant, they were produced and purified to
homogeneity. To determine the enantioselectivity of
the LanDo variant, transformation of the racemic 1-
(p-methoxyphenyl)-ethanol was followed by chiral-
HPLC (Figure 4) and 100% conversion with an
enantiomeric excess (ee) >99% was achieved after a
effect that A77V, R80C and V340A exerted on
LanDo´s overall activity (Table 4).
7
This article is protected by copyright. All rights reserved.

10.1002/adsc.201900134
Advanced Synthesis & Catalysis
Figure 4. Chiral HPLC analysis. HPLC elution profiles
after the reaction of the AAO variant LanDo (1 µM) with
racemic 1-(p-methoxyphenyl)-ethanol (2.5 mM). Reactions
were performed at room temperature in 100 mM phosphate
buffer pH 6.0 with continuous shaking and the aliquots
were analyzed by chiral HPLC at different times. The
separation of the R and S enantiomers in the negative
control is represented in black, whereas the 15, 45, and 90
min reactions are represented in red, blue and green,
respectively.
Table 3. Initial turnover rates for secondary alcohols.
p-Fluoro-α-methylbenzyl
1-Phenylethanol
1-Phenylpropanol
alcohol
0.10 ± 0.03
n.m.
0.35 ± 0.01
7.3 ± 0.1
wtAAO
LanDo
10 ± 0.5
1 ± 0.1
Turnover rates (min-1) were estimated in 100 mM phosphate buffer pH 6.0 containing 5 mM of each secondary alcohol
with the exception of 1-phenylpropranol (2.5 mM). Each reaction was performed by triplicate and substrate conversion
was followed by measuring the absorption at 563nm (ε563 = 56000 M-1 cm-1) using the HRP/Amplex red coupled assay as
described in the experimental section.
Table 4. Kinetic parameters for AAO variants.
Substrate
Kinetic constants
wtAAO**
LanDo
Km (mM)*
kcat (s-1)
24.9 ± 1.1
0.18 ± 0.002
0.007
0.65 ± 0.1
4.9 ± 0.1
7.5
1-(p-methoxyphenyl)-ethanol
kcat/Km (mM-1 s-1)
Km (mM)
kcat (s-1)
0.027 ± 0.004
142 ± 5
0.02 ± 0.003
72 ± 3
p-methoxybenzyl
kcat/Km (mM-1 s-1)
5233
3600
alcohol**
Km (mM)
kcat (s-1)
0.094 ± 0.005
119 ± 2
0.095 ± 0.006
40.9 ± 0.7
430.5
2,4-hexadien-1-ol**
kcat/Km (mM-1 s-1)
1271
AAO kinetic constants were measured in 100 mM phosphate buffer pH 6.0 at 25°C. All reactions were performed by
triplicate. *Referred to the S enantiomer, as 50% of the racemic mixture. **Calculated for wtAAO as described previously
[22,26]
where both proton transfer from the hydroxyl group
to the catalytic base His502 and hydride abstraction
from the benzylic position by the flavin are taking
place at the same reaction step.[11] PELE results were
plotted placing both catalytic distances in the X and
Y axis, and the interaction energy between the protein
and the ligand was represented by colors (Figure 5).
Computational analysis
In order to rationalize the effect of the mutations
identified, PELE simulations were run for wtAAO
and the variants obtained in the different rounds of
directed evolution. The oxidation of alcohols by
AAO involves a non-synchronous concerted reaction,
8
This article is protected by copyright. All rights reserved.

10.1002/adsc.201900134
Advanced Synthesis & Catalysis
Figure 5. Substrate diffusion computational PELE simulations. Plots represent the PELE simulations relating catalytic
distances (X and Y) and interaction energies (color scheme, right Y axis) for different AAO variants: (A) wtAAO; (B) the
FX9 secretion mutant; (C) the 15G12 mutant; (D) the final evolved benzyl alcohol oxidase, LanDo mutant.
No significant differences were evident between the
wtAAO and the secretion mutant FX9 (H91N-
L170M) in these plots, consistent with the
experimental evidence that this variant does not
improve the activity on secondary alcohols but does
increase expression and stability. The 15G12 variant
included the I500M and F501W mutations on top of
the previous ones, accumulating a total of 4 mutations
in the mature protein. In this case, the plot shows how
the ligand can reach catalytic positions 2.5 Å away
from both the FAD and the histidine at the same time,
producing better catalytic constants than those of the
wtAAO. In addition, a minimum could be seen where
the ligand-histidine distance was ~2.2 Å, although the
interaction energies were much higher and they were
therefore less accessible.
allow the ligand to achieve even smaller catalytic
distances, up to ~2 Å for the histidine and 2.4 Å for
the FAD, with reasonable interaction energies.
Considering that closer catalytic distances imply a
decrease in energy barriers, this agreed well with the
higher kinetic constants. Moreover, the increase in
the number of structures with good catalytic distances
could reflect the ease with which the ligand can find
catalytic positions, explaining the lower Km values for
this variant.
Despite the large number of mutations in LanDo (7
in total, excluding the mutations in the chimeric
prepro-leader), there are no major conformational
changes in the protein or in the positioning of the
ligand. Nevertheless, we did note subtle
modifications that were sufficient to improve the
catalytic position of (S)-1-(p-methoxyphenyl)-ethanol.
In particular, the R80C mutation was found
repeatedly in independent libraries during evolution
(Figure 2) and it created an empty space at the top of
The largest improvement came after introducing an
additional three mutations in the LanDo variant:
A77V, R80C, and V340A. These three substitutions
9
This article is protected by copyright. All rights reserved.

10.1002/adsc.201900134
Advanced Synthesis & Catalysis
the FAD cofactor. Consequently, the backbone
containing H502 shifts in that direction (Figure 6A).
Moreover, Arg80 interacts with the oxygen of the
backbone of residue 501, such that this mutation frees
Trp501 to form a hydrogen bond with the oxygen
backbone of Val77 (Figure 6B). All these subtle
adjustments allow the ligand to adopt conformations
with better catalytic distances (Figures 5, 6).
variant
providing
maintained
strong
enantioselectivity,
a
potential catalyst for chiral de-
racemization. The complex enzyme-substrate
relationships of this enzyme were highlighted by an
enhancement of three orders of magnitude in the
catalytic efficiency, an effect produced by a
combination of bulky substitutions in the catalytic
cavity and other unpredicted changes. Paradoxically,
a steric problem appeared to be resolved by
introducing bulkier residues, something difficult to
anticipate from a rational point of view. It is worth
noting that the resolution of molecules like 1-
phenylpropanol could be of use to obtain moiety
precursors of serotonin/norepinephrine reuptake
inhibitors like Fluoxetine or Atmoxetine. [6]
The results obtained here also highlight the
importance and efficacy of S. cerevisiae as a platform
for both the functional expression of eukaryotic genes
and as a molecular tool-box to generate DNA
libraries for directed evolution campaigns. The data
presented open new opportunities for the evolution of
AAO, which include the oxidation of furfural
derivatives for the synthesis of biopolymers or the in
situ
production
of
H2O2
in
cascade
oxyfunctionalization reactions by peroxygenases.[25]
Experimental Section
Materials
All chemicals were reagent-grade purity. Benzyl
alcohol,
p-chlorobenzyl
alcohol,
3,4-
dimethoxybenzyl alcohol, p-Methoxybenzyl alcohol,
1-phenylethanol,
1-phenylpropanol,
1-(p-
methoxyphenyl)-ethanol,
4-fluoro-α-methylbenzyl
alcohol, (R)-4-fluoro-α-methylbenzyl alcohol, (S)-4-
fluoro-α-methylbenzyl alcohol, 2,4-hexadien-1-ol,
cinnamyl alcohol, Horseradish peroxidase (HRP),
Taq polymerase and the Yeast Transformation Kit
were purchased from Sigma (Madrid, Spain).
Amplex®
Red
reagent
(10-acetyl-3,7-
dihydroxyphenoxazine) was obtained from Biogen
(Madrid, Spain). Zymoprep Yeast Plasmid Miniprep,
Yeast Plasmid Miniprep Kit I and Zymoclean Gel
DNA Recovery Kit were from Zymo Research
(Orange, CA, USA). Restriction enzymes BamHI and
XhoI were from New England Biolabs (Hertfordshire,
UK). I-Proof high-fidelity DNA polymerase was
from Biorad (Hercules, CA, USA). The episomal
shuttle vector pJRoC30 was from the California
Institute of Technology (CALTECH, USA). E. coli
XL2-Blue competent cells were from Stratagene (La
Jolla, CA, USA). Primers were acquired from Isogen
Life Science (Barcelona, Spain) and are included in
Table S1.
Figure 6. Conformational changes at the catalytic pocket
of AAO. (A) Backbone displacement (green) to better
position (S)- 1-(p-methoxyphenyl)-ethanol (pMPE). (B)
Interruption of the interaction of Arg80 with the backbone
of residue 501, and the formation of a hydrogen bond
between Val77 and Trp501.
Conclusion
Focusing evolution on structural elements made it
possible to identify mutations in the catalytic pocket
and access channel that allowed an AAO to be
designed that acts on different secondary benzyl
alcohols. This final secondary benzyl alcohol oxidase
10
This article is protected by copyright. All rights reserved.

10.1002/adsc.201900134
Advanced Synthesis & Catalysis
described below. Each construct was recovered and
its sequence confirmed by DNA sequencing.
Culture media
Minimal medium SC contained 100 mL 6.7% (w:v)
sterile yeast nitrogen base, 100 mL 19.2 g/L sterile
yeast synthetic drop-out medium supplement without
uracil, 100 mL sterile 20% raffinose (w:v), 700 mL
sddH2O and 1 mL 25 g/L chloramphenicol. YP
medium contained 10 g yeast extract, 20 g peptone
and ddH2O to 650 mL whereas YPD medium also
contained 20% glucose (w:v). AAO expression
medium contained 144 mL YP 1.55x, 13.4 mL 1 M
KH2PO4 pH 6.0 buffer, 22.2 mL 20% galactose (w:v),
0.222 mL 25 g/L chloramphenicol and ddH2O to 200
mL. Luria Broth (LB) medium contained 10 g sodium
chloride, 5 g yeast extract, 10 g peptone, 1 mL 100
mg/mL ampicillin and ddH2O to 1 L. AAO selective
expression medium (SEM) contained 100 mL 6.7%
(w:v) sterile yeast nitrogen base, 100 mL 19.2 g/L
sterile yeast synthetic drop-out medium supplement
without uracil, 100 mL sterile 20% galactose (w:v),
100 mL 1 M KH2PO4 pH 6.0, 600 mL sddH2O and 1
mL 25 g/L chloramphenicol.
Directed evolution
First
generation:
combinatorial
saturation
mutagenesis at positions Ile500 and Phe501: Two
PCR reactions were carried out in a final volume of
50 μL containing 3% DMSO, 0.8 mM dNTPs (0.2
mM each), 0.03 U/μL iproof DNA polimerase, and
0.2 ng/μL FX9 template and different primers
according to the 22-trick protocol.[13] PCR1 contained
0.25 μM RMLN and 0.25 μM mix of reverse primers:
22c1R, 22c2R, 22c3R, 22c4R, 22c5R, 22c6R, 22c7R,
22c8R and 22c9R. PCR2 contained 0.25 μM RMLC
and 0.25 μM mix of forward primers: 22c1F, 22c2F,
22c3F, 22c4F, 22c5F, 22c6F, 22c7F, 22c8F and
22c9F. Amplification reactions were carried out in a
thermal cycler MycyclerTM (BIO-RAD, USA) with
the following PCR program: 98°C for 30 seconds (1
cycle); 98°C for 10 seconds, 50°C for 25 seconds and
72°C for 60 seconds (28 cycles); and 72°C for ten
minutes (1 cycle). After purification, PCR products
(400 ng each) were mixed with the linearized
pJRoC30 (100 ng; ratio PCR product: vector = 4:1)
and transformed in yeast for in vivo cloning.
According to the 22-trick protocol, a library of 3066
individual colonies was screened as described below.
Construction of variants I500A, F501A and I500A-
F501A
All the PCR products were cleaned, concentrated,
loaded onto a preparative agarose gel (1%, w:v) and
purified using the Zymoclean Gel DNA Recovery kit
before being cloned into the shuttle vector pJRoC30
under the control of the GAL1 promoter. BamHI and
XhoI were used to linearize the plasmid pJRoC30 and
to remove the parent gene. FX9 variant was amplified
from pJRoC30-FX9[12b] with two PCR reactions for
each mutant containing overhang segments for the
whole plasmid to be reassembled in the yeast. PCR
reactions were carried out in a final volume of 50 μL
containing 3% DMSO, 0.8 mM dNTPs (0.2 mM
each), 0.03 U/μL Iproof DNA polymerase, and 0.2
ng/μL template. The oligos used for each PCR
reaction were: For I500A, PCR1 (oligo sense RMLN
and oligo antisense I500Ar), PCR2 (oligo sense
I500Af and oligo antisense RMLC). For F501A,
PCR1 (RMLN and oligo antisense F501Ar), PCR2
(oligo sense F501Af and RMLC). For double mutant
I500A-F501A, PCR1 (RMLN and oligo antisense
DM500-1Ar), PCR2 (oligo sense DM500-1Af and
RMLC). Amplification reactions were carried out in
a thermal cycler MycyclerTM (BIO-RAD, USA) with
the following PCR program: 98°C for 30 seconds (1
cycle); 98°C for 10 seconds, 50°C for 25 seconds and
72°C for 60 seconds (28 cycles); and 72°C for ten
minutes (1 cycle). After purification, PCR products
(400 ng each) were mixed with the linearized
pJRoC30 (100 ng; ratio PCR product: vector = 4:1)
and transformed in yeast (Yeast transformation kit)
for the recombination and in vivo cloning. 176
colonies were picked, expressed and screened as
Second generation: MORPHING: The 15G12 variant
was used as the parental template for focused random
mutagenesis technique MORPHING (Mutagenic
Organized Recombination Process by Homologous
IN vivo Grouping).[14] Three libraries were
constructed independently targeting 3 protein blocks:
MA, MB, and MC. Additionally, a mutagenic library
subjecting the whole AAO fusion was prepared by
error prone PCR (ep-PCR). Primers were designed to
create homologous overlapping areas of ~50 bp for
the whole gene to be reassembed in vivo upon
transformation in S. cerevisiae. i) ep-PCR for
MORPHING blocks and whole AAO gene were
carried out in a final volume of 50 µL containing: 90
nM oligo sense (FM1 for MA block, FM2 for MB
block, FM3 for MC block and RMLN for whole gene
amplification), 90 nM oligo antisense (RM1 for
fragment MA, RM2 for fragment MB, RM3 for
fragment MC and RMLC for whole gene
amplification), 0.3 mM dNTPs (0.075 mM each), 3%
DMSO, 0.05 or 0.01 mM MnCl2, 1.5 mM MgCl2,
0.05 U/ µL Taq polymerase DNA, and 0.92 ng/
µL15G12 template. The amplification parameters
were 95°C for 2 min (1 cycle); 95°C for 45 s, 50°C
for 45 s, and 74°C for 2 min (28 cycles); and 74°C
for 10 min (1 cycle). Concentrations of 0.05 and 0.01
mM MnCl2 were used for MORPHING and full gene
ep-PCR, respectively, to adjust the mutational rate to
1-3 mutations per gene. ii) High-fidelity PCRs for
11
This article is protected by copyright. All rights reserved.

10.1002/adsc.201900134
Advanced Synthesis & Catalysis
MORPHING were carried out in a final volume of 50
μL containing 3% DMSO, 0.8 mM dNTPs (0.2 mM
each), 0.03 U/μL iproof DNA polimerase, 0.25 μM
PCR1 used oligo sense RMLN and oligo antisense
R116 (containing position 39). PCR2 was performed
with oligo sense F116 (containing position 39) and
oligo sense (RMLN for HA1, HB1 and HC fragments, oligo antisense R5 (containing positions 76, 77, 80,
FHF1 for HA2 and FHF2 for HB2), 0.25 μM oligo
antisense (RHF1 for HA1, RHF2 for HB1, RHF3 for
MC and RMLC for HA2 and HB2 fragments) and 0.2
ng/μL template. High-fidelity PCR was performed
using the following parameters: 98°C for 30 s (1
cycle); 94°C for 10 s, 48°C for 30 s, 72°C for 30 s
(30 cycles); and 72°C for 10 min (1 cycle). The
assembly of the fragments for the different libraries is
described in Figure 1B. PCR products were cleaned,
concentrated, loaded onto a preparative agarose gel,
purified, mixed in equimolar amounts, (200 ng
mutagenic fragment and 200 ng non-mutagenic
fragment) and transformed with linearized pJRoC30
(200 ng) into chemically competent cells, as
described above.
83 and 90). PCR3 used oligo sense F5 (containing
positions 76, 77, 80, 83 and 90) and oligo antisense
R521 (containing position 174). PCR4 used oligo
sense F521 (containing position 174) and oligo
antisense R9419 (containing positions 332 and 340).
PCR5 was performed with oligo sense F9419
(containing positions 332 and 340) and oligo
antisense R1397 (containing position 466). PCR6 was
performed with oligo sense F1397 (containing
position 466) and oligo antisense RMLC, (Figure S1,
Table S1). For the in vivo assembly of the whole
gene, the fragments were designed with overhangs of
around 40 bp between them. Amplification reactions
were carried out in a thermal cycler and the PCR
program was: 98°C for 30 seconds (1 cycle); 98°C
for 10 seconds, 50°C for 25 seconds and 72°C for 30
seconds (28 cycles); and 72°C for ten minutes (1
cycle). After purification, PCR products (400 ng
each) were mixed with the linearized pJRoC30 (100
ng; ratio PCR product: vector = 4:1) and transformed
in yeast for in vivo cloning. A library of 3070
individual colonies was screened as described below.
Third generation: mutagenic shuffling: ep-PCR
reactions were performed separately with mutants
3F10, 11H2, 12D2 and 6G3. Reaction mixtures were
prepared in a final volume of 50 µL containing DNA
template (0.92 ng/µL), 90 nM oligo sense RMLN (5’-
CCTCTATACTTTAACGTCAAGG-3’), 90 nM
Reverse
primer
RMLC
(5’-
GGGAGGGCGTGAATGTAAGC-3’), 0.3 mM
dNTPs (0.075 mM each), 3% (v:v) dimethylsulfoxide
(DMSO), 1.5 mM MgCl2, 0.1 mM MnCl2 and 0.05
U/µL Taq DNA polymerase. PCRs were performed
in a thermocycler (Mycycler, Bio-Rad, Hercules, CA,
USA) and parameters were: 95°C for 2 min (1 cycle);
95°C for 45 s, 50°C for 45 s, 74°C for 45 s (28
cycles); and 74°C for 10 min (1 cycle). The PCR
products were mixed with linearized pJRoC30 (at a
PCR product/linearized plasmid ratio of 4:1) and
transformed into competent S. cerevisiae cells to
promote in vivo DNA shuffling. Transformed cells
were plated on SC (synthetic complete) drop-out
plates and incubated for 3 days at 30°C. Colonies
containing the whole autonomously replicating vector
were picked and subjected to high-throughput
screening as described below.
High-throughput screening (HTS) assay
Transformed cells were plated in SC drop-out plates
and incubated for 3 days at 30°C, individual clones
were fermented in sterile 96-well plates containing
200 μL of SEM medium. Plates were sealed and
incubated at 30°C, 225 rpm and 80% relative
humidity in a humidity shaker (Minitron-INFORS,
Biogen, Spain) for 72 hours. Aliquots of 20 μL of
yeast supernatants were transferred to a 96-well plate
using a robotic station for liquid handling Freedom
EVO (Tecan, Männedorf, Switzerland) and 180 μL of
HRP-Amplex Red reagent for secondary alcohol
activity detection. The final concentrations in the well
were 5 mM 1-(p-methoxyphenyl)-ethanol, 70 µM
Amplex Red, 3 μg/mL HRP in 100 mM phosphate
buffer pH 6.0. Reagents were dispensed with
Multidrop™ Combi Reagent Dispenser (Thermo
Scientific, Massachusetts, USA). The plates were
incubated at room temperature and activity with the
chiral alcohol was determined as H2O2 production
coupled to the oxidation of Amplex Red reagent by
the HRP and measured at 563nm (εresorufin 563 = 56000
M-1 cm-1). Reaction mixture with p-methoxybenzyl
alcohol (1 mM final concentration) was also prepared
to determine activity with a primary alcohol. One unit
of AAO activity is defined as the amount of enzyme
that converts 1 µmol of alcohol to aldehyde or ketone
with the stoichiometric formation of H2O2 per min
under the reaction conditions. The HTS-assay
incorporated two consecutive re-screenings to rule
Fourth generation: site-directed recombination: The
3C11 variant was used as template and primers were
designed for the 10 selected mutations (E39G, I76V,
A77V, R80C, M83I, V90A, Q174R, F332L, V340A,
and Q466R) for the in vivo site-directed
recombination of mutations vs. corresponding
reversions at each position of the combinatorial
library. A total of 6 PCR reactions were performed.
Reactions were carried out in a final volume of 50 μL
containing 3% DMSO, 0.8 mM dNTPs (0.2 mM
each), 0.03 U/μL iproof DNA polimerase, and 0.2
ng/μL 3C11 with 0.25 μM of the following oligos:
12
This article is protected by copyright. All rights reserved.

10.1002/adsc.201900134
Advanced Synthesis & Catalysis
out the selection of false positives as described in
previous work.[12a]
U/mL HRP and 75 µg/mL Amplex Red final
concentrations) following activity at 563nm (ε563
=
56000 M-1 cm-1).
Protein production and purification
HPLC analysis and optical rotation: The
enantioselectivity of LanDo variant was analyzed by
chiral HPLC with equipment consisting of a tertiary
pump (Varian/AgilentTechnologies) coupled to an
autosampler (Merck Millipore) and a Lux 5 µm
Cellulose-1 column (Phenomenex). For the mobile
phase hexane and isopropanol in a ratio 9:1 was used.
The separation of the enantiomers was performed at a
flow rate of 1.0 mL/min. The rotary polarization was
measured with a JASCO P-2000 polarimeter. After
full conversion, a reaction mixture with the remaining
alcohol (1mg/mL of 1-(p-methoxyphenyl)-ethanol)
was extracted with ethyl acetate and dissolved in
methanol. The measurement was made at 25 °C with
a sodium lamp at 589 nm.
The native AAO, heterologously expressed in E. coli
and in vitro refolded (wtAAO), was produced and
purified as described elsewhere.[15] Evolved variants
were produced in yeast[12a] and purified by cationic
exchange
chromatography,
anion
exchange
chromatography and size-exclusion chromatography
(ÄKTA purifier, GE Healthcare, WI, US). The crude
extract from S. cerevisiae cultures was concentrated
and dialyzed in 20 mM sodium phosphate/citrate at
pH 3.3 (buffer A) by tangential ultrafiltration
(Pellicon; Millipore, USA) through a 10-kDa-pore-
size membrane (Millipore, USA) by means of a
peristaltic pump (Masterflex easy load; Cole-Parmer,
USA). The sample was filtered and loaded onto a
strong cation-exchange column (HiTrap SP FF; GE
Healthcare) pre-equilibrated with buffer A and
coupled to the ÄKTA purifier system. The proteins
were eluted with a linear gradient of buffer P (20 mM
piperazine buffer pH 5.5) + 1M NaCl in two phases at
a flow rate of 1 ml/min: from 0 to 50 % during 15
min and from 50 to 100% in 2 min. Fractions with
AAO activity were pooled, dialyzed against buffer P,
concentrated and loaded onto a high-resolution resin,
strong-anion-exchange column (Biosuite MonoQ 10
cm; Waters, USA) pre-equilibrated in buffer P. The
proteins were eluted with a linear gradient from 0 to
0.5 M of NaCl in two phases at a flow rate of 1
ml/min: from 0 to 50 % during 20 min and from 50 to
100% in 2 min. Fractions with AAO activity were
pooled, dialyzed against 20 mM phosphate buffer pH
6.0, concentrated and further purified by a Superose
12 HR 10/30 molecular exclusion column
(Amersham Bioscience; Amersham, UK) pre-
equilibrated with 150 mM NaCl in 20 mM phosphate
buffer pH 6.0 at a flow rate of 0.5 ml/min. The
fractions with AAO activity were pooled, dialyzed
against buffer 20 mM phosphate buffer pH 6.0,
concentrated and stored.
Protein modeling: A structural model of the AAO
from P. eryngii crystal structure at a resolution of
2.55 Å (Protein Data Bank Europe [PDB] accession
number 3FIM,[16] was used as scaffold for the wild
type AAO model and the homology models for
different mutants were made from 3FIM by PyMol
(Schrodinger LLC.; http://www.pymol.org).
DNA sequencing: All genes were verified by DNA
sequencing (BigDye Terminator v3.1 Cycle
Sequencing Kit) using the following primers: primers
sense, RMLN and AAOsec1F and primers antisense
RMLC, AAOsec2R, and AAOsec3R.
Computational methods
The diffusion and binding of the secondary alcohol
(S)-1-(p-methoxyphenyl)-ethanol were studied using
the new adaptive-PELE (Protein Energy Landscape
Exploration) software.[17] This adaptive protocol
offers improved sampling by running multiple and
consecutive short PELE simulations (epochs), setting
the initial conditions of each epoch following a
reward function that favours sampling of unexplored
areas. Even though the ligands were parameterized
based on the OPLS2005 force field, the electrostatic
charges were derived from the electrostatic potential
obtained through quantum calculations at the M06/6-
31G* level with Jaguar[18], and its rotamer library was
built with Macromodel.[19] Similarly, the FAD
cofactor was optimized with a mixed QM/MM
calculation at the same level of theory using Qsite.
A single adaptive-PELE simulation was enough to
explore the diffusion and binding of (S)-1-(p-
methoxyphenyl)-ethanol for the native protein and
the different variants. The ligand was initially placed
between residues Gly52 and Asn56 in all cases. Each
simulation used 192 processors, producing 60 epochs
of 20 PELE steps each. During the simulation, ligand
perturbations (rotations and translations) are coupled
Biochemical characterization
Steady-state kinetic constants: Alcohol oxidation
kinetics for 1-(p-methoxyphenyl)-ethanol, p-
methoxybenzyl alcohol and 2,4-hexadien-1-ol were
measured in 100 mM phosphate buffer pH 6.0 at
25°C in air-saturated conditions (0.256 mM O2
concentration). Reactions were performed by
triplicate and substrates oxidations were followed by
measuring the absorption at 285 nm for p-
methoxybenzyl alcohol, ε285 = 16,950 M−1 cm−1 and
at 280 nm for 2,4-hexadien-1-ol, ε280 = 30,140 M−1
cm−1. The oxidation of 1-(p-methoxyphenyl)-ethanol
was measured indirectly coupled with saturated
conditions of HRP and Amplex Red substrate (4.5
13
This article is protected by copyright. All rights reserved.

10.1002/adsc.201900134
Advanced Synthesis & Catalysis
[10] F.J. Ruiz-Dueñas, A.T. Martinez, Microb. Biotechnol.
2009, 2, 164-77.
[11] P. Ferreira, A. Hernandez-Ortega, B. Herguendas,
A.T. Martinez, M. Medina, J Biol. Chem. 2009, 284,
24840-24847.
[12] a) J. Viña-Gonzalez, D. Gonzalez-Perez, P. Ferreira,
A.T. Martinez, M. Alcalde, Appl Environ Microbiol. 2015,
81(18), 6451-62; b) J. Viña-Gonzalez, K. Elbl, X. Ponte, F,
Valero, M. Alcalde, Biotechnol. Bioeng. 2018, 115(7),
1666-1674.
with the backbone perturbation to allow protein
flexibility. This is achieved by using an anisotropic
network model[20] applied to every Cα atom, while all
sidechain conformations within 6 Å of the ligand
were predicted each step. An epsilon value of 0.2 was
used in the adaptive protocol, meaning that 20% of
the processors started each epoch from the structure
with the best ligand-FAD distance.
[13] S. Kille, C.G. Acevedo-Rocha, L.P. Parra, Z.G.
Zhang, D.J. Opperman, M.T. Reetz, J.P. Acevedo. ACS
Synth. Biol. 2013, 2, 83-92.
[14] D. Gonzalez-Perez, P. Molina-Espeja, E. Garcia-Ruiz,
M. Alcalde, PLoS One. 2014, 9:e90919.
[15] F.J. Ruiz Dueñas, P. Ferreira, M.J. Martínez, A.T.
Martínez, Prot. Exp. Purif. 2006, 45, 191-199.
[16] I.S. Fernandez, F.J. Ruiz-Dueñas, E. Santillana, P.
Ferreira, M.J. Martinez, A.T. Martinez, A. Romero, Acta
Crystallogr. D: Biol Cristallogr. 2009, 65, 1196-1205.
[17] D, Lecina, J. F. Gilabert, V. Guallar, Sci. Rep. 2017,
7, 8466.
[18] A. D. Bochevarov, E. Harder, T. F. Hughes, J. R.
Greenwood, D. A. Braden, D. M. Philipp, D. Rinaldo, M.
D. Halls, J. Zhang, R. A. Friesner. Int. J. Quantum Chem.
2013, 113, 2110-2142.
[19] a) R. B. Murphy, D. M. Philipp, R. A. Friesner, J.
Comput. Chem. 2000, 21, 1442-1457; b) G. M. Sastry, M.
Adzhigirey, T. Day, R. Annabhimoju, W. Sherman, J.
Comput. Aided Mol. Des. 2013, 27, 221-234.
[20] A.R. Atilgan, S.R. Durell, R.L. Jernigan, M.C.
Demirel, O. Keskin, I. Bahar, Biophys. J. 2001, 80, 505-
515.
[21] A. Hernandez-Ortega, P. Ferreira, A.T. Martinez,
Appl. Microbiol. Biotechnol. 2012, 93, 1395-1410.
[22] A. Serrano, F. Sancho,a J. Viña-González, J. Carro,
M. Alcalde, V. Guallar, A.T. Martínez, Catal. Sci.
Technol. 2019, 9, 833-841.
[23] A. Hernandez-Ortega, P. Ferreira, P. Merino, M.
Medina, V. Guallar, A.T. Martinez, ChemBioChem. 2012,
13, 427-435.
[24] a) P. Ferreira, F.J. Ruiz-Dueñas, M.J. Martinez,
W.J.H. van Berkel, A.T. Martínez. FEBS J. 2006, 273,
4878-4888; b) A. Hernandez-Ortega, F. Lucas, P. Ferreira,
M. Medina, V. Guallar, A.T. Martinez, Biochemistry.
2012, 51, 6595-6608; c) J. Carro, P. Amengual-Rigo, F.
Sancho, M. Medina, V. Guallar, P. Ferreira, A.T. Martínez,
Sci Rep. 2018, 8, 8121.
[25] J. Carro, E. Fernandez-Fueyo, C. Fernandez-Alonso,
J. Cañada, R. Ullrich, M. Hofrichter, M. Alcalde, P.
Ferreira, A. T. Martinez, Biotechnol Biofuels. 2018, 11, 86.
[26] P, Ferreira, M, Medina, F, Guillén, M.J, Martínez,
W.J, Van Berkel, A.T.Martínez, Biochem J. 2005, 1,
389(Pt 3), 731-8.
Acknowledgements
This research was supported by the EU project FP7-
KBBE-2013-7-613549-INDOX, by the Spanish
Government projects BIO2016-79106-R-Lignolution,
CTQ2016-74959-R (MINECO/FEDER, EU) and by
the Comunidad de Madrid project Y2018/BIO4738-
EVOCHIMERA.
References
[1]
https://www.statista.com/statistics/263102/pharmaceutical-
market-worldwide-revenue-since-2001
[2] a) J. Caldwell, Hum Psychopharmacol. 2001, 16(S2),
S67-S71; b) B.S. Sekhon, J. Mod. Med. Chem. 2013, 1, 10-
36; c) N. Chhabra, M.L. Aseri, D. Padmanabhan, Int J
App Basic Med Res. 2013, 3(1), 16-18.
[3] N.M. Maier, P. Franco, W. Lindner, J Chromatogr A.
2001, 12;906(1-2):3-33.
[4] a) K. Faber, Chemistry. 2001, 3;7(23):5004-10; b) O.
Pàmies, J.E. Bäckvall,. Trends Biotechnol. 2004, 22(3),
130-5; c) N.J. Turner, Curr Opin Chem Biol. 2004, 8(2),
114-9.
[5] a) G. Fantin, M. Fogagnolo, A. Medici, P. Pedrini, S.
Poli, M. Sinigaglia, Tetrahedron Lett. 1993, 34(5), 883-
884; b) Y.L. Li, J.H. Xu, Y. Xu, J Mol Catal B Enzym.
2010, 64, 48-52; c) C. Voss, C. Gruber W. Kroutil, Angew.
Chem. Int. Ed. 2008, 47, 741-745.
[6] T. Matsuda, R. Yamanaka, K. Nakamura, Tetrahedron:
Asymmetry. 2009, 20, 513-557.
[7] a) A. Ghanem, H.Y. Aboul-Enein, Chirality. 2005,
17(1), 1-15; b) S. Qin, Y. Zhao, B. Wu, B. He, Appl
Biochem Biotechnol. 2016, 180(7), 1456-1466.
[8] a) F. Escalettes, N.J. Turner, Chembiochem. 2008,
14;9(6), 857-60; b) M. M. Musa, K. I. Ziegelmann-Fjeld,
C. Vieile, J. G. Zeikus and R. S. Phillips, J. Org. Chem.
2007, 72, 30-34.
[9] a) W. Kroutil, H. Mang, K. Edegger K. Faber, Adv.
Synth. Catal. 2004, 346, 125-142; b) E. W. van
Hellemond, L. Vermote, W. Koolen, T. Sonke, E.
Zandvoort, D. P. Heuts, D. B. Janssen, M. W. Fraaije, Adv.
Synth. Catal, 2009, 351, 1523-1530; c) W. P. Dijkman, C.
Binda, M. W. Fraaije, A. Mattevi, ACS Catal. 2015, 5,
1833-1839.
14
This article is protected by copyright. All rights reserved.

10.1002/adsc.201900134
Advanced Synthesis & Catalysis
FULL PAPER
Structure-Guided Evolution of Aryl Alcohol
Oxidase from Pleurotus eryngii for the Selective
Oxidation of Secondary Benzyl Alcohols
Adv. Synth. Catal. Year, Volume, Page – Page
Javier Viña-Gonzalez, Diego Jimenez-Lalana,
Ferran Sancho, Ana Serrano, Angel T. Martinez,
Victor Guallar and Miguel Alcalde*
15
This article is protected by copyright. All rights reserved.