Journal of Bioscience and Bioengineering

VOL. 115 No. 1, 43e49, 2013

www.elsevier.com/locate/jbiosc

Kinetics of D-glucose and D-fructose conversion during the alcoholic fermentation

promoted by Saccharomyces cerevisiae
Angela Zinnai, Francesca Venturi, Chiara Sanmartin, Mike F. Quartacci, and Gianpaolo Andrich*
Department of Crop Plant Biology, University of Pisa, Via del Borghetto 80, I-56124 Pisa, Italy
Received 11 June 2012; accepted 10 August 2012
Available online 15 September 2012

Although many studies on the different aspects of alcoholic fermentation are available in the literature, it is still
difcult to identify the possible causes of the slowing-down or stuck of fermentations, even if the change of some
compositional parameters (D-glucose/D-fructose and glycerine produced/hexoses converted ratios) could be assumed as
sound signals of a possible deviation from the usual Saccharomyces metabolic pathways. The reason why alcoholic yeasts
preferably metabolise D-glucose rather than D-fructose was investigated by a kinetic model based on six functional
parameters having a well-dened chemicalephysical meaning. The time evolution of different initial concentrations of
D-glucose and D-fructose, dissolved in a model solution simulating a must (citrate buffer at pH 3.4 inoculated by
a commercial strain of Saccharomyces cerevisiae), was investigated adding or not ethanol to the reaction medium. When
a reduced amount of ethanol was dissolved in the reaction medium, the time evolution of the fermentation rates of
these two sugars did not differ signicantly, to diversify rather strongly when the alcoholic concentration increased. The
hypothesised mathematical model accounts for this particular kinetic behaviour. In fact, only the sensitivity to ethanol
showed by the enzymatic protein involved in the limiting steps of the fermentation process of these two sugars differed
signicantly, the enzymatic transformation of D-fructose being more sensitive to ethanol than D-glucose. This difference
was able to justify the different kinetic behaviours shown by the two sugars when ethanol concentration in the reaction
medium increased.
2012, The Society for Biotechnology, Japan. All rights reserved.
[Key words: Hexose metabolisation; Kinetic model; Stuck of fermentation; Saccharomyces cerevisiae; Wine; Alcoholic fermentation; Yeast]

As widely reported in the literature, the lack of micro- and

macronutrients necessary for yeasts, unsuitable reaction temperatures, too low pH values, the presence of signicant concentrations
of inhibitors (ethanol, phenols, etc.) in the reaction medium, the
development of dangerous microorganisms as well as the alteration
of ionic equilibrium can induce a deep modication of the alcoholic
fermentation kinetics (1,2).
Over the past two decades, wine producers aimed to produce
grapes with increased sugar to total acid ratios to obtain higher
concentrations of phenols and aromatic compounds in order to
increase wine quality. As a consequence, the musts obtained by
these grapes are more difcult to process because they present
unsuitable conditions for yeast reproduction, particularly in the
warmer climates (3). The factors above cited could justify the
increasing number of stuck of fermentations that occurs in several
wine-producing countries, similarly to what observed for beer
fermentation (4).
The cause of the decline of the fermentation rate is not fully
understood. The increase in the alcoholic fermentation rate obtained by the addition of selected yeast strains to the must/wine
* Corresponding author. Tel.: 39 050 2216624; fax: 39 050 2216636.
E-mail addresses: azinnai@agr.unipi.it (A. Zinnai), fventuri@agr.unipi.it
(F. Venturi), csanmartin@agr.unipi.it (C. Sanmartin), mfquart@agr.unipi.it
(M.F. Quartacci), gandrich@agr.unipi.it (G. Andrich).

could often result ineffective if the residual sugars are used by

contaminating microorganisms able to carry on unwanted metabolic pathways (5,6). In these conditions some heterofermentative
lactic acid bacteria strains could signicantly increase volatile
acidity with a consequent loss of the sensory quality of the alcoholic beverage. Moreover, favourable conditions for yeasts lysis and
release of intracellular compounds e as it often occurs at the end of
the alcoholic fermentation when reduced amounts of molecular
SO2 are dissolved in the liquid phase e may strongly stimulate the
growth of Brettanomyces spp. (5).
Although many studies on the different aspects of alcoholic
fermentation are available in the literature, it is still difcult to
identify the possible causes of the slowing-down or stuck of
fermentations (7), even if the change of some compositional
parameters (D-glucose/D-fructose and glycerine produced/
hexoses converted ratios) or an accumulation of intermediates of
sugar catabolism could be assumed as sound signals of
a possible deviation from the usual Saccharomyces metabolic
pathways (6).
In particular, it is interesting to nd the reason why alcoholic
yeasts preferably metabolise D-glucose rather than D-fructose
(2,8e11), and to investigate the kinetic aspects related to their
conversion in order to optimise specic substrate consumption
rates. Moreover, this kinetic approach appears to be potentially able
to clarify many metabolic aspects related to hexoses conversion

1389-1723/$ e see front matter 2012, The Society for Biotechnology, Japan. All rights reserved.
http://dx.doi.org/10.1016/j.jbiosc.2012.08.008

44

ZINNAI ET AL.

J. BIOSCI. BIOENG.,

with the aim to better control alcoholic fermentation and to avoid

that unwanted dangerous transformations take place.
To reduce the large number of variables able to inuence the
kinetics of alcoholic fermentation, the time evolution of different
initial concentrations of D-glucose and D-fructose, dissolved in
a model solution simulating a must (citrate buffer at pH 3.4
inoculated by a commercial strain of Saccharomyces cerevisiae),
have been investigated in the presence or not of ethanol.
MATERIALS AND METHODS
The kinetic runs were carried out at 27.0  1.5 C using a 500 mL batch reactor. To
ensure anaerobic and sterilised conditions the whole experimental apparatus was
autoclaved and subjected to three cycles of vacuum following by replacement with
nitrogen sterilised by ltration. Thus, the presence of undesired microorganisms and
the aerobic utilisation of sugars by yeasts were ruled out.
The characteristics of this bioreactor, realised at the Department of Crop Plant
Biology of the University of Pisa, were already reported in a previous paper (12). The
fermentation temperature was maintained constant by a heat exchanger, whereas
the homogeneity of the reaction medium was ensured by a magnetic stirrer. The
bioreactor was initially lled with 250 mL of a citrate buffer aqueous solution
(pH 3.4) containing D-glucose and/or D-fructose (at three different concentrations:
900, 1111, and 1667 mmol L1) added or not with 1120 or 1340 mmol L1 of ethanol,
respectively, that was sterilised by ltration. To the reaction medium containing only
buffer and sugars (and ethanol when added) about 3.1 g (6.2 g/L) of a lyophilised
yeast commercial strain (S. cerevisiae Actiore BJL source p1, Laffort Oenologie) were
directly added to ensure a number of colony forming units (CFU) ranging from 1010
to 1011. This addition represented the initial time of all kinetic determinations.
The time evolution of both CFU and concentrations of both reagents (D-glucose
and/or D-fructose) and their products (glycerine and ethanol) were evaluated by
a total plate count or utilising specic commercial enzymatic kits (Megazyme),
respectively (13).
The identication of the best values to be assigned to the model parameters was
carried out by the specic statistical programme BURENL (14) able to identify in
a space of j-dimensions (where j is equal to the number of model parameters) the
minimum value of the F function, which is given by the sum of squares of differences
occurring among experimental (Yi, exper.) and calculated (Yi,calc.) data:
F

N 
X
2
Yi;calc:  Yi;exper:

(1)

i1

where N represents the total number of experimental determinations. The values

assumed by the model parameters at the minimum of the F function represent the
best values.
For each experimental run the calculation of the three parameters related to the
time evolution of the yeast cells (kY, kinetic constant of yeast population inactivation; [Y]t0, yeast density at the initial run time; KY$E, constant related to the
equilibrium occurring between alcoholic yeasts and ethanol) was carried out using
the experimental data deriving from the determination of the microbial density. To
evaluate the kinetic constants related to the time evolution of the hexose under
investigation ([H]t0, concentration of hexose initially added to the reaction
medium; kH, specic kinetic activity shown by a single cell; KH$E, constant related to
the equilibrium occurring between ethanol and the enzymatic protein (Penz.)
involved in the rate limiting step of sugar fermentation) the experimental data
concerning hexose (D-glucose) decrease and both ethanol and glycerine accumulations were used.

down (16e18). In addition, the alcohol level gradually increases

becoming toxic to the yeast cells, and the use of fructose is even
more compromised (15).
To avoid that yeast replication in the different experimental
conditions could signicantly affect the rate of sugar consumption
a high concentration of lyophilised yeasts (6.2 g compared to
0.2 g L1 suggested by the manufacturer) was initially added to the
reaction medium without any preventive rehydration. Thus, the
lack of oxygen and above all the absence of essential nutrients
signicantly reduced the yeast replication rate, while the high CFU
number ensured a sensible conversion of sugars added. The
absence of a sigmoidal evolution of the experimental data related to
sugar consumption validates the previous statement. This procedure was adopted to evidence whether ethanol accumulation could
be able to explain the different rates of utilisation of the two sugars.
According to the stoichiometry of alcoholic fermentation, the
sum of the analytical data related to the concentrations of unconverted sugars, accumulated glycerine and half of ethanol formed
did not vary signicantly with time, assuming values very close to
the initial concentration of sugar used. As a consequence, a possible
signicant accumulation of a different intermediate reaction can be
ruled out.
The analytical points describing the decrease of concentrations
of the two monosaccharides (D-glucose and D-fructose) as a function of fermentation time when initial concentrations of 200 g L1
(1111 mmol L1) were used are reported in Fig. 1. While during the
rst phases of the fermentative process no signicant differences
between the metabolisation rates of the two sugars could be
observed, when reaction time increased and ethanol accumulation
became more relevant D-fructose consumption run more slowly
compared to D-glucose, although it reached very low concentration
values, consistent with an almost complete conversion of this
monosaccharide. This phenomenon was more evident when an
initial value of 300 g L1 (1667 mmol L1) of the two sugars was
tested (Fig. 2). In these conditions also D-glucose needed a reaction
time more than double to reach concentration values close to zero
(200 vs. 80 h necessary with an initial amount of 1111 mmol L1).
D-fructose conversion became even more difcult and its concentration assumed with time an asymptotic minimum value so that
about 20% of the initial amount remained unconverted in the
reaction medium. According to the literature, these preliminary
results conrm that ethanol inhibited the activity of the yeast
microbial population which, in its turn, induced an analogous
reduction of the conversion of both sugars. When the concentration
of ethanol in the fermentation medium increased the conversion
rates of the two sugars diversied signicantly, D-fructose being
converted more slowly than D-glucose (Figs. 1 and 2).

RESULTS AND DISCUSSION

The kinetics of sugar utilisation by S. cerevisiae during fermentation is largely driven by sugar transport, and glucose is typically
consumed at a faster rate than fructose (15). According to the
literature the ability shown by the yeast population to metabolise
the two sugars largely depends on the temperature and the
compositions of culture media (sugar levels, D-glucose to D-fructose
ratio as well as yeast-assimilable nitrogen) (10,16). In particular, the
synthesis of proteins involved in the transport of sugars into the cell
deeply affects their following utilisation (15).
During winemaking, sugars are consumed mainly during the
stationary phase when the available nitrogen gradually becomes
less available. Since nitrogen is an essential nutrient involved in the
transport of sugars into the cell via protein synthesis, this partially
explains why both yeast replication and fermentation activity slow

FIG. 1. Time evolution of the ratio between the concentration of D-glucose and
1
D-fructose at a random time t t ([H]tt) and the initial value ([H]t0) of 200 g L
(1111 mmol L1).

VOL. 115, 2013

KINETICS OF HEXOSE CONVERSION IN WINE BY S. CEREVISIAE

dHtt =dt kP $Penz: tt $Htt

45
(8)

where kP is the kinetic constant, [Penz.]tt is the concentration of

the free form of the enzyme involved in the rate limiting step at the
reaction time t t, and [H]tt represents the concentration of the
hexose at the reaction time t t. If this inhibitory effect can be
regarded as competitive, the following equilibrium reaction can be
introduced:
E Penz: 4E$Penz:

(9)

and KH$E (KG$E for D-glucose and KF$E for D-fructose), the constant
related to this equilibrium, will be equal to:


KH$E E$Penz tt = Ett $Penz: tt
(10)
FIG. 2. Time evolution of the ratio between the concentration of D-glucose and
1
D-fructose at a random time t t ([H]tt) and the initial value ([H]t0) of 300 g L
(1667 mmol L1).

In the experimental conditions used, the density of total cells

([Y]tt) able to promote the alcoholic fermentation inevitably
decreased with time as the reaction medium was devoid of many
components essential for promoting the growth of the yeast population (e.g., nitrogen sources for protein synthesis). If a rst order
kinetic equation is used to describe the time evolution of microbial
inactivation, the following mathematical relation can be introduced:
dYtt =dt kY $Ytt

(2)

This differential equation can be easily integrated and the

following relation obtained:
Ytt Yt0 $eky$t

(3)

Ethanol (E) produced by the alcoholic fermentation might

interact with yeasts and induce their progressive inactivation by
the following equilibrium reaction:
*

Y E4Y$E

(4)

If KY$E represents the constant related to the proposed equilibrium [Y$E]/([Y*]$[E]), and [Y*] the density of yeast cells still active in
the alcoholic medium, the number of yeast cells inactivated by
ethanol will be equal to:
Y$Ett Y*tt $Ett $KY$E

(5)

Thus, the total number of microbial cells able to form colonies,

present in the reaction medium at a random time t t ([Y]tt), will
be equal to the sum of the active fraction ([Y*]tt) and the one
which interacted with ethanol ([Y$E]tt):
Ytt Y*tt Y$Ett Y*tt Y*tt $Ett $KY$E


Y*tt $ 1 Ett $KY$E

(6)

and the following relation can be obtained:





Y*tt Ytt = 1 Ett $KY$E Yt0 $eky$t = 1 Ett $KY$E
(7)

from which:
E$Penz: tt Penz: tt $Ett $KH$E

(11)

The total concentration of the enzyme dissolved in the reaction

medium will be given by the sum of the free enzyme fraction and
the one bound to ethanol:
Penz: tot;tt Penz: tt E$Penz: tt
Penz: tt Penz: tt $Ett $KH$E


Penz: tt $ 1 Ett $KH$E

(12)

As the total concentration of the enzyme involved in the limiting

step is proportional to the density of the active fraction of yeasts
([Y*]) present in the reaction medium at that time:


Penz: tot;tt kPenz $Y*tt Penz: tt $ 1 Ett $KH$E
(13)
it follows that:


Penz: tt kPenz $Y*tt = 1 Ett $KH$E

(14)

and the rate related to hexose conversion (d[H]tt/dt) will be

equal to:
dHtt =dt kP $Penz: tt $Htt


kP $kPenz $Y*tt = 1 Ett $KH$E $Htt

(15)

As the product of two constants is equal to a new constant

(kP$kPenz kH), the following expression can be obtained:


dHtt =dt kH $Y*tt = 1 Ett $KH$E $Htt

(16)

In order to describe the time evolution of the concentrations of

reagents (D-glucose and D-fructose) and their products (ethanol and
glycerine) as well as that of yeast density, the following system of
three equations (Eqs. 3, 17 and 18) should be solved:
 
Ytt Yt0 $eky$t 3




Y*tt Ytt = 1 2a0E $ Ht0  Htt $KY$E

(17)





dHtt =dt kH = 1 2a0E $ Ht0  Htt $KH$E $Y*tt $Htt
(18)

Ethanol, interacting with an enzymatic protein (Penz.) involved

in one of the several metabolic steps of sugar fermentation (active
transport into the cell, glycolysis, pyruvate decarboxylation and
acetaldehyde reduction), may give rise to an inactive form (E$Penz.)
and slow down the rate of a specic step so to make it the limiting
step of all the fermentative process. In this case, the metabolisation
rate (d[H]tt/dt) of the hexose under investigation would coincide with that of the limiting step, and the following equation could
be introduced:

where a0E represents the fraction of hexose converted to ethanol

(selectivity to ethanol), whereas the difference ([H]t0  [H]tt) is
the amount of sugar converted at the reaction time t t. As,
according to the stoichiometry of this fermentation, from every unit
of hexose converted 2 mol of alcohol are produced, the amount of
ethanol produced can be easily calculated:


Ett 2a0E $ Ht0  Htt

(19)

46

ZINNAI ET AL.

J. BIOSCI. BIOENG.,

To describe the decrease of hexose concentration and the

accumulation of both ethanol and glycerine with fermentation
time, the numerical integration of the differential equation introduced was carried out. According to this approach the following
relation can be written:

.
t2  t1
dHtt =dt Htt2  Htt1
n

0
kH = 1 2aE $ Ht0

o
 Htt1 $KH$E $Y*tt2 $Htt1
(20)
hence:
n

Htt2 Htt1  kH = 1 2a0 $ Ht0

o
 Htt1 $KH$E $Y*tt2 $Htt1 $t2  t1

(21)

If t2  t1 Dt, then t2 t1 Dt. Thus, the following sequence of

equations is involved in the iterative calculation connected to the
numerical integration:
t2 t1 Dt

(22)





Ytt2 Yt0 $eky$t2 = 1 2a0E $ Ht0  Htt1 $KY$E

(23)


o
n

Htt2 Htt1  kH = 1 2a0E $ Ht0  Htt1 $KH$E $Ytt2
 Htt1 $Dt

(24)



Ett2 2a0E $ Ht0  Htt2

(25)



GY tt2 a00Gly $ Ht0  Htt2

(26)

Htt1 Htt2

(27)

t1 t2
where [Y]tt is density of yeast viable cells present in a liter of
fermentation medium at a random time t t able to promote

hexose fermentation (CFU L1); [Y]t0 is density of yeast viable

cells initially (t 0) present in a liter of fermentation medium
(CFU L1); kY is kinetic constant related to yeast inactivation (h1);
t is reaction time; a0E is hexose fraction converted to ethanol;
[H]t0 is sugar concentration initially (t 0) present in the
reaction medium (mmol L1); [H]tt is sugar concentration
present in the reaction medium at a random time t t (mmol L1);
KY$E is constant related to the equilibrium occurring between
ethanol and alcoholic yeasts (L mmol1); kH is kinetic constant
related to hexose conversion (h1 L CFU1); KH$E is constant
related to the equilibrium occurring between ethanol and the
enzymatic protein (Penz.) involved in the limiting step of sugar
fermentation (L mmol1); and a00Gly is hexose fraction converted to
glycerine.
When a further decrease of the Dt value does not induce any
signicant difference on the time evolution of the active fraction of
yeast as well as on the concentrations of reagents and products, the
numerical integration can be assumed to supply a realistic picture
of the process. Thus, knowing the initial conditions (t1 0;
Ytt1 Yt0 ; Htt1 Ht0 ) it is possible to determine the
evolution of all the involved components ([Y]tt, [H]tt, [E]tt and
[G]tt).
The identication of the best values to assign to the eight
functional parameters involved in the proposed kinetic model was
carried out using the experimental data describing the time course
of yeast population ([Y]tt), hexose consumption ([H]tt), as well
as ethanol ([E]tt) and glycerine ([Gly]tt) accumulations (Figs. 1
and 2). The values of the eight functional parameters calculated
by BURENL are reported in Table 1 as a function of the hexose
utilised and of its initial concentration.
Although the kinetic constant of the time evolution of yeast
population (kY) was not affected by a high variability, the squares of
the correlation coefcients of the microbial evolution showed
a remarkable variation ranging from 0.30 to 0.96 (Table 1). The
number of yeast cells did not change signicantly during some
experimental runs and this could be the main reason of the relatively low values of the squares of correlation coefcients. On the
other hand, the high values assumed by the squares of correlation
2 , r 2 , r 2 and r 2 ) related to the time evolution of
coefcients (rG
F
E
Gly
D-glucose, D-fructose, ethanol and glycerine concentrations
(Table 1) conrm the suitability of the kinetic equations as well as
the validity of the hypothesis introduced.

TABLE 1. Values assigned to the functional parameters involved in the kinetic model used to describe alcoholic fermentation following the addition to the reaction medium of
glucose and fructose at two different initial concentrations in the absence of ethanol.
Parameter

kY
KY$E
kG
KG$E
kF
KF$E
rY2

(3.41  0.01)$1011
1023.9  0.1
e
0.90  0.01
0.14  0.01
(5.61  2.23)$107
(4.48  3.08)$106
(8.70  0.04)$1014
(0.97  0.08)$104
e
e
0.37

(1.90  0.01)$1011
1659.8  0.2
e
0.85  0.01
0.14  0.01
(5.61  4.01)$107
(4.48  3.36)$106
(7.23  0.03)$1014
(1.32  0.05)$104
e
e
0.96

(4.72  0.01)$1011
e
1159.2  0.1
0.90  0.01
0.14  0.01
(5.62  2.32)$107
(4.54  1.75)$106
e
e
(8.75  0.01)$1014
(9.57  0.03)$104
0.30

(1.63  0.01)$1011
e
1782.0  0.1
0.85  0.01
0.16  0.01
(5.61  2.23)$107
(4.54  2.15)$106
e
e
(6.73  0.02)$1014
(9.34  0.05)$104
0.89

[Y]t0
[G]t0
[F]t0

a0E
a00Gly

2
rG

0.95

0.98

rF2

0.97

0.98

rE2

0.99

0.98

0.99

0.99

2
rGly

0.96

0.98

0.93

0.94

a0E ,

hexose fraction converted to

Data are expressed as means  condence intervals (p 0.05). Y, yeast active fraction; G, glucose; F, fructose; E, ethanol; Gly, glycerine;
ethanol; a00Gly , hexose fraction converted to glycerine; kY, kinetic constant related to yeast inactivation; KY$E, constant related to the equilibrium occurring between ethanol and
yeasts; kG, kinetic constant related to glucose conversion; KG$E, constant related to the equilibrium occurring between ethanol and the enzymatic protein involved in the
limiting step of glucose fermentation; kF, kinetic constant related to fructose conversion; KF$E, constant related to the equilibrium occurring between ethanol and the
enzymatic protein involved in the limiting step of fructose fermentation; r, correlation coefcients related to the time evolution of D-glucose, D-fructose, ethanol and glycerine
concentrations.

VOL. 115, 2013

The kinetic constants representing the time evolution of the
yeast population (kY) did not change signicantly as a function of
the monosaccharide used (Table 1) and showed a similar sensitivity
to ethanol (compare the values of the KY$E constant). In all runs
a slight decrease of yeast populations was observed (from an initial
1010e1011 to a nal 108e109 CFU mL1). Also the kinetic constants
related to the fermentation rates of the two sugars (kG and kF) did
not vary signicantly and assumed very similar values. As easily
predictable, the production of ethanol (a0E ) and glycerine (a00Gly ) did
not change with the fermentation substrate. Only the sensitivity to
ethanol showed by the enzymatic protein involved in the limiting
steps of the two sugars differed signicantly, the enzymatic
transformation of D-fructose being more sensitive to ethanol than
D-glucose. This difference, by itself, would be able to justify the
different time evolutions shown by the two sugars as demonstrated
by the high degree of overlap between experimental and calculated
values (Figs. 3A, B and 4A, B) and by the high values (Table 1) of the
2 , r 2 , r 2 and r 2 ) connected to
squares of correlation coefcients (rG
F E
Gly
the reagents and products.
To verify how the presence of ethanol in the reaction medium
can justify the different experimental behaviours shown by the two
monosaccharides, ethanol was initially added to the aqueous
solution of the two sugars (1120 and 1340 mmol L1 for glucose and
fructose, respectively). Table 2 reports the values calculated for the
model parameters when yeasts were added to a hydroalcoholic
solution. The evolution with fermentation time of reagents and
products are shown in Fig. 5A, B and the good degree of overlap
between calculated and experimental values give a measure of the
capacity shown by the model to describe the time evolution of
reagents and products. Moreover, the high values of the squares of
correlation coefcients conrm the suitability of the mathematical
model to describe the time evolution of the species involved in the
fermentation process also when ethanol was initially added to the
reaction medium.

KINETICS OF HEXOSE CONVERSION IN WINE BY S. CEREVISIAE

47

FIG. 4. Experimental points (squares, sugar; triangles, ethanol; circles, glycerine) and
theoretic values calculated by the kinetic model as a function of fermentation time. (A)
[D-glucose]t0 1667 mmol L1; (B) [D-fructose]t0 1667 mmol L1.

TABLE 2. Values assigned to the functional parameters involved in the kinetic model
used to describe alcoholic fermentation following the addition to the reaction
medium of glucose and fructose in the presence of ethanol.
GE

Parameter

kY
KY$E
kG
KG$E
kF
KF$E
rY2

(8.75  0.02)$10
1099.5  0.1
e
1120.9  0.1
0.86  0.01
0.14  0.02
(6.40  3.27)$107
(5.49  2.32)$106
(8.55  0.03)$1014
(1.73  0.03)$104
e
e
0.45

(2.01  0.01)$1011
e
1094.2  0.2
1344.5  0.2
0.90  0.01
0.16  0.01
(6.40  4.23)$107
(4.51  3.32)$106
e
e
(8.34  0.03)$1014
(9.47  0.03)$104
0.76

2
rG

0.99

rF2

0.96

rE2

0.98

0.95

2
rGly

0.91

0.93

[Y]t0
[G]t0
[F]t0
[E]t0

a0E
a00Gly

FIG. 3. Experimental points (squares, sugar; triangles, ethanol; circles, glycerine) and
theoretic values calculated by the kinetic model as a function of fermentation time. (A)
[D-glucose]t0 1111 mmol L1; (B) [D-fructose]t0 1111 mmol L1.

FE
10

Data are expressed as means  condence intervals (p 0.05). Y, yeast active

fraction; G, glucose; F, fructose; E, ethanol; Gly, glycerine; a0E , hexose fraction converted to ethanol; a00Gly , hexose fraction converted to glycerine; kY, kinetic constant
related to yeast inactivation; KY$E, constant related to the equilibrium occurring
between ethanol and yeasts; kG, kinetic constant related to glucose conversion; KG$E,
constant related to the equilibrium occurring between ethanol and the enzymatic
protein involved in the limiting step of glucose fermentation; kF, kinetic constant
related to fructose conversion; KF$E, constant related to the equilibrium occurring
between ethanol and the enzymatic protein involved in the limiting step of fructose
fermentation; r, correlation coefcients related to the time evolution of D-glucose,
D-fructose, ethanol and glycerine concentrations.

48

ZINNAI ET AL.

J. BIOSCI. BIOENG.,
initially added to the reaction medium, are the average of only three
values.
The two kinetic constant kG and kF, which were expected to
assume different values as a function of the sugar utilised, did not
differ signicantly (Tables 1 and 2). This particular situation might
nd a possible explanation assuming that in the presence of very
low concentrations of ethanol, the limiting step of both monosaccharides remained the same, being connected to one of the
sugar fermentation reactions. Only when the concentration of
ethanol in the reaction increased signicantly, its effect on the
transformation rates of these two sugars became more relevant.
The hypothesised mathematical model accounts for this
particular kinetic behaviour. In fact, the kinetic constants related to
0
0
D-glucose (kG ) and D-fructose (kF ) conversions are equal to the ratio
between a constant kH connected to the limiting step common to
the two sugars, and the sum of 1 and the concentration of ethanol
multiplied by a constant the value of which varies signicantly as
a function of the monosaccharide:


k0G kH = 1 EtOHtt $KG$E
(28)
and


k0F kH = 1 EtOHtt $KF$E

(29)

At low ethanol concentrations (1 [EtOH]tt$KG$E w 1) the rate

limiting steps of the two sugars coincide, and the following situation occurs:


k0G kH = 1 EtOHtt $KG$E kH =f1 0$KG$E g kH
(30)
FIG. 5. Experimental points (squares, sugar; triangles, ethanol; circles, glycerine) and
theoretic values calculated by the kinetic model when ethanol was initially added to
the reaction medium. (A) [D-glucose]t0 1111 mmol L1 plus 1120 mmol L1 ethanol;
(B) [D-fructose]t0 1111 mmol L1 plus 1340 mmol L1 ethanol.

The parameters related to yeast evolution (kY and KY$E) and those
connected to the conversion of the two sugars (kG, KG$E, kF, and KF$E)
did not differ signicantly from the ones reported in Table 1.
Ethanol addition was able to modify the kinetics of fructose
fermentation immediately after its addition when the replication
rate of yeast cells can be surely disregarded. Thus, the effect of
ethanol is able to justify the alteration of fructose metabolisation
before a signicant modication of the protein synthesis responsible of an altered transport of this sugar could occur.
Table 3 reports the mean values of the parameters involved in the
kinetic model calculated by the elaboration of the experimental runs
previously reported. While for the rst four parameters (a0E , a00Gly , kY,
KY$E) the values represent the average of six data, the last two
(KG$E, KF$E), which are strongly related to the monosaccharide
TABLE 3. Values of the functional parameters involved in the kinetic model used to
describe alcoholic fermentation calculated by elaboration of experimental runs.
Parameter

a0E
a00Gly
kY
KY$E
kH
KG$E
KF$E



k0F kH = 1 EtOHtt $KF$E kH =f1 0$KG$E g kH

(31)

thus, according to what experimentally found, k0G k0F kH .

On the contrary, when ethanol concentration increases the rate
limiting steps related to the transformations of the two sugars are
connected to different partial transformations. In particular, the
rate limiting step concerning D-fructose conversion might be linked
either to the active transport of this sugar from the reaction
medium into the yeast cell or to its following isomerisation to
produce D-glucose. In this case:


k0G kH = 1 EtOHtt $KG$E skH

(32)



k0F kH = 1 EtOHtt $KF$E skH

(33)

Moreover, as the KF$E constant is greater than the one related to

(KF$E/KG$E 9.46$104/1.34$104 w 7), the ratio between
the two kinetic constants:

D-glucose




k0F =k0G 1 KG$E $EtOHtt
1 KF$E $EtOHtt



1 KG$E $EtOHtt
1 7$KG$E $EtOHtt

(34)

Units
0.88  0.05
0.15  0.02
(5.87  0.82)$107
(4.67  0.81)$106
(8.05  1.75)$1014
(1.34  1.11)$104
(9.46  0.34)$104

e
e
h1 L CFU1
L mmol1
h1 L CFU1
L mmol1
L mmol1

Data are expressed as means  condence intervals (p 0.05). Y, yeast active

fraction; G, glucose; F, fructose; E, ethanol; Gly, glycerine; a0E , hexose fraction converted to ethanol; a00Gly , hexose fraction converted to glycerine; kY, kinetic constant
related to yeast inactivation; KY$E, constant related to the equilibrium occurring
between ethanol and yeasts; kH, kinetic constant related to hexose conversion; KG$E,
constant related to the equilibrium occurring between ethanol and the enzymatic
protein involved in the limiting step of glucose fermentation; KF$E, constant related
to the equilibrium occurring between ethanol and the enzymatic protein involved in
the limiting step of fructose fermentation.

will decrease when the concentration of ethanol increases and thus

with fermentation time.
The mean values of the model parameters were used to calculated the theoretical evolution of the components involved in an
experimental run where active cells of the commercial strain of
S. cerevisiae (w7$1011 CFU L1) were added to an aqueous solution
containing an equal amount of the two sugars (900 mmol L1). The
evolution with time of experimental and calculated points are reported in Fig. 6. The high degree of overlap and the high values of
the squares of correlation coefcients of the three species involved
2 0:99; r 2 0:98; r 2 0:98) give a measure of the suitability
(rG
F
E
of the kinetic model, which could be effectively used to describe the
time evolution of reagents and products involved in the alcoholic
fermentation.

VOL. 115, 2013

FIG. 6. Time evolution of both theoretical and experimental points (squares, D-fructose;
rhombus, D-glucose; triangles, ethanol) of equivalent amounts (900 mmol L1) of
D-glucose and D-fructose, and the corresponding accumulation of ethanol. The theoretical values were calculated by the kinetic model and using the mean values of the
functional parameters reported in Table 3.

In conclusion, ethanol accumulation in the reaction medium

was able to justify the different metabolisation rates observed for
the two sugars (D-glucose and D-fructose) involved in the alcoholic
fermentation, while at low ethanol concentration the rates related
to the two sugar conversion did not differ signicantly. When
alcohol concentration increased the D-glucose metabolisation rate
became faster than fructose one. The hypothesised mathematical
model accounts for this particular kinetic behaviour. In particular,
at low ethanol concentration the two sugars have a same limiting
step located in the common part of their pathways, while when
alcohol concentration increased the rate limiting step of fructose
metabolisation appears to be located in a step different of that of
D-glucose. This step is likely located along the active transport of the
sugar through the cellular membranes of yeasts or in its isomerisation to produce D-glucose.

References
1. Sablayrolles, J. M.: Control of alcoholic fermentation in winemaking: current
situation and prospect, Food Res. Int., 42, 418e424 (2009).

KINETICS OF HEXOSE CONVERSION IN WINE BY S. CEREVISIAE

49

2. Tronchoni, J., Gamero, A., Arroyo-Lpez, F. N., Barrio, E., and Querol, A.:
Differences in the glucose and fructose consumption proles in diverse
Saccharomyces wine species and their hybrids during grape juice fermentation,
Int. J. Food Microbiol., 134, 237e243 (2009).
3. Loureiro, V. and Malfeito-Ferreira, M.: Spoilage yeasts in the wine industry,
Int. J. Food Microbiol., 86, 23e50 (2003).
4. Jones, G. V., White, M. A., Cooper, O. R., and Storchmann, K.: Climate change
and global wine quality, Climatic Change, 73, 319e343 (2005).
5. Ribreau-Gayon, P., Dubourdier, D., Donche, B., and Lonvaud, A.: Handbook
of Oenology, vol. 1. The microbiology of wine and vinications. John Wiley and
Sons, West Sussex (2005).
6. Urtubia, A., Hernndez, G., and Roger, J. M.: Detection of abnormal fermentations in wine process by multivariate statistics and pattern recognition
techniques, J. Biotechnol., 159, 336e341 (2012).
7. Emparn, M., Simpson, R., Almonacid, S., Teixeira, A., and Urtubia, A.: Early
recognition of problematic wine fermentations through multivariate data
analyses, Food Control, 27, 248e253 (2012).
8. Bauer, F. F. and Pretorius, I. S.: Yeast stress response and fermentation efciency: how to survive the making of wine e a review, S. Afr. J. Enol. Vitic., 21,
27e51 (2000).
9. Perez, M., Luyten, K., Michel, R., Riou, C., and Blondin, B.: Analysis of
Saccharomyces cerevisiae hexose carrier expression during wine fermentation:
both low- and high-afnity Hxt transporters are expressed, FEMS Yeast Res., 5,
351e361 (2005).
10. Guillaume, C., Delobel, P., Sablayrolles, J. M., and Blondin, B.: Molecular basis
of fructose utilization by the wine yeast Saccharomyces cerevisiae: a mutated
HXT3 allele enhances fructose fermentation, Appl. Environ. Microbiol., 73,
2432e2439 (2007).
11. Berthels, N. J., Cordero Otero, R. R., Bauer, F., Pretorius, I. S., and
Thevelein, J. M.: Correlation between glucose/fructose discrepancy and
hexokinase kinetic properties in different Saccharomyces cerevisiae wine yeast
strains, Appl. Microbiol. Biotechnol., 77, 1083e1091 (2008).
12. Andrich, G., Casella, S., Fiorentini, R., and Spettoli, P.: A tentative model to
evaluate the kinetics of malolactic fermentation, Ann. NY Acad. Sci., 542,
356e359 (1989).
13. Zinnai, A., Venturi, F., Quartacci, M. F., and Andrich, G.: A mathematical
model to describe malolactic fermentation, Ital. J. Food Sci., 23, 80e89 (2011).
14. Buzzi Ferraris, G. and Manca, D.: BURENL. Politecnico, Dipartimento di
Ingegneria Chimica G. Natta, Milan (1996).
15. Dumont, A., Rayanal, C., Raginel, F., and Ortiz-Julien, A.: La capacit de
consommation du fructose par les levures nologiques, La Revue des Oenologues, 129, 15e18 (2008).
16. Salmon, J. M.: Effect of sugar transport inactivation in Saccharomyces cerevisiae
on sluggish and stuck enological fermentations, Appl. Environ. Microbiol., 55,
953e958 (1989).
17. Weusthuis, R. A., Pronk, J. T., Van Der Broeck, P. J. A., and Van Dijkenl, J. P.:
Chemostat cultivation as a tool for studies on sugar transport in yeasts,
Microbiol. Rev., 58, 616e630 (1994).
18. Sablayrolles, J., Dubois, C., Manginot, C., Roustan, J., and Barre, P.: Effectiveness of combined ammoniacal nitrogen and oxygen additions for
completion of sluggish and stuck wine fermentations, J. Ferment. Bioeng., 82,
377e381 (1996).