Appl Microbiol Biotechnol (2008) 80:579–587
DOI 10.1007/s00253-008-1573-4
BIOTECHNOLOGICAL PRODUCTS AND PROCESS ENGINEERING
Production of the aroma chemicals 3-(methylthio)-1-propanol
and 3-(methylthio)-propylacetate with yeasts
M. M. W. Etschmann & P. Kötter & J. Hauf &
W. Bluemke & K.-D. Entian & J. Schrader
Received: 30 April 2008 / Revised: 9 June 2008 / Accepted: 10 June 2008 / Published online: 3 July 2008
# Springer-Verlag 2008
Abstract Yeasts can convert amino acids to flavor alcohols
following the Ehrlich pathway, a reaction sequence com-
prising transamination, decarboxylation, and reduction. The
alcohols can be further derivatized to the acetate esters by
alcohol acetyl transferase. Using L-methionine as sole
nitrogen source and at high concentration, 3-(methylthio)-
1-propanol (methionol) and 3-(methylthio)-propylacetate
(3-MTPA) were produced with Saccharomyces cerevisiae.
Methionol and 3-MTPA acted growth inhibiting at concen-
trations of >5 and >2 g L−1, respectively. With the wild type
strain S. cerevisiae CEN.PK113-7D, 3.5 g L−1 methionol
and trace amounts of 3-MTPA were achieved in a
bioreactor. Overexpression of the alcohol acetyl transferase
gene ATF1 under the control of a TDH3 (glyceraldehyde-
3-phosphate dehydrogenase) promoter together with an
M. M. W. Etschmann : J. Schrader (*)
Biochemical Engineering Group,
Theodor-Heuss-Allee 25,
60486 Frankfurt am Main, Germany
e-mail: schrader@dechema.de
P. Kötter : K.-D. Entian
Institute of Molecular Biosciences,
Molecular Genetics and Cellular Microbiology,
Max-von-Laue-Str. 9,
60438 Frankfurt am Main, Germany
J. Hauf
SRD—Scientific Research and Development GmbH,
Koehlerweg 20,
61440 Oberursel, Germany
W. Bluemke
Evonik Degussa GmbH,
Postcode 1024-319, Rodenbacher Chaussee 4,
63457 Hanau-Wolfgang, Germany
optimization of the glucose feeding regime led to product
concentrations of 2.2 g L−1 3-MTPA plus 2.5 g L−1
methionol. These are the highest concentrations reported
up to now for the biocatalytic synthesis of these flavor
compounds which are applied in the production of savory
aroma compositions such as meat, potato, and cheese
flavorings.
Introduction
The aroma compounds 3-(methylthio)-1-propanol (methionol)
and 3-(methylthio)-propylacetate (3-MTPA) are derived from
the sulfur amino acid methionine and both impart a powerful
odor reminiscent of soup, meat, onions, and potatoes
(Arctander 1969). Naturally they occur in many fruits, beer,
and malt whisky. Methionol was also found in soy sauce,
tomatoes, cheese, cured pork, rye bread, and seafood
(Burdock 2004). A prominent product of chemical me-
thionine conversion via Strecker degradation in food is
3-(methylthio)-1-propionaldehyde (methional) which is re-
sponsible for light-induced off-taste in milk and dairy
products (Ballance 1961). In beer and wine, the methionine
derivatives are also considered as off-flavors (Gijs et al. 2000;
Moreira et al. 2005; Landaud et al. 2008), whereas in cheese
they are important compounds of characteristic aroma profiles
(Arfi et al. 2006; Bonnarme et al. 2000; Kagkli et al. 2006).
As important constituents of savory flavor preparations, the
“natural” versions of the aroma chemicals are desirable for
the producers, as consumers are increasingly demanding
naturally flavored products (Schreiber et al. 1997). Chemi-
cally identical molecules can yield a much higher price if they
are made from natural educts, e.g., fermentatively produced
amino acids, by a biochemical process, e.g., yeast cultivation
(‘natural’ flavor compounds), as if the same molecule was
580
synthesized by a chemical process (‘nature identical’ flavor
compound).
One of the few articles about the deliberate, yeast-based
production of methionine-derived flavors was published
more than 30 years ago (Schreier et al. 1976). As the
methionine-derived compounds contribute essentially to the
flavor of soy sauce, research does exist in this direction
(Hayashibe et al. 1970; Aoki and Uchida 1991). Further-
more, there is a high-yielding baker’s yeast process whose
product 3-(methylthio)-propionic acid is the oxidation
product of methionol (Whitehead and Ohleyer 1993). A
more recent patent describes methionol production by
fermentation of dairy products supplemented with methio-
nine; however, with 0.1 g L−1 the product concentration
remains relatively low (Liu and Crow 2007).
In the biological formation of methionine-derived fla-
vors, the Ehrlich pathway (Fig. 1) is often involved. This
pathway, which is prevalent in yeasts and has been
intensively studied for the last 100 years (Hazelwood et
al. 2008), is especially active if the amino acid is the sole
nitrogen source for the organism. The amino acid is
transaminated and decarboxylated to the corresponding
aldehyde which is then reduced to the higher alcohol. If
alcohol acetyl transferase activity is present in the organ-
ism, the alcohol can be partially transesterified to the
acetate ester.
The catabolism of the amino acids leucine, valine,
isoleucine, phenylalanine, tryptophan, and methionine in
Saccharomyces cerevisiae and its genetic background was
investigated thoroughly (Dickinson et al. 1997, 1998, 2000,
2003), whereupon methionine degradation turned out to be
more complex than that of the other amino acids (Perpete et
al. 2006). The Ehrlich products of leucine, valine, and
isoleucine are 3-methylbutanol (isoamyl alcohol), isobuta-
nol, and 2-methylbutanol (active amyl alcohol), respective-
ly. These are considered as cheap “fusel alcohols” whereas
the phenylalanine-derived 2-phenylethanol (2-PE) and its
acetate ester 2-phenylethylacetate (2-PEAc) are high-value
natural flavor and fragrance compounds due to their rose-
like odor. Extensive optimization of bioprocess parameters
Fig. 1 Ehrlich pathway in COOH transaminase COOH decarboxylase
general, leading from the amino
acid to the higher alcohol with R NH2 2-ketoglutarate R O
subsequent esterification to the L-glutamate
acetate and reaction principle
amino acid
applied to L-methionine
HO HO
O O
O O
H2N
S S
4-methylthio-
L-methionine
2-oxo-butanoate
Appl Microbiol Biotechnol (2008) 80:579–587
of the cultivation of yeasts with L-phenylalanine as the sole
nitrogen source and the evaluation of different in situ
product removal techniques resulted in a process that
yielded maximum product concentrations of 26.5 g L−1 2-
PE and 6.1 g L−1 2-PEAc in the organic phase (Etschmann
et al. 2003, 2004, 2005; Etschmann and Schrader 2006).
In the present work, L-methionine was selected as
nitrogen source for the cultivation of S. cerevisiae CEN.
PK113-7D and the optimization of methionol and ester
production was carried out combining biochemical and
metabolic engineering. By overexpressing the ATF1 gene
coding for alcohol acetyl transferase 1 and adaptation of the
glucose feed mode, the production of the acetate ester was
considerably enhanced. We therefore report for the first
time the production of methionol and 3-(methylthio)-
propylacetate on the grams per liter scale.
Materials and methods
Plasmid construction and genetic manipulations
Established protocols for molecular biology were followed
(Ausubel et al. 1989). Transformation of yeast strains
utilized the lithium acetate procedure (Schiestl and Gietz
1989). Yeast chromosomal DNA was isolated by the
method described in Hoffman and Winston (1987). The S.
cerevisiae strains used in this study are listed in Table 1.
The S. cerevisiae TDH3 promoter fragment (TDH3
encodes glyceraldehyde-3-phosphate dehydrogenase isoen-
zyme 3) consisting of 671 bp 5′-upstream sequence and the
first two base pairs of the TDH3 open reading frame was
amplified by PCR using the primers TDH3-SalI and TDH3-
BsiWI (see Table 2) and the CEN.PK113-7D genomic
DNA as template. The PCR product was digested with SalI/
BsiWI and subsequently ligated into plasmid pUG6
(Güldener et al. 1996) opened with the same enzymes
resulting in plasmid pPK421.
All genetic modifications of the S. cerevisiae genome
were performed via homologous recombination using PCR
O dehydrogenase OH alcoholacetyltransferase
R
R H R O O
CO2 NADH + H + Acetyl-CoA CoA
NAD+
aldehyde alcohol acetate ester
O
O OH
S S
S
methional = methionol =
3-methylthio-
3-methylthio- 3-methylthio-
propylacetate
propionaldehyde 1-propanol
Appl Microbiol Biotechnol (2008) 80:579–587 581

Table 1 Yeast strains used in this study

Name Genotype Source

CBS 6412 WT (Kyokai No. 7) CBS

RAK1857 Kyokai No. 7 after self-cloning TDH3p::ATF1 Hirosawa et al. 2004
CEN.PK122 MATa/MATα HIS3/HIS3 LEU2/LEU2 TRP1/TRP1 URA3/URA3 MAL2-8cc SUC2/SUC2 Entian and Kötter 2007
CEN.PK113-7D MATa HIS3 LEU2 TRP1 URA3 MAL2-8c SUC2 Entian and Kötter 2007
CEN.PK834-1C MATa HIS3 LEU2 TRP1 URA3 MAL2-8c SUC2 loxP-KanMX4-loxP-TDH3p::(-50,-2)ATF1 This study

The numbers in parentheses indicate the substituted nucleotides of the 5′-UTR of the ATF1 gene. CBS Centraalbureau voor Schimmelcultures,
Utrecht, The Netherlands

products with 40-bp short flanking homologies to the target shown as mean values with error bars. For the bioreactor
site (Wach et al. 1994). cultivations, the results of representative experiments are
The kanMX4-TDH3p cassette of plasmid pPK421 was shown.
amplified by PCR using primers (see Table 2) matching the The yeasts were kept at 4°C on yeast medium agar slants
kanMX4-TDH3p cassette and the 5′-UTR of the ATF1 gene with the following composition per liter: 3 g yeast extract,
(ATF1-G1/ATF1-T2) and pPK421 as a template. After 3 g malt extract, 5 g casein peptone, pancreatic digest, 10 g
transformation of the diploid prototrophic S. cerevisiae glucose, and 15 g agar. The inoculum was prepared by
strain CEN.PK122 using the amplified kanMX4-TDH3p growing cells from a fresh agar slant in 100 mL medium in
cassette, correct integration in G418-resistant colonies was a 300 mL shake flask at 30°C at 180 rpm (at an amplitude
verified by diagnostic PCR. Subsequently, the verified of 25 mm) to the exponential phase.
diploid transformants were sporulated and the haploid strain For the shake flask experiments, a synthetic medium was
CEN.PK834-1C was obtained by tetrad dissection. used: 10 g L-methionine, 1.24 g KH2PO4, and 0.16 g
K2HPO4 (resulting in pH 5) were autoclaved in 0.8 L water.
Cultivation After cooling, a filter-sterilized solution of 1.7 g Difco
Yeast Nitrogen Base without amino acids and ammonium
All chemicals were analytical grade and purchased from sulfate in 0.1 L water was added as well as 0.1 L autoclaved
Fluka (Buchs, Switzerland) or Sigma (Unterhaching, glucose solution. For the inhibition studies, methionine was
Germany), except for 3-MTPA, which was a donation of replaced by 1.84 g L−1 (NH4)2SO4.
R.C. Treatt & Co. Ltd. (Bury St. Edwards, UK). The shake For bioreactor and minireactor cultivations, 20 g L−1 L-
flask experiments were done in duplicate and the results are methionine, 12.4 g L−1 KH2PO4, and 1.6 g L−1 K2HPO4
were sterilized in deionized water. After cooling, an
appropriate amount of autoclaved glucose solution as well
Table 2 Primer used in this study as 0.04 L L−1 of a filter-sterilized vitamin and trace mineral
concentrate was added, whose composition is described
Name Sequence 5′–3′ elsewhere (Etschmann et al. 2004). Bioreactor cultivations
were carried out at 30°C in a 2.4 L bioreactor KLF 2000
TDH3-SalI TCAGGTCGACTTTATCAT
TATCAATACTGCC (Bioengineering, Wald, Switzerland) which was aerated at
TDH3-BsiWI AACTCGTACGATTTTGTTT 1 vvm and stirred between 400 and 1,000 rpm to maintain a
GTTTATGTGTGTTTATTCG dissolved oxygen concentration of 50% saturation with air.
ATF1-T2 TTTTGAAGCCAACCCAAC The pH was controlled at 5.0 by 1 M NaOH.
AAAAATTCGAGACAAGAAA The glucose concentration in the reactor was monitored
ATAGCATAGGCCACTAGTG by a ProcessTrace online measuring and control unit (Trace
GATCTG
GmbH, Braunschweig, Germany) equipped with a dialysis
ATF1-G1 TTGCACGGGGGCCTGATTTT
TCTCATCGATTTCATTCATG
probe. The ProcessTrace unit was coupled with an Ismatec
TTTGTTTGTTTATGTGTGT Reglo Digital MS-2/6 peristaltic pump (Ismatec, Wertheim,
TTATTCG Germany), which delivered concentrated glucose solution
ATF1-A1 GCACAGGACTATTCCACCC to keep the glucose concentration at the desired value.
ATF1-A2 CAGATCCCATACGCCGAGC Minireactor cultivations were carried out at 30°C in a
TDH3-A7 CTTCTGCTCTCTCTGATTTGG Dasgip FedBatch Pro 4-fold parallel cultivation system
K2 GTTTCATTTGATGCTCGATGAG (Dasgip GmbH, Juelich, Germany) equipped with 0.45 L
Restriction sites are in italics, sequences complementary to plasmid culture vessels at a working volume of 0.2 L. These were
pPK421 are in bold aerated at 1 vvm and stirred between 200 and 1,000 rpm to
582
maintain a dissolved oxygen concentration of 50% satura-
tion with air. The pH was controlled at 5.0 by 1 M NaOH.
Inhibition studies
S. cerevisiae CEN.PK113-7D was grown in 300 mL unbaffled
shake flasks in inhibition medium without L-methionine,
which prevented the yeasts from producing methionol. The
methionol concentration was adjusted before cultivation start
by external dosage. The specific growth rate was determined
from the increase in biomass at intervals of 1 h. The
inhibition kinetics of 3-MTPA was determined similarly.
Analytics
Sample preparation
The samples were centrifuged immediately in closed
reaction tubes for 2 min at 16,000 ×g (Eppendorf 5415 R,
Hamburg, Germany) and the supernatant used for further
analysis.
For the quantification of methionine, 10 μL of the
supernatant was injected into a RP-C18 column (250×
4.6 mm, 5 μm, Alltech, Unterhaching, Germany) with an
eluant consisting of 10% methanol and 90% water,
acidified to pH 3 with H3PO4, at 2 mL min−1. Methionol
and 3-MTPA were analyzed on a RP-C18 column (250×
4.6 mm, 10 μm, Alltech) with 40% methanol 60% water (v/v) at
1.5 mL min−1. In both cases, diode array detection was carried
out at 216 nm (SPD M10 advp, Shimadzu, Kyoto, Japan).
Ethanol and glucose were analyzed by HPLC by
injecting 10 μL into an ion exchange column (Aminex
HPX-87H 300×7.8 mm, Bio-Rad, Munich, Germany), with
3 mmol H2SO4 as eluant at 0.6 mL min−1, 27°C, and
detection by refraction index (RID 10 A, Shimadzu).
Quantification of all substances was performed by
calibration and verification with external standards.
For GC–MS analysis, the supernatant was extracted with
an equal volume of hexane by intensive mixing on a vortex
shaker for 1 min. The hexane extract was dried over sodium
sulfate and 1 μL was injected into a GC-17 A with QP5050
MS (Shimadzu) equipped with a DB-5 column (30 m×0.25
mm×0.25 μm, Agilent, Santa Clara, USA). The carrier gas
flow was 0.8 mL min−1 He; split injection (5:1); temper-
ature program from 60°C with 3°C min−1 to 180°C, injector
220°C, detector 240°C; EI 70 eV.
The product 3-MTPA was identified by comparison of its
retention index and mass spectrum with those of the
reference substance. Retention index: 1123. Mass spectrum
m/z (relative intensity): 45 (30), 46 (22), 47 (26), 48 (11), 60
(11), 61 (71), 73 (100), 75 (23), 88 (92), 148 (24).
If the glucose concentration was not analyzed online
with the ProcessTrace system, it was determined offline
Appl Microbiol Biotechnol (2008) 80:579–587
with a YSI 2700 glucose analyzer (Yellow Springs Instru-
ments Inc., Yellow Springs, USA).
For the determination of cell dry weight, 10 mL culture
broth was filtered over a 0.45 μm nitrocellulose filter,
washed with deionized water, and dried in a moisture
analyzer (MA 100, Sartorius, Göttingen, Germany).
Results
Strain characterization
As the toxic effects of methionol and 3-MTPA are strain-
specific, their inhibition kinetics for S. cerevisiae CEN.
PK113-7D had to be determined first.
The maximum specific growth rate (0.53 h−1) was achieved
in the absence of any product (μ/μmax =1) (Fig. 2). Growth
was significantly impaired by concentrations >5 g L−1
methionol. The inhibition increased steadily and reached its
maximum at 40 g L−1 methionol with a complete cessation of
growth.
The inhibition kinetics of 3-MTPA with μmax =0.32 h−1,
revealed growth impairment from 2 g L−1. A higher toxicity
of 3-MTPA can be concluded from the fact that total
inhibition of growth occurred at about 6 g L−1 3-MTPA.
Cultivation experiments
Preliminary shake flask cultivations of S. cerevisiae CEN.
PK113-7D had shown that precursor concentrations of up
to 40 g L−1 methionine did not impair yeast growth (data
not shown) thus all further experiments were executed with
a concentration of 10 g L−1 methionine. These cultivations
yielded product concentrations of 1.5 g L−1 methionol in
1.0
Methionol
0.8 3-MTPA
0.6
µ µ max-1
0.4
0.2
0.0
0 5 10 15 20 25 30 35 40
Product [g L-1 ]
Fig. 2 Relative growth rate of S. cerevisiae CEN.PK113-7D as a
function of the methionol (●) and 3-MTPA (▪) concentration at 30°C in
inhibition medium. Maximum absolute growth rates of μmax =0.53 h−1
and 0.32 h−1 were achieved in the absence of methionol and 3-MTPA,
respectively. Experiments were done in duplicate
Appl Microbiol Biotechnol (2008) 80:579–587 583

7 70
6 60
Methionol, Cdw [gL-1]
To investigate the efficiency of the overexpression, the
newly constructed mutant CEN.PK834-1C and its wild type
CEN.PK113-7D were compared with the ATF1 overexpres-

5 50 sion mutant RAK 1857, which had been designed by

Ethanol, Glucose [gL-1]

Hirosawa and co-workers (2004) from the wild type CBS
4 40
6412.
3 30 In Fig. 4, the final product concentrations of a shake
flask cultivation of both wild types as well as both mutants
2 20 are shown. CBS 6412 produces 1.5 times as much
methionol as CEN.PK113-7D and in both wild types the
1 10
3-MTPA concentration is below the detection limit of the
0 0 HPLC system. The mutant RAK 1857 produces 1.4 times
0 10 20 30 40 50 60 70 as much methionol as CEN.PK834-1C. In contrast, CEN.
Cdw Methionol
Time [h]
Ethanol Glucose
PK834-1C generates ten times more 3-MTPA (0.53 g L−1)
than RAK 1857 (0.05 g L−1).
Fig. 3 Bioprocess kinetics of the production of methionol with S. At high glucose concentrations, S. cerevisiae produces
cerevisiae CEN.PK113-7D during batch cultivation in a 0.45 L
bioreactor with 0.2 L working volume. Glucose was fed at a constant ethanol even under aerobic conditions due to the Crabtree
flow rate between t=26 and 36 h effect (Berry and Brown 1987). As ethanol reportedly
enhances the inhibiting effects of other higher alcohols such
as 2-phenylethanol (Seward et al. 1996) chances are that
about 18 h, after which the glucose reservoir was depleted. this is also true for methionol. Therefore, we aimed at
As this methionol concentration was not yet in the range preventing ethanol formation during the cultivation of
where product inhibition effects were to be expected, CEN.PK834-1C in a bioreactor by keeping the glucose
measures for in situ product removal were foregone for concentration low, which meant 1–2 g L−1. Figure 5 shows
the time being. The presence of the acetate ester 3-MTPA the kinetics of this process.
was verified by GC–MS but the concentration was below Methionol formation started promptly and reached its
the detection limit of the HPLC of 50 mg L−1. maximum of 1.6 g L−1 after 40 h, as did ethanol with a
To eliminate carbon limitation and for improved process maximum concentration of 14 g L−1. Formation of 3-MTPA
control, the strain was cultivated in a bioreactor and the ran parallel with an offset of 20 h and yielded 1.1 g L−1
kinetics are shown in Fig. 3. The formation of methionol product.
started promptly and in a logarithmic fashion for the first As ethanol formation could not be eliminated complete-
24 h and shows association with the biomass formation. ly, a new approach was to use the ethanol as an additional
Both continued, albeit at a lower slope, until 64 h. Apart carbon source so as not to waste any energy invested in the
from 3.5 g L−1 methionol, 6.2 g L−1 cell dry weight and process in the form of glucose. This was achieved by
46 g L−1 ethanol were obtained. deliberately letting the yeasts deplete the glucose reservoir

Strain optimization by genetic engineering 2.5 0.6

Methionol
3-MTPA
The existence of 3-MTPA in the cultivation broth showed 2.0
0.5
that S. cerevisiae CEN.PK113-7D features at least one of
Methionol [g L-1]

0.4
3-MTPA [g L ]

the two alcohol acetyl transferase genes ATF1 and ATF2.

-1

1.5
Overexpression of these genes should thus lead to an
0.3
enhanced production of the acetate ester.
1.0
For this, strain CEN.PK834-1C was constructed by 0.2
substitution of the ATF1 promoter against the strong and
constitutively expressed promoter of the TDH3 gene (see 0.5
0.1
“Materials and methods”). The kanMX4-TDH3p cassette
was genomically integrated by double homologous recom- 0.0
CEN.PK113-7D CEN.PK834-1C CBS 6412 RAK 1857
0.0
bination. The correct integration of both recombination Yeast strain
sites was analysed by diagnostic PCR. Using primer pairs
Fig. 4 Final product concentrations of wild types (CBS 6412 and
ATF1-A1/K2 and TDH3-A7/ATF1-A2 for the PCR resulted
CEN.PK113-7D) and ATF1 overexpression mutants (RAK 1857 and
in the expected PCR products of 639 and 319 bp, CEN.PK834-1C) after 42 h shake flask cultivation in synthetic
respectively (data not shown). medium at 30°C. Experiments were done in duplicate
584 Appl Microbiol Biotechnol (2008) 80:579–587

22 1.8
If the ATF1 overexpression mutant CEN.PK834-1C was
20 1.6 cultivated at a constantly low glucose concentration, the
18
generation of the unwanted byproduct ethanol was low
Ethanol, Methionine, Cdw [gL-1]

1.4
16 (max. 14 g L−1) but so was the generation of the products

Methionol, 3-MTPA [gL-1]

1.2
14 methionol (1.6 g L−1) and 3-MTPA (1.1 g L−1). The final
12 1.0 concentration of 3-MTPA as well as its space–time yield
10 0.8 could be doubled to 2.2 g L−1 and 0.04 g L−1 h−1,
8 respectively, by forcing the yeasts to use the ethanol
0.6
6 generated as carbon source.
0.4
4
2 0.2

0 0.0
0 10 20
Methionol 3-MTPA
Fig. 5 Bioprocess kinetics of S. cerevisiae CEN.PK834-1C in a
bioreactor at constant glucose concentration. The glucose concentra-
tion was kept between 1 and 2 g L−1 by automatic dosage
thus forcing them to metabolize ethanol for a certain
amount of time. The effect of a subsequent addition of
glucose was then investigated and the results are shown in
Fig. 6.
The initially present 18 g L−1 glucose were converted to
6 g L−1 ethanol (and about 3 g L−1 cdw) by S. cerevisiae
CEN.PK834-1C within 20 h. Methionol formation set in
after a short lag phase. The ethanol was metabolized during
the following 8 h and significant amounts of biomass and
methionol were synthesized. However, the rate of the
methionol formation was considerably lower during growth
on ethanol than during growth on glucose. This became
especially apparent upon the glucose pulse at t=28 h, when
the product formation was even faster than during the first
glucose phase. The same applies to the ethanol generation,
while biomass growth seemed to be unaffected by the
change in substrate. Methionol accumulation stagnated at a
maximum of 2.5 g L−1 when the glucose reservoir was
depleted. At the same time, 3-MTPA formation set in.
Discussion
30 40 50 60 70
Time [h]
Ethanol L-Methionine Cdw
In literature, there are already examples for the successful
overexpression of one of the two alcohol acetyl transferase
genes in yeasts. First, in the wine yeast S. cerevisiae VIN13, the
overexpression of ATF1 under the control of the constitutive
PGK1 promoter led to up to 12-fold higher concentrations of
certain acetate esters in wine (Lilly et al. 2000). Second, the
ATF1 overexpression mutant RAK 1857 of the sake yeast S.
cerevisiae Kyokai No. 7 (available at CBS as CBS 6412)
produced nearly five times as much isoamyl acetate
(24.2 ppm) than the wild type strain (4.9 ppm) during
cultivation on rice mash (Hirosawa et al. 2004).
Due to the physiological amino acid concentrations in
both grape and rice mash, the product concentrations in the
examples cited above were naturally low. With the present
work, we investigated whether the principle of ATF1
overexpression could be transferred to S. cerevisiae CEN.
PK113-7D, a prototrophic haploid yeast strain well suited for
bioprocess development (van Dijken et al. 2000). At the
same time, the Ehrlich pathway was harnessed for efficient
methionol-type flavor production after process optimization.
26
2.5
24
22
20 2.0
Glucose, Ethanol, Cdw [gL ]

Methionol, 3-MTPA [gL-1]

-1

During the degradation of the accumulated 12 g L−1 ethanol 18

and the next glucose pulse, a total of 2.2 g L−1 3-MTPA 16
1.5
were generated. Over the course of the experiment of 53 h, 14
12 g L−1 dry biomass were generated. In a control 12
10 1.0
experiment with CEN.PK834-1C under conventional non-
8
limiting glucose feeding conditions, maximum methionol
and 3-MTPA concentrations of 2.0 g L−1 after 30 h and
6
0.5
4
2.1 g L−1 after 70 h were obtained, respectively (data not 2
shown). 0 0.0
Table 3 summarizes the results of the cultivation experi- 0 10 20 30 40 50
Time [h]
ments for the process optimization. The highest methionol Methionol 3-MTPA Glucose Ethanol Cdw
concentration (3.5 g L−1) as well as the highest product
Fig. 6 Bioprocess kinetics of S. cerevisiae CEN.PK834-1C in a
yield (0.64 mol mol−1) was achieved with the fed-batch bioreactor. Glucose was fed manually at t=8, 29, and 45 h. The
cultivation of the wild type strain CEN.PK113-7D. glucose concentration in the reactor was monitored online
Appl Microbiol Biotechnol (2008) 80:579–587
Table 3 Summary of the key figures of the bioreactor cultivations of S. cerevisiae CEN.PK113-7D and CEN.PK834-1C in different feeding
modes
Strain CEN.PK834-1C
Cultivation mode Constant glucose concentration
Max. ethanol concentration [g L−1] 14
Product Methionol 3-MTPA
Concentration [g L−1] 1.6 1.1
Process duration [h] 40 65
Space–time yield [g L−1 h−1] 0.04 0.02
yP/X [g (g cdw)−1] 0.22 0.18
yP/S [mol mol−1] 0.21 0.09
Using L-methionine as sole N-source and at high
concentration, a methionol yield of 0.64 mol mol−1 and a
final methionol concentration of 3.5 g L−1 were obtained in
a fed-batch cultivation of the wild type strain S. cerevisiae
CEN.PK113-7D. This is competitive with the only pub-
lished bioprocess of industrial relevance whose product,
however, is not methionol but its oxidation product 3-
(methylthio)-propionic acid (Whitehead and Ohleyer 1993).
After optimization by biochemical and genetic engineer-
ing, the acetate ester of methionol was accessible concom-
itantly and maximum product concentrations of 2.5 g L−1
methionol and 2.2 g L−1 3-MTPA were obtained with the
ATF1 overexpression mutant (Fig. 6). To the authors’
knowledge, concentrations of these methionol-type flavor
compounds as high as reported in this work have not been
cited in literature before. The 3-MTPA formation never set
in before a methionol concentration of >1.0 to 1.5 g L−1 in
the supernatant had been reached (cf. Figs. 5 and 6)
indicating that not only overexpression but also a critical
precursor concentration is necessary to thrive on Atf1
activity. With respect to 3-MTPA production, the mutant
CEN.PK834-1C outperformed RAK 1857 (Hirosawa et al.
2004) by a factor of ten when cultivated under production
conditions, although essentially the same genetic approach
of overexpressing ATF1 was followed. The reason why S.
cerevisiae CEN.PK122 is significantly better suited for 3-
MTPA production through overexpression of ATF1 than S.
cerevisiae CBS 6412 still remains elusive.
Especially the shift from glucose as the sole carbon
source to glucose and ethanol as alternate carbon sources
caused a remarkable increase in product concentrations and
yields YP/S with CEN.PK834-1C. Being forced to re-
metabolize the ethanol produced, the yeast derived more
total energy from the C-source glucose than in case of a
continuous glucose feeding. This became manifest in the
fact that significantly more biomass was synthesized from
the same amount of glucose when it was intermittently
dosed. The enhanced nitrogen demand associated with the
boosted cell growth caused an increased amino acid
consumption by the culture and, thus, a higher product
585
CEN.PK113-7D
Intermittent glucose addition Fed-batch
Ethanol produced served as interim C-source 46
Methionol 3-MTPA Methionol
2.5 2.2 3.5
33 53 64
0.07 0.04 0.05
0.37 0.19 0.50
0.35 0.15 0.64
concentration, because alcohol formation in S. cerevisiae
via the Ehrlich pathway has been shown to be strongly
growth-associated (Stark et al. 2002). Furthermore, with
alternate C-sources the flux through the Ehrlich pathway
was obviously enhanced at the expense of other degrada-
tion pathways since higher molar product yields were
achieved, too. Interestingly, when pulsing glucose for the
second time, methionol formation rate was even higher than
after the first glucose pulse despite a constant biomass
accumulation. By this means, an even higher methionol
concentration was obtained than in a control experiment
with a continuous non-limiting glucose feeding (2.5 g L−1
vs. 2.0 g L−1). When product formation ceased, there was
no limitation in carbon or nitrogen and, from the inhibition
kinetics (Fig. 2), only a small effect caused by 3-MTPA was
to be expected. Ethanol had accumulated to a concentration
of 22 g L−1, which also is no reason for complete cessation
of growth and a synergistic effect of ethanol and methionol
was excluded experimentally (data not shown). As the
positive effect of the alternate C-sources is not explicable so
far, further experiments are necessary. Ideally, these should
include analyses of the cellular metabolite pools and of key
enzyme activities involved in product formation to gain a
better understanding of the bioprocess kinetics observed
when applying the alternate C-sources cultivation method.
By legislation, an enzymatically or microbially produced
flavor compound, which is known to be present in nature, is
only “natural” if the raw material used as substrate itself is
from natural origin (Mueller 2007). In principle, two major
sources of natural L-methionine exist: from animal/plant
source or microbial production. Plants do not contain
sufficient L-methionine to justify extraction, but it can be
isolated from animal feathers or hairs, and this is still done
for feed and pharma grade L-methionine. But, as for
gelatine, the customer acceptance of animal-derived amino
acids has considerably decreased making these sources less
important in the future.
Attempts for microbial L-methionine fermentation actu-
ally lack on titer and yield (Kumar and Gomes 2005), but
considerable improvement was made. Thus, L-methionine
586
from microbial fermentation may be competitive with
current production (enzymatic, animal derived) in the future
and will lead to reasonable prices for natural methionine.
These efforts are supported by the attractive and growing
markets for L-methionine (pharma, food).
Taking into account that natural flavor compounds are
typically sold at prices at least one order of magnitude
higher than those of their chemical counterparts (Schrader
et al. 2004), the results presented in this work may serve as
a promising starting point for further R&D activities
towards an industrial production of natural methionol and
3-MTPA.
However, especially European consumers are very
critical towards gene technology in general, even more so
if it concerns foods and thus tend to avoid it; US consumers
on the other hand are much more open-minded towards this
topic (Lusk and Rozan 2005). In Japan, yeasts such as RAK
1857 are regarded as “self-cloning” as the vector sequences
are removed and the strain contains only its own DNA but
none from a different organism. Thus, the yeast does not
have to be labeled as genetically modified (Hirosawa et al.
2004). In conclusion, it is not only the cost effectiveness of
a newly developed biotechnological production process that
decides about the success of the product, the latter also has
to be accepted by the consumer.
Acknowledgements We thank Dr. Rinji Akada for the donation of
the yeast strain RAK 1857, R.C. Treatt & Co. Ltd. for the donation of
the 3-MTPA standard, and the German Federal Ministry for Food,
Agriculture and Consumer Protection for funding the project (No.
22008803).
References
Aoki T, Uchida K (1991) Enhanced formation of 3-(methylthio-)-
1-propanol in a salt-tolerant yeast, Zygosaccharomyces rouxii,
due to deficiency of S-adenosylmethionine synthase. Agric Biol
Chem 55:2113–2116
Arctander S (1969) Perfume and flavor chemicals. Aroma chemicals.
Published by the author, Montclair, USA
Arfi K, Landaud S, Bonnarme P (2006) Evidence for distinct L-
methionine catabolic pathways in the yeast Geotrichum candidum
and the bacterium Brevibacterium linens. Appl Environ Microbiol
72:2155–62
Ausubel FM, Brent R, Kingston RE, Moore DD, Seidman JG, Smith
JA, Struhl K (1989) Current protocols in molecular biology.
Wiley, New York
Ballance PE (1961) Production of volatile compounds related to the
flavour of foods from the Strecker degradation of DL-methionine. J
Sci Food Agric 12:532–536
Berry DR, Brown C (1987) Physiology of yeast growth. In: Berry DR,
Russel I, Steward GG (eds) Yeast biotechnology. Allen and
Unwin, London, pp 159–199
Bonnarme P, Psoni L, Spinnler HE (2000) Diversity of L-methionine
catabolism pathways in cheese-ripening bacteria. Appl Environ
Microbiol 66:5514–7
Appl Microbiol Biotechnol (2008) 80:579–587
Burdock GA (2004) Fenaroli’s handbook of flavor ingredients. CRC,
Boca Raton, p 1320
Dickinson JR, Lanterman MM, Danner DJ, Pearson BM, Sanz P,
Harrison SJ, Hewlins MJ (1997) A 13C nuclear magnetic
resonance investigation of the metabolism of leucine to isoamyl
alcohol in Saccharomyces cerevisiae. J Biol Chem 272:26871–8
Dickinson JR, Harrison SJ, Hewlins MJ (1998) An investigation of
the metabolism of valine to isobutyl alcohol in Saccharomyces
cerevisiae. J Biol Chem 273:25751–6
Dickinson JR, Harrison SJ, Dickinson JA, Hewlins MJ (2000) An
investigation of the metabolism of isoleucine to active Amyl
alcohol in Saccharomyces cerevisiae. J Biol Chem 275:10937–42
Dickinson JR, Salgado LE, Hewlins MJ (2003) The catabolism of
amino acids to long chain and complex alcohols in Saccharomyces
cerevisiae. J Biol Chem 278:8028–34
van Dijken JP, Bauer J, Brambilla L, Duboc P, Francois JM, Gancedo
C, Giuseppin ML, Heijnen JJ, Hoare M, Lange HC, Madden EA,
Niederberger P, Nielsen J, Parrou JL, Petit T, Porro D, Reuss M,
van Riel N, Rizzi M, Steensma HY, Verrips CT, Vindelov J,
Pronk JT (2000) An interlaboratory comparison of physiological
and genetic properties of four Saccharomyces cerevisiae strains.
Enzyme Microb Technol 26:706–714
Entian K-D, Kötter P (2007) 25 Yeast genetic strain and plasmid
collections. Meth Microbiol 36:629–666
Etschmann MMW, Sell D, Schrader J (2003) Screening of yeasts for
the production of the aroma compound 2-phenylethanol in a
molasses-based medium. Biotech Lett 25:531–536
Etschmann MMW, Sell D, Schrader J (2004) Medium optimization for
the production of the aroma compound 2-phenylethanol using a
genetic algorithm. J Mol Catal B 29:187–193
Etschmann MM, Sell D, Schrader J (2005) Production of 2-phenylethanol
and 2-phenylethylacetate from L-phenylalanine by coupling whole-
cell biocatalysis with organophilic pervaporation. Biotechnol
Bioeng 92:624–634
Etschmann MM, Schrader J (2006) An aqueous-organic two-phase
bioprocess for efficient production of the natural aroma chem-
icals 2-phenylethanol and 2-phenylethylacetate with yeast. Appl
Microbiol Biotechnol 71:440–443
Gijs L, Perpète P, Timmermans A, Collin S (2000) 3-Methylthiopro-
pionaldehyde as precursor of dimethyl trisulfide in aged beers. J
Agric Food Chem 48:6196–6199
Güldener US, Heck T, Fielder J, Beinhauer, Hegemann JH (1996) A
new efficient gene disruption cassette for repeated use in budding
yeast. Nucleic Acids Res 24:2519–24
Hayashibe M, Katoda S, Owada H, Yoshida H, Katayosa A,
Terashima T (1970) Methionine metabolism in yeast. J Ferment
Technol 48:22–28
Hazelwood LA, Daran JM, van Maris AJ, Pronk JT, Dickinson JR
(2008) The Ehrlich pathway for fusel alcohol production: a
century of research on Saccharomyces cerevisiae metabolism.
Appl Environ Microbiol 74:2259–66
Hirosawa I, Aritomi K, Hoshida H, Kashiwagi S, Nishizawa Y, Akada
R (2004) Construction of a self-cloning sake yeast that over-
expresses alcohol acetyltransferase gene by a two-step gene
replacement protocol. Appl Microbiol Biotechnol 65:68–73
Hoffman CS, Winston F (1987) A ten-minute DNA preparation from
yeast efficiently releases autonomous plasmids for transformation
of Escherichia coli. Gene 57:267–272
Kagkli DM, Tache R, Cogan TM, Hill C, Casaregola S, Bonnarme P
(2006) Kluyveromyces lactis and Saccharomyces cerevisiae, two
potent deacidifying and volatile-sulphur-aroma-producing micro-
organisms of the cheese ecosystem. Appl Microbiol Biotechnol
73:434–442
Kumar D, Gomes J (2005) Methionine production by fermentation.
Biotechnol Adv 23:41–61
Appl Microbiol Biotechnol (2008) 80:579–587
Landaud S, Helinck S, Bonnarme P (2008) Formation of volatile
sulfur compounds and metabolism of methionine and other sulfur
compounds in fermented food. Appl Microbiol Biotechnol
77:1191–205
Lilly M, Lambrechts MG, Pretorius IS (2000) Effect of increased
yeast alcohol acetyltransferase activity on flavor profiles of wine
and distillates. Appl Environ Microbiol 66:744–753
Liu SQ, Crow VL (2007) Dairy product and process. WO 2007/
136280A1
Lusk JL, Rozan A (2005) Consumer acceptance of biotechnology and
the role of second generation technologies in the USA and
Europe. Trends Biotechnol 23:386–387
Moreira N, Mendes F, Hogg T, Vasconcelos I (2005) Alcohols, esters
and heavy sulphur compounds production by pure and mixed
cultures of apiculate wine yeasts. Int J Food Microbiol 103:285–
294
Mueller DA (2007) Flavours: the legal framework. In: Berger RG (ed)
Flavours and fragrances. Chemistry, bioprocessing and sustain-
ability. Springer, Berlin, pp 15–24
Perpete P, Duthoit O, De Maeyer S, Imray L, Lawton AI,
Stavropoulos KE, Gitonga VW, Hewlins MJ, Dickinson JR
(2006) Methionine catabolism in Saccharomyces cerevisiae.
FEMS Yeast Res 6:48–56
587
Schiestl RH, Gietz RD (1989) High efficiency transformation of intact
yeast cells using single stranded nucleic acids as a carrier. Curr
Genet 16:339–46
Schrader J, Etschmann MM, Sell D, Hilmer JM, Rabenhorst J (2004)
Applied biocatalysis for the synthesis of natural flavour compounds—
current industrial processes and future prospects. Biotechnol Lett
26:463–72
Schreiber WL, Scharpf LG, Katz I (1997) Future needs of chemistry
in flavors and fragrances. Perfumer & Flavorist 22:11–16
Schreier P, Drawert FD, Junker A, Barton H, Leupold G (1976) Über
die Biosynthese von Aromastoffen durch Mikroorganismen. Z
Lebensm Unters Forsch 162:279–284
Seward R, Willetts JM, Dinsdale MG, Lloyd D (1996) The effects of ethanol,
hexan-1-ol, and 2-phenylethanol on cider yeast growth, viability,
and energy status; synergistic inhibition. J Inst Brew 102:439–443
Stark D, Münch T, Sonnleitner B, Marison IW, Stockar von U (2002)
Extractive bioconversion of 2-phenylethanol from L-phenylalanine
by Saccharomyces cerevisiae. Biotechnol Prog 18:514–523
Wach A, Brachat A, Pohlmann R, Philippsen P (1994) New
heterologous modules for classical or PCR-based gene disruptions
in Saccharomyces cerevisiae. Yeast 10:1793–808
Whitehead IM, Ohleyer E (1993) Microbial carboxylic acid production
method. WO9308293