ARTICLE IN PRESS

Metabolic Engineering 11 (2009) 355–366

Contents lists available at ScienceDirect

Metabolic Engineering
journal homepage: www.elsevier.com/locate/ymben

Metabolic engineering of the complete pathway leading to heterologous

biosynthesis of various flavonoids and stilbenoids in Saccharomyces cerevisiae
Emmanouil Trantas a,b, Nickolas Panopoulos b,c, Filippos Ververidis a,
a
Laboratory of Plant Biochemistry & Biotechnology, Department of Plant Sciences, Technological Educational Institute of Crete, P.O. Box 1939, Heraklion GR 71004, Greece
b
Department of Biology, University of Crete, P.O. Box 2208, Heraklion GR 71409, Greece
c
Institute of Molecular Biology and Biotechnology, Foundation for Research and Technology Hellas, P.O. Box 1527, Heraklion GR 71110, Greece

a r t i c l e in fo abstract

Article history: Chemical or biological synthesis of plant secondary metabolites has attracted increasing interest due to
Received 21 February 2009 their proven or assumed beneficial properties and health promoting effects. Resveratrol, a stilbenoid,
Received in revised form naringenin, a flavanone, genistein, an isoflavone, and the flavonols kaempferol and quercetin have been
12 June 2009
shown to possess high nutritional and agricultural value. Four metabolically engineered yeast strains
Accepted 16 July 2009
harboring plasmids with heterologous genes for enzymes involved in the biosynthesis of these
Available online 22 July 2009
compounds from phenylalanine have been constructed. Time course analyses of precursor utilization
Keywords: and end-product accumulation were carried out establishing the production of 0.29–0.31 mg/L of
Flavonoids trans-resveratrol, 8.9–15.6 mg/L of naringenin, 0.1–7.7 mg/L of genistein, 0.9–4.6 mg/L of kaempferol and
Isoflavonoids
0.26–0.38 mg/L of quercetin in defined media under optimal growth conditions. The recombinant yeast
Stilbenoids
strains can be used further for the construction of improved flavonoid- and stilbenoid-overproducers.
Metabolic engineering
Metabolic reconstitution & 2009 Elsevier Inc. All rights reserved.
Plant biosynthesis
Genistein
Kaempferol
Naringenin
Quercetin
Resveratrol

1. Introduction of recorded effects on human health (Manach et al., 2004),

Phenylpropanoid synthesis in plants constitutes a major
anabolic process where the amino acid phenylalanine serves as
a precursor molecule for the biosynthesis of lignins, stilbenoids,
flavonoids and condensed tannins. Among the phenylpropanoid-
derived groups, flavonoids and stilbenoids have diverse functions
in plants, acting as antimicrobial agents in plant defense (Dixon
et al., 1996; Jeandet et al., 1997; Hammerschmidt, 1999),
photoreceptors, visual attractors (Winkel-Shirley, 2001), UV
protectants (Stafford, 1991), feeding repellants (insect and
herbivore protectants), signals in the early steps of rhizobia–le-
gume symbiosis (Dixon et al., 1996), regulators of auxin transport
(Havsteen, 2002) and stimulators of pollen germination
(Mascarenhas, 1993). In addition, these substances have a range
Abbreviatios: PAL, phenylalanine-ammonia lyase; C4H, cinnamic acid
4-hydroxylase; 4CL, 4-coumaric acid: CoA ligase; RS, resveratrol synthase; CPR,
cytochrome P450 reductase; CHS, chalcone synthase; CHI, chalcone isomerase; IFS,
isoflavone synthase; F3H, flavanone 3-hydroxylase; F30 H, flavonoid 30 hydroxylase;
FLS, flavonol synthase
 Corresponding author. Fax: +30 2810 318204.
E-mail address: ververidis@teicrete.gr (F. Ververidis).
involving antioxidant (Pietta, 2000; Williams et al., 2004; Noel
et al., 2005), antiallergic (Nowakowska, 2007), anti-inflammatory,
antithrombotic and antioncogenic activities (Kris-Etherton et al.,
2002; Kampa et al., 2008).
In a recent review (Ververidis et al., 2007a), we have amended
the grouping terminology amongst the distinct biosynthetic
pathways of phenylpropanoid metabolism leading to stilbenoids,
major flavonoids and isoflavonoids synthesis. The diverse physio-
logical properties of members of these groups of metabolites have
stimulated scientists to search and identify new member
compounds belonging to those groups, with over 8000 known
substances on record in year 2000 (Harborne and Williams, 2000;
Pietta, 2000) and growing (NAPRALERT database at http://www.
napralert.org). The structural multiplicity of these compounds
derives from modifications of the basic flavonoid backbone that
give the diversity of flavonoids and flavonoid derivatives and
are carried out by a variety of isoforms of enzymes participating
in the respective metabolic pathways. The enzymes include
members of the cytochrome P450 hydroxylase requiring the
activity of an NADPH-cytochrome P450 reductase (CPR), NADPH-
dependent reductase, 2-oxoglutarate-dependent dioxygenase
(ODDs), O-methyltransferase (OMT), acyl and glycosyltransferase

1096-7176/$ - see front matter & 2009 Elsevier Inc. All rights reserved.
doi:10.1016/j.ymben.2009.07.004
 ARTICLE IN PRESS
356 E. Trantas et al. / Metabolic Engineering 11 (2009) 355–366
(UGT) families, which appear to have been recruited into new
functions during the rapid evolution that accompanied the
domination of plants in moist terrestrial environments (Stafford,
1991; Ververidis et al., 2007a). Furthermore, recent data has
shown that despite such diversity in secondary plant products,
metabolic channeling and metabolon formation enables plants to
effectively synthesize specific natural products and hinder
enzymes from accessing unwanted substrates (Winkel, 2004;
Jorgensen et al., 2005). These evolutionary effects were further
empowered by the ability of several enzymes of the same
metabolic step to catalyze the formation of different intermedi-
ates or end-products using different starting substrates or
precursors. This has been shown from metabolic engineering
studies with several enzymes, such as 4-coumaric acid ligase
(4CL), chalcone isomerase (CHI), isoflavone synthase (IFS) (Liu
et al., 2007) and many others that are able to utilize different
substrates within the same dissected metabolic pathway, con-
verting them into different products, when expressed in microbial
model (Kang and Back, 2009).
Different phenylpropanoid acids delivered to 4CL provide
flavanone-chalcone as well as 5-deoxyflavanone-chalcone to CHI
which converts them to the corresponding flavanone or
5-deoxyflavanone (Allina et al., 1998; Hwang et al., 2003; Yan
et al., 2005a). Likewise, IFS enzymes convert flavanones or
5-deoxyflavanones to the corresponding isoflavones (Steele
et al., 1999). These examples suggest that the cyclization of
chalcones and the aryl migration from C-2 position to C-4 as well
as aromatization, hydroxylation, glycosylation, acylation, prenyla-
tion, sulfation, and methylation are regio-specific reactions (Noel
et al., 2005).
The significance of plant secondary metabolism products is
reflected in the numerous plant-based medicines containing
phenylpropanoid-derived active components that have long been
used by humans. The benefits of specific flavonoids and other
phenylpropanoid-derived compounds to human health and their
potential long-term health profits have been recognized relatively
recently (Ververidis et al., 2007a, 2007b). Phenylalanine is the
connecting metabolite between primary metabolism and a vast
array of secondary metabolites. Initially, phenylalanine is deami-
nated to trans-cinnamic acid by the action of phenylalanine-
ammonia lyase (PAL). Synthesis of p-coumaric acid follows
through the hydroxylation of trans-cinnamic acid at the para-
position of the aromatic ring, by cinnamic acid 4-hydroxylase
(C4H) (Vannelli et al., 2007). The activity of 4CL leads to the
4-coumaroyl-coenzyme A, a nodal compound of phenylpropanoid
metabolism which leads to either stilbenoids or flavonoids
(including anthocyanins and catechins) or lignins. With the action
of resveratrol synthase (RS) and 3 molecules of malonyl-
coenzyme A the cascade enters the pathway of stilbenoids with
trans-resveratrol to be the first but one of the most interesting
compounds of the group. On the alternative metabolic route, the
action of chalcone synthase (CHS) with the aid of 3 molecules of
malonyl-coenzyme A guides the cascade to the general flavonoid
pathway with the production of naringenin-chalcone (referred as
chalcone in many cases). The establishing of the main flavonoid
chassis is created through a cyclization reaction of flavanone-
chalcones (e.g. naringenin-chalcone) to give rise to flavanones by
the action of CHI enzyme. Flavanones are also nodal compounds
because they are precursors for the synthesis, among other, of
isoflavones (e.g. genistein) and of flavonols (e.g. kaempferol,
quercetin) (Ververidis et al., 2007a, 2007b).
In this paper, we present the construction of a series of
metabolically engineered Saccharomyces cerevisiae strains and
discuss the results from the heterologous complete reconstitution
of the various biosynthetic pathways utilizing phenylalanine as
the initial precursor, leading to formation of various pathway end-
products such as the stilbenoid resveratrol, the flavanone
naringenin, the isoflavone genistein and the flavonols kaempferol
and quercetin. Although some of these compounds have been
previously produced in Escherichia coli, we report a different
approach for their synthesis in S. cerevisiae by engineering the
complete pathway for each plant end-product and thus testing the
effect of the number of genes involved as well as the substrate
fluxes at certain enzymatic steps in the metabolon formation in
relation to the precursor molecule used.
2. Materials and methods
2.1. Culture media and microbial growth conditions
Bacterial cell cultures were grown in Luria-Bertani broth
(10 g/L bacto-tryptone, 5 g/L yeast extract, 10 g/L sodium chloride)
with pH adjusted to 7 while yeast cell suspensions were cultured
either in Yeast Peptone Dextrose medium (YPD, consisted of 10 g/L
yeast extract, 20 g/L bacto-peptone, 20 g/L glucose) or in
auxotrophic Complete Minimal medium (CM, consisted of
6.7 g/L yeast nitrogen base without amino acids, 20 g/L glucose,
10 mg/L isoleucine, 150 mg/L valine, 20 mg/L adenine hemisulfate,
20 mg/L arginine-HCl, 30 mg/L lycine-HCl, 20 mg/L methionine,
200 mg/L threonine, 30 mg/L tyrosine, 50 mg/L phenylalanine and
optionally supplemented with leucine, histidine, uracile and
tryptophane at concentrations of 100, 20, 20 and 20 mg/L
respectively) with pH adjusted to 5.8. Bacteria were cultivated
at 37 1C and yeast at 28–30 1C.
2.2. Genetic material, microbial hosts and cloning vectors
Total plant RNA of Glycine max and Solanum tuberosum was
extracted with standard isolation procedures (Ausubel et al., 1994)
using the TRIs Reagent RNA Isolation Reagent (Sigma). Total plant
RNA of Vitis vinifera cv. Sultanina was isolated according to a
specific protocol for isolation of functional RNA from grapevine
tissues (Skopelitis et al., 2006). Total RNA extracted from each
plant source was treated with DNaseI which was then removed
with a phenol–chloroform extraction step. Reverse transcriptase
reaction (MMLV enzyme) for cDNA synthesis was performed using
1 mg of total RNA with 2 mM 18-mer oligo-dT, 0.5 mM dNTPs, 1 
MMLV buffer and 200 units MMLV. One-tenth volume of the cDNA
reaction mixture (cDNA 0.5 mg, 1  buffer GC, 150 mM dNTPs,
0.2 mM forward and reverse primers and 2U Phusions High-
Fidelity DNA Polymerases (Finnzymes)) was used further in the
presence of the appropriate PCR primers (Table 1) to amplify the
appropriate plant gene clone of interest. PCR was programmed to
run the first cycle at 94 1C for 30 s, the next 35 cycles with 3 steps
each starting at 94 1C for 10 s, followed by an appropriate
annealing temperature for each primer for 30 s and 72 1C for
1.5 min, and a 5-min final cycle at 72 1C.
The bacterial strain used for accomplishing the cloning
strategy was the JM83 (F, ara, D(lac-proAB), rpsL (Strr), [F80,
lacZDM15] thi) while the yeast strain was the YPH499 (mat a,
ura3–52, lys2–801amber, ade2–1011chre, trp1–D63, his3–D200,
leu2–D1).
All subsequent cloning and subcloning steps were done using
pGEM T-Easy vector (Promega), pBluescriptII KS+ (Stratagene) and
pT20, a vector that resulted from the subcloning of the EcoRI-
HindIII multicloning site of pT3T7lac into the pSPTBM20 vector
(Boehringer). A series of four epitope-tagging (pESC, Stratagene)
vectors carrying GAL1 and GAL10 yeast promoters, were used for
expression and functional analysis of the plant genes in the S.
cerevisiae strains. These pESC vectors feature an extensive
 ARTICLE IN PRESS
E. Trantas et al. / Metabolic Engineering 11 (2009) 355–366 357

Table 1 pESC-URA-C4H vector creating the vector pESC-URA-C4H-PAL

Oligonucleotides used as primers to amplify by PCR and/or RT-PCR genes involved (Fig. 1A). 4CL and RS, obtained from G. max and V. vinifera cv.
in this study.
Soultanina, respectively, were subcloned into pESC-HIS (Fig. 1B),
Gene/source/ Oligo sequence (50 -30 ) NCBI while subcloning of CPR, obtained from same source as PAL, was
amplicon code inserted into pESC-TRP vector (Fig. 1C). The 4CL gene was cloned
length (bp) from a leaf cDNA pool of G. max by PCR (Table 1) and transferred
first to pGEM T-Easy and then (EcoRI-ClaI fragment) to pESC-HIS,
PAL/Poplar Upstream TGTAATCCATCGATTCTAGAAAAATGG L11747
hybrid Populus Downstream
giving pESC-HIS-4CL. The RS gene was cloned from a leaf cDNA
trichocarpa  P. CTCTAGAAGCTTAGCAAATAGGAAGAGG pool of V. vinifera cv. Soultanina (Table 1) into pGEM T-Easy
deltoides/2181 bp cloning vector. This pESC vector construct was completed with the
subcloning of BamHI-ApaI RS gene fragment of pGEM T-Easy
C4H/Glycine Upstream GTCGACAATGGATCTCCTCCTTCTGGA FJ770468
creating the plasmid pESC-HIS-4CL-RS (Fig. 1B). The ApaI-SalI CPR
max/1530 bp Downstream
CTCGAGTCTAAAATGACCTTGGCTTTGCC gene fragment, derived from another donated plasmid pESC-LEU-
C4H-CPR2 (Ro et al., 2002), was subcloned into pESC-TRP at the
4CL/Glycine max/ Upstream ATGATAACTCTAGCTCCTTCTCTTGATAC FJ770469 respective sites creating the pESC-TRP-CPR vector (Fig. 1C).
1697 bp Downstream The IFS gene was PCR cloned (Table 1) into pT20 at EcoRV site,
GCTATCGATTAAGGCGTCTGAGTGGCGGCGG
giving the pT20-IFS vector. The BamHI fragment of this vector
IFS/Glycine max/ Upstream GGATCCACGATGTTGCTGGAACTTGCAC FJ770473 comprising the IFS gene was subcloned into pESC-HIS-4CL giving
1565 bp Downstream GATGGCAAGACACTACTATTG the pESC-HIS-4CL-IFS plasmid (Fig. 1D). A CHS leaf cDNA of G. max
was cloned into pGEM T-Easy (Table 1) and then (NotI fragment)
CHS/Glycine max/ Upstream ATGGTCAGTGTTGAAGAGATCC FJ770471 into pESC-LEU giving the pESC-LEU-CHS vector. A CHI cDNA from
1165 bp Downstream TTAGACAGTGACACTGCGGAG
the same cDNA pool of G. max was PCR cloned into pT20 at EcoRV
CHI/Glycine max/ Upstream ATGGCATTTCCGTCCGTAACC FJ770472 site (Table 1). The BamHI-HindIII CHI gene fragment was then
680 bp Downstream TCACTTATGAGATTGAGGGTCAC subcloned into the pESC-LEU-CHS vector giving plasmid pESC-
LEU-CHS-CHI (Fig. 1E).
F3H/Glycine max/ Upstream CAACATATGGCACCAACAGC FJ770474 The FLS gene was derived from a leaf cDNA pool of S. tuberosum
1128 bp Downstream TTAAGCAAGAATCTCCTTCAAAGG
and was PCR cloned into pGEM T-Easy (Table 1) and the NotI
F30 H/Glycine Upstream TGTAATCCATCGATTCTAGAAAAATGG FJ770476 fragment comprising the FLS open reading frame was subcloned
max/1540 bp Downstream GTCGACAATTATAAGGAAG into pT20 at the respective site. Subsequently, the BamHI-ApaI
fragment carrying the FLS gene was subcloned into the respective
FLS/Solanum Upstream ATGAAAACAATTCAAGGTCAG FJ770475 restriction sites of pESC-HIS-4CL creating the pESC-HIS-4CL-FLS
tuberosum/ Downstream CCCATAATTCTTCACTGAGG
vector (Fig. 1F). The F3H gene was derived from leaf cDNA pool of
1049 bp
G. max and was PCR cloned (Table 1) into pGEM T-Easy. The NotI
CPR/Poplar Upstream AF302497 fragment containing the open reading frame was then subcloned
hybrid Populus TACAGCGGGCCCAACATGCAATCATCAAGCAGCTCG into the pESC-TRP vector creating the pESC-TRP-F3H plasmid. The
trichocarpa  P. Downstream SalI-ApaI fragment of the vector pESC-LEU-C4H-CPR2 (Ro et al.,
deltoides/2140 bp TACAGCGTCGACCCATACACGCAGATACCTGC
2002), carrying the CPR gene, was subcloned into the pESC-TRP-
F3H plasmid giving the pESC-TRP-F3H-CPR vector (Fig. 1G).
The F30 H gene was PCR cloned from a leaf cDNA pool of G. max
at the EcoRV site of pT20 (Table 1). A fragment containing the
polylinker sequence while each one contains one of the four open reading frame of F30 H was generated by digestion with SalI
different yeast-selectable markers (HIS3, TRP1, LEU2, or URA3) in and was subcloned at the respective site of pESC-TRP-F3H vector
the same vector backbone, which allows expression analysis of (described earlier), generating the pESC-TRP-F3H-F30 H plasmid
two different co-expressed genes per vector in a single yeast cell. (Fig. 1H).

2.3. Gene cloning and construction of engineered yeast plasmids
with plant genes
All engineered yeast plasmids constructed in this study were
based on the various available pESC vectors and the construction
strategy and steps for each engineered plasmid used in this study,
are listed in Fig. 1. All plant genes used were cloned by PCR from
various sources. To confirm each gene’s identity the cloned PCR
fragments were initially sequenced and compared in silico with
the NCBI deposited sequence of this gene (Table 1).
PAL gene that was obtained from a Populus hybrid (Populus
trichocarpa  P. deltoids, gift from Prof. Douglas, UBC, Canada)
and C4H, obtained from G. max, were subcloned into pESC-URA
(Fig. 1A). The C4H gene (cDNA) was initially cloned by PCR (see
primer sequences in Table 1) from leaf cDNA pool of G. max at the
EcoRV restriction site of pT20. The SalI fragment of pT20 was
subcloned into pESC-URA thus creating the pESC-URA-C4H vector.
Similarly, PAL was subcloned by PCR (Table 1) from the donated
plasmid pESC-HIS-PAL2 (Ro and Douglas, 2004) to the pGEM
T-Easy vector and then subcloned as a NotI fragment into the
2.4. Flavonoid extraction and analysis method
The isolation of the newly synthesized compounds from
genetically modified yeast strains was accomplished by double
liquid–liquid extraction of 1 ml of yeast culture suspension with
0.7 ml of ethyl acetate. The resulting 1.4 ml ethyl acetate fraction
was lyophilized in a vacuum controlled centrifugal lyophilizer and
the pellet was dissolved in 60 ml of 70% ethanol.
Qualitative and quantitative analysis of samples for stilbenoids
and flavonoids were carried out in an Agilent Capillary Electro-
phoresis system coupled with a diode array detector (CE-DAD). A
bare fused silica capillary column was used with extended light
path, effective length of 72 cm and inner diameter of 50 mm. The
analysis was carried out in three steps: preconditioning, injection
and analysis. The preconditioning step involved an initial flush
with 1 N sodium hydroxide for 2 min followed by a flush with
analysis buffer (25 mM tetrasodium borate decahydrate, pH 9.2)
for 4 min. The injection step comprised a 50 mbar pressure for 4 s
and the analysis was performed at 29 kV for 16 min, time
sufficient for maximum resolution of the compounds analyzed
 ARTICLE IN PRESS
358 E. Trantas et al. / Metabolic Engineering 11 (2009) 355–366

Fig. 1. Construction strategy of all engineered pESC vectors (gray shadowed area) with plant genes (colored arrows). Steps for the construction of: (A) pESC-URA-PAL-C4H,
(B) of pESC-HIS-4CL-RS, (C) pESC-TRP-CPR, (D) pESC-HIS-4CL-IFS, (E) pESC-LEU-CHS-CHI, (F) pESC-HIS-4CL-FLS, (G) pESC-TRP- F3H-CPR, and (H) pESC-TRP- F3H-F30 H. (For
interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

in this study. The temperature of the column cassette was kept
constantly at 25 1C.
Initial poor reproducibility of the peaks resulted from electro-
lytic phenomena occurring in the running buffer and solvent
evaporation. Both phenomena were diminished by adding two
extra steps between the preconditioning and the injection steps.
The first intermediate step involved buffer replenishment of inlet
and outlet vials followed by a ‘‘buffer customization’’ step using
20 kV voltage for 2 min before each analysis. The sample tray was
kept constantly at 6 1C to prevent sample evaporation.
Flavonoids and stilbenoids generated from S. cerevisiae
fermentations were identified by matching the retention time,
UV-absorbance spectrum, and co-chromatography with authentic
chemicals. trans-Resveratrol was quantified by absorbance at
290 nm while the isoflavone genistein and the flavonols kaemp-
ferol and quercetin were quantified at 270 nm. The precursor
phenylalanine was quantified at 210 nm, p-coumaric acid at
290 nm and naringenin at 320 nm. Calibration curves were
obtained with authentic flavonoid/stilbenoid solutions of various
concentrations. The retention times under the CE-DAD conditions
were: trans-resveratrol 8 min, genistein 10 min, naringenin
10.3 min, kaempferol 12.7 min, quercetin 14.3 min, phenylalanine
7.8 min.
2.5. Yeast total RNA extraction
Yeast cells grown to OD60041 in 50 ml of yeast media were
collected by centrifugation at 1500g for 3 min. The supernatant
was discarded and cells were washed with half volume of sterile
distilled water (SDW). The washing step was repeated with 1 ml
SDW. The final cell pellet was resuspended in 500 ml of lysis buffer
(10 mM Tris–HCl pH 7.5, 10 mM EDTA, 0.5% SDS). Equal volume of
water-saturated phenol was added and samples were vortexed for
1.5 min at maximum speed. Samples were then kept at 65 1C for
2 h with periodic vortexing and then placed on ice for 10 min
before adding 1/10th of sample volume of 3 M sodium acetate pH
5.2. Insoluble materials were discarded after centrifugation for
10 min at maximum speed and 4 1C. The supernatant was washed
three times with an equal volume chloroform/water-saturated
phenol. Soluble nucleic acids were pelleted with 2.5 volumes of
ethanol or an equal volume of isopropanol by a 30-min
 ARTICLE IN PRESS
E. Trantas et al. / Metabolic Engineering 11 (2009) 355–366 359

centrifugation at 15,000g. Finally, nucleic acids were dissolved in
50–100 ml of SDW or TE at the desired concentration.
2.6. Gene expression analysis
Expression analysis of genetically transformed yeast strains
was performed by semi-quantitative PCR. Total RNA extracted
from each strain was treated with DNase I, which was then
removed with a phenol–chloroform extraction step. Reverse
transcriptase reaction (MMLV enzyme) for cDNA synthesis was
performed using 1 mg of total RNA with 2 mM 18-mer oligo-dT,
0.5 mM dNTPs, 1  MMLV buffer and 200 units MMLV. One-tenth
volume of the cDNA reaction mixture (cDNA 0.5 mg, 1  reaction
buffer, 1.5 mM MgCl2, 150 mM dNTPs, 0.2 mM forward and reverse
primers and 1 U Taq polymerase (HyTest Ltd)) was used further in
the presence of the appropriate PCR primers (Table 1) to detect
plant gene transcripts in the transformed yeast strain. PCR was
programmed to run the first cycle at 94 1C for 30 s, the next 35
cycles with three steps each starting at 94 1C for 45 s, followed by
an appropriate annealing temperature for each primer for 45 s and
72 1C for 1.5 min, and a 5-min final cycle at 72 1C.
3. Results
3.1. Factors affecting flavonoid/stilbenoid synthesis and pathway
fluxes
made it critical to examine the need for increased initial
concentration of phenylalanine in order to determine the best
regime for optimal growth and highest flavonoid production.
To optimize this parameter we used the yeast strain COUM11
(Table 2) that was capable of producing p-coumaric acid,
representing an early biosynthetic intermediate (Fig. 2). Meta-
bolic engineering of yeast to produce p-coumaric acid required the
insertion of three plant genes leading to the construction of
COUM11 (Fig. 2 and Table 2). The vectors pESC-URA-C4H-PAL and
pESC-TRP-CPR, carrying the necessary plant genes were used to
transform the YPH499 yeast strain. The presence and the
transcriptional activity of the three plant genes (PAL, C4H and
CPR) were demonstrated by semi-quantitative PCR (data not
shown). The COUM11 strain was grown in suitable CM media that
were supplemented with various concentrations of phenylalanine
as precursor. In this way, we estimated the maximal production in
the yeast strain of p-coumaric acid converted from phenylalanine.
Results in Fig. 3, demonstrate that p-coumaric acid production
from strain COUM11 grown in different phenylalanine
concentrations can be increased substantially by increasing the
initial concentration of phenylalanine from 0.3 to 20 mM.
However, even though this resulted in a substantial increase in
the production rate of p-coumaric acid, it also showed that 10 mM
of phenylalanine is an optimal concentration for efficient
substrate fluxes in flavonoid production, as further increase in
phenylalanine did not result in higher p-coumaric acid production
(Fig. 3).

3.1.2. P450 enzyme requirements for CPR

3.1.1. Competition for phenylalanine
The reconstituted part of the phenylpropanoid pathway in
In the basal CM medium (see Section 2) that is used for the
yeast appeared to need additional adjustments and gene require-
expression of foreign genes in yeast, phenylalanine is normally
ments. Metabolon formation of this pathway in plants includes a
supplemented with a standard concentration of 0.3 mM plus other
special group of cytochromes P450 enzymes that require the
amino acids (Ausubel et al., 1994). However, in our metabolic
coexpression of a cytochrome P450 reductase (CPR, Fig. 2) for
engineering strategy, phenylalanine played a dual role as it was
the other enzymes to efficiently carry out their functions
also utilized by the engineered yeast strains as an initial precursor,
(Werck-Reichhart and Feyereisen, 2000; Winkel, 2004; Ayabe
feeding the entire heterologous plant flavonoid/stilbenoid path-
and Akashi, 2006).
way. In essence, phenylalanine was a linking molecule between
Although S. cerevisiae carries a chromosomal CPR gene, we
primary metabolism for cellular growth and as an initial precursor
checked the productivity of C4H from G. max with (strain COUM11,
molecule for flavonoid/stilbenoid biosynthesis. Although yeasts
Table 2) and without (COUM12, same genotype as COUM11 strain
are capable of producing phenylalanine (Braus, 1991), the fact that
but lacking the pESC-TRP-CPR plasmid construct) a heterologous
in this work yeast is bioconverting the amino acid to flavonoids
CPR gene derived from a Populus hybrid (Ro et al., 2002). When
this CPR was co-expressed, the production of p-coumaric acid was
Table 2
enhanced 4-fold (data not shown) indicating that the efficient
Metabolically engineered yeast stains constructed in this work (derived from
parental strain YPH499) to produce heterologously plant-like secondary metabo- metabolon formation required the support of a plant CPR.
lites of phenylpropanoid pathway. Evidently, the activity of endogenous yeast CPR was not sufficient
to support maximal substrate fluxes from phenylalanine to
Metabolically pESC plasmids contained Plant metabolite p-coumaric acid through 4-hydroxylase activity of the cloned C4H.
engineered yeast (corresponding figure) produced (genes
strains involved)
3.2. Construction of trans-resveratrol producing yeast strain
COUM11 pESC-URA-PAL-C4H, pESC-TRP-CPR p-Coumaric acid (PAL,
(Fig. 1A, C) C4H, CPR)
COUM12 pESC-URA-PAL-C4H (Fig. 1A) p-Coumaric acid (PAL, Metabolic engineering of yeast to produce trans-resveratrol
C4H) required insertion of five plant genes leading to the construction
RESV11 pESC-URA-PAL-C4H, pESC-HIS-4CL- trans-Resveratrol (PAL, of strain RESV11 (Fig. 2 and Table 2). The vectors pESC-URA-C4H-
RS, pESC-TRP-CPR (Fig. 1A–C) C4H, CPR, 4CL, RS)
NAR12 pESC-URA-PAL-C4H, pESC-HIS-4CL, Naringenin (PAL, C4H,
PAL, pESC-HIS-4CL-RS and pESC-TRP-CPR, carrying the necessary
pESC-LEU-CHS-CHI, pESC-TRP-CPR CPR, 4CL, CHS, CHI) plant genes were used to transform the YPH499 yeast strain. A
(Fig. 1A, C, D, E) positive (prototrophic) clone, designated RESV11 (Table 2) was
GEN23 pESC-URA-PAL-C4H, pESC-HIS-4CL- Genistein (PAL, C4H, confirmed to carry the PAL, C4H, 4CL, RS and CPR genes in a
IFS, pESC-LEU-CHS-CHI, pESC-TRP- CPR, 4CL, CHS, CHI, IFS)
transcriptionally active state, by semi-quantitative PCR (not
CPR (Fig. 1A, C, D, E)
KAE34 pESC-URA-PAL-C4H, pESC-HIS-4CL- Kaempferol (PAL, C4H, shown). This strain produced trans-resveratrol in CM growth
FLS, pESC-LEU-CHS-CHI, pESC-TRP- CPR, 4CL, CHS, CHI, F3H, medium supplemented with either p-coumaric acid or phenyla-
F3H-CPR (Fig. 1A, E, F, G) FLS) lanine, though with different efficiencies.
QUE44 pESC-URA-PAL-C4H, pESC-HIS-4CL- Quercetin (PAL, C4H, When the suspension culture was supplemented with 1 mM
FLS, pESC-LEU-CHS-CHI, pESC-TRP- 4CL, CHS, CHI, F3H, F30 H,
F3H-F30 H (Fig. 1A, E, F, H) FLS)
p-coumaric acid, trans-resveratrol was produced at a level of
0.31 mg/L (1.36 mM, Fig. 4) after approx. 75 h of cultivation. When
 ARTICLE IN PRESS
360 E. Trantas et al. / Metabolic Engineering 11 (2009) 355–366

Fig. 2. Schematic representations of all the engineered plant biosynthetic pathways, shown as a whole, transferred to various S. cerevisiae strains leading in each strain to an
end-product synthesis such as p-coumaric acid (strains COUM11 and COUM12), trans-resveratrol (strain RESV11), naringenin (strain NAR12), genistein (strain GEN23),
kaempferol (strain KAE34) and quercetin (strain QUE44), shown in frame-folders.

Fig. 4. Heterologous production of resveratrol (K), naringenin (%), genistein (E),

Fig. 3. Dependence of strain p-coumaric acid production by COUM11 in, relation to kaempferol (m) and quercetin ($) from metabolically engineered S. cerevisiae
the initial phenylalanine concentration used. strains RESV11, NAR12, GEN23, KAE34 and QUE44 respectively, grown in CM
medium, fed with 1 mM p-coumaric acid as precursor.

the CM culture medium was supplemented with 10 mM of
phenylalanine, the strain produced trans-resveratrol at a final
concentration of 0.29 mg/L after approximately 120 h of
cultivation (1.27 mM, Fig. 5). The essentially identical levels of
trans-resveratrol produced in culture, when supplemented with
phenylalanine or p-coumaric acid as the precursor molecule may be
explained by the fact that phenylalanine can be easily converted to
p-coumaric acid, ‘‘saturating’’ the 4CL enzyme to the same extend as
when p-coumaric acid is supplied externally in the medium.
Furthermore it was observed that the final concentration of trans-
resveratrol in the media is declining most probably because of the
labile nature of the compound, a phenomenon that has been
observed previously (Beekwilder et al., 2006).
and Table 2) was obtained after transformation of strain YPH499
with the plasmids pESC-URA-PAL-C4H, pESC-HIS-4CL, pESC-LEU-
CHS-CHI, pESC-TRP-CPR, collectively carrying the six genes
required for naringenin biosynthesis. Positive transformants were
confirmed to carry and express all six genes (PAL, C4H, CPR, 4CL,
CHS and CHI) by semi-quantitative PCR (not shown).
In the same manner as we followed for trans-resveratrol and the
other flavonoids, we analyzed the production of flavanone nar-
ingenin. When the added precursor was 1 mM p-coumaric acid,
supplied periodically (every approx. 40 h), naringenin was maxi-
mally produced after over 100 h at approximately 15.6 mg/L (57 mM,
Fig. 4). When phenylalanine served as the precursor molecule at a
concentration of 10 mM, it was converted to naringenin at the level
of 8.9 mg/L (33 mM, Fig. 5) after about 150 h.

3.3. Construction of naringenin producing yeast strain 3.4. Construction of genistein producing yeast strain

Biosynthetic engineering of yeast to produce naringenin Engineering yeast to produce genistein required the transfer of
required the introduction of six plant genes. Strain NAR12 (Fig. 2 seven plant genes. Strain GEN23 (Fig. 2 and Table 2) was obtained
 ARTICLE IN PRESS
E. Trantas et al. / Metabolic Engineering 11 (2009) 355–366
Fig. 5. Heterologous production of resveratrol (K), naringenin (%), genistein (E),
kaempferol (m) and quercetin ($) from metabolically engineered S. cerevisiae
strains RESV11, NAR12, GEN23, KAE34 and QUE44 respectively, grown in CM
medium and fed with 10 mM phenylalanine as precursor.
Fig. 6. Heterologous production of genistein (E), kaempferol (m) and quercetin
($) from metabolically engineered S. cerevisiae strains GEN23, KAE34 and QUE44,
respectively, grown in CM medium, fed with 0.5 mM naringenin as precursor.
after transformation of the yeast strain YPH499 with the plasmids
pESC-URA-PAL-C4H, pESC-TRP-CPR, pESC-LEU-CHS-CHI and pESC-
HIS-4CL-IFS, which collectively carry the plant genes involved. The
presence and transcriptional activity of the seven genes (PAL, C4H,
CPR, CHS, CHI, 4CL and IFS) was demonstrated by semi-quantita-
tive PCR (not shown).
The GEN23 strain was grown in suitable CM media that were
supplemented appropriately with either phenylalanine, p-couma-
ric acid or naringenin as starting precursor compound. In the
presence of 0.5 mM naringenin the strain produced 7.7 mg/L of
genistein after approx. 180 h (28.5 mM, Fig. 6). When 1 mM
p-coumaric was supplied periodically (every 2 days), 0.14 mg/L
of genistein was synthesized after about 90 h (0.52 mM, Fig. 4).
Similarly, phenylalanine supplementation at a 10 mM concen-
tration resulted in the production of 0.1 mg/L of genistein at about
90 h of cultivation (0.4 mM, Fig. 5).
3.5. Construction of kaempferol producing yeast strain
Biosynthetic engineering of yeast to produce kaempferol
required the introduction of eight plant genes. Strain KAE34
361
(Fig. 2 and Table 2) was obtained after transformation of strain
YPH499 with the plasmids pESC-URA-PAL-C4H, pESC-HIS-4CL-
FLS, pESC-LEU-CHS-CHI, pESC-TRP-F3H-CPR, collectively carrying
the eight genes required for kaempferol biosynthesis. Positive
transformants were confirmed to carry and express all eight genes
by semi-quantitative PCR (not shown).
In the same manner as we followed for trans-resveratrol,
naringenin and genistein, we analyzed the production of flavonol
kaempferol. When the added precursor was 0.5 mM naringenin,
kaempferol was produced at about 4.6 mg/L after about 70 h
(16 mM, Fig. 6). However, when p-coumaric acid was supplied
periodically (every 2 days) as a precursor, kaempferol was
produced after 100 h at approximately 0.9 mg/L (3.1 mM, Fig. 4).
When phenylalanine served as the precursor molecule at a
concentration of 10 mM, it was converted to kaempferol at
1.3 mg/L after about 160 h (5 mM, Fig. 5).
3.6. Construction of quercetin producing yeast strain
The quercetin producing yeast strain was obtained after
transformation of strain YPH499 with plasmids pESC-URA-PAL-
C4H, pESC-HIS-4CL-FLS, pESC-LEU-CHS-CHI and pESC-TRP-F3H-
F30 H (Fig. 2 and Table 2). A positive clone, QUE44, was confirmed
by semi-quantitative PCR to carry in transcriptionally active form
all eight genes necessary for quercetin biosynthesis (PAL, C4H, 4CL,
FLS, CHS, CHI, F3H and F30 H) (not shown).
The strain was checked for its ability to produce quercetin by
growing in CM media supplemented with naringenin, p-coumaric
acid or phenylalanine. When 0.5 mM naringenin was added, the
levels of quercetin reached 0.38 mg/L after about 70 h (1 mM,
Fig. 6), while when 1 mM of p-coumaric acid was added each
second day of growth, quercetin levels reached 0.26 mg/L after
about 150 h (0.9 mM, Fig. 4). With phenylalanine supplementation
(10 mM) only trace amounts of quercetin were produced (Fig. 5)
too low to be quantified with the detection hardware used
(CE-DAD). In all cases kaempferol was also detected in consider-
able amounts (1.47, 1.29, 20.6 mM when the culture was supplied
with 10 mM phenylalanine, 1 mM p-coumaric acid added
periodically and 0.5 mM naringenin, respectively), indicating the
regio-specificity of FLS in dihydroflavonol biosynthesis (Noel et al.,
2005; Leonard et al., 2006).
4. Discussion
The exceptional properties of flavonoids in relation to human
and plant health, together with their relatively low abundance in
many foods, have stimulated interest in the heterologous
biosynthesis of such compounds in microorganisms. In earlier
work, expression of single genes of the phenylpropanoid pathway
was achieved in E. coli, leading to heterologous production of a
semantic stilbenoid and/or chalcone (Schulz et al., 1989; Melchior
and Kindl, 1990). Since then, a plethora of studies reporting the
production of various subgroups of stilbenoids and flavonoids
have been published as reviewed recently (Ververidis et al., 2007a,
2007b).
S. cerevisiae as a eukaryotic organism has transcriptional and
translational mechanisms similar in basic respects to those of
plants and this would make yeast a suitable single-celled
organism for the production of secondary metabolites through
the heterologous expression of plant genes. However, this
expectation has never been confirmed, as in many cases the
production of flavonoids in bacteria is more efficient than in
eukaryotic organisms. A detailed update of all the progress
concerning the level of phenylpropanoid pathway products
biosynthesized in model-microbe heterologous systems (E. coli,
 ARTICLE IN PRESS
362 E. Trantas et al. / Metabolic Engineering 11 (2009) 355–366

Table 3
Reports showing categorized experimental details of the heterologous production of flavonoids and stilbenoids (as end-products) in E. coli and S. cerevisiae.

End- Precursor No. of Gene sources Host organism Level of Reference

product molecule genes produc-
tion (mg/L)

Resveratrol p-Coumaric acid 2 Nicotiana tabacum (4CL), Vitis vinifera (RS) E. coli 16 Beekwilder et al. (2006)
p-Coumaric acid 2 Arabidopsis thaliana (4CL), Arachis hypogaea E. coli 104.5 Watts et al. (2006)
(RS)
p-Coumaric acid 3 Populus trichocarpa  P. deltoides (4CL), Vitis S. cerevisiae 1.45  103 Becker et al. (2003)
vinifera (RS)
p-Coumaric acid 2 Nicotiana tabacum (4CL), Vitis vinifera (RS) S. cerevisiae 6 Beekwilder et al. (2006)
p-Coumaric acid 2 A. thaliana (4CL), V. vinifera (RS) S. cerevisiae 5.25 Zhang et al. (2006)
Phenylalanine 5 Populus trichocarpa  P. deltoides (PAL, CPR), S. cerevisiae (strain 0.29 This work
Glycine max (C4H, 4CL), Vitis vinifera (RS) RESV11)
p-Coumaric acid 5 Populus trichocarpa  P. deltoides (PAL, CPR), S. cerevisiae (strain 0.31 This work
Glycine max (C4H, 4CL), Vitis vinifera (RS) RESV11)

Naringenin tyrosine 3 Rhodotorula rubra (PAL), Streptomyces E. coli 0.45 Hwang et al. (2003)
coelicolor (4CL), Glycyrrhiza echinata (CHS)
Without any 2 Rhodobacter Sphaeroides (TAL), Arabidopsis E. coli 20.8 Watts et al. (2004)
precursor molecule thaliana (CHS)
Tyrosine 5 R. rubra (PAL), S. coelicolor (4CL), Glycyrrhiza E. coli 57 Miyahisa et al. (2005)
echinata (CHS), Pueraria lobata (CHI),
Corynebacterium glutamicum (ACC)
p-Coumaric acid 5 P. crispum (4CL), Petunia hybrid (CHS, CHI), E. coli 119 Leonard et al. (2007)
Photorhabdus luminescens (ACC), E. coli
(ACS)
p-Coumaric acid 5 P. crispum (4CL), Petunia hybrid (CHS, CHI), E. coli 155 Leonard et al. (2008)
Rhizobium trifolii (MATB, MATC)
Phenylalanine 3 Rhodosporidium toruloides (PAL), A. thaliana S. cerevisiae 7 Jiang et al. (2005)
(4CL), Hypericum androsaemum (CHS)
Phenylalanine 6 P. trichocarpa  P. deltoides (PAL, CPR), G. S. cerevisiae 8.9 This work
max (C4H, 4CL, CHS, CHI)
p-Coumaric acid 6 P. trichocarpa  P. deltoides (PAL, CPR), G. S. cerevisiae 15.6 This work
max (C4H, 4CL, CHS, CHI)
p-Coumaric acid 4 Arabidopsis thaliana (C4H), Petroselinum S. cerevisiae 28.3 Yan et al. (2005b)
crispum (4CL), Petunia hybrid (CHS, CHI)

Genistein N- 3 Glycyrrhiza echinata (CHS, IFS), Pueraria E. coli 0.34 Katsuyama et al. (2007)
acetylcysteamine- lobata (CHI)
attached p-
coumaric acid
Tyrosine 6 Rhodotorula rubra (PAL), Streptomyces E. coli and S. 6 Katsuyama et al. (2007)
coelicolor (4CL), Glycyrrhiza echinata (CHS, cerevisiae co-
IFS), Pueraria lobata (CHI), Corynebacterium incubation
glutamicum (ACC)
Naringenin 2 Trifolium pratense (IFS), Oryza sativa (CPR) S. cerevisiae 20.8 Kim et al. (2005)
Phenylalanine 7 P. trichocarpa  P. deltoides (PAL, CPR), G. S. cerevisiae (strain 0.1 This work
max (C4H, 4CL, CHS, CHI, IFS) GEN23)
p-Coumaric acid 7 P. trichocarpa  P. deltoides (PAL, CPR), G. S. cerevisiae (strain 0.14 This work
max (C4H, 4CL, CHS, CHI, IFS) GEN23)
Naringenin 7 P. trichocarpa  P. deltoides (PAL, CPR), G. S. cerevisiae (strain 7.7 This work
max (C4H, 4CL, CHS, CHI, IFS) GEN23)

Kaempferol Tyrosine 8 Rhodotorula rubra (PAL), Streptomyces E. coli 15.1 Miyahisa et al. (2006)
coelicolor (4CL), Citrus cinensis (F3H, FLS),
Glycyrrhiza echinata (CHS), Pueraria lobata
(CHI), Corynebacterium glutamicum (ACC)
p-Coumaric acid 6 Petroselinum crispum (4CL), Petunia hybrid E. coli 0.3 Leonard et al. (2006)
(CHS, CHI), Malus domestica (F3H), A.
thaliana (FLS)
Naringenin 6 Petroselinum crispum (4CL), Petunia hybrid E. coli 0.8 Leonard et al. (2006)
(CHS, CHI), Malus domestica (F3H), A.
thaliana (FLS)
Phenylalanine 8 P. trichocarpa  P. deltoides (PAL, CPR), G. S. cerevisiae (strain 1.3 This work
max (C4H, 4CL, CHS, CHI, F3H), Solanum KAE34)
tuberosum (FLS)
p-Coumaric acid 8 P. trichocarpa  P. deltoides (PAL, CPR), G. S. cerevisiae (strain 0.9 This work
max (C4H, 4CL, CHS, CHI, F3H), S. tuberosum KAE34)
(FLS)
Naringenin 8 P. trichocarpa  P. deltoides (PAL, CPR), G. S. cerevisiae (strain 4.6 This work
max (C4H, 4CL, CHS, CHI, F3H), S. tuberosum KAE34)
(FLS)

Quercetin p-Coumaric acid 7 Malus domestica (F3H), A. thaliana (FLS), E. coli 0.05 Leonard et al. (2006)
Petunia hybrid (CHS, CHI), Petroselinum
crispum (4CL), Catharanthus, Roseus (F30 50 ,
CPR)
Naringenin 7 Malus domestica (F3H), A. thaliana (FLS), E. coli 0.18 Leonard et al. (2006)
Petunia hybrid (CHS, CHI), Petroselinum
 ARTICLE IN PRESS
E. Trantas et al. / Metabolic Engineering 11 (2009) 355–366 363

Table 3 (continued )

End- Precursor No. of Gene sources Host organism Level of Reference

product molecule genes produc-
tion (mg/L)

crispum (4CL), Catharanthus, Roseus (F30 50 ,

CPR)
Phenylalanine 8 P. trichocarpa  P. deltoides (PAL), G. max S. cerevisiae (strain traces This work
(C4H, 4CL, CHS, CHI, F3H, F30 H), S. tuberosum QUE44)
(FLS)
p-Coumaric acid 8 P. trichocarpa  P. deltoides (PAL), G. max S. cerevisiae (strain 0.26 This work
(C4H, 4CL, CHS, CHI, F3H, F30 H), S. tuberosum QUE44)
(FLS)
Naringenin 8 P. trichocarpa  P. deltoides (PAL), G. max S. cerevisiae (strain 0.38 This work
(C4H, 4CL, CHS, CHI, F3H, F30 H), S. tuberosum QUE44)
(FLS)

S. cerevisiae), including the findings of this work, is presented in

Table 3. However, although Beekwilder et al. (2006) (Table 3)
compared the production levels of trans-resveratrol from yeast
and bacteria and found that the two organisms are comparable (6
and 16 mg/L, respectively), other researchers managed to produce
quantities higher than 100 mg/L in E. coli (Watts et al., 2006) while
the highest production published in yeast is about 6 mg/L
(Beekwilder et al., 2006) (Table 3). In the case of flavonols, using
E. coli as the production system, kaempferol was produced up to
15.1 mg/L (Miyahisa et al., 2006) when fed with tyrosine and
quercetin was produced up to 0.05 mg/L when the cells were fed
with p-coumaric acid (Leonard et al., 2006). In the case of
genistein, there are no full reports for the reconstitution of the
metabolic pathway for its biosynthesis, rather only reports of
functional activity of IFS (Akashi et al., 1999) and of a fusion of IFS
with CHI (Tian and Dixon, 2006).
We have successfully expressed all the biosynthetic genes and
engineered the complete pathway leading to the synthesis of a
stilbenoid and various flavonoids in S. cerevisiae for the first time,
Fig. 7. Flavonoid and stilbenoid end-product levels heterologously synthesized
starting from phenylalanine as well as additionally utilizing from each genetically engineered yeast strain (shown in parentheses).
alternate starting precursor molecule (Fig. 7). All the modified
yeast strains were tested for the presence of heterologous plant
genes through the semi-quantitative RT-PCR. The genes that were not crucial for the activity of the enzymes, as in silico protein
included in the strategy followed in this work for the production analysis showed that none was located in the active center of
of flavonoids and other phenylpropanoid pathway products were these enzymes.
cloned, sequenced and compared with the corresponding genes However, final yields of some end-products were lower than
deposited in NCBI database. In silico translation of the cloned expected particularly when phenylalanine was the starting
sequences showed some differences in the sequence of the coded biosynthetic precursor, probably because the fluxes of substrates
proteins for some of the genes. The proteins coded by PAL, C4H, did not reach the metabolon at constant saturating rates to ensure
CHI, IFS, F3H, F30 H showed no differences with deposited maximal activity of the enzyme involved in the sequential
sequences (comparisons were based on protein sequences avail- biosynthetic steps. This is more obvious for the end-products
able as GenBank formats in National Center for Biotechno- that needed more than three biosynthetic steps from the point of
logy Information (NCBI); all GenBank/NCBI codes are shown in their supplied precursor. For instance, strain RESV11 (Table 3)
Table 1), while the rest showed the following differences: produced after 120 h (on average) of incubation 0.29 mg/L
4CL[123(Asn-Ser), 140(Ala-Glu), 143(Ile-Val), 231(Gln-Lys), (1.27 mM) trans-resveratrol when fed with phenylalanine (four
232(Val-Ile)], CHS:[258(Gly-Arg)] and FLS:[254(+Gly-Leu-Gln), enzymatic steps involved) and 0.31 mg/L (1.36 mM) trans-resver-
321(Asn-Lys)]. atrol when fed with p-coumaric acid as a starting precursor (2
We were able to detect and quantify phenylpropanoid pathway biosynthetic steps, Table 3). This is still much lower production
intermediates and end-products (p-coumaric acid, trans-resvera- than previously reported by other groups (Table 3).
trol, naringenin, genistein, kaempferol, quercetin) in the culture Similarly, when strain QUE44 was fed with phenylalanine, no
media of the appropriate engineered yeast strains. However, no quercetin could be detected (eight enzymes involved, Fig. 2).
traces of the same compounds could be found in the wild type However, when p-coumarate was the fed precursor, quercetin
(wt) yeast that contained no plasmid or exogenous transgenes, biosynthesis occurred (six enzymes involved, Table 3) at 0.26 mg/L
showing that these wt strains could not metabolize the exogen- (0.9 mM). Moreover, when QUE44 was fed with naringenin,
ously added precursors to the targeted compounds (data not quercetin synthesis reached 0.38 mg/L (1.3 mM, Table 3, three
shown). This indicates that the compounds produced by cell enzymatic steps involved, Fig. 2). These levels of quercetin bio-
suspensions of the metabolically engineered strains result from synthesis are still higher than those previously reported (Table 3)
the activity of the newly introduced enzymes, implying that the using E. coli (Leonard et al., 2006). However, it should also be
deviations in the primary protein sequences mentioned above are noted that QUE44 strain utilized the endogenous CPR yeast gene
 ARTICLE IN PRESS
364 E. Trantas et al. / Metabolic Engineering 11 (2009) 355–366
for quercetin biosynthesis instead of the heterologous plant CPR,
as this was exploited in the construction of the other engineered
yeast strains (COUM11, RESV11, GEN23 and KAE34). This is
because there was no other pESC vector available as all selectable
markers of the four commercially offered pESC vectors (see
Section 2) were occupied with the basic eight genes involved in
the engineered quercetin biosynthesis (Fig. 2 and Table 3). Similar
biosynthetic patterns in relation to the involvement of the
number of biosynthetic steps were obtained with strain GEN23
leading to isoflavone genistein synthesis.
Analogously, when GEN23 was fed with phenylalanine,
genistein synthesis (six biosynthetic steps, with heterologous
plant CPR also present) was 0.1 mg/L (0.4 mM), while when fed
with p-coumaric (four steps) was improved to 1.4 mg/L (0.52 mM)
and reached an optimum when fed with naringenin (one
enzymatic step) where genistein production was estimated at
7.7 mg/L (28.5 mM). The fact that yeast can produce genistein up to
20 mg/L if engineered with only the necessary IFS/CPR pair and
fed only with naringenin (Kim et al., 2005), shows that GEN23
might have been genetically overloaded with the heterologous
path prior to naringenin biosynthetic step. Regardless of this, our
work shows that the entire biosynthetic pathway from plants
leading to genistein synthesis can be reconstructed in yeast.
Genistein could also be produced by E. coli as shown in Table 3,
but at much lower levels, even when much fewer genes were
employed.
Finally, strain KAE34 was stoichiometrically a much better
kaempferol producer as compared to the quercetin producer
strain QUE44 (Table 3). The kaempferol pathway (in KAE34)
shares the six first biosynthetic steps as identical to those
involved in the quercetin pathway (in QUE44). It appears that
an additional plant CPR existing in KAE34 and not in QUE44
(needed for C4H and F30 H enzymatic activity, as explained above)
might be critical for higher levels of heterologous biosynthesis of
quercetin in yeast. This is because, when KAE34 was fed either
with phenylalanine (involvement of seven biosynthetic steps with
participation of eight genes, Fig. 2), or p-coumarate (five
biosynthetic steps) or naringenin (two biosynthetic steps) the
resulted kaempferol levels were in average 1.3 mg/L (5 mM),
0.9 mg/L (3.1 mM) and 4.6 mg/L (16 mM), respectively. Similarly,
strain QUE44 produced nearly undetectable traces, 0.26 mg/L
(0.9 mM) and 0.38 mg/L (1.3 mM) of quercetin when was fed with
phenylalanine, p-coumarate and naringenin, respectively. This
4- to 12-times drop in flavonoid biosynthesis between KAE34 and
QUE44 is also in agreement with the nearly 4-fold decreased
production of coumarate between strains COUM11 (with plant
CPR) and COUM12 (without plant CPR) as analyzed in the Results
section.
Continuous monitoring of the exogenously added commer-
cially available precursors, like phenylalanine or coumarate
ensured the constant availability of these molecules to the
pathway. However, the continuous feeding of precursors did not
result in any dramatic improvement of their end-product yields in
engineered strains that were low producers (such as GEN23,
KAE34 and QUE44, Table 3). The level of consumption of each
precursor was monitored by Capillary Electrophoresis measure-
ments. This strengthens our initial hypothesis that there must be
one or more enzymatic steps in which the substrate saturation of
key enzymes is suboptimal.
To elaborate this further, we compared the product fluxes
produced in each engineered strain versus the molar stochiometry
at each biosynthetic step. In this respect we found that for the
heterologous biosynthesis of their end-product metabolites,
engineered yeast strains NAR12, GEN23, KAE34 and QUE44
consumed, on average, 3.4 mmoles/L of the precursor molecule
phenylalanine. Given the fact that phenylalanine is also
used in primary metabolism (protein synthesis), this quantity
was not exclusively consumed in the heterologous metabolism of
phenylpropanoids. By comparing the consumption of phenylala-
nine, needed for the growth of a wild type parental yeast
strain (YPH499), grown in the same conditions and nutritional
requirements, we found that YPH499 consumed 2.6 mmoles/L of
phenylalanine. Therefore, the difference of 0.8 mmoles obtained,
should be solely credited in the consumption of phenylalanine
associated with the heterologous biosynthesis of the above
metabolites. Because the stoichiometry of reactants to products
in all the biosynthetic steps, studied in this work, is one to
one (Fig. 2), the consumption of 0.8 mmoles of phenylalanine on
the heterologous metabolism in yeast, if there were ideal
conditions, should lead to the production of 0.8 mmoles of
each end-product, that is 0.8 mmoles of p-coumaric acid (from
strain COUM11), 0.8 mmoles of trans-resveratrol (from strain
RESV11), 0.8 mmoles of naringenin (from strain NAR12),
0.8 mmoles of genistein, 0.8 mmoles of kaempferol (from strain
KAE34) and 0.8 mmoles of quercetin (from strain QUE44).
However, as it is shown in Fig. 7, it produced only 662 mmoles of
p-coumaric acid, corresponding to 83% of flux efficiency when
reaching the second biosynthetic step (PAL-C4H/CPR). Although
the remaining 17% in flux difference could probably be due
to the previous intermediate involved (trans-cinnamic acid), it
appears that overall the PAL and C4H genes and their protein
products function relatively well (83% flux yield). Examining
further the flux efficiency of the pathway, it appears that 4CL that
catalyzes the third enzymatic step also functioned well. In
cultures of yeast strains NAR12, GEN23, KAE34 and QUE44 the
exogenously supplied p-coumaric acid, as precursor molecule, was
rapidly consumed for the production of 4-coumaroyl-CoA. This
was deduced from the quick consumption to nearly zero of
1 mmol/L p-coumaric acid added to cultures, when added
approximately every 40 h, and not from the direct measurement
of 4-coumaroyl-CoA, which was not easily measured with our
analytical capacity.
At this point, 4-coumaroyl-CoA, is a connecting molecule
feeding the stilbenoid synthesis as well as the flavonoid synthesis
(Fig. 2). Thus, we studied in yeast the fluxes of the heterologous
biosynthetic path separately by monitoring the end-product
accumulation in the strains RESV11 (for stilbenoid synthesis)
and NAR12, GEN23, KAE34 and QUE44 (for the other flavonoid
synthesis, Fig. 2). Analyzing the stilbenoid biosynthesis in RESV11,
appears this strain to be a low producer by synthesizing averagely
1.32 mmoles/L trans-resveratrol giving an flux efficiency of 0.16%
(as 662 mmoles were expected). This resveratrol production
was approximately the same whether using phenylalanine
(1.27 mmoles/L) as the precursor compound or p-coumaric acid
(1.36 mmoles/L), showing that the RS enzyme, catalyzing the final
step to the synthesis of trans-resveratrol is not fully efficient.
Given the fact that RS utilizes additionally 3 moles of malonyl-CoA
for the conversion of every mole of 4-coumaroyl-CoA into trans-
resveratrol (Fig. 2) and that by engineering yeast to improve its
intracellular malonyl-CoA pool this results in an increase in
flavanone production up to 576% (Leonard et al., 2007), we
conclude that maybe the poor availability of malonyl-CoA (that is
also needed for the basic metabolism of yeast) is the rate limiting
factor for the resveratrol heterologous biosynthesis. On the other
hand, malonyl-CoA does not appear to be a rate limiting factor
when involved in the flavonoid synthesis in strain NAR12 (end-
product naringenin, Fig. 2). We found that NAR12 is producing 33
and 57 mmoles/L naringenin, when fed with phenylalanine and
p-coumarate, respectively. This indicates a flux efficiency of more
than 8%, showing also that in this case, contrary to our suggestions
for RS which for the same number of biosynthetic steps showed a
flux efficiency of 0.16%.
 ARTICLE IN PRESS
E. Trantas et al. / Metabolic Engineering 11 (2009) 355–366
Furthermore, in both strains NAR12 and GEN23, the enzymes
CHS (malonyl-CoA dependent) and CHI synthesize naringenin
chalcone and naringenin, respectively (Fig. 2). For these genes
conclusions cannot be drawn because of the inability to quantify
the intermediate compounds 4-coumaroyl-CoA and naringenin
chalcone. The fact that a strain capable of producing naringenin
(NAR12) is not high (approx. 5–8% efficient), indicates that both
CHS and CHI enzymes do not function at their maximum activity.
The abovementioned conclusion that 4CL enzyme shows good
activity means that the CHS enzyme has the needed optimal
concentration of its substrate for the conversion into naringenin
chalcone. The inability of quantification of naringenin chalcone
and/or naringenin, did not allow us to conclude about the activity
of CHS enzyme. In addition, CHS and CHI enzymes belong to
multigene families in the soybean genome and there is insuffi-
cient information to assess the degree of specificity for its
substrate, the 4-coumaroyl-CoA (Wingender et al., 1989). The
same goes for CHI and its substrate, the naringenin chalcone
(Shimada et al., 2003).
With regard to the enzyme IFS, when the strain GEN23 was
provided exogenously with 0.5 mM of naringenin, it was not
entirely converted into genistein, only a portion of 28.5 mmoles
(yield 3.6%) was produced, although naringenin was not toxic in
the concentration of 0.5 mM (data not shown). When the
engineered GEN23 strain was fed with phenylalanine and
p-coumaric acid as precursors then 0.4 and 0.52 mmoles of
genistein were produced, respectively (Fig. 7). This was expected,
as the pool of the intermediate naringenin, reached levels
estimated to be 9–15 times lower than the optimal concentration
500 mM needed for IFS saturation (data not shown). This
demonstrates that the CHS and CHI enzymes do not have the
optimal performance and thus supplying the IFS with the needed
substrate.
The bioconversion of naringenin into genistein found to be a
two-step reaction. In the first step naringenin is converted into
2-hydroxy-isoflavanone through the activity of IFS enzyme and in
the second step the 2-hydroxy-isoflavanone loses a water
molecule and is transformed into genistein. The latter step is
carried out either by the action of an dehydratase, IFD or
spontaneously (Akashi et al., 2005; Davies and Schwinn, 2006).
In the strategy we followed a dehydratase was not used and might
be the reason for the low efficiency of the system.
Finally, from the analyses mentioned above, it appears that as
the number of genes that are necessary for the biosynthesis of a
compound is increasing the heterologous biosynthesis system’s
performance is decreasing. This is clear in Fig. 7 in which it
appears that the final concentrations of p-coumaric acid, narin-
genin and kaempferol are 662, 33 and 5 mM, respectively, when
the precursor molecule used was phenylalanine.
This further strengthens our initial hypothesis and let us to
conclude that metabolons are not efficiently organized in
metabolically engineered yeast strains, in order to allow physical
interactions to take place between the intermediate compounds
produced and the sequential enzymes in each pathway. If that was
not true greater flux rates should exist.
Variants of strains KAE34, GEN23 and QUE44 are currently
being constructed to allow us to determine the rate limiting
step(s) in the metabolically engineered S. cerevisiae.
Acknowledgments
Authors thank Prof. Carl J. Douglas (University of British
Coloumbia, Canada) for providing PAL and CPR genes, Dr. A.
Tampakaki (Agricultural University of Athens, GR) for helpful
advice, Dr. Andreas Doulis (National Agricultural Research
365
Foundation), for providing helpful technical support. E.T. is final
year doctorate student and this paper is part of his thesis
requirements. This work was supported by grants of the General
Secretariat Research & Technology (GSRT) of Greece (PENED 03ED
776, co-financed by E.U.-European Social Fund (75%) and the
Greek Ministry of Development-GSRT (25%).) to F.V. and N.P., an
EPEAEK grant (ARCHIMIDIS) to F.V. and University Student
Entrepreneurship (UNISTEP PLUS) to E.T.
References
Akashi, T., Aoki, T., Ayabe, S., 1999. Cloning and functional expression of a
cytochrome P450 cDNA encoding 2-hydroxyisoflavanone synthase involved in
biosynthesis of the isoflavonoid skeleton in licorice. Plant Physiol. 121,
821–828.
Akashi, T., Aoki, T., Ayabe, S., 2005. Molecular and biochemical characterization of
2-hydroxyisoflavanone dehydratase. Involvement of carboxylesterase-like
proteins in leguminous isoflavone biosynthesis. Plant Physiol. 137, 882–891.
Allina, S.M., Pri-Hadash, A., Theilmann, D.A., Ellis, B.E., Douglas, C.J., 1998. 4-
Coumarate: coenzyme A ligase in hybrid poplar. Properties of native enzymes,
cDNA cloning, and analysis of recombinant enzymes. Plant Physiol. 116,
743–754.
Ausubel, F.M., Brent, R., Kingston, R.E., Moore, D.D., Seidman, J.G., Smith, J.A., Struhl,
K., 1994. Current Protocols in Molecular Biology. Wiley, New York.
Ayabe, S.-i., Akashi, T., 2006. Cytochrome P450s in flavonoid metabolism.
Phytochem. Rev. 5, 271–282.
Becker, J.V., Armstrong, G.O., van der Merwe, M.J., Lambrechts, M.G., Vivier, M.A.,
Pretorius, I.S., 2003. Metabolic engineering of Saccharomyces cerevisiae for the
synthesis of the wine-related antioxidant resveratrol. FEMS Yeast Res. 4,
79–85.
Beekwilder, J., Wolswinkel, R., Jonker, H., Hall, R., de Vos, C.H., Bovy, A., 2006.
Production of resveratrol in recombinant microorganisms. Appl. Environ.
Microbiol. 72, 5670–5672.
Braus, G.H., 1991. Aromatic amino acid biosynthesis in the yeast Saccharomyces
cerevisiae: a model system for the regulation of a eukaryotic biosynthetic
pathway. Microbiol. Rev. 55, 349–370.
Davies, K.M., Schwinn, K.E., 2006. Molecular biology and biotechnology of
flavonoid biosynthesis. In: Andersen, O.M., Markham, K.R. (Eds.), Flavonoids:
Chemistry, Biochemistry, and Applications. CRC, Taylor & Francis, Boca Raton,
FL, pp. 143–218.
Dixon, R.A., Lamb, C.J., Masoud, S., Sewalt, V.J., Paiva, N.L., 1996. Metabolic
engineering: prospects for crop improvement through the genetic manipula-
tion of phenylpropanoid biosynthesis and defense responses—a review. Gene
179, 61–71.
Hammerschmidt, R., 1999. PHYTOALEXINS: what have we learned after 60 years?.
Annu. Rev. Phytopathol. 37, 285–306.
Harborne, J.B., Williams, C.A., 2000. Advances in flavonoid research since 1992.
Phytochemistry 55, 481–504.
Havsteen, B.H., 2002. The biochemistry and medical significance of the flavonoids.
Pharmacol. Ther. 96, 67–202.
Hwang, E.I., Kaneko, M., Ohnishi, Y., Horinouchi, S., 2003. Production of plant-
specific flavanones by Escherichia coli containing an artificial gene cluster. Appl.
Environ. Microbiol. 69, 2699–2706.
Jeandet, P., Breuil, A.C., Adrian, M., Weston, L.A., Debord, S., Meunier, P., Maume, G.,
Bessis, R., 1997. HPLC analysis of grapevine phytoalexins coupling photodiode
array detection and fluorometry. Anal. Chem. 69, 5172–5177.
Jiang, H., Wood, K.V., Morgan, J.A., 2005. Metabolic engineering of the phenylpro-
panoid pathway in Saccharomyces cerevisiae. Appl. Environ. Microbiol. 71,
2962–2969.
Jorgensen, K., Rasmussen, A.V., Morant, M., Nielsen, A.H., Bjarnholt, N., Zagrobelny,
M., Bak, S., Moller, B.L., 2005. Metabolon formation and metabolic channeling
in the biosynthesis of plant natural products. Curr. Opin. Plant Biol. 8, 280–291.
Kampa, M., Nifli, A.P., Notas, G., Castanas, E., 2008. Polyphenols and cancer cell
growth. Rev. Physiol. Biochem. Pharmacol. 159, 79–113.
Kang, K., Back, K., 2009. Production of phenylpropanoid amides in recombinant
Escherichia coli. Metab. Eng. 11, 64–68.
Katsuyama, Y., Miyahisa, I., Funa, N., Horinouchi, S., 2007. One-pot synthesis of
genistein from tyrosine by coincubation of genetically engineered Escherichia
coli and Saccharomyces cerevisiae cells. Appl. Microbiol. Biotechnol. 73,
1143–1149.
Kim, D.H., Kim, B.G., Lee, H.J., Lim, Y., Hur, H.G., Ahn, J.H., 2005. Enhancement of
isoflavone synthase activity by co-expression of P450 reductase from rice.
Biotechnol. Lett. 27, 1291–1294.
Kris-Etherton, P.M., Hecker, K.D., Bonanome, A., Coval, S.M., Binkoski, A.E., Hilpert,
K.F., Griel, A.E., Etherton, T.D., 2002. Bioactive compounds in foods: their role in
the prevention of cardiovascular disease and cancer. Am. J. Med. 113 (Suppl.
9B), 71S–88S.
Leonard, E., Lim, K.H., Saw, P.N., Koffas, M.A., 2007. Engineering central metabolic
pathways for high-level flavonoid production in Escherichia coli. Appl. Environ.
Microbiol. 73, 3877–3886.
 ARTICLE IN PRESS
366 E. Trantas et al. / Metabolic Engineering 11 (2009) 355–366
Leonard, E., Yan, Y., Fowler, Z.L., Li, Z., Lim, C.-G., Lim, K.-H., Koffas, M.A.G., 2008.
Strain improvement of recombinant Escherichia coli for efficient production of
plant flavonoids. Mol. Pharm. 5, 257–265.
Leonard, E., Yan, Y., Koffas, M.A.G., 2006. Functional expression of a P450 flavonoid
hydroxylase for the biosynthesis of plant-specific hydroxylated flavonols in
Escherichia coli. Metab. Eng. 8, 172–181.
Liu, R., Hu, Y., Li, J., Lin, Z., 2007. Production of soybean isoflavone genistein in non-
legume plants via genetically modified secondary metabolism pathway.
Metab. Eng. 9, 1–7.
Manach, C., Scalbert, A., Morand, C., Remesy, C., Jimenez, L., 2004. Polyphenols:
food sources and bioavailability. Am. J. Clin. Nutr. 79, 727–747.
Mascarenhas, J.P., 1993. Molecular mechanisms of pollen tube growth and
differentiation. Plant Cell 5, 1303–1314.
Melchior, F., Kindl, H., 1990. Grapevine stilbene synthase cDNA only slightly
differing from chalcone synthase cDNA is expressed in Escherichia coli into a
catalytically active enzyme. FEBS Lett. 268, 17–20.
Miyahisa, I., Funa, N., Ohnishi, Y., Martens, S., Moriguchi, T., Horinouchi, S., 2006.
Combinatorial biosynthesis of flavones and flavonols in Escherichia coli. Appl.
Microbiol. Biotechnol. 71, 53–58.
Miyahisa, I., Kaneko, M., Funa, N., Kawasaki, H., Kojima, H., Ohnishi, Y., Horinouchi,
S., 2005. Efficient production of (2S)-flavanones by Escherichia coli containing
an artificial biosynthetic gene cluster. Appl. Microbiol. Biotechnol. 68,
498–504.
Noel, J.P., Austin, M.B., Bomati, E.K., 2005. Structure–function relationships in plant
phenylpropanoid biosynthesis. Curr. Opin. Plant Biol. 8, 249–253.
Nowakowska, Z., 2007. A review of anti-infective and anti-inflammatory chalcones.
Eur. J. Med. Chem. 42, 125–137.
Pietta, P.G., 2000. Flavonoids as antioxidants. J. Nat. Prod. 63, 1035–1042.
Ro, D.K., Douglas, C.J., 2004. Reconstitution of the entry point of plant
phenylpropanoid metabolism in yeast (Saccharomyces cerevisiae): implications
for control of metabolic flux into the phenylpropanoid pathway. J. Biol. Chem.
279, 2600–2607.
Ro, D.K., Ehlting, J., Douglas, C.J., 2002. Cloning, functional expression, and
subcellular localization of multiple NADPH-cytochrome P450 reductases from
hybrid poplar. Plant Physiol. 130, 1837–1851.
Schulz, W., Eiben, H.-G., Hahlbrock, K., 1989. Expression in Escherichia coli of
catalytically active phenylalanine ammonia-lyase from parsley. FEBS Lett. 258,
335–338.
Shimada, N., Aoki, T., Sato, S., Nakamura, Y., Tabata, S., Ayabe, S., 2003. A cluster of
genes encodes the two types of chalcone isomerase involved in the
biosynthesis of general flavonoids and legume-specific 5-deoxy(iso)flavonoids
in Lotus japonicus. Plant Physiol. 131, 941–951.
Skopelitis, D.S., Paranychianakis, N.V., Paschalidis, K.A., Pliakonis, E.D., Delis, I.D.,
Yakoumakis, D.I., Kouvarakis, A., Papadakis, A.K., Stephanou, E.G., Roubelakis-
Angelakis, K.A., 2006. Abiotic stress generates ROS that signal expression of
anionic glutamate dehydrogenases to form glutamate for proline synthesis in
tobacco and grapevine. Plant Cell 18, 2767–2781.
Stafford, H.A., 1991. Flavonoid evolution: an enzymic approach. Plant Physiol. 96,
680–685.
Steele, C.L., Gijzen, M., Qutob, D., Dixon, R.A., 1999. Molecular characterization of
the enzyme catalyzing the aryl migration reaction of isoflavonoid biosynthesis
in soybean. Arch. Biochem. Biophys. 367, 146–150.
Tian, L., Dixon, R.A., 2006. Engineering isoflavone metabolism with an artificial
bifunctional enzyme. Planta.
Vannelli, T., Wei, Qi, W., Sweigard, J., Gatenby, A.A., Sariaslani, F.S., 2007. Production
of p-hydroxycinnamic acid from glucose in Saccharomyces cerevisiae and
Escherichia coli by expression of heterologous genes from plants and fungi.
Metab. Eng. 9, 142–151.
Ververidis, F., Trantas, E., Douglas, C., Vollmer, G., Kretzschmar, G., Panopoulos, N.,
2007a. Biotechnology of flavonoids and other phenylpropanoid-derived
natural products. Part I: chemical diversity, impacts on plant biology and
human health. Biotechnol. J. 2, 1214–1234.
Ververidis, F., Trantas, E., Douglas, C., Vollmer, G., Kretzschmar, G., Panopoulos, N.,
2007b. Biotechnology of flavonoids and other phenylpropanoid-derived
natural products. Part II: reconstruction of multienzyme pathways in plants
and microbes. Biotechnol. J. 2, 1235–1249.
Watts, K.T., Lee, P.C., Schmidt-Dannert, C., 2004. Exploring recombinant flavonoid
biosynthesis in metabolically engineered Escherichia coli. Chembiochem. 5,
500–507.
Watts, K.T., Lee, P.C., Schmidt-Dannert, C., 2006. Biosynthesis of plant-specific
stilbene polyketides in metabolically engineered Escherichia coli. BMC
Biotechnol. 6, 22.
Werck-Reichhart, D., Feyereisen, R., 2000. Cytochromes P450: a success story.
Genome Biol. 1 Reviews: 3003.1–9.
Williams, R.J., Spencer, J.P., Rice-Evans, C., 2004. Flavonoids: antioxidants or
signalling molecules?. Free Radic. Biol. Med. 36, 838–849.
Wingender, R., Rohrig, H., Horicke, C., Wing, D., Schell, J., 1989. Differential
regulation of soybean chalcone synthase genes in plant defence,
symbiosis and upon environmental stimuli. Mol. Gen. Genet. 218,
315–322.
Winkel-Shirley, B., 2001. Flavonoid biosynthesis. A colorful model for
genetics, biochemistry, cell biology, and biotechnology. Plant Physiol. 126,
485–493.
Winkel, B.S., 2004. Metabolic channeling in plants. Annu. Rev. Plant Physiol. Plant
Mol. Biol. 55, 85–107.
Yan, Y., Chemler, J., Huang, L., Martens, S., Koffas, M.A., 2005a. Metabolic
engineering of anthocyanin biosynthesis in Escherichia coli. Appl. Environ.
Microbiol. 71, 3617–3623.
Yan, Y., Kohli, A., Koffas, M.A.G., 2005b. Biosynthesis of natural flavanones in
Saccharomyces cerevisiae. Appl. Environ. Microbiol. 71, 5610–5613.
Zhang, Y., Li, S.Z., Li, J., Pan, X., Cahoon, R.E., Jaworski, J.G., Wang, X., Jez, J.M., Chen,
F., Yu, O., 2006. Using unnatural protein fusions to engineer resveratrol
biosynthesis in yeast and mammalian cells. J. Am. Chem. Soc. 128,
13030–13031.