Impact of mixed Torulaspora delbrueckiiSaccharomyces cerevisiae

culture on high-sugar fermentation

Marina Bely
a,

, Philippe Stoeckle
a
, Isabelle Masneuf-Pomarde
a,b
, Denis Dubourdieu
a
a
UMR 1219 oenologie, Universit Bordeaux 2, INRA, ISVV, 351 cours de la Libration, 33400 Talence, France
b
ENITA de Bordeaux, 1 cours du Gnral de Gaulle, CS 40201, 33175 Gradignan cedex, France
Received 10 May 2007; received in revised form 23 November 2007; accepted 30 December 2007
Abstract
Conventional wine yeasts produce high concentrations of volatile acidity, mainly acetic acid, during high-sugar fermentation. This alcoholic
fermentation by-product is highly detrimental to wine quality and, in some cases, levels may even exceed legal limits. In this study, a non-
conventional species, Torulaspora delbrueckii, was used, in pure cultures and mixed with Saccharomyces cerevisiae yeast, to ferment botrytized
musts. Fermentation rate, biomass growth, and the formation of volatile acidity, acetaldehyde, and glycerol were considered. This study
demonstrated that T. delbrueckii, often described as a low acetic producer under standard conditions, retained this quality even in a high-sugar
medium. Unlike S. cerevisiae, this species did not respond to the hyper-osmotic medium by increasing acetic production as soon as it is inoculated
into the must. Nevertheless, this yeast produced low ethanol and biomass yields, and the fermentation was sluggish. As a result, T. delbrueckii
fermentations do not reach the required ethanol content (14%vol.), although this species can survive at this concentration. A mixed culture of T.
delbrueckii and S. cerevisiae was the best combination for improving the analytical profile of sweet wine, particularly volatile acidity and
acetaldehyde production. A mixed T. delbrueckii/S. cerevisiae culture at a 20:1 ratio produced 53% less in volatile acidity and 60% less
acetaldehyde than a pure culture of S. cerevisiae. Inoculating S. cerevisiae after 5 days' fermentation by T. delbrueckii had less effect on volatile
acidity and acetaldehyde production and resulted in stuck fermentation. These results contribute to a better understanding of the behaviour of non-
Saccharomyces and their potential application in wine industry.
2008 Elsevier B.V. All rights reserved.
Keywords: Wine; High-sugar fermentation; Torulaspora delbrueckii; Mixed culture; Volatile acidity; Acetaldehyde
1. Introduction
Spontaneous alcoholic fermentation of grape must is a com-
plex process, carried out by sequential action of several yeast
genera and species, found on grapes and in must. Saccharo-
myces cerevisiae is not the only species that grows during
natural wine fermentations. The early stages in fermentation are
dominated by the growth of non-Saccharomyces yeasts, char-
acterized by low fermentative capacity. After the first few days
of fermentation, they die off, due to the increasing concentration
of ethanol (Heard and Fleet, 1985; Heard and Fleet, 1986).
Subsequently, Saccharomyces, the most strongly fermenting,
ethanol-tolerant species, take over the fermentation. Recent
quantitative studies of winemaking ecology showed that sig-
nificant levels of certain non-Saccharomyces species survive for
longer periods than previously thought (Fleet et al., 1984; Heard
and Fleet, 1985; Cabrera et al., 1988; Pardo et al., 1989; Herraiz
et al., 1990; Lema et al., 1996; Egli et al., 1998; Fleet, 2003;
Pina et al., 2004). Furthermore, several authors have reported
on the influence of non-Saccharomyces yeast species on wine
quality under usual winemaking conditions (Cabrera et al.,
1988; Herraiz et al., 1990; Ciani and Picciotti, 1995; Lema
et al., 1996; Ciani and Ferraro, 1998; Romano et al., 2003;
Fleet, 2003). One of these non-Saccharomyces, Torulaspora
delbrueckii (formerly Saccharomyces rosei), is reported to have
Available online at www.sciencedirect.com
International Journal of Food Microbiology 122 (2008) 312320
www.elsevier.com/locate/ijfoodmicro

Corresponding author. Tel.: +33 5 4000 6489; fax: +33 5 4000 6468.
E-mail address: marina.bely@univ-pau.fr (M. Bely).
0168-1605/$ - see front matter 2008 Elsevier B.V. All rights reserved.
doi:10.1016/j.ijfoodmicro.2007.12.023
a positive effect on the taste and aroma of alcoholic beverages
(Herraiz et al., 1990; Moreno et al., 1991; Ciani and Picciotti,
1995, Ciani and Maccarelli, 1998) and exhibits low produc-
tion of acetaldehyde, acetoin, acetate, and ethyl acetate (Cabrera
et al., 1988; Herraiz et al., 1990; Martinez et al., 1990, Ciani and
Ferraro, 1998). Due to its high fermentation purity, its use un-
der standard conditions, in mixed or sequential culture with
S. cerevisiae, has been proposed as a way of reducing the acetic
acid content in wine (Herraiz et al., 1990; Ciani and Picciotti,
1995; Ciani et al., 2006). Holm Hansen et al. (2001) indicated
that T. delbrueckii was less tolerant to low available oxygen
conditions than S. cerevisiae. On the other hand, little is known
about the potential of non-Saccharomyces wine yeasts in high-
sugar fermentations. Laffon-Lafourcade et al. (1981) evaluated
the possibility of using S. rosei (T. delbrueckii) to improve the
quality of botrytized wine. These wines are commonly made
from grapes infected with Botrytis cinerea (noble rot), which
results in concentrations of grape sugars ranging from 350
to 450 g/L. Sipiczki et al. (2001), Naumov et al. (2002) and
Sipiczki (2003) reported a high frequency of Saccharomyces
bayanus, which dominated the latter stages of fermentation, and
the presence of Candida stellata, a high acid producer in the
fermenting yeast flora of Tokay wine. T. delbrueckii was also
detected (1%) in botrytized must during the early stages of
fermentation (Sipiczki et al., 2001; Antunovics et al. 2003).
The amount of volatile acidity (mainly acetate) plays a sig-
nificant role in wine aroma and excessive concentrations of this
alcoholic fermentation by-product are highly detrimental to
wine quality. The quantity of volatile acidity produced by S.
cerevisiae is usually low (0.25 to 0.50 g/L expressed in acetic
acid), but may be higher under certain fermentation conditions.
In particular, during fermentation of high-sugar media, such as
botrytized musts, the volatile acidity content may be over 1.8 g/
L or even higher, i.e. over the EEC legal limit of 1.5 g/L. S.
cerevisiae produce acetic acid as a by-product of the hyper-
osmotic stress response to the high-sugar concentrations in
grape must (Blomberg and Alder, 1992).
Recently Erasmus et al. (2004), comparing the suitability
of different commercial S. cerevisiae wine yeast strains for ice-
wine production, confirmed that wine yeast strains varied greatly
in their capacity to form acetic acid. The S. cerevisiae Zymaflore
ST strain was identified as one of the most suitable for ice-wine
production, confirming previous findings in botrytized must
(Barbe et al., 2000; Masneuf and Dubourdieu, 2000). It was also
recently established that the initial assimilable nitrogen in the
must and the physiological state of the inoculum affected this
strain's volatile acidity production in high-sugar fermentations
(Bely et al., 2003; Bely, Masneuf-Pomarde and Dubourdieu,
2005). The optimumnitrogen concentration was 190 mgN/L and
direct inoculation of rehydrated active dry yeasts produced the
most volatile acidity, while yeast pre-culture for 24 h reduced the
final production.
Nevertheless, in spite of more detailed knowledge of the S.
cerevisiae Zymaflore ST strain, volatile acidity production in
sweet wine may still be problematic, particularly in dry years in
the Sauternes area, when initial sugar concentrations are over
350 g/L.
The aim of this work was to evaluate the ability of a non-
conventional, non-Saccharomyces yeast species, T. delbrueckii,
often described as a low acid acetic producer, to ferment high-
sugar must. This paper reports the impact of pure, mixed, and
sequential cultures of T. delbrueckii and S. cerevisiae on fer-
mentation behaviour, biomass growth, and analytical profile.
2. Materials and methods
2.1. Microorganisms and media
Two T. delbrueckii strains from the MUCL collection
(Mycothque de l'Universit Catholique de Louvain-la-Neuve),
T. delbrueckii 27828 and T. delbrueckii 31703, and one S.
cerevisiae strain, Zymaflore ST(Laffort Oenologie, France), were
used. The Zymaflore ST strain was selected from native micro-
flora during spontaneous fermentation of a botrytized must for its
ability to produce small amounts of volatile acidity and SO
2
-
binding compounds (Barbe et al., 2000, Masneuf and Dubour-
dieu, 2000), and is commonly used in Sauternes.
Yeasts were grown at 30 C on complete YPD medium
(1% yeast extract, 1% peptone, 2% dextrose) solidified with 2%
agar.
2.2. Cell population count
Cell populations were determined by measuring absorbance
at 610 nm. One absorbance unit corresponded to 6

10
6
cfu/mL
for the two T. delbrueckii strains and to 7

10
6
cfu/mL for S.
cerevisiae ST. Cell population was measured every day until no
further increase in O.D was observed (O.D610 max.).
Yeast growth was also determined by plate counting in some
cases. Samples were withdrawn throughout fermentation and
diluted appropriately in dilution medium. Non-Saccharomyces
cells were counted using lysine agar (LA) (Per litre: 66 g lysine
medium (Oxoid), 10 mL 50% potassium lactate, and 1 mL 10%
lactic acid, pH 4.8). Total yeast cells were counted using YPDA
medium supplemented with chloramphenicol (100 mg/L). The
number of S. cerevisiae was given as the difference between the
total plate count using YPDA and the plate count using LA.
YPDA and LA plates were incubated at 25 C for 4 days before
counting.
2.3. Must
The musts were obtained from botrytized Semillon grapes,
harvested in vineyards in the Sauternes and Barsac appellations
(20032005 vintages). They were collected in the cellar after
settling and frozen. Initial sugar concentrations were 360 g/L.
Musts were supplemented with nitrogen by adding ammonium
sulphate (NH4)
2
SO
4
to a level of 190 mgN/L. According to
previous studies (Bely et al., 2003), this nitrogen concentration
is optimum for minimising volatile acidity production by this
strain. The indigenous yeast population was estimated by plate
counting.
Pasteurized musts were heated to 90 C for 15 min and the
effectiveness of this treatment was verified by plate counting.
313 M. Bely et al. / International Journal of Food Microbiology 122 (2008) 312320
2.4. Fermentation conditions
Triplicate experiments were carried out in 375 mL sterile
bottles (320 mL must per bottle) at 23 C, inoculated with 24 h
pre-culture (grown in the same medium diluted 1:1 at 23 C)
to obtain an initial level of 2

10
6
cells/mL. This pre-culture
preparation is essential to minimise volatile acidity production
in high-sugar fermentations (Bely et al., 2005). Mixed
fermentation trials were simultaneously inoculated with 10
7
(or 5.10
6
) T. delbrueckii 31703 cells/mL plus S. cerevisiae
ST at ratios varying from 5:1 to 100:1. Sequential fermenta-
tions were inoculated with 2

10
6
cells/mL of T. delbrueckii,
followed by 2

10
6
cells/mL of S. cerevisiae after 5 days'
fermentation.
Fermentations under static conditions were monitored by
CO
2
release, determined by measuring weight loss every 24 h.
Fermentations were stopped by adding 250 mg/L SO
2
, when a
minimumof 35g CO
2
per bottle had beenreleased, corresponding
to an alcohol concentration of approximately 14 %vol., in accor-
dance with usual winemaking practices in Sauternes wineries.
2.5. Must and wine analysis
Ammonia was assayed by an enzymatic method (Roche-
Biopharm), while primary amino acids were evaluated using an
OPA/NAC spectrophotometric assay, as described by Dukes
and Buzkle (1998).
Ethanol concentrations (%vol.) were measured by infrared
refractance (Infra Analyser 450, Technicon, Plaisir, France).
Sugar (g/L) and volatile acidity (expressed in g/L acetic acid)
were determined chemically by colorimetry (A460 nm) in con-
tinuous flux (Sanimat, Montauban, France). Glycerol and acetal-
dehyde were analysed by enzymatic methods (Roche-Biopharm,
kit. n10148270035, 10668613035). These analyses were
carried out in the Sarco Laboratory (Floirac, France), according
to quality assurance procedures approved by the French Ac-
creditation committee COFRAC.
2.6. Statistical analysis
Volatile acidity, acetaldehyde, ethanol, glycerol, nitrogen, and
sugar concentrations were subjected to single-factor variance
analysis with three repetitions (=1%.), followed by a Newman-
Keuls average comparison test (=5%) (ANOVA, Stabox
Software , Grimmer Logiciels).
3. Results and discussion
3.1. Fermentation behaviour of pure culture
Several fermentations were carried out in a pasteurized must
containing 360 g/L sugar. The influence of different yeasts on
fermentation kinetics and on the main oenological character-
istics is shown in Fig. 1 and Table 1. All data correspond to the
average values obtained from triplicate fermentations. Our
results on fermentation kinetics confirmed that S. cerevisiae ST
was the only yeast capable of producing the minimum amounts
of CO
2
required (35 g/bottle, corresponding to an ethanol con-
centration of approximately 14%vol.), while both T. delbrueckii
27828 and T. delbrueckii 31703 stopped fermenting when only
18.7 and 23.9 g/bottle CO
2
, respectively, had been released.
The fermentation rate varied markedly depending on the yeast
used (Fig. 1). Alcoholic fermentation with S. cerevisiae ST
was completed 11 days after inoculation, whereas both T.
delbrueckii 27828 and T. delbrueckii 31703 strains had slower
fermentation rates and fermentations stopped after 20 and
26 days, respectively. It is worthwhile noting that, in both T.
delbrueckii, the fermentation rate was the same during the first
15 g/bottle CO
2
released (corresponding to an ethanol content
of 6%vol.). S. cerevisiae ST had the highest cell densities (12.4)
after 3 days (not shown), while both T. delbrueckii 27828 and
T. delbrueckii 31703 stopped growing after 4 days' fermenta-
tion at cell densities of 4 and 8.4, respectively (Table 1). This
is consistent with previous results, which reported that T.
delbrueckii species had lower growth rates than S. cerevisiae
(Mauricio et al., 1991; Ciani and Picciotti, 1995; Ciani, 1997).
However, this reduced growth activity was not due to a deficit
of assimilable nitrogen compounds, as they did not consume all
that was available. Similar behaviour was reported by Ciani
et al. (2006).
The higher ethanol concentration in the S. cerevisiae ST
culture (15.1% vol.) was not the only difference. In comparison
with the S. cerevisiae ST strain, the two T. delbrueckii fer-
mentations also had a lower ethanol/sugar yield, which differs
froma previous study using a low-sugar must (Ciani and Picciotti,
1995). Glycerol production by T. delbrueckii 27828 was lower
than that of the other two yeasts.
T. delbrueckii 27828 and T. delbrueckii 31703 produced 2.7-
and 1.9-fold less volatile acidity, respectively, than S. cerevisiae
ST, confirming the specific characteristic of low acetic acid
production of these yeasts species, as already reported in previ-
ous studies under standard winemaking conditions (Cabrera
et al., 1988; Herraiz et al., 1990; Martinez et al., 1990; Ciani and
Picciotti, 1995; Ciani, 1997; Ciani and Ferraro, 1998).
Fig. 1. Fermentation courses, in pasteurized must, at 23 C. Pure S. cerevisiae
STculture (S). Pure T. delbrueckii culture (T). Means of triplicate fermentations,
max. SD0.9.
314 M. Bely et al. / International Journal of Food Microbiology 122 (2008) 312320
Yeast responds to increased external osmolarity by enhanced
production and intracellular accumulation of glycerol to coun-
terbalance the osmotic pressure (Blomberg and Alder, 1989;
Blomberg and Alder, 1992; Myers et al., 1997). In S. cerevisiae
yeast, this regulation mechanism is due to the expression of
certain osmoresponsive genes including glycerol 3-phosphate
dehydrogenase, encoded by GPD1 (Albertyn et al., 1994;
Attfied and Kletsas, 2000), as well as two types of aldhehyde
dehydrogenase, encoded by ALD2 and ALD3 (Navarro-Avino
et al., 1999; Erasmus et al., 2003; Pigeau and Inglis, 2005). In
order to maintain the intracellular redox balance, yeast cells
regenerate an equimolar amount of cytoplasmic NADH. This
requirement seems to be partially met by a decreased reduction
of acetaldehyde to ethanol, on the one hand, and an increased
oxidation to acetate, on the other (Blomberg and Alder, 1989).
Thus, it is clear that overproduction of glycerol and acetate is
unavoidable in high-sugar fermentation by S. cerevisiae (Caridi
et al., 1999). The low volatile acidity production of T.
delbrueckii and the non-negligible amount of glycerol suggest
that this species does not produce acetic acid as a by-product of
the hyper-osmotic stress response to high-sugar concentrations.
The direct consequence is a decrease in volatile acidity pro-
duction. Actually, this species is known to be highly-resistant to
stress conditions and is, therefore, often isolated in juice fruits
and bakery (Legan and Voysey, 1991), but there are no previous
findings concerning its resistance mechanism in high-sugar
media.
In both T. delbrueckii trials, acetaldehyde production was
lower than in pure S .cerevisiae fermentations, as reported by
Ciani and Picciotti (1995), Lema et al. (1996), Herraiz et al.
(1990), and Ciani and Maccarelli (1998), but contrary to more
recent findings (Ciani et al., 2006). Acetaldehyde is an important
fermentation by-product, which affects wine quality. Barbe et al.
(2000) stated that a high concentration of this compound in wine
frombotrytized grapes reduced the amount of free SO
2
, an active
antimicrobial agent used to stop the fermentation. T. delbrueckii
27828 and T. delbrueckii 31703 produced 4.4- and 3-fold less
acetaldehyde, respectively, than S. cerevisiae ST. It is worth
pointing out that the S. cerevisiae ST strain was selected for its
low volatile acidity and acetaldehyde production (Barbe et al.,
2000, Masneuf and Dubourdieu, 2000).
As expected, inoculating high-sugar musts with T. del-
brueckii yeast resulted in stuck fermentations. Although this
yeast species had a low ethanol yield, it had the best secondary
metabolite profile. Low acetic acid and acetaldehyde production
is a positive feature for fermenting botrytized must, so a com-
bination of T. delbrueckii species and S. cerevisiae was envis-
aged to improve wine quality, while ensuring that fermentation
would be completed. T. delbrueckii 31703 was chosen for the
following study as it had a higher ethanol production.
3.2. Fermentation behaviour of mixed culture
Pure and mixed fermentations were carried out using T.
delbrueckii 31703 and S. cerevisiae ST in pasteurized (PM) or
non-pasteurized must (NPM). Non-pasteurized musts were used
to mimic winemaking practices as closely as possible. Pure
cultures were inoculated with 2

10
6
cells/mL. Mixed fermenta-
tion trials were inoculated with 5.10
6
cells/mL of T. delbrueckii
31703 and 10
6
cells/mL of S. cerevisiae ST (ratio T. delbrueckii/
S. cerevisiae 5:1).
Yeast growth was monitored by plate counting. Non-Sac-
charomyces cells (corresponding to T. delbrueckii 31703 in
pasteurized must) were counted using lysine agar (LA). The
number of Saccharomyces cells was calculated as the difference
between the total plate count using YPDA and LA. In
pasteurized must, the estimated Saccharomyces count corre-
sponded to the S. cerevisiae ST population. The Saccharomyces
and non-Saccharomyces populations in the non-pasteurized
must were estimated at 2

10
5
cfu/mL and 1.4

10
4
cfu/mL,
respectively.
The required amounts of ethanol (up to 14%vol.) were ob-
tained in all fermentations except the pure T. delbrueckii 31703
culture in PM, which stopped fermenting at 9.7%vol. (not
shown). This result suggested that the indigenous yeasts
contributed sufficiently to fermentation behaviour to produce
the amount required of ethanol for the pure culture of T.
delbrueckii 31703 in NPM. Fermentation time varied, depend-
ing on the must treatment and yeast culture (Fig. 2). The fastest
fermentation occurred in the pure S. cerevisiae ST culture in
NPM, followed by S. cerevisiae ST in PM and mixed culture in
NPM, then mixed culture in PM. The pure T. delbrueckii 31703
cultures in NPM and PM took the longest to ferment (22 days).
As described above, the pure T. delbrueckii 31703 cultures
had the lowest maximum cell populations (O.D610 max.) and
the highest nitrogen residues (Table 2). However, nitrogen
consumption was higher in NPM, indicating competition be-
tween T. delbrueckii 31703 and the indigenous yeasts in the
NPM. On the contrary, in pure and mixed S. cerevisiae ST
cultures, irrespective of the must treatment, the nitrogen was
almost completely consumed and maximum cell populations
were higher.
Biomass evolutions during fermentation are shown in Fig. 3.
No significant differences in the Saccharomyces populations
were observed between the pasteurized and non-pasteurized
Table 1
Main oenological characters of pure fermentations at 23 C
S. cerevisiae ST T. delbrueckii T. delbrueckii
27828 31703
Max.cell population
(O.D
610
max)
12.4 4 8.4
Residual nitrogen (mgN/L) 12.001.41
c
77.005.66
b
100.502.12
a
Residual sugar (g/L) 106.331.15
c
219.633.79
a
187.0010.82
b
Ethanol (% vol.) 15.100.07
a
7.400.21
c
9.380.25
b
Glycerol (g/L) 17.240.34
a
14.050.92
b
15.950.07
a
Volatile acidity
(g/L acetic acid)
1.170.08
a
0.430.01
c
0.620.01
b
Acetaldehyde (mg/L) 80.716.31
a
18.43.50
b
27.003.00
b
Yield ethanol/sugar (g/g) 0.48 0.43 0.46
Initial sugar concentration: 360 g/L. Initial nitrogen concentration: 190mgN/L.
Initial glycerol concentration: 5.7 g/L.
Means of triplicate fermentations SD. Values followed by different superscript
letters, within each line, are different according to the Newman-Keuls test (=5%).
315 M. Bely et al. / International Journal of Food Microbiology 122 (2008) 312320
treatments (Fig. 3a). The maximum concentration of Sacchar-
omyces cells varied from 6.9 to 7.9 Log (cfu/mL), with the
highest population in pure S. cerevisiae ST cultures (in NPM or
PM) and the lowest in NPM inoculated with pure T. delbrueckii
31703. Intermediate populations were observed in mixed cultures,
suggesting possible competition between non-Saccharomyces
and Saccharomyces species, at critical nitrogen concentrations,
which restricted Saccharomyces cell growth.
The biomass evolutions of non-Saccharomyces species
(Fig. 3b) revealed similar maximum cell populations, 7.5 Log
(cfu/mL), except for the pure S. cerevisiae ST inoculum. In PM,
the pure T. delbrueckii 31703 culture slowly began to die off
after the level reached 6%vol. (corresponding to 12 g/bottle
CO
2
released), while in NPM the non-Saccharomyces yeasts
population subsisted at a higher level, 6.6 Log(cfu/mL), until
fermentation was stopped. It is noteworthy that, in PM the pure
T. delbrueckii 31703 culture stopped fermenting when 20 g/
bottle CO
2
had been released (at a cell concentration of 6.6 Log
cfu/mL), while, in mixed culture, it was still present in similar
cell concentrations at the end of fermentation. This suggests
that, although T. delbrueckii 31703 is a low ethanol producer, it
is able to survive at high ethanol concentrations, up to 14%vol.
In NPM, the non-Saccharomyces cells in the mixed culture were
present at maximum concentrations throughout fermentation.
These findings show that non-Saccharomyces species can
survive at significant levels during high-sugar fermentation,
irrespective of the must treatment. In mixed culture in NPM, the
non-Saccharomyces and Saccharomyces cell populations
reached similar levels at the end of fermentation. The pure S.
cerevisiae culture did not suppress the growth of indigenous
non-Saccharomyces yeasts. The reduced growth of Saccharo-
myces species in mixed cultures, was the most significant
Table 2
Main oenological characters of pure and mixed fermentations at 23 C
Culture trial Max.cell population Residual nitrogen Glycerol Volatile acidity
O.D
610
max. (mgN/L) (g/L) (g/L acetic acid)
Pasteurized must Pure S. cerevisiae (S) 13.8 16.670.58
b
17.070.81
a
1.000.06
a
Pure T. delbrueckii (T)

8.4 130.53.54
a
14.700.53
b
0.560.10
b
Mixed S /T 14.2 11.001.00
c
16.170.23
a
0.730.09
b
Non-pasteurized must Pure S. cerevisiae (S) 14.8 25. 330.58
b
17.050.35
a
0.940.06
a
Pure T. delbrueckii (T) 9.6 68.3310.02
a
17.031.25
a
0.790.02
b
Mixed S/T 15.0 21.000.00
b
16.400.99
a
0.750.11
b
Initial sugar concentration: 360 g/L. Initial nitrogen concentration: 190mgN/L. Initial glycerol concentration: 6.6 g/L.

Stuck fermentation.
Means of triplicate fermentations SD. Values followed by different superscript letters, within each column, are different according to the Newman-Keuls test
(=5%).
Fig. 2. Fermentation courses at 23 C. Pasteurized (open symbols) or non-
pasteurized must (filled symbols). Pure S. cerevisiae ST culture (S). Pure T.
delbrueckii culture (T). Mixed T. delbrueckii 31703/S. cerevisiae ST (T/S).
Means of triplicate fermentations, max. SD0.9.
Fig. 3. Evolutions of the biomass in function of CO
2
released at 23 C. Sac-
charomyces populations (Fig. 3a) or non-Saccharomyces populations (Fig. 3b).
Pasteurized (open symbols) or non-pasteurized must (filled symbols). Pure S.
cerevisiae ST culture (S). Pure T. delbrueckii 31703 culture (T). Mixed T.
delbrueckii 31703/S. cerevisiae ST (T/S). Means of duplicate fermentations,
max. SD0.15.
316 M. Bely et al. / International Journal of Food Microbiology 122 (2008) 312320
finding. These results differ from those reported by Nissen et al.
(2004); Nissen and Arneborg (2003). In their research, the
authors found that, in mixed cultures, T. delbrueckii arrested
growth earlier than S. cerevisiae. This phenomenon was ap-
parently due to a cellcell contact mechanism, dependent on the
presence of high concentrations of viable S. cerevisiae cells.
These divergences were probably due to significant differ-
ences in fermentation conditions: nitrogen, sugar, and inoculum
concentrations.
The main oenological characteristics of the fermentations
are presented in Table 2. The amount of glycerol was similar,
except in the stuck fermentation with the pure T. delbrueckii
31703 culture in PM.
There were no significant differences between the pasteur-
ized and non-pasteurized treatments in terms of volatile acidity
production, except for the T. delbrueckii 31703 trial in PM.
Volatile acidity concentrations in the wine varied considerably,
ranging from 0.56 to 1 g/L acetic acid. Inoculation with a pure
S. cerevisiae ST culture always produced more volatile acidity,
irrespective of the must treatment. On the contrary, pure or
mixed cultures with T. delbrueckii 31703 always produced less
volatile acidity than pure S. cerevisiae ST. The volatile acidity
production kinetics for one of the triplicate fermentations are
shown in Fig. 4, revealing that the differences in final volatile
production between the two pure cultures, mentioned above,
were present from the onset of fermentation. In fact, in the pure
S. cerevisiae ST culture, over 35% of the final amount was
produced during the first few hours, when under 0.2 g/bottle
CO
2
had been released. In contrast, in the pure T. delbrueckii
31703 culture, volatile acidity was produced gradually through-
out fermentation. Intermediate behaviour was observed in
mixed cultures, irrespectively of the must treatment. After a
burst of volatile acidity production at the beginning of the
fermentation, concentrations remained constant until 5 g/bottle
CO
2
had been released, then increased slowly until the end of
fermentation. Note that, after the first third of the fermentation
period, the slopes of the volatile acidity/CO
2
curves were simi-
lar, regardless of the type of culture. All these results confirmed
that T. delbrueckii 31703 does not response to increased
external osmolarity in the same way as S. cerevisiae ST, i.e. by
producing volatile acidity as soon as it is inoculated into a high-
sugar must. The direct consequence is lower volatile acidity
production than in pure S. cerevisiae ST cultures. In mixed
cultures, the high T. delbrueckii 31703 concentration reduced
acetate production during the first stage of fermentation.
The advantage of mixed culture was demonstrated in this
experiment, as volatile acidity production was reduced, while
alcoholic fermentation was successfully completed in both NPM
and PM. However, only one T. delbrueckii 31703/S. cerevisiae
ST ratio was tested, so further experiments were carried out to
optimized this ratio.
3.3. Optimizing the mixed culture ratio
Pure cultures were carried out using pasteurized (PM) and
non-pasteurized must (NPM). In order to mimic winemaking
practice as closely as possible, mixed and sequential fermenta-
tions were carried out in non-pasteurized must. The initial
Saccharomyces and non-Saccharomyces populations of the
must were estimated at 2

10
4
cfu/mL and 2

10
3
cfu/mL,
respectively. Pure cultures were inoculated with 2

10
6
/mL.
Mixed fermentation trials were simultaneously inoculated with
10
7
cells/mL T. delbrueckii 31703, together with S. cerevisiae
ST in ratios ranging from 5:1 to 100:1. Sequential fermentations
were inoculated with 2

10
6
cells/mL T. delbrueckii 31703 and
2

10
6
cells/mL S. cerevisiae ST after 5 days' fermentation,
when 5%vol. of ethanol had been produced. CO
2
release and the
main oenological parameters are reported in Figs. 5 and 6 and
Table 3.
The required amounts of ethanol (14%vol.) were obtained in
all fermentations except the pure T. delbrueckii culture in PM and
NPM, and the sequential fermentation (Table 3), which stopped at
6.92, 10.47, and 13.15%vol., respectively. Fermentations using T.
delbrueckii 31703, in pure or mixed cultures, had lower ethanol/
sugar yields than those with S. cerevisiae STalone. Fermentation
rates varied depending on the must treatment (Fig. 5): fermenta-
tions were slower in PM than NPM. This suggested that the
indigenous yeasts contributed to fermentation kinetics, but did not
Fig. 4. Volatile acidity productions in function of CO
2
released, at 23 C.
Pasteurized (open symbols) or non-pasteurized must (filled symbols). Pure S.
cerevisiae ST culture (S). Pure T. delbrueckii 31703 culture (T). Mixed T.
delbrueckii 31703/S. cerevisiae ST (T/S). Means of duplicate fermentations.
max. SD0.06.
Fig. 5. Fermentation courses at 23 C. Pasteurized (open symbols) or non-
pasteurized must (filled symbols). Pure S. cerevisiae ST culture (S). Pure T.
delbrueckii culture (T). Means of triplicate fermentations, max. SD0.6.
317 M. Bely et al. / International Journal of Food Microbiology 122 (2008) 312320
improve them sufficiently for the pure T. delbrueckii 31703
culture to produce the required amount of ethanol, which differs
from the previous result. This phenomenon was due to the lower
indigenous yeast concentration in this must (2.2

10
4
cfu/mL
compared to 2.1

10
5
cfu/mL).
In NPM, all mixed cultures had longer fermentation times,
varying from 15 to 32 days, compared with 11 days for the pure
S. cerevisiae culture (Fig. 6). The shortest multistarter fermen-
tation was obtained with the highest concentrations of S.
cerevisiae ST in the mixed inoculum.
As expected, regardless of the must treatment, the T.
delbrueckii culture had the lowest maximum cell population
and the highest residual nitrogen. On the other hand, mixed
cultures produced higher cell concentrations, reaching the same
levels as the pure S. cerevisiae ST culture at a T. delbrueckii
31703/S. cerevisiae ST ratio of 5:1. The nitrogen residues
decreased in inverse proportion to the S. cerevisiae ST concen-
tration in the inoculum. The pure T. delbrueckii 31703 and
sequential culture exhibited similar behaviour, suggesting that
S. cerevisiae ST was unable to use the nitrogen available in
the must when it was introduced after 5 days' fermentation with
T. delbrueckii 31703. The direct consequence was reduced
growth cell, resulting in stuck fermentation. Further investiga-
tion is required to clarify the mechanisms causing this reduction
in S. cerevisiae ST fermentation capacity (presence of toxic
compounds, nutrient depletion, etc.).
Glycerol concentrations in wines fermented with pure S.
cerevisiae STand mixed cultures did not differ significantly. On
the contrary, pure T. delbrueckii 31703 and the sequential fer-
mentations produced less glycerol.
Once again, volatile acidity production by T. delbrueckii
31703 in PM was very low compared to that of S. cerevisiae ST.
In NPM, in both mixed and sequential cultures values were
lower than those in pure S. cerevisiae culture, with a signifi-
cant difference at 95% probability. Ciani et al. (2006) recently
reported similar results in pasteurized must containing less
sugar. There were similar reductions in mixed cultures at ratios
of 5:1 to 20:1, with up to 55% less volatile acidity in the wine. In
sequential culture, volatile acidity production was only reduced
by 37%. T. delbrueckii 27828 strain was also tested in same
conditions and comparable results were observed (not shown).
The significant variations were observed in acetaldehyde
production, but these were not correlated with volatile acidity
levels. Concentrations varied from to 14.77 to 45.63 mg/L, with
the highest concentrations in both pure cultures of S. cerevisiae
ST and T. delbrueckii 31703 in NPM. Acetaldehyde production
was correlated with the proportion of S. cerevisiae ST in the
mixed cultures. Up to 68% less acetaldehyde was produced in
mixed cultures with a T. delbrueckii 31703/S. cerevisiae ST
ratio of 50:1. Recently, Cheraiti et al. (2005) showed up ac-
etaldehyde exchanges between S. cerevisiae and S. uvarum in
mixed culture under oenological conditions. Further investiga-
tion is needed in our conditions.
The results show that, unlike pure T. delbrueckii, mixed
cultures were capable of completing fermentation. Sequential
cultures gave the least positive results, in terms of volatile
acidity and acetaldehyde production, while mixed cultures were
Table 3
Main oenological characters of pure, mixed and sequential fermentations at 23 C
Culture trial Max.cell
population
Residual
nitrogen
Ethanol Glycerol Volatile acidity Acetaldehyde Yield
ethanol
A610max (mgN/L) (% vol.) (g/L) (g/L acetic acid) (mg/L) (g/g)
Pasteurized must Pure S. cerevisiae 11.8 16.002.00 14.230.11 15.800.36 1.050.02 38.505.33 0.47
Pure T. delbrueckii

6.2 107.0010.00 6.920.33 10.931.02 0.270.10 25.205.00 0.44

Non-pasteurized must Pure S. cerevisiae 11.2 21.672.08
d
14.100.35
a
15.601.06
a
0.890.16
a
45.635.86
a
0.48
Pure T. delbrueckii

7.2 79.009.54
a
10.470.35
c
13.470.38
b
0.470.04
cde
41.833.06
a
0.45
Mixed T/S 5:1 11.1 23.671.53
d
14.270.35
a
16.000.10
a
0.410.03
e
30.904.00
b
0.46
Mixed T/S 10:1 10.8 28.330.58
cd
14.430.04
a
15.900.30
a
0.400.03
e
25.903.90
b
0.47
Mixed T/S 20:1 11.0 35.002.65
c
13.920.08
a
15.070.91
a
0.420.04
de
18.201.47
c
0.47
Mixed T/S 50:1 9.4 45.003.61
b
14.150.10
a
15.300.44
a
0.500.01
bcd
14.770.81
c
0.47
Mixed T/S 100:1 9.4 48.674.93
b
14.080.08
a
15.901.51
a
0.520.05
bc
16.470.12
c
0.47
Sequential T/S

nd 85.005.66
a
13.150.41
b
13.170.15
b
0.560.04
b
20.301.84
c
0.46
Initial sugar concentration: 360 g/L. Initial nitrogen concentration: 190mgN/L. Initial glycerol concentration: 5.5 g/L.

Stuck fermentation.
S. cerevisiae ST (S); T. delbrueckii 31703 (T); Mixed T. delbrueckii 31703/S. cerevisiae ST (T/S) with 10
7
T. delbrueckii cells/mL.
Means of triplicate fermentations SD. Values followed by different superscript letters, within each column, are different according to the Newman-Keuls test
(=5%).
Fig. 6. Fermentation courses, in non-pasteurized must, at 23 C. Pure S.
cerevisiae ST culture (S). Pure T. delbrueckii 31703 culture (T). Mixed T.
delbrueckii 31703/S. cerevisiae ST (T/S) with 10
7
T. delbrueckii 31703 cells/mL
Means of triplicate fermentations, max. SD0.6.
318 M. Bely et al. / International Journal of Food Microbiology 122 (2008) 312320
the most advantageous, irrespective of the S. cerevisiae con-
centration. Final volatile acidity and acetaldehyde concentra-
tions were up to 55% and 68% lower, respectively. Considering
the reduction in both volatile acidity and acetaldehyde pro-
duction, the best multistarter was T. delbrueckii 31703 mixed
with S. cerevisiae at a ratio of 20:1, with 10
7
T. delbrueckii
cells/mL.
4. Conclusion
This study of fermentation in high-sugar must shows that T.
delbrueckii is characterized by pure fermentation, with very
low volatile acidity and acetaldehyde production, confirming
previous studies under standard winemaking conditions. Even
if it is a low ethanol producer, T. delbrueckii still has useful
potential in sweet wine fermentation, as it does not seem to
respond to osmotic stress in the same way as S. cerevisiae.
Volatile acidity production remains constant throughout fer-
mentation, in contrast to S. cerevisiae, where over 35% of the
total production occurs in the initial stage of fermentation.
Several studies have linked this overproduction by S. cerevisiae
to increased glycerol production, induced by osmotic stress.
One solution to the problem of excessive volatile acidity
formation in high-sugar fermentations was to use mixed cultures
of T. delbrueckii and S. cerevisiae, with a higher concentration of
T. delbrueckii to promote its growth. This mixed inoculumresults
in combined, rather than successive, fermentations, producing
lower levels of acetic acid and acetaldehyde without affecting
the glycerol content. Reduction of the final volatile acidity and
acetaldehyde productions up to 55% and 68% were obtained,
respectively. This phenomenon depends on biochemical mechan-
isms involving both yeast metabolisms. We show that a high
population of T. delbrueckii in mixed culture reduces the acetate
production during the first stage of fermentation. Further inves-
tigations are required to clarify the role of mixed culture in
reducing volatile acidity production: could it be that the reduced
growth of the S. cerevisiae induced by a high concentration of
T. delbrueckii, may explain the acetate reduction or are there any
acetate exchanges between the two yeasts?
Acknowledgements
The authors thank Chteau Rayne Vigneau for its collab-
oration and acknowledge SARCO for technical assistance.
References
Albertyn, J., Hohmann, S., Thevelein, J.M., Prior, B.A., 1994. GPD1, which
encodes glycerol-3-phosphate dehydrogenase is essential for growth under
osmotic stress in Saccharomyces cerevisiae and its expression is regulation
by the high-osmolarity glycerol response pathway. Molecular and Cellular
Biology 14, 41354144.
Antunovics, Z., Csoma, M., Sipiczki, M., 2003. Analyse molculaire et
gntique de la flore levurienne des vins botrytiss de Tokay. Bulletin de
l'Office International de la Vigne et du Vin 867, 383397.
Attfied, P.V., Kletsas, S., 2000. Hyperosmotic stress response by strains of bakers'
yeasts in high sugar concentration medium. Letters in Applied Microbiology
31, 323327.
Barbe, J.C., de Revel, G., Joyeux, A., Lonvaud-Funel, A., Bertrand, A., 2000.
Role of carbonyl compounds in SO
2
binding phenomena in musts and wines
from botrytized grapes. Journal of Agricultural and Food Chemistry 48,
34133419.
Bely, M., Rinaldi, A., Dubourdieu, D., 2003. Influence of assimilable nitrogen
on volatile acidity production by Saccharomyces cerevisiae during high
sugar fermentation. Journal of Bioscience and Bioengineering 96, 507512.
Bely, M., Masneuf-Pomarde, I., Dubourdieu, D., 2005. Influence of physi-
ological state of inoculum on volatile acidity production by Saccharomyces
cerevisiae during high sugar fermentation. Journal International des
Sciences de la Vigne et du Vin 39, 191197.
Blomberg, A., Alder, L., 1989. Role od glycerol and glycerol-3-phosphate de-
hydrogenase (NAD+) in acquired osmotolerance of Saccharomyces cerevisiae.
Journal of Bacteriology 171, 10871092.
Blomberg, A., Alder, L., 1992. Physiology of osmotolerance in fungi. Advanced
in Microbiology and Physiology 33, 145212.
Cabrera, M.J., Moreno, J., Ortega, J.M., Medina, M., 1988. Formation of
ethanol, higher alcohols, esters and terpenes by five yeast strains in must
from pedo Ximenez grapes in various degrees of ripeness. American Journal
of Enology and Viticulture 39, 283287.
Caridi, A., Crucitti, P., Ramondino, D., 1999. Winemaking of must at high osmotic
strength by thermotolerant yeasts. Biotechnology Letters 21, 617620.
Ciani, M., 1997. Role, enological properties and potential use of non-Saccharomyces
wine yeast. Recent Research and Development in Microbiology 1, 317331.
Ciani, M., Picciotti, G., 1995. The growth kinetics and fermentation behaviour
of some non-Saccharomyces yeasts associated with wine-making. Biotech-
nology Letters 17, 12471250.
Ciani, M., Ferraro, L., 1998. Combined used of immobilized Candida stellata
cells and Saccharomyces cerevisiae to improve the quality of wines. Journal
of Applied Microbiology 85, 247254.
Ciani, M., Maccarelli, F., 1998. Oenological properties of non-Saccharomyces
yeasts associated with wine-making. World Journal of Microbiology and
Biotechnology 14, 199203.
Ciani, M., Beco, L., Comitini, F., 2006. Fermentation behaviour and metabolic
interactions of multistarter wine yeast fermentations. International Journal of
Food Microbiology 108, 239245.
Cheraiti, N., Guezenec, S., Salmon, J.M., 2005. Redox interactions between
Saccharomyces cerevisiae and Saccharomyces uvarum in mixed culture
under enological conditions. Applied and Environmental Microbiology 71,
225260.
Dukes, B.C., Buzkle, C.E., 1998. Rapid determination of primary amino acids in
must using an OPA/NAC spectrophotometric assay. American Journal of
Enology and Viticulture 49, 125133.
Egli, C.M., Edinger, W.D., Mitrakul, C.M., Henick-Kling, T., 1998. Dynamics
of indigenous and inoculated yeast populations and their effect on the
sensory character of Riesling and Chardonnay wines. Journal of Applied
Microbiology 85, 779789.
Erasmus, D.J., van der Merwe, G.K., van Vuuren, H.J.J., 2003. Genome-wide
expression analyses: metabolic adaptation of Saccharomyces cerevisiae to
high sugar stress. FEMS Yeast Research 3, 375399.
Erasmus, D.J., Cliff, M., van Vuuren, H.J.J., 2004. Impact of yeast strain on the
production of acetic acid, glycerol, and the sensory attributes of Icewine.
American Journal of Enology and Viticulture 55, 371378.
Fleet, G.H., 2003. Yeast interactions and wine flavour. International Journal of
Food Microbiology 86, 1122.
Fleet, G.H., Lafon-Lafourcade, S., Ribreau-Gayon, P., 1984. Evolution of
yeasts and lactic acid bacteria during fermentation and storage of Bordeaux
wines. Applied and Environmental Microbiology 48, 10341038.
Heard, G.M., Fleet, G.H., 1985. Growth of natural yeast flora during the
fermentation of inoculated wines. Applied and Environmental Microbiology
50, 727728.
Heard, G.M., Fleet, G.H., 1986. Occurrence and growth of yeast species during
the fermentation of some Australian wines. Food Technology in Australia
38, 2225.
Herraiz, T., Reglero, G., Herraiz, M., Martin-Alvarez, P.J., Cabezudo, M., 1990.
The influence of the yeast and type of culture on the volatile composition of
wine fermented without sulphur dioxide. American Journal of Enology and
Viticulture 41, 313318.
319 M. Bely et al. / International Journal of Food Microbiology 122 (2008) 312320
HolmHansen, E., Nissen, P., Sommer, P., Nielsen, J.C., Arneborg, N., 2001. The
effect of oxygen on the survival of non-Saccharomyces yeasts during mixed
culture fermentations grape juice with Saccharomyces cerevisiae. Journal of
Applied Microbiology 91, 541547.
Laffon-Lafourcade, S., Lucmaret, V., Joyeux, A., Ribreau-Gayon, P., 1981.
Utilisation de levains mixtes dans l'laboration des vins de pourriture
noble, en vue de rduire l'acidit volatile. Compte Rendu de l'Acadmie
d'Agriculture de France 67, 616622.
Legan, J.D., Voysey, P.A., 1991. Yeast spoilage of bakery products and ingre-
dients. Journal of Applied Bacteriology 70, 361371.
Lema, C., Garcia-Jares, C., Orriols, I., Angulo, L., 1996. Contribution of Sac-
charomyces and non-Saccharomyces populations to the production of some
components of Albarino wine aroma. American Journal of Enology and
Viticulture 47, 206216.
Martinez, J., Toledano, F., Millan, C., 1990. Development of alcoholic fermen-
tation in non-sterile musts from Pedro Ximenez grapes inoculated with pure
cultures of selected yeasts. Food Microbiology 7, 217225.
Masneuf, I., Dubourdieu, D., 2000. Rle de la souche de levure sur les com-
binaisons du dioxyde de souffre des vins issus de raisins botrytiss et
passerills. Journal International Sciences de la Vigne et du Vin 34, 2731.
Mauricio, J.C., Guijo, S., Ortega, J.M., 1991. Relationship between phospho-
lipid and sterol contents in Saccharomyces cerevisiae and Torulaspora
delbrueckii and their fermentation activity in grape musts. American Journal
of Enology and Viticulture 42, 301308.
Moreno, J.J., Millan, C., Ortega, J.M., Medina, M., 1991. Analytical differ-
entiation of wine fermentations using pure and mixed yeast cultures. Journal
of Food Microbiology 7, 181190.
Myers, D.K., Lawlor, D.T.M., Attfield, P.V., 1997. Influence of invertase
activity and glycerol synthethis and retention on fermentation of media with
a high sugar concentration by. Applied and Environmental Microbiology 63,
145150.
Naumov, G.I., Naumova, E.S., Antunovics, Z., Sipiczki, M., 2002. Saccharo-
myces bayanus var.uvarum in Tokaj winemaking of Slavakia and Hungary.
Applied Microbiology and Biotechnology 59, 727730.
Navarro-Avino, J.P., Prasad, R., Miralles, V.J., Benito, R.M., Serrano, R., 1999.
A proposal for nomenclature of aldehyde dehydrognase in Saccharomyces
cerevisiae and characterization of stress inducible ALD2 and ALD3 genes.
Yeast 15, 829842.
Nissen, P., Arneborg, N., 2003. Characterization of early deaths of non-
Saccharomyces yeasts in mixed cultures with Saccharomyces cerevisiae.
Archives of Microbiology 180, 257263.
Nissen, P., Nielsen, D., Arneborg, N., 2004. The relative glucose uptake abilities
of non-Saccharomyces yeasts play a role in their coexistence with Sacchar-
omyces cerevisiae in mixed cultures. Applied Microbiology and Biotechnol-
ogy 64, 543550.
Pardo, I., Garcia, M.J., Zuniga, M., Urubury, F., 1989. Dynamics of microbial
populations during fermentation of wines from Utiel-Requena region of
Spain. Applied and Environmental Microbiology 55, 539541.
Pigeau, G.M., Inglis, D.L., 2005. Upregulation of ALD3 and GPD1 in Sac-
charomyces cerevisiae during icewine fermentation. Journal of Applied
Microbiology 99, 112125.
Pina, C., Santos, C., Couto, J.A., Hogg, T., 2004. Ethanol tolerance of five non-
Saccharomyces wine yeasts in comparison with a strain of Saccharomyces
cerevisiaeinfluence of different culture conditions. Food Microbiology
21, 439447.
Romano, P., Fiore, C., Paraggio, M., Caruso, M., Capece, A., 2003. Function of
yeast species and strains in wine flavour. International Journal of Food
Microbiology 86, 169180.
Sipiczki, M., 2003. Candida zemplinina sp.nov., an osmotolerant and psy-
chrotolerant yeast that ferments sweet botrytzed wines. International Journal
of Systematic and Evolutionary Microbiology 53, 20792083.
Sipiczki, M., Romano, P., Lipani, G., Miklos, I., Antunovics, Z., 2001. Analysis
of yeasts derived from natural fermentation in a Tokay winery. Antonie van
Leeuwenhoek 79, 97105.
320 M. Bely et al. / International Journal of Food Microbiology 122 (2008) 312320