Enzymatic conversion of flavonoids using bacterial chalcone isomerase and enoate reductase

Article abstract of DOI:10.1002/anie.201306952

Flavonoids are a large group of plant secondary metabolites with a variety of biological properties and are therefore of interest to many scientists, as they can lead to industrially interesting intermediates. The anaerobic gut bacterium Eubacterium ramulus can catabolize flavonoids, but until now, the pathway has not been experimentally confirmed. In the present work, a chalcone isomerase (CHI) and an enoate reductase (ERED) could be identified through whole genome sequencing and gene motif search. These two enzymes were successfully cloned and expressed in Escherichia coli in their active form, even under aerobic conditions. The catabolic pathway of E. ramulus was confirmed by biotransformations of flavanones into dihydrochalcones. The engineered E. coli strain that expresses both enzymes was used for the conversion of several flavanones, underlining the applicability of this biocatalytic cascade reaction.

Full text of DOI:10.1002/anie.201306952

Angewandte
Chemie
DOI: 10.1002/anie.201306952
Biocatalysis
Enzymatic Conversion of Flavonoids using Bacterial Chalcone
Isomerase and Enoate Reductase**
Mechthild Gall, Maren Thomsen, Christin Peters, Ioannis V. Pavlidis, Patrick Jonczyk,
Philipp P. Grꢀnert, Sascha Beutel, Thomas Scheper, Egon Gross, Michael Backes,
Torsten Geißler, Jakob P. Ley,* Jens-Michael Hilmer, Gerhard Krammer, Gottfried J. Palm,
Winfried Hinrichs, and Uwe T. Bornscheuer*
Abstract: Flavonoids are a large group of plant secondary
metabolites with a variety of biological properties and are
therefore of interest to many scientists, as they can lead to
industrially interesting intermediates. The anaerobic gut bacte-
rium Eubacterium ramulus can catabolize flavonoids, but until
now, the pathway has not been experimentally confirmed. In
the present work, a chalcone isomerase (CHI) and an enoate
reductase (ERED) could be identified through whole genome
sequencing and gene motif search. These two enzymes were
successfully cloned and expressed in Escherichia coli in their
active form, even under aerobic conditions. The catabolic
pathway of E. ramulus was confirmed by biotransformations
of flavanones into dihydrochalcones. The engineered E. coli
strain that expresses both enzymes was used for the conversion
of several flavanones, underlining the applicability of this
biocatalytic cascade reaction.
as antioxidants[1] and flavor enhancers.[2] Their biosynthesis in
plants is well documented in the literature, and the basic
scaffold is formed from malonyl and coumaroyl precursors in
the presence of 4-coumaryl-CoA ligase and a chalcone
synthase. In the next step, chalcone isomerase (CHI) cata-
lyzes the stereospecific formation of the tricyclic flavanones
such as naringenin (Scheme 1). These are further functional-
Scheme 1. Postulated pathway for the degradation of flavonoids.
CHI=chalcone isomerase, ERED=enoate reductase. The equilibrium
lies strongly on the side of flavanone formation.
F
lavonoids are polyphenolic compounds that occur in
plants; they are involved in plant coloration and act as
biochemical sensing molecules, but also have important roles
ized by other enzymes, such as hydroxylases or glycosidases,
to create the diverse family of flavonoids. For plant-derived
CHIs, their evolutionary origin,[3] reaction mechanism,[4] and
structures[5] have been extensively studied. More recently, it
was found that CHI can also be found in some gammapro-
teobacteria and ascomycetes, but their physiological role
remains to be determined.[6] One exception is the anaerobic
gut microorganism Eubacterium ramulus. It has been
reported that this strain is able to degrade a range of
flavonoids, including naringenin-7-neohesperidoside, a glyco-
sylated derivative of naringenin.[7] The authors suggested
a possible degradation pathway, but did not present exper-
imental verification of key intermediate steps. Furthermore,
phloretin (R1 = OH, R2 = H; Scheme 1) is hydrolyzed in
E. ramulus to 3-(4-hydroxyphenyl)propanoic acid and phlo-
roglucinol, which is an undesirable pathway, as phloretin, as
a degradation product of naringenin, is a very important
flavor compound.[8] Although phloretin can in principle be
chemically produced by reduction of the corresponding
chalcone or by a Friedel–Crafts-type acylation of phenol
with dihydrocinnamic acid,[9] these chemically synthesized
products cannot be used as natural flavoring substances
according to European flavor legislation. Therefore, an
enzymatic or fermentative process for phloretin production
would hold additional value for the flavor and fragrance
[*] Dipl.-Biol. M. Gall, Dipl.-Biochem. C. Peters, Dr. I. V. Pavlidis,
Prof. Dr. U. T. Bornscheuer
Institute of Biochemistry
Department of Biotechnology & Enzyme Catalysis
Greifswald University
Felix-Hausdorff-Strasse 4, 17487 Greifswald (Germany)
E-mail: uwe.bornscheuer@uni-greifswald.de
Homepage: http://biotech.uni-greifswald.de
Dipl.-Biochem. M. Thomsen, Dr. G. J. Palm, Prof. Dr. W. Hinrichs
Institute of Biochemistry, Department of Structural Biology
Greifswald University
Felix-Hausdorff-Strasse 4, 17487 Greifswald (Germany)
M. Sc. P. Jonczyk, M. Sc. P. P. Grꢀnert, Dr. S. Beutel,
Prof. Dr. T. Scheper
Institute of Technical Chemistry
Gottfried Wilhelm Leibniz University of Hannover
Callinstrasse 5, 30167 Hannover (Germany)
E. Gross, Dr. M. Backes, Dr. T. Geißler, Dr. J. P. Ley, Dr. J.-M. Hilmer,
Dr. G. Krammer
Symrise, P.O. Box 1253, 37603 Holzminden (Germany)
E-mail: jakob.ley@symrise.com
[**] We thank the “Bundesministerium fꢀr Bildung und Forschung” for
financial support within the “Biokatalyse 2021” cluster (FKZ:
0315365 and 031A109). M.T. thanks the “Landesgraduiertenkolleg
of Mecklenburg-Vorpommern” for financial support.
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201306952.
Angew. Chem. Int. Ed. 2013, 52, 1 – 5
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
1
These are not the final page numbers!

.
Angewandte
Communications
industry. Alternatively, glycosides of phloretin, such as
phlorizin, can be obtained from Malus ssp.,[10] and are
subsequently hydrolyzed to yield the aglycon phloretin.[11]
However, this established process is time- and cost-intensive
and also depends on the seasonal availability of the starting
material. To avoid the seasonal impact and to simultaneously
reduce production costs and waste streams, a biotechnological
route for this important flavor compound is of high interest.
The cultivation of the strictly anaerobic E. ramulus
DSM 16296 for the setup of an industrially useful bioprocess
is an obstacle, as the growth is very slow, and whole-cell
biotransformations lead to further degradation of phlore-
tin.[12] These problems can be overcome by the heterologous
expression of enzymes in microorganisms. This technique is
already in use for the production of flavonoids from various
substrates, such as l-phenylalanine, tyrosine, or cinnamic
acid, but only for genes derived from plants.[13] In our present
study, we came across several challenges, including the
identification of the genes from the E. ramulus genome that
encode the enzymes that are involved in the biocatalytic
transformation of naringenin into phloretin, and the func-
tional expression of the proteins in a recombinant host.
Herles and co-workers[14] determined a fragment of 15
amino acids at the N terminus of the CHI from E. ramulus
DSM 16296 that is proposed to be involved in the conversion
of naringenin into phloretin. This peptide sequence was used
in the present study for the identification of the whole gene.
Therefore, we sequenced the entire genome of E. ramulus
DSM 16296 and used the resulting contigs to identify the
encoding gene for the CHI. Only one hit was found, which
aligned twelve consecutive residues of the published sequence
directly after the starting methionine. This open reading
frame (ORF; GenBank Accession Number: KF154734)
corresponds to a protein of 32.5 kDa, which correlates well
with the size that Herles and co-workers determined by SDS-
PAGE for one subunit of the enzyme.[14] Interestingly,
a BLAST[15] search did not identify any other CHIs from
plant or other sources with a sequence identity of ꢀ 10%
compared to this bacterial CHI; hence, this protein is unique
(Supporting Information, Figure S1).
from E. ramulus; however, lower specific activities were
found (Tables S1 and S2). The characterization of CHI
revealed that the recombinant CHI maintains more than
90% of its activity at pH 6.4–7.6 (Figure S5), whereas the
optimal operational temperature is 458C (Figure S6). The
recombinant CHI also exhibited satisfactory stability, main-
taining more than 50% of its initial activity after incubation at
418C for six hours (Figure S7).
The kinetic constants that were determined for the
recombinant CHI underline the high activity of the enzyme
(Figure S8). Although the Km value of the recombinant
bacterial CHI (36.9 mm) lies between those of the plant
CHIs, which vary from 2 mm to 112 mm for the CHI from
Glycine max[17] and Medicago sativa,[4] the turnover number is
significantly higher. The kcat values that have been reported in
the literature for plant CHIs vary from 186 sÀ1[4] to 833 sÀ1 [18]
,
whereas a value of 4483 sÀ1 was determined for the recombi-
nant CHI from E. ramulus. Hence, the recombinantly
expressed bacterial CHI shows a catalytic efficiency of 1.2 ꢀ
108 m-1 s-1, which is 75-times higher than the plant one from
Medicago sativa, whose catalytic efficiency is 1.6 ꢀ 106 m-1 s-1.[4]
When we compared the recombinantly expressed CHI with
the native CHI that was produced in E. ramulus, the kinetic
values differed. Herles and co-workers determined the Km to
be 42.7 mm with a kcat value of 2300 sÀ1, which resulted in
a catalytic efficiency of 0.5 ꢀ 108 mÀ1 sÀ1,[14] which is signifi-
cantly lower than the one of the recombinant enzyme; this
difference is mainly due to the different turnover numbers.
The higher catalytic activity of the recombinant CHI can be
attributed to the higher purity or to the formation of different
multimers.
Analysis of the X-ray crystal structure (PDB code: 3zph)
revealed that the recombinant CHI forms hexamers as
trimers of three dimer units (Figure 1), whereas Herles and
For the enoate reductase (ERED), no gene information
was available. For this reason, we aligned 34 sequences of
known EREDs and found a conserved motif that served as
a basis to also identify the ERED-encoding gene in the
E. ramulus DSM 16296 genome. Again, we found only one
ORF (GenBank Accession Number: KF154735), its amino
acid sequence shared low identity (ꢁ 29%; Figure S2) with
known EREDs.[16] However, all of these other EREDs have
been only insufficiently characterized.
Figure 1. Crystal structure (left: top view; right: turned by 908) of
a hexamer with D3 symmetry of the recombinant CHI from E. ramulus;
PDB code: 3zph.
The genes that encode the bacterial CHI and ERED were
cloned into common pET vectors, and functional expression
in E. coli Rosetta was performed. We were pleased to find
that the CHI from E. ramulus was expressed in soluble form
under aerobic conditions in the E. coli Rosetta (DE3) strain
(Figure S3). Slight modifications of the purification protocol
described by Herles and co-workers[14] provided us with pure
recombinant CHI for biochemical characterization and sub-
sequent crystallographic analysis (Figure S4). The same
purification protocol was also applied to the wild-type protein
co-workers stated that only tetramers were observed in
E. ramulus.[14] Regardless of this observation, the structure
of this bacterial CHI differs significantly from those of plant
CHIs; the protein–protein interactions of a dimer unit of the
recombinant bacterial CHI are very strong owing to the
incorporation of the C terminus of one protein chain into the
second protein chain, whereas the plant enzymes mainly
occur as monomers.[5] Moreover, in contrast to plant CHIs, an
internal symmetry exists in the tertiary structure of a mono-
2
www.angewandte.org
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2013, 52, 1 – 5
These are not the final page numbers!

Angewandte
Chemie
mer, as it can be divided into two domains with similar folds,
which are rotated by 908 against each other.
pET22b_sERED/CHI were cultivated aerobically (see the
Supporting Information), and biocatalytic transformations of
three structurally similar flavanones were performed by
whole cells under nitrogen atmosphere, as the aerobic process
led to lower and reversible activity. High conversions were
observed for naringenin, eriodictyol, and homoeriodictyol
(Scheme 1) within a short period of time (Table 1). The
In contrast to the expression of the CHI, expression of the
ERED was more challenging. As the ERED gene with the
native nucleotide sequence showed only a low expression
level in E. coli Rosetta, a synthetic codon-optimized gene of
ERED (sERED) was cloned into vector pET22b, which led
to overexpression of soluble sERED at 208C under aerobic
conditions. The protein had the expected molecular weight of
70 kDa, as determined by SDS-PAGE, which correlates well
with the theoretical size of 75 kDa. Unfortunately, only low
activity was observed for the conversion of the naringenin
chalcone into phloretin. As the enzyme originates from
a strictly anaerobic bacterium, we performed the expression
under anaerobic conditions. Under nitrogen atmosphere, the
highest expression level was observed at 258C. After ensuring
that all steps (cell harvest, lysis, and biocatalysis) were
performed under nitrogen atmosphere, significant enzymatic
activity in the reduction of the chalcone to phloretin could be
monitored. Furthermore, a crude cell extract that contained
the sERED could be added to (purified) CHI, and conversion
from naringenin into phloretin was observed. Pleasingly, the
undesired enzymatic degradation of phloretin that was
observed with E. ramulus did not take place with the E. coli
system, thus enabling the synthesis of the target product.
The high oxygen sensitivity of the ERED from E. ramulus
might be explained by its sequence similarity to a 2,4-dienoyl-
CoA reductase and the anaerobic enoate reductase from
Clostridium sp.[16a] These two enzymes are described as multi-
domain proteins with a barrel domain that is related to the
“Old Yellow Enzyme” with strong sequence conservation in
a core region of approximately 40 amino acids. The second
domain of this ERED is also redox-active and related to
glutathione reductases with an iron–sulfur cluster, and could
therefore be the oxygen-sensitive part of the protein. In this
sequence, the four cysteines that are required for the iron–
sulfur cluster were found in the typical motif
CXXCXXC(X)22C, a motif that also appears in the ERED
from E. ramulus, starting at C361 (Figure S2).[14b] It should be
noted that the sequence motif GXGXXG(X)17E for the
NADH and FAD binding site in glutathione reductase[19] is
found twice in the sequence of E. ramulus ERED.
Table 1: Conversion [%] of flavanones into their respective dihydrochal-
cones.
Substrate
Reaction time
2 h
1 h
17 h
Naringenin
Eriodictyol
Homoeriodictyol
69Æ1%
46Æ2%
47Æ3%
86Æ1%
51Æ4%
52Æ1%
93Æ1%
72Æ2%
63Æ2%
hydroxyl group in the para position of the phenyl ring appears
to be crucial for CHI activity; docking experiments of
naringenin at the active site of the CHI showed that this
hydroxyl group participates in a hydrogen-bonding network
with Asp79 and Gln101, which is crucial for the right
orientation of the substrate. On the other hand, these results
show that the CHI seems to be tolerant towards substituents
at the meta position of the phenol ring. Even though
naringenin is the preferred substrate, as it only bears a hydro-
gen atom at this position, it seems that the active site of CHI
can accommodate bulkier substituents, such as hydroxyl
(eriodictyol) or methoxy (homoeriodictyol) groups.
In summary, the results of our experiments confirmed the
metabolic pathway that was proposed by Herles and co-
workers[14] for the degradation of flavonoids by the identifi-
cation and successful recombinant expression of the chalcone
isomerase and an enoate reductase from the anaerobic
bacterium Eubacterium ramulus. The engineered E. coli
strain that expresses both enzymes can be used for the
conversion of several flavanones, which underlines the
applicability of the biocatalytic system that was developed
in this study.
Received: August 7, 2013
Revised: November 14, 2013
Published online: && &&, &&&&
To facilitate the industrial application of this bioprocess,
we envisaged the simultaneous expression of both enzymes in
an E. coli strain. To achieve this, the CHI gene was cloned
with a ribosome binding site directly behind the sERED gene
on the pET22b vector. With this new construct (pET22b_
sERED/CHI), the soluble and active expression of both
enzymes under anaerobic cultivation conditions was possible.
Biocatalysis with the crude cell extract allowed the produc-
tion of 50 mm phloretin after one hour under anaerobic
conditions. Under aerobic conditions, both enzymes can be
solubly expressed, although a significant amount of sERED is
produced in its insoluble form (Figure S9). However, the
amount of soluble sERED that was produced with this
cultivation method is large enough to push the equilibrium
towards the production of phloretin.
Keywords: chalcone isomerase · enoate reductase ·
.
enzyme biocatalysis · Eubacterium ramulus · flavonoids
[1] A. Crozier, I. B. Jaganath, M. N. Clifford, Nat. Prod. Rep. 2009,
26, 1001 – 1043.
[2] a) J. P. Ley, G. Krammer, G. Reinders, I. L. Gatfield, H.-J.
Bertram, J. Agric. Food Chem. 2005, 53, 6061 – 6066; b) S.
Scharbert, T. Hofmann, J. Agric. Food Chem. 2005, 53, 5377 –
5384; c) M. S. J. Simmonds, Phytochemistry 2001, 56, 245 – 252.
[3] a) M. N. Ngaki, G. V. Louie, R. N. Philippe, G. Manning, F. Pojer,
M. E. Bowman, L. Li, E. Larsen, E. S. Wurtele, J. P. Noel, Nature
2012, 485, 530 – 533; b) W. A. Peer, Nat. Chem. Biol. 2012, 8,
607 – 608.
To evaluate the substrate specificity of this system and
thus its versatility, E. coli BL21(DE3) cells with the construct
[4] J. M. Jez, J. P. Noel, J. Biol. Chem. 2002, 277, 1361 – 1369.
Angew. Chem. Int. Ed. 2013, 52, 1 – 5
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
3
These are not the final page numbers!

.
Angewandte
Communications
[5] J. M. Jez, M. E. Bowman, R. A. Dixon, J. P. Noel, Nat. Struct.
Biol. 2000, 7, 786 – 791.
[6] M. Gensheimer, A. Mushegian, Protein Sci. 2004, 13, 540 – 544.
[7] H. Schneider, M. Blaut, Arch. Microbiol. 2000, 173, 71 – 75.
[8] “Flavor suppression and enhancement”: J. P. Ley, K. Reichelt, G.
Krammer in Food Flavors. Chemical, Sensory and Technological
[13] a) M. Koffas, E. Leonard, Y. Yan, J. Chemler, Patent WO/2006/
010117, 2006; b) C. Schmidt-Dannert, K. Watts, Patent WO/
2005/084305, 2005.
[14] C. Herles, A. Braune, M. Blaut, Arch. Microbiol. 2004, 181, 428 –
434.
[15] C. Camacho, G. Coulouris, V. Avagyan, N. Ma, J. Papadopoulos,
K. Bealer, T. Madden, BMC Bioinf. 2009, 10, 421.
[16] a) R. E. Williams, N. C. Bruce, Microbiology 2002, 148, 1607 –
1614; b) F. Rohdich, A. Wiese, R. Feicht, H. Simon, A. Bacher, J.
Biol. Chem. 2001, 276, 5779 – 5787.
´
Properties (Ed.: H. H. Jelen), CRC Press, Boca Raton, USA,
2012, pp. 155 – 175.
[9] V. Siddaiah, C. V. Rao, S. Venkateswarlu, G. V. Subbaraju,
Tetrahedron 2006, 62, 841 – 846.
[10] F. Will, H. Zessner, H. Becker, H. Dietrich, Food Sci. Technol.
LEB 2007, 40, 1344 – 1351.
[17] L. Ralston, S. Subramanian, M. Matsuno, O. Yu, Plant Physiol.
2005, 137, 1375 – 1388.
[11] J. Ley, J.-M. Hilmer, I. L. Gatfield, European Patent 2017272,
2008.
[18] R. A. Bednar, J. R. Hadcock, J. Biol. Chem. 1988, 263, 9582 –
9588.
[12] P. Jonczyk, A. Schmidt, I. Bice, M. Gall, E. Gross, J.-M. Hilmer,
U. Bornscheuer, S. Beutel, T. Scheper, Chem. Ing. Tech. 2011, 83,
2147 – 2152.
[19] O. Dym, D. Eisenberg, Protein Sci. 2001, 10, 1712 – 1728.
4
www.angewandte.org
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2013, 52, 1 – 5
These are not the final page numbers!

Angewandte
Chemie
Communications
Biocatalysis
M. Gall, M. Thomsen, C. Peters,
I. V. Pavlidis, P. Jonczyk, P. P. Grꢁnert,
S. Beutel, T. Scheper, E. Gross, M. Backes,
T. Geißler, J. P. Ley,* J.-M. Hilmer,
G. Krammer, G. J. Palm, W. Hinrichs,
U. T. Bornscheuer*
&&&& — &&&&
The biocatalytic metabolic pathway for
the degradation of flavonoids of Eubac-
terium ramulus was identified. A chalcone
isomerase and an enoate reductase were
successfully cloned, subsequently
aerobic conditions, although E. ramulus is
Enzymatic Conversion of Flavonoids
using Bacterial Chalcone Isomerase and
Enoate Reductase
a strictly anaerobic bacterium. The engi-
neered E. coli strain that expresses both
enzymes was used for the conversion of
several flavanones.
expressed in E. coli, and used under
Angew. Chem. Int. Ed. 2013, 52, 1 – 5
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
5
These are not the final page numbers!