Dihydrogen-Driven NADPH Recycling in Imine Reduction and P450-Catalyzed Oxidations Mediated by an Engineered O2-Tolerant Hydrogenase

Article abstract of DOI:10.1002/cctc.202000763

The O₂-tolerant NAD⁺-reducing hydrogenase (SH) from Ralstonia eutropha (Cupriavidus necator) has already been applied in vitro and in vivo for H₂-driven NADH recycling in coupled enzymatic reactions with various NADH-dependent oxidoreductases. To expand the scope for application in NADPH-dependent biocatalysis, we introduced changes in the NAD⁺-binding pocket of the enzyme by rational mutagenesis, and generated a variant with significantly higher affinity for NADP⁺ than for the natural substrate NAD⁺, while retaining native O₂-tolerance. The applicability of the SH variant in H₂-driven NADPH supply was demonstrated by the full conversion of 2-methyl-1-pyrroline into a single enantiomer of 2-methylpyrrolidine catalysed by a stereoselective imine reductase. In an even more challenging reaction, the SH supported a cytochrome P450 monooxygenase for the oxidation of octane under safe H₂/O₂ mixtures. Thus, the re-designed SH represents a versatile platform for atom-efficient, H₂-driven cofactor recycling in biotransformations involving NADPH-dependent oxidoreductases.

Full text of DOI:10.1002/cctc.202000763

The European Society Journal for Catalysis
Supported by
Accepted Article
Title: Dihydrogen-driven NADPH recycling in imine reduction and
P450-catalyzed oxidations mediated by an engineered O2-
tolerant hydrogenase
Authors: Janina Preissler, Holly A. Reeve, Tianze Zhu, Jake Nicholson,
Kouji Urata, Lars Lauterbach, Luet L. Wong, Kylie A. Vincent,
and Oliver Lenz
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: ChemCatChem 10.1002/cctc.202000763
Link to VoR: https://doi.org/10.1002/cctc.202000763
01/2020

10.1002/cctc.202000763
ChemCatChem
FULL PAPER
Dihydrogen-driven NADPH recycling in imine reduction and P450-
catalyzed oxidations mediated by an engineered O2-tolerant
hydrogenase
Janina Preissler,[a] Holly A. Reeve,*[b] Tianze Zhu,[b] Jake Nicholson,[b] Kouji Urata,[b] Lars
Lauterbach,*[a] Luet L. Wong,[b] Kylie A. Vincent,*[b] and Oliver Lenz*[a]
[a]
[b]
Dr. J. Preissler, Dr. L. Lauterbach, Dr. O. Lenz
Institute of Chemistry, Biophysical Chemistry
Technische Universität Berlin
Straße des 17. Juni 135, 10623 Berlin, Germany
E-mail: lars.lauterbach@tu-berlin.de, oliver.lenz@tu-berlin.de
Dr. Holly A. Reeve, T. Zhu, J. Nicholson, K. Urata, Prof. L. L. Wong, Prof. K. A. Vincent
Department of Chemistry,
University of Oxford
Inorganic Chemistry Laboratory, South Parks Road, Oxford, OX1 3QR, UK
E-mail: kylie.vincent@chem.ox.ac.uk
Supporting information for this article is given via a link at the end of the document.
Abstract: The O2-tolerant NAD+-reducing hydrogenase (SH) from
Ralstonia eutropha (Cupriavidus necator) has already been applied in
vitro and in vivo for H2-driven NADH recycling in coupled enzymatic
reactions with various NADH-dependent oxidoreductases. To expand
the scope for application in NADPH-dependent biocatalysis, we
introduced changes in the NAD+-binding pocket of the enzyme by
rational mutagenesis, and generated a variant with significantly higher
affinity for NADP+ than for the natural substrate NAD+, while retaining
native O2-tolerance. The applicability of the SH variant in H2-driven
NADPH supply was demonstrated by the full conversion of 2-methyl-
1-pyrroline into a single enantiomer of 2-methylpyrrolidine catalysed
by a stereoselective imine reductase. In an even more challenging
reaction, the SH supported a cytochrome P450 monooxygenase for
the oxidation of octane under safe H2/O2 mixtures. Thus, the re-
designed SH represents a versatile platform for atom-efficient, H2-
driven cofactor recycling in biotransformations involving NADPH-
dependent oxidoreductases.
at the expense of glucose-(6P) oxidation.[4] The gluconolactone
thus formed spontaneously hydrolyses to gluconic acid, rendering
the reaction irreversible and leads to acidification of the reaction
solution, which necessitates continuous pH monitoring and
adjustment. The large quantity of gluconic acid waste also lowers
the atom efficiency of the reaction, complicates recovery of water-
soluble products and limits the potential environmental benefits of
biocatalysis. Alternative systems for cofactor recycling are
phosphite dehydrogenase and alcohol dehydrogenase, which,
however,
produce phosphate and aldehydes/ketones,
respectively, as unwanted by-products during NAD(P)H
recycling.[5] Formate dehydrogenase (FDH) represents another
widely used cofactor regeneration system. FDH couples NAD(P)+
reduction with the oxidation of formate, but also suffers from low
enzymatic activity and formation of the by-product CO2.[6]
In contrast, the use of H2 gas as a source of reducing equivalents
allows atom-efficient recycling of NAD(P)H without any by-
product that may alter the pH or complicate product recovery.[7]
Indeed, two different NAD(P)+-linked soluble hydrogenases (SHs)
have been demonstrated for H2-driven NADH or NADPH recycling
to support the activity of a range of NAD(P)H-dependent
enzymes.[8,9] One of them is the archaeal NADP+-linked [NiFe]-
hydrogenase (I) from Pyrococcus furiosus (PfSHI), which works
well in a wide temperature range of 30–80 °C and reduces NADP+
rather than NAD+. Although described as partially O2-tolerant,[10]
electrochemical measurements on PfSHI at 60 °C, show that the
H2 oxidation activity of the enzyme at 1 bar H2 drops rapidly when
1% O2 is introduced, settling at a level corresponding to about 8%
of the anaerobic activity, while at 25 °C the activity drops to zero
under 1% O2.[11] Sustained H2-driven NADP+ reduction in the
presence of O2 has not been demonstrated for PfSHI. The O2
sensitivity of the Pf enzyme represents a major drawback for
hydrogenase-linked cofactor recycling as reactions need to be
carried out under anaerobic conditions. It clearly cannot support
cofactor supply for O2-dependent enzymes such as
monooxygenases that utilise NAD(P)H to supply electrons for
Introduction
Biocatalysis is becoming established as a valuable tool in the
production of complex chemicals.[1] In particular biocatalytic
transformations for asymmetric reductions including ketone,
alkene and imine reductions, and for controlled oxidations such
as terminal alcohol to aldehyde conversions and C–H bond
activations, are becoming accepted as promising alternatives to
traditional chemical methods.[2] However, the enzymes that
catalyze these transformations rely on electron transfer to or from
redox equivalents which in most cases are the expensive
biological cofactors NAD(P)+ or NAD(P)H.[3] To develop viable
biocatalytic processes, it is essential to have efficient systems for
recycling these cofactors.
One of the most commonly used cofactor recycling systems for
NAD(P)H is glucose-(6P) dehydrogenase which reduces NAD(P)+
1
This article is protected by copyright. All rights reserved.

10.1002/cctc.202000763
ChemCatChem
FULL PAPER
reductive activation of O2 for oxygen atom insertion into C–H or Results and Discussion
C–C bonds.[3,12,13]
The second promising candidate for both in vitro [9,14] and in vivo
cofactor recycling[15] is the soluble NAD+-reducing [NiFe]-
hydrogenase from Ralstonia eutropha (ReSH; R. eutropha is also
known as Cupriavidus necator[16]) which is able to catalyze H2-
driven NAD+ reduction even in the presence of ambient O2.[17,18]
Whole cells of Pseudomonas putida engineered to heterologously
co-produce both the NADH- and O2-dependent P450
monooxygenase (CYP153A) from Polaromonas sp. JS666 and
the ReSH were shown to oxidise octane to 1-octanol.[15] However,
many P450 enzymes of biotechnological interest are dependent
on the phosphorylated derivative of NADH, NADPH. The ReSH is
highly specific for NAD+, and H2-driven NADP+ reduction by ReSH
is not detected.[19] Thus, the native enzyme is unsuitable for
application in NADPH recycling and there have been no reports
of isolated enzymes suitable for NADP+ reduction in the presence
Construction of NADP+-reducing SH variants by rational
design
Altering substrate specificity of NAD+-reducing SH by rational
design required the identification of amino acid residues that are
involved in substrate binding. Unfortunately, a crystal structure of
SH from R. eutropha H16 is not yet available. The structure for
the closely related SH from Hydrogenophilus thermoluteolus has
been reported,[29] which, however, does not contain the NAD(H)
cofactor, whose precise orientation is essential for rational design
of the corresponding binding pocket. Therefore, we used two
crystal structures of Complex I from Thermus thermophilus (PDB
accession numbers 2FUG and 3IAM) as templates for homology
modelling (SWISS-MODEL) of HoxF (Figure 1a).[30,31] The Nqo1
subunit of Complex I shows 62% sequence similarity with the C-
terminal part of the HoxF subunit of ReSH (Figure S1) and, more
importantly, has been crystallised in the presence of NADH.[30] In
the homology model, the NADH binding site (Figure 1b) is located
in an unusual Rossmann fold featuring a glycine-rich loop that
facilitates simultaneous binding of NADH and FMN.[31] Stacking
interactions between the nicotinamide ring of NADH and the
isoalloxazine moiety of FMN as well as aromatic side chains of
conserved amino acid residues guarantee precise cofactor
positioning (Figure 1c). Several hydrogen bonds are formed
between NADH and the protein backbone. One of these is formed
between the carboxyl group of Glu341 and the 2’-hydroxyl group
of the adenosine ribose of NADH (Figure 1c). In case of NADPH,
this hydroxyl group is replaced by a bulky, negatively charged
phosphate group (Figure 1d), which would result in a steric clash
with Glu341. Notably, this glutamate residue is absent in NADPH-
specific dehydrogenases. We decided to engineer the specificity
of the SH towards NADP+ by exchanging Glu341, which interacts
with NADH, for the smaller alanine residue (E341A), to allow
accommodation of the NADP+-specific phosphate group. In
NADP(H)-converting enzymes, a positively charged arginine is
found in the neighbouring position to the alanine and stabilises
NADP(H) binding by forming an ionic bond to the negatively
charged phosphate group.[22] We thus replaced Ser342 of the SH
with an arginine (S342R, Figure 1d) to counterbalance the
negatively charged phosphate group of NADPH. Construction of
a single S342R variant was not considered because neighbouring
negatively charged (D, E) and positively charged (R, K) residues
were not observed in the cofactor binding site any NAD(P)(H)-
converting enzyme.[32]
of O2.
[20]
Since NADPH is generally less stable
and more expensive
than NADH there have been many reports, dating back to 1990,
on changing the cofactor specificity of technologically useful
enzymes from NADPH to NADH.[21] There have been some
notable successes, including structure-guided, semi-rational
strategies for reversing nicotinamide cofactor specificity, as
summarised by Cahn and coworkers.[22] However, significant loss
of catalytic activity tends to occur on attempting to alter cofactor
selectivity and the need to engineer each oxidoreductase
individually represents a significant burden. Thus, there is a clear
need for improved O2-tolerant NADPH cofactor recycling systems
that operate with high atom economy and high activity to support
NADPH-dependent reductases and monooxygenases.
Here, we describe site-directed amino acid exchanges in ReSH
which resulted in synthetic H2-dependent NADP+ reduction
activity which was still O2 tolerant. The nicotinamide cofactor
binding site of the ReSH is part of a non-classical Rossmann fold
including a flavin mononucleotide (FMN). This arrangement is
similar to that in the orthologous counterparts of respiratory
Complex I, certain FDH enzymes, and [FeFe]-hydrogenases.[23,24]
During H2-driven NAD+ reduction by the ReSH, the FMN
accumulates two electrons, which are subsequently transferred
as a hydride to NAD+ resulting in the reduced form, NADH.[25] In
order to convert ReSH into a NADP+-reducing enzyme, we first
constructed a homology model for the NAD(H)-binding subunit,
HoxF, to identify promising sites for amino acid exchanges. This
homology model is based on the crystal structure of the Nqo1
subunit from Complex I of Thermus thermophilus. A glutamate
residue in the NADH-binding pocket of the corresponding subunit
in Complex I from E. coli has been described to be crucial for
cofactor specificity.[26] We used this observation to modify the
HoxF subunit of ReSH towards NADP+ binding. Furthermore,
classical Rossmann motifs involved in nicotinamide cofactor
binding have already been targeted to change cofactor specificity
in many different dehydrogenases.[22] For example, Lerchner et al.
showed that by removing a positively charged residue, the
NADP+-specific alcohol dehydrogenase from Ralstonia sp. was
able to efficiently accept NAD+.[27] Here, we used these findings
to rationally design NADPH recycling SH variants that are able to
support a stereoselective imine reductase and an octane-
oxidising P450BM3 monooxygenase (CYP102A1)[28]. The latter
case represents the first example of H2-driven NADPH recycling
to support an O2-dependent biocatalytic process.
Finally, we considered amino acid residues, which do not directly
interact with the cofactor, but contribute to overall charge of the
binding pocket. The R. eutropha SH possesses a surface-
exposed, negatively charged aspartate residue at position 467
(Figure S2), while the cyanobacterial soluble hydrogenases,
which are reported to interact with both NADH and NADPH, carry
either serine or threonine at the corresponding position [33] (Figure
S1). In order to assist NADP+ binding, a D467S exchange was
also introduced in the SH. The corresponding HoxF variants
E341A,
D467S,
E341A/D467S,
E341A/S342R,
and
E341A/S342R/D467S were constructed by genetic engineering,
and the resulting variant SH proteins, henceforth designated
SHE467A, SHE341A/D467S, etc., were purified as described in the
Experimental Section.
2
This article is protected by copyright. All rights reserved.

10.1002/cctc.202000763
ChemCatChem
FULL PAPER
Figure 1. Subunit and cofactor composition of the soluble NAD+-reducing [NiFe]-hydrogenase (SH) of R. eutropha H16. (a) Schematic representation of the
hydrogenase module (HoxH and HoxY subunits in blue) and the NAD+-reductase module (HoxF and HoxU subunits in orange). The latter carries two copies of the
non-essential HoxI subunit. The iron-sulfur cluster relay connecting the two catalytic centres of the SH is shown as spheres. (b) Homology model of the HoxF subunit
(orange) based on the crystal structure of the Nqo1 subunit (green) of Complex 1 from Thermus thermophilus (2FUG).[31,34] The flavin mononucleotide and NADH
molecules (shown as sticks) are taken from pdb entry 3IAM. (c) The NAD(H)-binding pocket of native SH. Important hydrogen bonding and -stacking interactions
are shown as dashed lines. (d) Model of the NAD(P)(H) binding pocket of the SHE341A/S342R variant (carbon atoms of exchanged amino acids in pink). Note the
stabilising effect of the guanidinium group of R342 on the additional phosphate group of NADP+. The location of residue D467 is depicted in Figure S2.
Biochemical and electrochemical properties of SH variants
of native SH. The pH optimum of 7.5 of the NADP+ reduction
activity of SHE341A/S342R was found to be slightly lower than that (pH
8.0) for NAD+ reduction. Since all three exchanges had a positive
effect on the NADP+ reduction of the SH, we combined them in a
SHE341A/S342R/D467S variant. This variant, however, showed a
significantly lower NADP+ reduction activity than the two double
exchange variants (Figure 2a). The low activity might be due to
structural distortions of the cofactor binding site. The
SHE341A/S342R/D467S variant was therefore disregarded for the
subsequent experiments.
Next, the NAD(P)+ reduction activity was determined
electrochemically by means of protein film voltammetry. This
method allows analysis of the NAD(P)+/NAD(P)H cycling moiety
independently of the H+/H2 cycling module and relies on interfacial
First we measured the H2-dependent NAD(P)+ reductase
activities with 1 mM NAD(P)+ at pH 8. As shown in Figure 2a, all
the SH variants showed a dramatic decrease in H2-driven NAD+
reduction activity (orange bars). The lowest activity was
measured for the SHE341A protein, with only 10% of the activity of
native SH. However, NADP+ reduction activity was already clearly
detectable for this variant, which was not the case for the native
SH. The NADP+ reduction activity increased in the order SH <
SHD467S < SHE341A < SHE341A/D467S < SHE341A/S342R (Figure 2a). The
NADP+ reduction activity of the SHE341A/S342R variant of 1 U ∙ mg–1
represents approximately 10% of the NAD+ reduction activity of
this variant, and approximately 1% of the NAD+ reduction activity
3
This article is protected by copyright. All rights reserved.

10.1002/cctc.202000763
ChemCatChem
FULL PAPER
electron transfer between the electrode and the immobilised
enzyme. The activity of the enzyme film is proportional to the
current, but film-to-film variations in enzyme coverage on the
electrode make quantitative conclusions on the activity less
reliable from this method.[24] The electrochemical results (Figure
2b,c) are generally consistent with the biochemical assay data.
Although the interaction between the protein and the electrode is
not fully understood and a single amino acid exchange at the
enzyme surface (e.g. D467S) might have an impact on
immobilisation to the electrode, this is unlikely to be a significant
effect due to the high number of protein-electrode interactions of
the large SH enzyme. In fact, all variants show lowered current for
NAD+ reduction, and increased current for NADP+ reduction, with
the highest NADP+ reduction observed for the double variant
SHE341A/D467S. The electrochemical measurements also show that
the onset potential for NAD(P)+ reduction is not significantly
affected by the amino acid exchanges. The voltammogram
recorded at the electrode modified with SHD467S shows a slight
redox peak at approximately –250 mV; this is likely due to
dissociated FMN adhering to the electrode,[24] suggesting
increased flavin loss by this variant.
We then determined the Michaelis-Menten parameters for
NAD(P)+ reduction for the SH enzymes. The results clearly show
that all the exchanges significantly increased the affinity of SH for
NADP+ (Table 1). Even the D467S variant, with a less than 2-fold
increase in kcat for NADP+ reduction compared to native SH,
NADP
already showed a KM
of 3.4 mM, supporting the previous
observation that remote alterations can affect the NAD(P)H
binding properties of the catalytic centre.[35] The E341A exchange
NAD
led to an increase of KM
and a concomitant loss of NAD+
reduction activity (Figure 2a,b), presumably due to insufficient
substrate stabilisation as the hydrogen bond between the E341
side chain and the 2’-hydroxyl group of the adenosine ribose of
NAD+ is lost (Figure 1c,d). On the other hand, the much smaller
alanine side chain provides enough space for the additional 2’-
phosphate group of NADP+, which becomes accepted as a
substrate with a KM of 4.8 mM (Table 1); the kcat of 3.3 s–1 is more
than 30-fold higher than for native SH (Figure 2a,c). An additive
effect was achieved by combining the E341A and D467S
NADP
exchanges, with the SHE341A/D467S variant showing a KM
of
2.3 mM and kcat of 8.5 s–1. Even more specific binding of NADP+
(KMNADP = 0.6 mM) was observed upon introduction of the S342R
exchange to the SHE341A variant although the kcat was slightly
lowered to 6.0 s–1 (Table 1). The positively charged side chain of
Arg342 is proposed to form hydrogen bonds to the 2’-phosphate
group of NADP+.[36] The highest catalytic efficiency of the
SHE341A/S342R variant for H2-driven NADP+ reduction and the lowest
for NAD+ reduction are consistent with this proposal. Thus, the
cofactor preference of SH was switched by just two exchanges in
the SHE341A/S342R variant.
Figure 2. Comparison of the activity of the SH variants for NAD+ and NADP+
reduction. (a) Spectrophotometric determination of the H2-driven NAD+ (grey)
and NADP+ (blue) reduction activity. Conditions: 50 mM Tris-HCl buffer, pH 8,
1 mM NAD(P)+, 1 µM FMN and 1 mM dithiothreitol (DTT), 30 °C. H2-dependent
NADP+ reduction was not detectable for native SH. Means and error bars
(calculated standard deviation) are derived from three biological replicates
which were measured at least twice. In the case of NADP+ reduction by the
SHD467S variant the error bar is too small to be visible. (b) Electrode-driven NAD+
and, (c), NADP+ reduction activity measured at an enzyme-modified pyrolytic
graphite electrode, rotated at 2000 rpm. Conditions: 100 mM Tris-HCl buffer,
pH 8.0, 5 mM cofactor, scan rate 10 mV s–1, 30 °C.
4
This article is protected by copyright. All rights reserved.

10.1002/cctc.202000763
ChemCatChem
FULL PAPER
Table 1. KM and kcat values for H2-driven NAD+ and NADP+ reduction by SH variants
NAD+  NADH[a]
kcat (s–1
NADP+  NADPH
SH variant
KM (mM)
)
kcat/KM
(s–1/mM)
KM (mM)
kcat (s–1
)
kcat/KM
(s–1/mM)
native
0.55 ± 0.08
0.61 ± 0.11
0.66 ± 0.18
0.65 ± 0.18
1.32 ± 0.29
690 ± 63
20 ± 2.6
55 ± 6.9
68 ± 7.4
25 ± 4
1255
32.6
83.0
36.9
18.9
> 50
< 0.1
< 0.01
0.68
0.06
3.7
E341A
4.8 ± 1.7
3.4 ± 1.0
2.3 ± 1.1
0.60 ± 0.33
3.3 ± 1.0
0.2 ± 0.04
8.5 ± 1.6
6.0 ± 0.4
D467S
E341A/D467S
E341A/S342R
9.9
[a] Values were determined biochemically using 2 biological replicates for SH, SHE341A, and SHD467S and 3 biological replicates for the variant SH proteins carrying
double exchanges. Conditions: 740 µM H2, 0.1 – 6 mM NAD(P)+, approx. 40 nM enzyme in 25 mM Tris-HCl, pH 8.0 at 30 °C. Data are given as mean ± SD.
H2-driven NADPH recycling for the reduction of cyclic imines
Scheme 1. IRED-catalysed conversion of 2-methyl-1-pyrroline into (R)-2-
methylpyrrolidine. H2-driven NADPH recycling is mediated by the SHE341A/S342R
protein.
The most promising variant, SHE341A/S342R, showed good activity
for both H2-driven (kcat = 6.0 s–1) and electrode-driven NADP+
reduction and had an acceptable Michaelis constant of 600 µM
for NADP+. These properties made it an attractive candidate to
The efficiency of the SHE341A/S342R in NADPH cofactor
investigate H2-driven regeneration of NADPH to support the
regeneration
was
demonstrated
with
four
different
activity of redox enzymes such as imine reductases. We chose
an imine reductase (IRED) that catalyses the asymmetric
reduction of 2-methyl-1-pyrroline to a single enantiomer of 2-
methylpyrrolidine (Scheme 1).
biotransformations (Table 2, entries 1-4), all of which contained
1 mg IRED and 0.1 mg SHE341A/S342R in a 1 mL-reaction mixture
with varying substrate and cofactor concentrations at 2 bar H2.
Full conversion (> 99%) of 15 mM substrate as starting material
was achieved with 1 mM of NADP+ in an overnight reaction (16 h,
Table 2, entry 1). An increase of substrate concentration (25 mM)
led to a concomitant increase in the total turnover number of the
SH (Table 2, entry 2). The highest total turnover number of SH
(35000) was obtained with a higher substrate concentration and
a prolonged reaction time of 40 h. A still significant conversion of
34 % was reached after 40 h with 15 mM substrate and just
0.1 mM of cofactor. Under these conditions, the highest cofactor
total turnover number of 51 was achieved. Notably, the
enantiomeric excess in all reactions was greater than 99.
Table 2. H2-driven NADP+ reduction by SHE341A/S342R to support an IRED reaction.
Entry[a]
[Substrate]/
mM
[Cofactor]/
mM
Time/
h
Conversion/
%
[Product]/
g ∙ L–1
Cofactor
TTN
SH
TTN
ee
1
2
3
4
15
25
25
15
1
1
1
0.1
16
16
40
40
99
67
88
34
1.3
1.4
2.0
0.4
99
99
99
99
15
17
22
51
24000
27000
35000
8000
[a] reaction conditions: 15 - 25 mM substrate, 0.1 - 1 mM NADP+ cofactor, 1 mg (0.04 U) imine reductase, 0.1 mg SHE341A/S342R in 1 mL Tris-HCl (100 mM, pH 8),
incubated for 16 - 40 h at 200 rpm under 2 bar H2. See Table S1 for control reactions. The abbreviation “TTN” refers to the total turnover number of 2-methyl-1-
pyrrolidine converted per mol of cofactor/enzyme.
H2-driven NADP+ reduction in the presence of O2
O2, equivalent to the O2 level in aqueous solution equilibrated in
air. The results are shown in Table 3 and clearly demonstrate that
the SH variants are active for NADPH regeneration in the
presence of this level of O2. Although the O2 tolerance was
compromised somewhat in the single variants, importantly, the
double variants (which are most active for NADP+ reduction)
showed similar O2 tolerance compared to native SH.
The native SH is well-known for its capacity to maintain its H2-
dependent NAD+ reduction activity in the presence of atmospheric
O2 concentrations.[17,18] To test the effect of the amino acid
exchanges on O2 tolerance, we surveyed the H2-driven NADP+
reduction capacity of the SH variants in the presence of 0.23 mM
5
This article is protected by copyright. All rights reserved.

10.1002/cctc.202000763
ChemCatChem
FULL PAPER
oxidation in air versus 10 % air in N2, using GDH/glucose for
NADPH regeneration. Both enzymes gave the same product
concentrations and distributions, within experimental error, in air
and in 10 % air in N2 (Figure S3, Table S2), showing that P450BM3
is able to operate under a sub-ambient O2 partial pressure, as
described previously.[39] Next, we compared octane oxidation for
native P450BM3 and the A330P variant under 90 % H2:10 % air
using GDH/glucose or the SHE341A/S342R as the NADPH
regeneration system. For each P450BM3 enzyme, the total product
concentrations and distributions under the different conditions
were identical, within experimental error (Figure 3, Table S2). As
expected, no product formation was detected in the absence of
NADP+. We conclude that the O2-tolerant SHE341A/S342R variant
supports H2-driven NADPH-dependent substrate oxidation by
P450BM3 using safe H2/air mixtures as effectively as the
GDH/glucose system, under these reaction conditions.
Table 3. Oxygen tolerance of NADP+-reducing SH variants
SH Variant
Condition[a]
Hydrogenase
activity[a] / U ∙ mg–1
Hydrogenase
activity / %
SH
– O2
+ O2
– O2
+ O2
– O2
+ O2
– O2
+ O2
– O2
+ O2
53.6 ± 2.3[b]
46.9 ± 0.5[b]
0.78 ± 0.01
0.43 ± 0.01
0.22 ± 0.07
0.10 ± 0.02
0.93 ± 0.01
0.83 ± 0.01
1.22 ± 0.01
1.15 ± 0.02
100
87
SHE341A
100
56
SHD467S
100
46
SHE341A/D467S
SHE341A/S342R
100
89
100
94
[a] H2-dependent reduction of NAD(P)+ was measured biochemically in the
presence of either 50% H2/50% N2 (– O2) or 50% H2/30% N2/20% O2 (+ O2,
corresponding to 0.23 mM O2). [b] As there is no NADP+ reduction activity
detectable for native SH, NAD+ was used as the substrate.
SH-mediated NADPH recycling for H2-driven alkane oxidation
by P450BM3
The fact that the variant SHE341A/S342R remained catalytically active
in the presence of O2 prompted us to investigate H2-driven
regeneration of NADPH to support the activity of enzymes such
as cytochrome P450 monooxygenases, which require O2 as co-
substrate. P450 enzymes catalyse the selective oxidative
functionalisation of unactivated C–H bonds to the alcohol
functionality – one of the most difficult reactions in classical
synthesis. P450-catalysed reactions have huge potential
application in synthesis (e.g. for generating drug metabolites and
fine chemicals).[37] All eukaryotic and mitochondrial P450s utilise
electrons from NADPH to activate O2. Whilst the majority of
bacterial P450s require NADH, a significant number are NADPH-
dependent including P450BM3 (CYP102A1) from Bacillus
megaterium, a highly active, catalytically self-sufficient single-
polypeptide enzyme that has been extensively engineered for the
oxidation of unnatural substrates.[13,28,38]
Figure 3. GC analysis of octane oxidation by the A330P variant
of P450BM3 with different cofactor recycling systems. The total
reaction volume was 1.5 mL including 100 mM phosphate buffer,
pH 7.4, 0.5 µM P450BM3, 1 mM NADP+ (or in black the control
reaction without NADP+), 5 mM octane, and 20 µg ∙ mL–1 bovine
liver catalase under 90 % H2:10 % air (i.e. 2 % O2) with either, in
blue, glucose dehydrogenase (GDH, 10 U ∙ mL–1) and glucose
(0.2 M) or, in red, SHE341A/S342R (1 U/mL), for NADPH regeneration.
The total integrated areas of the product peaks for the
GDH/glucose and SH/H2 reactions are within 5 % of each other.
Asterisks denote background contaminant peaks.
Native P450BM3 and its A330P variant were studied for octane
oxidation (Scheme 2) with H2-driven NADPH regeneration with
the SHE341A/S342R variant.
Conclusion
Our results show that all of the engineered SH variants catalyse
H2-dependent NADP+ reduction, which is not the case for native
SH. The cofactor preference was switched from NAD+ to NADP+
in the case of the SHE341A/S342R and SHE341A/D467S proteins, and
importantly, these double variants of the SH are still able to
oxidise H2 in the presence of O2, with an O2-tolerance comparable
to the native SH. Notably, a further increase in NADP+-reducing
activity was not achieved with the SHE341A/S342R/D467S variant, which
determines the current limit for rational design. Further
improvements, however, might be possible by directed evolution
using a microorganism with NADPH deficiency, which can be
compensated by NADPH-producing SH. Nonetheless, the
increased catalytic efficiency for H2-driven NADP+ reduction
enables H2-driven biotransformations with enzyme cascades
containing NADPH-dependent oxidoreductases, such as imine
reductase and monooxygenase. In fact, the O2-tolerant
Scheme 2. P450BM3-catalysed oxidation of n-octane into 2-octanol. H2-driven
NADPH recycling is mediated by the SHE341A/S342R protein.
The native P450BM3 enzyme and the A330P variant produce
different ratios of octanols and octanones in air when glucose
dehydrogenase (GDH) is used for NADPH regeneration,[28] and
therefore provide a useful test pair for exploring whether the
product distribution is perturbed by applying the SHE341A/S342R. A
safe mixture of 90 % H2:10 % air (i.e. approximately 2 % O2) was
selected for this work, and experiments were carried out in air at
ambient temperature. We first set out to determine whether 10%
air is sufficient to support P450BM3 activity by comparing octane
6
This article is protected by copyright. All rights reserved.

10.1002/cctc.202000763
ChemCatChem
FULL PAPER
SHE341A/S342R variant supported H2-driven NADPH-dependent
octane oxidation by cytochrome P450BM3 as effectively as the
industry standard GDH/glucose system with the advantage of
guaranteeing 100% atom efficiency and no generation of carbon-
based by-products from the cofactor recycling. The
biotransformation was performed with a safe gas mixture
containing 2 % O2 and 90 % H2 (remainder N2). To our knowledge,
this is the first demonstration of H2-driven NADPH recycling by a
hydrogenase in the presence of O2. Notably, use of H2 and other
flammable reagents is well established in continuous flow
technologies where the continuous addition of low levels of
reactant provides for safe operation. Using H2 as reductant, we
have been able to translate biocatalytic reactions into continuous
processing,[40] and this approach would likely enable the O2-
dependent reactions to be carried out with improved control,
safety and productivity. The ability to use H2 as a reducing agent
for NADPH recycling opens up opportunities for highly atom-
efficient catalysis by NADPH-dependent enzymes, providing a
versatile platform for future biotransformations with NADPH
dependent oxidoreductases.
press cell at 18,000 psi, debris was separated from the soluble extract by
centrifugation at approx. 140,000 x g for 45 min. The supernatant was
loaded onto 2 mL Strep-Tactin Superflow columns (IBA), which were
previously equilibrated with resuspension buffer. After washing with at
least 6 column-volumes of resuspension buffer the protein was eluted in
resuspension buffer containing 5 mM desthiobiotin. A final concentration
of 2 – 6 mg mL–1 of purified protein was achieved using Ultra Centrifugal
Filter Units (Amicon). From 1 g of cell mass (wet weight), we usually
obtained 1 mg of purified SH. The protein concentration was determined
according to the method of Bradford.[43]
Hydrogenase activity measurements
H2-dependent NAD(P)+ reduction activities were determined as described
previously.[44] The reaction was performed in rubber-sealed cuvettes filled
with 2 mL of 50 mM Tris-HCl, pH 8.0, 1 mM NAD(P)+, 1 mM DTT, and
1 µM FMN. The solution was flushed with H2, and the reaction was started
by adding 2 – 5 µL of purified protein. The absorbance change through
accumulation of NAD(P)H was measured at 340 nm with a Cary 50
(Varian) spectrophotometer. For determination of Michaelis-Menten
parameters, initial reaction velocities at varying NAD(P)+ concentrations
were recorded. The velocities were plotted against substrate
concentrations and fitted to the Michaelis-Menten equation using Origin
9.1. Activity tests in the presence of O2 were performed in rubber-sealed
cuvettes containing appropriate mixtures of gas-saturated (H2, N2, and O2)
buffers.[18]
Experimental Section
Genetic construction of SH derivatives
To facilitate mutagenesis of the genes encoding the NAD+ reductase
module, HoxFU, of the Ralstonia eutropha SH, a broad-host-range vector
harbouring just the hoxFU genes under control of the SH promoter was
constructed. A 2823-bp XbaI-BamHI fragment was excised from plasmid
pJH#2904 and ligated into the vector pCM66 digested with the same
restriction endonucleases. The resulting plasmid, pJP66, encodes the N-
terminally Strep-tagged HoxF subunit as well as the HoxU protein and
served as template for site-directed mutagenesis. Introduction of site-
Electrochemical activity measurements
Electrochemical experiments were performed following previously
[24]
described procedures
in an anaerobic glove box (mBraun, O2 < 0.1
ppm) using an Autolab potentiostat (EcoChemie 128N). The pyrolytic
graphite edge rotating disc electrode (RDE) with planar electrode surface
area of 0.03 cm2 was prepared ‘in house’, and was rotated at 2000 rpm for
all experiments using an EG&G electrode rotator model 636 to ensure
efficient mass transport of substrate and product to and from the enzyme
films. Before each scan the electrode was polished using a slurry of 1 µm
alumina, and was sonicated in purified water (MilliQ, 18 MΩ ∙ cm) for 10 s.
The electrode was then rinsed and dried before application of an enzyme
film by spotting enzyme (0.5 µL) onto the electrode and leaving it to adsorb
over ~20 seconds. The electrode was then placed into buffered electrolyte
(100 mM Tris-HCl buffer, pH 8.0) containing 5 mM cofactor. Cyclic
voltammograms were recorded at a scan rate of 10 mV ∙ s–1. The
reference electrode was a saturated calomel electrode (SCE, BAS) and
potentials were converted to units of Volts (V) vs. SHE using E(SHE) =
E(SCE) + 0.242 V at 25 °C. The counter electrode was a Pt wire.
specific mutations was performed by
a modified QuikChange™
protocol.[41] For the E341A and D467S exchanges in HoxF, primers 5’-
ATC TGC GGC GAC gca TCG GCG CTC ATC- 3’ and 5’- CGC AAG CTC
GCG TAC GAA TCG CTT tcg TGC AAT GGC GCC-3’, respectively, were
used (exchanged bases in lower case letters). Desired nucleotide
exchanges in the resulting plasmids, pJP01 and pJP02, respectively, were
confirmed by sequence analysis using primers 5’-GTT ATC GCC GCG
TAT GC-3’ and 5’-GAT CGA ACG ATG GCA GC-3’. Plasmid pJP01 then
served as template to introduce the D467S exchange using the
aforementioned primer. The resulting plasmid encoding HoxFE341A/D467S
-
HoxU was named pJP03. To construct the E341A/S342R double
exchange, the primer 5’-CTT ATA TCT GCG GCG ACG CAc gc G CGC
TCA TCG AGT CCT GCG-3’ was designed and used with pJP1 to yield
pJP04. Plasmid pJP04, in turn, served as the template to introduce a
further D467S exchange by employing primer 5’- CGC AAG CTC GCG
TAC GAA tcg CTT TCG TGC AAT GGC GCC-3’. This yielded plasmid
pJP5 encoding HoxFE341A/S342R/D467S. Plasmids pJP66 and pJP01-05 were
transferred via conjugation from E. coli S17-1 to Ralstonia eutropha strain
HF903,[24] which carries in-frame deletions in hoxFU and the large subunits,
hoxG and hoxC, of the membrane-bound (MBH) and regulatory
hydrogenase (RH), respectively. Thus, this strain is SH-, MBH- and RH-
negative but possesses all other genes that are necessary for
hydrogenase maturation.
H2-driven IRED reaction
Reaction mixtures with a total volume of 1 mL consisted of H2 saturated
Tris-HCl buffer (100 mM, pH 7.5), 0.1-1 mM NADP+, 1.3–2 g ∙ L–1
substrate, 1 mg ∙ mL–1 IRED (Prozomix Ltd., PRO-CoE-IRED(046)) and
0.11 mg ∙ mL–1 SHE341A/S342R variant (1 U ∙ mg–1). These were prepared in
a stainless steel pressure vessel (Tinyclave steel, Büchi AG) fitted with a
50 mL glass insert equipped with a magnetic stirrer bar and sealed in a H2
atmosphere (2 bar). The reaction mixtures were stirred (200 rpm) at 32 °C
for 16–40 h and aliquots were removed at the end point for analysis by GC
and 1H NMR.
Production and purification of NADP+-reducing SH derivatives
Control reactions (Table S1) were performed on a 1 mL scale in round
bottom flasks (10 mL) and included a magnetic stirrer bar. Initially, reaction
mixtures consisting of 1 mM NAD(P)+ in Tris-HCl buffer (100 mM, pH 8.0)
were saturated with H2, and then SH or SH variant (to a concentration
0.11 mg mL–1, 0.11 U) was added. After 5 minutes, IRED was added
(100 µL from 10 × concentrated stock solution, 1 mg ∙ mL–1, 0.04 U). After
a further 5 min, substrate was added (2-methyl-1-pyrroline, 25 mM) and
the reaction was stirred at 200 rpm. The end solution was analysed by GC
and 1H NMR.
R. eutropha HF903 transconjugants carrying either of pJP66 and pJP01-
04 were grown heterotrophically in a mineral salts medium containing a
mixture of 0.05 ꢀ% (w/v) fructose and 0.4ꢀ % (v/v) glycerol (FGN medium)
at 30 °C until the optical density at 436 nm was 9 – 11.[42] Cells were
harvested by centrifugation at 11,500 x g for 15 min at 4 °C. The cell pellets
were resuspended in 50 mM Tris-HCl, pH 8, containing 150 mM KCl, 5 %
(v/v) glycerol, 5 mM NAD+ and protease-inhibitor cocktail (EDTA-free
Protease Inhibitor, Roche). After two passages through a chilled French
7
This article is protected by copyright. All rights reserved.

10.1002/cctc.202000763
ChemCatChem
FULL PAPER
Glucose dehydrogenase (GDH) control reactions were performed on a
1 mL scale in capped glass vials (2 mL) and stirred with a magnetic stirrer
bar. The reaction was initiated by addition of 1 mg of GDH (Codexis, GDH-
105) and 1 mg of IRED (both added as 100 µL from 10 × concentrated
stock solutions) to a solution of 1 mM NADP+, 20 mM glucose, and 5 mM
imine in Tris-HCl (100 mM, pH 8.0) and stirred at 200 rpm overnight. The
end solution was analysed by 1H NMR.
10 % air and with GDH/glucose as the NADPH regeneration system, the
mixtures contained 10 U · mL–1 GDH and 0.2 M glucose in place of the SH
variant. Such high concentration of glucose is used in order to ensure that
cofactor recycling is not limiting the reaction. Control reactions included
mixtures without NADP+ but under 2 % O2, and those using GDH/glucose
for NADPH regeneration in air.
Gas chromatography for IRED reaction
Acknowledgements
Following the IRED reactions described above, a 100-µL aliquot of the
solution was basified by addition of 10 µL of 10 N NaOH. The resulting
solution was extracted by addition of 200 µL EtOAc (containing 2 mM
undecane as internal standard), then centrifuged at 10 000 × g for 3 min.
The organic layer was dried over MgSO4, then centrifuged at 10,000 × g
for 3 min. Subsequently, 100 μL of the dried organic layer was heated to
60 °C with 15 µL of acetic anhydride and 7 µL of pyridine for 1 h in a
capped glass GC vial. The resulting solution was analysed by chiral GC-
FID (Figures S4-S6) using an Agilent CP-Chirasil-Dex CB column (25 m
length, 0.25 mm diameter, 0.25 μm film thickness), fitted with a guard of
10 m deactivated fused silica of the same diameter. Carrier gas: He (CP
grade), 2 mL ∙ min–1 (constant flow). Inlet temperature = 200 °C. Injection
volume = 0.1 μL.
We are indebted to Janna Schoknecht for skillful assistance.
Research by J.P., L.L. and O.L. was funded by the Deutsche
Forschungsgemeinschaft (DFG, German Research Foundation)
under Germany´s Excellence Strategy – EXC 2008 – 390540038
– UniSysCat., and the research project 405325648 (to L.L.).
Research by H.A.R. and K.A.V. was supported by an industry
interaction voucher from the Metals in Biology BBSRC NIBB
BB/L013711/1, and by EPSRC IB Catalyst award EP/N013514/1,
which also supported J.N.. K.U. was supported by EPSRC DTA
studentship, 2012, EP/K503113/1, and J.P. was supported by a
scholarship from the Berlin International Graduate School for
Natural Science & Engineering (BIG-NSE).
Oven heating profile:
Time / min
Temperature
Keywords: hydrogenase, metalloenzyme, cytochrome P450,
monooxygenase, oxidoreductase, imine reductase, octane
oxidation, nicotinamide cofactor, NADH, NADPH, cofactor
recycling, biotransformation
0  5
Hold at 70 ˚C
Ramp to 180 °C at 5 °C ∙ min–1
5 27
27  29
Hold at 180 °C for 2 min
[1] a) A. Liese, M. V. Filho, Curr. Opin. Biotechnol. 1999, 10,
595; b) A. J. J. Straathof, S. Panke, A. Schmid, Curr. Opin.
Biotechnol. 2002, 13, 548; c) B. M. Nestl, B. A. Nebel, B.
Hauer, Curr. Opin. Biotechnol. 2011, 15, 187; d) D. J.
Pollard, J. M. Woodley, Trends Biotechnol. 2007, 25, 66.
[2] a) D. Muñoz Solano, P. Hoyos, M. J. Hernáiz, A. R.
Alcántara, J. M. Sánchez-Montero, Bioresour. Technol.
2012, 115, 196; b) W. Kroutil, H. Mang, K. Edegger, K.
Faber, Curr. Opin. Chem. Biol. 2004, 8, 120.
[3] R. Bernhardt, V. B. Urlacher, Appl. Microbiol. Biotechnol.
2014, 98, 6185.
[4] C.-H. Wong, G. M. Whitesides, J. Am. Chem. Soc. 1981,
103, 4890.
NMR spectroscopy for IRED reaction
All NMR spectroscopy was carried out on a Bruker Avance III HD nanobay
(400 MHz) or a Bruker Avance III HD (500 MHz) (Figures S7-S10).
Following removal of denatured enzyme by centrifugation, 450 μL of the
sample solution was transferred to a Norell® SelectSeries™ 5 mm
400 MHz sample tube. A further 50 μL of D2O was added for field locking.
H2-driven octane oxidation by P450BM3
A mixture with a total volume of 1.5 mL consisting of 100 mM phosphate
buffer, pH 7.4, 0.5 µM P450BM3 enzyme, 1 mM NADP+, 20 µg ∙ mL–1
catalase, and 100 µL of a solution of the SHE341A/S342R variant (NAD+
reduction activity = 1 nmol · min–1 · µL–1) was prepared in a 6 mL (23 × 38
mm) glass vial in a glove box. O2 was removed from all the solutions, and
the phosphate buffer was saturated with H2 for 30 min before being taken
into the glove box. The glass vial was sealed with a suba seal, and the
reaction mixture was bubbled with H2 for a further 30 s and H2 was flushed
into the headspace for 15 seconds. The sealed glass vial was taken out of
the glove box and octane was added as a 0.2 M stock in DMSO to a final
concentration of 5 mM using a gas-tight syringe. The glass vial was then
injected with 10 % air to give 2 % O2 overall in both the solution and the
headspace. The vial, with the suba seal in place, was wrapped in Parafilm
to minimise gas loss. The reaction mixture was stirred for 2 h at ambient
temperature, after which a 1 mL aliquot was removed and extracted with
300 µL ethyl acetate. After vortex mixing and centrifugation (14,333 × g)
for 2 min, the organic layer was removed and analysed by gas
chromatography on a TRACE GC instrument (ThermoFinnegan) with a
SPB-1 fused silica capillary column (60 m × 0.53 mm i.d. and 1 µm film
thickness) and flame ionization detection. Helium was used as the carrier
gas (flow rate 3.5 mL · min–1). The injector was at 230 °C and the FID at
250 °C. The oven temperature was held at 100 °C for 40 min; the retention
times were: 3-octanone, 34.4 min; 2-octanone, 35.2 min; 4-octanol,
35.9 min; 3-octanol, 37.3 min; 2-octanol, 38.1 min. For the reactions under
[5] a) J. M. Vrtis, A. K. White, W. W. Metcalf, W. A. van der
Donk, Angew. Chem. Int. Ed. 2002, 41, 3257; b) E. Keinan,
E. K. Hafeli, K. K. Seth, R. Lamed, J. Am. Chem. Soc.
1986, 108, 162.
[6] K. Hoelsch, I. Sührer, M. Heusel, D. Weuster-Botz, Appl.
Microbiol. Biotechnol. 2013, 97, 2473.
[7] a) D. Holtmann, M. W. Fraaije, Arends, Isabel W C E, D. J.
Opperman, F. Hollmann, Chem. Commun. (Camb) 2014,
50, 13180; b) L. Lauterbach, O. Lenz, K. A. Vincent, FEBS
J. 2013, 280, 3058.
[8] a) L. Greiner, I. Schröder, D. H. Müller, A. Liese, Green
Chem. 2003, 5, 697; b) R. Mertens, L. Greiner, van den
Ban, Eyke C.D, H. B.C.M. Haaker, A. Liese, J. Mol. Catal.
B Enzym 2003, 24-25, 39; c) R. Mertens, A. Liese, Curr.
Opin. Biotechnol. 2004, 15, 343; d) J. Ratzka, L.
Lauterbach, O. Lenz, M. B. Ansorge-Schumacher, J. Mol.
Catal. B Enzym 2012, 74, 219.
[9] J. Ratzka, L. Lauterbach, O. Lenz, M. B. Ansorge-
Schumacher, Biocatal. Biotransformation 2011, 29, 246.
[10] C.-H. Wu, P. M. McTernan, M. E. Walter, M. W. W. Adams,
Archaea (Vancouver, B.C.) 2015, 2015, 912582.
[11] P. Kwan, C. L. McIntosh, D. P. Jennings, R. C. Hopkins, S.
K. Chandrayan, C.-H. Wu, M. W. W. Adams, A. K. Jones,
J. Am. Chem. Soc. 2015, 137, 13556.
8
This article is protected by copyright. All rights reserved.

10.1002/cctc.202000763
ChemCatChem
FULL PAPER
[12] a) S. G. Bell, X. Chen, F. Xu, Z. Rao, L.-L. Wong, Biochem.
Soc. Trans. 2003, 31, 558; b) C. F. Harford-Cross, A. B.
Carmichael, F. K. Allan, P. A. England, D. A. Rouch, L. L.
Wong, Protein Eng. 2000, 13, 121; c) L. L. Wong, Curr.
Opin. Chem. Biol. 1998, 2, 263.
[39] S. Castrignanò, G. Di Nardo, S. J. Sadeghi, G. Gilardi, J.
Inorg. Biochem. 2018, 188, 9.
[40] C. Zor, H. A. Reeve, J. Quinson, L. A. Thompson, T. H.
Lonsdale, F. Dillon, N. Grobert, K. A. Vincent, Chem.
Commun. (Camb.) 2017, 53, 9839.
[13] C. J. C. Whitehouse, S. G. Bell, L.-L. Wong, Chem. Soc.
Rev. 2012, 41, 1218.
[41] T. Pfirrmann, A. Lokapally, C. Andreasson, P. Ljungdahl, T.
Hollemann, PLoS One 2013, 8, e64870.
[14] a) A. K. Holzer, K. Hiebler, F. G. Mutti, R. C. Simon, L.
Lauterbach, O. Lenz, W. Kroutil, Org. Lett. 2015, 17, 2431;
b) H. A. Reeve, L. Lauterbach, O. Lenz, K. A. Vincent,
ChemCatChem 2015, 7, 3480; c) H. A. Reeve, L.
Lauterbach, P. A. Ash, O. Lenz, K. A. Vincent, Chem.
Commun. (Camb) 2011, 48, 1589.
[42] E. Schwartz, U. Gerischer, B. Friedrich, J. Bacteriol. 1998,
180, 3197.
[43] M. M. Bradford, Anal. Biochem. 1976, 72, 248.
[44] T. Burgdorf, O. Lenz, T. Buhrke, E. van der Linden, A. K.
Jones, S. P. Albracht, B. Friedrich, J. Mol. Microbiol.
Biotechnol. 2005, 10, 181.
[15] T. H. Lonsdale, L. Lauterbach, S. Honda Malca, B. M.
Nestl, B. Hauer, O. Lenz, Chem. Commun. 2015, 51,
16173.
[16] P. Vandamme, T. Coenye, Int. J. Syst. Evol. Microbiol.
2004, 54, 2285.
[17] K. Schneider, H. G. Schlegel, Biochem. J. 1981, 193, 99.
[18] L. Lauterbach, O. Lenz, J. Am. Chem. Soc. 2013, 135,
17897.
[19] a) T. Burgdorf, E. van der Linden, M. Bernhard, Q. Y. Yin,
J. W. Back, A. F. Hartog, A. O. Muijsers, C. G. de Koster,
S. P. J. Albracht, B. Friedrich, J .Bacteriol. 2005, 187,
3122; b) K. Schneider, H. G. Schlegel, Biochim. Biophys.
Acta 1976, 452, 66.
[20] J. T. Wu, L. H. Wu, J. A. Knight, Clinical Chem. 1986, 32,
314.
[21] N. S. Scrutton, A. Berry, R. N. Perham, Nature 1990, 343,
38.
[22] J. K. B. Cahn, C. A. Werlang, A. Baumschlager, S.
Brinkmann-Chen, S. L. Mayo, F. H. Arnold, ACS Synth.
Biol. 2017, 6, 326.
[23] a) S. J. Pilkington, J. M. Skehel, R. B. Gennis, J. E. Walker,
Biochemistry 1991, 30, 2166; b) R. G. Efremov, L. A.
Sazanov, Curr. Opin. Struct. Biol. 2011, 21, 532.
[24] L. Lauterbach, Z. Idris, K. A. Vincent, O. Lenz, PLoS One
2011, 6, e25939.
[25] M. Horch, L. Lauterbach, O. Lenz, P. Hildebrandt, I.
Zebger, FEBS Lett. 2012, 586, 545.
[26] K. Morina, M. Schulte, F. Hubrich, K. Dorner, S. Steimle, S.
Stolpe, T. Friedrich, J. Biol. Chem. 2011, 286, 34627.
[27] A. Lerchner, A. Jarasch, W. Meining, A. Schiefner, A.
Skerra, Biotechnol. Bioeng. 2013, 110, 2803.
[28] C. J. C. Whitehouse, S. G. Bell, H. G. Tufton, R. J. P.
Kenny, L. C. I. Ogilvie, L.-L. Wong, Chem. Commun.
(Camb.) 2008, 966.
[29] Y. Shomura, M. Taketa, H. Nakashima, H. Tai, H.
Nakagawa, Y. Ikeda, M. Ishii, Y. Igarashi, H. Nishihara, K.-
S. Yoon et al., Science 2017, 357, 928.
[30] J. M. Berrisford, L. A. Sazanov, J. Biol. Chem. 2009, 284,
29773.
[31] L. A. Sazanov, P. Hinchliffe, Science 2006, 311, 1430.
[32] R. Takase, A. Ochiai, B. Mikami, W. Hashimoto, K. Murata,
Biochim. Biophys. Acta 2010, 1804, 1925.
[33] O. Schmitz, G. Boison, H. Salzmann, H. Bothe, K. Schutz,
S.-h. Wang, T. Happe, Biochim. Biophys. Acta 2002, 1554,
66.
[34] P. Hinchliffe, L. A. Sazanov, Science 2005, 309, 771.
[35] E. Aubert-Jousset, M. Cano, G. Guedeney, P. Richaud, L.
Cournac, FEBS J. 2011, 278, 4035.
[36] O. Carugo, P. Argos, Proteins 1997, 28, 10.
[37] F. P. Guengerich, Nat. Rev. Drug Discov. 2002, 1, 359.
[38] a) Z. J. Wang, H. Renata, N. E. Peck, C. C. Farwell, P. S.
Coelho, F. H. Arnold, Angew. Chem. Int. Ed. Engl. 2014,
53, 6810; b) G. Di Nardo, V. Dell’Angelo, G. Catucci, S. J.
Sadeghi, G. Gilardi, Arch. Biochem. Biophys. 2016, 602,
106; c) J. A. Dietrich, Y. Yoshikuni, K. J. Fisher, F. X.
Woolard, D. Ockey, D. J. McPhee, N. S. Renninger, M. C.
Y. Chang, D. Baker, J. D. Keasling, ACS Chem. Biol. 2009,
4, 261.
9
This article is protected by copyright. All rights reserved.

10.1002/cctc.202000763
ChemCatChem
FULL PAPER
Entry for the Table of Contents
Insert graphic for Table of Contents here.
Insert text for Table of Contents here.
By protein engineering, we converted the NAD+ binding pocket of an O2-tolerant [NiFe]-hydrogenase into a NADP+ binding pocket.
The resulting biocatalyst supported H2-dependent NADPH recycling in coupled enzyme reactions involving imine reductase and P450
monoxygenase. The modified hydrogenase mediates atom-efficient cofactor recycling in challenging biotransformations with
oxidoreductases that depend on O2 and NADPH as co-substrates.
10
This article is protected by copyright. All rights reserved.