Finding the Switch: Turning a baeyer-villiger monooxygenase into a NADPH Oxidase

Article abstract of DOI:10.1021/ja508265b

By a targeted enzyme engineering approach, we were able to create an efficient NADPH oxidase from a monooxygenase. Intriguingly, replacement of only one specific single amino acid was sufficient for such a monooxygenase-to-oxidase switch - a complete transition in enzyme activity. Pre-steady-state kinetic analysis and elucidation of the crystal structure of the C65D PAMO mutant revealed that the mutation introduces small changes near the flavin cofactor, resulting in a rapid decay of the peroxyflavin intermediate. The engineered biocatalyst was shown to be a thermostable, solvent tolerant, and effective cofactor-regenerating biocatalyst. Therefore, it represents a valuable new biocatalytic tool.

Full text of DOI:10.1021/ja508265b

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
Communication
pubs.acs.org/JACS
Finding the Switch: Turning a Baeyer−Villiger Monooxygenase into a
NADPH Oxidase
Patrícia B. Brondani,† Hanna M. Dudek,† Christian Martinoli,‡ Andrea Mattevi,*,‡
and Marco W. Fraaije*,†
†Molecular Enzymology Group, University of Groningen, Nijenborgh 4, 9747AG Groningen, The Netherlands
‡Department of Biology and Biotechnology, University of Pavia, Via Ferrata 1, 27100 Pavia, Italy
S
* Supporting Information
ABSTRACT: By a targeted enzyme engineering ap-
proach, we were able to create an efficient NADPH
oxidase from a monooxygenase. Intriguingly, replacement
of only one specific single amino acid was sufficient for
such a monooxygenase-to-oxidase switcha complete
transition in enzyme activity. Pre-steady-state kinetic
analysis and elucidation of the crystal structure of the
C65D PAMO mutant revealed that the mutation
introduces small changes near the flavin cofactor, resulting
in a rapid decay of the peroxyflavin intermediate. The
engineered biocatalyst was shown to be a thermostable,
solvent tolerant, and effective cofactor-regenerating
biocatalyst. Therefore, it represents a valuable new
biocatalytic tool.
Figure 2. Catalytic mechanism of phenylacetone monooxygenase
showing the Baeyer−Villiger oxidation and uncoupling reaction (red)
routes. Several flavin intermediates are highlighted in color (oxidized
flavin in green, reduced flavin in magenta, and peroxyflavin in red) and
correlate to the colors used in Figure 3.
any redox enzymes depend on nicotinamide cofactors.
MSuch biocatalysts promote, for example, reduction of
ketones or oxidation of alcohols, including stereo- and
regioselective transformations.1 To avoid the use of equimolar
known, and only some of them can be produced and isolated.
Some NADH oxidases can be used with both NADH and
amounts of these expensive reagents, efficient and sustainable
strategies for cofactor regeneration are needed, which is one of
NADPH, but the activity for NADPH oxidation is usually low
the current challenges in biocatalysis. The most promising
when compared to NADH oxidation, and the respective
approach is the use of an ancillary enzyme at the expense of a
NAD(P)H oxidases may not be very robust.4
sacrificial cosubstrate.2
Baeyer−Villiger monooxygenases are flavin-containing en-
zymes that use O2 and NADPH to catalyze, among others, the
oxidation of ketones to esters.5 The oxygenation reactions
catalyzed by these biocatalysts proceed via a reactive flavin
intermediate, a flavin-peroxide, that is formed after the reduction
of the flavin cofactor by NADPH and a subsequent reaction with
O2 (Figure 2). In the absence of a suitable substrate, the
characteristic oxygen-activating peroxyflavin intermediate can
decay to yield oxidized flavin and hydrogen peroxide at the
expense of NADPH (Figure 2).6 The process of oxidizing
NADPH without catalyzing substrate oxygenation, referred to as
uncoupling, is typically very slow (<0.1 s−1). Albeit slow, this
unproductive mode of action of BVMOs effectively corresponds
to an NADPH oxidase activity.
While the regeneration of the reduced nicotinamide cofactors
(NAD[P]H) by enzymatic reactions is well established (for
example, by formate or phosphite dehydrogenases), regeneration
of the oxidized counterparts (NAD[P]+) still represents a
challenge.3 Some enzymes that regenerate NAD(P)+ by reducing
molecular oxygen (O2) have been reported. Such NAD(P)H
oxidases are typically flavin-containing enzymes where the flavin
prosthetic group acts as an electron mediator, transferring the
electrons from NAD(P)H to O2 (Figure 1). Although several
NADH oxidases are available, only a few NADPH oxidases are
Phenylacetone monooxygenase (PAMO) from Thermobifida
fusca is a thermostable and well-characterized BVMO.7 Besides
the exploration of its biocatalytic properties, several crystal
Figure 1. Example of a biocatalytic process in which a NADPH oxidase
Received: August 12, 2014
Published: November 25, 2014
catalyzes the regeneration of NADP+.
© 2014 American Chemical Society
16966
dx.doi.org/10.1021/ja508265b | J. Am. Chem. Soc. 2014, 136, 16966−16969

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
Journal of the American Chemical Society
Communication
structures of PAMO have been solved, and its catalytic
mechanism has been elucidated in detail (Figure 2).8 The
available structural information and overall robustness of PAMO
make this biocatalyst an ideal candidate for enzyme engineering.
In fact, we and other groups have succeeded in generating a
number of PAMO-based biocatalysts that can be used for a
variety of oxygenation reactions.9 During these studies we
observed that some PAMO mutants displayed slightly increased
uncoupling rates. For example, the Q152F and Q152F/L153G/
M446G PAMO mutants catalyze futile NADPH oxidation with a
rate of 0.5−0.6 s−1, while for wild-type PAMO the uncoupling
rate is only 0.01−0.02 s−1 (Table 1).7,8a,10 Careful screening of
SI, Table S2). These specific residue couples form distinct parts
of the active site of PAMO. The three libraries were screened
using our recently developed assay for detecting (1) NADPH
consumption by an enzyme coupled assay and (2) peroxide
detection by using the fluorogenic chemical probe Peroxy Green,
which specifically reacts with hydrogen peroxide.10 The Q152/
L153 library yielded no active mutants, whereas screening of the
two other libraries produced several clear hits with potentially
enhanced uncoupling. After sequencing of these clones, the
C65D/I67 V, C65D/I67S, C65D, and Q152Y/M446W
mutations were identified as the most promising hits. The
corresponding proteins were tested for Baeyer−Villiger mono-
oxygenase and NADPH oxidase activities. Although all mutants
achieved full substrate conversion after 24 h, uncoupling rates
were high, even compared to the previous generation of mutants
that displayed uncoupling activity. The best hit was C65D (Table
1), which exhibited a strong NADPH oxidase activity, and this is
in line with the fact that this amino acid replacement is shared by
three out of the four identified mutants.
Although the single C65D mutation on its own featured
interesting properties, we sought to explore the effects of
combined mutations. Because we did not observe C65D/I67P as
a hit in the screening, this mutant was constructed. In parallel, we
constructed a triple mutant combining C65D/Q152F/M446W
mutations. Although these mutants did not yield any Baeyer−
Villiger oxidation product (SI, Table S1), their NADPH oxidase
activities were also low (Table 1). We also introduced a
glutamate instead of an aspartate at position 65 because this
mutation was not present in the library, but no improvement was
observed. In line with these experiments, we also attempted
introducing an aspartate residue at homologous positions in
other BVMOs. However, the T56D mutant of cyclohexanone
monooxygenase showed no NADPH oxidase activity. Mutations
F54D and V190D introduced in acetone monooxygenase and 4-
hydroxyacetophenone monooxygenase, respectively, resulted in
impaired FAD binding. In summary, the C65D mutant of PAMO
turned out to be a unique hit, representing a potent NADPH
oxidase.
To explore the properties and biocatalytic potential of this
engineered NADPH oxidase, further characterization studies
were performed. The C65D PAMO mutant could be easily
produced and purified using the same protocol as used for wild-
type PAMO. By measuring the absorbance spectrum in the range
of 250−600 nm, it was confirmed that the FAD cofactor is well
bound to the protein (A280/A440 = 12).11 By performing oxygen
measurements we confirmed that, in the absence of a ketone,
C65D PAMO readily forms H2O2. When using 1.0 mM
phenylacetone, we observed that a significant amount (∼50%)
of H2O2 was still formed. This is in sharp contrast with wild-type
PAMO, which does not show any appreciable uncoupling activity
in the presence of phenylacetone. Using different concentrations
of NADPH and in the absence of phenylacetone, KM,NADPH = 3.5
μM and kcat = 5.0 s−1 were determined (Table 2). Thus, the
C65D mutation increases the rate of uncoupling by 2−3 orders
of magnitude (5.0 vs 0.01−0.02 s−1).7,8a These data also show
that this mutant enzyme behaves as a NADPH oxidase by
displaying a high oxidase activity and a relatively low Baeyer−
Villiger monooxygenase activity. In fact, the kinetic parameters
for the C65D PAMO mutant for oxidation of NADPH are in the
same range when compared with those of two bacterial
NAD(P)H oxidases (Table 2).
Table 1. Observed Rate of NADPH Oxidation by PAMO
a
Mutants
activity (s−1
)
in the absence of
phenylacetone
in the presence of
phenylacetone
enzyme variant
wild-type PAMO
I67P PAMO
0.02
0.10
0.17
n.d.
3.0
0.09
0.18
M446W PAMO
b
Q152F/L153G/M446G
PAMO
0.6
Q152F
0.5
0.5
Q152F/M446W PAMO
I67P/Q152F PAMO
C65D/I67V PAMO
C65D/I67S PAMO
C65D PAMO
0.38
0.13
2.1
0.47
0.15
2.2
2.6
2.0
4.9
3.8
Q152Y/M446W PAMO
C65D/I67P PAMO
1.3
1.1
0.05
0.43
0.05
0.27
C65D/Q152F/M446W
PAMO
C65E/PAMO
0.11
0.43
a
Activity was measured by monitoring the rate of absorbance decrease
at 340 nm at 25 °C for 3 min. The experiment was performed in the
presence (Baeyer−Villiger monooxygenase and/or NADPH oxidase
activity) or absence (NADPH oxidase activity) of phenylacetone. The
reaction mixture contained 0.1 mM NADPH, (1.0 mM phenyl-
acetone), 0.1−0.5 μM enzyme in 50 mM Tris-HCl pH 7.5 (final
b
volume 1.0 mL). Taken from ref 9a.
recently created site-saturation libraries led us to uncover even
more enzyme variants with increased NADPH oxidase
activities.10 Among them, M446W and I67P stood out, having
higher uncoupling rates (Table 1) and less (M446W) or no
(I67P) conversion of phenylacetone when compared with wild-
type PAMO (Supporting Information (SI), Table S1). These
results prompted us to embark on a dedicated search for PAMO
mutants with an enhanced and biocatalytically relevant NADPH
oxidase activity. As a first strategy, we decided to combine the
previously identified oxidase-promoting single-site mutations
into double mutants. However, no considerable increase in
NADPH oxidase activity was obtained for all tested double
mutants (Table 1). Therefore, we designed and constructed
three mutant libraries in which two residues were simultaneously
and randomly mutated: Q152&M446, C65&I67, and
Q152&L153. I67, Q152, L153, and M446 were chosen because
mutating these residues resulted in some increase in uncoupling
activity (Table 1 and refs 9 and 10), while C65 and I67 are the
residues closest to the locus of the peroxy moiety of the
peroxyflavin. At each position, 12 or 13 different amino acids
were allowed by choosing specific degenerate primers (details in
To probe the effect of the C65D mutation on biocatalyst
stability, we used the ThermoFAD method.12 This revealed an
16967
dx.doi.org/10.1021/ja508265b | J. Am. Chem. Soc. 2014, 136, 16966−16969

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
Journal of the American Chemical Society
Communication
NADPH oxidase. Wild-type PAMO has a very narrow substrate
acceptance range; therefore, the PAMO-derived NADPH
oxidase will be compatible with most reactions that require
NADP+ recycling.
To explain the specific and drastic effect of the C65D
mutation, we studied the catalytic mechanism of the engineered
NADPH oxidase in more detail. By using the stopped-flow
technique, we established that the rate of formation of
peroxyflavin is similar to that of wild-type PAMO (A→B
transition in Figure 3, see also Figure 2).8a However, the rate of
Table 2. Steady-State Kinetic Parameters for NAD(P)H
Oxidation by Wild-Type PAMO, C65D PAMO, and Two
Bacterial NADPH Oxidases
KM,NADPH
(μM)
kcat/KM
enzyme
kcat (s−1
)
(s−1 μM−1
)
PAMO
0.016
5.0
0.7
3.5
6.1
6
0.02
1.4
C65D PAMO
a
LsNOX
52
8.5
b
V193R/V194H SmNOX
20
3.3
a
LsNOX: NAD(P) oxidase from Lactobacillus sanfranciscus; kinetic
b
data taken from ref 4b. V193R/V194H SmNOX is a recently
engineered mutant of NADH oxidase from Streptococcus mutant which
accepts NADPH as substrate; kinetic data taken from ref 4d.
apparent melting temperature of C65D PAMO of 53.5 °C, which
is somewhat lower than that of the wild-type protein (59 °C).13
Most importantly, since many alcohol dehydrogenases are more
active at temperatures above room temperature, we investigated
the thermostability of the C65D mutant by incubation at 40 °C.
This revealed that the enzyme retained >80% activity for at least
72 h at this temperature (SI, Table S3). Incubating C65D PAMO
in 10% methanol or DMSO for 24 h also did not have a
detrimental effect on its activity (>90% activity was retained, SI,
Table S4). Clearly, the PAMO-based NADPH oxidase has
fundamentally retained the robustness of its parent enzyme.7,14
We applied C65D PAMO in reactions either using phosphite
and phosphite dehydrogenase, or using an organic substrate and
an alcohol dehydrogenase, in which the NADP+ cofactor is
regenerated by the engineered NADPH oxidase. The reactions
and the results are presented in Table 3, and further details about
Figure 3. Stopped-flow measurement of the reaction of 10 μM reduced
C65D PAMO upon rapid mixing with 120 μM dioxygen. Only seven
selected spectral scans are shown of the 1000 collected scans (total
reaction time 1.0 s). The inset shows the deconvolution of the spectra
and clearly shows, starting for reduced C65D PAMO (A), the initial
formation of the peroxyflavin intermediate (B) at a rate of 45 s−1, with its
absorbance maximum at 370 nm. The colors of the spectral species
comply with the color coding in Figure 2. The peroxyflavin subsequently
decays at a rate of 25 s−1 to form oxidized flavin (C) and H2O2.
Table 3. Using C65D PAMO as NADP+-Regenerating
a
Biocatalyst by a Combined Use with Dehydrogenases
decay of the peroxyflavin has increased tremendously (B→C
transition in Figure 3). While this reaction proceeds at a rate of
0.014 s−1 in wild-type PAMO, C65D PAMO is extremely rapid in
generating hydrogen peroxide (25 s−1).8a Such a drastic change is
in accordance with the observation that the NADPH oxidase
activity of C65D PAMO (5.0 s−1) is even higher when compared
to the Baeyer−Villiger monooxygenase activity of wild-type
PAMO (3.0 s−1).
b
c
entry
substrate
phosphite
dehydrogenase
conversion (%)
1
2
3
4
PTDH
ADHT
ADHE
ADHE
100
The crystal structure of the mutant enzyme was elucidated in
order to gain a further understanding of the specific effect of the
C65D mutation (crystallographic details in Supporting In-
formation, Table S5). Cys65 belongs to a loop that is in direct
contact with the C4a-N5 edge of the flavin (Figure 4). Cys and
Asp have almost isosteric side chains, and, indeed, the crystal
structures of C65D PAMO in both the oxidized and reduced
states are virtually identical to those of the wild-type enzyme
(rms deviations of 0.2 Å for all Cα atoms). The only variation can
be observed in the crystal structure of the oxidized mutant, where
Asp66 rotates toward the flavin as opposed to the conformation
observed in both the oxidized and reduced wild-type structures,
as well as the reduced C65D enzyme (Figure 4). Interestingly,
the same conformation has been observed in the complex of
PAMO with APADP+, a NADP(H) analogue which can function
in flavin reduction but causes substantial uncoupling due to a
reduced stability of the flavin peroxide.15 The reorientation of
Asp66 may favor intermediate decay by the combination of two
factors: (1) steric hindrance due to direct contact with the C4a
locus of the flavin, and (2) triggering of the protonation of the
intermediate proximal oxygen with consequent formation of
2-butanol
100
cyclohexanol
rac-1-phenylethanol
100
50 (ee >99%)
a
The reactions were performed for 24 h with 20 mM substrate and 0.5
b
mM NADP+. See Supporting Information Notes for details. Three
different dehydrogenases were used: phosphite dehydrogenase
(PTDH), alcohol dehydrogenase from T. brockii (ADHT), and alcohol
c
dehydrogenase evo-1.1.270 (ADHE). 100% conversion means the
substrate is fully converted.
the procedure can be found in the SI (Notes S1−S3). C65D
PAMO was effective in regenerating the cofactor, and no side
products from Baeyer−Villiger oxidation were observed, even
when different types of substrates were used. Moreover, we used
the enzyme in a kinetic resolution, leading to an enantiopure
chiral alcohol (Table 3, entry 4). The addition of catalase did not
affect the outcome of the reactions, and this suggests that the
tested biocatalysts are not very H2O2-sensitive. The observation
that no Baeyer−Villiger oxidation products were found illustrates
that wild-type PAMO is a good starting point to engineer a
16968
dx.doi.org/10.1021/ja508265b | J. Am. Chem. Soc. 2014, 136, 16966−16969

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
Journal of the American Chemical Society
Communication
AUTHOR INFORMATION
■
Corresponding Authors
andrea.mattevi@unipv.it
m.w.fraaije@rug.nl
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
The research leading to these results has received funding from
the 7th EU Framework Programme for research, technological
development and demonstration under grant agreement no.
212281. P.B.B. thanks CNPq for financial support.
Figure 4. Structural studies of C65D PAMO: oxidized C65D PAMO
(green carbons) superposed on the Cys65-Asp66 side chains of wild-
type PAMO (magenta carbons). Carbon atoms of the FAD and NADP+
are in yellow and cyan, respectively. Asp65 is engaged in H-bonds with a
water molecule and the backbone nitrogen of Ala91, which are not
shown for sake of clarity. See also Figure S9 and Table S5 in the SI.
REFERENCES
■
(1) Faber, K. Biotransformation in Organic Chemistry; Springer Verlag:
Berlin, 2002; pp 177−219.
(2) (a) Kara, S.; Schrittwieser, J. H.; Hollmann, F.; Ansorge-
Schumacher, M. B. Appl. Microbiol. Biotechnol. 2014, 98, 1517.
(b) Weckbecker, A.; Groger, H.; Hummel, W. Adv. Biochem. Eng.
̈
Biotechnol. 2010, 120, 195.
H2O2.15,16 It is very difficult to assign specific causes for the small
change in the Asp66 conformation, which likely reflects the
altered balance of electrostatic, H-bonding, and van der Waals
interactions with nearby protein, cofactor, and solvent atoms
caused by the C65D mutation. However, it appears that the
properties of the C65D PAMO mutant can be ascribed mainly to
stabilization of this uncoupling-promoting conformation of
Asp66. Interestingly, introducing an aspartate at the same
position in sequence-related monooxygenases did not result in
the creation of efficient NADPH oxidases. This indicates that the
effect of the C65D mutation in PAMO is subtle and unique.
In this work we report on an enzyme activity switch,
converting a thermostable and solvent-tolerant monooxygenase
into an oxidase. Intriguingly, it was sufficient to mutate only one
amino acid in the parent enzyme to achieve this switch. The
observation that a single mutation in a flavoprotein mono-
oxygenase results in a drastic boost in uncoupling activity also
presents a warning for enzyme engineering approaches that
concern flavoproteins. Yet, the observation that similar
mutations in sequence-related monooxygenases did not yield
NADPH oxidases may suggest that PAMO is rather unique. In
fact, we have generated numerous PAMO mutants in previous
studies and never observed such high NADPH oxidase activities.
The generated PAMO-based robust NADPH oxidase represents
a useful biocatalyst for cofactor regeneration, as only very few
alternative (bio)catalysts are available.17 Several specific features
(thermostability, solvent tolerance, tight binding of the flavin
cofactor, ease of production) make the engineered NADPH
oxidase a promising candidate for biocatalytic processes.
Furthermore, the biocatalyst may also develop into a useful
tool in cell biology research because NADPH oxidase activities
are crucial in triggering signal transduction pathways.18 More-
over, intracellular NADPH levels are crucial regulators of
metabolism.19 The created NADPH oxidase may also develop
as a tool in metabolic engineering to manipulate the intracellular
NADP+/NADPH ratio.
(3) Kochius, S.; Magnusson, A. O.; Hollmann, F.; Schrader, J.;
Holtmann, D. Appl. Microbiol. Biotechnol. 2012, 93, 2251.
(4) (a) Rocha-Martín, J.; Vega, D.; Bolivar, J. M.; Godoy, C. A.;
́ ́
Hidalgo, A.; Berenguer, J.; Guisan, J. M.; Lopez-Gallego, F. BMC
Biotechnol. 2011, 11, 101. (b) Lountos, G. T.; Jiang, R.; Wellborn, W. B.;
Thaler, T. L.; Bommarius, A. S.; Orville, A. M. Biochemistry 2006, 45,
9648. (c) Riebel, B. R.; Gibbs, P. R.; Wellborn, W. B.; Bommarius, A. S.
Adv. Synth. Catal. 2002, 344, 1156. (d) Petschacher, B.; Staunig, N.;
Muller, M.; Schurmann, M.; Mink, D.; De Wildeman, S.; Gruber, K.;
Glieder, A. Comput. Struct. Biotechnol. J. 2014, DOI: 10.5936/
csbj.201402005.
(5) de Gonzalo, G.; Mihovilovic, M. D.; Fraaije, M. W. ChemBioChem
2010, 11, 2208.
(6) Sheng, D.; Ballou, D. P.; Massey, V. Biochemistry 2001, 40, 11156.
(7) Fraaije, M. W.; et al. Appl. Microbiol. Biotechnol. 2005, 66, 393.
(8) (a) Torres Pazmino, D. E.; Baas, B. J.; Janssen, D. B.; Fraaije, M. W.
Biochemistry 2008, 47, 4082. (b) Malito, E.; Alfieri, A.; Fraaije, M. W.;
Mattevi, A. Proc. Natl. Acad. Sci. U.S.A. 2004, 101, 13157. (c) Orru, R.;
Dudek, H. M.; Martinoli, C.; Torres Pazmino, D. E.; Royant, A.; Weik,
M.; Fraaije, M. W.; Mattevi, A. J. Biol. Chem. 2011, 286, 29284.
̈
̈
̃
̃
(9) (a) Torres Pazmino, D. E.; Snajdrova, R.; Rial, D. V.; Mihovilovic,
̃
M. D.; Fraaije, M. W. Adv. Synth. Catal. 2007, 349, 1361. (b) Parra, L. P.;
Agudo, R.; Reetz, M. T. ChemBioChem 2013, 14, 2301. (c) Dudek, H.
M.; Fink, M. J.; Shivange, A. V.; Dennig, A.; Mihovilovic, M. D.;
Schwaneberg, U.; Fraaije, M. W. Appl. Microbiol. Biotechnol. 2014, 98,
4009.
(10) Dudek, H. M.; Popken, P.; van Bloois, E.; Duetz, W. A.; Fraaije, M.
W. J. Biomol. Screen. 2013, 18, 678.
(11) Dudek, H. M.; de Gonzalo, G.; Pazmino, D. E.; Stepniak, P.;
Wyrwicz, L. S.; Rychlewski, L.; Fraaije, M. W. Appl. Environ. Microbiol.
2011, 77, 5730.
(12) Forneris, F.; Orru, R.; Bonivento, D.; Chiarelli, L. R.; Mattevi, A.
FEBS J. 2009, 276, 2833.
̃
(13) van Beek, H. L.; de Gonzalo, G.; Fraaije, M. W. Chem. Commun.
2012, 48, 3288.
(14) de Gonzalo, G.; Ottolina, G.; Zambianchi, F.; Fraaije, M. W.;
Carrea, G. J. Mol. Catal. B: Enzym. 2006, 39, 91.
(15) Martinoli, C.; Dudek, H. M.; Orru, R.; Edmondson, D. E.; Fraaije,
M. W.; Mattevi, A. ACS Catal. 2013, 3, 3058.
(16) Bach, R. D.; Mattevi, A. J. Org. Chem. 2013, 78, 8585.
(17) Maid, H.; Bohm, P.; Huber, S. M.; Bauer, W.; Hummel, W.; Jux,
̈
ASSOCIATED CONTENT
■
N.; Groger, H. Angew. Chem., Int. Ed. 2011, 50, 2397.
̈
S
* Supporting Information
(18) Lambeth, J. D. Nat. Rev. Immunol. 2004, 4, 181.
(19) (a) Kleijn, R. J.; Liu, F.; van Winden, W. A.; van Gulik, W. M.; Ras,
C.; Heijnen, J. J. Metab. Eng. 2007, 9, 112. (b) Fan, J.; Ye, J.; Kamphorst,
J. J.; Shlomi, T.; Thompson, C. B.; Rabinowitz, J. D. Nature 2014, 510,
298.
Primer sequences, steady-state kinetic data of the reported
enzymes, crystallographic statistics, and experimental details
concerning conversions. This material is available free of charge
via the Internet at http://pubs.acs.org.
16969
dx.doi.org/10.1021/ja508265b | J. Am. Chem. Soc. 2014, 136, 16966−16969