Journal of Biotechnology 83 (2000) 219 230

www.elsevier.com/locate/jbiotec

Effect of pH and stirring rate on itaconate production by

Aspergillus terreus
Emanuele Riscaldati a, Mauro Moresi a,*,1, Federico Federici,
Maurizio Petruccioli b
b

a
Istituto di Tecnologie Agroalimentari, Uni6ersity of Tuscia, Via S. Camillo de Lellis, 01100 Viterbo, Italy
Dipartimento di Agrobiologia ed Agrochimica, Uni6ersity of Tuscia, Via S. Camillo de Lellis, 01100 Viterbo, Italy

Received 28 January 2000; received in revised form 5 June 2000; accepted 28 June 2000

Abstract
The production of itaconic acid from glucose-based media by Aspergillus terreus NRRL 1960 was found to be
controlled by stirring rate and pH. When the phosphorous (P) level in the production medium was reduced to less
than 10 mg l 1, the fungal mycelium exhausted its primary growth and started to excrete itaconic acid, while it
continued its secondary growth at the expense of ammoniacal nitrogen. The fermentation exhibited a mixed-growthassociated product formation kinetics, the non-growth associated production term (mI) being practically zero only
when the pH was left free to change from 3.4 down to 1.85. On the contrary, when the pH was kept reducing up to
a constant value by automatic addition of KOH 4 mol l 1, the itaconate yield coefficient on the initial glucose
supplied (YI/So) and mI and were 0.53 g g 1 and 0.028 h 1 at pH 2.4 and 320 rev min 1 and 0.5 g g 1 and 0.036
h 1 at pH 2.8 and 400 rev min 1, respectively. Although the differences between mI and YI/So were statistically
insignificant at the 95% confidence level, the net difference in the corresponding yield coefficients for itaconic acid on
mycelial biomass resulted in a maximum itaconate production rate of 0.41 g l 1 h 1 at pH 2.8 and 400 rev min 1,
thus showing that this operating condition is no doubt optimal for the process under study. 2000 Elsevier Science
B.V. All rights reserved.
Keywords: Itaconic acid; Production; Aspergillus terreus; pH; Stirring rate; Mathematical modelling

1. Introduction

* Corresponding author. Tel.: + 39-0761-357494; fax: +390761-357498.

E-mail address: mmoresi@unitus.it (M. Moresi).
1
Present address: Istituto di Tecnologie Agroalimentari,
University of Tuscia, Via San Camillo de Lellis, 01100
Viterbo, Italy.

Owing to its conjugated double bonds and two

carboxyl groups, itaconic acid is a very reactive
compound and is used as monomer or comonomer for plastics, resins, synthetic fibers and
elastomers (Tate, 1981; Milson and Meers, 1985).
Laboratory and pilot-plant production of
itaconic acid by surface and submerged-cul-

0168-1656/00/$ - see front matter 2000 Elsevier Science B.V. All rights reserved.
PII: S 0 1 6 8 - 1 6 5 6 ( 0 0 ) 0 0 3 2 2 - 9

220

E. Riscaldati et al. / Journal of Biotechnology 83 (2000) 219230

ture fermentations was firstly reported by the

Research Laboratories of the US Department of
Agriculture (Lockwood and Ward, 1945; Nelson
et al., 1952) and the strain Aspergillus terreus
NRRL 1960 has been so far used as a target
microorganism for optimising the fermentation
process. Typical yields of 45 54 g of itaconic acid
per 100 g of glucose supplied are generally obtained in 46 days of fermentation (Nelson et al.,
1952).
Nowadays, itaconic acid is produced by submerged culture fermentation with A. terreus in a
medium containing molasses as the sugar source
at 3240C and pH of 1.8 4.0 under 0.25 0.5
volumes of air per volume of medium per minute
(l l 1 min 1) for 48 72 h (Milson and Meers,
1985; Roehr and Kubicek, 1996). The mycelial
growth is sensitive to the ferrous ion content and
limited by the phosphate concentration in the
fermentation medium (Roehr and Kubicek, 1996).
Several byproducts, such as erythritol and a-ketoglutaric, L-malic, succinic and itatartaric acids,
have been reported to be produced during the
itaconic acid fermentation (Larsen and Eimhjellen, 1955; Kobayashi, 1967). However, their formation can be suppressed by controlling the pH
of the medium at 3.0 5.0 once the itaconate
concentration reaches 4.6% (Batti, 1964).
Although several studies have been so far carried out to optimise the fermentation process
(Milson and Meers, 1985; Gyamerah, 1995; Roehr
and Kubicek, 1996; Yahiro et al., 1997), some
operating parameters are not fully defined yet. In
fact, their so called optimal values vary in quite a
large range and, sometimes, are even discordant
(Roehr and Kubicek, 1996).
As for the pH, for instance, good itaconic acid
yields have been claimed on condition that the pH
of the production medium be kept at 1.8 2.2
(Larsen and Eimhjellen, 1955; Elnaghy and
Megalla, 1975) or 2.9 3.5 (van der Westhuizen et
al., 1951; Rychtera and Wase, 1981).
The effects of dissolved oxygen concentration
(DO) and stirring speed at pH 2.0 were assessed
by Park et al. (1993). A maximum yield coefficient
of 0.55 g of itaconic acid per g of glucose consumed was obtained provided that DO was about
20% of the saturation value and the peripheral

speed (6I) (60-mm Rushton impellers) was about

0.94 m s 1, equivalent to a stirring rate of 320 rev
min 1. Any increase or reduction in stirring rate
would lower the itaconate production because of
either mechanical damage of mycelial structure or
insufficient oxygen transfer rate. On the contrary,
Rychtera and Wase (1981) found that the best
conditions for itaconate production were a pH
between 2.9 and 3.1 and a peripheral impeller
speed of 0.71 m s 1.
The main aim of this work was to maximise
itaconic acid yield and productivity with respect
to pH and stirring rate. Therefore, a series of
fermentations was firstly carried out to assess the
effect of pH on the itaconic fermentation under
constant stirring rate. Secondly, another series of
trials was designed to assess the effect of stirring
speed at the optimal pH suggested by Rychtera
and Wase (1981). Finally, mathematical modelling
of itaconate formation kinetics was used to
confirm the optimal operating conditions for this
production.

2. Materials and methods

2.1. Microorganism
A. terreus NRRL 1960 was obtained from the
culture collection of the North Regional Research
Laboratory (Peoria, IL). During this work, the
culture was maintained on malt extract agar at
46C and subcultured every month.

2.2. Apparatus
A 15-l fermenter (Applikon, Schiedam, NL)
with internal diameter 0.211 m and two turbines
with six flat blades (diameter 75 mm) was used to
study the fermentation process. The fermenter
was equipped for the control of stirrer rate, temperature, pH, dissolved oxygen (DO) and air flow.
Foaming or pH, when needed, was automatically
controlled by adding a silicon antifoam (Antifoam
A, Fluka) or an alkaline solution (4 mol l 1
KOH), respectively.

E. Riscaldati et al. / Journal of Biotechnology 83 (2000) 219230

2.3. Culture medium and growth conditions

The medium A for vegetative seed culture consisted of (per litre of de-ionised water): glucose,
20.0 g; (NH4)2SO4, 0.47 g; KH2PO4, 0.022 g;
MgSO47H2O, 0.415 g; CaCl2 2H2O, 0.026 g;
5H2O, 0.04 mg;
NaCl, 0.015 g; CuSO4
FeSO47H2O, 1.1 mg; MnCl24H2O, 0.14 mg;
ZnSO47H2O, 0.25 mg. The production medium B
contained (per litre of de-ionised water): glucose
100.0 g; (NH4)2SO4, 2.36 g; KH2PO4, 0.11 g;
MgSO47H2O, 2.08 g; CaCl22H2O, 0.13 g; NaCl,
0.074 g; CuSO4
5H2O, 0.2 mg; FeSO4
7H2O, 5.5 mg; MnCl2 4H2O, 0.7 mg; ZnSO4
7H2O, 1.3 mg. The pH of medium A or B was
adjusted to 4.2 or 3.4, respectively. All media were
sterilised at 121C for 20 min.
For inoculation, 5 ml of spore suspensions containing about 4 107 spores ml 1 were transferred into 1-l Erlenmeyer flasks containing 200
ml of sterile medium A and incubated on a rotary
shaker at 200 rev min 1 and 35C for 24 h. Then,
1 l of the vegetative seed culture were aseptically
transferred to the fermentation vessel previously
filled with 9 l of sterile medium B.
The itaconate fermentation trials were performed batchwise under the following operating
conditions: 35C, initial pH 3.4, and 0.4 l l 1
min 1. During the itaconate production, the pH
tended to reduce and was left free to change up to
a final value of 1.85. Alternatively, as the pH was
1.95, 2.2, 2.4 or 2.89 0.05 pH units, it was kept
constant by automatic addition of 4 mol l 1
KOH. If not indicated otherwise, the stirring rate
was 320 rev min 1, this being equivalent to a
peripheral impeller speed of 1.256 m s 1. Any
experimental trial was replicated twice, thus involving a coefficient of variation (i.e. the ratio
between the standard deviation and sample mean)
for all stoichiometric and kinetic parameters estimated here ranging from 0.05 to 0.1.

2.4. Analytical methods

Samples were withdrawn at intervals of 6 8 h,
diluted 1:5 with boiling distilled water, and
filtered on pre-weighed Whatman GF/D discs to
recover the mycelial biomass. The solid fraction

221

was washed twice with distilled water, dried at

105C for 24 h and weighed to yield the mycelial
biomass (X) concentration. Residual glucose (S)
in the supernatant was spectrophotometrically
measured by the DNSA method (Miller, 1959),
while residual ammoniacal nitrogen (N) and phosphorous (P) in the supernatant were spectrophotometrically assayed as described by Strickland
and Parson (1972). Itaconic acid (I) in the supernatant was determined by using a Violet Instruments (Rome, Italy) high-performance liquid
chromatograph. The apparatus was equipped
with a high-pressure volumetric pump (type Clar
003, Violet Instruments, Rome, Italy), a UV detector (Violet SP 430) and a 20-cm reverse phase
LiChrospher, 100 RP-8, 5 mm (Merck, Darmstadt,
Germany) column at room temperature. A buffer,
0.05 mol l 1 KH2PO4 at pH 3, was used as eluent
at a flow rate of 1.2 ml min 1 and itaconic acid
detection was performed at 210 nm.

3. Results and discussion

To assess the optimal operating conditions for
the itaconate production by A. terreus NRRL
1960 on glucose-based media, two series of fermentation trials were performed under constant
stirring rate and pH, respectively.

3.1. Effect of pH
Several fermentation trials were performed by
varying the pH in the range 1.852.8 under a
constant stirring rate of 320 rev min 1. Despite
the corresponding peripheral impeller speed
(1.256 m s 1) was greater than that (0.942 m s 1)
suggested by Park et al. (1993), such agitation rate
was chosen to enhance mixing and oxygen transfer, being the working volume of the stirred fermenter used here 5 times greater than that used
by Park et al. (1993).
The mean values of the main results [initial (ti)
and final (tf) times of the itaconate production
phase together with the corresponding concentrations of mycelial biomass (Xi, Xf), glucose (SiF,
SfF) and itaconic acid (Ii, If), and estimated values
of itaconate production (RI), glucose consumption

222

E. Riscaldati et al. / Journal of Biotechnology 83 (2000) 219230

(RS) and mycelial growth (RX) rates and of yield

coefficients for itaconic acid on glucose consumed
(YI/S) or supplied (YI/So) and for mycelial biomass
on glucose consumed (YX/S)] of all these fermentation runs are shown in the upper section of Table
1. As reported in Section 2, the coefficient of

Fig. 1. Time course of itaconic acid production by Aspergillus

terreus NRRL 1960 at 320 rev min 1 when the pH was left
free to change: (a) Concentrations of glucose (S : ), itaconic
acid (I : ") and mycelial biomass (X : ) as a function of the
fermentation time (t); (b) Concentrations of dissolved oxygen
(DO : D), ammoniacal nitrogen (N :
) and phosphorous (P :
), and pH (2) versus t. The continuous lines were calculated
by using Eqs. (6) and (7) together with the empirical coefficients listed in Table 2.

variation for all the estimated parameters in Table

1 fell within 0.05 and 0.1.
Itaconic acid yield (YI/S) and productivity (RI)
appeared to be function of the pH of the medium
and varied in the ranges of 0.330.63 g per gram
of glucose consumed and 0.120.25 g l 1 h 1,
respectively.
Of the two itaconate yield coefficients estimated
here, that on the initial glucose supplied (YI/So)
appeared to be the most suitable for assessing the
optimal pH for the fermentation process. When
the pH was uncontrolled and reached a final value
of 1.85, the acidity of the medium markedly limited the fungal utilisation of the glucose initially
supplied and YI/So was only 0.21 g g 1. On the
contrary, when the pH was controlled at 2.4 the
glucose was practically exhausted at the end of
the fermentation (Table 1) and YI/So was 0.53 g
g 1.
Figs. 1 and 2 illustrate the typical time course
of the itaconate fermentation under the operating
conditions mentioned above.
As shown in Figs. 1a and 2a, after about 24 h
of fermentation, the phosphorous (P) concentration in the production medium fell to about 10 mg
l 1 and the fungus started to excrete itaconic
acid. During the production phase, the nitrogen
(N) concentration continued to decrease from
about 20 to less than 1 mg l 1, while mycelial
biomass increased from 3 to 1011 g l 1.
The rapid phosphorous consumption during the
first 24 h was accompanied by a dramatic fall in
dissolved oxygen (DO) from 100 to 1015% of
the saturation value. After that time, DO remained almost constant up to the overall exhaustion of the nitrogen source, then started to
increase again, thus tending to an equilibrium
value when all mycelial metabolic activity was
completely exhausted (Figs. 1b and 2b).
Despite the above optimal pH of 2.4 is in line
with several other reports (Larsen and Eimhjellen,
1955; Elnaghy and Megalla, 1975), it is worth
noting that the itaconate yield coefficient (YI/S =
0.33) determined here (when using a 75-mm Rushton-type turbine operating at 6I = 1.256 m s 1)
was half that (YI/S = 0.64) obtained by Rychtera
and Wase (1981), who worked at a pH of 2.93.1
with a 85-mm single flat-blade impeller operating
at 6I = 0.71 m s 1.

24.5
22.0
26.5
23.0
19.0
23.0
15.0
22.0

1.85a
1.95
2.2
2.4
2.8
2.9
2.8
2.8
165.0
141.0
209.0
251.0
137.0
166.5
121.0
128.5

tf (h)

3.1
2.5
2.6
1.9
3.0
4.5
2.6
4.0

XI
(g
l1)
11.2
11.8
8.6
10.6
8.0
10.7
9.3
9.7

Xf
(g
l1)
78.0
86.7
73.6
107.4
75.0
85.0
86.0
86.0

SiI
(g
l1)
36.8
48.1
17.0
1.5
24.5
3.0
5.5
13.0

SfI
(g
l1)
0.00
0.04
1.50
0.00
1.04
0.50
0.00
0.80

Ii
(g
l1)
16.5
21.0
36.9
57.2
17.7
25.2
43.1
33.1

If
(g
l1)

0.12
0.18
0.19
0.25
0.14
0.17
0.41
0.30

RI
(g l1
h1)

Referred to the final pH of the medium when the fermentation was carried out with no pH control.

320
320
320
320
320
240
400
480

tI (h)

pH () Stirring
rate (rev
min1)

-0.29
-0.32
-0.31
-0.46
-0.43
-0.57
-0.76
-0.69

RS
(g l1
h1)

0.06
0.08
0.03
0.04
0.04
0.04
0.06
0.05

RX
(g l1
h1)

0.40
0.54
0.63
0.54
0.33
0.30
0.54
0.44

YI/S
(g
g1)

0.21
0.24
0.48
0.53
0.22
0.29
0.50
0.38

YI/So
(g
g1)

0.20
0.24
0.11
0.08
0.10
0.08
0.08
0.08

YX/S
(g
g1)

Table 1
Mean values of the main results of several itaconic acid production by A. terreus NRRL 1960 replicated twice: Effect of pH and stirring rate on the initial (ti) and
final (tf) times of the itaconate production phase together with the corresponding concentrations of mycelial biomass (Xi, Xf), glucose (Si, Sf) and itaconic acid (Ii, If),
and estimated values of itaconate production (RI), glucose consumption (RS) and mycelial growth (RX) rates and yield coefficient for itaconic acid on glucose consumed
(YI/S) or initially supplied (YI/So) and for mycelium on glucose consumed (YX/S). The coefficient of variation for all these stoichiometric and kinetic parameters ranged
from 0.05 to 0.1a

E. Riscaldati et al. / Journal of Biotechnology 83 (2000) 219230

223

224

E. Riscaldati et al. / Journal of Biotechnology 83 (2000) 219230

Fig. 2. Time course of itaconic acid production by A. terreus

NRRL 1960 at 320 rev min 1 when the pH was left free to
change up to 2.4: (a) Concentrations of glucose (S : ),
itaconic acid (I : ") and mycelial biomass (X : ) as a function
of the fermentation time (t); (b) Concentrations of dissolved
oxygen (DO : D), ammoniacal nitrogen (N :
) and phosphorous (P : ), and pH (2) versus t. The continuous lines were
calculated by using Eqs. (6) and (7) together with the empirical
coefficients listed in Table 2.

3.2. Effect of stirring rate

To establish if the above experimental result
(YI/S =0.33) was due to insufficient oxygen transfer in the fermenter used, a second series of trials
was carried out by setting the stirring rate to 240,
400 and 480 rev min 1 (these values corresponding to peripheral impeller speeds of 0.94, 1.57 and
1.88 m s 1) and pH to about the optimal value

suggested by Rychtera and Wase (1981). The

mean values of their main results are given in the
lower section of Table 1.
By increasing the stirring rate from 240 to 480
rev min 1, the yield coefficients YI/So and YI/S
firstly increased reaching a maximum value of
0.50 and 0.54 g g 1 at 400 rev min 1, respectively, and then decreased. On the contrary, YX/S
appeared to be independent of agitation rate and
practically equal to 0.08 g g 1.
Fig. 3 shows the typical time course of the
itaconate fermentation at pH 2.8 and 400 rev
min 1. By only increasing the stirring rate to 400
rev min 1, it was possible to confirm previous
findings by Rychtera and Wase (1981), thus revealing a significant dependence of the itaconate
yields upon the aeration-agitation conditions
within the fermenter used. Although there are a
number of reports in the literature of hyphal
damage due to high agitation intensity [see for
general reference the well known paper by Dion et
al. (1955) and more specifically for the fermentation under study that by Park et al. (1993)], these
results clearly show that the mycelium was not
damaged by the shear stress exerted by the impeller blades at peripheral impeller speeds up to
1.57 m s 1. This is in strict agreement with van
Suijdam and Metz (1981) results on the effect of
agitation intensity on the length of mycelial hyphae and with Moresi et al. (1987) on the effect of
6I in the range 1.763.8 m s 1 on the stoichiometry and kinetics of untreated orange peel utilisation by Fusarium a6enaceum.
The fact that the strain used in this work at pH
2.8 exhibited a higher resistance to shearing forces
than that used by Park et al. (1993) at pH 2.0
might be in all probability attributed to the higher
pH of the fermenting media.
To test whether the experimental mean values
of the yield coefficients for itaconic acid on glucose supplied (YI/So), that were observed at pH 2.8
and 400 rev min 1 and at pH 2.4 and 320 rev
min 1, were significantly different, their standardised normal difference [d= (Y1 Y2)/
((s 21/p1)+
(s 22/p2)) where Yj, sj, and pj are the mean value,
standard deviation and overall number of trials
for the variables tested] was compared with the
classic inequality of the hypothesis test for means,

E. Riscaldati et al. / Journal of Biotechnology 83 (2000) 219230

[i.e. d5ta,f the one-sided Student t-test for a

confidence level a and f ( =p1 +p2 2) degrees
of freedom] (Bennett and Franklin, 1954).
In this way, for s1 :s2 =0.03, p1 =p2 =2 and
t0.05,2 =4.3, it was possible to reject the hypothesis
of inequality (d=1.0 B4.3) at a probability level
of 0.05. Similarly, the same inequality test performed on the itaconate formation rate (RI) values obtained at 320 rev min 1 and pH ranging
from 1.95 to 2.8 (upper section of Table 1) yielded

225

that there was no statistically significant difference

between the yields and rates examined, thus justifying the discordant literature indications about
the optimal pH of the medium during the itaconate production phase (Larsen and Eimhjellen,
1955; Elnaghy and Megalla, 1975; Rychtera and
Wase, 1981).
However, when the stirring rate was increased
from 320 to 400 rev min 1 at pH 2.8, the itaconate formation rate (RI) was found to increase
from 0.259 0.02 to 0.419 0.03 g l 1 h 1 (Table
1). Since the difference in RI (6.3\ 4.3) was statistically significant at the 95% confidence level, the
kinetics of this fermentation is no doubt controlled by the effective oxygen transfer capability
of the fermenter used.
In conclusion, despite the operation at pH 2.4
and 320 rev min 1 would seem to be more convenient because the smaller the stirring speed the
smaller the power input per unit itaconate formed
and the smaller the pH of the medium the smaller
the alkaline reactant consumption and the microbial contamination risk, the operation at pH 2.8
and 400 rev min 1 has to be regarded as the
optimal one because of the greater itaconate formation rate.

3.3. Kinetics of itaconate production

Among the numerous models reported in the
literature to describe quantitatively the formation
rate of a metabolic compound (Moser, 1985), the
now classic Luedeking and Piret (1959), kinetics
has so far proved extremely useful and versatile in
fitting product formation data from many different fermentations:
RI = dI/dt = YI/X(dX/dt)+mI(X)

Fig. 3. Time course of itaconic acid production by A. terreus

NRRL 1960 at 400 rev min 1 when the pH was left free to
change up to 2.8: (a) Concentrations of glucose (S : ),
itaconic acid (I : ") and mycelial biomass (X : ) as a function
of the fermentation time (t); (b) Concentrations of dissolved
oxygen (DO : D), ammoniacal nitrogen (N :
) and phosphorous (P : ), and pH (2) versus t. The continuous lines were
calculated by using Eqs. (6) and (7) together with the empirical
coefficients listed in Table 2.

(1)

where mI is the specific product formation rate at

zero mycelial growth rate and YI/X the yield factor
for itaconic acid on unit mycelial biomass. In
particular, the first term indicates the product
formation rate in association with mycelial
growth rate; while the second one is the nongrowth associated product formation rate.
To apply this two-parameter kinetic expression,
it is necessary to know the mycelial growth rate
(RX), which can be generally described as:

E. Riscaldati et al. / Journal of Biotechnology 83 (2000) 219230

226

Fig. 4. Itaconic acid production by A. terreus NRRL 1960 at

pH= 2.8 and 400 rev min 1: logarithm of (X/Xo : ") and
concentrations of phosphorous (P : ) and ammoniacal nitrogen (N :
) versus time. The continuous line was calculated by
using Eq. (4) together with the empirical coefficients listed in
Table 2.

RX = dX/dt =m(X)

(2)

where X is the instantaneous mycelial concentration and m is the actual specific cell growth rate
which is a more or less complex function of
several variables such as, for instance, the limiting
substrate concentration in the case of the Monod
equation or the maximum mycelial concentration
in the case of the logistic equation (Moser, 1985).
By assuming no explicit relationship for m, Eq. (2)
can be integrated as
ln

X
=
X0

&

m dt

(3)

If the mycelium exponentially duplicates, the

specific mycelial growth rate will coincide with the
maximum specific one and the plot of the natural
logarithm of the actual to the initial mycelial
concentrations against time will be a straight line
passing through the origin of the plot.
In this specific fermentation, the general plot of
ln(X/X0) versus time had the pattern shown in
Fig. 4, which may be approximated by using at
least two different straight lines, the first and
second ones fitting the experimental data at the
beginning and at the end of the fermentation,
respectively. This means that mycelial growth may

be roughly described as double-substrate growth

(Moser, 1985), the first limiting substrate being
the phosphorous and the second one the nitrogen
(Fig. 4). In fact, the ln(X/X0) deviates from the
first straight line having slope m1 as the P content
in the medium is smaller than about 10 mg l 1
(Fig. 4). When the mycelium has adapted to the
new cultural condition, it reduces its instantaneous specific growth rate and starts to excrete
itaconic acid. In the circumstance, being the
medium still rich in carbon and nitrogen sources,
the mycelial weight concentration X will continue
to increase at the expense of another limiting
substrate, that is the ammoniacal nitrogen. Such a
secondary mycelial growth is characterised by a
secondary specific growth rate m2, which was
found to be about one tenth of the primary one
(m1).
This microbial behaviour is not specific of the
A. terreus strain used, since it was already observed in Yarrowia lipolytica (Moresi, 1994) and
A. niger (Verhoff and Spradlin, 1976; Briffaud
and Engasser, 1979) during the citric acid formation and in Saccharomyces cere6isiae (Martini et
al., 1976) during the aerobic growth, both on
media rich in glucose and devoid of the nitrogen
source.
To describe such double-substrate growth, the
following empirical regression was used:
ln

X(t) q
= %jbj (1e mj t)
Xo
1

(4)

where q is the number of statistically significant

limiting substrates, mj the j-th maximum specific
mycelial growth rate and bj the j-th empirical
regression coefficient.
By using the same procedure developed by
Nussinovitch et al. (1989) and modified by
Mancini et al. (1999) to correlate stress relaxation
data, it was possible to fit all the experimental
data by assigning initially an arbitrary spectrum
of reaction times:
1
tj = = a x10z j
mj

(5)

where a is an integer index ranging from 1 to 9

and z is an empirical exponent assigned to make
the maximum reaction time coincident with the

E. Riscaldati et al. / Journal of Biotechnology 83 (2000) 219230

overall duration of the fermentation process. For

instance, if the process lasts 300 h, z is equal to 4
so as to make the first reaction time (t1) varying
from 1000 and 9000 h.
By using a linear regression method, it was
possible to discard firstly the reaction times, the
contribution of which to the ln(X/Xo)-t relationship was statistically insignificant at the confidence level of 95%, and then to determine the
optimal a value by maximising the coefficient of
determinations (r 2) of Eq. (4).
By applying the above operating procedure,
there was no statistically significant improvement
in fitting the time course of any experimental
[ln(X/Xo)t] data by considering more than two
maximum specific cell growth rates in agreement
with the aforementioned verbal kinetic model.
Under the assumption that all the stoichiometric and kinetic parameters were constant, the
above set of differential equations can be integrated, thus yielding the following:

X0
q
X=
X0 exp %jbj (1 e uj t)

1
I = Io + mI


&

for t 5t0
for t \t0

(6)

X dt +YI/X (X Xi)

(7)

tI

where the second term at the right hand side of

Eq. (7) can be numerically integrated once X(t) is

227

known. By applying the analytical expressions

given above in conjunction with the least squares
method, it was possible to fit all the fermentation
data collected and evaluate the kinetic and stoichiometric parameters listed in Table 2.
The primary maximum specific cell growth rate
(m1) ranged from 0.025 to 0.045 h 1, while the
secondary one (m2) was one tenth of m1. Moreover, the itaconate production by A. terreus
NRRL 1960 exhibited a mixed-growth-associated
product formation kinetics, the contribution of
growth associated production being related to the
secondary mycelial growth only and that of nongrowth associated production depending on pH
and stirring rate.
More specifically, at constant stirring rate of
320 rev min 1 the specific product formation rate
at zero mycelial growth rate (mI) was practically
zero when the pH was left reducing without any
control. This revealed a growth-associated
product formation only when the mycelium was
severely stressed by as low a pH as 1.85.
By controlling the pH of the medium during the
fermentation, mI progressively increased from as
low as 0.004 h 1 at pH 1.95 to a maximum value
of 0.028 h 1 at pH 2.4. At pH 2.8, mI was found
to be also a function of the stirring rate, its
maximum value (0.036 h 1) being associated with
a stirring rate of 400 rev min 1. However, the
difference in the last two maximum values of mI
(d=3.1 B 4.3) was found to be statistically in-

Table 2
Main kinetic and stoichiometric parameters of several itaconic acid production by A. terreus NRRL 1960 replicated twice: Effect of
pH and stirring rate on the primary (m1) and secondary (m2) maximum specific mycelial growth rates and their corresponding
dimensionless contribution (bj ), specific product formation rate at zero mycelial growth rate (mI) and yield factor for itaconic acid
on unit mycelial biomass (YI/X) together with their standard deviations and or coefficients of determinations (r 2)a
pH ()

1.85a
1.95
2.2
2.4
2.8
2.9
2.8
2.8
a

Stirring rate
(rev min1)
320
320
320
320
320
240
400
480

b1 ()

m1 (h1) b2 ()

m2 (h1) r 2

mI (h1)

YI/X (g g1)

r2

0.01 90.3
0.59 0.6
1.1 9 0.3
0.3 90.2
0.49 0.2
0.1 90.2
0.1 90.1
0.1 90.5

0.04
0.03
0.04
0.02
0.04
0.04
0.04
0.04

0.004
0.003
0.004
0.003
0.005
0.004
0.005
0.004

0.00290.001
0.004 9 0.002
0.012 9 0.003
0.028 90.002
0.003 90.003
0.013 9 0.001
0.036 90.003
0.023 90.003

2.4 90.1
2.4 9 0.2
3.9 9 0.6
1.0 90.2
2.4 90.3
1.6 9 0.1
2.2 90.3
1.5 90.3

0.97
0.99
0.97
0.98
0.90
0.99
0.99
0.97

3.4890.09
3.659 0.2
1.329 0.12
2.989 0.07
1.83 9 0.08
2.009 0.08
2.199 0.05
2.2890.15

0.99
0.96
0.94
0.99
0.99
0.99
0.99
0.97

Referred to the final pH of the medium when the fermentation was performed with no pH control.

228

E. Riscaldati et al. / Journal of Biotechnology 83 (2000) 219230

significant at the 95% confidence level. On the

contrary, at the same probability level the
difference in the corresponding YI/X values
(Table 2) resulted to be statistically significant
(d= 4.7\ 4.3). Therefore, the net improvement in
the RI value observed at pH 2.8 and
400 rev min 1 with respect to that obtained at pH
2.4 and 320 rev min 1 (Table 1) is most likely
due to the difference in the yield coefficient for
itaconic acid on mycelial biomass formed (Table
2).
The continuous lines plotted in Figs. 1 3 were
calculated by using Eqs. (6) and (7) together with
the empirical coefficients reported in Table 2 and
show a remarkable agreement between calculated
and experimental X and I data. Therefore, all
these results are consistent with the present hypothesis of mixed kinetics for itaconate
production.

progressively increased from 0.004 h 1 at pH 1.95

to 0.028 h 1 at pH 2.4, even if such a maximum
value was also associated with higher values of
pH (2.8) and stirring rate (400 rev min 1). Of
these two operating conditions, the latter has to
be definitively regarded as the optimal one for the
itaconic acid fermentation owing to the greater
product formation rate.

5. Notation
a
d
DO
I
mI
N

4. Conclusions
P
The discordant literature indications about the
optimal pH of the medium during the production
of itaconic acid from glucose-based media by A.
terreus NRRL 1960 may be attributed to the fact
that this fermentation is controlled by stirring rate
and pH.
A maximum yield coefficient (YI/S) of about
0.54 g of itaconic acid per g of glucose consumed
was obtained under different operating conditions, such as pH 2.4 and 320 rev min 1 or pH
2.8 and 400 rev min 1, while their corresponding
itaconate formation rates (RI) were 0.25 and 0.41
g l 1 h 1, respectively.
This fermentation exhibited a mixed-growth-associated product formation kinetics, even if the
contribution of non-growth associated production
was mainly affected by the pH during the itaconate production phase and secondly by stirring
rate. When no control of the pH was assured, the
mycelium resulted to be so severely stressed as to
stop itaconate production at pH 1.85, thus involving a strict growth-associated production kinetics.
On the contrary, under pH control and constant
stirring rate (320 rev min 1) the specific product
formation rate at zero mycelial growth rate (mI)

pj
q
r2
RI
RS
RX
S
sj
t
ta,f
6I
X
YI/S
YI/So
YI/X

integer index of Eq. (5).

standardised normal difference between means.
dissolved oxygen concentration
(%).
itaconic acid concentration (g l1).
specific product formation rate at
zero mycelial growth rate (h1).
ammonia nitrogen concentration
(mg l1).
phosphorous concentration (mg
l1).
overall number of trials for the jth parameter tested.
number of statistically significant
limiting substrates.
coefficient of determinations.
itaconate production rate (g l1
h1).
glucose consumption rate (g l1
h1).
mycelial growth rate (g l1 h1).
glucose concentration (g l1).
standard deviation of the j-th
parameter tested.
fermentation time (h).
one-sided Student t-test.
peripheral impeller speed (m s1).
mycelial biomass concentration (g
l1).
yield coefficients for itaconic acid
on glucose consumed (g g1).
yield coefficients for itaconic acid
on glucose supplied (g g1).
yield factor for itaconic acid on
unit mycelial biomass (g g1).

E. Riscaldati et al. / Journal of Biotechnology 83 (2000) 219230

Yj
YX/S

mean value of the j-th parameter

tested.
yield coefficients for mycelial
biomass on glucose consumed (g
g1).

Greek symbols
a
confidence level.
bj
generic j-th empirical regression
coefficient of Eq. (4).
f
degrees of freedom.
m
specific cell growth rate (h1).
mj
generic j-th maximum specific
mycelial growth rate (h1).
tj
generic j-th reaction time (h).
z
empirical exponent of Eq. (5).
Subscript
o
I
f

referred to the initial value.

referred to the beginning of the
itaconate production phase.
referred the end of the itaconate
production phase.

Acknowledgements
This research work was supported by a grant
from the National Research Council (CNR) of
Italy: Target Project on Biotechnology.

References
Batti, M.A., 1964. Process for the production of itaconic acid.
US Patent number, 3,162,582, (Dec. 22, 1964).
Bennett, C.A., Franklin, N.L., 1954. Statistical Analysis in
Chemistry and the Chemical Industry. Wiley, New York,
pp. 176 185.
Briffaud, J., Engasser, M., 1979. Citric acid production from
glucose. I. Growth and excretion kinetics in a stirred
fermentor. Biotechnol. Bioeng. 21, 20832092.
Dion, W.N., Carilli, A., Sermonti, G., Chain, E.B., 1955. The
effect of mechanical agitation on the morphology of Penicillium chrysogenum Thom in stirred fermenters. In: Rendiconto dellIstituto Superiore di Sanita`, English, 17,
187 205.
Elnaghy, M.A., Megalla, S.E., 1975. Itaconic acid production
by a local strain of Aspergillus terreus. Eur. J. Appl.
Microbiol. 2, 159 172.

229

Gyamerah, M.H., 1995. Oxygen requirement and energy relations of itaconic acid fermentation by Aspergillus terreus
NRRL 1960. Appl. Microbiol. Biotechnol. 44, 20 26.
Kobayashi, T., 1967. Itaconic acid fermentation. Process
Biochem. 2 (9), 61 65.
Larsen, H., Eimhjellen, K.E., 1955. The mechanism of itaconic
acid formation by Aspergillus terreus. I. The effect of
acidity. Biochem. J. 60, 135 147.
Lockwood, L.B., Ward, G.E., 1945. Fermentation process for
itaconic acid. Indus. Eng. Chem. 37, 405 406.
Luedeking, R., Piret, E.L., 1959. A kinetic study of the lactic
fermentation. Batch process at controlled pH. J. Biochem.
Microb. Technol. Eng. 1, 393 412.
Mancini, M., Moresi, M., Rancini, R., 1999. Uniaxial compression and stress relaxation tests on alginate gels. J.
Texture Studies 30 (6), 639 657.
Martini, A., Marocchini, C., Federici, F., 1976. Composizione
in acidi grassi delle cellule di Saccharomyces bayanus e S.
cere6isiae nel corso della diauxia glucosio-etanolo. (Fatty
acid composition of Saccharomyces bayanus and S. cere6isiae during diauxic growth on glucose-ethanol). Annali
Facolta` di Agraria dellUniversita` di Perugia, 31, 491 501.
Miller, G.L., 1959. Use of dinitrosalycilic acid reagent for
determination of reducing sugars. Anal. Chem. 31, 426
430.
Milson, P.E., Meers, J.L., 1985. Gluconic and itaconic acid.
In: Blanch, H.W., Drew, S., Wang, D.I. (Eds.), Comprehensive Biotechnology. vol. 3. The Practice of Biotechnology. Pergamon Press, Oxford, pp. 681 700.
Moresi, M., 1994. Effect of glucose concentration on citric
acid production by Yarrowia lipolytica. J. Chem. Technol.
Biotechnol. 60, 387 395.
Moresi, M., Clementi, F., Rossi, J., 1987. Kinetics of untreated orange peel utilisation by F. a6enaceum. J. Chem.
Technol. Biotechnol. 40, 233 249.
Moser, A., 1985. Kinetics of batch fermentation. In: Brauer,
H. (Ed.), Biotechnology: Fundamentals of Biochemical
Engineering, vol. 2. VCH, Weinheim, pp. 243 283.
Nelson, G.E.N., Traufler, D.H., Kelley, S.E., Lockwood, L.B.,
1952. Production of itaconic acid by Aspergillus terreus in
20-l fermentors. Ind. Eng. Chem. 44, 1166 1168.
Nussinovitch, A., Peleg, M., Normand, M.D., 1989. A
modified Maxwell and a nonexponential model for characterization of the stress relaxation of agar and alginate gels.
J. Food Sci. 54, 1013 1016.
Park, Y.S., Ohta, N., Okabe, M., 1993. Effect of dissolved
oxygen concentration and impeller tip speed on itaconic
acid production by Aspergillus terreus. Biotechnol. Lett.
15, 583 586.
Roehr, M., Kubicek, C.P., 1996. Further organic acids. In:
Rehm, H.J., Reed, G. (Eds.), Biotechnology. Products of
Primary Metabolism, vol. 6, second ed. VCH Weinheim,
pp. 364 379.
Rychtera, M., Wase, D.A.J., 1981. The growth of Aspergillus
terreus and the production of itaconic acid in batch and
continuous cultures. The influence of pH. J. Chem. Technol. Biotechnol. 31, 509 521.

230

E. Riscaldati et al. / Journal of Biotechnology 83 (2000) 219230

Strickland, J.D.H., Parson, T.R., 1972. A Practical Handbook

of Seawater Analysis. Fisheries Research Board of Canada,
Canada.
Tate, B.E., 1981. Itaconic acid and derivatives. KirchOthmer.
In: Mark, H.F., Othmer, D.F., Overberger, C.G., Seaborg,
G.T. (Eds.), Encycl. Chem. Technol, vol. 13, third. Wiley,
New York, pp. 865 873.
van der Westhuizen, G.C.A., Spruit, C.J.P., Sephton, H.H,
1951. Itaconic acid. I. Production by Aspergillus terreus from
unrefined cane sugar. J. Appl. Chem. 1, 356360.

van Suijdam, J.C., Metz, B., 1981. Influence of engineering

variables upon the morphology of filamentous molds. Biotechnol. Bioeng. 23, 111 148.
Verhoff, F.H., Spradlin, J.E., 1976. Mass and energy balance
analysis of metabolic pathways applied to citric acid production by Aspergillus niger. Biotechnol. Bioeng. 18, 425 432.
Yahiro, K., Takahama, T., Jai, S., Park, Y., Okabe, M., 1997.
Comparison of air-lift and stirred tank reactors for itaconic
acid production by Aspergillus terreus. Biotechnol. Lett. 19,
619 621.