Effects of Agitation Intensity on Mycelial

Morphology and Protein Production in

Chemostat Cultures of Recombinant
Aspergillus oryzae
A. Amanullah, R. Blair, A. W. Nienow, C. R. Thomas

Centre for Bioprocess Engineering, School of Chemical Engineering, The

University of Birmingham, Edgbaston, Birmingham B15 2TT, UK;
telephone: 44-121-414-5355; fax: 44-121-5324; e-mail: C.R.
THOMAS@bham.ac.uk
Received 19 February 1998; accepted 14 August 1998

Abstract: The effects of agitation on fragmentation of a

recombinant strain of Aspergillus oryzae and its consequential effects on protein production have been investigated. Constant mass, 5.3-L chemostat cultures at a dilution rate of 0.05 h1 and a dissolved oxygen level of
75% air saturation, have been conducted at 550, 700, and
1000 rpm. These agitation speeds were chosen to cover
a range of specific power inputs (2.2 to 12 kW m3) from
realistic industrial levels to much higher values. The use
of a constant mass chemostat linked to a gas blender
allowed variation of agitation speed and hence gas holdup without affecting the dilution rate or the concentration
of dissolved oxygen. The morphology of both the freely
dispersed mycelia and clumps was characterized using
image analysis. Statistical analysis showed that it was
possible to obtain steady states with respect to morphology. The mean projected area at each steady state under
growing conditions correlated well with the energy dissipation/circulation function, [P/(kD 3tc)], where P is the
power input, D the impeller diameter, tc the mean circulation time, and k is a geometric constant for a given
impeller. Rapid transients of morphological parameters
in response to a speed change from 1000 to 550 rpm
probably resulted from aggregation. Protein production
(-amylase and amyloglucosidase) was found to be independent of agitation speed in the range 550 to 1000
rpm (P/V = 2.2 and 12.6 kW m3, respectively), although
significant changes in mycelial morphology could be
measured for similar changes in agitation conditions.
This suggests that mycelial morphology does not directly affect protein production (at a constant dilution
rate and, therefore, specific growth rate). An understanding of how agitation affects mycelial morphology and
productivity would be valuable in optimizing the design
and operation of large-scale fungal fermentations for the
production of recombinant proteins. 1999 John Wiley &

Correspondence to: C. R. Thomas

organic acids, proteins and food. During the last two decades, filamentous fungi have been used increasingly as
eukaryotic hosts for foreign gene expression, for which they
have several attractions (Jeenes et al., 1991). Firstly, due to
their saprophytic life they are capable of secreting large
quantities of proteins (Van Brunt, 1986). Posttranscriptional
modifications of proteins such as glycosylation are important capabilities offered by these hosts (Mackenzie et al.
1993). In addition, many species are generally regarded as
safe by regulatory authorities.
Despite the widespread industrial use and potential of
fungal strains for heterologous protein production, relatively little is known about the influence of engineering
variables, such as agitation conditions, upon the morphology of such organisms in submerged cultures. In many fungal fermentations, the high apparent viscosities and the nonNewtonian behavior of the broths necessitate the use of high
agitation speeds to provide adequate mixing and oxygen
transfer. However, mycelial damage at high stirrer speeds
(or power input) can limit the acceptable range of speeds,
and hence, consequently, the oxygen transfer capability and
the volumetric productivity of the fermentor. The effects of
mechanical forces (shear) on fungal physiology, in particular branching, tip extension, and vacuolation, are poorly
understood.
Fungal morphology can be classified as dispersed or pelleted. The dispersed form can be divided further into freely
dispersed hyphae, and clumps (Paul and Thomas, 1998).
Although many studies have been conducted to investigate
the effects of mechanical forces on mycelial morphology
and productivity (Ayazi Shamlou et al., 1994; Dion et al.,
1954; Makagiansar et al., 1993; Metz et al., 1981; Nielsen et
al., 1995; Reuss, 1988; Smith et al., 1990; van Suijdam and
Metz, 1981; Ujcova et al., 1980), they generally suffer from
two limitations. First, due to the lack of suitable methods for
characterizing clumps (Tucker et al., 1992), only the freely
dispersed form has been considered, although it may only
account for only a small fraction of the biomass (Justen et
al., 1996; Tucker et al., 1992). Second, it has not been

1999 John Wiley & Sons, Inc.

CCC 0006-3592/99/040434-13

Sons, Inc. Biotechnol Bioeng 62: 434446, 1999.

Keywords: agitation; morphology; heterologous protein

production; Aspergillus oryzae

INTRODUCTION
The industrial importance of filamentous fungi is illustrated
by applications ranging across the production of antibiotics,

possible to dissociate the influence of agitation from mass

transfer effects. The dependence of product formation rates
on impeller shear has been observed for a wide variety of
filamentous fungi (Braun and Vecht-Lifshitz, 1991; Markl
et al., 1991; Merchuk, 1991; Smith et al., 1990; Vardar and
Lilly, 1982; Ujcova et al., 1980). There are also reports that
mycelial fragmentation is dependent on the physiological
state of the microorganisms (Smith et al., 1990; Paul et al.,
1994).
Using well-established image analysis methodologies
(Paul and Thomas, 1998; Tucker et al., 1992), Justen et al.
(1996) were able to make quantitative measurements on the
breakage of clumps and, therefore, to take into account the
influence of realistic agitation conditions on the whole of
the biomass. In off-line agitation studies, it was demonstrated that P. chrysogenum morphological data, using both
radial and axial flow impellers of very different geometries
and power numbers, could be correlated with an energy
dissipation/circulation function developed from the earlier
work of Smith et al. (1990) and Makagiansar et al. (1993).
This is defined as [P/(kD3tc)], where P is the power input, D
the impeller diameter, tc, the mean circulation time, and k is
a geometric constant for a given impeller. This function
arises from a consideration of the energy dissipation in the
impeller-swept volume and the frequency of mycelial circulation through that volume. Although other correlating
parameters, such as impeller tip speed and specific power
input, were also considered, they were inferior to the energy
dissipation/circulation function. The broader validity of
these correlations was also verified in fragmentation studies
at scales up to 180 L. Justen et al. (1998) modeled the
results of Justen et al. (1996), and suggested that clump
fragmentation was the main cause of morphological change,
and that the freely dispersed form was dominated by short
fragments originating from clumps. It should be recognized
that these studies were conducted in off-line vessels under
nongrowing conditions. However, the correlation of morphology and energy dissipation/circulation function may
also be applicable for fed-batch P. chrysogenum fermentations (Justen, 1997). There are no published data concerning
the use of this correlating variable for other species, nor on
production of metabolites other than penicillin.
The present study reports on the influence of agitation
conditions on the morphology and protein production capability of a genetically modified industrial strain of Aspergillus oryzae. This is an important issue, as it has been
suggested that protein secretion in fungi occurs only at the
tips (Wessels, 1990, 1993), so that there might be a strong
dependence of the secretion rate on mycelial morphology.
The aim of the experiments was to discover whether and
how recombinant protein production depended on morphology and, hence, impeller-induced fragmentation. This culture was grown in a chemostat using specific energy dissipation rates from around those used industrially to significantly higher values. A controlled dissolved oxygen level
was used throughout. Although there is very limited use of
continuous culture systems in industry, they are extremely

useful research tools. Steady-state continuous cultivation

can give precise information on the influence of a single
variable. Transient experiments, where the dynamic response of cultures to a perturbation can readily be studied,
are particularly useful in such systems. The industrial recombinant strain of Aspergillus oryzae used in this study
produces -amylase (homologous protein) and amyloglucosidase (heterologous protein). These two enzymes are
mainly used for the conversion of starch into sugars, syrups,
and dextrins. -Amylase is an endo-attacking enzyme hydrolyzing 1,4-- linkages in starch, whereas amyloglucosidase is an exo-attacking enzyme hydrolyzing 1,4-- as well
as 1,6-- linkages in starch.
MATERIALS AND METHODS
Media
Batch fermentation (in g/L): Maltodextrin, 12.0; citric acid,
2.0; MgSO4.7H2O, 2.0; KH2PO4, 2.0; K2SO2, 2.0; NH4SO4,
3.0; CaCl2, 0.8; yeast extract, 5.0; trace metal solution (citric acid, 3.0; ZnSO4, 0.29; Fe2SO4, 0.28; CuSO4, 0.25;
MnSO4, 0.25; NiCl, 0.05), 0.5 mL/L; Pluronic P6100, 1.0
mL/L.
Continuous culture feed (in g/L): Maltodextrin, 3.33;
NH4SO4, 8.0; K2SO4, 1.5; MgSO4.7H2O, 2.0, 1.0; NaCl,
0.5; CaCl2.2H2O, 0.1; trace metal solution (citric acid, 3.0;
ZnSO4, 0.29; Fe2SO4, 0.28; CuSO4, 0.25; MnSO4, 0.25;
NiCl, 0.05) 1.0 mL/L; Pluronic P6100, 0.25 mL/L.
The feed medium used in the chemostat contained an
excess of all but one of the nutrients essential for growth. To
avoid catabolite repression of -amylase and -glucosidase
production (Harvey and McNeil, 1994), maltodextrin,
which has a low content of mono- and disaccharides, was
used as a limiting carbon source to control the steady state
biomass concentration. Maltodextrin is a starch hydrolysate
containing approximately 0.5% glucose and 99.5% oligoand polysaccharides. The amyloglucosidase produced by
the mycelia degrades the latter to glucose, maltose, and
smaller oligosaccharides.
Shake Flask Cultures
The recombinant strain of Aspergillus oryzae was supplied
by Novo Nordisk A/S (Baegsvard, Denmark). Spores were
propagated from freeze-dried ampoules onto agar slopes.
Following incubation for 7 days at 30C, the slopes were
washed with 10 mL of 0.1% Tween-80 solution. The average spore concentration was 4.3 1.0 107 spores/mL. Nine
milliliters of the spore solution were used per 200 mL of
shake flask media (adjusted to pH 3.5 using 5% w/v aqueous H3PO4) resulting in a shake flask inoculum level of 1.9
0.4 109 spores/L medium. The shake flasks were placed
in an orbital shaker at 200 rpm and 30C for 24 h. Two
flasks were used to inoculate the fermentor. In these experi-

AMANULLAH ET AL.: MORPHOLOGY AND PROTEIN PRODUCTION OF ASPERGILLUS ORYZAE

435

ments, to avoid pellet formation by aggregation of spores,

the initial batch medium was inoculated with dispersed mycelia.
Chemostat Cultures
Fermentations were conducted in a 6-L bioreactor (LSL
Biolafitte, Luton, UK). The internal diameter of the vessel,
T, was 0.153 m and it had an approximate working volume
of 5.3 L. The base of the vessel was supported on a load cell.
Agitation was provided by two Rushton turbines with a D/T
ratio of 0.5. The lower impeller was 0.39T from the bottom
of the vessel, and the impeller spacing was adjusted to equal
the vessel diameter. A pipe sparger was used to aerate the
culture at a rate of 0.5 vvm. The broth temperature was
maintained at 32C (0.2C), and the pH at 5.0 (0.05)
using 10% w/v H3PO4 and 10% w/v NaOH solutions. It was
important to maintain good pH control, because the activities of the expressed enzymes are known to be pH sensitive.
Thus, the broth pH was confirmed by two independent in
situ pH probes. The readings of the probes never differed by
more than 0.05 pH units. A previously calibrated peristaltic
pump (32 rpm, Watson-Marlow 101U, Falmouth, UK) was
used to commence feeding at 12 h (at the end of the batch
phase) at a flow rate of 265 mL/h, resulting in a dilution rate
of 0.05 h1. To minimize growth of cells back up the feed
line during long-term chemostat operation, a double dripfeed assembly was incorporated into the feed line. The device prevents aerosols from entering the feed line and acts
as a break in feed flow. Feed, acid, base, and antifoam
additions were added via a branched fitting terminating in a
single tube in the zone of high-energy dissipation close to
the lower impeller. A second peristaltic pump (WatsonMarlow 505U) was used to withdraw broth from the fermentor. The exit pipe was located approximately 10 cm
below the broth surface near the upper impeller. The load
cell was linked to the WM 505U peristaltic pump via a TCS
(Turnbull Control Systems, Worthing, UK) control unit.
The speed of this pump was varied according to the output
from the load cell, allowing the fermentor to be operated as
a chemostat with a constant mass of 5.3 kg at a precision
greater than 0.02 kg. Thus, the dilution rate of the chemostat remained independent of gas hold-up effects caused by
changes in impeller speed.
Dissolved oxygen was measured by a polarographic
probe (Ingold, Switzerland) and controlled at 75% (1%) of
air saturation by blending nitrogen with air using a gas
blender during continuous fermentation. Thus, the agitation
speed could be varied without affecting either the dissolved
oxygen concentration in the broth or the gas flow rate. The
inlet and exit gases from the bioreactor were measured using a VG M8-80 mass spectrometer (VG Gas Analysis,
Middlewich, UK). The SETCIM (AspenTech, Boston, MA)
process management system was used for monitoring and
control. In an effort to reduce the possibility of wall growth
in the headspace encountered in preliminary trials, coolant
at 18C, provided by a chiller unit using 60% water/40%
ethylene glycol, was passed through a fine silicone tube

436

wound around the outside of the fermentor (above the level

of the broth, i.e., the headspace). Extra care was also taken
when setting up the fermentor, to thoroughly clean the walls
in the headspace, and also the inside of the headplate, to
remove any deposits that might aid attachment of the hyphae. In addition, the chemostat was operated with broth 1
cm from the top of the headplate.

Agitation Conditions
Fermentation 1 (C1). The agitation speed was maintained at
1000 (5) rpm (12 kW m3) until 348 h, when it was decreased to 550 (5) rpm (2.2 kW m3). This fermentation
was terminated at 650 h.
Fermentation 2 (C2). This was a repeat of C1 except that
the reduction in speed from 1000 to 550 rpm was made at
279 h. This fermentation was terminated at approximately
660 h.
Fermentation 3 (C3). Agitation speed was maintained at
700 (5) rpm (4.3 kW m3) until the end of the fermentation
at 250 h.
The choice for these speeds has already been discussed;
these speeds result in specific energy dissipation rates (2.2
to 12.6 kW m3) of the order of those used in industry, or
higher. The gassed power input at these speeds was measured separately in water (at 0.5 vvm) using a frictionless air
bearing dynamometer developed by Nienow and Miles
(1969). These measurements could be applied directly to the
agitation conditions in the chemostat, because the broth (2
g/L) was Newtonian with a viscosity similar to that of water.
The flow regime in all cases was turbulent (Reynolds numbers between 52,000 and 94,000). This is realistic, as in
practice, large-scale industrial fermentations, involving
even pseudoplastic broths, are often turbulent (or, at worst,
transitional) due to the use of large-diameter impellers.
Samples were taken periodically for measurements of
biomass concentration, morphological parameters and enzyme analysis. The sampling frequency was increased to
examine the dynamic response of the culture to a step
change in agitation speed. Samples were taken at 1, 5, 15,
30, and 60 min following the reduction in agitation speed
from 1000 to 550 rpm (in C1). Biomass concentration was
measured by filtering a known weight of broth using a preweighed filter paper, followed by drying at 100C for 24 h.
Morphology was characterized as described later, and enzymes were assayed as described next.

Enzyme Analysis
Twenty-milliliter broth samples were filtered using a syringe filled with Ballotini balls (1 to 2 cm in depth) and two
10-mL aliquots of the cell-free filtrate stored in freezer at
8C for later analysis of enzyme titers. -Amylase activity
was measured with a kit (MPR2 1442295) from Boehringer
Mannheim (Mannheim, Germany) using a computercontrolled Uvikon 922 double-beam spectrophotometer
(Kontron Instruments, Watford, UK). The frequency of

BIOTECHNOLOGY AND BIOENGINEERING, VOL. 62, NO. 4, FEBRUARY 20, 1999

sampling was 50/min. Approximately 200 readings of absorbance were made and the slope of the linear plot (absorbance vs. reaction time) was used to determine the enzyme
activity. Regression coefficients always remained >0.98.
The analysis was calibrated with an -amylase standard
from Novo Nordisk. The activity of the enzyme is expressed
arbitrarily in FAU, which is defined by Novo Nordisk as the
amount of enzyme that hydrolyses 5.26 g of starch per hour
at 30C. A linear correlation (r2 > 0.99) was found between
the increase in absorbance at 405 nm/min and the -amylase
concentration in the range of 0.022 to 0.884 FAU/mL. Repeated analysis showed a maximum variation of 0.03
FAU/mL (typically 0.02 FAU/mL).
The assay activity of amyloglucosidase (AMG) is based
on a manual method reported by Holm (1986). This method
is based on a spectrophotometeric analysis of p-nitrophenol
(yellow solution under alkaline conditions) formed by the
reaction of amyloglucosidase with the colorless pnitrophenyl--glycoside, pNPG (Merck, Darmstadt, Germany). Samples were diluted fivefold and 1 mL was used to
react with 2 mL of 0.1% pNPG at pH 4.3 to 4.4 (using 1 M
acetate buffer). Following incubation for 20 min at 30C,
the reaction was stopped using 3 mL 0.1 M borax solution.
The absorbance of the resulting yellow solution was measured immediately at 400 nm. The contribution of both
pNPG and borax was subtracted as blanks from this measurement. The activity of this enzyme is expressed in an
arbitrary unit, AGU. This unit is defined by Novo Nordisk
as the amount of enzyme that hydrolyzes 1 mol of maltose
per minute at 30C. The analysis was calibrated using an
AMG standard from Novo Nordisk. A linear correlation (r2
>0.99) was established between the absorbance at 400 nm
and the AMG concentration, in the range 0.005 to 1.696
AGU/mL. Analysis of repeated samples did not show variations in excess of 0.03 AGU/mL.
Image Analysis

Measurement of Hyphal Morphology

Samples were diluted 30-fold using 20% sucrose solution
fixed, and stained using lactophenol cotton blue (Fluka Chemie, Buchs, Switzerland). This dilution, resulting in a typical biomass concentration of 0.2 g/L on the slide, was also
necessary to minimize any artifacts caused by hyphal overlapping. Mycelial morphology was measured on a Quantimet 570 image analyzer (Leica, Cambridge, UK) connected to a Polyvar optical microscope (Reichert-Jung, Austria) using the method of Tucker et al. (1992) at a
magnification of 40. Additional measurements of hyphal
widths were made at a magnification of 250. A 30-L
aliquot from a previously Whirly-mixed vial was placed on
a microscope slide and covered with a cover slip. Pipette
tips with cut-off ends were used to dispense samples to
eliminate preferential mycelial size selection during dilution
and slide preparation. Analysis of fields on the slide was

partly automated, with manual intervention for editing and

rejection of empty fields, the latter to minimize analysis
time. As a compromise between analysis time (typically
between 2 and 3 h per sample) and accuracy of results, a
total of 200 elements were analyzed. These elements included clumps (dense aggregates of entangled hyphae) as
well as the freely dispersed form. Morphological parameters
of interest for the freely dispersed mycelia were mean total
hyphal length, mean projected area, and the number of tips
per hypha. Clump morphology was quantified in terms of
mean projected area (Tucker et al., 1992). The mean projected area of all elements was taken as a measure of the
total biomass (Packer and Thomas, 1990).
RESULTS
Continuous cultures at a dilution rate of 0.05 h1 (residence
time 20 h) were successfully conducted for up to 650 h
without contamination. A summary of all the experiments
and results is given in Table I. However, a detailed description of the experimental results is restricted to C1. The
steady-state values of morphological parameters, biomass
concentration, and protein activity measured before and after each change in agitation speed were compared using a
t-test of statistical significance. The null hypothesis that a
given parameter was independent of agitator speed was rejected when P < 0.05. The values used in the statistical test
were obtained after five to six residence times of a change
in operating conditions. The standard errors quoted for the
mean values of both morphological parameters as well as
for enzyme analysis at each steady state have been calculated from the measured errors according to Levitt (1973).
Figure 1 shows the profiles of agitator speed, biomass concentration, carbon dioxide production rate, dissolved oxygen, and pH in C1. Dissolved oxygen and pH (Fig. 1c) could
generally be maintained independently of the agitation
speed at 1% and 0.1 of the set-point, respectively. The
change in agitation speed did not influence either biomass
concentration or the carbon dioxide production rate (Fig.
1b). A steady-state biomass concentration of 2.0 0.1 g/L
was obtained, resulting in a yield on substrate of 0.61 g
biomass/g substrate. Morphological measurements (projected area) of the biomass revealed that approximately 33
2% of the mycelia existed in the freely dispersed form at
1000 rpm, with clumps accounting for the remainder (Fig.
2). The change in agitation speed to 550 rpm resulted in a
rapid increase in the freely dispersed fraction to 42%, followed by a gradual decrease to a steady state value of 15
2%. It is likely that the fragmentation of aggregates at the
higher speed increased the proportion of biomass in the
freely dispersed form. These results also clearly highlight
that representative morphology of the biomass for this strain
(at the given operating conditions) can only be obtained by
considering the clump morphology in addition to the freely
dispersed.
Although, in reality, such defined subpopulations may
not have physiological significance, they are nevertheless

AMANULLAH ET AL.: MORPHOLOGY AND PROTEIN PRODUCTION OF ASPERGILLUS ORYZAE

437

Table I.

Summary of chemostat experiments at different agitation speeds (values quoted are at steady state).
Morphological parameters
Freely dispersed
Agitation conditions

Enzyme activity

Agitation
speed (rpm)

P/V
(kW m3)

P/(kD tc)
(kW m3 s1)

Clumps + freely
dispersed: mean
projected area (m2)

Fermentation C1
1000a
1000b
550

12.6
12.6
2.2

950
950
90

6500 1400
6100 1100
16,500 3800

3300 800
3200 700
5100 1300

410 70
400 50
530 80

5.2 0.8
5.1 0.7
7.4 1.3

79 12
78 12
69 13

0.37 0.06
0.31 0.06
0.37 0.06

0.77 0.08
0.36 0.07
0.38 0.05

Fermentation C2 (repeat of C1)

1000
12.6
550c
2.2

950
90

7600 1800
22,100 6000

4100 1000
3100 600

340 60
410 70

5.7 1.0
6.6 0.9

62 13
62 14

0.44 0.06
0.46 0.06

0.80 0.09
0.85 0.08

Fermentation C3
700

230

13,900 3000

9000 2500

460 80

7.7 1.6

63 15

0.42 0.06

0.95 0.09

4.3

Mean
projected
area (m2)

Mean total
length (m)

Mean
number
of tips

Hyphal
growth unit
(m/tip)

-Amylase
(FAU/mL)

AMG
(AGU/mL)

P/V: specific power input; P/(kD3tc): energy dissipation/circulation function.

Before loss in gene copynumber.
b
After loss in gene copynumber.
c
Data shown are for measurements up to five residence times after speed change from 1000 to 500 rpm.
a

very useful for measuring relative changes in mycelial morphology. The mean projected area of the total biomass was
found to increase significantly from 6100 1100 m2
(mean standard error) at 1000 rpm to 16,500 3800 m2
at 550 rpm (Fig. 3). The establishment of the new steady
state after the step change in speed occurred within six
residence times (120 h). It is clear from Figure 3 that the
mean projected areas of the two samples preceding the
change in agitation conditions were in close agreement, and
that the area seemed to have reached another, higher value
by 500 h. The normalized size distributions of the mycelia
(clumps and freely dispersed) were also in good agreement
for each pair of samples (Fig. 4). Figure 4 also shows that
the size distribution was skewed to smaller sizes at the
higher speed. The maximum projected area of clumps did
not exceed 5 104 m2 at 1000 rpm, whereas a considerable number of clumps with projected areas between 5 to 9
104 m2 was measured at 550 rpm.

Figure 1. Profiles of (a) agitator speed, (b) biomass concentration, carbon dioxide production rate, and (c) pH and dissolved oxygen concentration as a function of fermentation time in a chemostat with a dilution rate
0.05 h1 (experiment C1).

438

Figure 2. The variation of percent of clumps and freely dispersed with

agitation speed (C1).

BIOTECHNOLOGY AND BIOENGINEERING, VOL. 62, NO. 4, FEBRUARY 20, 1999

Figure 3. The variation of mean projected area of the biomass with

agitation speed (C1).

The mean total hyphal length and mean projected area of

the freely dispersed mycelia as a function of the agitation
speed are shown in Figure 5a. A reduction in the agitation
speed from 1000 to 550 rpm caused an increase in the mean
total length from 400 50 m to 530 80 m. These
changes resulted in an increase in the mean projected area
from 3200 730 m2 to 5100 1300 m2. The measurements of the mean number of tips per hypha suggest that the
microorganisms become more branched at the lower speed.

Figure 5. Changes in freely dispersed morphology with agitation speed:

(a) mean total length and mean freely dispersed area; (b) mean number of
tips per mycelium; and (c) hyphal growth unit (C1).

Figure 4. (a) Normalized size distributions of mycelia at 1000 rpm (330and 348-h samples) and (b) 550 rpm (505- and 557-h samples) demonstrating morphological steady states (C1).

The steady-state values were constant at 5.1 0.7 and 7.4

1.3 tips per hypha at 1000 and 550 rpm, respectively (Fig.
5b). The hyphal growth unit (Caldwell and Trinci, 1973;
Trinci, 1974), HGU, determined as the quotient of the mean
total length and the mean number of tips, is shown in Figure
5c. Due to the relatively large standard errors of the mean
values, the hyphal growth unit at 1000 rpm was statistically
indistinguishable at 78 12 m/tip to the value of 69 13
m/tip at 550 rpm. At 700 rpm (fermentation C3, see Table
I), the measured steady-state value of HGU was 63 15
m/tip.
The activities of -amylase and AMG for C1 are shown
in Figures 6a and b, respectively. Both -amylase and AMG
activity remained fairly constant at 0.37 0.06 FAU/mL
and 0.77 0.08 AGU/mL, respectively, until 183 h. Thereafter, there was a reduction in the activity of both enzymes.
This decrease may possibly be due to a loss in the measured
AMG gene copy number and is discussed later, although it
is important to point out that this decrease did not occur
around the changes in agitation speed. Incidentally there
was no change in the morphological state of the culture
in the same period. Referring to the 70-h period (or

AMANULLAH ET AL.: MORPHOLOGY AND PROTEIN PRODUCTION OF ASPERGILLUS ORYZAE

439

3.5 residence times) before the speed change, it seems that

the enzyme activities had stabilized. Following the reduction in the agitation speed, -amylase activity increased
from 0.31 0.06 FAU/mL to 0.38 0.06 FAU/mL (the
mean steady-state values and their standard errors are
quoted). The AMG activity increased from 0.36 0.07 to
0.38 0.05 AGU/mL. These differences were not statistically significant, suggesting that the agitation speed did not
affect the enzyme production rates, whereas a similar decrease in agitation speed resulted in nearly 2.7 increase in
the mean projected area (see Fig. 3). Because it could be
argued that the 70 h before the speed change may strictly
not have been enough to attain steady state, the results obtained in experiment C1 were verified from a repeat experiment, C2, where the agitation speed was also reduced from
1000 to 550 rpm, although at 279 h, revealing a similar
trend in -amylase activity (Fig. 7a). Following the immediate decrease in activity after the speed change, the steady
state value at 550 rpm (0.44 0.06 FAU/mL) remained
statistically indistinguishable from the value at 1000 rpm
(0.46 0.06 FAU/mL). As with -amylase, AMG activity
also decreased immediately from 0.85 0.08 AGU/mL following the reduction in agitation speed (Fig. 7b). Thereafter, it increased to a steady-state value of 0.80 0.09 AGU/
mL at 550 rpm and remained at this value until 450 h (8.5
residence times). The AMG activity at 550 rpm was also
statistically indistinguishable from the value at 1000 rpm.
These results are shown in Table I.

Figure 7. The variation of: (a) -amylase; and (b) AMG activities with
agitation speed (experiment C2). Agitation speed was decreased from 1000
to 550 rpm at 279 h.

DISCUSSION
Morphology

Steady State

Figure 6. The variation of: (a) -amylase; and (b) AMG activities with
agitation speed (experiment C1). Agitation speed was decreased from 1000
to 550 rpm at 348 h (C1).

440

The successful operation of chemostat cultures of filamentous fungi requires that differential retention of mycelia by
size does not occur in the bioreactor. There is a possibility
of this phenomenon in chemostat cultures operated with an
overflow weir. In the present study, however, the combination of well-mixed conditions, the location of the exit pipe
next to the upper impeller, and the rapid removal of broth by
the pump when required should have prevented differential
washout. This was confirmed by morphological measurements of broths sampled simultaneously from the vessel
(near the bottom impeller) and the overflow; the results did
not differ at the 95% confidence level. Furthermore, it is
clear from the results shown in Figures 3 and 4 that both the
mean projected area as well as the normalized size distribution of mycelia were similar in samples taken at steady
state. Thus, it was possible to obtain morphological steady
states. The precise control of fermentation parameters including dissolved oxygen and pH has allowed investigation
of the influence of agitation speed alone on mycelial morphology and productivity. Problems of wall growth encountered in initial experiments were resolved in the present
study.

BIOTECHNOLOGY AND BIOENGINEERING, VOL. 62, NO. 4, FEBRUARY 20, 1999

Dependence of Mycelial Fragmentation (Clumps

and Freely Dispersed) on Energy Dissipation/
Circulation Function
Prolonged cultivation of filamentous fungi in continuous
culture requires the creation of new centers of growth and
this is achieved by mycelial fragmentation. Under equilibrium conditions, such as those found in a chemostat under
normal operation, hyphae appear to grow toward a maximum length (van Suijdam and Metz, 1981). According to
the hypothesis proposed by van Suijdam and Metz, hyphae
shorter than this maximum length will not fragment. However, it is difficult to reconcile the use of mean morphological data with the concept of a maximum size. It is unclear
how any individual hypha could ever grow longer than the
maximum size under this hypothesis, unless one invokes
some unquantified biological variability.
A more important objection to the hypothesis of van Suijdam and Metz (1981) is that it does not allow for clump
fragmentation. Following the suggestion of Justen et al.
(1998), it is proposed that clump fragmentation may lead to
equilibrium sizes in the clump as well as the freely dispersed data. The mean projected area of the biomass and
mean total hyphal length of the freely dispersed mycelia at
steady state for 550, 700, and 1000 rpm are shown in Table
I. These steady-state morphological parameters have been
correlated with both the energy dissipation/circulation function (EDC) as well as specific power input, as shown in
Figure 8a and b, respectively. The regression coefficients
and slopes of the plots shown in Table II suggest that the
EDC function is an excellent correlator of hyphal fragmentation even under growing conditions. This confirms the
data of Justen (1997) for fed-batch P. chrysogenum fermentations. Because the EDC function implies irregular passage
of the mycelia through the impeller region of high-energy
dissipation (two in this case), a distribution of sizes could
occur, especially if there is also clump formation due to
aggregation. The equilibrium mean total hyphal lengths
may be a consequence of equilibrium in clump formation
and breakage rather than any balance between growth and
fragmentation of the freely dispersed mycelia.
It should be noted that there is no fundamental understanding of how clumps might be broken by agitation, nor
why EDC works well in growing cultures. From Figure 8b
and Table II, it can be seen that the mean projected area and
the mean total hyphal length could be equally well be correlated with specific power input, because, at a given scale

Table II. Dependence of morphological parameters on energy dissipation/circulation function and specific power input in chemostat cultures.
Correlator of
mycelial damage

Mean
projected area

Mean
total length

P/(kD3 tc)

r2 0.95
Slope 0.44
r2 0.95
Slope 0.59

r2 0.99
Slope 0.12
r2 0.99
Slope 0.16

P/V

Figure 8. Correlation of the mean projected area of the biomass (clumps

+ freely dispersed) and mean total length (freely dispersed) with: (a)
P/(kD3tc) the energy dissipation /circulation function; and (b) P/V the specific power input.

using one impeller type, both correlators are simply functions of impeller speed [P/(kD3tc)] N4, (P/V) N3. Therefore, although the slopes are different, the regression coefficients are identical. However, previous work by Justen et
al. (1996) has established the generality of this function
rather than specific power input or impeller tip speed in
correlating mycelial damage. It is for this reason that EDC
is considered here to be superior to specific power input as
a correlator of hyphal breakage.
van Suijdam and Metz (1981) reviewed and conducted
studies on the effects of agitation on mycelial morphology
of P. chrysogenum. The steady-state mean total length of
the freely dispersed elements at different agitation intensities (using a single Rushton turbine) at a dilution rate of
0.055 h1 was used to assess mycelial damage. Because data
for the mean clump-projected area were not available from
their studies, their mean total hyphal length data have been
compared to those obtained in the present study (Fig. 9).

AMANULLAH ET AL.: MORPHOLOGY AND PROTEIN PRODUCTION OF ASPERGILLUS ORYZAE

441

Figure 9. The variation of the mean total length (freely dispersed) with
specific power input in the present study with Aspergillus oryzae (D
0.05 h1) and in the studies of van Suijdam and Metz (1981) with Penicillium chrysogenum (D 0.055 h1).

The large standard errors associated with van Suijdam and

Metzs data in Figure 9 were due to the relatively small
number of mycelia analyzed by these investigators. Although the freely dispersed data must also be treated with
caution, because it is probable they represent only a small
fraction of the biomass, the dependencies of the mean total
hyphal length on power input for A. oryzae and P. chrysogenum are given by the slopes 0.16 and 0.18, respectively.
This suggests that the mechanisms of agitation-related fragmentation might be similar for both strains. It must be
stressed, however, that it is vital to include clump morphological data to assess fragmentation properly.
Number of Tips (Freely Dispersed Form)
There are several reports in the literature (Dion et al., 1954;
Metz et al., 1981; van Suijdam and Metz, 1981) which
suggest that hyphae are shorter, thicker, and more highly
branched at high agitation speeds compared to low speeds,
although the mean number of tips per hypha was not reported. These studies only considered the freely dispersed
form, which may not have constituted most of the biomass.
Shorter hyphae could have resulted from fragmentation of
the clumped form as well as any direct breakage of the
freely dispersed form. It is also important to note that, in the
work of Dion et al. (1954), dissolved oxygen was not controlled independently of agitation speed. Indeed, the introduction of oxygen instead of air at low agitation speeds also
resulted in the highly branched form of the mycelia. The

442

studies of Metz et al. (1981) and van Suijdam (1981) were

conducted with P. chrysogenum. There are no published
reports on effects of agitation intensity on Aspergillus oryzae.
In the present study, the mean number of tips per hyphal
element (of the freely dispersed form) was found to be
higher at 550 compared to 1000 rpm (see results for fermentations C1 and C2 in Table I). This implies less highly
branched hyphae at higher speeds. At 700 rpm (fermentation C3), the value was similar to that at 550 rpm. It is
proposed that the balance between growth ( 0.05 h1)
and fragmentation was relatively similar at 550 rpm (P/V
2.2 kW m3) and 700 rpm (P/V 4.3 kW m3) giving a
similar mean number of tips per hypha, whereas, at 1000
rpm (P/V 12.6 kW m3) fragmentation dominated
growth, resulting in a lower mean value of tips. It is possible
that the combination of high agitation rates and very low
specific growth rates (as in many chemostat cultures and in
the production phase of fed-batch cultures) leads to fragmentation dominating growth, giving rise to fewer tips per
hypha.
Makagiansar (1991) reported such a phenomenon in 5-L
fed-batch cultures of P. chrysogenum, observing that the
mean number of tips per hypha was slightly reduced (in the
production phase) when the agitation speed was increased in
the range 700 to 1300 rpm. In general, the mean number of
tips is also likely to be influenced by the specific growth
rate of the culture and nutrient availability, and is probably
also strain-dependent. Measurements of the hyphal growth
unit (Fig. 5c) or the average length of hypha associated with
each tip at different agitation speeds do not appear to be
very reliable due in part to the large standard errors. In this
study, the values of HGU at each agitation speed tested
remained statistically indistinguishable from one another.
van Suijdam and Metz (1981) also did not find much variation in hyphal growth unit with agitation speed. The hyphal
growth unit as measured in their study was 57 40 and 49
16 m/tip at specific power inputs of 1.46 and 15.06 kW
m3, respectively. The hyphal growth unit volume, HGV
(Caldwell and Trinci, 1973; Trinci, 1974), might in fact be
a better parameter to consider, especially since variations in
mycelial diameter with agitation speed have been observed
(Dion et al., 1954). Morphological measurements at a much
higher magnification (250) are required for the accurate
discrimination of hyphal diameters. Using such a magnification in this study, hyphal widths of steady-state samples
of 1000 and 550 rpm were 2.27 0.04 and 2.38 0.05 m,
respectively. Using these values, it was shown that the
HGVs also remained statistically indistinguishable from one
another with changes in agitation speed. Except in its reflection of the fragmentation of clumps, measurements on
the freely dispersed form alone are probably of little value.
Morphological Transients
The mean projected area considering both the clumped and
freely dispersed mycelia (Fig. 3) increased from 5300 400

BIOTECHNOLOGY AND BIOENGINEERING, VOL. 62, NO. 4, FEBRUARY 20, 1999

m (at 1000 rpm) to 9400 900 m (at 550 rpm) in just 1

h following the reduction in agitation speed. The doubling
time of the culture can be calculated as approximately 14 h
at a specific growth rate of 0.05 h1. If it is assumed that no
fragmentation followed the change in agitation speed from
1000 to 500 rpm at 348 h (this assumption might be applied
here, given the significant lowering of agitation speed), and
that the density and width of the mycelia were constant, it
can be shown that the measured changes would require
approximately 10 h if the increase in the mean projected
area was due to growth alone. Instead, these changes were
observed over 1 h. Clearly, a physical rather than a biological mechanism must be responsible for the rapid morphological changes. It is suggested that an aggregation process
with a time constant in minutes causes the initial increase in
the mean projected area. In addition, the morphological
transients of the freely dispersed mycelia at 1, 5, 15, 30, 60,
and 540 min following the step change in agitation speed
from 1000 to 500 rpm at 348 h are shown in Figure 10a and
b. A rapid increase in the mean total length from 380 20
to 460 20 m was measured just 1 min after the speed
change (Fig. 10a). Similar values were obtained at up to 1 h.
The mean total length increased to a maximum of 510 30
m by 9 h, before decreasing to 480 30 m at 376.5 h.
The mean number of tips (Fig. 10b) showed very similar
behavior. The changes measured in the first hour were too
rapid for a biological response of the culture. This is demonstrated by the following calculation. Assuming that the
hyphal diameter and density were constant and no fragmentation took place, the mean tip extension rate (Nielsen,
1992; Nielsen and Krabben, 1995) can be calculated from:

Figure 10. Morphological transients of the freely dispersed mycelia after

a step change in speed from 1000 to 550 rpm: (a) mean total length; and (b)
mean number of tips per mycelium. Data points shown are for 1, 5, 15, 30,
60, and 540 min after the change in speed.

Lt
rtip =
Nt

(1)

where rtip is the tip extension rate, the specific growth

rate, Lt, the mean total length, and Nt the mean number of
tips per hypha. The above equation is strictly valid for freely
dispersed mycelia that do not undergo fragmentation. The
application of Eq. (1) also assumes that the measurements of
the freely dispersed class are not contaminated by fragments
breaking off clumps. The mean tip extension rate calculated
using Eq. (1) immediately preceding the step change in
impeller speed was 3.7 m/tip per hour (using a measured
mean value of tips per hypha of 5.2). This implies that it
would require approximately 3.8 h to account for the measured increase in mean total length. Mycelia continue to
grow in size due to the combined effects of both initial
aggregation and proceeding growth and approach a new
steady-state size that is determined by the balance of growth
and fragmentation at the lower speed. The transient response of the culture to a step increase in speed from 550 to
1000 rpm was found to be much faster (data from experiment C2, not shown) than a step decrease.
PROTEIN PRODUCTION
The decrease in -amylase and AMG activities between 183
and 300 h at 1000 rpm in C1 (Fig. 6) was of concern,
especially because it occurred after a relatively short duration of chemostat operation. The decrease in AMG is
thought to be related to a loss in the measured gene copy
numbers expressing this heterologous protein. This was
confirmed by a Southern blot analysis on samples removed
from the chemostat and flash frozen in liquid nitrogen. The
analysis revealed that, relative to the -amylase gene, the
AMG copy number decreased from 4 in the shake flask
inoculum to between 1 and 2 at 376 h of chemostat operation. The three additional samples analyzed after 376 h and
the end of the cultivation showed that the AMG gene copy
number remained at this value. Unfortunately, samples prior
to 376 h were not taken from the chemostat for similar
analysis. The AMG copy number analysis is in broad agreement with the results of the AMG activity, where a decrease
from what seemed to be a steady state value of 0.77 0.08
AGU/mL measured up to 183 h appeared to have stabilized
at 0.36 0.07 AGU/mL between 300 and 348 h (Table I).
The increase in AMG activity following the step change in
agitation speed from 1000 to 550 rpm was found to be
insignificant at 0.38 0.05 AGU/mL.
The consequences of a similar speed change in C2 (at 279
h) was also found to result in an insignificant decrease from
0.85 0.08 to 0.80 0.09 AGU/mL (steady-state values are
quoted). The data shown in Table I for C2 refer to measurements up to five residence times after speed change
from 1000 to 550 rpm. No loss of AMG gene copy number
is thought to have occurred over the duration of these measurements. Although not measured, and also not relevant to
the results in this study, it is postulated that a similar loss in

AMANULLAH ET AL.: MORPHOLOGY AND PROTEIN PRODUCTION OF ASPERGILLUS ORYZAE

443

AMG gene copy number occurred in C2 (probably to a

different extent compared to C1) after approximately 450 h,
where a substantial decrease in AMG activity occurred. The
AMG activity at 700 rpm (C3) was 0.95 0.09 AGU/mL
(see Table I).
The results of the Southern test in C1 could not be used
to determine if there was a loss in genes expressing the
homologous -amylase, because this itself was used as a
reference for AMG analysis. The reasons for the small decrease in -amylase activity from 0.37 0.06 to 0.31 0.06
FAU/mL between 183 and 300 h remain unclear, but may
be linked to the decrease in AMG activity. Because maltodextrin is broken down to smaller sugars such as glucose
and maltose by AMG, a decrease in AMG activity may have
reduced the availability of the smaller sugars to cells, resulting in a decrease in the -amylase activity. The small
increase in activity from 0.31 0.06 to 0.38 0.05 FAU/mL
following the reduction in speed is difficult to explain, especially because the results of C2 showed that -amylase
activity was not influenced by agitation speeds of 1000 or
550 rpm, remaining at approximately 0.45 0.06 FAU/mL.
A value of 0.42 0.06 FAU/mL was measured at 700 rpm.
However, it is most likely that such small variations in
enzyme activity with agitation speed may also be also inherent (as experimental error) in continuous cultivations of
fungi. Carlsen (1994) reported variations in -amylase activity of 0.10 FAU/mL at steady state using a dilution rate
of 0.10 h1. Carlsen et al. (1994) reported that the amylase activity secreted by the wild-type strain of A. oryzae was 0.55 FAU/mL at a dilution rate of 0.05 h1. It is
concluded that, within experimental error, the expression of
AMG and -amylase remains independent of agitator speed
in the range 550 to 1000 rpm.
The small differences in steady-state values at the different agitation speeds were not significant, suggesting that
agitation speed did not affect the enzyme production rates,
whereas a similar decrease in agitation speed from 1000 to
550 rpm resulted in nearly a 2.7-fold increase in the mean
projected area. Thus, very significant changes in mycelial
morphology (at a constant biomass concentration) as a consequence of changes in agitation speed do not affect enzyme
production. These results can also be analyzed in the context
of Wesselss model for protein secretion (Wessels, 1990,
1993, Wosten et al., 1991). According to Wessels, protein
secretion in filamentous fungi occurs by a bulk flow mechanism through the plasticized cell membrane of mycelial tips.
This has more recently been confirmed by Nielsen (1997)
using MITC-conjugated antibodies raised against amylase produced by an A. oryzae strain cultivated on solid
medium. The plasticized cell membrane area in turn is dependent upon the tip extension rate. Thus, protein production according to Wesselss hypothesis is determined by the
product of the tip extension rate and the total number of
actively growing tips.
Because the value of the hyphal growth unit (Lt/Nt) was
found to be essentially constant at each of the agitation
speeds used in this study, it follows from Eq. (1) that, at a

444

constant dilution rate (and hence specific growth rate), the

tip extension rate remains constant. This in turn implies that,
for the experimentally determined protein production to remain independent of agitation speed, the total number of
actively growing tips also remains approximately constant.
The results of C1 between 183 and 300 h indicate that,
although it was possible to obtain morphological steady
states, the same could not be said for protein expression.
This has important consequences for prolonged continuous
culture cultivations. Due to the uncertainties with respect to
heterologous DNA loss, chemostat cultures operated for
minimal amounts of time are recommended to obtain single
steady states. Several studies have demonstrated the problems of genetic mutations in chemostat cultures of filamentous fungi (Christensen et al., 1995; Righelato, 1976; Trinci
et al., 1990). Loss of heterologous DNA has been reported
for chymosin production by A. niger due to the accumulation of a low-producing mutant (Dunn-Coleman et al.,
1993). Withers et al. (1995) showed that, although the original strain of P. chrysogenum had been completely replaced
by spontaneous mutants in a chemostat operated up to 1600
h, no loss of heterologous gene expression could be detected.
CONCLUSIONS
The effects of agitation intensity on mycelial morphology
and recombinant protein production have been investigated
independently of dilution rate and dissolved oxygen effects.
To date, this is one of the most carefully controlled experiments reported in the literature in relation to the effects of
agitation intensities in fungal chemostats. Quantitative measurements using image analysis revealed that representative
morphology of this culture can only be obtained by considering both the clumps and freely dispersed class of mycelia.
It was possible to obtain morphological steady states at
different agitation intensities. Mycelial fragmentation was
characterized by the mean projected area of mycelia, which
could be correlated to the energy dissipation/circulation
function. Variations in agitation intensity resulted in significant changes in mycelial morphology. Morphological transients suggested that mycelial aggregation may be responsible for the rapid and large changes measured following a
reduction in speed.
The expression of both homologous and heterologous
proteins remained independent of agitation speed in the
range 550 to 1000 rpm (2.2 to 12.6 kW m3). Thus, the most
important finding of this study is that significant variations
in mycelial morphology as a consequence of changes in
agitation speed do not affect protein production (at a constant specific growth rate). Presuming secretion to be dependent on the number of actively growing hyphal tips, this
result also suggests that protein secretion is not a bottleneck
in this strain. Studies linking protein secretion to actively
growing mycelial tips should perhaps be conducted using
strains with very high gene copy numbers, where protein
secretion may become the limiting step. The practical im-

BIOTECHNOLOGY AND BIOENGINEERING, VOL. 62, NO. 4, FEBRUARY 20, 1999

plication of the results of this study is that the agitation

intensity in such fungal fermentations can be manipulated to
meet process requirements in terms of dissolved oxygen
levels and bulk mixing, and possibly to control broth rheology by changing morphology without compromising recombinant protein production. However, it is possible that
changes in agitation intensity can influence the specific
growth rate of the culture in fed-batch fermentations and,
therefore, growth-related protein production. This will be
the subject of further studies.
NOMENCLATURE
C
CPR
D
Fl
HGU
HGV
k
Lt
Nt
P
P/(kD3tc)
Po
rtip
T
tc
V
Vi
W

vessel off-bottom clearance (m)

carbon dioxide production rate (mmol/L per hour)
impeller diameter (m)
impeller flow number
hyphal growth unit (m/tip)
hyphal growth unit volume (m3/tip)
impeller geometrical parameter(W/4D)
mean total length of freely dispersed mycelia (m)
mean number of tips per hypha
power input (kW m3)
energy dissipation/circulation function (kW m3s1)
power number
tip extension rate (m/tip per hour)
vessel diameter (m)
mean circulation time (s)
volume of liquid in vessel (m3)
impeller swept volume (m3)
impeller blade height (m)
specific growth rate (h1)

References
Ayazi Shamlou P, Makagiansar HY, Ison AP, Lilly MD. 1994. Turbulent
breakage of filamentous micro-organisms in submerged culture in mechanically stirred bioreactors. Chem Eng Sci 49:26212631.
Braun S, Vecht-Lifshitz SE. 1991. Mycelial morphology and metabolite
production. Trends Biotechnol 9:6368.
Caldwell IY, Trinci APJ. 1973. The growth unit of the mould Geotrichum
candidum. Arch Microbiol 88:110.
Carlsen M. 1994. Ph.D. thesis, Department of Biotechnology, Technical
University of Denmark, Lyngby, Denmark.
Carlsen M, Spohr A, Mrkeborg R, Nielsen J, Villadsen J. 1994. Growth
and protein formation of recombinant Aspergillus: utility of morphological characterisation by image analysis, In: Galindo E and Ramirex
OT, editors. Advances in bioprocess engineering. Dordrecht: Kluwer.
p 19202.
Christensen LH, Henriksen CM, Nielsen J, Villadsen J, Egelmitani M.
1995. Continuous cultivation of P. chrysogenum. Growth on glucose
and penicillin production. J Biotechnol 42:95107.
Dion WM, Carilli A, Sermonti G, Chain EB. 1954. The effect of mechanical agitation on the morphology of Penicillium chrysogenum Thom in
stirred fermenters. Rend First Super Sanita 17:187205.
Dunn-Coleman NS, Bodie E, Carter GL, Armstrong GL. 1993. Stability of
recombinant strains under fermentation conditions, pp. 152174. In:
Kinghorn JR and Turner G, editors. Applied molecular genetics of
filamentous fungi. Glasgow: Blackie.
Harvey LM, McNeil B. 1994. Liquid fermentation systems and product
recovery of Aspergillus, In: J. E. Smith, editor. Aspergillus. New
York: Plenum. p 141176.
Holm KA. 1986. Automatic spectrometric determination of amylogluco-

sidase activity using p-nitrophenyl--D-glucopyranoside and a flow

injection analyser. Analyst 111:927929.
Jeenes DJ, Mackenzie DA, Roberts IN, Archer DB. 1991. Heterologous
protein production by filamentous fungi. Biotechnol Gen Eng Rev
9:327367.
Justen P, Paul GC, Nienow AW, Thomas CR. 1996. Dependence of mycelial morphology on impeller type and agitation intensity. Biotechnol
Bioeng 52:672684.
Justen P. 1997. Ph.D. thesis. University of Birmingham, Birmingham, UK.
Justen P, Paul GC, Nienow AW, Thomas CR. 1998. A mathematical model
for agitation induced fragmentation of Penicillium chrysogenum. Bioproc Eng 18:716.
Levitt BP. 1973. The accuracy of measurements and treatment of results,
In: Levitt BP, reviewer. Findlays practical physical chemistry. 9th
edition. London: Longman. p 125.
Mackenzie DA, Jeenes DJ, Belshaw NJ, Archer DB. 1993. Regulation of
secreted protein production by filamentous fungi: recent development
and perspectives. J Gen Microbiol 139:22952307.
Makagiansar HY. 1991. Ph.D. thesis. University College, London, UK.
Makagiansar HY, Ayazi Shamlou P, Thomas CR, Lilly MD. 1993. The
influence of mechanical forces on the morphology and penicillin production of Penicillium chrysogenum. Bioproc Eng 9:8390.
Markl, H, Bronnenmeier R, Wittek B. 1991. The resistance of microorganisms to hydrodynamic stress. Int Chem Eng 31:185.
Merchuk JC. 1991. Shear effects on suspended cells. Adv Biochem Eng
Biotechnol 44:6695.
Metz B, de Bruijn EW, van Suidjam JC. 1981. Method for quantitative
representation of the morphology of molds. Biotechnol Bioeng 23:
149163.
Nielsen J. 1992. Modelling the growth of filamentous fungi. Adv Biochem
Eng 46:187223.
Nielsen J, Johansen CL, Jacobsen M, Krabben P, Villadsen J. 1995. Pellet
formation and fragmentation in submerged cultures of Penicillium
chrysogenum and its relation to penicillin production. Biotechnol Prog
11:9398.
Nielsen J, Krabben P. 1995. Hyphal growth and fragmentation of Penicillium chrysogenum in submerged cultures. Biotechnol Bioeng 46:
588598.
Nielsen J. 1997. Physiological engineering. Proceedings of the 8th European Congress on Biotechnology, Budapest, Hungary, August 1997.
Nienow AW, Miles D. 1969. A dynomometer for the accurate measurement of mixing torque. J Sci Inst 2:994995.
Packer HL, Thomas CR. 1990. Morphological measurements on filamentous micro-organisms by fully automatic image analysis. Biotechnol
Bioeng 35:870881.
Paul GC, Kent CA, Thomas CR. 1994. Hyphal vacuolation and fragmentation in Penicillium chrysogenum. Biotechnol Bioeng 44:655660.
Paul GC, Thomas CR. 1998. Characterisation of mycelial morphology
using image analysis. Adv Biochem Eng 60:159.
Reuss M. 1988. Influence of mechanical stress on the growth of Rhizopus
nigricans in stirred bioreactors. Chem Eng Technol 11:178187.
Righelato RC. 1976. Selection of strains of Penicillium chrysogenum with
reduced penicillin yields in continuous cultures. J Appl Chem Biotechnol 26:153159.
Smith JJ, Lilly MD, Fox RL. 1990. Morphology and penicillin production
of Penicillium chrysogenum. Biotechnol Bioeng 35:10111023.
Trinci APJ. 1974. A study of kinetics of hyphal extension and branch
initiation of fungal mycelia. J Gen Microbiol 81:225236.
Trinci APJ, Robson GD, Wiebe MG, Cunliffe B, Naylor TW. 1990.
Growth and morphology of Fusarium graminearum and other fungi in
batch and continuous culture, In: Poole RK, Bazin MJ, and Keevil
CW, editors. Microbial growth dynamics. IRL Press: Oxford. p 1738.
Tucker KG, Kelly T, Delgrazia P, Thomas CR. 1992. Fully automatic
measurement of mycelial morphology by image analysis. Biotechnol
Prog 8:353359.
Ujcova E, Fencl Z, Musilkova M, Seichert L. 1980. Dependence of release
of nucleotides from fungi on fermentor turbine speed. Biotechnol Bioeng 22:237241.

AMANULLAH ET AL.: MORPHOLOGY AND PROTEIN PRODUCTION OF ASPERGILLUS ORYZAE

445

van Suidjam JC, Metz B. 1981. Influence of engineering variables upon the
morphology of filamentous molds. Biotechnol Bioeng 23:111148.
van Brunt J. 1986. Fungi: the perfect hosts? Bio/Technology 4:10571062.
Vardar F, Lilly MD. 1982. Effect of cycling of dissolved oxygen on product formation in penicillin fermentations. Eur J Appl Microbiol Biotechnol 14:203211.
Wessels JGH. 1990. Role of cell wall architecture in fungal tip growth
generation, In: Heath IB, editor. Tip growth in plant and fungal walls.
San Diego, CA: Academic Press. p 129.

446

Wessels JGH. 1993. Wall growth, protein excretion morphogenesis in

fungi. New Phytol 123:397413.
Withers JM, Wiebe MG, Robson GD, Osborne D, Turner G, Trinci APJ.
1995. Stability of recombinant protein production by Penicillium
chrysogenum in prolonged chemostat culture. FEMS Microbiol Lett
133:245251.
Wosten HAB, Moukha SM, Sietsma SM, Wessels JGH. 1991. Localisation
of growth and secretion of proteins in Aspergillus niger. J Gen Microbiol 137:20172023.

BIOTECHNOLOGY AND BIOENGINEERING, VOL. 62, NO. 4, FEBRUARY 20, 1999