Oxygen Supply Without GasLiquid Film

Resistance to Xanthomonas
campestris Cultivation
G. Sriram, Y. Manjula Rao, A. K. Suresh, G. K. Sureshkumar

Biochemical Engineering Group, Department of Chemical Engineering,

Indian Institute of Technology, Bombay, Powai, Mumbai 400 076, India;
telephone: 91-(22)-576 7208; fax: 91-(22)-578 3480;
e-mail: gksuresh@che.iitb.ernet.in
Received 27 August 1997; accepted 9 February 1998

Abstract: Alternative methods of oxygen supply are of

crucial importance, especially in viscous fermentations
and shear-sensitive fermentations. A method of oxygen
supply that completely eliminates the gasliquid transport resistance has been presented. The method involves
a need-based liquid-phase decomposition of hydrogen
peroxide to provide the necessary oxygen. When Xanthomonas campestris was cultivated (viscous cultivation) using this method of oxygen supply, dissolved oxygen (DO) levels were maintained above the setpoint of
50% throughout the cultivation, whereas the conventional cultivation was able to meet culture oxygen demand only for about 6 h in a 72-h fermentation. Furthermore, the maximum specific growth rate and xanthan
yields in the novel cultivation were 89% and 169%, respectively, of those obtained in conventional cultivation.
A mathematical model was also developed to simulate
and predict results in fermentations employing the presented methodology. In addition, studies with HOCl pretreatments indicated that monofunctional catalase may
be responsible for the decomposition of H2O2 supplied
externally to cells; HOCl pretreatments also increased the
tolerance of cells to H2O2. The decomposition kinetics of
externally supplied H2O2 was MichaelisMenten in nature with vmax = 1.196 106 M s1 and Km = 0.21 mM.
The catalase concentration was estimated to be 3.4
1010 mol/g of cells. 1998 John Wiley & Sons, Inc. Biotechnol Bioeng 59: 714723, 1998.

Keywords: oxygen-supply; bioreactors; H2O2 decomposition; xanthomonas; catalase; HOCl pretreatment

INTRODUCTION
The performance of aerobic bioreactors, especially on largescale, may be suboptimal due to inadequate oxygen supply
rates. Insufficient oxygen supply could lead to suboptimal
productivities (Shu and Yang, 1990) as well as products of
low quality (Suh et al., 1990). Inadequate oxygen supply
results due to the low solubility of oxygen in the medium
and the large gasliquid film resistance for oxygen transport, when aeration is used.
Researchers have tried to enhance the gasliquid oxygen
transport to meet the oxygen requirement of the culture
Correspondence to: G. K. Sureshkumar
Contract grant sponsor: Department of Science and Technology, India

1998 John Wiley & Sons, Inc.

through several means, such as different impeller or sparger

designs (Tecante and Choplin, 1993), modified reactor designs (Kawase and Tsujimura, 1992; Panda et al., 1989), by
using emulsified oxygen vectors (Rols, 1990) or through
hollow fiber membranes (Matsuoka, 1992). Another approach to enhance oxygen supply is to increase the partial
pressure of oxygen in the gas phase by increasing the oxygen composition in the inlet gas. This results in an increase
in the driving force and, therefore, leads to enhanced oxygen transport. However, this methodology may lead to undesirable changes in cell metabolism (Onken, 1990), loss of
culture growth, and hence, bioreactor productivity (Cacciuttolo et al., 1993; Liefke and Onken, 1992).
The methods just described have attempted to improve
oxygen transport rates, but still retain the basic limitation of
gasliquid oxygen transport. Hence, they have been only
partly successful and may not meet the requirement in
large-scale bioreactors and thus may make the systems
partly anaerobic. This could lead to formation of undesirable products in the system, which would question the economic viability of the process itself. Further, if either the
fermentation is viscous or the microorganisms are shearsensitive, such methods of enhancing oxygen availability
may be ineffective. Therefore, if oxygen transport can be
achieved with complete elimination of gasliquid transport,
it would greatly improve the economics of the process.
The methodology demonstrated in this article to overcome gasliquid transport resistance is through a liquid
phase conversion of hydrogen peroxide (H2O2) to oxygen
and water. This reaction is catalyzed by the enzyme catalase
available from the culture itself; no external catalase is
needed. In this case, the oxygen molecule being in liquid
phase is ready for consumption by cells and hence gas
liquid oxygen transport is nonexistent. Although H2O2 is
used to kill cells, the concentrations required for killing cells
are in the percent solution range, whereas the methodology
presented here employs H2O2 on a need basis, at concentrations at least 150 times lower. Therefore, cell death need
not be a concern.
The particular reaction for oxygen evolution was chosen
because it occurs naturally: hydrogen peroxide is generated

CCC 0006-3592/98/060714-10

in the cell, either directly by the divalent reduction of oxygen or indirectly by the dismutation of superoxide radicals
(Gonza`lez-Flecha and Demple, 1995); the generated H2O2
is scavenged through catalase (Fridovich, 1976). Also, most
aerobes contain catalase (Aebi, 1974), which is used as a
defense against H2O2 generated in the cell (Hassan and
Fridovich, 1984).
Thus far, hydrogen peroxide has not been used as an
oxygen source for bacterial cultivations although it has been
used for yeast cultivation (Zhang, 1994) with moderate success. The system of interest in this study is a bacterial,
aerobic, viscous fermentation system: Xanthomonas
campestris producing xanthan gum. Xanthan gum is an important product of the bioindustry with wide applications in
the food, pharmaceutical, oil, and other industries.

ter at 250 rpm with 1 L min1 air flow rate, was 28.8 h1 and
the mixing time in distilled water at 250 rpm was 18 seconds. The cell concentration was measured using a spectrophotometer (Jasco, Japan) through cell scatter at 650 nm,
using a calibration curve of cell concentration versus optical
density. The glucose concentration was determined using
the standard o-toluidine method. The xanthan concentration
was determined by dry weight measurement after isolation
through alcohol precipitation followed by centrifugation at
10,000g for 45 min (Garca-Ochoa et al., 1993).

MATERIALS AND METHODS

1. Feasibility experiment to demonstrate the availability of

oxygen from H2O2.
2. Xanthomonas campestris cultivation with aeration as the
oxygen source (referred to as conventional cultivation).
3. Xanthomonas campestris cultivation with H2O2 pulses as
the oxygen source (referred to as cultivation with
H2O2).
4. Experiments to estimate the catalase concentration in
cells.

Culture
The microorganism used in this study was Xanthomonas
campestris (MTCC 2286, IMTECH, Chandigarh, India)
producing xanthan gum. The stock cultures were maintained at 4C on agar slants containing 25 g L1 glucose, 8
g L1 yeast extract, 10 g L1 CaCO3, 5 g L1 K2HPO4, 0.6
g L1 MgSO4 7H2O and 20 g L1 agar. The growth medium consisted of 25 g L1 glucose, 8 g L1 yeast extract, 5
g L1 K2HPO4, 0.6 g L1 MgSO4 7H2O, and 0.4 g L1 urea
in tap water (to provide trace elements; however, the medium was sterilized by autoclaving without glucose before
inoculation). The glucose solution pH was adjusted to 4.0,
autoclaved, and was then added to the medium under aseptic
conditions (Brauer, 1985). The cells were grown at 30C in
a shaker incubator or in a 1-L bioreactor, depending on the
requirement.
Equipment, Reagents, and Analyses
It is commonly known that H2O2 is toxic to cells at high
concentrations. Therefore, all experiments employing H2O2
were performed at low enough concentrations of H2O2. The
experiments to check feasibility of oxygen availability from
H2O2 (S.D. Fine-Chem Ltd., Boisar, India) and cultivation
experiments were carried out in a 1-L bioreactor (fermentor)
(Vaspan Industries, Mumbai, India) equipped with temperature and rpm controls. The impeller was a Rushton turbine
type. The dissolved oxygen (DO) level was measured using
a polarographic probe (NBS Instruments) and pH was measured off-line with a pH probe (Control Dynamics Instrumentation, Bangalore, India). The data were acquired with a
data acquisition system (S.C.R. Elektroniks, Mumbai, India). The DO probe was calibrated before each experiment
with 0% corresponding to saturation with nitrogen and
100% corresponding to air saturation. The average volumetric oxygen transfer coefficient (kLa) of the bioreactor (measured using the dynamic response method), in distilled wa-

Bioreactor Experiments
The experiments in the bioreactor were carried out at 30C
and 250 rpm. Four types of experiments were performed in
the bioreactor:

The feasibility experiments were performed with mid-logphase cells obtained from shake flasks; Here, a 200 M
pulse of hydrogen peroxide was provided to two mediaa
diluted broth containing 0.78 g L1 cells (referred to as
medium with cells), and to a control broth that did not
contain cells, but only the cell-free supernatant, obtained
through centrifugation at 10,000g for 40 min (referred to as
medium without cells).
For the cultivation experiments the initial cell concentration was approximately 0.2 g L1. During cultivation with
H2O2, the magnitudes of the H2O2 pulses were chosen such
that, upon addition, they would result in H2O2 concentrations in the range 0.5 to 4.0 mM in the bioreactor. The
magnitude of the pulse was based on the growth phase of
the microorganism, and, consequently, the oxygen demand
of the culture. The inoculum was added to air-saturated
(100% DO) medium. When the DO level dropped to 50% a
2 mM H2O2 pulse was added during the cultivation with
H2O2. Subsequent H2O2 pulses were added manually when
DO was below the setpoint during the cultivation with
H2O2. The 50% DO value was chosen as the setpoint because it has been reported that there is no variation in the
specific xanthan formation rate between 20% and 80% DO
(Shu and Yang, 1990). Also, during earlier experiments in
which the setpoint was 80%, severe DO overshoots occurred, which consequently kept the DO above 100% air
saturation for most of the time (unpublished results) and
hence affected xanthan productivity. In either cultivation,
samples were drawn aseptically and analyzed for cell and
xanthan concentrations.

SRIRAM ET AL.: OXYGEN SUPPLY WITHOUT AERATION TO BIOREACTORS

715

HOCl Pretreatment Experiments

The HOCl pretreatment procedure for Escherichia coli, reported by Dukan and Touati (1996), was employed, in outline. However, the HOCl concentrations and growth phases
used were different in this study. Cells in the required
growth phase, at a concentration of 2 g L1, were resuspended in 0.05 M phosphate buffer, taken in several 25-mL
shake flasks. The flasks were thoroughly cleaned before
use, including a sulfachromic acid wash. Freshly prepared
hypochlorous acid, the concentrations of which were determined by iodometry, were added to the cultures in two
stages. The first stage consisted of a 15 mM HOCl treatment
followed by 40 min of incubation in the dark, at 30C, with
gentle shaking. The second stage was different for different
shake flasks. It consisted of adding either 49 mM, 65 mM,
or 80 mM HOCl to different shake flasks, followed by 40
min of incubation in the dark, at 30C, with gentle shaking.
After the second incubation, the free chlorine left in the
flasks was quenched by using sterile sodium thiosulfate.
Culturable bacteria were assayed by plating on LB agar
(Miller, 1992), and colonies were counted after 10 h of
incubation.
The live colonies were isolated and further grown in
shake flasks at 30C with an initial glucose concentration of
40 g L1. The other media components were the same as in
the growth medium mentioned previously. In addition, the
oxygen supply for these experiments was through liquid
surface air transfer in shake flasks.
All experiments described in various sections above were
at least done in triplicate. The simulations were carried out
using programs written in TURBOC++. These and other computational analyses were conducted on an IBM-PC 486 machine.
MATHEMATICAL MODEL
A model was developed, as explained next, to predict DO
levels in Xanthomonas campestris fermentation, when oxygen was supplied either by aeration or hydrogen peroxide
decomposition, or a combination of both.
A basic material balance of dissolved oxygen around the
fermenter gives the accumulation of dissolved oxygen in the
fermentation broth as the sum of the dissolved oxygen supplied by sparging and that generated by the decomposition
of peroxide, minus the oxygen consumed by the microorganisms for cell growth and maintenance. Thus:
dCO2
dt

= kLa CO*2 CO2 + rO2,(H2O2 decomposition)

rO2,(growth + maintenance)

(1)

The first term on the right represents the rate of mass transfer of oxygen from the gas to the liquid. If DO levels should
increase above 100% air saturation due to the H2O2 decomposition reaction, a similar term accounts for a desorption of
gas from the liquid. Although a large extent of reaction
could lead to the nucleation of additional bubbles in prin-

716

ciple, such a mechanism is not considered here, because the

H2O2 addition strategy, by design, would seek to minimize
such wasteful evolution of oxygen.
The availability of oxygen from H2O2 decomposition is
given by:
CEO CHOO + CH2O2
rO2,(H2O2 decomposition) = kapp
2

(2)

This arises from the following two-step mechanism for the

action of the enzyme catalase in decomposing H2O2, and the
corresponding kinetics reported by Schonbaum and Chance
(1976):
E
catalase
EO
compoundI

k1
EO
+ H2O2
compoundI + H2O

(3)

k2
E
+ H2O2
catalase + O2 + H2O

(4)

The reaction in Eq. (4) follows the kinetics in Eq. (2), for
which kapp
is the apparent rate constant. The concentration
2
of H2O2 was therefore taken to vary as:
dCH2O2
dt

= k1CH2O2 Ccatalase kapp

2 CEOCHOO + CH2O2

(5)
which is different from the kinetics normally assumed for
H2O2 decomposition (Aebi, 1974; Turhan and Uslan, 1990):
dCH2O2
dt

= k0CH2O2 Ccatalase

(6)

The HOO ion concentration was calculated from the equi+

librium H2O2 *
) H + HOO , and the equilibrium constant,
Ka, which was obtained from Schonbaum and Chance
(1976).
The pH of the medium was assigned values in the range
4.5 to 6.7, which were interpolated from experimental data.
The total concentration of the enzyme, E0 (i.e., CE +
CEO), which is a constant, is a fitted parameter in the model.
The best value of E0 which fit the data, as obtained through
a least-squares analysis, was used.
The following correlation, suggested by Hocker et al.
(1981), was used for the volumetric mass transfer coefficient, kLa (in the case of conventional cultivation):
.
V
kLa = 0.105
V

Pg
g
.
V L
L

0.67

0.6

LDL

0.3

(7)

The power required for agitation, Pg, will be less than the
ungassed power, Pf, which is supplied to the gas-free liquid.
However, due to lack of suitable estimates, a constant value
of Pg was assumed.
The broth viscosity, , was determined from (Herbst et
al., 1981):
= s1 + iPexpkiP

(8)

The product concentration, P, was estimated from the following logistic equation:

BIOTECHNOLOGY AND BIOENGINEERING, VOL. 59, NO. 6, SEPTEMBER 20, 1998

P = P0 + 0.869ePt 1

(9)

in which the value of p was estimated from shake-flask

experimental data.
The oxygen consumed for cell growth and maintenance
was given by (Pinches and Pallent, 1986):
rO2,growth + maintenance = c

dX
+eX
dt

(10)

The exponential phase growth expression was used for the

biomass growth, assuming no lag phase. Therefore:
X = X0e

maxt

(11)

The differential equations constituting the model were

solved by the fourth-order RungeKutta method. The parameter values used in the simulations are given in Table I.
RESULTS AND DISCUSSION
Inadequacy of Aeration
The dissolved oxygen (DO) profile when oxygen was supplied through aeration to a Xanthomonas campestris culture
is presented in Figure 1. It can be seen from the figure that
the DO drops to less than 20% in about 6 h when the broth
is initially 90% air saturated. It has been reported (Suh et al.,
1990) that lower than 20% DO levels have an adverse effect
on the xanthan productivities. Furthermore, Xanthomonas
campestris is an obligate aerobe (Pons et al., 1989) and
hence oxygen supply is necessary for its growth. Thus, the
conventional method of oxygen supply is inadequate for this
cultivation. It should be noted that the bioreactor kLa for the
conditions employed in the conventional cultivation was
28.8 h1, which compares well with 21.6 h1, obtained at a
Table I.

Values of parameters used in the mathematical model.

Parameter

Value

Basis data

EO
CH2O2
c
e
DL
Km
k1
kapp
2
P0
Pg
pKa
V

V
vmax
X0
L

0.228 nM
0.554 mM
1.84 102 mol g1
5.01 104 mol g1 h1
1.6 109 m2 s1
2.04 104 M
6 106 M1 s1
1.8 107 M1 s1
0.0 g L1
3600 W m3
11.6
1L
5 105 m3 s1
1.196 106 M s1
0.2 g L1
103 kg m3

max
P

0.111 h1
2.3 103 h1

This work
This work
Pinches and Pallent (1986)
This work
Bailey and Ollis (1986)
This work
Schonbaum and Chance (1976)
Schonbaum and Chance (1976)
This work
This work
Schonbaum and Chance (1976)
This work
This work
This work
This work
Reasonable assumption because
the product and cell
concentrations are small
This work
This work

Figure 1. Experimental and simulation results for DO variation in X.

campestris cultivation using aeration.

higher rpm (450 rpm), in another stirred tank used for xanthan gum studies reported in the literature (Li et al., 1995).
Thus, the kLa employed was a reasonable value. A higher
kLa would only mean that oxygen limitation occurs at a later
time, and thus kLa improvement, per se, for the conventional
cultivation, was not in the scope of this study.
The simulation results from the mathematical model developed are shown as a solid line in Figure 1. A good
agreement was obtained between simulation results and experimentally observed values.
As aeration is inadequate to meet the oxygen demand
even for 7 h it is desirable to study alternative means of
providing oxygen. In addition, one of the aims in this study
is to overcome the gasliquid film resistance which is the
largest resistance when aeration is employed for oxygen
supply.
Feasibility Demonstration of Oxygen Supply
Using Hydrogen Peroxide
The feasibility of providing oxygen through liquid phase
decomposition of hydrogen peroxide was studied by adding
a 1-mL pulse of hydrogen peroxide (H2O2) to growing Xanthomonas campestris culture in a bioreactor. The concentration of the pulse was adjusted such that the concentration
of H2O2 in the bioreactor was 200 M upon pulse addition.
The results are presented as change in DO versus time in
Figure 2. The pulse was added at time zero. As is evident
from Figure 2, there is an increase of about 20% DO in
about 60 min for the medium containing cells, but no significant increase for the control medium devoid of cells.
Because this experiment was performed in the absence of
aeration, the increase in DO for the medium with cells is
clearly due to oxygen availability from H2O2 decomposition
by cells. Other sources of DO increase (such as oxygen
transfer from the bioreactor headspace to an initially unsaturated broth) can be ruled out, because these would have
manifested in the case of the medium without cells also. In
addition, because X. campestris is an aerobe, the cells would

SRIRAM ET AL.: OXYGEN SUPPLY WITHOUT AERATION TO BIOREACTORS

717

model developed. The results of the simulation involving

need-based H2O2 pulses to the culture in its growth phase,
presented in Figure 3b, show that it is possible to maintain
DO at 50% through the proposed methodology. A one-time
supply of H2O2 cannot be made because high H2O2 concentrations are toxic to cells.
Cultivation with H2O2 Pulses

Figure 2. Feasibility experiment to demonstrate availability of oxygen

from H2O2 decomposition by cells.

have consumed oxygen during the experiment and therefore, the actual amount of oxygen generated is higher than
the observed (20%) increase in DO.
The cell concentration increased from 0.078 g L1 to
0.095 g L1 during the feasibility experiment and hence
H2O2, at the concentration added, did not cause culture
growth cessation. This suggests that the free radicals generated from 200 M H2O2 were either not enough to cause
growth cessation or were effectively rendered harmless by
the cells. It is known that cells possess effective defense
mechanisms against oxy-free radicals such as superoxide
dismutase (Halliwell and Gutteridge, 1985; Hassan and Fridovich, 1984).
Prediction of Oxygen Availability from a
H2O2 Pulse
When H2O2 is used as the sole oxygen source, the simulation results presented in Figure 3a show that a 550 M pulse
of H2O2 can sustain the oxygen demand of the culture for
about 45 min during the stationary phase. The pulse lasts for
about 25 to 35 min during the initial stages of the fermentation (simulation results not shown). However, if another
pulse of H2O2 is supplied when the DO falls to a setpoint
value, followed by such other need-based pulses, then oxygen demand can be expected to be met. The suitability of
such a methodology was verified by using the mathematical

Figure 3.

718

After the feasibility of providing oxygen through H2O2 decomposition was demonstrated, cultivations were carried
out using H2O2 as the sole oxygen source. Figure 4 depicts
the DO profiles for the H2O2-based cultivation and the conventional cultivation. It can be seen that, for the conventional cultivation, the oxygen supplied falls short of the
oxygen demand of the culture within 7.5 h; that is, within
barely 10% of the total fermentation time. In contrast, the
H2O2-based cultivation is able to maintain DO above the
setpoint (50%) for the entire length of the fermentation. The
decreases in DO below the setpoint value during the cultivation were chiefly because of human limitations (H2O2
pulses were added manually), and not because of limitations
of the methodology. This problem can easily be overcome
by automated additions of H2O2.
It was observed during conventional cultivation that there
was an increase in DO level from zero after about 50 h. This
may be due to formation of surfactants that resulted in surface tension or other liquid property changes and hence an
increased oxygen transfer coefficient (Hunt et al., 1971).
Also, it is of interest to note that, in other H2O2-based
cultivations, the DO level could be recovered back to the
setpoint even after it had been zero for over 30 min by a
H2O2 pulse addition (data not shown).
Starting with an initial cell concentration of around 0.2 g
L1, the maximum cell concentration in the H2O2-based
cultivation was 1.6 g L1, which was 55% of the maximum
cell concentration of 2.9 g L1 obtained in conventional
cultivation (Fig. 5a). However, the maximum specific
growth rate, for the H2O2-based cultivation, max of 0.098
h1, was 89% of that for the conventional cultivation, 0.11
h1 (Table II). Hence, the H2O2-based oxygen supply gave
specific growth rates comparable to conventional cultiva-

DO variation during X. campestris cultivation using H2O2: simulated results. (a) Single pulse; (b) multiple pulses.

BIOTECHNOLOGY AND BIOENGINEERING, VOL. 59, NO. 6, SEPTEMBER 20, 1998

Figure 4.

DO variation in conventional and H2O2-based cultivations.

tion, but the culture growth stopped earlier. Culture growth

cessation may have resulted from the free radicals generated
by H2O2 decomposition and work is already in progress in
our laboratory on aspects of free radicals in bioreactors,
with a view to improve cell yields.
In contrast, the maximum xanthan concentration obtained
in H2O2-based cultivation was 8.3 g L1 (Fig. 5b), which
was 69% higher than that obtained in conventional cultivation. Also, it can be seen from Table II that the xanthan gum
yield per unit substrate, YP/S, for the cultivation with H2O2,
was about 1.75 times higher than the conventional cultivation. The lower yield, and thus the lower productivity observed in the conventional cultivation, may be a result of the
severe oxygen limitation that had occurred during this cultivation. It is known that, under oxygen limitation, the specific xanthan production rate is proportional to the oxygen
transfer rate (Pons et al., 1989), which in turn is limited by
gasliquid oxygen transfer. As the cultivation with H2O2
overcame gasliquid transport resistance, and hence oxygen
limitation, higher xanthan yields were observed. As an
aside, it should be noted that in conventional cultivations the
actual magnitudes of the final xanthan concentrations were
dependent on the strain used and the medium composition
(Pinches and Pallent, 1986; Souw and Demain, 1979). However, here the comparisons between conventional and H2O2based cultivations were made under the same conditions,
except in regard to oxygen availability.
As the maximum cell concentration in the H2O2-based
cultivation was lower than in the conventional cultivation,
the higher xanthan yield in the H2O2-based cultivation may
have resulted from a shift to a more efficient biochemical
pathway for xanthan production, when oxygen is supplied
through H2O2 decomposition. This can be suspected from
the fact that the medium pH decreased from an initial value
of 6.5 to 4.7 in the cultivation with H2O2, whereas, in the
conventional cultivation, it did not drop below 6.0 (Fig. 6).
The decrease in pH could also have occurred due to the
formation of acid molecules during the biosynthesis of catalase. It is known that 3 mol of hydrogen ions are produced
per mole of catalase synthesized inside the cell (Dounce,
1983). An estimate of the magnitude of the pH drop from
this source, without any buffer effect considerations, suggests a decrease of only 0.008 units, and hence does explain

the decrease of 1.8 units. This suggests metabolic acid production by the microorganism, and hence a change in culture metabolism compared with the conventional cultivation
case.
Furthermore, no foam was observed during the H2O2based cultivation, whereas antifoam was needed in the conventional cultivation case. Therefore, an added advantage of
this methodology is the prevention of cell loss to foam.
During the cultivation with H2O2, the total volume of
30% (w/v) H2O2 solution used was 24.4 mL. This translates
into a raw material cost, for oxygen supply, of Rs. 0.71 or
US$ 0.02 per gram (on dry mass basis) of xanthan gum
produced.
At large scales, this methodology is expected to have
good applicability, because the power required to maintain
rpm at a level adequate to mix the broth is much lower than
the power required for rpm to maintain high kLa for an
adequate supply of oxygen through aeration alone.
Inferences on Catalasic Action When H2O2 Is
Supplied Externally to Cells
The oxygen supply strategy presented thus far assumed that
oxygen was made available to cells through H2O2 decomposition, which was catalyzed by catalase. Catalase is an
important protective enzyme against toxicity due to internally generated, metabolic H2O2 (Mongkolsuk et al., 1996).
Although catalasic action on metabolic H2O2 is known to a
reasonable extent, the catalasic action on externally supplied H2O2 is not known. Here we discuss the type of catalase that may be responsible for the decomposition of, and
its growth-phase-dependent induction in response to, externally supplied H2O2. Also, the decomposition kinetics of
externally supplied H2O2 is discussed. Another aspect of
interest examined is the H2O2 tolerance induced by HOCl in
Xanthomonas campestris.
Monofunctional Catalase May Be Responsible
There are three groups of bacterial catalases known: the
monofunctional catalases; the catalase-peroxidases (which
have a peroxidatic mode of reaction in which an organic
electron donor or halide ion is employed in the reduction of
H2O2); and the non-heme, Mn-containing catalases (Loewen, 1997). Among these catalases, we deduce here that the
monofunctional catalase may be the predominant catalase in
the breakdown of H2O2 supplied externally. In experiments
in which wild-type Xanthomonas campestris cells were
treated with HOCl, it was found that HOCl conferred higher
tolerance to H2O2. For example, HOCl-treated cells gave
more than 200 colonies on agar plates even in the presence
of 20% (w/v) H2O2, whereas the wild-type cells did not
give any colonies when plated under the same conditions.
It has been shown in another bacterium, E. coli, that the
genes mediating HOCl tolerance overlap those mediating
H2O2 tolerance (Dukan and Touati, 1996), namely the ones

SRIRAM ET AL.: OXYGEN SUPPLY WITHOUT AERATION TO BIOREACTORS

719

Figure 5. Time-profiles of cell concentrations (a) and xanthan gum concentrations (b) obtained from conventional (open circles) and H2O2-based (closed
circles) cultivations.

under the control of OxyR or rpoS transcriptional activators.

Also, it is known that these tolerance genes give rise to
amino acid sequences that are highly conserved among microbes (Mongkolsuk, 1996). Hence, it can be suggested that
the HOCl and H2O2 tolerance genes overlap in Xanthomonas campestris also. Furthermore, it is known that, in
other Xanthomonas strains such as Xanthomonas oryzae,
superoxide generators induce synthesis of 80 kDa monofunctional catalase (Chamnongpol et al., 1995). In addition,
it is well known that H2O2 is also a superoxide generator.
Therefore, it can be suspected that monofunctional catalase
is the predominant catalase in the breakdown of externally
supplied H2O2. However, the mechanism of catalase availability for decomposition of H2O2 supplied externally is not
very clear.
Catalase InductionStationary Phase versus
LogPhase
In bacteria, protective genes under the control of rpoS are
induced against the metabolic H2O2, generated when cells
enter the stationary phase (Dukan and Touati, 1996). Therefore, catalase activity is higher in the stationary phase of
most bacteria compared with their log phase. In contrast, in
Xanthomonas, catalase activity is highest in the log phase
(Mongkolsuk et al., 1996). This ability may have evolved in
Xanthomonas because it is a plant pathogen (Mongkolsuk et

al., 1990) and hence needs to survive oxidative species generated by the plant during its attack.
When Xanthomonas cells are cultivated, if catalase genes
are induced in the stationary phase also, then the cells can
better tolerate the ill effects of metabolic H2O2 and hence
the growth and product yield can be improved. To test this
possibility, we studied whether catalase can be induced in
the stationary phase cells of Xanthomonas campestris and
its effect on growth and xanthan production. Results of the
experiments in which HOCl treatments were made (see Materials and Methods) show that the treated stationary-phase
cells grew better (450 colonies) compared with wild-type
cells cells (100 colonies) under similar conditions (Table
III). This suggests that HOCl treatment induced catalase
genes in the stationary phase and hence the cells countered
metabolic H2O2 better. Also, the growth of stationaryphase-treated cells was better than the log-phase-treated
cells when the cells were treated with 15 mM HOCl followed by 80 mM HOCl (Table III). This observation may be
because, in the log phase, catalase is expected to be induced,
naturally, and hence HOCl treatment may have produced
only an increment in the induction level. Whereas, in the
stationary phase, catalase is not induced naturally and thus
HOCl treatment may have resulted in a higher catalase in-

Table II. Comparison of performance parameters between conventional

and H2O2-based cultivations.
Quantity

Conventional
cultivation

H2O2-based
cultivation

max (h1)
Yield per unit substrate (YP/S)
Yield per unit biomass (YP/X)
Final xanthan concentration (g L1)

0.111
0.1657
1.2740
4.89

0.098
0.2870
10.560
8.30

720

Figure 6. Medium pH variation in conventional and H2O2-based cultivations.

BIOTECHNOLOGY AND BIOENGINEERING, VOL. 59, NO. 6, SEPTEMBER 20, 1998

Table III Number of colonies obtained when wild-type and HOCltreated Xanthomonas campestris cells were grown on agar plates.

HOCl treatment

Treated logphase cells

Treated stationaryphase cells

None
15 mM + 80 mM

100 10
300 15

100 10
450 15

duction. However, when cells were treated with 15 mM

HOCl, followed by either 49 mM HOCl or 65 mM HOCl,
there was no difference in the number of colonies per plate
between the treated log-phase cells and treated stationaryphase cells (data not shown), although they were higher than
the number of colonies obtained with wild-type cells.
Results presented in Table IV show that the maximum
cell concentration reached when the HOCl-treated stationary-phase cells were grown in shake flasks was 5.72 times
that for untreated cells. The maximum cell concentration
reached with HOCl-treated log-phase cells was 2.86 times
that for untreated cells. The maximum xanthan concentrations also showed a similar trend1.6 times that for HOCltreated stationary-phase cells and 1.25 times that for HOCltreated log-phase cells compared with untreated cells. The
higher cell concentrations observed may be due to the enhanced ability of HOCl-treated Xanthomonas campestris
cells to counter the free radicals generated internally during
the stationary phase, as in the case of growth on agar plates.
The higher cell concentrations may have also resulted in
higher xanthan concentrations because xanthan is a secondary metabolite. Another possibility is that higher expression
of catalase genes also resulted in higher expression of xanthan genes, but this needs to be verified.

Figure 7. Variation of DO with time from a single H2O2 pulse in the

stationary phase of a H2O2-based cultivation: experimental (points) and
simulated (line) results.

was higher than the rate of oxygen production from H2O2

decomposition. For a first approximation, it was assumed
that the rate of oxygen production by H2O2 decomposition
was zero during the latter part of the pulse. (The DO profiles
in the latter part of the many pulses showed that the DO
decrease was linear.) Therefore, considering the oxygen uptake to be adequately described by Eq. (10), the constant, e,
was calculated from the slope of the DO profile (latter part)
and the experimental value of the cell concentration, for
various pulses. This was possible, because, during the stationary phase, no growth occurred and hence there was no
contribution from the growth-linked term in Eq. (10). The
constant, e, as evaluated in this manner, was found to be
5.01 104 mol g1 h1 5%.
When an alternative model for H2O2 decomposition
(Kremer, 1983) was used, it predicted the observed DO
variation better (thick line, Fig. 7) than the previously assumed kinetics. The mechanism is as follows:
k1,k2

E + H2O2 *
) EOI + H2O

Kinetics of H2O2 Decomposition:

an Alternative Model

k3
EOI
EOII

The experimental and the simulated variations of DO with

time, from a single H2O2 pulse, in the stationary phase
during H2O2-based cultivation, are presented in Figure 7.
The simulations were performed using the Schonbaum and
Chance model presented in Materials and Methods. The
agreement between the experimental and simulated profiles
(thin line, Fig. 7) is evident. The value of the oxygen consumption coefficient, e, was obtained from this DO profile,
as follows: During the latter part of the pulse, the DO decreased because the rate of oxygen consumption by cells
Table IV. Ratios of maximum cell concentrations and xanthan concentrations to those obtained in wild-type growth, when cultivated in shake
flasks.

Cell type

Maximum cell
concentration ratio

Xanthan
concentration ratio

Wild-type
Log-phase treated
Stationary-phase treated

1.00
2.86
5.72

1.00
1.25
1.60

k4
E + H2O + O2
EOII + H2O2

Making a quasi steady-state approximation for (EO)I and

(EO)II and assuming k1 k4 we get:

dH2O2
2k3 + k2 H2O2 E0
=
dt
H2O2 + 2k3 + k2 k1

which is of the MichaelisMenten form with vmax (2k3 +

k2)E0 and Km (2k3 + k2)/k1. The values of vmax 1.196
x 106 M s1 and Km 0.21 mM were obtained from a
least-squares fit.
Catalase Concentration
We estimated the catalase concentration available for externally supplied H2O2 decomposition by cells. At higher
H2O2 concentrations, the decomposition rate was found to
be independent of H2O2 concentration, a fact that is consistent with the MichaelisMenten kinetics described in the
previous section. Hence, a 3.5 mM H2O2 (much higher than

SRIRAM ET AL.: OXYGEN SUPPLY WITHOUT AERATION TO BIOREACTORS

721

g
Ka

Figure 8. Variation of estimated catalase concentration in the bioreactor

with cell concentration.

2Km) pulse was provided separately to four different cell

concentrations (1.192 g L1 was diluted to yield 0.894,
0.596, and 0.298 g L1), in separate bioreactors. The cells
were in the stationary phase.
Due to the nonavailability of a value for k2, Kremers
model was not used to estimate the catalase concentrations
and the estimate was made by employing least-squares
analysis with the Schonbaum and Chance model, as explained earlier. The catalase concentrations were plotted
against the cell concentration (Fig. 8) and the variation was
linear, as expected, because the cells are the source of catalase. The slope of the curve yields a catalase concentration
of 3.4 1010 mol/g of cells.
CONCLUSIONS
It is possible to employ H2O2 decomposition for oxygen
supply to Xanthomonas campestris cultivation. This method
of oxygen supply yielded lower cell yields, but higher xanthan yields compared with conventional cultivation under
otherwise similar conditions. However, the specific growth
rate in the H2O2-based cultivation was comparable to that
obtained in the conventional cultivation. Monofunctional
catalase may be responsible for decomposition of externally
supplied H2O2 and its expression can be altered through
pretreatment with HOCl. The MichaelisMenten parameters for externally supplied H2O2 decomposition were
found and the catalase concentration was estimated.
The authors thank A. Banerjee, Department of Chemical Engineering, I.I.T. Bombay, India, for his help.

NOMENCLATURE
C
CO2
C*O2
c
e
DL
E
EO
EOI, EOII
E0

722

concentration (of the suffixed species) (M)

dissolved oxygen (DO) in water (M)
saturation dissolved oxygen (DO) (from air) in water (M)
constant in Eq. (10) (mol O2 g1)
constant in Eq. (10) (mol O2 g1 s1)
fluid diffusivity (m2 s1)
catalase
catalaseH2O2 addition compound, compound I
isomers of compound I
total concentration of enzyme (CE + CEO) (M)

k0
k1, k2, k3, k4
kapp
2
kLa
k
P
Pg
rO2
S
t
V

V
vmax
X
YP/S
YP/X

acceleration due to gravity (m s2)

dissociation constant for the reaction H2O2 H+ +
HOO (M)
rate constant for first order decomposition of H2O2, (s1)
rate constants for second order reactions (M1 s1)
apparent rate constant for the reaction in Eq. (4) (M1 s1)
volumetric mass transfer coefficient (s1)
constant in Eq. (8) (L s1 g1 Pa1)
product (xanthan) concentration (g L1)
power input into liquid with dispersed gas, (W)
rate of oxygen availability/consumption (mol time1)
substrate (glucose) concentration (g L1)
time (s)
bioreactor dispersion volume (m3)
volumetric gas flow rate into bioreactor (m3 s1)
maximum (H2O2 decomposition) rate (M s1)
biomass concentration (g L1)
yield with respect to substrate (g xanthan g glucose1)
yield with respect to cell mass (g xanthan g cell1)
fluid viscosity (Pa s)
fluid intrinsic viscosity (Pa s)
(without subscript) specific growth rate (1/X dX/dt)
(h1)
(subscript P) specific product formation rate ( 1/P dP/
dt) (h1)
fluid density (kg m3)

Subscripts
0
P

initial value (at t 0)

product

Superscripts
max

maximum

References
Aebi, H. E. 1974. Catalase, pp. 273286. In: H. U. Bergmeyer (ed.), Methods of enzymatic analysis (vol. 2). Verlag Chemie, Weinheim.
Bailey, J. E., Ollis, D. F. 1986. Biochemical engineering fundamentals.
2nd edition. McGraw-Hill, New York.
Brauer, H. (ed.) 1985. pp. 619620. In: Biotechnology (vol. 2). VCH,
Weinheim.
Cacciuttolo, M. A., Trinh L., Lumpkin, J. A., Rao, G. 1993. Hyperoxia
induces DNA damage in mammalian cells. Free Rad. Biol. Med. 14:
267276.
Chamnongpol, S., Vattanaviboon, P., Lopraset, S., Mongkolsuk, S. 1995.
Atypical oxidative stress regulation of Xanthomonas oryzae pv. oryzae
monofunctional catalase. Can. J. Microbiol. 41: 541547.
Dounce, A. L. 1983. A proposed mechanism for the catalatic action of
catalase. J. Theor. Biol. 105: 553561.
Dukan, S., Touati, D. 1996. Hypochlorous acid stress in Escherichia coli:
resistance, DNA damage and comparison with hydrogen peroxide
stress. J. Bacteriol. 178: 61456150.
Fridovich, I. 1976. In: W. A. Pryor (ed.), Free radicals in biology (vol. 1).
Academic Press, New York.
Gonza`lez-Flecha, B., Demple, B. 1995. Metabolic sources of hydrogen
peroxide in aerobically growing Escherichia coli. J. Biol. Chem. 273:
1368113687.
Garca-Ochoa, Casas, J. A., Mohedano, A. F. 1993. Xanthan precipitation
from solutions and fermentation broths. Sep. Sci. Technol. 28:
13031313.
Halliwell, B., Gutteridge, J. M. 1985. Free radicals in biology and medicine. Clarendon Press, Oxford.
Hassan, H. M., Fridovich, I. 1984. Oxygen toxicity in prokaryotes. In: J. V.

BIOTECHNOLOGY AND BIOENGINEERING, VOL. 59, NO. 6, SEPTEMBER 20, 1998

Bannister and W. H. Bannister (eds.), The biology and chemistry of

active oxygen. Elsevier, New York.
Herbst, H., Schumpe, A., Deckwer, W. D. 1992. Xanthan production in
stirred tank fermentors. Chem. Eng. Technol. 15: 425432.
Hocker, H., Langer, G., Werner, U. 1981. Mass transfer in aerated newtonian and non-newtonian liquids in stirred reactors. German Chem.
Eng. 4: 5462.
Hunt, G., Reismann, H. B., Lago, J. 1971. Oxygen mass transfer in biological systems at zero dissolved oxygen concentration. AIChE Symp.
Ser. 67: 6065.
Kawase, Y., Tsujimura, M. 1994. Enhancement of oxygen transfer in
highly viscous non-newtonian fermentation broths. Biotechnol. Bioeng. 44: 11151121.
Kremer, M. L. 1983. Mechanism of catalase action. J. Chem. Soc. Faraday
Trans. 79: 21252131.
Li, G., Qiu, H. Zheng, Z., Cai, Z., Yang, S. 1995. Effect of fluid rheological
properties on mass transfer in a bioreactor. J. Chem. Tech. Biotechnol.
62: 385391.
Liefke, E., Onken, U. 1992. Influence of total and oxygen partial pressure
on growth and metabolism of Methylomonas clara. Biotechnol. Bioeng. 40: 719724.
Loewen, P. C. 1997. Bacterial catalases. In: Oxidative stress and the molecular biology of antioxidant defenses. Cold Spring Harbor Laboratory Press, Cold Spring, NY.
Matsouka, H. 1992. High oxygen transfer rate in a new aeration system
using a hollow fiber membrane. Biotechnol. Bioeng. 40: 346351.
Miller, J. H. 1992. A short course in bacterial genetics. In: A laboratory
manual and handbook for Escherichia coli and related bacteria. Cold
Spring Harbor Laboratory Press, Cold Spring, NY.
Mongkolsuk, S., Loprasert, S., Vattanaviboon, P., Chanvanichayachai, C.,
Chamnongpol, S., Supsamran, N. 1996. Heterologous growth phaseand temperature-dependent expression and H2O2 toxicity protection of
a superoxide-inducible monofunctional catalase gene from Xanthomonas oryzae pv. oryzae. J. Bacteriol. 178: 35783584.

Onken, U. 1990. Batch and continuous cultivation of Pseudomonas fluorescens at increased pressure. Biotechnol. Bioeng. 35: 983989.
Panda, A. K., Mishra, S., Bisaria, V. S., Bhojwani, S. S. 1989. Plant cell
reactorsa perspective. Enz. Microb. Technol. 11: 389397.
Pinches, A., Pallent, L. J. 1986. Rate and yield relationship in the production of xanthan gum by batch fermentation using complex and chemically defined growth media. Biotechnol. Bioeng. 26: 14841496.
Pons, A., Dussap, C. G., Gros, J. B. 1989. Modelling Xanthomonas
campestris batch fermentations in a bubble column. Biotechnol. Bioeng. 33: 394405.
Rols, J. C. 1990. Mechanism of enhanced oxygen transfer in fermentation
using emulsified oxygen vectors. Biotechnol. Bioeng. 35: 454461.
Schonbaum, G. R., Chance, B. 1976. Catalase. In: P. D. Boyer (ed.), The
enzymes (vol. 13). Academic Press, New York.
Souw, P., Demain, A. L. 1979. Nutritional studies on xanthan production
by Xanthomonas campestris NRRL B1459. Appl. Environ. Microbiol.
37: 11861192.
Shu, C-H., Yang, S-T. 1990. Effects of temperature on cell growth and
xanthan production in batch cultures of Xanthomonas campestris. Biotechnol. Bioeng., 35: 454468.
Suh, I-S., Herbst, H., Schumpe, A., Deckwer, W-D. 1990. The molecular
weight of xanthan polysaccharide produced under oxygen limitation.
Biotechnol. Lett. 12: 201206.
Tecante, A., Choplin, L. 1993. Liquid mass transfer in non-newtonian
fluids in a tank stirred with a helical ribbon screw impeller. Can. J.
Chem. Eng. 71: 859865.
Turhan, L., Uslan, A. H. 1990. Characterisation and operational stability of
immobilised catalase. Proc. Biochem. 25: 1418.
Zhang, H. 1994. Effects of endogenous and exogenous oxidative free radicals on cellular molecules, M.S. thesis, Drexel University, Philadelphia, PA, USA.

SRIRAM ET AL.: OXYGEN SUPPLY WITHOUT AERATION TO BIOREACTORS

723