Process Biochemistry 40 (2005) 1071–1083

Modelling the fed-batch production of pediocin using

mussel processing wastes
Nelson P. Guerra, Ana Torrado Agrasar, Cristina López Macı́as, Lorenzo Pastrana∗
Departamento de Bioquı́mica, Xenética e Inmunoloxı́a, Facultade de Ciencias de Ourense, Universidade de Vigo,
As Lagoas, 32004 Ourense, Spain

Received 9 December 2003; received in revised form 5 March 2004; accepted 23 March 2004

Abstract

Cell growth and pediocin production by Pediococcus acidilactici NRRL B-5627 were compared using batch (on MRS broth and a culture
medium from mussel processing wastes (MPW)) and two fed-batch fermentations on MPW with re-alkalization cycles. This last fermentation
technique yielded the highest biomass and pediocin productions as compared to batch fermentations. Mathematical models were set up to
describe fed-batch production of biomass and pediocin by P. acidilactici. While cell growth was dependent on pH change, nitrogen and
phosphorous availability and product inhibition (lactic acid, ethanol and butane-2,3-diol), pediocin production was dependent on both growth
and the final pH reached in each re-alkalization period. The models developed offer a better fit and a more realistic description of the
experimental biomass and pediocin production data than the logistic and Luedeking and Piret model. Therefore, they could be used to design
feeding strategies for enhancing and controlling fed-batch pediocin production.
© 2004 Elsevier Ltd. All rights reserved.

Keywords: Pediocin; Fed-batch fermentation; Mussel processing wastes; Modelling; Homolactic fermentation; Mixed acid fermentation

1. Introduction Since bacteriocin synthesis is considered to be growth

In recent years, there has been an increased interest in the
use of bacteriocins as natural food preservatives and antimi-
crobial agents [1]. Bacteriocins are biologically active pro-
teins that have an antibacterial activity against Gram-positive
bacterial species related to the producer strain. Some of these
have a very narrow spectrum of activity but others have a
relatively broad spectrum of antibacterial activity [2–4].
Pediocins (that are produced by Pediococcus strains, or-
ganisms generally recognized as safe (GRAS)) have a wide
inhibitory spectrum of activity which includes both spoilage
and pathogenic organisms, such as Listeria monocytogenes,
Enterococcus faecalis, Staphylococcus aureus and Clostrid-
ium perfringens [5]. The use of these bacteriocins in combi-
nation with other stress inducing processes (such as heating,
freezing, acid treatment, chelating agents, high hydrostatic
pressure and electroporation) can also be effective against
Gram-negative or resistant Gram-positive bacteria [6–9].
∗ Corresponding author. Tel.: +34 988 387062; fax: +34 988 387001.
E-mail address: pastrana@uvigo.es (L. Pastrana).
associated, factors affecting biomass production (such as
temperature, pH, media composition, availability of certain
essential compounds, presence of inhibitory compounds)
also affect bacteriocin production [10,11]. However, in some
cases the pH can also produce a specific effect on bacteriocin
synthesis without affecting biomass production [10,12–14].
In batch fermentations, when the optimal initial culture con-
ditions for cell growth are fixed (pH, temperature and me-
dia composition), bacteriocin production was found to be
strongly dependent on the initial pH [15–18], the pH-time
course [13,19], the pH drop [12,19] and the final pH reached
in the cultures [19–22]. However, this effect of pH on bac-
teriocin production depends on the culture medium [17,18]
and the bacteriocin-producing strains or species [10,23].
Studies on factors affecting bacteriocin production by lac-
tic acid bacteria (LAB) are usually carried out in rich, un-
defined media. However, such media are too expensive and
have high peptone contents, which make difficult the sub-
sequent purification of bacteriocins [10]. For these reasons,
residual effluents from food industry have been used as in-
expensive substrates for bacteriocin production [13,14]. In
a previous work [14], we demonstrated the suitability of

0032-9592/$ – see front matter © 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.procbio.2004.03.014
1072 N.P. Guerra et al. / Process Biochemistry 40 (2005) 1071–1083

glycogen-rich mussel processing wastes [(MPW); average
COD: 25 g O2 /l; glycogen as main component: 5–10 g/l] to
be used as culture medium for batch production of nisin and
pediocin. Since the production of some bacteriocins can be
increased by employing fed-batch rather than batch fermen-
tation methods [25–27], we used a fed-batch fermentation
technique based on successive re-alkalizations of the culture
medium [12] to improve nisin production on mussel pro-
cessing wastes [24].
Some mathematical models have been proposed to de-
scribe bacteriocin production in batch cultivations [16,19].
These models were set up based on the Luedeking and
Piret-like equation [28], with a term for growth associated
bacteriocin production and a term for bacteriocin degrada-
tion or adsorption. However, models describing fed-batch
bacteriocin production are lacking and those used in batch
cultivations were not suitable for the simulation of fed-batch
bacteriocin production [12,26].
In this paper, we report on the use of the fed-batch culture
technique with re-alkalization cycles to increase pediocin
production by Pediococcus acidilactici NRRL B-5627 on
mussel-processing wastes. Furthermore, mathematical mod-
els were set up to describe the kinetics of P. acidilactici
NRRL B-5627 cell growth and bacteriocin synthesis during
fed-batch fermentation.
2. Materials and methods
2.1. Microorganisms
Pediococcus acidilactici NRRL B-5627, the pediocin-
producing strain, and Carnobacterium piscicola CECT
4020, the target organism, were obtained from the Northern
Regional Research Laboratory (NRRL, Peoria, IL, USA)
and the Spanish Type Culture Collection (CECT), respec-
tively. These bacteria were maintained at 4 ◦ C on agar slants
(MRS).
2.2. Culture medium, inoculum preparation and
fermentation conditions
The composition of the mussel processing wastes used as
fermentation medium is shown in Table 1. The preparation
of this waste to be used as culture medium was described
previously [14].
Table 1
Mean composition (g/l) of the culture media obtained from MPW
MPW
Total sugars 5.33
Reducing sugars 5.33
Proteins 1.82
Total nitrogen 0.65
Total phosphorous 0.14
Batch cultures of P. acidilactici on MRS and MPW
medium were performed in 250 ml Erlenmeyer flasks con-
taining 50 ml of the fermentation medium, on a rotary
shaker (200 rpm) at 30 ◦ C for 18 h. The inoculum consisted
on 2% (v/v) of an exponentially growing culture on MRS
broth or MPW medium. Samples were withdrawn (each 2
or 4 h) during incubation periods to perform the analytical
determinations.
Fed-batch and re-alkalized cultures on MPW medium
were carried out at a controlled temperature of 30 ◦ C in a 6 l
bench top fermentor (New Brunswick Scientific, New Jer-
sey) with an agitation of 200 rpm and continuous-record of
pH. The fermentor was filled with 4 l working volume of
medium. The aeration level (0.5 l/h) was obtained by con-
trolling the air supply by a flowmeter. Stepwise-pH profiles
were obtained by re-alkalizing the cultures repeatedly up to
initial pH (6.3) with 4N NaOH each 8 h. The feeding sub-
strate was added at the same time as re-alkalizations. The
inoculum consisted of 2% (v/v) of an exponentially grow-
ing culture on MPW medium. Samples (100 ml) were with-
drawn at 8 h intervals to perform analytical determinations.
The total sugar determination was used to calculate the fresh
feeding substrate volumes needed to restore the initial total
sugars concentration (5.33 g/l) in the fermentation medium.
The culture volume was kept constant by feeding the fer-
mentor with 100 ml fresh substrates containing the amounts
of sugars consumed in each cycle and distilled water when
this was necessary.
Two fresh substrates were used to feed the fermentor: a
240 g/l concentrated glucose (fermentation I), and a concen-
trated MPW (CMPW, Table 1) medium (fermentation II)
which was prepared in the same conditions as the fermenta-
tion medium (MPW). The feeding media were added to the
fermentor using a peristaltic pump (LKB, Pharmacia).
2.3. Analytical determinations
Methods for determination of cell growth, total phospho-
rus, nitrogen, protein and sugars were described or referred
to in a previous paper [24].
The concentrations of lactic and acetic acids, butane-2,3-
diol and ethanol were measured by HPLC using an ION-300
Organic Acids column (length 300 mm, internal diam-
eter 7.8 mm) with a pre-column IONGUARDTM (poly-
meric guard column), both obtained from Tecknokroma
S. Coop. C. Ltda., Barcelona, Spain. The mobile phase
consisted of 0.006N H2 SO4 at a flow rate of 0.4 ml/min
at (60–65) ◦ C and the refractive index of the peaks was
measured by a refractometer with a refractive-index detec-
tor [24]. All analytical determinations were performed in
CMPW triplicate.
101.33
101.33 2.4. Antibacterial activity assay
3.47
0.54 Aliquots from P. acidilactici NRRL B-5627 cultures were
0.06
adjusted to pH 3.5 with 5N HCl to avoid the adsorption
 N.P. Guerra et al. / Process Biochemistry 40 (2005) 1071–1083 1073

of molecules of bacteriocin onto the producer cell surfaces
[22]. Subsequently, they were heated at 100 ◦ C for 3 min to
kill the cells and centrifuged at 27 200 ×gfor 15 min at 4 ◦ C.
The culture supernatants (cell-free broth) containing overall
antibacterial activity were frozen until further use.
Quantitative antibacterial activities of cell-free broths
were estimated using a photometric assay on culture tubes
[24,29] using C. piscicola CECT 4020 as target organ-
ism. Antibacterial activity was expressed as Activity Units
(AU/ml, 1 AU/ml = amount of antibacterial compound
causing 50% growth inhibition (inhibitory dose 50: ID50
obtained from triplicate samples)), compared with a control
without bacteriocins.
3. Results and discussion
3.1. Batch cultures of Pediococcus acidilactici NRRL
B-5627 on MRS broth and MPW medium
The ability of P. acidilactici NRRL B-5627 to produce
bacteriocins was firstly tested in batch cultures on MRS
broth and on two supplemented MPW media (Fig. 1). In
the first medium the concentrations of biomass and pe-
diocin were higher than those reached in the unsupple-
mented MPW medium after 18 h of incubation (Table 2).
Since bacteriocin production is strongly affected by the type
and level of carbon, nitrogen and phosphate sources, cations
and inhibitors [10], the different pediocin productions ob-
served can be attributed to the different composition of the
media.
However, when the MPW medium was supplemented
with 2% complex nitrogen sources like Casitone or yeast ex-
tract (Fig. 1), the pediocin levels were two-fold higher than
those obtained in MRS broth. Thus, the pediocin-promoting
properties of both nitrogen sources could be related with
the additional supplement of vitamins, minerals and amino
acids [30].
Nevertheless, both yeast extract and Casitone are too
expensive nitrogen sources for the large-scale production
of pediocin. For this reason, it was necessary to find other
fermentation techniques more productive than the batch
mode. Taking into account that the fed-batch fermenta-
tion enhanced nisin production on MPW [24] and on TGE

Fig. 1. Kinetics of growth of Pediococcus acidilactici NRRL B-5627 on MRS broth (䊉), non-supplemented MPW (䊊) and MPW supplemented with
a 2% yeast extract (䉱) and 2% Casitone (). CDW: cell dry weight; Ped: pediocin titres; LA: lactic acid; Pr: proteins; RS: reducing sugars; TN: total
nitrogen; TP: total phosphorous. Means of three analytical replications.
1074 N.P. Guerra et al. / Process Biochemistry 40 (2005) 1071–1083
Table 2
Growth and fermentation parameters of non-realkalized (batch) and realkalized (fed-batch) cultures of Pediococcus acidilactici NRRL B-5627 for pediocin
production
Parameter Batch cultures
MRS MPW
CDW∗ (g/l) 1.76 0.79
Ped∗ (AU/ml) 493.21 368.40
YPed/TSc (AU/mg) 75.07 151.41
YPed/TNc (AU/mg) 604.3 3185.1
ETS 0.339 0.537
ETN 0.268 0.207
CDW∗ and Ped∗ are the maximum biomass and pediocin levels produced in the cultures. YPed/TSc and YPed/TNc are the pediocin yield coefficients
expressed as the amount of pediocin produced (AU) per mg of total sugars (TSc) or nitrogen (TNc) consumed. ETS and ETN are the efficiencies expressed
as g of substrate consumed per g of substrate supplied.
broth [12], two fed-batch and re-alkalized fermentations for
pediocin production on MPW were carried out.
3.2. Re-alkalized culture of Pediococcus acidilactici NRRL
B-5627 on MPW with feeding with glucose (fermentation I)
The growth kinetics of P. acidilactici NRRL B-5627 in
the fed-batch culture on MPW fed with a concentrated glu-
cose solution (240 g/l) and re-alkalized up to initial pH (6.3)
is shown in Fig. 2. From a kinetic point of view, it can be
noted that biomass and pediocin production and nutrients
consumption described wavy profiles. Moreover, the final
pHs at the end of each re-alkalization cycle increased almost
linearly throughout the fermentation and the pH recovery
capacity collapsed after 96 h of incubation when the growth
stopped. However, the final productions of biomass and pe-
diocin (2.28 g/l and 1034.01 AU/ml) were higher than those
obtained in batch culture on MRS broth and on the two sup-
plemented MPW media.
During the first 56 h of cultivation, P. acidilactici NRRL
B-5627 developed a homolactic fermentation (only lactic
acid was produced). In this period, lactic acid and glucose
were produced and consumed at constant rates of 0.18
and 0.20 g l−1 h−1 , respectively. The biomass concentra-
tion profile displayed two exponential growth phases (0–24
and 40–56 h) with similar biomass production rates (0.028
and 0.026 g l−1 h−1 ), separated by an intermediate station-
ary phase (24–40 h of incubation). The final biomass level
reached in this homolactic phase was 1.35 g/l. In addition,
the profiles of nutrient (protein (Pr), total phosphorous
(TP) and nitrogen (TN)) consumptions paralleled with that
described by biomass production.
In contrast, pediocin production increased to a level of
608.9 AU/ml, describing a logistic-type profile. A steep ini-
tial slope in the first 16 h of fermentation, when the final
pH in each re-alkalization cycle reached a value of 4.0,
was followed by a second slow but steadily near-linear bac-
teriocin synthesis, owing to increasing levels in the final
pH levels (from 16 to 56 h). Therefore, the pediocin pro-
duction rates decreased from 28.02 UA ml−1 h−1 (0–16 h)
to 4.01 UA ml−1 h−1 (16–56 h), even though the producing
Fed-batch cultures
MPW + 2% YE MPW + 2% CT MPW (I) MPW (II)
1.39 1.41 2.28 2.82
939.27 912.11 1034.01 1395.14
210.15 216.33 27.13 77.98
806.19 695.44 2839.2 2790.9
0.507 0.258 0.951 0.670
0.360 0.335 0.552 0.655
strain continued growing. Due to this specific effect of pH on
pediocin synthesis, a deviation in linearity between biomass
and pediocin production was produced (Fig. 2).
The same specific effect of final pH on pediocin produc-
tion has been observed before [13,14,20,21,24]. These re-
searchers obtained the highest levels of pediocin at final pH
values below 5.0 as compared to the negligible amounts of
bacteriocin obtained at final pH values above 5.0, even in the
presence of a substantial increase in cell density. This was
related to the need of low pH for posttranslational process-
ing of prepediocin to produce the active pediocin [20,21].
As it can be observed in Fig. 2, a mixed acid fermenta-
tion phase began after 56 h of cultivation, with the accumu-
lation of butane-2,3-diol in the medium. In addition, ethanol
became detectable in the medium after 88 h of incubation.
As a consequence of the production of these mixed acid
metabolites, the glucose consumption rate throughout this
phase (0.48 g l−1 h−1 ) was higher than that observed in the
previous phase.
The nitrogen (35 mg l−1 h−1 ) and phosphorus (383.4
mg l−1 h−1 ) consumption rates, which were constant from
56 to 96 h, decreased to 0.76 and 38.85 mg l−1 h−1 until the
112 h of cultivation.
The profiles described by both nitrogen and phosphorus
consumption paralleled the evolution of biomass and pe-
diocin production throughout the mixed acid fermentation
phase. Biomass and pediocin production rates were constant
(0.021 g l−1 h−1 and 8.99 AU ml−1 h−1 ) from 56 to 96 h, but
they decreased to 0.006 g l−1 h−1 and 4.09 AU ml−1 h−1 af-
terwards. A similar trend was observed for the final pH
reached at the end of each re-alkalization cycle.
The decrease in biomass production observed from 96 h
could not be associated with the complete exhaustion of the
nitrogen source, since at the end of the cultivation, the re-
maining concentration of nitrogen was 0.17 g/l. Neither bac-
teriocin autoinhibition nor the accumulations of by-products
with antibacterial activity seem to be causes for the cessa-
tion of growth. This is because the producing strains possess
specific immunity to their own bacteriocin [31,32] and the
amounts of inhibitory by-products synthesized (lactic acid,
butane-2,3-diol or ethanol) produced by P. acidilactici were
 N.P. Guerra et al. / Process Biochemistry 40 (2005) 1071–1083 1075

Fig. 2. Time course of re-alkalized cultures of Pediococcus acidilactici NRRL B-5627 on MPW with feeding with a 240 g/l concentrated glucose. B:
butane-2,3-diol; Et: ethanol; Pr, TS, TN, TP: proteins, total sugars, nitrogen and phosphorous supplemented () remaining (䊐) and consumed (䊊). I:
phase of homolactic fermentation; II: phase of mixed acid fermentation. Other notations as in Fig. 1.

lower than those considered damaging for lactic acid bacte-
ria [17].
In this case, a possible explanation could be the low avail-
ability or the exhaustion of one or several amino acids,
peptides, vitamins or cations essential for the growth of P.
acidilactici [30]. In this way, Kim et al. [32] also observed
that the growth of Lactococcus lactis and nisin production
decreased with decreasing organic nitrogen concentrations.
This was attributed to low nutrient availability, which oc-
curs when a strain grows well in a medium and quickly ex-
hausts the available nutrients leading to a cessation in both
the growth and bacteriocin production.
Other possible explanation for growth cessation could be
the lower phosphorus concentration (0.016 g/l) reached at
the end of the fermentation. This observation agrees with the
results obtained by De Vuyst and Vandamme [33] for L. lac-
tis subsp. lactis. These researchers observed a drastic limi-
tation in the cell yield and growth rate of this strain at phos-
phate concentrations lower than 1%. In contrast, increasing
the initial phosphate concentrations to values between 1 and
5% led to an increase in both the growth and nisin activity
levels. This fact was related to a possible stimulatory effect
of high phosphate concentrations on the formation of ATP,
which leads to a high energy charge of the cells [33].
The growth of P. acidilactici NRRL B-5627 on MPW
seems to be regulated and controlled by the nutrient (ni-
trogen and phosphorus sources) availability or limitation,
whereas pediocin production, which depends on biomass
production, is also regulated by the final pH reached in each
re-alkalization cycle.
Taking into account that both nitrogen and phosphorus
were almost exhausted due to both the consumption and
the successive extraction of samples, another fed-batch and
re-alkalized culture of P. acidilactici NRRL B-5627 was
carried out, using a concentrated MPW medium as a feeding
substrate. Thus, this feeding medium was not only used to
1076 N.P. Guerra et al. / Process Biochemistry 40 (2005) 1071–1083
Fig. 3. Time course of re-alkalized cultures of Pediococcus acidilactici NRRL B-5627 on MPW with feeding with CMPW (101.33 g of glucose/l).
Notations as in Figs. 1 and 2.
replenish the glucose consumed in each re-alkalization cycle,
but also as an additional source of nitrogen and phosphorus.
3.3. Fed-batch culture for pediocin production by
Pediococcus acidilactici on MPW medium with feeding
with CMPW medium (fermentation II)
The results of this second fed-batch fermentation are
shown in Fig. 3. As a consequence of feeding the fermenta-
tion medium with a concentrated MPW, biomass (2.82 g/l)
and pediocin (1395.1 AU/ml) levels were, respectively, 1.26
and 1.37 times higher than those obtained in the previous
culture with the same pH decrease values. Moreover, from
96 h of incubation, the final pH before each re-alkalization
cycle was stabilised in a value slightly higher than 4.5.
In addition, probably as a consequence of the joint sup-
plement of glucose, nitrogen and phosphorous throughout
the feeding with CMPW medium, the pediocin producing
strain developed a metabolism typically homofermentative
in this second fermentation.
In Fig. 3, it can be noted that at 64 h of incubation, the
remaining concentration of nitrogen in the medium reached
the same value (0.21 g/l) as that observed at the end of
the former fed-batch fermentation. However, P. acidilactici
continued growing and consuming nitrogen and phospho-
rous until the end of the fermentation. Thus, nitrogen and
phosphorus consumption were so highly stimulated, that the
final concentrations of both nutrients after 120 h of incuba-
tion were 0.08 and 0.014 g/l, respectively. In contrast, in the
first culture, the growth stopped (from the 96 h of incuba-
tion), when the concentrations of nitrogen and phosphorous
in the medium were 0.17 and 0.017 g/l. This suggests that
the CMPW medium provided growth-stimulating molecules
(essential amino acids, peptides, vitamins or cations (Mg2+ ,
Ca2+ , Mn2+ )) which probably were exhausted in the first
cultivation.
In this way, it has been observed that for growth of lac-
tococci on 10 g lactose/l, the essential amino acids were
required at different concentrations. While glutamic acid
was needed in the largest amounts (77 mg/l), the other
amino acids were required in concentrations between 20
and 40 mg/l [34]. However, other researchers [35] observed
that for optimal growth of L. lactis and L. cremoris at a
lactose concentration of 12 g/l, the minimal concentration
of essential amino acids was in the range of 10–80 mg/l.
In addition, when the MPW was supplemented with a
2% of glutamic acid [36], biomass production by L. lac-
tis subsp. lactis CECT 539 was four-fold higher than that
obtained in the non-supplemented MPW medium. From
these observations, it could be suggested that the relative
amounts of the essential amino acids seem to be probably
more important than their actual concentrations [37].
 N.P. Guerra et al. / Process Biochemistry 40 (2005) 1071–1083 1077

The stimulatory effect of pH decrease on pediocin synthe- tions, which can be described by the following stoichiomet-
sis was also observed in this fermentation. Thus, pediocin ric equations [42]:
production increased quickly after 16 h of incubation in con-
comitance with a rapid increase in biomass production and Glucose + 2ADP + 2Pi → 2lactate + 2ATP + 2H2 O
the highest pH decreases. However, after this time, the final
pHs in each cycle were always higher than 4.0 and bacteri- Glucose + 2ADP + 2Pi → butane-2, 3-diol + 2CO2
ocin production slowed down.
+ 2ATP + 2H2 O
As shown in Table 2, the results of this study showed that
the fed-batch and re-alkalized cultures (fermentations I and
II) enhanced pediocin production on MPW as compared to Glucose + 2ADP + 2Pi → 2ethanol + 2CO2
batch fermentations on MRS and MPW (non-supplemented +2ATP + 2H2 O
and supplemented with a 2% of yeast extract or Casitone).
The values of YPed/TSc were lower in the two cultivations From these equations, it possible to obtain expressions
with steps of pH with respect to the four batch cultures. for determining the glucose utilization rate (rG ) for biomass
This could probably be due to an overconsumption of sug- (X), by-product formation (lactic acid (LA), butane-2,3-diol
ars to support higher growth, metabolite production (lactic (B) and ethanol (Et)), CO2 and maintenance energy:
acid, ethanol and butane-2,3-diol) as well as to compensate
for fermentation I:
the stress produced by the re-alkalizations each 8 h [24].
Contrarily, the values of YPed/TN in the fed-batch cultures 1 1 1
−rG = rLA + rEt + rB + mX
were higher than those values obtained in batch cultivations. YLA/G YEt/G YB/G
This fact suggests a more economic utilization of the ni- 1 1
trogen sources for pediocin production in the fed-batch and + rCO2 + rX (1)
YCO2 /G YX/G
re-alkalized cultures.
With regard to nutrient consumptions, it can be appreci- for fermentation II:
ated that the efficiencies (ETSc and ETNc ) in the batch cul- 1 1 1
tures on MRS and MPW were lower than those obtained in −rG = rLA + mX + rCO2 + rX
YLA/G YCO2 /G YX/G
the two fed-batch cultures (Table 2). Thus, this last fermen-
tation technique contributes to obtaining lower amounts of (2)
nutrients (total sugars and nitrogen) in the medium at the
end of the fermentation. As reported by Murado et al. [38] The term mX accounts for the amount of glucose used
the mussel processing wastes represent a serious environ- for maintenance energy. The molar coefficients YLA , YB , YEt
mental problem in the Rias Baixas (Galicia, NW of Spain). and YCO2 were calculated from the above stoichiometric re-
Therefore, the use of this effluent as culture media for bac- lationships. The molar coefficient for biomass production
teriocin production could be a suitable treatment to reduce (YX ) was obtained from bibliography data and it was as-
its contamination effect. sumed to be 0.5. In the same way, rX was calculated assum-
The beneficial effect that this fed-batch fermentation ing a molecular weight for biomass of 25 g/mol [43].
technique with re-alkalization cycles produced on pediocin Since the experimental design used did not allow the cal-
production seems to be a common phenomenon among the culation of the glucose consumed for the maintenance and
lactic bacteria, since the same positive effects were observed the production of CO2 , both parameters were calculated
before for nisin production on TGE [12] and MPW [24]. jointly from the Eqs. (1) and (2) as follows:
for fermentation I:
3.4. Stoichiometry

1
It has been reported that the carbohydrates and the mX + rCO2
YCO2 /G
aminoacids: arginine (which is converted to ornithine with 
the production of 1 mol of ATP per mol of arginine utilised) 1 1 1
= −rG − rLA + rEt + rB
and serine are the only energy sources for lactic acid bac- YLA/G YEt/G YB/G

teria [39,40]. Taking into account the low carbohydrate 1
content (5 g/l) in the MPW medium (fermentations I and + rX (3)
YX/G
II), it is reasonable to suppose that P. acidilactici obtained
additional ATP from the catabolism of both amino acids, for fermentation II:
which are present in this medium [41], for both growth and  
1
maintenance. mX + rCO2
YCO /G
In an attempt to prove the above observations, the theo-  2 
retical glucose utilization by P. acidilactici in MPW cultures 1 1
= −rG − rLA + rX (4)
was determined by using the corresponding metabolic reac- YLA/G YX/G
1078 N.P. Guerra et al. / Process Biochemistry 40 (2005) 1071–1083
Fig. 4. Glucose utilization rates (γ G ) for both the maintenance and CO2
production, and nitrogen consumption rates (NCR) determined from the
re-alkalized and fed-batch cultures of Pediococcus acidilactici NRRL
B-5627 on MPW. (A) Fermentation I and (B) fermentation II. In the
dotted lines γ G = 0.
Then, the parameter (mX + rCO2 (1/YCO2 /G )) was named
as γ G and plotted versus the time of fermentation (Fig. 4).
The value of γ G was always negative for the two fed-batch
fermentations (upper parts of Fig. 4A and B), indicating that
the glucose levels supplied were not enough to support both
growth and metabolite production by P. acidilactici through-
out the fermentations I and II. In addition, it can be noted
that the nitrogen consumption rate (NCR) increased (lower
parts of Fig. 4A and B) as the parameter γ G decreased, sug-
gesting that P. acidilactici NRRL B-5627 utilized the argi-
nine and serine to obtain additional energy for growth and
maintenance.
From the above discussion, it can be pointed out that the
nitrogen consumption rate is an important variable that may
be taken into account in the development of a model for de-
scribing the kinetics of biomass production in the fed-batch
fermentations on MPW. In addition, since the phosphate
source seems to play an important role on the growth of P.
acidilactici NRRL B-5627, this variable was considered in
the overall model developed for cell growth.
3.5. Development of the kinetic models
In this work, the volume of medium in the two fed-batch
fermentations was maintained constant ((dV/dt) = 0) by
matching the feeding volume with the sampling volume.
Biomass and pediocin production in the fermentor are
given by the following expressions:
   
VS
X(tn ) = X(tn−1 ) 1 − + rX dt (5)
VF
   
VS
Ped(tn ) = Ped(tn−1 ) 1 − + rPed dt (6)
VF
where X(tn ) and Ped(tn ) are the biomass (in g/l) and pediocin
(in AU/ml) produced at the end of each re-alkalization pe-
riod. X(tn−1 ) and Ped(tn−1 ) are the biomass (in g/l) and pe-
diocin (in AU/ml) at the beginning of each re-alkalization
period. VF is the volume (in l) of medium in the fermen-
tor and VS is the volume (in l) of sampling. rX and rPed
are the biomass and pediocin production rates and dt is the
re-alkalization period (in h).
Usually, logistic-type models and Luedeking and Piret
model [28] have been, respectively, used to model bacte-
rial growth and bacteriocin production in batch fermenta-
tion [44,45]. However, these models have been found to be
unsatisfactory to describe biomass and bacteriocin produc-
tion in fed-batch cultures [26] and in fed-batch cultures with
re-alkalization cycles [12]. This indicates that these mod-
els did not include all the variables affecting both biomass
and bacteriocin production. For these reasons, some modi-
fications for both the logistic-type and Luedeking and Piret
models have been suggested recently [12,26].
As was discussed above, the growth rate of P. acidilactici
growing in MPW medium seems to be a function of pH
change, nutrient limitation (nitrogen and phosphorous) and
product inhibition (lactic acid, ethanol and butane-2,3-diol),
variables that change in concentration in the course of the
fermentation. In addition, pediocin production seemed to
be affected by the final pH reached in each re-alkalization
cycle. Therefore, in a first step, an overall equation was set
up to describe the effects of the main factors affecting the
growth of P. acidilactici. A modification of the Luedeking
and Piret model including a term for the specific effect of
pH on pediocin production was then developed.
Although these models have little biological meaning be-
cause their empirical nature, they can be used to develop
an appropriate control strategy for optimizing the fed-batch
production of pediocin.
3.5.1. Effect of pH on the growth of Pediococcus acidilactici
Although the optimal pH for growth and product forma-
tion by lactic acid bacteria has found to be around 6.0, a rapid
 N.P. Guerra et al. / Process Biochemistry 40 (2005) 1071–1083 1079

decline in the growth is produced as pH decrease [15,46–49]. With regard to lactic acid, there are discrepancies related
Thus, an inadequate pH can cause the precipitation of nec- to its main mechanism of inhibition on cell growth. Thus,
essary nutrients, ions, etc. the denaturation of enzymes and some researchers consider that the inhibitory action lies in
structural proteins or cause membrane disruption or inacti- the presence of both the undissociated and dissociated forms
vation of transport or other protein functions [46,50–52]. of the molecule [48,49,57], but others believe that the inhi-
Since the growth rate is a function of pH [53], the model bition is produced only by the undissociated form [58,59],
proposed by Presser et al. [48], was set up to describe the or by the total concentration of the acid [52].
relationship between both variables, as follows: Taking into account these discrepancies, the inhibitory
  effect of the undissociated and dissociated forms of lactic
10pHmin
RX = b0 1 − ; being : b0 = rX × 10−pHmin acid on cell growth as proposed by Presser et al. [48] were
10pHt considered in the following form:
(7)

[LA]
where RX is the growth rate (g l−1 h−1 ) affected by the pH, γ[ULA] = d 1 −   ;
[ULAmin ] 1 + 10pHt −pKa
pHmin is the minimum pH value beyond which no growth is
possible, pHt is the pH over time. rX is the absolute growth being : d = c0 ULAmin (12)
rate (in g l−1 h−1 ) in each re-alkalization cycle.
The numeric integration of Eq. (7) with respect to time  
[LA]
provided the estimated values of biomass concentration over γ[DLA] = p 1 − ;
[DLAmin ](1 + 10pKa −pHt )
time:
  being : p = c1 DLAmin (13)
t=t
 t=t
10pHmin
XR = RX = b0 1 − (8) where γ[ULA] and γ[DLA] are the inhibitory functions for
10pHt
t=0 t=0 the inhibition of the undissociated and dissociated forms of
where XR is the biomass concentration (g/l) and t is the time the lactic acid, c0 and c1 are constants of proportionality to be
(h). Other terms are as previously defined. experimentally determined. [LA] is the total concentration
of lactic acid, [ULAmin ], [DLAmin ] are, respectively, the
3.5.2. Effect of nitrogen and phosphorus limitation minimum concentrations of undissociated and dissociated
Biomass production was strongly influenced by the time forms of the acid which completely prevents the growth, pKa
course of both the nitrogen and phosphorus consumption is the pH at which the concentrations of the two forms are
profiles (Figs. 2 and 3). From the observation that biomass equal and pHt is the final pH in each re-alkalization cycle.
production and both nitrogen and phosphorous consumption Including the effects of the undissociated and dissociated
are directly correlated (Figs. 2 and 3), the simplest forms to species of lactic acid in Eq. (11) gives:
describe these relationships are: t=t
 t=t
  
10pHmin
t=t
 XR = RX = b0 1− (1 + b1 rNc )
RX = rX (1 + b1 rNc ) (9) 10pHt
t=0 t=0
t=0 × (1 + b2 rPc )γ[ULA]γ[DLA] (14)
t=t
 In the same way, the inhibitory effect of the total con-
RX = rX (1 + b2 rPc ) (10) centration of lactic acid was modelled using the following
t=0
expression [53,59]:
where rNc and rPc are the nitrogen and phosphorous con-  
[LA]
sumption rates, b1 and b2 are constants of proportionality. γ[LA] = 1 − (15)
[LAmax ]
Combining the effects of pH and the rates of nitrogen and
phosphorous consumption, the overall form of the model is: where γ[LA] is an inhibition function that accounts for
t=t
 t=t
   the inhibitory effect of the total lactic acid concentration,
10pHmin [LAmax ] is the maximum total lactic acid concentration from
XR = RX = b0 1 − (1 + b1 rNc )
10pHt which the growth stopped and [LA] is as previously defined.
t=0 t=0
× (1 + b2 rPc ) (11) Therefore, including Eq. (15) into model (11):
t=t
 t=t
  
10pHmin
3.5.3. Effect of product formation XR = RX = b0 1− (1 + b1 rNc )
Taking into account that product accumulation (or- 10pHt
t=0 t=0
ganic acids (mainly lactic and acetic acids), ethanol and × (1 + b2 rPc )γ[LA] (16)
butane-2,3-diol) produces inhibitory effects on bacterial
growth [54–56], in this work, the effects of these variables The ability of ethanol to produce perturbation in cell mem-
were also considered in the growth model. branes or the alteration in fatty acid composition contributes
1080 N.P. Guerra et al. / Process Biochemistry 40 (2005) 1071–1083

to an increase in the uptake of the undissociated form of lac- 3.5.4. Specific effect of the optimum final pH value on
tic acid [55,56]. On the other hand, it has been reported that pediocin production
butane-2,3-diol is able to deactivate enzymes from several As was discussed above, in the two fed-batch cultures of
microorganisms [60]. P. acidilactici, pediocin synthesis was strongly dependent on
The negative influence of the concentrations of ethanol or both biomass production and the final pH value reached at
butane-2,3-diol produced in the cultures may be described the end of each re-alkalization period. It was hypothesised
in the same form to that given for the effect of the total lactic that the increase in pediocin concentration is proportional to
acid concentration, so that: the increase in the pH drop rate (∂pH) which is less than the
  maximum ∂pH from which bacteriocin synthesis stops. To
[Et]
γ[Et] = 1 − (17) describe the effects of ∂pH on rPed , the equation proposed
[Etmax ]
by Guerra and Pastrana [19] was used:
 
γ[B] = 1 −
[B]
(18) rPed (∂pH) = rPed e−b3 (∂pHmax −∂pH) (25)
[Bmax ]
where rPed (∂pH) is the pediocin production rate affected by
where γ[Et] and γ[B] are the inhibition functions for ethanol the pH drop rate in each re-alkalization cycle, b3 a constant
and butane-2,3-diol inhibition. [Et] and [B] are the total con- of proportionality, ∂pHmax is the maximum decrease in pH
centrations (mM) of ethanol and butane-2,3-diol. [Etmax ] and per unit of time, from which bacteriocin production stopped.
[Bmax ] are, respectively, the maximum concentrations (mM) The combination of Eqs. (24) and (25) yields Eq. (26):
of ethanol and butane-2,3-diol, which completely prevent
t=t
 t=t

the growth. Including the effects of both variables in mod-
els (14) and (16), biomass formation could be expressed by Ped = rPed (αRX + βXR )(1 + e−b3 (∂pHmax −∂pH) )
t=0 t=0
the following equations:
(26)
t=t
 t=t
  
10pHmin Then, the effects of all factors affecting biomass produc-
XR = RX = b0 1 − (1 + b1 rNc )
10pHt tion by P. acidilactici were tested by fitting models (21) and
t=0 t=0
× (1 + b2 rPc )γ[ULA]γ[DLA]γ[B]γ[Et] (22) to the experimental cell growth data. Using the biomass
formation estimates obtained from these two models, pe-
(19) diocin production was then modelled using the models (24)
and (26).
t=t
 t=t
   For optimal estimation of the model parameters, a
10pHmin
XR = RX = b0 1− (1 + b1 rNc ) non-linear regression technique (method of Newton by
10pHt
t=0 t=0 using a minimal non-linear squares method), was used to
× (1 + b2 rPc )γ[LA]γ[B]γ[Et] (20) minimise the deviations between the model predictions and
experimental fed-batch data.
Then, if µ was constant, Eqs. (19) and (20) could be
expressed in the integral form:
K
XR = (21)
1 + e(c−µ((1−(10 min /10 t ))(1+b1 rNc )(1+b2 rPc )γ[ULA]γ[DLA]γ[B]γ[Et])t)
pH pH

K
XR = (22)
1 + e(c−µ((1−(10 min /10 t ))(1+b1 rNc )(1+b2 rPc )γ[LA]γ[B]γ[Et])t)
pH pH

The introduction of the values of RX and XR into the

Luedeking and Piret expression [28] results in: As it can be seen in Figs. 5 and 6 and Table 3, mod-
els (21) and (22) described the trend well. From Table 3,
rPed = αRX + βXR (23) it can be noted that the inhibitory action of the dissociated
form of the lactic acid was lower than the inhibitory effect
where rPed is the pediocin (Ped) production rate, α a produced by the undissociated specie, since a higher con-
growth-associated constant (BU/mg) that corresponds to centration of the dissociated form was needed to inhibit the
the yield product on biomass formed (YB/X ) for primary growth of P. acidilactici. Nevertheless, the inhibitory con-
metabolites, and β is the non-growth-associated constant centrations of both the undissociated and dissociated forms
(BU mg−1 h−1 ). of the acid predicted by model (21) were always higher than
Therefore, numeric integration of Eq. (23) results in those amounts produced in the two fed-batch fermentations
Eq. (24), which provides the pediocin estimates (Ped): (Tables 3 and 4).
t=t
 t=t
 When the inhibitory action is considered to be due to the
Ped = rPed = (αRX + βXR ) (24) total concentration of lactic acid, the maximum inhibitory
t=0 t=0 concentration obtained from model (22) was also higher
 N.P. Guerra et al. / Process Biochemistry 40 (2005) 1071–1083 1081

Table 3
Parameters in the models for fed-batch production of biomass (models 21 and 22) and pediocin (model 26) by Pediococcus acidilactici NRRL B-5627
in fermentations I and II
Parameter Fermentation I Fermentation II

Eqs. (21) and (26) Eqs. (22) and (26) Eqs. (21) and (26) Eqs. (22) and (26)

K (g/l) 1.89 1.94 2.89 2.94

µ (h−1 ) 0.055 0.058 0.047 0.031
pHmin 3.86 3.88 3.96 3.88
b1 0.31 0.36 0.31 2.67
b2 3.15 3.70 1.73 0.45
ULAmin (mM) 68.90 – 68.92 –
pKa 3.86 – 3.86 –
DLAmin (mM) 298.35 – 359.30 –
LAmax (mM) – 277.93 – 519.85
Etmax (mM) 150.45 150.45 150.45 150.45
Bmax (mM) 438.97 341.26 438.97 341.26
α (BU/mg) 2255.78 2253.55 2725.80 2620.54
β (BU/mg−1 h−1 ) 0 0 0 0
b3 10.89 11.54 4.22 3.33
∂pHmax 0.273 0.273 0.274 0.274
rX
2 0.9858 0.9851 0.9880 0.9861
rPed
2 0.9847 0.9836 0.9710 0.9747

rX
2 and r 2
Ped are the correlation coefficients between expected and experimental biomass and pediocin values.

than those amounts produced by P. acidilactici in the two
fed-batch fermentations (Tables 3 and 4).
The inhibitory effects of both butane-2,3-diol and ethanol
in the first fed-batch culture seem to be very low, since
the inhibitory levels obtained from models (21) and (22)
were higher than those levels reached at the end of both
cultivations (Table 3 and Fig. 2).
On the other hand, pediocin production was best mathe-
matically described using Eq. (26) as compared to Eq. (24)
in both fermentations. From the values of ∂pHmax calcu-
lated, the minimum pH (pHmin ) from which bacteriocin pro-
duction stopped was determined as 4.03 in fermentations I
and II. This pH value is in accord with the optimal final pH
value (below 5.0) reported by other researchers for pediocin
production [20,21]. In addition, these observations support
the proposal of the existence of an optimum final pH for
bacteriocin production [13,14,20,21].
On the other hand, taking into account that in both
fed-batch cultures β = 0, and pediocin production was
dependent on the final pH reached in each re-alkalization

Fig. 5. Experimental data (symbols) of biomass (left part) and pediocin
(right part) production by Pediococcus acidilactici NRRL B-5627 on
MPW (fed-batch fermentation I). The curves drawn through the biomass
data were obtained according to the models 21 (A) and 22 (B). The
curves drawn through the pediocin data were obtained according to the
models 24 (dashed lines) and 26 (solid line). Notations as in Fig. 1.
Fig. 6. Experimental data (symbols) of biomass (left part) and pediocin
(right part) production by Pediococcus acidilactici NRRL B-5627 on
MPW (fed-batch fermentation II). The curves drawn through the biomass
data were obtained according to the models 21 (A) and 22 (B). The
curves drawn through the pediocin data were obtained according to the
models 24 (dashed lines) and 26 (solid line). Notations as in Fig. 1.
1082 N.P. Guerra et al. / Process Biochemistry 40 (2005) 1071–1083
Table 4
Concentrations of undissociated (U) and dissociated (D) forms of lac-
tic (LA) acid produced by Pediococcus acidilactici NRRL B-5627
at the end of both fed-batch fermentations calculated by using the
Henderson–Hasselbach equation
Fermentation I Fermentation II
ULA (mM) 1.53 17.49
DLA (mM) 211.50 241.40
TLA (mM) 213.03 258.89
Et (mM) 44.66 – [10]
B (mM) 157.76 – bacteriocins from lactic acid bacteria. Appl Microbiol Biotechnol
TLA is the total concentration of lactic acid. Et and B are the concentrations
of ethanol and butane-2,3-diol reached at the end of the cultures.
cycle, this bacteriocin can be classified as a primary metabo-
lite pH dependent.
4. Conclusions
The kinetic data from fed-batch and re-alkalized cultiva-
tion of P. acidilactici NRRL B-5627 were used to develop
mathematical models for describing both growth and pe-
diocin productions. Model parameters were obtained by the
non-linear regression technique assisted by a computer pro-
gram. Excellent agreements were found between fed-batch
experimental data and predictions from the proposed mod-
els. This showed that the developed models for biomass
and pediocin production by P. acidilactici NRRL B-5627
can be used to design feeding strategies and develop a
control-system for fed-batch pediocin fermentation.
Acknowledgements
The research presented in this paper was financially
supported by the Instituto Nacional de Investigación y
Tecnologı́a Agraria y Alimentaria (INIA), Spain (project
CAL01-045-C2-2) and The Xunta de Galicia, Spain (project
PGIDT00BIO1E).
References
[1] Abee T, Krockel L, Hill C. Bacteriocins: modes of action and poten-
tials in food preservation and control of food poisoning. Int J Food
Microbiol 1995;28:169–85.
[2] Tagg JR, Dajani AD, Wannamaker LW. Bacteriocins of gram-positive
bacteria. Bacteriol Rev 1976;40:722–56.
[3] Marugg JD. Bacteriocins, their role in developing natural products.
Food Biotechnol 1991;5:305–12.
[4] Barefoot SF, Nettles CG. Antibiosis revisited: bacteriocins produced
by dairy starter cultures. J Dairy Sci 1993;76:2366–79.
[5] Bhunia AK, Johnson MC, Ray B. Purification, characterization and
antimicrobial spectrum of a bacteriocin produced by Pediococcus
acidilactici. J Appl Bacteriol 1988;65:261–8.
[6] Kalchayanand N, Hanlin MB, Ray B. Sublethal injury makes
Gram-negative and resistant Gram-positive bacteria sensitive to
the bacteriocins pediocin AcH and nisin. Lett Appl Microbiol
1992;15:239–43.
[7] Kalchayanand N, Sikes T, Dunne CP, Ray B. Hydrostatic pressure and
electroporation have increased bactericidal efficiency in combination
with bacteriocins. Appl Environ Microbiol 1994;60:4174–7.
[8] Delves-Broughton J, Blackburn P, Evans RJ, Hugenholtz J. Applica-
tions of the bacteriocin, nisin. Antonie van Leeuw 1996;69:193–202.
[9] Soomoro AH, Masud T, Anwaar K. Role of lactic acid bacteria
(LAB) in food preservation and human health. A review. Pakistan J
Nutr 2002;1:20–4.
Parente E, Ricciardi A. Production, recovery and purification of
1999;52:628–38.
[11] Leroy F, De Vuyst L. Growth of bacteriocin-producing Lactobacillus
sakei strain CTC 494 in MRS broth is strongly reduced due to
nutrient exhaustion: a nutrient depletion model for the growth of
lactic acid bacteria. Appl Environ Microbiol 2001;67:4407–13.
[12] Cabo ML, Murado MA, González Ma P, Pastoriza L. Effects of
aeration and pH gradient on nisin production. A mathematical model.
Enzyme Microb Technol 2001;29:264–73.
[13] Guerra NP, Rua ML, Pastrana L. Nutritional factors affecting the
production of two bacteriocins from lactic acid bacteria on whey.
Int J Food Microbiol 2001;70:271–85.
[14] Guerra NP, Pastrana L. Production of bacteriocins from Lacto-
coccus lactis subsp. lactis CECT 539 and Pediococcus acidilac-
tici NRRLB-5627 using mussel-processing wastes. Biotechnol Appl
Biochem 2002;36:119–25.
[15] Parente E, Ricciardi A, Addario G. Influence of pH on growth and
bacteriocin production by Lactococcus lactis subsp. lactis 140NWC
during batch fermentation. Appl Microbiol Biotechnol 1994;41:388–
94.
[16] Parente E, Brienza C, Ricciardi A, Addario G. Growth and bacte-
riocin production by Enterococcus faecium DPC1146 in batch and
continuous culture. J Ind Microbiol Biotechnol 1997;18:62–7.
[17] Matsusaki H, Endo N, Sonomoto K, Ishizaki A. Lantobiotic nisin
Z fermentative production by Lactococcus lactis IO-1: relationship
between production of lantibiotic and lactate and cell growth. Appl
Microbiol Biotechnol 1996;45:36–40.
[18] Chichanoti N, Zaima T, Matsusaki H, Sonomoto K, Ishikazi A.
Relationship between nisin Z fermentative production and aeration
condition using Lactococcus lactis IO-1. J Fac Agric Kyushu Univ
1997;43:437–48.
[19] Guerra NP, Pastrana L. Influence of pH drop on both nisin and pe-
diocin production by Lactococcus lactis and Pediococcus acidilactici.
Lett Appl Microbiol 2003;37:51–5.
[20] Biswas SR, Ray P, Johnson MC, Ray B. Influence of growth con-
ditions on the production of a bacteriocin, pediocin AcH, by Pedio-
coccus acidilactici H. Appl Environ Microbiol 1991;57:1265–7.
[21] Yang R, Ray B. Factors influencing production of bacteriocins by
lactic acid bacteria. Food Microbiol 1994;11:281–91.
[22] Yang R, Johnson MC, Ray B. Novel method to extract large amounts
of bacteriocins from lactic acid bacteria. Appl Environ Microbiol
1992;58:3355–9.
[23] Ennahar S, Aoude-Werner D, Sorokine O, Van Dorsselaer A, Bringel
F, Hubert J-C, et al. Production of pediocin AcH by Lactobacillus
plantarum WHE92 isolated from cheese. Appl Environ Microbiol
1996;64:3275–81.
[24] Guerra NP, Pastrana L. Enhancement of nisin production by Lacto-
coccus lactis in periodically re-alkalized cultures. Biotechnol Appl
Biochem 2003;38:157–67.
[25] Paik HD, Glatz BA. Enhanced bacteriocin production by Pro-
pionibacterium thoenii in fed-batch fermentation. J Food Prot
1997;60:1529–33.
[26] Callewaert R, De Vuyst L. Bacteriocin production with Lactobacil-
lus amylovorus DCE 471 is improved and stabilized by fed-batch
fermentation. Appl Environ Microbiol 2000;66:606–13.
 N.P. Guerra et al. / Process Biochemistry 40 (2005) 1071–1083
[27] De Vuyst L, Vandamme EJ. Microbial manipulation of nisin biosyn-
thesis and fermentation. In: Jung G, Sahl HG, editors. Nisin novel
lantibiotics. Leiden, The Netherlands: ESCOM Science Publishers
BV; 1991. p. 397–409.
[28] Luedeking R, Piret EL. Transient and steady states in continuous
fermentation. Theory and experiment. J Biochem Microbiol Technol
Eng 1959;1:431–59.
[29] Cabo ML, Murado MA, González Ma P, Pastoriza L. A method for
bacteriocin quantification. J Appl Microbiol 1999;87:907–14.
[30] Van Niel EWJ, Hahn-Hägerdal B. Nutrient requirements of lactococci
in defined growth media. Appl Microbiol Biotechnol 1999;52:617–
27.
[31] Nes IF, Diep DB, Havarstein LS, Brurberg MB, Eijsink V, Holo H.
Biosynthesis of bacteriocins in lactic acid bacteria. Anton Leeuw Int
J G 1996;70:113–28.
[32] Kim WS, Hall RJ, Dunn NW. The effect of nisin concentration and
nutrient depletion on nisin production of Lactococcus lactis. Appl
Microbiol Biotechnol 1997;48:449–53.
[33] De Vuyst L, Vandamme EJ. Influence of the phosphorus and nitrogen
source on nisin production in Lactococcus lactis subsp. lactis batch
fermentations using a complex medium. Appl Microbiol Biotechnol
1993;40:17–22.
[34] Reiter B, Oram JD. Nutritional studies on cheese starters. J Dairy
Res 1962;29:63–77.
[35] Law BA, Sezgin E, Sharpe ME. Amino acid nutrition of
some commercial cheese starters in relation to their growth in
peptone-supplemented whey media. J Dairy Res 1976;43:291–300.
[36] Guerra NP, Pastrana L. Nisin and pediocin production on a mus-
sel processing waste supplemented with glucose and five nitrogen
sources. Lett Appl Microbiol 2002;34:114–8.
[37] De Vuyst L. Nutritional factors affecting nisin production by Lac-
tococcus lactis subsp. lactis NIZO 22186 in a synthetic medium. J
Appl Bacteriol 1995;78:28–33.
[38] Murado MA, Siso Ma IG, González Ma P, Montemayor Ma I, Pas-
trana L, Mirón J. Characterization of microbial biomasses and amy-
lolytic preparations obtained from mussel-processing waste treatment.
Biores Technol 1993;43:117–25.
[39] Yamada T, Carlsson J. Regulation of lactate dehydrogenase and
change of fermentation products in streptococci. J Bacteriol
1975;124:55–61.
[40] Garrigues C, Loubiere PN, Lindley D, Cocaign-Bousquet M. Con-
trol of shift from homolactic acid to heterolactic fermentation in
Lactococcus lactis: predominant role of the NADH/NAD+ ratio. J
Bacteriol 1997;179:5282–7.
[41] Murado MA, González Ma P, Pastrana L. Mussel processing wastes as
a fermentation substrate. In: Martin AM, editor. Fisheries processing:
biotechnological applications. London: Chapman and Hall; 1994.
p. 311–43.
[42] Cocaign-Bousquet M, Garrigues C, Lubiere P, Lindley N. Physiology
of pyruvate metabolism in Lactococcus lactis. Antonie van Leeuw
1996;70:253–67.
[43] Blanch HW, Clark DS. Biochemical engineering. USA: Marcel
Dekker; 1996.
[44] Lejeune R, Callewaert R, Crabbé K, De Vuyst L. Modelling the
growth and bacteriocin production by Lactobacillus amylovorus DCE
471 in batch cultivation. J Appl Bacteriol 1998;84:159–68.
1083
[45] Leroy F, De Vuyst L. Temperature and pH conditions that prevail dur-
ing fermentation sausages are optimal for production of the antilis-
terial bacteriocin sakacin K. Appl Environ Microbiol 1999;65:974–
81.
[46] Bibal B, Goma G, Vayssier Y, Pareilleux A. Influence of pH, lactose
and lactic acid on the growth of Streptococcus cremoris: a kinetic
study. Appl Microbiol Biotechnol 1989;28:340–4.
[47] Wijtzes T, De Wit JC, Huis in’t Veld JHJ, van’t Riet K, Zwi-
etering MH. Modelling bacterial growth of Lactobacillus curvatus
as a function of acidity and temperature. Appl Environ Microbiol
1995;61:2533–9.
[48] Presser KA, Ratkowsky DA, Ross T. Modelling the growth rate of
Escherichia coli as a function of pH and lactic acid concentration.
Appl Environ Microbiol 1997;63:2355–60.
[49] Åkerberg C, Hofvendahl K, Zacchi G, Hahn-Hägerdal B. Modelling
the influence of pH, temperature, glucose and lactic acid concen-
trations on the kinetics of lactic acid production by Lactococcus
lactis ssp. lactis ATCC 19435 in whole wheat flour. Appl Microbiol
Biotechnol 1998;49:682–90.
[50] Hofvendahl K, van Niel EWJ, Hahn-Hägerdal B. Effect of tem-
perature and pH on growth and product formation of Lactococcus
lactis ssp. lactis ATCC 19435 growing on maltose. Appl Microbiol
Biotechnol 1999;51:669–72.
[51] Poolman B, Konings WN. Relation of growth of Streptococcus lactis
and Streptococcus cremoris to amino acid transport. J Bacteriol
1988;170:700–7.
[52] Gonçalves LMD, Ramos A, Almeida JS, Xavier AMRB, Carrondo
MJT. Elucidation of the mechanism of lactic acid growth inhibition
and production in batch cultures of Lactobacillus rhamnosus. Appl
Microbiol Biotechnol 1997;48:346–50.
[53] Passos FV, Fleming HP, Ollis DF, Felder RM, McFeeters RF. Kinetics
and modelling of lactic acid production by Lactobacillus plantarum.
Appl Environ Microbiol 1994;60:2627–36.
[54] Jay JM. Antimicrobial properties of diacetyl. Appl Environ Microbiol
1982;44:525–32.
[55] Kadner RJ. Cytoplasmic membrane. In: Neidhardt FC, editor. Es-
cherichia coli and Salmonella: cellular and molecular biology. Wash-
ington, DC: American Society for Microbiology; 1996. p. 58–
87.
[56] Jordan SL, Glover J, Malcolm L, Thomson-Carter FM, Booth IR,
Park SF. Augmentation of killing of Escherichia coli O157 by com-
binations of lactate, ethanol, and low-pH conditions. Appl Environ
Microbiol 1999;65:1308–11.
[57] Venkatesh KV, Okos MR, Wankat PC. Kinetic model of growth and
lactic acid production from lactose by Lactobacillus bulgaricus. Proc
Biochem 1993;28:231–41.
[58] Ahamad MR, Marth EH. Behaviour of Listeria monocyto-
genes at 7, 13, 21, and 35 ◦ C in tryptose broth acidified
with acetic, citric or lactic acid. J Food Prot 1989;52:688–
95.
[59] Besser RE, Lett SM, Weber JT, Doyle MP, Barrett TJ, Wells JG, et
al. An outbreak of diarrhea and hemolytic uremic syndrome from
Escherichia coli o157:H7 in fresh-pressed apple cider. J Am Med
Assoc 1993;269:2217–20.
[60] Jay JM. Do background microorganisms play a role in the safety of
fresh food? Trends Food Sci Technol 1997;8:421–4.