Photoenzymatic reduction of C=C double bonds

Article abstract of DOI:10.1002/adsc.200900560

A simplified procedure for cell-free biocatalytic reductions of conjugated C=C double bonds using old yellow enzymes (OYEs) is reported. Instead of indirectly regenerating YqjM (an OYE homologue from B. subtilis) or NemA (N-ethylmaleimide reductase from E

Full text of DOI:10.1002/adsc.200900560

FULL PAPERS
DOI: 10.1002/adsc.200900560
Photoenzymatic Reduction of C=C Double Bonds
a
a
a
b
Maria Mifsud Grau, John C. van der Toorn, Linda G. Otten, Peter Macheroux,
c
d
a
Andreas Taglieber, Felipe E. Zilly, Isabel W. C. E. Arends,
and Frank Hollmann *
Delft University of Technology, Department of Biotechnology, Biocatalysis and Organic Chemistry Group, Julianalaan
36, 2628 BL Delft, The Netherlands
a,
a
1
Fax : (+31)-152-781-415; phone: (+31)-152-781-957; e-mail: f.hollmann@tudelft.nl
b
c
Graz University of Technology, Institute of Biochemistry, Petersgasse 12, 8010 Graz, Austria
Firmenich SA, Corporate R&D, Biotechnology & Bioengineering, Route des Jeunes 1, 1211 Geneva 8, Switzerland
Max-Planck-Institut fꢀr Kohlenforschung, Kaiser-Wilhelm-Platz 1, 45470 Mꢀlheim an der Ruhr, Germany
d
Received: August 8, 2009; Revised: November 22, 2009; Published online: December 10, 2009
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/adsc.200900560.
Abstract: A simplified procedure for cell-free biocat- scheme was characterized. Up to 65% rates of the
alytic reductions of conjugated C=C double bonds NADH-driven reaction were obtained while preserv-
using old yellow enzymes (OYEs) is reported. In- ing enantioselectivity. The chemoselectivity of the
stead of indirectly regenerating YqjM (an OYE ho- novel approach was exclusive. Even when using
mologue from B. subtilis) or NemA (N-ethylmale- crude cell extracts as biocatalyst preparations, only
ACHTUNGTRENNUNGi mide reductase from E. coli) via regeneration of re- C=C bond reduction was observed while ketone and
duced nicotinamide cofactors, we demonstrate that aldehyde groups remained unaltered. Overall, a
direct regeneration of catalytically active reduced simple and practical approach for photobiocatalytic
flavins is an efficient and convenient approach. Re- reductions is presented.
ducing equivalents are provided from simple sacrifi-
cial electron donors such as ethylenediaminetetra- Keywords: alkene reduction; asymmetric catalysis;
A
C
H
T
U
N
G
T
R
E
N
N
U
N
G
acetate (EDTA), formate, or phosphite via photoca- cofactor regeneration; enzyme catalysis; old yellow
talytic oxidation. This novel photoenzymatic reaction enzymes; photoenzymatic reduction
Introduction
The active reductant within OYEs is a reduced
flavin (reduced flavin adenine mononucleotide,
[
19,20]
Asymmetric reduction of C=C double bonds is one of FMNH2)
transferring a hydride in a Michael-type
the most widely used methodologies for the genera- reaction to the substrateꢁs b-C-atom. The resulting
[
1]
tion of chiral molecules and synthons. In addition to enolate anion is protonated in a trans-fashion (by a
the well-established transition metal-catalyzed (trans- tyrosine residue) at a-C, leaving the reduced product
[
2,3]
fer) hydrogenations
new methodologies such as or- and the oxidized flavin prosthetic group. A new cata-
[4]
[5]
ganocatalytic or biocatalytic approaches are con- lytic cycle is initiated by nicotinamide [NAD(P)H]-
stantly being added to the toolbox. For the latter, so- dependent reduction of the flavin (Scheme 1).
called enoate reductases [E.C. 1.3.1.X] (or old yellow
enzymes, OYEs) have recently experienced a re- tutes a major challenge for OYEs en route to efficient
This dependence on NAD(P)H, however, consti-
[6]
newed interest as catalysts for the asymmetric reduc- biocatalytic reductions. NAD(P)H is expensive there-
[
5]
tion of activated C=C-double bonds. For example, fore necessitating catalytic use and efficient in situ re-
+
enantiospecific reductions of a,b-unsaturated alde- generation from NAD(P) . In principle, a broad
[
7–12]
[13,14]
hydes and ketones,
acids and esters,
as well as range of NAD(P)H regeneration systems are avail-
[
15–17]
[18]
nitro compounds
ed. Especially their broad substrate scope combined tion can be chosen.
or nitriles have been report- able from which the most suitable for the given reac-
[
21–26]
To avoid additional efforts
with the often excellent enantioselectivity makes and expenditures of external NAD(P)H regeneration,
OYEs attractive biocatalytic counterparts to the exist- asymmetric bioreductions using OYEs are preferen-
ing chemical methods.
tially performed in vivo, thus utilizing the metabolic
Adv. Synth. Catal. 2009, 351, 3279 – 3286
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
3279

Maria Mifsud Grau et al.
FULL PAPERS
[19]
Scheme 1. Catalytic cycle of OYEs. EWG=electron-withdrawing group, for example, ÀCHO, ÀCOR, ÀCOOH/R, ÀNO2
etc.
2
+ [28–31]
NAD(P)H regeneration capacity of the living such as [Cp*Rh
cell.
A
H
U
G
R
N
N
(bpy)
A
H
U
G
R
N
U
G
(2) direct electro-
2
[9–11,14,27]
[32]
However, ubiquitous NAD(P)H-depen- chemical regeneration, and, (3) light-driven reduc-
[33–36,39,48]
dent ketoreductases can cause undesired carbonyl re- tion exploiting the photoexitability of flavins.
duction of substrates and products leading to lower Due to its simplicity, we decided to evaluate the
yields and complex product mixtures. Even using iso- latter strategy to directly regenerate OYEs
[
34]
lated OYEs together with an enzymatic NAD(P)H re- (Scheme 2).
Such an approach circumvents the
generation system the procedure can suffer from this costly and instable nicotinamide cofactor as well as a
[
8]
side reaction due to contaminating ketoreductases.
corresponding enzymatic regeneration system. As a
An alternative approach circumventing this overre- result a highly simplified and more chemoselective
duction challenge would be to directly regenerate the production system is obtained (Scheme 2).
flavin prosthetic group of OYEs. So, ketoreductases
present in the reaction mixture would not be regener- from Bacillus subtilis (YqjM, E.C. 1.6.99.1).
ated due to their exclusive NAD(P)H dependency. a model substrate we chose ketoisophorone.
As a model enzyme we chose the OYE homologue
[
37,40,41]
As
[
8,34]
For this, 3 methods are available: (1) chemical regen-
The aim of this study was to evaluate the efficiency
eration of FMNH for example, by transition metals of this photoenzymatic reduction and to investigate
2
Scheme 2. Comparison of: the ꢃtraditionalꢁ regeneration approach for OYEs utilizing (enzymatic) regeneration of NAD(P)H
(
left) and the proposed light-driven direct regeneration of OYEs utilizing the prosthetic group as photocatalyst (right). For
[35]
an overview over the proposed catalytic mechanisms, please refer to the Supporting Information.
3280
asc.wiley-vch.de
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2009, 351, 3279 – 3286

Photoenzymatic Reduction of C=C Double Bonds
FULL PAPERS
the influence of the single reaction parameters on its experiments; therefore, we exclude in situ racemiza-
performance.
tion as a major factor influencing enantioselectivity.
The rate of the non-enzymatic reduction was exam-
ined either in the absence of CE or using thermally
inactivated CE (4 h at 908C). In both cases, a back-
ground activity of less than 0.27 turnovers per minute
was determined for FMN in the non-enzymatic reduc-
tion, also not significantly contributing to the non-ex-
Results and Discussion
Enzyme Preparation
Envisaging a most simple catalyst preparation, we de- clusive enantioselectivity. Furthermore, increasing
cided to heterologously express YqjM in Escherichia [FMN], increasing the rate of the background reac-
coli (E. coli) and use the crude cell extracts (CE) as tion, did not negatively influence enantioselectivity
biocatalyst preparation. The CE contained approxi- (vide infra). However, using stoichiometric NAD(P)H
mately 27% YqjM (w/w of total soluble protein, Sup- as source of reducing equivalents, practically the same
porting Information). We estimated the concentration enantioselectivity (72% ee) was observed. Hence, we
À1
of active YqjM to be 0.17 mgmL (4.57 mM).
concluded that the enantioselectivity of the biocata-
lyst preparation was not optimal (vide infra) and that
the ꢃunnatural reaction conditionsꢁ of the simplified
reaction set-up did not impair enantioselectivity.
Reaction Set-Up
Initial photoenzymatic experiments were performed
at ambient temperature and atmosphere using FMN The Photocatalyst
as externally added photocatalyst. Under these condi-
tions only trace amounts of product were observed. Changing the photocatalyst from FMN to FAD or ri-
We attribute this to the high reactivity of reduced fla- boflavin as photocatalyst resulted in slight changes of
vins with O and concomitant formation of hydrogen both rate and enantioselectivity (Figure 1).
2
[42]
peroxide.
Deaeration of the reaction mixture
There was, however, no clear correlation between
rate and enantioselectivity on the one side and struc-
proved to be essential to achieve conversion.
Also the light source had a significant influence on ture of the photocatalyst on the other side. We attri-
the rate of the photoenzymatic reaction. Changing bute this variability to differences in the electron
the light source from a conventional projector lamp transfer rates between external reduced flavins and
(
40 W, set-up 1) to a more intensive light bulb (250 W, YqjM-bound FMN (the KD value of FMN in YqjM
[
37]
set-up 2) increased the productivity, under otherwise lays in the lower mM range ).
identical conditions, by more than 50% (Supporting
In all cases, anaerobic conditions were crucial to
Information). This set-up was used for all subsequent observe conversion. This becomes obvious consider-
reactions.
ing the high reactivity of reduced flavins with molecu-
lar oxygen yielding oxidized flavins and hydrogen per-
[
42]
oxide. However, when using synthetic, O -stable 5-
2
[
38,39]
Chemo- and Enantioselectivity
deazaflavin as photocatalyst,
rapid reduction of
Especially aldehydes are subject to undesired ketore-
[
8]
ductase-catalyzed reduction to alcohols. Therefore,
we tested citral (3,7-dimethyl-2,6-octadienal) as sub-
strate in the photoenzymatic reduction. Full conver-
sion into citronellal was observed without detectable
traces of either citronellol or geraniol or nerol. Thus
we conclude that even by using crude E. coli cell ex-
tracts containing various endogeneous ketoreductases,
chemoselective photobiocatalytic reduction of the
C=C-double bond is feasible. By selectively reducing
the flavin cofactor only enoate reductases such as
YqjM are regenerated while NAD(P)H-dependent
ketoreductases remain inactive.
The photoenzymatic reduction of ketoisophorone
proceeded with comparably poor enantioselectivity
Figure 1. Influence of flavin type on the rate (black) and
enantioselectivity (grey) of the photoenzymatic reduction
reaction. Riboflavin, FMN, and FAD (100 mM each): an-
(
69% ee). Therefore, we investigated a range of po-
tential causes. The optical purity of the product de-
creased by only 1–2% ee over the time course of the (100 mM each): aerobic conditions.
AHCTUNGTRENNUGNa erobic conditions; deazaflavin and deazaflavin+FMN
Adv. Synth. Catal. 2009, 351, 3279 – 3286
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
3281

Maria Mifsud Grau et al.
FULL PAPERS
ketoisophorone was observed even in the presence of
O albeit forming racemic product. Enantioselectivity
2
could partially be reconstituted (50% ee, Figure 1) by
addition of calaytic amounts of FMN to saturate
YqjMꢁs active site. This may be explained by the spe-
cificity of YqjM for FMN on the one hand and by
more negative reduction potential of deazaflavin
(
>100 mV more negative than, for example, ribofla-
[46]
vin) on the other hand. Reduced deazaflavin seems
to be a potent reductant for at least ketoisophorone
without being recognized at the chiral active site of
YqjM. Thus, in the absence of FMN only non-enan-
tioselective background reduction occurs. If YqjM is
saturated with FMN, also electron transfer from dea-
zaflavin to YqjM-bound FMN occurs resulting in
enzyme-related reduction competing with deazafla-
vin-mediated non-enantioselective reduction.
Figure 3. Influence of the reaction temperature on the maxi-
mal conversion rate (black) and optical purity (grey).
Initial rates correlated with the reaction tempera-
ture between 20 and 508C (Figure 3). However at
508C, productivity caved-in by more than 90% after
only 30 min. Likewise, only racemic product was pro-
Influence of Reaction pH and Temperature
The influence of pH was examined between pH 5.1 duced from this time on resulting in an overall ee of
and 9.4 showing a relatively broad activity maximum less than 25% after 4 h. Obviously, the enzyme stabili-
in the slightly alkaline range. Enantioselectivity slight- ty was greatly impaired at 508C leading to fast inacti-
ly decreased with increasing pH, which is most likely vation. Below 508C, only a slight decrease of enantio-
explained by increased racemization via enolization selectivity was observed over time indicating signifi-
(
Figure 2).
cantly higher enzyme stability.
Interestingly, also at pH 5, a significant reduction
activity was observed contradicting earlier observa-
[
35]
tions by Frisell et al. who explained poor photoca- Rate-Dependence on [EDTA] and [FMN]
talytic activity at low pH values with unfavourable N-
protonation resulting in ꢃmaskedꢁ N-lone pairs re- In the experiments performed so far, a nominal YqjM
À1
quired for efficient electron transfer to the photoex- activity of approximately 0.5 Umg was estimated
°
cited oxidized FMN (FMN ). Obviously, in case of corresponding to 7.1% of the activity determined in
+
EDTA [pK
A
C
H
T
U
N
G
T
R
E
N
N
U
N
G
(NH )=6.13 and 10.37] already at pH 5 the NAD(P)H-dependent activity assay under compa-
a
the concentration of unprotonated nitrogen is suffi- rable conditions. We suspected [FMNH ] to be signifi-
2
°
ciently high to enable efficient reduction of FMN . cantly below saturation concentration for YqjM.
The activity decrease at elevated pH (>8.5) is most Therefore a series of experiments investigating the
likely to be explained by decreasing YqjM activity. It effect of varying [EDTA] and [FMN] on the overall
should be mentioned, however, that so far a distinct reduction rate were performed to clarify this issue
pH (and T) profile of YqjM is lacking.
(Figure 4 and Figure 5).
The concentration of the sacrificial electron donor
(
[EDTA]) had no influence on the initial rate of the
photoenzymatic reduction [A
Umg , Figure 4]. Apparently, under the given condi-
tions [EDTA] does not influence the formation rate
spec A( YqjM)=0.64Æ0.03
H
U
G
R
N
U
G
À1
of FMNH suggesting flavin photoexcitation to be
2
rate-limiting.
EDTA contains per molecule 4 oxidizable N-alkyl-
amino acid moieties. Therefore, principally 0.25 equiv-
alents (based on substrate) should be sufficient to
obtain full conversion. Under these conditions, how-
ever, the conversion rate leveled off significantly
above 40–50% indicating that only 2 of the 4 oxidiza-
ble functionalities are available for FMN photoreduc-
Figure 2. Influence of pH on activity (^) and enantioselec-
tivity (^) of the photoenzymatic reduction of ketoisophor-
one. 100% corresponds to
a
initial productivity of tion.
À1
6.38 mMh
3282
asc.wiley-vch.de
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2009, 351, 3279 – 3286

Photoenzymatic Reduction of C=C Double Bonds
FULL PAPERS
Figure 6. Comparison of rates (diamonds) and enantioselec-
tivities (squares) photoenzymatic reductions of ketoisophor-
one using crude extracts of E.coli containing YqjM (filled
symbols) and devoid of YqjM (open symbols).
Most interestingly, however, was that the enantiose-
lectivity of the photoenzymatic reduction increased
with increasing FMN concentration. This was unex-
pected assuming an increased background reduction
Figure 4. Influence of [EDTA] on the rate of the photoenzy-
matic reduction. [EDTA]=2.5 mM (&), 5 mM (^), 10 mM
(
~).
at increased FMNH concentrations. Obviously, in situ
2
formed FMNH is efficiently used by YqjM for reduc-
2
tion. Even more astonishingly, inversion of enantiose-
lectivity was observed at FMN concentrations below
1
0 mM. One plausible explanation involves the pres-
ence of another, stereocomplementary biocatalyst
with higher affinity towards or faster electron transfer
rates with FMNH than YqjM (vide infra).
2
Enoate Reductase Activity in E. coli
As stated above, the peculiar inversion of enantiose-
lectivity at low [FMN] may be explained by the pres-
ence of an endogeneous E. coli enoate reductase ex-
hibiting inverse enantioselectivity and higher affinity
to FMNH . In fact, a putative OYE homologue,
2
[44]
Figure 5. Influence of [FMN] on the rate (^) and enantiose- NemA, had been reported in E. coli by Miura et al.
lectivity (^) of the photoenzymatic reduction reaction.
Therefore, as a control experiment, we performed the
light-driven reduction of ketoisophorone using a
A plausible explanation for this may be the in- crude cell extract from E. coli cells devoid of the plas-
creased basicity of the secondary amines generated in mid encoding for YqjM as biocatalyst. Indeed, under
the photooxidation process. Thus, the protonation these conditions reduction of ketoisophorone oc-
equilibrium may be shifted towards the ammonium curred with opposite enantioselectivity compared to
salts thereby slowing down the rate of the next elec- the use of YqjM as biocatalyst (Figure 6).
tron transfer step to photoexcited FMN (Supporting
Information).
Thus, we have strong indications that NemA is the
actual reason for the comparably poor enantioselec-
Quite surprising results were obtained varying tivity observed in the reduction of ketoisophorone.
FMN] (Figure 5). In accordance with our initial hy-
[
pothesis, the overall rate correlated with [FMN] sug-
gesting the formation rate of FMNH to be overall Alternative Electron Donors
2
rate-limiting. Accordingly, the specific activity deter-
À1
mined for YqjM increased to 4.6 Umg (65% of the Finally, we performed a preliminary screening for
nominal activity) in the presence of 0.5 mM FMN. suitable electron donors for the photoenzymatic re-
Adv. Synth. Catal. 2009, 351, 3279 – 3286
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
3283

Maria Mifsud Grau et al.
FULL PAPERS
Table 1. Alternative sacrificial electron donors tested.
Table 2. Performance summary of the photobiocatalytic re-
duction system.
[a]
Electron donor
Rate [%]
ee [%]
À1
À1
TF [min ]
TTN [molmol ]
EDTA
NaHCO2
100
51
69
1
0
FMN
YqjM
4
210
1000
10.900
Na HPO3
26
2
Methionine
Triethyl phosphite
Nicotine
14
20
20
>1
n.d.
30
n.d.
n.d.
The influence of various reaction parameters on
rate and enantioselectivity has been investigated put-
ting a solid basis for further preparative-scale applica-
tions of the photoenzymatic reduction reaction. Al-
ready at this early stage of development, the catalytic
Sarcosine
[
a]
The rate of using EDTA was set to 100%.
duction. EDTA may not be the optimal sacrificial re- performance of the photo- and biocatalyst are promis-
ductant for a variety of reasons. From an atom effi- ing, as shown in Table 2. A direct comparison with
ciency point-of-view the necessity of at least the reported traditional approaches is not straightfor-
0.6 equivalents for full conversion appears questiona- ward due to a lack of exact values.
ble. Furthermore, formaldehyde and amines as side One point of interest concerns the stability of the
products are also not unproblematic from an environ- flavin photocatalyst. Isoalloxazines such as Rf, FMN
mental and toxicological point of view. We therefore and FAD have been reported to be rather labile
[
47]
performed a preliminary screening of other potential- under irradiation conditions.
In contrast, the high
ly more suited electron donors such as formate or total turnover number observed for FMN in this
[
48]
phosphite (Table 1). In fact, EDTA could be substitut- study and some of our recent work suggest that the
À
2À
ed by HCO2 or HPO3 to achieve photoenzymatic reported photoinstability does not pose a severe limi-
reduction. tation for our system. This is also in line with the pro-
Rates were somewhat decreased compared to posed flavin-degradation products being photoactive
[
47]
EDTA but, more importantly, enantioselectivity was isoalloxazines.
reduced as well (Supporting Information). Currently Further work will concentrate on detailed analysis
we are lacking a plausible explanation for this obser- of the reaction mechanism including photodegrada-
vation. Perhaps the enzyme enantioselectivity is im- tion and enzyme kinetics using purified preparations
paired by the presence of these salts. Alternatively, of YqjM and NemA. Furthermore, we will substanti-
the reduced in situ concentration of FMNH (due to ate the very promising results using alternative, envi-
2
reduced reduction rate) may have a similar effect on ronmentally benign electron donors such as formate
enantioselectivity as observed with low total FMN or phosphite.
concentrations.
Thus, we are convinced to obtain a simple, prepara-
tively and economically attractive methodology for
enantioselective C=C bond reductions.
Conclusions
Light-driven reduction of flavins is a feasible method Experimental Section
for direct, cofactor-independent regeneration of
OYEs such as YqjM and NemA. Thus, photoenzymat- Production of YqjM
ic reductions circumvent the need for the nicotina-
Fermentative production of YqjM was performed according
mide cofactor and a corresponding (enzymatic) nicoti-
namide regeneration system. Using the simplified, ar-
tificial regeneration system rates of up to 65% of the
[37]
to the method reported by Macheroux et al. In brief, E.
TM
coli BL21-codon plus -DE3 ACHUTNGRENNUG( RIL) (Stratagene) containing
pET21a encoding the gene for YqjM was fermented on 10-L
ꢃnativeꢁ, NADH-driven reaction have been reached scale at 378C using LB medium. YqjM production was in-
while preserving the enantioselectivity. Another bene- duced at OD600 =0.6 with 1 mMfinal IPTG and fermentation
fit of the simplified system is that NAD(P)H-depen- was continued for 4 h. The harvested cells were resuspended
in potassium phosphate buffer (50 mM, pH 7.5) and broken
dent reduction of keto groups is circumvented yield-
by two passages through a French press and centrifuged.
ing a more chemoselective alternative to ꢃclassicalꢁ re-
The resulting clear crude cell extract was supplemented with
generation approaches. Unfortunately, enantioselec-
5
mMfinal FMN and stored at À208C. Under these conditions
tivity in the present system is impaired by the pres-
ence of two, enantiocomplementary enoate reductases
enzyme activity was not impaired for at least 6 months.
(
recombinant YqjM and endogeneous NemA). Using
purified enzyme preparations, we expect also to
obtain a highly enantioselective system.
3284
asc.wiley-vch.de
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2009, 351, 3279 – 3286

Photoenzymatic Reduction of C=C Double Bonds
FULL PAPERS
Activity Assay
[8] M. Hall, C. Stꢀckler, H. Ehammer, E. Pointner, G.
Oberdorfer, K. Gruber, B. Hauer, R. Stꢀrmer, W.
Kroutil, P. Macheroux, K. Faber, Adv. Synth. Catal.
NADH-dependent YqjM activity was assessed spectropho-
tometrically at 340 nm (e=6.22 mM ) using cyclohexenone
À1
2
008, 350, 411–418.
1
1
00 mM) as substrate. A specific C=C-bond reduction of
.2 UmL was determined by subtracting substrate-inde-
[
9] M. A. Swiderska, J. D. Stewart, J. Mol. Catal. B:
À1
Enzym. 2006, 42, 52–54.
10] A. Mꢀller, B. Hauer, B. Rosche, J. Mol. Catal. B:
Enzym. 2006, 38, 126–130.
11] M. Wada, A. Yoshizumi, Y. Noda, M. Kataoka, S. Shi-
mizu, H. Takagi, S. Nakamori, Appl. Environm. Micro-
biol. 2003, 69, 933–937.
12] F. J. Chaparro-Riggers, T. A. Rogers, E. Vazquez-Fig-
ueroa, K. M. Polizzi, A. S. Bommarius, Adv. Synth.
Catal. 2007, 349, 1521–1531.
pendent aerobic NADH oxidation from the total activity.
Based on these numbers and literature values [M ACHNTUGRNENUG( YjqM)=
[
[
À1 [37]
37.4 kDa, Aspec =7.03 Umg ] , the concentration of YqjM
in the stock solution was estimated to be 4.57 mM).
Light-Driven Reduction Reactions
[
Unless stated otherwise photoenzymatic reduction reactions
were performed on a 10-mL scale using potassium phos-
phate (100 mM, pH 7) at ambient temperature supplement-
ed with 100 mMfinal FMN, 25 mM EDTA, and 10 mM keto-
[13] C. Stueckler, M. Hall, H. Ehammer, E. Pointner, W.
Kroutil, P. Macheroux, K. Faber, Org. Lett. 2007, 9,
5409–5411.
ACHTUNGTRENNUNGi sophorone. Prior to enzyme addition, the reaction mixture
was degassed by leading N gas through the reaction mixture
[14] A. Kurata, T. Kurihara, H. Kamachi, N. Esaki, Tetrahe-
dron: Asymmetry 2004, 15, 2837–2839.
2
for 15 min. Throughout the reaction, a gentle N stream was
2
passed over the reaction mixture. Afterwards, 0.914 mMfinal
YqjM was added and the reaction was started with insetting
illumination (250 W bulb, Philips 7748 XHP). At intervals,
samples were withdrawn, extracted with one aliquot of ethyl
acetate (containing 5 mM dodecane as internal standard),
[15] R. E. Williams, D. A. Rathbone, N. S. Scrutton, N. C.
Bruce, Appl. Environm. Microbiol. 2004, 70, 3566–
3
574.
[
16] Y. Meah, B. J. Brown, S. Chakraborty, V. Massey, Proc.
Natl. Acad. Sci. USA 2001, 98, 8560–8565.
dried over MgSO and analyzed by GC. Alterations to this
4
[17] Y. Meah, V. Massey, Proc. Natl. Acad. Sci. USA 2000,
7, 10733–10738.
protocol are indicated in the text. Semi-quantitative deter-
mination of hydrogen peroxide was performed using Quan-
tofix (Sigma–Aldrich).
9
[
[
[
[
[
18] B. Kosjek, F. J. Fleitz, P. G. Dormer, J. T. Kuethe, P. N.
Devine, Tetrahedron: Asymmetry 2008, 19, 1403–1406.
19] R. E. Williams, N. C. Bruce, Microbiology 2002, 148,
Synthesis of Deazaflavin
1
607–1614.
20] P. A. Karplus, K. M. Fox, V. Massey, FASEB J. 1995, 9,
1518–1526.
21] H. Zhao, W. A. van der Donk, Curr. Opin. Biotechnol.
5
-Deazariboflavin (deazaflavin) was synthesized according
[45]
to literature methods.
2
003, 14, 583–589.
22] F. Hollmann, K. Hofstetter, A. Schmid, Trends Biotech-
nol. 2006, 24, 163–171.
Acknowledgements
[23] F. Hollmann, A. Schmid, Biocatal. Biotransform. 2004,
2, 63–88.
2
We thank Adrie Straathof (TUDelft) for kind provision with
standard substances. M. M. G. thanks the Spanish Ministry
for Science and Education for a postdoctoral stipend.
[24] W. Liu, P. Wang, Biotechnol. Adv. 2007, 25, 369–384.
[25] W. A. van der Donk, H. Zhao, Curr. Opin. Biotechnol.
2003, 14, 421–426.
[
26] H. Chenault, G. Whitesides, Appl. Biochem. Biotech-
nol. 1987, 14, 147–197.
[
27] A. Mꢀller, R. Stꢀrmer, B. Hauer, B. Rosche, Angew.
Chem. 2007, 119, 3380–3382; Angew. Chem. Int. Ed.
References
2
007, 46, 3316–3318.
[
[
[
[
1] R. A. Sheldon, I. W. C. E. Arends, U. Hanefeld, Green
Chemistry and Catalysis, Wiley-VCH, Weinheim, 2007.
2] W. S. Knowles, Angew. Chem. 2002, 114, 2096–2107;
Angew. Chem. Int. Ed. 2002, 41, 1998–2007.
3] R. Noyori, Angew. Chem. 2002, 114, 2108–2123;
Angew. Chem. Int. Ed. 2002, 41, 2008–2022.
4] J. W. Yang, M. T. Hechavarria Fonseca, N. Vignola, B.
List, Angew. Chem. 2005, 117, 110–112; Angew. Chem.
Int. Ed. 2005, 44, 108–110.
[
[
[
[
[
[
28] F. Hollmann, B. Witholt, A. Schmid, J. Mol. Catal. B:
Enzym. 2002, 19–20, 167–176.
29] F. Hollmann, P. C. Lin, B. Witholt, A. Schmid, J. Am.
Chem. Soc. 2003, 125, 8209–8217.
30] S. Unversucht, F. Hollmann, A. Schmid, K.-H. van Pꢄe,
Adv. Synth. Catal. 2005, 347, 1163–1167.
31] G. de Gonzalo, G. Ottolina, G. Carrea, M. W. Fraaije,
Chem. Commun. 2005, 3724–3726.
32] F. Hollmann, K. Hofstetter, T. Habicher, B. Hauer, A.
Schmid, J. Am. Chem. Soc. 2005, 127, 6540–6541.
33] F. Hollmann, A. Taglieber, F. Schulz, M. T. Reetz,
Angew. Chem. 2007, 119, 2961–2964; Angew. Chem.
Int. Ed. 2007, 46, 2903–2906.
[
[
[
5] R. Stꢀrmer, B. Hauer, M. Hall, K. Faber, Curr. Opin.
Chem. Biol. 2007, 11, 203–213.
6] O. Warburg, W. Christian, Biochem. Z. 1933, 263, 228–
2
29.
7] M. Hall, C. Stꢀckler, B. Hauer, R. Stꢀrmer, T. Frie-
drich, M. Breuer, W. Kroutil, K. Faber, Eur. J. Org.
Chem. 2008, 2008, 1511–1516.
[34] A. Taglieber, F. Schulz, F. Hollmann, M. Rusek, M. T.
Reetz, ChemBioChem 2008, 9, 565–572.
Adv. Synth. Catal. 2009, 351, 3279 – 3286
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
3285

Maria Mifsud Grau et al.
FULL PAPERS
[
[
[
[
[
[
35] W. R. Frisell, C. W. Chung, C. G. Mackenzie, J. Biol.
Chem. 1959, 234, 1297–1302.
36] V. Massey, M. Stankovich, P. Hemmerich, Biochemistry
[42] V. Massey, J. Biol. Chem. 1994, 269, 22459–22462.
[43] V. Massey, P. Hemmerich, J. Biol. Chem. 1977, 252,
5612–5614.
[44] K. Miura, Y. Tomioka, H. Suzuki, M. Yonezawa, T.
Hishinuma, M. Mizugaki, Biol. Pharmaceut. Bull. 1997,
20, 110–112.
1
978, 17, 1–8.
37] T. B. Fitzpatrick, N. Amrhein, P. Macheroux, J. Biol.
Chem. 2003, 278, 19891–19897.
38] V. Massey, P. Hemmerich, J. Biol. Chem. 1977, 252,
[
[
[
45] E. Carlson, L. Kiessling, J. Org. Chem. 2004; 69, 2614–
617.
46] P. Macheroux, S. Bornemann, S. Ghisla, R. N. F. Thor-
neley, J. Biol. Chem. 1996, 271, 25850–25858.
47] W. Holzer, J. Shirdel, P. Zirak, A. Penzkofer, P. Hege-
5
612–5614.
2
39] F. Zilly, A. Taglieber, F. Schulz, F. Hollmann, M. T.
Reetz, Chem. Commun. 2009, 7152–7154.
40] T. B. Fitzpatrick, S. Auweter, K. Kitzing, T. Clausen, N.
Amrhein, P. Macheroux, Prot. Express. Purif. 2004, 36,
AHCTUNGTRENNUNG
2
80–291.
2
[41] K. Kitzing, T. B. Fitzpatrick, C. Wilken, J. Sawa, G. P.
[
48] D. I. Perez, M. Mifsud Grau, I. W. E. C. Arends, F.
Bourenkov, P. Macheroux, T. Clausen, J. Biol. Chem.
Hollmann, Chem. Commun. 2009, 6848–6850.
2
005, 280, 27904–27913.
3286
asc.wiley-vch.de
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Adv. Synth. Catal. 2009, 351, 3279 – 3286