One-Pot Conversion of Cinnamaldehyde to 2-Phenylethanol via a Biosynthetic Cascade Reaction

Article abstract of DOI:10.1021/acs.orglett.9b02611

A novel biosynthetic pathway for the production of natural 2-phenylethanol from cinnamaldehyde is reported. An ene-reductase (OYE)-mediated selective hydrogenation of cinnamaldehyde to hydrocinnamaldehyde is followed by a regioselective Baeyer-Villiger oxidation (BVMO) to produce the corresponding formate ester that either spontaneously hydrolyzes to 2-phenylethanol in water or is assisted by a formate dehydrogenase (FDH). This cascade reaction is performed in a one-pot fashion at ambient temperature and pressure. High selectivity and complete conversion were achieved.

Full text of DOI:10.1021/acs.orglett.9b02611

Letter
Cite This: Org. Lett. XXXX, XXX, XXX−XXX
pubs.acs.org/OrgLett
One-Pot Conversion of Cinnamaldehyde to 2‑Phenylethanol via a
Biosynthetic Cascade Reaction
Amanda Vorster, Martha S. Smit, and Diederik J. Opperman*
Department of Biotechnology, University of the Free State, Bloemfontein 9300, South Africa
*
S Supporting Information
ABSTRACT: A novel biosynthetic pathway for the produc-
tion of natural 2-phenylethanol from cinnamaldehyde is
reported. An ene-reductase (OYE)-mediated selective hydro-
genation of cinnamaldehyde to hydrocinnamaldehyde is
followed by a regioselective Baeyer−Villiger oxidation
(
BVMO) to produce the corresponding formate ester that
either spontaneously hydrolyzes to 2-phenylethanol in water
or is assisted by a formate dehydrogenase (FDH). This
cascade reaction is performed in a one-pot fashion at ambient
temperature and pressure. High selectivity and complete
conversion were achieved.
2
-Phenylethanol (2-PE) is an aromatic alcohol with a roselike
ally, 2-PE can also be further metabolized and degraded. 2-PE
production is also eventually limited by its toxicity to growing
cells. For the efficient production of 2-PE from L-Phe in situ
product removal to avoid 2-PE toxicity is thus essential. Much
research has been done on improving the productivity of this
1
aroma. It is an important chemical used in the food and
fragrance industry, with annual production of 2-PE estimated
at more than 10000 tons. Traditionally natural 2-PE is
extracted and purified from flowers, specifically the hydro-
distillation of rose petals. This natural route, however, yields
2
3
18−21
biological route, e.g., biphasic or in situ product removal
very low product recovery with very high cost implications. To
meet the current global demands for 2-PE, most 2-PE is thus
currently synthesized chemically. 2-PE can be chemically
to overcome the toxicity of 2-PE and genetic engineering of
2
2
yeasts strains to increase space−time yields (STY).
Engineered bacterial strains mimicking the Ehrlich pathway
have also been created for the production of 2-PE from L-
4
synthesized by a Grignard reaction from chlorobenzene or via
5
2
3−25
Friedel−Crafts alkylation of benzene. Both of these chemical
Phe
pathway.
styrene monooxygenase (SMO), styrene oxide isomerase
SOI), and phenylacetaldehyde reductase (PAR) was shown
or from glucose by exploiting the shikimate
2
6−28
routes have several drawbacks including the use of hazardous
or corrosive chemicals, difficult separation mixtures, and low
selectivity. Alternatively, 2-PE production has been demon-
strated via the catalytic hydrogenation of styrene oxide.
Originally proven using Raney nickel as catalyst and hydrogen
More recently, an E. coli strain coexpressing
(
29
to catalyze the hydration of styrene to 2-PE. 2-PE production
via this styrene pathway was also recently extended, enabling
the conversion of L-Phe to styrene by introducing phenyl-
alanine ammonia lyase (PAL) and phenylacrylic acid
decarboxylase (PAD). Similarly, an engineered styrene
producing E. coli strain was further modified through the
6
7−9
gas, other nonpyrophoric catalysts have been developed.
Although the cost of chemical 2-PE is significantly lower than
that of natural 2-PE, chemically synthesized 2-PE is limited in
its use as an aroma compound in food, beverages, and
cosmetics. Not only do these reactions rely on petrochemical
feedstocks, but the formation of various side products, which at
even very low concentrations can destroy the aroma of 2-PE.
The increased demand for natural products has seen the
30
31
introduction of SMO and SOI for 2-PE production from
32
−1
glucose. Although product titers of ca. 2 g L were reached,
high glucose loading was required with yields of only 61 mg 2-
−
1
2,10−12
PE g glucose. 2-PE production from glucose could be
significantly improved (ca. 5-fold) by utilizing two E. coli
strains to couple L-Phe production from glucose and its further
conversion to 2-PE. Despite improved 2-PE titers, the system
likewise required high glucose concentrations with L-Phe yields
rapid development of biotechnological routes to 2-PE.
The US Food and Drug Administration and European
legislation state that products from biotechnological (enzy-
matic or microbiological) processes can be classified as natural
33
13,14
if the substrate used is of natural origin.
Yeasts such as
−
1
of only 60 mg g glucose.
Saccharomyces cerevisiae and Kluyveromyces marxianus can
15−17
We propose a new synthetic route for natural 2-PE
production from inexpensive and abundant cinnamaldehyde
convert L-phenylalanine via the Ehrlich pathway
to 2-PE
via phenylpyruvate and phenylacetaldehyde when actively
metabolizing cells are given L-Phe as the sole nitrogen source.
The intermediates can, however, be overoxidized via
endogenous dehydrogenases present in these yeasts. Addition-
Received: July 25, 2019
©
XXXX American Chemical Society
A
DOI: 10.1021/acs.orglett.9b02611
Org. Lett. XXXX, XXX, XXX−XXX

Organic Letters
Scheme 1) mediated via a biocatalytic cascade reaction
utilizing a novel Baeyer−Villiger monooxygenase (BVMO).
Letter
(
glycerol to allow cofactor recycling via E. coli central
metabolism. Similar low conversion and high side-product
formation were observed.
In an effort to avoid side-product formation and increase 2-
PE yields, we decided to change to an in vitro system using
purified biocatalysts. Reaction mixtures (1 mL) contained 2
μM of ER and BVMO and 1 U of purified glucose
dehydrogenase (BmGDH) with 100 mM glucose for cofactor
regeneration. Disappointingly low concentrations of 2-PE were
again observed (<2 mM) but with side product formation
drastically reduced. Examination of the time-course analysis of
the intermediates of the cascade revealed the ER-mediated
reduction step to proceed rapidly, with the complete reduction
of the cinnamaldehyde within 1−2 h and the rate-limiting step
the autohydrolysis of the phenethyl formate to 2-PE (Figure 1.
Scheme 1. Biosynthetic Pathway of 2-Phenylethanol from
Cinnamaldehyde via an Enzymatic Cascade Involving an
Ene-Reductase from the OYE Family of Enzymes and a
Baeyer−Villiger Monooxygenase and Water-Assisted
Hydrolysis
Our previous investigations into BVMOs revealed
BVMOAFL838 from Aspergillus flavus to uniquely and preferen-
tially produce formyl esters rather than fatty acids from
3
4
aliphatic aldehydes. Testing of BVMOAFL838 and its
orthologue from A. oryzyae (BVMOAO) against hydro-
cinnamaldehyde gave exclusively the formyl ester which
spontaneously hydrolyzes in aqueous solution to 2-phenyl-
ethanol. It has also been reported extensively in literature that
ene-reductases (ERs) from the Old Yellow Enzyme (OYE)
family can reduce the α,β-unsaturated double bond of
35−38
cinnamaldehyde and its derivatives.
Indeed, screening of
five recombinant ERs (Figure S3) revealed the ERs from S.
cerevisiae (“classical” OYEs), also commonly referred to as
OYE2 and OYE3, to rapidly reduce the activated C−C double
bond of cinnamaldehyde. Further kinetic characterization
revealed OYE3 to have approximately three times higher
specific activity (Vmax) for cinnamaldehyde than OYE2, and
Figure 1. Time course of the conversion of cinnamaldehyde to 2-
phenylethanol via the biocatalytic cascade reaction using glucose
dehydrogenase for cofactor regeneration. Conditions: 50 mM Tris-
HCl buffer (pH 8), [ScOYE3] = 2 μM, [BVMO ] = 2 μM,
AO
−1
+
[
[
BmGDH] = 1 U mL , [glucose] = 100 mM, [NADP ] = 0.3 mM,
cinnamaldehyde] = 10 mM, T = 25 °C, shaking = 200 rpm.
also a lower K (0.07 mM). Despite mild product inhibition,
M
the catalytic efficiency of OYE3 was still higher than that of
OYE2 and was therefore selected (Figure S5).
Thus, as an initial proof-of-principle, an in vivo cascade was
constructed in E. coli, with simultaneous recombinant
A Tris adduct, formed as a Schiff base with hydro-
cinnamaldehyde, was also observed in the earlier stages of
the reaction. This reversible reaction, occurring at higher pH
values, decreased the initial effective hydrocinnamaldehyde
concentration, potentially alleviating the observed substrate
inhibition of BVMOAO with hydrocinnamaldehyde (Figure
S6).
Overall, the reaction leveled off after only 4 h, with no
further conversion of hydrocinnamaldehyde by the BVMO nor
autohydrolysis of the already produced phenethyl formate.
Evaluation of the pH after 24 h of biotransformation revealed
significant acidification of the reaction (pH < 4).
expression of OYE3 and BVMO . The pET-Duet-1 vector
AO
39
containing the open reading frames of both biocatalysts were
transformed into E. coli BL21(DE3) and grown for 12 h at 25
°C, after which 10 mM cinnamaldehyde was introduced to the
growing culture and conversion determined after 2 and 24 h.
Despite complete conversion of the cinnamaldehyde after 2 h,
only ca. 2.7 mM 2-PE was observed, which increased to ca. 4.2
mM after 24 h (Figure S1). Although 2-PE was obtained, the
low yields with various side products formed, such as the
corresponding alcohols from cinnamaldehyde and hydro-
cinnamaldehyde, could be attributed to the action of
endogenous enzymes of E. coli such as alcohol dehydrogenases
Glucose dehydrogenase is known to form gluconic acid
40
during cofactor recycling, lowering the pH after prolonged
reactions to below the operational levels for many biocatalysts.
However, considering the concentrations of the substrates and
intermediates utilizing NADPH during the biotransformation,
this atypically fast and drastic acidification could be attributed
to the uncoupling of the OYE. OYEs are known to also readily
(
ADHs). Despite many of these reactions being reversible,
phenacetaldehyde and benzyl alcohol were also observed,
suggesting the endogenous ADHs are also able to convert the
desired 2-PE to the corresponding aldehyde and further
conversion by the BVMO. Biotransformations under non-
growing conditions in only buffer were also evaluated. As both
of these enzymes require NADPH as cofactor, the reaction
mixtures were supplemented with 100 mM glucose and
35
reduce molecular oxygen, leading to the formation of reactive
oxygen species and the depletion of glucose (and thus
excessive gluconic acid production) even in the absence of
B
DOI: 10.1021/acs.orglett.9b02611
Org. Lett. XXXX, XXX, XXX−XXX

Organic Letters
Letter
substrate. Incubation of phenethyl formate in aqueous buffers
at different pH values also showed a significant pH dependence
of the rate of hydrolysis, with significantly lower rates observed
at neutral pH values and almost none at pH 6 (Figure S7). The
cascade was again tested with the buffering capacity increased
the molar equivalents of reduced cofactor (Scheme 2).
Additional formate was thus included as cosubstrate to redox
Scheme 2. Biosynthetic Pathway of 2-Phenylethanol from
Cinnamaldehyde via an Enzymatic Cascade Involving an
Ene-Reductase from the OYE Family of Enzymes and a
Baeyer−Villiger Monooxygenase with Formate
Dehydrogenase Mediated Cofactor Regeneration and
Oxidative Cleavage of Phenethyl Formate
(
200 mM Tris, pH 8). Nearly complete conversion of the
cinnamaldehyde to 2-PE was observed after only 8 h, with only
trace amounts of intermediates (with the exception of
phenethyl formate) observed after 4 h (Figure 2). Complete
balance the cascade. As the observed specific activity of
CbFDH is much lower than that of BmGDH, reactions were
constrained and contained only 0.2 U of CbFDH. Cofactor
regeneration now became the limiting factor, as cinnamalde-
hyde was only completely converted after 4 h, and significant
amounts of intermediates other than phenethyl formate were
observed at 8 h. Nearly complete conversion to 2-PE was,
however, still observed after 12 h (Figure 3) and no
acidification was observed, allowing for the cascade reaction
to proceed in a low concentration buffer.
Figure 2. Time course of the conversion of cinnamaldehyde to 2-
phenylethanol via the biocatalytic cascade reaction using glucose
dehydrogenase for cofactor regeneration under higher buffer
concentrations. Conditions: 200 mM Tris−HCl buffer (pH 8),
−
1
[
ScOYE3] = 2 μM, [BVMO ] = 2 μM, [BmGDH] = 1 U mL ,
AO
+
[
glucose] = 100 mM, [NADP ] = 0.3 mM, [cinnamaldehyde] = 10
mM, T = 25 °C, shaking = 200 rpm.
conversion was obtained after 12 h, but surprisingly, benzyl
alcohol was again observed as a minor byproduct (Figure S4).
As no ADHs are present and none of the enzymes have been
found to possess the ability to oxidize 2-PE, the manner for
benzyl alcohol formation is currently unknown.
To avoid the constraint of exceedingly high buffer
concentrations, we decided to replace the GDH with a
formate dehydrogenase (FDH). FDH is a common cofactor
+
regenerating enzyme utilizing formate to regenerate NAD to
NADH with only CO as byproduct. Wild-type FDH, however,
2
+
typically only accepts NAD and not its phosphorylated
+
41
counterpart NADP . Two mutants of the FDH from Candida
boidini (CbFDH) have been described in the literature with the
+
ability to also accept NADP . These two mutant CbFDHs
42
43
(
designated as CbFDH_P1 and P2 ) were thus created
Figure 3. Time course of the conversion of cinnamaldehyde to 2-
phenylethanol via the biocatalytic cascade reaction using formate
dehydrogenase for cofactor regeneration. Conditions: 50 mM Tris-
through site-directed mutagenesis for this study. Higher
turnover frequencies (TOFs) were observed with CbFDH_P2
under the tested conditions and were selected for NADPH
regeneration. Moreover, FDHs have previously been demon-
strated to accept formate esters as alternative substrates. This
HCl buffer (pH 8), [ScOYE3] = 2 μM, [BVMO ] = 2 μM,
AO
−
1
+
[
CbFDH_P2] = 0.2 U mL , [formate] = 50 mM, [NADP ] = 0.3
mM, [cinnamaldehyde] = 10 mM, t = 25 °C, shaking = 200 rpm.
oxidative ester cleavage also yields terminal alcohols and CO
2
44,45
instead of formic acid.
Phenethyl formate was tested
In summary, we report here a novel biocatalytic one-pot
cascade reaction for the conversion of cinnamaldehyde to 2-
phenylethanol. This cascade allows for the conversion of cheap
and abundant cinnamon to natural rose flavor, a high-value fine
chemical in the food and fragrance industry. Total turnover
+
against CbFDH (and the NADP -specific mutant P2), which
proved to be accepted as a substrate with oxidative hydrolysis
to 2-PE, albeit at very low reaction rates.
Although it would be advantageous to have a redox-balanced
cascade reaction, the proposed cascade reaction requires twice
numbers, with respect to either ScOYE3 or BVMO , of more
AO
C
DOI: 10.1021/acs.orglett.9b02611
Org. Lett. XXXX, XXX, XXX−XXX

Organic Letters
Letter
than 4000 were routinely obtained, and space time yields of
(18) Etschmann, M. M. W.; Schrader, J. Appl. Microbiol. Biotechnol.
between 0.07 and 0.09 g L−1
h
−1
(aqueous phase) were
2
(
2
(
006, 71, 440−443.
19) Etschmann, M. M. W.; Sell, D.; Schrader, J. Biotechnol. Bioeng.
005, 92, 624−634.
20) Wang, H.; Dong, Q.; Meng, C.; Shi, X.; Guo, Y. Enzyme Microb.
Technol. 2011, 48, 404−407.
achieved in these initial unoptimized proof-of-principle
experiments. Upon complete conversion, isolated yields of at
least 60% (6 mM 2-PE) were typically obtained.
(
21) Gao, F.; Daugulis, A. J. Biotechnol. Bioeng. 2009, 104, 332−339.
ASSOCIATED CONTENT
Supporting Information
■
(22) Kim, B.; Cho, B.-R.; Hahn, J.-S. Biotechnol. Bioeng. 2014, 111,
115−124.
(
*
S
23) Liu, J.; Jiang, J.; Bai, Y.; Fan, T.; Zhao, Y.; Zheng, X.; Cai, Y. J.
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/acs.or-
glett.9b02611.
Agric. Food Chem. 2018, 66, 3498−3504.
(24) Achmon, Y.; Ben-Barak Zelas, Z.; Fishman, A. Appl. Microbiol.
Biotechnol. 2014, 98, 3603−3611.
Biocatalyst production and purification, experimental
procedures, characterization data, and GC−MS spectra
of products and intermediates (PDF)
(25) Hwang, J.-Y.; Park, J.; Seo, J.-H.; Cha, M.; Cho, B.-K.; Kim, J.;
Kim, B.-G. Biotechnol. Bioeng. 2009, 102, 1323−1329.
(
(
26) Atsumi, S.; Hanai, T.; Liao, J. C. Nature 2008, 451, 86−89.
27) Guo, D.; Zhang, L.; Kong, S.; Liu, Z.; Li, X.; Pan, H. J. Agric.
Food Chem. 2018, 66, 5886−5891.
AUTHOR INFORMATION
Corresponding Author
(28) Kim, T.-Y.; Lee, S.-W.; Oh, M.-K. Enzyme Microb. Technol.
■
*
2
014, 61−62, 44−47.
(
29) Wu, S.; Liu, J.; Li, Z. ACS Catal. 2017, 7, 5225−5233.
Tel: +27 51 401 2714. E-mail: opperdj@ufs.ac.za.
(30) Lukito, B. R.; Wu, S.; Saw, H. J. J.; Li, Z. ChemCatChem 2019,
ORCID
11, 831−840.
(
31) McKenna, R.; Pugh, S.; Thompson, B.; Nielsen, D. R.
Diederik J. Opperman: 0000-0002-2737-8797
Notes
Biotechnol. J. 2013, 8, 1465−1475.
(32) Machas, M. S.; McKenna, R.; Nielsen, D. R. Biotechnol. J. 2017,
1
(
2, 1700310.
The authors declare no competing financial interest.
33) Sekar, B. S.; Lukito, B. R.; Li, Z. ACS Sustainable Chem. Eng.
2
019, 7, 12231−12239.
ACKNOWLEDGMENTS
■
(34) Ferroni, F. M.; Tolmie, C.; Smit, M. S.; Opperman, D. J.
ChemBioChem 2017, 18, 515−517.
We thank Mr. Sarel Marais (University of the Free State) for
GC−MS analyses. This study was funded by the Technology
Innovation Agency (TIA), South Africa.
(35) Toogood, H. S.; Gardiner, J. M.; Scrutton, N. S. ChemCatChem
2010, 2, 892−914.
(
36) Winkler, C. K.; Tasnad
Biotechnol. 2012, 162, 381−389.
37) Toogood, H. S.; Scrutton, N. S. ACS Catal. 2018, 8, 3532−
549.
38) Toogood, H. S.; Scrutton, N. S. Curr. Opin. Chem. Biol. 2014,
9, 107−115.
39) Bubb, W. A.; Berthon, H. A.; Kuchel, P. W. Bioorg. Chem. 1995,
3, 119−130.
40) Kaswurm, V.; Van Hecke, W.; Kulbe, K. D.; Ludwig, R. Adv.
́
i, G.; Clay, D.; Hall, M.; Faber, K. J.
REFERENCES
■
(
3
(
1
(
2
(
(
1) Fahlbusch, K.-G.; Hammerschmidt, F.-J.; Panten, J.;
Pickenhagen, W.; Schatkowski, D.; Bauer, K.; Garbe, D.; Surburg,
H. Flavors and Fragrances. In Ullmann’s Encyclopedia of Industrial
Chemistry; Wiley-VCH: Weinheim, 2003; Vol. 15, pp 735−768.
(
(
2) Hua, D.; Xu, P. Biotechnol. Adv. 2011, 29, 654−660.
3) Agarwal, S. G.; Gupta, A.; Kapahi, B. K.; Baleshwar; Thappa, R.
K.; Suri, O. P. J. Essent. Oil Res. 2005, 17, 265−267.
4) Weissenborn, A. U.S. Patent 2058373, 1936.
5) Schaarschmidt, A.; Hermann, L.; Szemzo, B. Ber. Dtsch. Chem.
Ges. B 1925, 58, 1914−1916.
6) Wood, T. F. U.S. Patent US252409A, 1950.
Synth. Catal. 2013, 355, 1709−1714.
(
(
(
(
41) Tishkov, V. I.; Popov, V. O. Biomol. Eng. 2006, 23, 89−110.
42) Rozzell, J.; Hua, L.; Mayhew, M.; Novick, S. U.S. Patent
̈
US2004/0115691 A1, 2004.
43) Wu, W.; Zhu, D.; Hua, L. J. Mol. Catal. B: Enzym. 2009, 61,
57−161.
44) Fro
941−7950.
45) Churakova, E.; Tomaszewski, B.; Buehler, K.; Schmid, A.;
Arends, I.; Hollmann, F. Top. Catal. 2014, 57, 385−391.
(
(
(
1
(
7
7) Sasu, A.; Dragoi, B.; Ungureanu, A.; Royer, S.; Dumitriu, E.;
Hulea, V. Catal. Sci. Technol. 2016, 6, 468−478.
8) Rode, C. V.; Telkar, M. M.; Jaganathan, R.; Chaudhari, R. V. J.
Mol. Catal. A: Chem. 2003, 200, 279−290.
9) Kirm, I.; Medina, F.; Sueiras, J. E.; Salagre, P.; Cesteros, Y. J.
Mol. Catal. A: Chem. 2007, 261, 98−103.
10) Etschmann, M. M. W.; Bluemke, W.; Sell, D.; Schrader, J. Appl.
Microbiol. Biotechnol. 2002, 59, 1−8.
11) Martínez-Avila, O.; Sanchez, A.; Font, X.; Barrena, R. Appl.
Microbiol. Biotechnol. 2018, 102, 9991−10004.
12) Chreptowicz, K.; Wielechowska, M.; Głow
Rybak, E.; Mierzejewska, J. Food Bioprod. Process. 2016, 100, 275−
81.
13) The European Parliament and the Council of the European
Union. Regulation (EC) No 1334/2008; European Union, 2008.
14) Food and Drug Administration. Code of Federal Regulations,
Title 21:101.22; FDA, 2018.
15) Hazelwood, L. H.; Daran, J.-M. G.; van Maris, A. J. A.; Pronk, J.
T.; Dickinson, J. R. Appl. Environ. Microbiol. 2008, 74, 2259−2266.
̈
hlich, P.; Albert, K.; Bertau, M. Org. Biomol. Chem. 2011, 9,
(
(
(
(
(
́
(
́
czyk-Zubek, J.;
2
(
(
(
̈
(
16) Neubauer, O.; Fromherz, K. Uber Den Abbau Der Amino-
sauren Bei Der Hefegarung. Hoppe-Seyler's Z. Physiol. Chem. 1910, 70,
26−350.
17) Ehrlich, F. Ber. Dtsch. Chem. Ges. 1907, 40, 1027−1047.
̈
̈
3
(
D
DOI: 10.1021/acs.orglett.9b02611
Org. Lett. XXXX, XXX, XXX−XXX