Coenzyme A-Conjugated Cinnamic Acids – Enzymatic Synthesis of a CoA-Ester Library and Application in Biocatalytic Cascades to Vanillin Derivatives

Article abstract of DOI:10.1002/adsc.201900892

We present a bioorthogonal method for the ligation of coenzyme A (CoA) with cinnamic acids. The reaction, which is the initial step in the biosynthesis of a multitude of bioactive secondary metabolites, is catalyzed by a promiscuous plant ligase and yields CoA conjugates with different functionalization in high purity and without formation of by-products. Its applicability in biosynthetic cascades is shown for the direct transformation of cinnamic acids into natural benzaldehydes (like vanillin) or artificial derivatives (e. g. ethylvanillin). (Figure presented.).

Full text of DOI:10.1002/adsc.201900892

Title: Coenzyme A-Conjugated Cinnamic Acids – Enzymatic Synthesis
of a CoA-Ester Library and Application in Biocatalytic Cascades
to Vanillin Derivatives
Authors: Martin Dippe, Anne-Katrin Bauer, Andrea Porzel, Evelyn
Funke, Anna Müller, Jürgen Schmidt, Maria Beier, and Ludger
A. Wessjohann
This manuscript has been accepted after peer review and appears as an
Accepted Article online prior to editing, proofing, and formal publication
of the final Version of Record (VoR). This work is currently citable by
using the Digital Object Identifier (DOI) given below. The VoR will be
published online in Early View as soon as possible and may be different
to this Accepted Article as a result of editing. Readers should obtain
the VoR from the journal website shown below when it is published
to ensure accuracy of information. The authors are responsible for the
content of this Accepted Article.
To be cited as: Adv. Synth. Catal. 10.1002/adsc.201900892
Link to VoR: http://dx.doi.org/10.1002/adsc.201900892

10.1002/adsc.201900892
Advanced Synthesis & Catalysis
COMMUNICATION
DOI: 10.1002/adsc.201((will be filled in by the editorial staff))
Coenzyme A-Conjugated Cinnamic Acids – Enzymatic
Synthesis of a CoA-Ester Library and Application in
Biocatalytic Cascades to Vanillin Derivatives
Martin Dippe, Anne-Katrin Bauer, Andrea Porzel, Evelyn Funke, Anna O. Müller,
Jürgen Schmidt, Maria Beier and Ludger A. Wessjohann*
Leibniz-Institute of Plant Biochemistry, Department of Bioorganic Chemistry, Weinberg 3, D-06120 Halle, Germany
Tel.: +49-345-55821301, Fax: +49-345-55821309, email: wessjohann@ipb-halle.de
Received: ((will be filled in by the editorial staff))
Supporting information for this article is available on the WWW under http://dx.doi.org/10.1002/adsc.201######.
starting point for the biosynthesis of the latter class of
compounds is the ATP-dependent conversion of
trans-4-coumaric acid (2, Scheme 1b) to its CoA
ester (4-coumaroyl-CoA) by 4-coumarate:CoA ligase
(4CL, E.C. 6.2.1.12) (Sche-me 1a).
Abstract. We present a bioorthogonal method for the
ligation of coenzyme A (CoA) with cinnamic acids. The
reaction, which is the initial step in the biosynthesis of a
multitude of bioactive secondary metabolites, is catalyzed
by a promiscuous plant ligase and yields CoA conjugates
with different functionalization in high purity and without
formation of by-products. Its applicability in biosynthetic
cascades is shown for the direct transformation of
cinnamic acids into natural benzaldehydes (like vanillin)
or synthetic derivatives (e.g. ethylvanillin).
Compared to other acyl donors involved in enzymatic
reactions (e.g. mixed anhydrides such as
phosphoacylates), CoA esters are much more stable
against hydrolytic cleavage under physiological
conditions. [10] This stability and the importance of the
compounds in a multitude of enzyme reactions
facilitated the development of chemosynthetic
strategies for their production, which are in general
based on the acylation of CoA or precursors with
reactive acyl chlorides, carbodiimide adducts[11] or
imidazolides.[10,12] Although high yields of the
thioesters can be achieved in these reactions, they
cannot be directly applied in bioorthogonal
(enzymatic) syntheses or in synthetic biology due to
the cross-reactivity of the used acyl donors with
reactants and enzymes. Thus, enzyme-mediated
syntheses seem to be the most convenient strategy for
in situ or in vivo generation of CoA conjugates in
biocatalytic cascades.
Keywords: Aldehydes; Biocatalysis; Carboxylic acids;
Ligases; Lyases
Coenzyme A (CoA) serves as a carrier and activator
for carboxylic acids in more than 100 types of
biochemical reactions.[1] Over 4 % of all known
enzymes catalyze transformations which involve such
CoA conjugates, e.g. the oxidoreductases acting in
fatty acid catabolism (such as beta-oxidation to
acetyl-CoA[2] and the subsequent oxidation of this
intermediate in the Krebs cycle)[3], in the reduction of
acids to aldehydes[4], or in hydroxylation of aromatic
acids.[5] The activation of the carboxylic acid
substrate by the reactive thioester bond of CoA esters
is also indispensable for acylation reactions on small
molecules and proteins, e.g. in gene-regulatory
histone acetylation.[6] In particular, a number of
primary and secondary metabolites are synthesized
via repeated condensation of CoA conjugates. These
reactions are the basis for several pathways to
polyketide natural products, such as antibiotics[7],
certain alkaloids and quinones[8] or flavonoids.[9] The
Ideally, a cascading reaction including a thioester
formation step should rely on a CoA ligase with
relaxed substrate specificity. The use of
a
promiscuous enzyme offers the possibility to react a
broad range of carboxylic acid substrates, providing
the basis for both the synthesis of natural or new-to-
nature derivatives in subsequent reaction steps. In the
search for a candidate with suitable substrate scope,
1
This article is protected by copyright. All rights reserved.

10.1002/adsc.201900892
Advanced Synthesis & Catalysis
the 4CL isoform 2 (4CL2) from the model plant
Arabidopsis thaliana
proved to be beneficial. The enzyme is characterized
by high activity and stability and – unlike other 4CL
enzymes which are prone to aggregation in
heterologous expression – can be produced in high
[14]
yield in Escherichia coli as production host.
In
particular, the ability of this ligase to convert the
three cinnamates 4-coumaric acid (2), caffeic acid (9)
and ferulic acid (10)[13] suggests a potential
promiscuity towards carboxylic acids. This tempted
us to probe the substrate scope of the enzyme.
Thus, the activity of 4CL2 was systematically tested
on a set of cinnamates, substituted either with one
hydroxyl (2 – 4), methyl (5), amino (6) or methoxyl
(7) group, or with several hydroxyl (9) and/or alkoxyl
substituents (10 – 16). Other compounds included in
the screening differed from the native substrate 2 in
the substitution of the phenyl group by the
heterocycle indole (8), or in saturation (17, 18) or
absence of the vinyl moiety (19). Strikingly,
conversion into the thioesters – which was detected
spectroscopically by depletion of the co-substrate
CoA – could be observed for all substrates except for
the highly substituted sinapic acid (16). Even benzoic
and arylacetic acids were accepted by the enzyme
(reaction rates at a substrate concentration of 600 μM
are given in brackets):
4CL2 transformed 4-
hydroxybenzoic acid (19) (27  1 nmol min-1 mg-1)
and indole-3-acetic acid (20) (2  1 nmol min-1 mg-1).
Among nine fatty acids with different chain lengths,#
only hexanoic acid (21) was converted (86  1 nmol
min-1 mg-1)*. The kinetic constants of 4CL2 (see Fig.
S1 in the Supporting Information) reflect its broad
substrate specificity but
Scheme 1. Synthesis of CoA esters in the biocatalytic
cascade to substituted benzaldehydes. a) Enzymatic
activation of cinnamic acids (left) yielding the
corresponding cinnamoyl-CoA thioesters (middle) by 4CL,
followed by FCoAHL-catalyzed hydration and retro-aldol
reaction to substituted benzaldehydes (right). b) Selected
acids used as 4CL substrates in this study. cinnamic acid
(1), 4-coumaric acid (2), 3-coumaric acid (3), 2-coumaric
acid (4), 4-methyl-cinnamic acid (5), 4-aminocinnamic
acid (6), 4-methoxy-cinnamic acid (7), 3-indoleacrylic acid
(8), caffeic acid (9), ferulic acid (10), isoferulic acid (11),
homoferulic acid (12), 4-ethoxy-3-hydroxycinnamic acid
(13), 3-hydroxy-4-n-propoxycinnamic acid (14), 5-hydro-
xyferulic acid (15), sinapic acid (16), dihydro-4-coumaric
acid (17), dihydrocaffeic acid (18), 4-hydroxybenzoic acid
(19), indole-3-acetic acid (20), hexanoic acid (21). The
yield (in %, calculated on the basis of CoA as the most
expensive component) in preparative transformations is
given in brackets. AMP, adenosine-5´-monophosphate;
CoA, coenzyme A; n.r., not reactive; ✔, reactive but not
used in preparative reaction due to low conversion rates
(see Tab. S1) or high susceptibility to oxidation. General
conditions: 50 mM Tris/HCl (pH 7.5), 2.5 mM MgCl2,
0.265 mM CoA, 2 mM acid substrate, 1.25 mM ATP, 25
µg ml-1 4CL2. Reaction time: 20 h. Fresh ligase enzyme
(25 µg ml-1 4CL2) and cosubstrates (0.265 mM CoA, 1.25
mM ATP) were added five hours after initiation of the
reactions. Products were purified by SPE (see Supporting
Information).
Figure 1. Conversion of cinnamic acids into vanillin and
related benzaldehydes by the 4CL2/FCoAHL cascade. The
chromatograms show the enzymatic formation of vanillin
(a) or ethylvanillin (b) from the corresponding cinnamic
acids (top row). Standards of the acid substrates and
vanillin products (structures depicted at top row) are
shown at the center and in the lowermost row
(chromatograms with dashed lines). The products of the
enzymatic reactions were identified by HR-MS/MS (gray
insets, see Supporting Information for methods). General
conditions: 50 mM Tris/HCl (pH 7.5), 2.5 mM MgCl2,
0.265 mM CoA, 2 mM acid substrate, 1.25 mM ATP, 25
2
This article is protected by copyright. All rights reserved.

10.1002/adsc.201900892
Advanced Synthesis & Catalysis
µg ml-1 4CL2. Reaction time: 20 h. Fresh ligase enzyme
(25 µg ml-1 4CL2) and cosubstrates (0.265 mM CoA, 1.25
mM ATP) were added five hours after initiation of the
reaction. The transformation into vanillins was started by
addition of FCoAHL
14 h).
OMe > p-OEt = p-OPr) (see Tab. S1 in the
Supporting Information).
The acceptance of the naturally occurring cinnamic
acids 2, 9 and 10 was also reported for other 4CL
enzymes which are involved in plant lignin
biosynthesis, such as the isoenzymes 4CL1 and 4CL3
(85 µg ml-1, incubation for
from A. thaliana[13], the orthologs from poplar[15]
,
rowanberry[16] and mulberry[17], or the structurally
characterized enzyme from tobacco[18]. In order to
gain insights into substrate recognition by these
promis-cuous enzymes, the active site architecture of
the tobacco protein was compared with structural
models of Arabidopsis 4CL2 and of the CoA ligase
from Scutellaria baicalensis. The substrate range of
the latter enzyme is restricted to unsubstituted
cinnamic acid 1 only.[19] The aromatic acid substrates
are bound in a cavity which is conserved in the three
homologs (Fig. 2). Interestingly, the binding sites of
the individual enzymes rather differ in amino acid
composition than in size (relative volume of the
cavity in the enzymes from A. thaliana/tobacco/S.
baica-lensis: 1/1.03/1.17). In the Arabidopsis enzyme,
the phenolic side chain of the carboxylic acid
substrate seems to be bound via a combination of a
serine, glutamine and tyrosine residue (S257, Y253,
Q227, see Fig. 2).[14a] These amino acids are also
present in the tobacco enzyme and other broad-
spectrum CoA ligases in homologous positions (see
Fig. S4 in the Supporting Information). The binding
cavity of S. baicalensis contrasts by being more
hydrophobic, with the respective residues being
replaced by valine, phenylalanine and isoleucine.
This hydrophobization seems to be responsible for
the enzyme’s selectivity towards unsubstituted
cinnamic acid (1), excluding more hydrophilic
carboxylic acids such as coumarates.[19] As reported
in previous studies, subtle differences in the active
sites strongly influence the reactivity of 4CL
enzymes[18] – a fact which suggests that both rational
engineering and activity screening of yet
uncharacterized proteins will facilitate access to
tailored ligase biocatalysts in the future.
Figure 2. Active sites of plant 4CL enzymes. Modelled
structures of 4CL2 from A. thaliana (dark blue) and
cinnamic acid-CoA ligase from S. baicalensis (yellow)
were aligned with the crystal structure of the tobacco
enzyme[18] (cyan). Upper panel: Substrate binding sites
occupied by the reaction intermediate 4-coumaroyl
adenosine 5´-mono-phosphate. Amino acids involved in
binding (numbering of positions indicated for the
Arabidopsis enzyme) are shown in stick representation.
Lower panel: Sphere model of the cavity accommodating
the aromatic acid moiety. The coloration of the amino acid
residues, bound molecules and sphere models is the same
as for the protein scaffolds.
To confirm formation of the thioesters in the
spectrophotometric assays
mentioned above,
preparative conversions of the cinnamic acid
derivatives 1 – 6, 8 – 15, and 18 were performed. The
one-pot synthetic reactions were incubated for a
prolonged reaction time (20 hours) and contained a
7.5-fold excess of the acid substrate to achieve
complete consumption of the added CoA (2.65 µmol).
The CoA esters were subsequently separated from
other components present in the samples in a single
simple purification step. The crude reaction mixtures
containing the thioester product were applied to solid
phase extraction (SPE) cartridges and buffer salts, the
also hints to a distinct selectivity towards the
different cinnamic acids. The maximum reaction rate
(vmax) specifically depends on the substitution pattern
of the aromatic ring, e.g. on the type of the
substituent (p-OH > p-H = p-Me (p-OH > m-OH = o-
OH) or alkoxy residues (p-OMe in 11 > m-OMe in
10)) and on the size of the attached alkyl chain (p-
3
This article is protected by copyright. All rights reserved.

10.1002/adsc.201900892
Advanced Synthesis & Catalysis
co-product adenosine-5´-mono-phosphate and 4CL2
were removed by a washing step. Pure CoA esters
were recovered from the SPE matrix in excellent
purity (see Supporting Information) and in good to
moderate yield (Scheme 1b). The identity of the
products was confirmed by HR-MS and NMR.
Formation of side products was not observed in any
of the enzymatic reactions.
A distinguished advantage of the incorporation of
CoA ligases in enzymatic multistep reactions is the
fact that their labile reaction products will be directly
transformed into follow-up products without the need
of purification or the risk of hydrolysis.[20] Even
combinatorial approaches, i.e. precursor-directed
biosyntheses using e.g. non-natural substrates, are
possible if the consecutive steps are catalyzed by
promiscuous biocatalysts. In order to demonstrate the
feasibility of such a cell-free system in a proof-of-
concept study, we combined 4CL2 with feruloyl-CoA
Footnotes
#
Saturated fatty acids having an even number (2 – 16) of
carbon atoms were used in the activity assays.
* This strict selectivity is probably caused by steric
hindrance which prevents occupation of long-chain fatty
acids in the active site. The calculated length of a
hexanoic acid molecule (8.90 Å) is similar to that of the
natural substrate 4-coumaric acid (2) (10.0 Å).
Experimental Section
All reagents were of the highest quality available and were
purchased from Sigma-Aldrich (St. Louis, USA) (cinnamic
acids, Ellman´s reagent) or Applichem (Darmstadt,
Germany) (CoA). 4CL2 and FCoAHL were produced as
recombinant proteins (as N-terminal fusion to maltose-
binding protein / with C-terminal decahistidine tag (4CL2),
or with an N-terminal hexahistidine tag (FCoAHL)) in E.
coli and were purified by metal affinity chromatography.
For a detailed description of experimental procedures see
Section “Supporting methods” in the Supporting
Information.
hydratase/lyase (FCoAHL, E.C. 4.2.1.101),
a
bacterial hydratase which catalyzes the Michael
addition of water to the acrylate moiety of feruloyl-
CoA. The intermediate β-hydroxypropionic thioester
then undergoes retro-aldol reaction to vanillin
(structure depicted in Fig. 1a) and the coupled by-
product acetyl-CoA.[21] Vanillin is produced when
FCoAHL is added to a 4CL2 reaction containing
ferulic acid 10 (Fig. 1a). As the hydratase showed
activity not only on naturally occurring CoA esters
(e.g. 4-coumaroyl- and feruloyl-CoA) but also
towards artificial derivatives (see Fig. S2 in
Supporting Information), we exploited the enzyme
cascade for direct synthesis of non-natural
benzaldehydes from the corresponding cinnamic
acids. Indeed, the formation of unprotected
Acknowledgements
We thank Anja Ehrlich and Julia Christke for excellent technical
assistance, and Annegret Laub and Dr. Andrej Frolov (IPB
Halle) for their help with the MS/MS measurements. Parts of the
research leading to these results have received funding from the
European Union´s Seventh Framework Program FP7/2007-2013
under grant agreement no 266025 (BIONEXGEN, MD), from the
BMBF (P34 project Biokatalyse 2021, AKB), from the Leibniz
Research Cluster “Bio/synthetic multifunctional micro
production units”, and from the ScienceCampus Halle - Plant-
Based Bioeconomy, cofinanced by the Land Sachsen-Anhalt and
the Leibniz Association.
ethylvanillin
–
an important artificial aroma
compound – was achieved in samples con-taining 12
as starting material via homoferulate-CoA (Fig. 1b).
References
In conclusion, broad-spectrum ligases like 4CL2 are
indispensable tools in the synthesis of activated
carboxylates, which are key intermediates in natural
or artificial pathways to various valuable naturally
produced products. In subsequent studies, the
reaction will be extended to multi-enzyme sequences
for the biosynthesis of non-natural products by
adding compatible catalysts such as cyclizing
enzymes (e.g. polyketide synthases) or oxidases (e.g.
coumarin biosynthetic enzymes). Moreover, such
cascade reactions can be used either for structural
diversification of intermediates and products (if the
reactions include promiscuous downstream enzymes)
or in the screening for biocatalysts acting on CoA
conjugates.
[1] R. Leonardi, Y. M. Zhang, C. O. Rock, S. Jackowski,
Prog. Lipid Res. 2005, 44, 125-153.
[2] S. M. Houten, R. J. A. Wanders, J. Inherit. Metab.
Dis. 2010, 33, 469-477.
[3] a) J. M. Berg, J. L. Tymoczko, L. Stryer, in Bio-
chemistry, 5th ed. (Eds.: J.M. Berg, J.L. Tymoczko, L.
Stryer), W. H. Freeman, New York, 2002, pp. 501-
527; b) D. Voet, J. G. Voet, in Biochemistry, 4th ed.,
John Wiley & Sons, Hoboken, 2011, pp. 789-823.
[4] W. Boerjan, J. Ralph, M. Baucher, Annu. Rev. Plant
Biol. 2003, 54, 519-46.
[5] K. Kai, M. Mizutani, N. Kawamura, R. Yamamoto,
M. Tamai, H. Yamaguchi, Plant J. 2008, 55, 989-999.
[6] a) S. M. Houten, R. J. A. Wanders, J. Inherit. Metab.
Dis. 2010, 33, 469-477; b) M. D. Resh, Biochim.
Biophys. Acta - Mol. Cell Res. 1999, 1451, 1-16.
[7] S. Ma, L. B. Pickens, Y. Tang, in Encyclopedia of
Industrial Biotechnology, 1st ed. (Eds.: M. C.
Flickinger, S. W. Drew, R. E. Spier), John Wiley &
4
This article is protected by copyright. All rights reserved.

10.1002/adsc.201900892
Advanced Synthesis & Catalysis
Sons, Hoboken, 2009, pp. 1-11.
D. Schomburg, E. Kombrink, H.-P. Stuible, Proc.
Natl. Acad. Sci. USA. 2003, 100, 8601-8606; b) H. P.
Stuible, E. Kombrink, J. Biol. Chem. 2001, 276,
26893-26897.
[8] a) R. B. Herbert, Nat. Prod. Rep. 1999, 16, 199-208;
b) A. J. Humphrey, D. O’Hagan, Nat. Prod. Rep.
2001, 18, 494-502; c) K. Springob, R. Lukačin, C.
Ernwein, I. Gröning, U. Matern, Eur. J. Biochem.
2000, 267, 6552-6559; d) M. F. Roberts, D. Strack, M.
Wink, in Annual Plant Reviews, vol. 40 (Ed.: M.
Wink), Wiley-VCH, Weinheim, 2008, pp. 20-91. e)
M. Petersen, J. Hans, U. Matern, in Ann. Plant Rev.,
vol. 40 (Ed.: M. Wink), Wiley-VCH, Weinheim,
2008, pp. 182-257.
[15] Y. Hu, Y. Gai, L. Yin, X. Wang, C. Feng, L. Feng, D.
Li, X.-N. Jiang, D.-C. Wang, Plant Cell 2010, 22,
3093-3104.
[16] M. M. Gaid, H. Scharnhop, H. Ramadan, T. Beuerle,
L. Beerhues, J. Plant Physiol. 2011, 168, 944-951.
[17] C.-H. Wang, J. Yu, Y. X. Cai, P.-P. Zhu, C.-Y. Liu,
A.-C. Zhao, R.-H. Lü, M.-J. Li, F.-X. Xu, M.-D. Yu,
PLoS ONE 2016, 11, e0155814.
[9] T. Vogt, Mol. Plant 2010, 3, 2–20.
[10] V. Agarwal, S. Diethelm, L. Ray, N. Garg, T.
Awakawa, P. C. Dorrestein, B.S. Moore, Org. Lett.
2015, 17, 4452-4455.
[11] a) P. K. Mishra, D. G. Drueckhammer, Chem. Rev.
2000, 100, 3283-3310; b) J. Stöckigt, M. H. Zenk, Z.
Naturforschung. C, A. J. Biosci. 1975, 30, 352-358.
[12] a) J. E. Bishop, A. K. Hajra, Anal. Biochem. 1980,
106, 344-350; b) A. Kawaguchi, T. Yoshimura, S.
Okuda, J. Biochem. 1981, 89, 337-339.
[18] Z. Li, S. K. Nair, Structure 2015, 23, 2032-2042.
[19] Q. Zhao,Y. Zhang, G. Wang, L. Hill, J. K. Weng,
X.Y. Chen, H. Xue, C. Martin, Plant Biochem. 2016,
2, e1501780.
[20] J. M. Luengo, J. Antibiot. (Tokyo) 1995, 48, 1195-
1211.
[21] A. Mitra, Y. Kitamura, M. J. Gasson, A. Narbad, A. J.
Parr, J. Payne, M. J. Rhodes, C. Sewter, N. J. Walton,
Arch. Biochem. Biophys. 1999, 365, 10-16.
[13] a) J. Ehlting, D. Büttner, Q. Wang, C. J. Douglas, I. E.
Somssich, E. Kombrink, Plant J. 1999, 19, 9-20; b)
M. Pietrowska-Borek, H.-P. Stuible, E. Kombrink, A.
Guranowski, Plant Physiol. 2003, 131, 1401-1410.
[14] a) K. Schneider, K. Hövel, K. Witzel, B. Hamberger,
5
This article is protected by copyright. All rights reserved.

10.1002/adsc.201900892
Advanced Synthesis & Catalysis
COMMUNICATION
Coenzyme A-Conjugated Cinnamic Acids
–
Enzymatic Synthesis of a CoA-Ester Library and
Application in Biocatalytic Cascades to Vanillin
Derivatives
Adv. Synth. Catal. Year, Volume, Page – Page
Martin Dippe, Anne-Katrin Bauer, Andrea Porzel,
Evelyn Funke, Anna O. Müller, Jürgen Schmidt,
Maria Beier and Ludger A. Wessjohann *
6
This article is protected by copyright. All rights reserved.