Sequential Enzymatic Conversion of α-Angelica Lactone to γ-Valerolactone through Hydride-Independent C=C Bond Isomerization

Article abstract of DOI:10.1002/cssc.201601363

A case of hydride-independent reaction catalyzed by flavin-dependent ene-reductases from the Old Yellow Enzyme (OYE) family was identified. α-Angelica lactone was isomerized to the conjugated β-isomer in a nicotinamide-free and hydride-independent process. The catalytic cycle of C=C bond isomerization appears to be flavin-independent and to rely solely on a deprotonation–reprotonation sequence through acid–base catalysis. Key residues in the enzyme active site were mutated and provided insight on important mechanistic features. The isomerization of α-angelica lactone by OYE2 in aqueous buffer furnished 6.3 mm β-isomer in 15 min at 30 °C. In presence of nicotinamide adenine dinucleotide (NADH), the latter could be further reduced to γ-valerolactone. This enzymatic tool was successfully applied on semi-preparative scale and constitutes a sustainable process for the valorization of platform chemicals from renewable resources.

Full text of DOI:10.1002/cssc.201601363

DOI: 10.1002/cssc.201601363
Communications
Sequential Enzymatic Conversion of a-Angelica Lactone to
g-Valerolactone through Hydride-Independent C=C Bond
Isomerization
[a]
[a]
[a, b]
[a]
[a]
Nikolaus G. Turrini, Elisabeth Eger, Tamara C. Reiter,
Kurt Faber, and Mꢀlanie Hall*
A case of hydride-independent reaction catalyzed by flavin-de-
pendent ene-reductases from the Old Yellow Enzyme (OYE)
family was identified. a-Angelica lactone was isomerized to the
conjugated b-isomer in a nicotinamide-free and hydride-inde-
pendent process. The catalytic cycle of C=C bond isomeriza-
tion appears to be flavin-independent and to rely solely on
a deprotonation–reprotonation sequence through acid–base
catalysis. Key residues in the enzyme active site were mutated
and provided insight on important mechanistic features. The
isomerization of a-angelica lactone by OYE2 in aqueous buffer
furnished 6.3 mm b-isomer in 15 min at 308C. In presence of
nicotinamide adenine dinucleotide (NADH), the latter could be
further reduced to g-valerolactone. This enzymatic tool was
successfully applied on semi-preparative scale and constitutes
a sustainable process for the valorization of platform chemicals
from renewable resources.
neutral C=C bond isomerization of a-methylene-g-butyrolac-
tone to thermodynamically more stable 3-methylfuran-2(5H)-
one was observed with OYE2 (Scheme 1B); this process re-
quires only catalytic amounts of NADH to activate the flavin by
means of reduction, thereby triggering intermolecular hydride
[5]
transfer from endo-Cb onto exo-Cb through FMN. In both
cases, the flavin acts as hydride shuttle between two substrate
molecules. In contrast to these redox-neutral isomerization re-
actions, C=C reduction of racemic g-substituted a,b-unsaturat-
ed lactones proceeds through kinetic resolution, and dynamic
kinetic resolution of a,g-disubstituted analogues was achieved
with the OYE homologue nicotinamide-dependent cyclo-
[6]
hexenone reductase (NCR) from Zymomonas mobilis. The re-
duction reactions resemble a 1,4-Michael-type addition and re-
quire
a stoichiometric amount of nicotinamide cofactor
(Scheme 1C). Until now, the catalytic cycle of ene-reductases
was reported to solely rely on hydride transfer using the re-
duced flavin, and no hydride-independent catalysis was report-
ed.
Ene-reductases from the Old Yellow Enzyme (OYE) family are
flavin-dependent enzymes that were well investigated for their
ability to catalyze the asymmetric reduction of activated C=C
Levulinic acid derived from cellulosic feedstocks represents
[7]
a major biobased platform chemical. Upon dehydration, a-
angelica lactone is formed, which can be isomerized to b-an-
gelica lactone or reduced to g-valerolactone, two versatile
building blocks with broad applications (e.g., precursors of bio-
based polymers and natural products, solvents, fuel addi-
[
1]
bonds at the expense of nicotinamide cofactor. In the first
reductive) half-reaction, flavin mononucleotide (FMN) is re-
duced by externally added nicotinamide adenine dinucleotide
phosphate) [NAD(P)H]. The reduced FMN, in turn, can transfer
a hydride onto the b-carbon of a C=C bond, which is activated
(
(
[8]
tives). Because OYEs can only reduce activated C=C bonds, a-
angelica lactone does not appear to be a suitable substrate,
but the C=C isomerization activity of some OYE homologues
on a,b-unsaturated cycloalkenones and lactones encouraged
us to investigate the reactivity of a-angelica lactone (1a) with
a panel of ene-reductases (OYE2, OYE3, NCR, YqjM, OPR1,
OPR3, EBP1, XenA).
[2]
by an electron-withdrawing group (oxidative half reaction). In
special cases, mostly on cyclic compounds such as a,b-unsatu-
rated cycloalkenones or lactones, distinct unusual catalytic be-
haviors of OYE homologues were observed: 1) Nicotinamide-
free dismutation catalyzed by OYE1 was identified using cyclo-
hexen-2-one and 3-oxodecalin-4-ene, in which one substrate
molecule is first dehydrogenated by oxidized flavin, followed
by reduction of a second molecule by the reduced flavin
The initial setup consisted of substrate solution (10 mm) in
Tris-HCl [tris(hydroxymethyl)aminomethane] buffer (pH 7.5,
[
3]
-1
(
Scheme 1A); this intermolecular reaction was extended to
50 mm), excess of NADH (1.5 equiv) and enzyme (100 mg mL ,
[
4]
a broad variety of substrates and OYE homologues. 2) Redox-
approximately 2.2 mm). Upon extraction of the aqueous solu-
tion with ethyl acetate after 24 h and subsequent GC–MS anal-
ysis, g-valerolactone (1c) was identified as single product from
the reaction with OYE2 from Saccharomyces cerevisiae whereas
no starting material could be recovered, indicating that OYE2
was able to convert a-angelica lactone. Other ene-reductases
appeared inactive. This result was surprising because reduction
of C=C bonds on analogous compounds was only reported on
enone derivatives so far (i.e., 1,4-addition on activated C=C
bonds). By analogy with the isomerization reaction of a-meth-
[a] Dr. N. G. Turrini, E. Eger, T. C. Reiter, Prof. K. Faber, Dr. M. Hall
Department of Chemistry
University of Graz
Heinrichstrasse 28, 8010 Graz (Austria)
E-mail: melanie.hall@uni-graz.at
[b] T. C. Reiter
ACIB GmbH, Department of Chemistry
University of Graz
Heinrichstrasse 28, 8010 Graz (Austria)
Supporting Information and the ORCID identification number(s) for the
author(s) of this article can be found under http://dx.doi.org/10.1002/
cssc.201601363.
[5]
ylene-g-butyrolactone catalyzed by OYE2, which requires cat-
alytic amounts of nicotinamide, the experiment was repeated
ChemSusChem 2016, 9, 1 – 5
1
ꢁ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
These are not the final page numbers! ÞÞ

Communications
Scheme 1. Versatile catalytic behavior of ene-reductases (ER) on a,b-unsaturated cycloalkenones and lactones. A) NADH-independent intermolecular hydrogen
transfer; B) NADH-catalyzed intermolecular hydrogen transfer; C) stoichiometric NADH-dependent C=C reduction. E-value: enantiomeric ratio.
with lower cofactor concentration. Upon reducing NADH con-
centration to substoichiometric amounts, isomerized product
b-angelica lactone (1b) could be detected, whereas consistent-
ly less saturated product 1c was produced (Table 1). At 10 mm
NADH (0.1 mol%), no reduction product was detected whereas
1
b was formed in 4 mm. The overall low recovery was attribut-
ed to reversible hydrolytic ring opening of a-angelica lac-
[
9]
tone, which appeared much more unstable than the b-
isomer 1b in aqueous solution (data not shown). The product
distribution depended highly on NADH concentration, and
clearly showed that 1c arises from reduction of 1b, which
itself derives from isomerization of 1a (Scheme 2). In addition,
the reaction of 1b (synthesized according to published proce-
Scheme 2. Sequential biotransformation of unactivated substrates 1a and
a by OYE2 from Saccharomyces cerevisiae to 1c and 2c, respectively.
2
were racemic. Reduction of rac-5-oxo-2,5-dihydrofuran-2-yl
acetate, chemically analogous to 1b, proceeded through kinet-
ic resolution with a handful of ene-reductases, but OYE2 af-
[
10]
dure,
Supporting Information) in presence of OYE2 and
[6]
excess of NADH led to formation of 1c as sole product. In
both cases (reaction starting from 1a or 1b), all compounds
forded the racemic product.
Finally, the reaction was performed in absence of NADH.
Almost full consumption of 1a (85%) could be observed
within 15 min, leading to the formation of 6.3 mm 1b. Among
all OYE homologues tested in absence of NADH, only OYE3
showed significant activity in the nicotinamide-free C=C bond
isomerization (35% conversion in 24 h, Table 2, entry 3). Clearly,
this enzymatic isomerization is distinct from the isomerization
of a-methylene-g-butyrolactone by OYE2, which requires cata-
Table 1. Product distribution after reaction of 1a (10 mm) with OYE2 at
[
a]
varying NADH concentrations.
Entry
Concentration [mm]
NADH
1a
1b
1c
[5]
1
2
3
4
5
6
7
8
10
5
2
0.5
0.1
n.d.
0.19
0.60
2.37
3.38
4.25
3.70
4.03
6.30
2.03
2.19
0.99
0.23
0.03
0.04
n.d.
lytic amounts of NADH cofactor to initiate hydride transfer.
0.13
0.57
0.82
1.05
0.83
0.93
1.50
5
By analogy with the isomerization mechanism of D -3-ketoste-
[11]
roid isomerase on steroids, the isomerization of the unacti-
vated C=C bond in a-angelica lactone by OYE2 most likely
relies on acid–base catalysis and is hydride independent. FMN
is naturally occurring in oxidized form in OYE homologues and
thus cannot initiate hydride addition.
0.05
0.01
[
b]
0
n.d.
[
a] Tris-HCl buffer (50 mm, pH 7.5), 308C, 120 rpm, 3 h reaction time;
To identify possible catalytic residues functioning as acid
and/or base, a mutational study was designed based on the
n.d.=not detected. [b] Reaction time: 15 min.
&
ChemSusChem 2016, 9, 1 – 5
www.chemsuschem.org
2
ꢁ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Communications
ing that the catalytic activity of the variant was severely affect-
ed [total turnover number (TTN) ꢀ4]. A control reaction in
presence of elevated bovine serum albumin (BSA) concentra-
tion or denatured OYE2 resulted in no product (Table 2, en-
tries 7 and 8), confirming that the low turnover did not result
from nonspecific amino-acid catalysis.
Table 2. Nicotinamide-free isomerization of 1a and 2a to 1b and 2b cat-
alyzed by OYEs.
[
a]
Entry Substrate Protein
Concentration [mm] Conv. [%]
1
2
3
4
5
6
7
8
9
1
1a
1a
1a
1a
1a
1a
1a
1a
2a
2a
2a
OYE2-wt
OYE2-wt
OYE3-wt
OYE2-H192A
OYE2-Y197F
OYE2-Y197F
BSA
OYE2 denatured
OYE2-wt
2.8
2.8
2.8
2.8
2.8
741
600
2.8
45
13
35
40
<1
30
<1
<1
49
25
15
[
[
b]
c]
Taken together, the data suggest that His192 is not the cata-
lytic base, although the role of Tyr197 remains unclear. Based
on the proposed mechanism (Scheme 3), an initial deprotona-
tion step would generate an aromatic anion as intermediate,
which resembles a phenolate. Phenols are notorious inhibitors
of OYEs owing to strong binding between phenolate and oxi-
[
d]
[
e]
[
[
f]
2.8
2.8
2.8
0
OYE2-H192A
OYE2-Y197F
g]
[15]
11
dized FMN. Similarly, the suggested anionic aromatic inter-
mediate would strongly interact with oxidized FMN and, in ab-
sence of a proton donor in OYE2-Y197F, cannot (or only
slowly) leave the active site, thereby preventing product for-
mation. In contrast, the reducing activity of variants lacking
the catalytic Tyr is not strongly affected (Supporting Informa-
tion) because the nonaromatic intermediate is easily released
[
[
4
9
[
a] Tris-HCl buffer (50 mm, pH 7.5), 2 h reaction time; wt=wild type.
b] pH 6. [c] 24 h reaction time. [d] Corresponds to TTN of approximately
(3 mm 1b). [e] Denaturation procedure: OYE2 sample was heated to
58C for 10 min, allowed to cool down, and used in the reaction as such.
f] No trace of 2a. [g] Control (no enzyme) also showed 15% of 2b.
[2a]
and further protonated by the solvent (water). The signifi-
cantly lower conversion level at pH 6 (Table 2, entry 2) strongly
supports the involvement of a base in the deprotonation step.
Although the reaction at higher pH values (8–9) led to forma-
tion of 1b, conversion levels could not be determined owing
to the instability of 1a in basic medium (data not shown). Elu-
cidation of the (yet unsolved) structure of OYE2 will enable
identification of additional point mutations for further mecha-
nistic studies. Interestingly, cases of flavin acting as acid/base
crystal structure of highly homologous OYE1 from Saccharomy-
[
12]
ces pastorianus (92% identity and 95% similarity). Two key
positions were targeted: 1) His192 (pKaꢀ6) is one of the two
residues involved in the binding of the substrate carbonyl
through hydrogen bonds and could act as a base using its un-
protonated nitrogen atom in the deprotonation of an acidic a-
[
2]
[13]
H of 1a; 2) Tyr197 (pKaꢀ9.1 ) is the proton donor for the
a-C in the reduction reaction and may function as proton
donor in the isomerization (Scheme 3).
[16]
were reported, but all require a reduced flavin.
Tyr197 was mutated to phenylalanine (Phe) to conserve simi-
lar steric properties but remove the hydrogen-donor capacity,
whereas His192 was mutated to alanine (Ala) to eliminate both
H-bonding capacity and basic character. Both single variants
were first evaluated in the reduction of N-phenyl-2-methyl-
maleimide. Successful reduction in the presence of NADH con-
firmed that both variants were still able to bind the substrate
The reaction was extended to 3-cyclopenten-1-one (2a),
which was isomerized to 2-cyclopenten-1-one (2b) in a nico-
tinamide-independent manner (Table 2, entry 9). Similar to 1a,
the conversion level with OYE2-H192A was affected by the mu-
tation and OYE2-Y197F was inactive (Table 2, entries 10 and
11). Finally, the reaction was scaled up to demonstrate the ap-
plicability of the cofactor-free enzymatic isomerization of a-an-
gelica lactone to its b-isomer. 1a (55 mg, 25 mm) was convert-
ed by OYE2 (0.03 mol%) in Tris-HCl buffer at 308C for 45 min
and allowed isolation of 18 mg 1b (32% overall yield, 98%
purity, Experimental Section).
(Supporting Information). OYE2-H192A could catalyze the iso-
merization of 1a to 1b in absence of NADH (Table 2, entry 4),
albeit with slightly decreased activity, going in line with lower
reduction activity and expected reduced binding affinity result-
ing from the change from two binding residues (H192/N195)
The catalytic activity of ene-reductases from the Old Yellow
Enzyme (OYE) family of flavoproteins was so far restricted to
flavin- and hydride-dependent processes in which reduction of
flavin mononucleotide (FMN) by an external hydride-cofactor is
the initial step required for activity. Asymmetric stoichiometric
[
14]
to only one. In contrast, OYE2-Y197F at first showed no ac-
tivity under standard reaction conditions (2.2 mm enzyme,
Table 2, entry 5). When the concentration was increased to
741 mm, 3 mm of 1b were obtained (Table 2, entry 6), indicat-
Scheme 3. Initial proposed mechanism for the isomerization of a-angelica lactone by OYE2. Return to histidine (His) and tyrosine (Tyr) initial states is solvent
mediated (buffer pH 7.5). Asn=asparagine.
ChemSusChem 2016, 9, 1 – 5
www.chemsuschem.org
3
ꢁ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
These are not the final page numbers! ÞÞ

Communications
nicotinamide-dependent reduction is broadly applicable to ac-
tivated a,b-unsaturated alkenes and includes (dynamic) kinetic
resolution of g-substituted substrates. Catalytic nicotinamide-
dependent isomerization was observed on a-methylene-g-bu-
tyrolactone, and nicotinamide-independent disproportionation
of enones relies on reduced FMN as hydride shuttle between
two substrates molecules. In contrast, the hydride- and nico-
tinamide-independent isomerization of unactivated C=C bonds
reported here represents a new catalytic activity of ene-reduc-
tases. The isomerization of a-angelica lactone to its b-isomer
by OYE2 enzyme was successfully performed on a preparative
scale, and the possibility of further reduction to g-valerolac-
tone in a cascade was demonstrated.
agreement n8115360, resources of which are composed of finan-
cial contribution from the European Union’s Seventh Framework
Programme (FP7/2007-2013) and EFPIA companies in kind contri-
bution. This research is part of project CHEM21 (https://www.che-
m21.eu/). E.E. was supported by the Austrian Science Fund (FWF,
project W9, DK Molecular Enzymology).
Keywords: biobased chemicals · biocatalysis · ene-reductases ·
isomerization · a-angelica lactone
[
[
1] K. Faber, M. Hall in Science of Synthesis: Biocatalysis in Organic Synthesis,
Vol. 2 (Eds.: K. Faber, W.-D. Fessner, N. J. Turner), Georg Thieme Verlag,
Stuttgart, 2015, pp. 213–260.
Experimental Section
2] a) B. J. Brown, Z. Deng, P. A. Karplus, V. Massey, J. Biol. Chem. 1998, 273,
Standard procedure for bioreduction reactions
32753–32762; b) P. A. Karplus, K. M. Fox, V. Massey, FASEB J. 1995, 9,
1518–1526; c) R. M. Kohli, V. Massey, J. Biol. Chem. 1998, 273, 32763–
32770.
To a solution of Tris-HCl buffer (800 mL, 50 mm, pH 7.5) containing
the substrate (10 mm) and the cofactor (concentration as indicated
in Table 1) was added an aliquot of the respective enzymes (final
[
3] A. D. N. Vaz, S. Chakraborty, V. Massey, Biochemistry 1995, 34, 4246–
256.
4] a) C. Stueckler, T. C. Reiter, N. Baudendistel, K. Faber, Tetrahedron 2010,
6, 663–667; b) C. K. Winkler, D. Clay, M. Entner, M. Plank, K. Faber,
Chem. Eur. J. 2014, 20, 1403–1409.
4
À1
protein concentration 100 mgmL , approximately 2.2 mm, unless
[
denoted otherwise). All samples were run in duplicates, and con-
trol experiments were included in every screening (no enzyme).
Substrates were added from a 0.5m stock solution in ethanol to
overcome solubility problems (final concentration of ethanol 2%).
The mixture was shaken at 308C and 120 rpm (reaction time as in-
dicated in the text and Tables 1 and 2). Afterwards, the reaction
was extracted with ethyl acetate (2ꢂ500 mL), and the combined or-
ganic layers were dried over Na SO and transferred into GC vials.
6
[5] K. Durchschein, S. Wallner, P. Macheroux, K. Zangger, W. M. F. Fabian, K.
Faber, ChemBioChem 2012, 13, 2346–2351.
[6] N. G. Turrini, M. Hall, K. Faber, Adv. Synth. Catal. 2015, 357, 1861–1871.
[
[
[
7] Top Value Added Chemicals from Biomass. Results of Screening for Poten-
tial Candidates from Sugars and Synthesis Gas, Vol. 1 (Eds.: T. Werpy, G.
Petersen), 2004, US Department of Energy. http://www.nrel.gov/docs/
fy04osti/35523.pdf.
2
4
Products were identified by comparison with authentic reference
material. Limonene (10 mm) was used as internal standard.
8] a) I. T. Horvꢃth, H. Mehdi, V. Fꢃbos, L. Boda, L. T. Mika, Green Chem.
2
008, 10, 238–242; b) J. Boukouvalas, N. Bruneau-Latour, Synlett 2013,
2
4, 2691–2694; c) J. H. Siitonen, P. M. Pihko, Synlett 2014, 25, 1888–
1890.
Procedure for scale-up
9] G. W. Huber, S. Iborra, A. Corma, Chem. Rev. 2006, 106, 4044–4098.
[
[
10] E. Guntrum, W. Kuhn, W. Spçnlein, V. Jꢄger, Synthesis 1986, 11, 921–
25.
11] Z. R. Wu, S. Ebrahimian, M. E. Zawrotny, L. D. Thornburg, G. C. Perez-Al-
To a solution of 1a (55.7 mg, 0.57 mmol, 25 mm) in Tris-HCl buffer
9
(
50 mm, pH 7.5) was added an aliquot of OYE2 (2.2 mm). The reac-
tion mixture was incubated at 308C and 120 rpm. Analysis of the
reaction (GC–MS) after 15 min reaction time showed fast isomeriza-
tion (1b/1a=90:10). Two aliquots of OYE2 (4.4 mm) were then
added (total 0.03 mol%) and the reaction was further incubated
for additional 30 min. Extraction was performed with ethyl acetate
varado, P. Brothers, R. M. Pollack, M. F. Summers, Science 1997, 276,
415–418.
[12] Y. S. Niino, S. Chakraborty, B. J. Brown, V. Massey, J. Biol. Chem. 1995,
270, 1983–1991.
[
[
13] Y. Meah, V. Massey, Proc. Natl. Acad. Sci. USA 2000, 97, 10733–10738.
14] H. Khan, T. Barna, N. C. Bruce, A. W. Munro, D. Leys, N. S. Scrutton, FEBS
J. 2005, 272, 4660–4671.
15] a) A. S. Abramovitz, V. Massey, J. Biol. Chem. 1976, 251, 5327–5336;
b) K. M. Fox, P. A. Karplus, Structure 1994, 2, 1089–1105.
16] a) C. J. Thibodeaux, S. O. Mansoorabadi, W. Kittleman, W.-C. Chang, H.-
W. Liu, Biochemistry 2008, 47, 2547–2558; b) C. T. Walsh, T. A. Wence-
wicz, Nat. Prod. Rep. 2013, 30, 175–200.
(
3ꢂ15 mL), and the combined organic layers were dried over
Na SO . Evaporation of the solvent yielded 18.1 mg 1b in 32%
2
4
[
[
overall yield (98% purity, Supporting Information).
Acknowledgements
The research leading to these results has received funding from
the Innovative Medicines Initiative Joint Undertaking under grant
Manuscript received: September 28, 2016
Final Article published: && &&, 0000
&
ChemSusChem 2016, 9, 1 – 5
www.chemsuschem.org
4
ꢁ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

COMMUNICATIONS
N. G. Turrini, E. Eger, T. C. Reiter, K. Faber,
M. Hall*
&& – &&
Sequential Enzymatic Conversion of a-
Angelica Lactone to g-Valerolactone
through Hydride-Independent C=C
Bond Isomerization
No hydride required: The sequential
enzymatic conversion of a-angelica lac-
tone to g-valerolactone by ene-reduc-
tase OYE2 proceeds through hydride-in-
dependent C=C bond isomerization to
b-Angelica lactone, a so far unreported
catalytic behavior of OYE homologues.
Upon addition of nicotinamide, the re-
duction reaction proceeds smoothly in
aqueous buffer from the b-isomer inter-
mediate to the final product.
ChemSusChem 2016, 9, 1 – 5
www.chemsuschem.org
5
ꢁ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
These are not the final page numbers! ÞÞ