Dielectric monitoring of growth and sporulation of Bacillus thuringiensis
M.H. Sarrafzadeh1,2, L. Belloy1, G. Esteban3, J.M. Navarro1& C. Ghommidh1,*
1UMR Inge´nierie de la Re´action Biologique – Bioproductions, Universite´ Montpellier II, 34095, Montpellier,
2Engineering Department, Persian Gulf University, Boushehr, Iran
3FOGALE Nanotech, Nıˆmes, France
*Author for correspondence (Fax: +33-467144292; E-mail: firstname.lastname@example.org)
Received 17 December 2004; Revisions requested 10 January 2005; Revisions received 11 February 2005; Accepted 14 February 2005
Key words: biomass measurement, capacitance, permittivity, spore-forming bacteria, segregated model
On-line permittivity and optical density measurements have been used to monitor biomass concentration
and sporulation status during growth of a spore-forming bacterium, Bacillus thuringiensis, in fed-batch
culture. The correlation between permittivity, optical density and other observations showed three distinct
phases of growth: growth itself, transition and sporulation. The permittivity variations during the transition
and sporulation phases could be related to the sporulation development: the evolution pattern of the ratio
of optical density to permittivity was representative of the culture state, and during the sporulation phase, a
permittivity index could be build to measure the extend of spore liberation.
Cell suspension dielectric permittivity is an indi-
cation of viable biomass concentration (Yardley
et al. 2000). Briefly, the plasma membranes of
living cells can be polarized under the influence
of an electrical field in the radio-frequency range
(between 0.1 and 10 MHz). The resulting capaci-
tance is correlated to the cell volume fraction i.e.
to the biomass concentration. The variation of
the dielectric permittivity of cell suspensions in
dependent not only on the cell concentration but
also on medium and cell cytoplasm conductivity,
on cell size and on membrane capacitance
dependent on the physiological state of the cells.
In this work, an attempt has been made to use
the permittivity as an indicator not only of the
cell concentration, but also of the development of
sporulation, using Bacillus thuringiensis as a
generally similar in terms of metabolic and physi-
ological responses to environmental conditions.
When starved, these bacteria can form spores in a
process that involves major changes of the cyto-
plasm membrane and construction of a protective
shell around the cell’s DNA, with a switch-off of
normal cell metabolism (Stragier & Losick 1996).
At the cell level, sporulation has been described
as a seven-stage process (Figure 1).
Many important bioproducts (solvents, antibi-
otics, enzymes, insecticides, etc.) are produced by
spore-formers. Typically, the maximal synthesis
rate of these products occurs at the onset and dur-
ing the sporulation process (Liu et al. 1994).
Therefore, the distinction of the various physiolog-
ical states is a critical step both for better fermen-
tation control and for further understanding of the
properties of these bacteria, where capacitance-
sensing technology could play a decisive role.
Materials and methods
Media: The composition of preculture, culture
and feeding media are shown in Table 1. The pH
of the media was adjusted to 6.8 with 3 M
NaOH. The media were heat-sterilised (121 ?C,
Biotechnology Letters (2005) 27: 511–517
? Springer 2005
15 min.). For fed-batch cultures, the concen-
trated glucose and yeast extract/hydrolisated
casein were sterilized separately (110 ?C, 15 min.)
then mixed aseptically.
Organism: Bacillus thuringiensis H14 (Ecautec,
Tahiti) was grown in a flask containing 500 ml
medium with shaking at 30 ?C for 9 h and used
to inoculate the fermenter.
Fermentation procedure: The fermenter (Biol-
afitte) had a maximum volume of 15 l with 9 l
culture medium. Three fermentations were initi-
ated in batch mode at 30 ?C. Dissolved O2(DO)
above 20% of saturation during the growth peri-
od. During the sporulation phase (U3) a different
oxygenation level was maintained. The pH was
controlled at 6.8 during the growth phase. Fed-
batch cultures were started from the 4th hour to
the 24th hour of fermentation at a constant feed
rate of 150 ml h)1. Halfway through the feeding
period, a concentrated salt solution (Table 1, to-
tal volume of 500 ml) was added aseptically to
the culture to avoid nutrient limitation.
Off-line biomass analysis
Biomass dry weight (DW) concentration was
determined by filtering a known amount of med-
ium and differential weighing of the filter before
and after drying at 104 ?C. Cell counts (vegeta-
tive and sporulated cells) were done using a
phase contrast microscope and a Thoma’s haem-
ocytometer. Total cells (CT) represented the sum
of the cells in any physiological state. Unlysed
sporulated cells were counted as spores. Total
sporulation (ST) was determined as the ratio of
lysed and unlysed sporulated cells, which can be
distinguished due to their refractility, over total
cell count. The percentage of sporulated lysed
cell vs. total cell count has been also calculated
as mature spore count (SM).
On-line biomass analysis
Dielectric permittivity: A Biomass System (BS,
Fogale Nanotech, Nıˆmes, France) fitted with a
standard 25 mm diameter probe was used. The
BS is a three-frequency capacitance analyser that
computes the biomass concentration from the
difference between measurements made at two
frequencies. The higher frequency (10 MHz) is
used as a reference, while the middle frequency is
Fig. 1. Different physiological states in a spore-forming bacte-
rium. X1: the biomass with ability of devision; X2: biomass in
sporulation process; X3: free spore; X4: sporangial debris. t0:
normal growth; t1: genetic system division; t2: asymmetric
septation; t3: engulfment; t4: cortex synthesis; t5: coat synthe-
sis; t6: lysis of mother cell; t7: spore liberation.
Table 1. Media composition.
Components (g l)1) PrecultureCulture Feeding Salt solution
MgSO4. 7 H2O
chosen at the fc characteristic frequency, in the
middle range of the b-dispersion to minimise the
influence of cell size variations. The third lowest
frequency is used to compensate the electrode po-
larisation in conductive media. The fc frequency
(2 MHz) was calculated according to the b-dis-
persion theory, assuming an average bacterial cell
size of 5 lm. The results were expressed as ‘‘abso-
lute dielectric permittivity increment’’ (pF/cm).
(Wedgewood BT65) was used. The response of
this sensor is not linear, and was corrected using
the empirical equation:
ODL¼ ODmþ 0:0288 e3:244ODm? 1
obtained from serial dilutions of a culture, in
which ODLwas the linearized optical density and
ODm, the measured optical density. Although the
sensor was operated in ‘‘bubble noise suppres-
sion’’ mode, the readings were further numeri-
cally filtered, by recording the maximal value
Fig. 2. Time course of a fed-batch culture of Bacillus thuringiensis: Variation of dry weight (•) and on-line measurements of optical
density ()), permittivity ()) and conductivity (—). The three culture phases (U1-3) have been specified on the figure.
Fig. 3. Time course of a fed-batch culture of Bacillus thuringiensis: Total (CT: n) and vegetative (CV: n) cell counts using Thoma
haemocytometer. Sporulation development as percentage of refractile lysed and unlysed sporulated cells presented as total sporula-
tion (ST: •), and percentage of mature spores (SM: •) versus total cells.
observed over a 1 min period, to reduce the noise
generated by air bubbles.
The results of a typical fed-batch culture of
Bacillus thuringiensis are presented in Figures 2
and 3. According to the on-line permittivity mea-
surements in Figure 2, three phases could be
distinguished over the fermentation time course:
a rapid growth phase until 14 h (U1), a phase of
slow growth between 14 and 24 h (U2) and
finally a decline phase which began with the
arrest of substrate feeding (U3). In contrast, the
on-line optical density, and biomass dry weight
measurements revealed only two phases: growth
The onset of U2was marked by a small but
noticeable decline of permittivity, from 11 to
10 pF cm)1. This event was not observed on on-
line OD and off-line measurements. A mineral
salt solution (growth activator) was added just
after. The permittivity continued to increase to
12.7 pF cm)1at 24 h of fermentation, when the
feeding pump was stopped thereby triggering the
start of U3. Then, the measurements of OD, DW,
permittivity and total cell counts declined stea-
dily until the end of the culture (t = 48 h.). As
shown on Figures 2 and 3, this decline was more
pronounced for permittivity measurements and
dry weight than for the other techniques.
Until the 14th hour of fermentation (Fig-
ure 3), corresponding to the end of U1, only veg-
etative cells (CV) were present in the culture.
During U2, CVstayed almost constant while the
total cell counts (CT) continued to rise and spor-
ulated cells began to appear. Mature spores (SM)
could be observed after the 18th hour, although
their liberation took place mainly during U3. To-
tal sporulated cell concentration (ST) increased
steadily and reached 84% at the end of the cul-
ture. Microscopy photos (Figure 4) present the
sporulation development and the variations of
cell morphology during the different fermentation
This pattern of evolution was reproduced dur-
ing two other experiments, differing only by the
oxygen level maintained during the sporulation
During the culture, the conductivity rose from
15 to 37 mS cm)1(Figure 2). During the growth
phase, this variation was due both to the feeding
and to the addition of NaOH to maintain a con-
stant pH. The sudden increase at about the 14th
hour was the result of the addition of the nutri-
ent salt solution, while the sudden drop at 24 h
was caused by an increase of the gas hold-up, re-
lated to a 20% withdrawal of culture medium, at
constant mixing rate.
The various biomass quantification techniques
used along this study gave a different perspective
of cell growth and state. A simple segregated
growth model with three populations in different
physiological state (Figure 1) was proposed to
make the result analysis easier:
1. Vegetative cells with ability to divide;
2. Sporulating cells from stages t1to t6of spor-
ulation (classification according to Losick
et al. 1986);
3. Mature spores (t7).
Fig. 4. Refractility evolution during fermentation of Bacillus
thuringiensis using phase-contrast microscopy. (a) Vegetative
cells at the beginning of the fermentation. (b) At about 14 h
of fermentation, small and slightly refractile phase-grey body
located at one end pole of the cells appears. (c) 23 h of fer-
mentation: the refractile unlysed sporulated cells and their
linked crystals. (d) 48 h of fermentation: fully mature spores,
freefrom the mothercell.
Scale bar(for allmicro-
The biomass concentration corresponding to
each population was defined as X1, X2, and X3
respectively. The total biomass (XT) in each
phase of fermentation was the sum of these three
populations, including also cell or sporangial
The following equations describe the possible
transformation routes of the different popula-
X1! X1þ X1 ð2Þ
ðvegetative cell growth and divisionÞ
X2! X3þ X4 ð4Þ
Regarding the measurement basis of each tech-
nique, the following rules could be asserted:
1. Optical density detected biomass in all states.
2. Capacitance measurement could only sense
X1 and X2 (because these two populations
had functional, insulating cytoplasmic mem-
branes), although the signal amplitude re-
different, since the cell structures differed
3. Dry weight measured X1, X2and X3and part
of X4, depending on cell debris size.
4. Microscopy cell counts could identify X1, X2
and X3, but could not reliably distinguish be-
tween X1and a part of X2before stage t4of
sporulation (because spores became refractile
only after this stage).
During U1, constant ratios were maintained be-
tween all the measurements. Since during this
phase, often described as ‘‘exponential’’, the cell
population was composed exclusively of X1cells
(Figure 3 and 4a), it was not surprising to find
the same general linear trend for all methods.
During U2, where the X2population began to
appear, the proportionality between the different
measured variables was lost. During this phase,
X1and X2were the main populations, and the
concentration of vegetative cells (CV) remained
nearly stable (Figure 3). Even with an overall
biomass increase, the permittivity signal could
not increase at the same relative rate than other
measurements since the ratio of X1and X2cells
A remarkable event during U2was the small
decline in permittivity observed at the 14th hour
(Figure 2). This event was also observed during
other cultures, and was probably related to the
onset of a nutrient limitation and to the begin-
ning of sporulation. None of the other tech-
niques had detected this event. The probable
physical basis of the permittivity decrease could
be the changes occurring in the cytoplasmic
membrane during pre-spore formation. As re-
ported by Losick et al. (1986), sporulation is
marked by several time-programmed stages of
morphological development taking place over a
period of at least 7 h. Therefore the perception
of sporulation was dependent on the stage that
might be detected by the applied technique. For
instance, sporulation estimation by microscopic
counts was possible only at a stage where pre-
spores became visible as refractile bodies using
phase-contrast microscopy. On another hand, the
variation of permittivity could happen at a stage
where the state of cell membrane (among other
physical parameters) changed. In the third stage
(t3) of sporulation, the cytoplasmic membrane
tears partially to separate the pre-spore compart-
ment (Piggot et al. 1994). Since this transforma-
tion is irreversible, it could be considered as the
real onset moment of sporulation. The detection
of this step was therefore a challenge in which
permittivity measurement could be a highly
(1994), the cells entering sporulation, become re-
fractile four hours later. Accordingly, the ob-
served 14thh decline of permittivity could be
related to the first emergence of refractile cells
(Figure 3) observed at the 18thhour. Subsequent
cultures have shown that this event was repro-
ducible, although with a variable amplitude.
During U3, the relative differences between
the various biomass determination techniques
were amplified by the appearance of free spores,
which refractility, size and structure were strong-
ly affecting measurements.
The ratios of OD, DW and CTto permittivity
were different during the various phases, and
could be used as tools to monitor the physiologi-
cal state of the culture. Since the permittivity
measurement was the most sensitive to physio-
logical state changes, it was used as the reference
(Figure 5) and of the ratio of OD to permittivity
(Figure 6), as functions of permittivity, could be
compared. The pattern similarity allowed an easy
identification of the growth phases: as long as
the ratio remained under or around 1, the culture
was in the U1rapid growth phase; between 1 and
2, while the permittivity was still increasing, the
culture was in the transition phase U2; finally,
when the permittivity decreased while the ratio
was increasing, the culture was in the U3sporula-
The three experiments reported on Figures 5
and 6 differed only by the dissolved oxygen
concentration maintained during the U3 phase.
Sporulation was more effective in anaerobic
conditions, and no significant variation of the
evolution pattern could be observed in 50% and
100% dissolved oxygen cultures.
During U3, a good agreement was found be-
tween the permittivity decline and the sporula-
tion development. Using equation 4 to describe
the dominant biological reaction, the permittivity
(e) variation could be directly related to cell lysis:
Fig. 5. Time-independent plot of sporulation as a function of permittivity during cultures of Bacillus thuringiensis under different
concentrations of dissolved oxygen.
Fig. 6. On-line detection of physiological phases of Bacillus
thuringiensis cultures, ran in different concentrations of dis-
solved oxygen, using ratio of OD to permittivity as a function
of permittivity. U1-3refer to the phases as in Figure 1.
Fig. 7. Relationship between permittivity as (1- et/emax) and
mature spore fraction during the sporulation phase (U3).
de Download full-text
dt¼ kdðX1þ X2Þ
the vegetativecells were absent
from equation 4:
! ðet? e24hÞ ¼ ?kðX3t? X324hÞ
If : e24h¼ emax;X324h¼ 0;
emax! ð1 ?
emaxÞ ¼ k0X3t
In Figure 7, the mature spore fraction during
U3has been plotted as a function of the permit-
tivity variation (1-et/emax). A linear correlation
was obtained, demonstrating the possibility of an
on-line detection of spore liberation rate using
this permittivity index.
This work is another demonstration of the useful-
ness of on-line dielectric permittivity measure-
ments in fermentation. Not only can the biomass
concentration be readily monitored, but physio-
logical transitions can also be detected. To our
knowledge, this is the first report of a real-time
detection of a transition phase during a culture of
sporulating bacteria. The benefits of the combina-
tion of on-line optical density and permittivity
measurements should not be limited to the moni-
toring and control of B. thuringiensis growth and
should be of more general applicability. Instead
of using a dual frequency measurement, a scan
over the whole frequency range of the b-disper-
sion could provide additional informations, such
as an indirect estimation of the cell size, and
avoid the selection of strain specific parameters.
This will be the purpose of future work.
The authors wish to thank S. Benoit for his skil-
led technical assistance and Pr. R. Sotudeh for
his valuable comments. This work was partially
supported by the Ecautec S.A. (Papeete, Tahiti)
and the Languedoc-Roussillon Regional Council.
Liu W, Bajpai R, Bihari V (1994) High-density cultivation of
sporeformers. Ann. N.Y. Acad. Sci. 721: 310–325.
Losick R, Youngman P, Piggot PJ (1986) Genetics of
endospore formation in Bacillus subtilis. Ann. Rev. Genet.
Piggot PJ, Bylund JE, Higgins ML (1994) Morphogenesis and
gene expression during sporulation. In: Piggot PJ, Moran CP
& Youngman P, eds. Regulation of Bacterial Differentiation.
Washington, DC: American Society for Microbiology, pp.
Schwan HP (1957) Electrical properties of tissue and cell
suspensions. Adv. Biol. Med. Phys. 5: 147–209.
Setlow P (1994) Mechanisms which contribute to the long term
survival of spores of Bacillus species. J. Appl. Bacteriol. 76:
Stragier P, Losick R (1996) Molecular genetics of sporulation in
Bacillus subtilis. Ann. Rev. Genet. 30: 297–341.
Yardley JE, Kell DB, Barrett J, Davey CL (2000) On-line, real-
time measurements of cellular biomass using dielectric
spectroscopy. Biotechnol. Gen. Eng. Rev. 17: 3–35.