Biotechnol. Prog.

2004, 20, 1757−1765 1757

Effect of Process Variables on Supercritical Fluid Disruption of

Ralstonia eutropha Cells for Poly(R-hydroxybutyrate) Recovery
Kianoosh Khosravi-Darani,† Ebrahim Vasheghani-Farahani,*,†
Seyed Abbas Shojaosadati,† and Yadollah Yamini‡
Department of Chemical Engineering, Tarbiat Modarres University, P.O. Box 14115-143, and
Department of Chemistry, Tarbiat Modarres University, P.O. Box 14115-175, Tehran, Iran

This research focuses on the disruption of the Gram-negative bacterium Ralstonia

eutropha cells by supercritical CO2 for poly(R-hydroxybutyrate) (PHB) recovery. The
variables affecting cell disruption such as drying strategy, type of modifier, and
cultivation time, as well as operating pressure, temperature, and repeated release of
supercritical CO2 pressure, have been studied. Effect of this disruption technique on
PHB molecular mass was also investigated. PHB recovery was examined using a
combination of this method and chemical pretreatments. For salt pretreatment, the
cells were exposed to 140 mM NaCl and heat (60 °C, 1 h). The cells were also exposed
to 0.2-0.8% (w/w) NaOH to examine the effect of alkaline pretreatment. Bacterial
cells treated in growth phase exhibited less resistance to disruption than nutrient-
limited cells in the stationary phase. It was also found that the wet cells could be
utilized to recover PHB, but purity of the product was lower than that obtained from
freeze-dried cells. Pretreatment with a minimum of 0.4% (w/w) NaOH was necessary
to enable complete disruption with two times pressure release. Salt pretreatment was
less effective; however, disruption was improved by the application of alkaline shock.
The proposed method is economic and comparable with other recovery methods in
terms of the percentage of PHB recovery and energy consumption, while it is
environmentally more benign.

1. Introduction
Products of recombinant DNA technology, the over-
production of proteins as inclusion bodies and the poten-
tial use of the intracellular storage products such as
poly(R-hydroxybutyrate) (PHB) have led to the increased
interest in efficient and cost-effective cell disruption to
enable the recovery of intracellular microbial products
in intact form (1). PHB, which is a polymer belonging to
the poly(hydroxyalkanoate) (PHA) group, accumulates as
distinct inclusions in the cell and comprises up to 80%
of cell dry weight for strains of Ralstonia eutropha under
conditions of nitrogen or phosphate limitation and excess
of carbon source (2). PHB has many applications in
medicine, agriculture, pharmacy, and packaging due to
its biodegradability and biocompatibility (3). Currently
the main problem that limits the widespread use of PHB
is its relatively high cost compared to petrochemical-
based polymers. It has been estimated that over half of
the production cost of PHAs is associated with the
recovery and purification processes (4). Several methods
have been developed for PHB recovery. Efficiency of the
oldest method, which involves extraction by chloroform
(5) with acetone pretreatment was 70%. However the
extracted polymer solution containing more than 5% (w/
v) PHB is very viscous, so the removal of cell debris is
difficult. Utilization of large amounts of toxic solvents
* To whom correspondence should be addressed. Phone: +98-
21-8005040. Fax: +98-21-8006544. E-mail: evf@modares.ac.ir.
† Department of Chemical Engineering.
‡ Department of Chemistry.
10.1021/bp0498037 CCC: $27.50 © 2004 American Chemical Society and American Institute of Chemical Engineers
Published on Web 10/27/2004
by this method not only increases the total cost but also
has adverse environmental consequences (6, 7). PHB
recovery by hypochlorite digestion (8) leads to low mo-
lecular mass PHB as a result of degradation by chemicals
(9). The use of sodium hypochlorite together with chloro-
form significantly reduces degradation but also requires
a large quantity of solvent (10, 11). Enzymatic digestion
method was also developed for the PHB recovery with
90% efficiency (7), but the use of expensive chemicals and
complex processes makes it economically unattractive
(12). Mechanical cell disruption methods appear to be
favored because of economic advantages for the recovery
of intracellular proteins. However PHB recovery by
mechanical methods has received less attention (13).
Tamer and Moo-Young (14) compared PHB recovery by
chemical digestion (sequential treatment with sodium
dodecyl sulfate (SDS) and sodium hypochlorite) and
mechanical (continuous-flow bead mill and high-pressure
homogenizer) methods. They concluded that chemically
induced disruption with SDS alone was not satisfactory,
but sequential treatment with SDS (1 h) followed by a
24 h sodium hypochlorite treatment resulted in 95%
removal of the cellular protein (a measure of disruption).
Homogenization after a suitable pretreatment is a proven
recovery method, although homogenizers are susceptible
to blockage (14) and only relatively dilute biomass
slurries can be satisfactorily processed. Harrison has
shown that chemical (addition of an anionic detergent
and monovalent cation) or enzyme pretreatment can
weaken the cell wall sufficiently to decrease energy
consumption and allows for operation at lower pressures
1758 Biotechnol. Prog., 2004, Vol. 20, No. 6

or fewer passes with the homogenizer (13). In view of

performance, relatively low power consumption, and
robustness, bead mill disruption was recommended for
PHB recovery (15). In this method the agitation speed
will affect the amount of energy consumption, but when
rotation was increased to a critical level, only a marginal
increase was found, with the excess energy being con-
verted to heat (16).
Recently, supercritical fluids (SCFs) have been of
interest in the biotechnological processes. They provide
solutions to drastic problems related to nonthermal cell
inactivation (17-19), enzyme inactivation (20), perme-
abilization (21, 22), and extraction of fermentation prod-
ucts (23-25). Application of SCF is simple, inexpensive,
and more importantly, noninjurious to the structure and
function of enzymes, e.g., alcohol dehydrogenase, inver-
tase, glucose-6-phosphate dehydrogenase, fumarase (26), Figure 1. Schematic diagram of experimental apparatus for
and protein activities (27-29). Supercritical carbon SC disruption of R. eutropha: (A) CO2 gas tank, (B) supercritical
dioxide (SC-CO2) is the most commonly used fluid and fluid pump, (C) disruption vessel, (D) oven, (E) 2-way valve.
its low critical temperature (31.1 °C) and pressure (73
bar) make it an ideal medium for processing volatile cals were the same as reported previously (38). Batch
products (30). Nontoxicity, nonflammability, the selectiv- culture was performed in 1000-mL flasks at 30 °C, the
ity of the process, and the ease of recovery are the most pH of the broth was adjusted to 7.0, and it contained 1.5%
important features (31). (w/v) fructose (as a sole carbon source) for different
To date, application of SCF in cell disruption (SC cultivation time intervals (20, 30, 40 or 50 h). Five
disruption) over a wide range of operating conditions has milliliters of sample of PHB containing biomass (18.5 g
been confined to yeasts (26, 32-35) and a few bacteria L-1) was taken at different time intervals, centrifuged
(27, 36, 37). The process involves a sudden release of the at 3000 × g for 20 min, and washed twice with deionized
applied SC-CO2 pressure that results in its penetration water. For larger quantities a centrifuge (G2-M1 Beck-
into the cells. After expansion of gas within the cells, man, USA) running at 16 250 × g for 15 min was
flash discharge of pressure forces the cell wall and causes employed. The biomass containing 82% (w/w) water was
cell disruption. The technique is relatively simple and either used directly for PHB recovery or dried in an oven
can easily be scaled up (27). Furthermore, the cells are at 60 °C or alternatively freeze-dried. The dried biomass
exposed to minimal shear forces and there is no heat containing 73% (w/w) PHB was kept at 4 °C until needed.
generation, which would otherwise affect the yield of Benzoic acid and PHB (Aldrich) were used as internal
temperature-sensitive and labile material recovery ad- and external standards for PHB purity analysis, respec-
versely. tively. A protein standard prepared from bovine serum
albumin was used for calibration. Absorbances were
A technique to exploit SC-CO2 disruption of R. eutro-
measured at 595 nm by spectrophotometer (Cary 50,
pha cells has been recently developed in our laboratory
Varian, Australia) after treatment with Coomassie bril-
(38). The effects of different variables such as exposure
liant blue G-250 (Sigma). Pure carbon dioxide (99.99%)
time, pressure, temperature, and volume of methanol (as
obtained from a local distributor (Sabalan, Tehran) was
a modifier) on SC disruption of suspension of bacterium
used in all experiments. Morphological changes in the
were investigated.
disrupted cells were assessed by direct observation with
In this study, we have extended our previous work to a scanning electron microscope (SEM) model Philips XL
obtain maximum recovery with minimum energy con- 30. For this purpose the cells were placed on a glass, dried
sumption. As a first step the effects of drying prior to in ambient temperature, and coated with a gold layer.
disruption and type of modifier on the PHB recovery were
2.2. Cell Disruption by SC-CO2. A Suprex MPS/225
studied. The exposure time was fixed at 40 min to
system (Pittsburgh, PA), illustrated in Figure 1, was used
minimize power consumption, and the effects of pressure,
for SC disruption of cells. It consists of two major
temperature, repeated release of SC-CO2 pressure, and
components: a syringe pump and a high-pressure vessel.
volume of modifier were optimized under the new condi-
CO2 was supplied from the gas cylinder (A), to the pump
tions. A two-level factorial design was employed to take
(B), which compressed CO2 to a desired pressure prior
into account the interaction of variables in generating
to injection into the pressure vessel (C). This vessel was
the process response (39). Then a three-level Taguchi
located in the oven (D). The cells were mixed well with
design was used to obtain more informative data about
glass beads and packed into the pressure vessel. This
the main effects (40). The strength of the cell wall is
process prevents channelling, increases the contact sur-
expected to have a significant influence on cell disruption
face between the cells and the SC-CO2, and consequently
by mechanical means. Hence a series of pretreatment
reduces the equilibrium time. Then SC-CO2 was rapidly
experiments were arranged to alter cell wall strength
released by a two-way valve (E). Sintered stainless steel
prior to disruption and to minimize repetition of pressure
filters were used to prevent any carryover of the cells.
release. The effect of cultivation time on the PHB
The volume of vessel C was 4 mL; however, other vessels
recovery was also studied. The proposed method is
with volumes of 1, 8, and 10 mL were also used to
compared with a mechanical method for cell disruption
compare energy consumption. The vessels and all other
from efficiency and economy points of view.
parts exposed to high pressures were made of stainless
steel. A pressure gauge was installed at the CO2 inlet of
2. Materials and Methods the vessel to calibrate the pressure.
2.1. Microorganisms and Culture Medium. The To start cell disruption, 2 g of the dried cells (or equal
microorganism (R. eutropha), culture medium, and chemi- amount in the case of wet biomass) was placed in the
Biotechnol. Prog., 2004, Vol. 20, No. 6
pressure vessel. Then 1 mL of phosphate buffer (0.1 M,
pH 7.0) and in some experiments 0.5-1% v/v modifier
(cosolvent; toluene, methanol, or acetone) was also added.
The vessel was then capped and placed in the thermostat
oven. The cells were exposed to SC-CO2 for 40 min. After
pressure release the vessel was washed with 4 mL of
phosphate buffer (0.1 M, pH 7.0) and centrifuged. Su-
pernatant and pellet were used for protein (41) and PHB
measurements (42), respectively.
2.3. Protein Assay. The extent of cell disruption was
determined by assessing total soluble protein release
from the bacteria by Bradford assay (41). The maximum
amount of releasable protein (10%) was measured when
further repeated release of SC-CO2 pressure yielded no
additional protein. The efficiency of protein recovery was
obtained by
% recovered protein (w/w) )
released protein
(
maximum releasable protein in biomass
× 100 ) mented with sodium chloride at a final concentration of
2.4. Determination of PHB Molecular Mass, Pu-
rity, and Recovery. To determine percentage of recov-
ered PHB from the cells, 2 mL of chloroform was added
to a pellet, transferred to a capped tube, and shaken for
15 min. The resulting mixture was centrifuged at 3000
× g for 20 min to obtain a bottom phase, which consisted
of extracted PHB in chloroform solution. Then the pellet
was removed to determine PHB purity and recovery. For
this purpose, an appropriate amount of benzoic acid (final
concentration of 1000 ppm) was added to 2 mL of the
pellet suspension in chloroform and subjected to metha-
nolysis in the presence of 3% (v/v) sulfuric acid (42). The
separation, identification, and determination of the re-
sulting methyl ester of 3-hydroxybutyrate was performed
using a Philips Scientific Model 4410 gas chromatograph
(GC) equipped with a flame ionization detector (FID) with
a split/splitless injector and fitted by a 25 m × 0.25 mm
i.d. fused silica capillary column coated with an im-
mobilized film of 14% cyanopropylphenyl dimethyl silox-
ane. Helium was used as the carrier gas at a flow rate of
1 mL min-1. Oven temperature was programmed to 90
°C for 3 min and then increased to a final temperature
of 150 °C at a rate of 8 °C min-1. The amount of resulting
methyl ester detected by gas chromatography reflects the
purity of recovered PHB. The percentage of the compound
was calculated by the area normalization method, with-
out considering response factors. The efficiency of PHB
recovery was obtained by the following equation:
% recovered PHB (w/w) )
released PHB
(
maximum releasable PHB in biomass
×100 ) method on the efficiency of PHB recovery, purity, and
The maximum amount of releasable protein and PHB
from dried biomass amounted to 10% and 78% (w/w),
respectively. These were further confirmed using chemi-
cal disruption methods such as chlorinated solvents (5)
and hypochlorite digestion methods (43).
The molecular mass of PHB was measured by a
Shimadzu 6A gel permeation chromatography (44) using
chloroform as an eluent. Samples of 0.1% (w/v) recovered
PHB in chloroform were prepared for this purpose. A
calibration curve was made using standard polystyrene
solution.
2.5. Chemical Pretreatments of Biomass. A series
of pretreatment experiments was arranged in combina-
tion with SC disruption. For this purpose the cells were
removed by centrifugation at 3000 × g for 20 min and
1759
Table 1. Effect of Cell Drying Method on PHB Purity,
Molecular Mass, and Efficiency of PHB Recoverya
PHB PHB % release (w/w)
biomass purity (%) hn
M PHB protein
wet biomass 77.6 715 000 68.3 75.5
dried at 60 °C 80.7 630 000 59.2 66.0
freeze-dried 83.9 760 000 64.6 70.9
a all data were obtained at 200 bar, 40 °C and 2.5% (w/w)
methanol as modifier for disruption of wet cells with an exposure
time of 100 min.
made into a slurry in phosphate buffer (0.1 M, pH 7.0)
as required. For alkaline pretreatment, sodium hydroxide
(3 M) was added to the cell suspension to obtain a
predetermined ratio of alkali to biomass (0.2-0.8% w/w).
The resulting suspension was processed immediately in
the SC apparatus. The entire process took about 45 min.
For salt pretreatment, the cell suspension was supple-
140 mM and incubated in a water bath (60 °C, 1 h). The
suspension was then cooled to 4 °C and centrifuged. A
combined pretreatment with salt, alkaline, and heat was
also tested. The cell suspension was supplemented with
sodium chloride (140 mM), the resulting suspension was
held in a water bath (60 °C, 1 h) and cooled to 4 °C, and
then saturated sodium hydroxide was added to the
biomass (0.4% w/w). After 1 min, the suspension was
neutralized to pH 7 with hydrochloric acid and centri-
fuged as explained above.
2.6. Design of Experiments and Analysis of Vari-
ance. As for any recovery process, the R. eutropha
disruption is affected by numerous variables including
operating pressure and temperature, volume of modifier,
and repeated release of SC-CO2 pressure. A two-level
factorial design of experiments is specially suitable and
sufficient to account for the parameters interactions;
hence it was used to study the interaction of variables.
Subsequently a three-level Taguchi design was used to
obtain more informative data about the main effects (40).
3. Results and Discussion
3.1. Effect of Process Variables on PHB Recovery.
3.1.1. Cell Drying Method. After fermentation, the
PHB-containing biomass is usually freeze-dried. Because
of the high energy requirement of the drying operation,
which must be considered for industrial production of
PHB, it is more desirable to use biomass directly for PHB
recovery. It is known that the drying method may affect
the product quality and recovery (45). Consequently, it
is necessary to investigate the effect of drying method
on PHB recovery and quality. The effect of cell drying
molecular mass (M h n) of the recovered product is given in
Table 1. These experiments were performed under the
optimized conditions obtained in our previous research
(38) that comprise a temperature of 40 °C, a pressure of
200 bar, a methanol content of 2.5% (v/v), and an
exposure time of 100 min. Results indicate that both PHB
purity and M h n were the highest in freeze-dried biomass,
which is in agreement with the results obtained by Chen
and co-workers (45), who investigated the effects of
cultivation time and biomass drying strategies on PHB
recovery using a surfactant-chelate aqueous system. A
higher purity was achieved for biomass dried at 60 °C
as compared to wet cells, but M h ndeclined to 630,000. The
reason for Mh n decrease when wet cells were employed is
unclear at this stage. A possible explanation for the
difference in product purity and recovery of samples dried
1760 Biotechnol. Prog., 2004, Vol. 20, No. 6

Figure 2. Release of total soluble protein from wet (b), 60 °C- Figure 3. Effect of modifier on supercritical cell disruption at
dried (0), and freeze-dried (2) biomass as a function of repeated temperature 40 °C, pressure 200 bar, and modifier 2.5% (v/v).
release of SC-CO2 pressure at temperature 40 °C; pressure 200
bar, and methanol 2.5% (v/v).

at 60 °C and freeze-dried samples is that PHB degrada-

tion may occur at 60 °C and also non-PHB cell materials
bind to PHB tightly, which results in reduced cell
permeability, cell disruption, and decrease of M h n(45).
Table 1 also implies that a better PHB recovery could be
achieved using the wet cells. A possible explanation for
this observation is that the dried R. eutropha are smaller
and more globular than wet cells, so they are hard to
disrupt by mechanical forces (46). Another reason is that
solubility of CO2 in the intracellular water content in the
wet cells is higher than that of dried biomass. This
observation is similar to those reported by Spilimbergo
and Bertucco (19) regarding the effect of water content
of cells on the SC-CO2 inactivation. This group has stated
that the amount of gas sorbed by cells increased with
increasing water content until free water appeared
around the cells (critical water content). It has been also
reported that wet cells of baker’s yeast with more than
60% moisture content were entirely destroyed when they
were saturated with CO2 gas at 40 bar and 40 °C (35).
Efficiency of cell disruption for 60 °C and freeze-dried
cells was low even after repeated release of pressure (as
shown in Figure 2). It can be concluded that cell disrup-
tion was more efficient using wet biomass, but purity of
recovered product decreased slightly. As a result, in the Figure 4. Scanning electron micrograph (×10,000) of intact
R. eutropha containing PHB (a) and SC-disrupted R. eutropha
subsequent experiments, in order to save the energy showing cell wall of disrupted bacterium (b).
required for cell drying, wet biomasses were used for PHB
recovery.
lustrates that after treatment, the appearance of
3.1.2. Type of Modifier. Acetone and methanol in-
crease CO2 polarity and solubility in the intracellular “wrinkles” was recognized in the cell surface (panel b),
aqueous compartments, and also toluene increases per- whereas untreated cells appeared as “rods” with a
meability of the cell wall. The results of the effects of particularly smooth surface (panel a). For most of our
acetone, methanol, and toluene as modifiers on cell disruption processes, the release of PHB was approxi-
disruption of wet cells are presented in Figure 3. The mately proportional to protein release (data not shown).
results were obtained under the optimum condition These results indicate that even if SC disruption causes
determined in our previous study (38) as explained above. a hole in the cell wall, it can be suitable and applicable
Figure 3 illustrates that the highest protein release in PHB recovery. Witte et al. (48) have reported that even
(77.2%) was achieved by addition of toluene as a modifier if tunnel structures with diameters of 40-200 nm are
as a result of the increased permeability of the cell formed in the cell envelope, the osmotic pressure differ-
membrane by solubilization of the lipidic components. ence between the cytoplasm and the outside medium
This may lead to increased CO2 absorption and conse- causes the driving force to squeeze the PHB granules
quently cell disruption. The effect of toluene as a modifier through the tunnel structures. Furthermore, Resch and
was predominant in these experiments, and it was colleagues showed that PHB granules do not possess rigid
selected for subsequent experiments. structures and are rather flexible, being able to be
3.2. Scanning Electron Microscopy (SEM). We squeezed out of the cell through the cell wall pores
speculated that SC disruption causes small pores in the produced by an enzyme in recombinant Escherichia coli
bacterial cell envelope with diameters less than that of (47). In some cases complete disruption of bacterium was
the PHB granules as was reported by Resch and co- observed (panel c). A possible explanation is that granules
workers (47) for cell lysis by enzyme digestion. This will with very large diameters rupture the bacterial envelope
raise the question of how such large particles can pass from the tunnel site while passing through the holes. On
through small pores. To address this question PHB- the basis of these facts and our own observation (Figure
containing cells were investigated by SEM as explained 4) in some experiments, only amounts of the released
above. Figure 4 shows micrographs of wet cells before protein were measured as criteria for cell disruption with
and after flash discharge of SC-CO2. This Figure il- the aim of PHB recovery.
Biotechnol. Prog., 2004, Vol. 20, No. 6 1761

Table 2. Analysis of Variance in a 23 Full Factorial Design (3 Variables in 2 Levels)

temp pressure % toluene % released % released
response (T) (°C) (P) (bar) (M) (v/v) protein (w/w) protein (w/w) F-valuea
1b 30 250 0 69.4 72.4
T 40 250 0 48.9 51.6 76.41
P 30 350 0 46.2 44.0 142.21
TP 40 350 0 47.5 50.4 60.94
M 30 250 1 79.1 76.8 220.09
TM 40 250 1 65.0 69.1 0.002
PM 30 350 1 65.0 67.5 5.24
TPM 40 350 1 60.0 61.4 25.64
a SSX ) {(r2n-1)2}/{r2n}; TSS ) {∑ X2i - ( ∑ Xi)2}/16 ) 1955.02; RSS ) TSS - ∑SSX ) 32.61. Error degree of freedom ) 9; error
variance ) 3.6; F1,9 ) 22.86 (R ) 0.001). b All variables are in their low levels.

Table 3. L-9 Taguchi Array for SC Disruption of R.

Figure 4 reveals that SC disruption has not caused eutropha
fragmentation in the cell wall of R. eutropha bacterium,
which facilitates subsequent product purification. Cell pressure temp no. of pressure toluene released
walls of SC-disrupted bacteria were less fragmented than trail (bar) (°C) releases (%) protein % (w/w)
those disrupted by conventional cell disruption tech- 1 200 25 1 0.5 71.8
niques, as has been reported also by Castor and co- 2 200 30 2 1 81.6
workers (32), while the yield of protein and PHB recovery 3 200 35 3 1.5 76.2
4 250 25 2 1.5 75.4
was the same for all disruption technique. 5 250 30 3 0.5 76.0
3.3. Analysis of Results in Full Factorial Design. 6 250 35 1 1 67.4
Table 2 shows possible combinations of three variables 7 300 25 3 1 73.3
at two levels and the corresponding protein release of 8 300 30 1 1.5 68.1
cells as a response to change in levels of variables, which 9 300 35 2 0.5 69.8
was proportional to PHB recovery. The data obtained for
Table 4. Analysis of Variance for PHB Recovery (%) in
the total soluble protein concentration were generally Taguchi Design
reproducible within (5%, given as a standard deviation
of replicated measurements. The analysis of the experi- variable sum of square variance percent
mental results is expressed in terms of F value in the pressure (bar) 45.2 22.6 33.1
last column of Table 2, which includes F values of the temperature (°C) 18.6 9.3 13.6
main effects of variables (temperature, T; pressure, P and no of pressure releases 71.3 35.6 52.2
modifier, M) and interaction effects between variables % toluene (v/v) 1.5 0.8 1.1
(TP, TM, PM, and TPM). This Table shows that F values
of all variables (rows 2, 3, 4, 5, and 8) are bigger than results in terms of the % released protein is also included
the critical value of F1,9 ) 22.86. Although the effect of in this table. The analysis of the results is shown in Table
all studied variables on the PHB recovery is highly 4. The results are presented as the sum of square,
significant (with 99.9% confidence), temperature-modi- variance, and percent of each variable in total response.
fier interaction is not significant, whereas the triple Apparently these values depend on the selected range
interaction in the last row is significant (with 95% for each variable. According to the analysis of the results,
confidence). In fact PHB recovery decreased with increas- the most effective variable on SC disruption of R.
ing pressure from 250 to 350 bar at constant temperature eutropha cells was repeated release of SC-CO2 pressure.
and also decreased with increasing temperature from 30 The analysis of experimental results is also expressed
to 40 °C at constant pressure. The modifier used in this in terms of main effects of each variable in Figure 5,
study, i.e., toluene, facilitated cell disruption at the which is defined as the average of the results for a
highest concentration examined. The effect of these variable at a specified level. Total soluble protein release
variables will be discussed later in more detail. Table 2 from the biomass was used to calculate the main effect
also implies that temperature-pressure interaction (F of each variable, as discussed below.
value of row 4) was significant. This is similar to results 3.3.1. Effect of Pressure. The bacterial cell disruption
reported by Spilimbergo and Bertucco (19) for the de- efficiency, as measured by total soluble protein release,
pendency of E. coli inactivation on SC-CO2. This inter- decreased with increasing pressure from 200 to 300 bar
action shows that at the static condition, diffusion of SC- (Figure 5a). A possible explanation is that pressure
CO2 into the cell wall increases with simultaneous increase causes increased viscosity and a decreased
increase of temperature and pressure, which in turn diffusion coefficient of the SC fluid and consequently
results in increased cell disruption. In fact the influence decreased mass transfer into the cells (50). The next
of pressure on the cell disruption is more pronounced at explanation is faster disruption of the cytoplasmic mem-
higher temperatures. The F value of the triple interaction brane than cell wall at high pressures (37). It is stated
between all variables (22.64) was also bigger than the that the cytoplasmic membrane can be disrupted more
critical value, but both of these interaction effects have rapidly than the cell wall, and hence at high pressure
F values (22.64 and 60.94) smaller than those of the main intracellular materials diffuse out, the cell wall shrinks,
effects (76.41, 142.21, and 220.09). Therefore interaction and consequently the efficiency of cell disruption de-
of these variables can be neglected in obtaining optimum creases (37).
condition (49). Other interaction effects (rows 6 and 7) 3.3.2. Effect of Temperature. Figure 5b shows that
have not caused any significant difference in the re- cell disruption increased with increasing temperature
sponse. from 25 to 30 °C and then decreased with further increase
Table 3 shows an L-9 Taguchi array that was used for of temperature to 35 °C. In fact by increasing tempera-
SC disruption of R. eutropha. Each experiment was ture at constant pressure, gas density decreases (36),
repeated two times, and the average of corresponding which leads to a decreasing net amount of gas absorbed
1762
Figure 5. Main effects of variables on the PHB recovery: (a) pressure, (b) temperature, (c) toluene % (v/v), and (d) number of
pressure release in full factorial design.
into the cells and consequently decreases the force caused
by flash discharge of pressure. It has been also reported
that temperature influences not only the gas absorption
but also physiological properties of the cell membranes
and walls (40). In addition, 30 °C is closer to the
supercritical temperature of the CO2 (TC ) 31.1 °C), and
flash discharge of pressure at this point will produce the
gas state with a maximum volume change, which in turn
causes more disruption of the cell wall. On the basis of
these results, 30 °C was selected as the optimum tem-
perature for cell disruption, which is close to optimal
growth conditions of bacterium. This allowed the disrup-
tion condition to be achieved at a temperature nonde-
structive to the cells or their contents.
3.3.3. Effect of Toluene Concentration. As shown
in Figure 5c cell disruption increased with increasing
toluene up to 1% (v/v), but further increase of it did not
affect protein release significantly. Initial increase of cell
disruption can be related to the increase of the cell
membrane permeability by removing its lipidic compo-
nents by toluene. Increased cell wall permeability leads
to increase mass transfer of SC-CO2 into the cells and
facilitates cell disruption by SC fluid.
3.3.4. Effect of Repeated Release of SC-CO2 Pres-
sure. Figure 5d reveals that cell disruption increased
with repeated release of SC-CO2 pressure. Indeed, SC
disruption procedure released nearly all of the total
soluble proteins within three times pressure release.
3.4. Cultivation Time. Because PHB content of
biomass and strength of the cell wall are determined by
the cultivation time, which in turn influences the recov-
ery and overall production expenditure, it is important
to study its effect on PHB recovery. The effect of cultiva-
tion time on the PHB content of freeze-dried biomass and
the efficiency of recovery and the purity of product are
given in Table 5. These results indicate that PHB content
increased with the increase of cultivation time and
consequently product purity was improved, i.e., the
higher the content of PHB, the higher the purity of
recovered product. However, at the late stage of fermen-
tation, PHB content did not increase significantly (with
95% confidence) but the efficiency of cell disruption
decreased. It is well-known that cell size and shape
influence the ease of cell disruption by mechanical
methods (46). Harrison reported that the size and shape
of R. eutropha vary with cell growth, which in turn affects
cell disruption by high-pressure homogenization (13). On
the basis of the results presented in Table 5 at the late
stage of cultivation time, globular cells containing more
PHB are hard to disrupt by SC disruption method. This
is similar to those obtained by other mechanical methods
(46) but opposite to results reported for the surfactant-
chelate method (50) because cell size and shape does not
influence the ease of cell disruption by chemical methods.
It has been also reported that young cells are more
susceptible to CO2 treatment than mature ones because
Biotechnol. Prog., 2004, Vol. 20, No. 6
Table 5. Effect of Cultivation Time on PHB Content,
Purity, and Efficiency of Recoverya
% release (w/w)
time (h) PHB content % PHB purity (%) PHB protein
20 41.2 83.4 73.7 81.1
30 73.7 93.6 69.3 79.0
40 77.1 95.0 64.4 72.9
50 77.5 95.2 61.9 70.4
a All data were obtained at 200 bar, 30 °C, and 1% (w/w) toluene
with an exposure time of 40 min.
Figure 6. Effect of vessel volume on energy consumption (2)
and efficiency of cell disruption (0).
bacteria synthesize new proteins, which protect cells
against a variety of adverse conditions such as high
temperature, oxidative stress, high salt concentration,
and high pressure (51, 52). It was also reported that for
maximum efficiency in yeast disruption requiring log
phase cells (37). Consequently, biomass produced after
30 h cultivation was used throughout our research to save
energy and time for PHB recovery.
3.5. Economical Aspect. The proposed technique was
compared with the bead mill recovery method, which in
view of performance and relatively low power consump-
tion has been recommended for PHB recovery (15). For
this purpose energy consumption in a single release of
SC-CO2 pressure in vessels with different volumes of 1,
4, 8, and 10 mL was investigated while the ratio of dried
cell weight per volume of the vessel was the same in all.
Figure 6 shows that energy consumption in SC-CO2
increases significantly (with 95% confidence) with de-
crease of vessel volume. In practice a suitable basis of
assessing operational efficiency is energy input per unit
mass of the disrupted cells (15). The energy input ξ to
the cell slurry was calculated from the voltage v and the
measured current I drawn by the pump:
(vItfilling + vItcompressing + vItoven)NF
ξ)
ηfm
where N is the flash discharge of pressure, Fand η (g
mL-1) are the concentrations of dried cell weight in the
fermenter and supercritical vessel, respectively, t (s) and
m (kg) are exposure time and dried cell weight of biomass
Biotechnol. Prog., 2004, Vol. 20, No. 6
Figure 7. Comparison of (a) alkaline pretreatment at concen-
tration of 0.2% (O), 0.4% (2), and 0.8% (0) (w/w) of NaOH with
untreated biomass (9); (b) salt pretreatment (O), salt and
alkaline (0.4% NaOH) pretreatment (b), and untreated biomass
(9) as a function of repeated release of SC-CO2 pressure.
in SC vessel, and f is the efficiency of disruption. The
pump operated at 220 V and drew 1.7 A current when
“filling” with CO2 and “compressing” the gas to its critical
pressure. In general, it took approximately 7 min for our
system to fill and attain 200 bar. The oven drew 2.5 A
current under the processing condition. As indicated in
Figure 6 the energy inputs for 40 min exposure time in
a 1-, 4-, 8-, and 10-mL vessel volume without any
repeated release of SC-CO2 pressure were 37.789, 9.815,
5.038, and 4.085 MJ kg-1. Disruption efficiency was
almost unaffected even in high vessel volume so long as
high vessel volume was used as in the present work.
Thus, if a bigger SC vessel can be used to disrupt 50 g of
biomass then the energy demand (even with 0.7 efficiency
of disruption) works out to only 0.431 MJ kg-1. Moo-
Young et al. have reported energy input of PHB recovery
from Alcaligenesis latus by bead mill disruption as 25.6
MJ kg-1 (15). They have suggested that for a very highly
concentrated slurry (500 kg m-3) energy consumption will
decrease to 3.38 MJ kg-1. In addition to energy consump-
tion, efficiency of disruption in all SC vessels in the
proposed technique is comparable with those reported by
the bead mill method.
As eq 1 illustrates, repeated release of SC-CO2 pressure
(N) increases the energy consumption. Thus pretreat-
ment of biomass is important to enhance the efficiency
of cell disruption without repeated release of SC-CO2
pressure.
3.6. Effect of Chemical Pretreatments. As shown
in Figure 7a for alkaline pretreatment alone, efficiency
of cell disruption improved by increasing the proportion
of sodium hydroxide relative to biomass. At a hydroxide
concentration of 0.4% and 0.8% (w/w), most of the protein
content released within two times pressure release; hence
0.4% (w/w) was selected as optimal pretreatment condi-
tion to reduce alkaline degradation of PHB. Alkaline
pretreatment improved protein release and consequently
reduced power consumptions by two-thirds that of un-
treated biomass. Cell disruption with a combined pre-
treatment with sodium chloride (140 mM), heat (60 °C,
1 h), and alkaline pH shock (pH 11.5 for 1 min) is shown
in Figure 7b. All of the data presented in this figure were
1763
obtained at 200 bar, 30 °C, and 1% (w/w) toluene for
disruption of wet cells with an exposure time of 40 min.
It can be seen that combined pretreatment of biomass
with sodium chloride resulted in increased efficiency of
disruption of cells (f ) 0.94) and decreased energy
consumption (3.25 MJ Kg-1) by a single flash discharge
of pressure.
4. Conclusion
The effect of different parameters on PHB recovery by
SC disruption of R. eutropha cells was studied. Wet cells,
cells dried at 60 °C, and freeze-dried cells were used to
explore a low-cost drying strategy for PHB recovery. It
was found that without adverse effect on M h n and a slight
decrease of product purity wet cells are more suitable for
PHB recovery. PHB recovery increased by addition of
toluene as a modifier to cells, because of increased
permeability of cell membrane resulted from solubiliza-
tion of cell membrane components by toluene.
On the basis of the results of this study, the optimum
conditions for cell disruption using SC-CO2 are 200 bar
of pressure, 30 °C temperature, and 1% (v/v) of toluene
with two times SC-CO2 pressure release. The optimum
conditions coincide with experiment 2 in Table 3, at
which the PHB recovery is 81.64% and comparable with
the results of existing methods for PHB recovery from
R. eutropha bacteria. The content of PHB increased with
the increase of cultivation time, and consequently product
purity was improved. At the late stage of fermentation
PHB content did not increase significantly (with 95%
confidence), but PHB recovery decreased as a result of
cell wall thickness and cell shape. All examined pretreat-
ments improved disruption to various degrees; however,
the sodium chloride pretreatments were less effective.
For otherwise fixed conditions increase of alkali-to-
biomass ratio (0.2-0.8% w/w) during alkaline pretreat-
ment enhanced SC-disruption efficiency. A combined salt
and alkaline pretreatment with 0.4% (w/w) NaOH fol-
lowed by a single flash discharge of SC-CO2 pressure
disrupted the cells completely. This recovery method is
recommended; however, special care is necessary to
ensure protection of disrupted homogenate against PHB
degradation by alkaline hydrolysis. It was found that
larger vessel volumes decrease energy consumption
significantly. Therefore a substantial improvement com-
pared to our previous results (38) for PHB recovery by
SC disruption of R. eutropha cells was achieved. In fact
energy consumption decreased by 200-fold (because of the
increased cell mass in the SC vessel and decreased
exposure time) to 23.76 MJ Kg-1; with advocated pre-
treatment and drying strategies energy consumption
decreased to 3.25 MJ Kg-1. Cell disruption by SC-CO2
does not fragment the cell wall, which facilitates subse-
quent PHB purification. From a practical point of view
(27) as well as for economical aspects, SC disruption is a
promising method for industrial applications.
Acknowledgment
This research project has been supported by Depart-
ment of Environment and National Research Council of
Islamic Republic of Iran (NRCI). We thank Dr. N.
Bahramifar in Department of Chemistry, Tarbiat
Modarres University (Tehran, Iran) for his help with the
supercritical fluid apparatus.
1764
References and Notes
(1) Lee, S. Y. Bacterial Poly(R-hydroxyalkanoates). Biotechnol.
Bioeng. 1996, 49, 1-14.
(2) Zinn, M.; Witholt, B.; Egli, T. Occurrence, Synthesis and
Medical Application of Bacterial Poly(β-hydroxyalkanoate).
Adv. Drug Delivery Rev. 2001, 53, 5-21.
(3) Scholz, C. Poly(β-hydroxyalkanoates) as Potential Biomedi-
cal Materials: An Overview. In Polymers from Renewable
Resources: Biopolyesters and Biocatalysts; Scholz, C., Gross,
R. A., Eds.; American Chemical Society: Washington, DC
2000; pp 328-334.
(4) Ling, Y.; Wong, H. H.; Thomas, C. J.; Williams, D. R. G.;
Middelberg, A. P. J. Pilot Scale Extraction of Poly(R-hydrox-
butyric acid) from Recombinant E. Coli by Homogenization
and Centrifugation. Bioseparation 1997, 7, 9-15.
(5) Ramsay, J. A.; Berger, E.; Voyer, R.; Chavarie, C.; Ramsay,
B. A. Extraction of Poly(R-hydroxybutyrate) Using Chlori-
nated Solvents. Biotechnol. Tech. 1994, 8, 589-594.
(6) Choi J, Lee SY. Factors Affecting the Economics of Poly(R-
hydroxyalkanoate) Production by Bacterial Fermentation.
Appl. Microbiol. Biotechnol. 1999, 51, 13-21.
(7) Hocking, P. J.; Marchessault, R. H. Biopolyesters. In
Chemistry and Technology of Biodegradable Polymers; Griffin,
G. J. L., Eds.; Chapman and Hall: London, 1994; pp 48-96.
(8) Braunegg, G.; Lefebvre, G.; Genser, K. F. Poly(R-hydroxy-
alkanotes), Biopolyesters from Renewable Resources: Physi-
ological and Engineering Aspects. J. Biotechnol. 1998, 65,
127-161.
(9) Harrison, S. T. L. Bacterial Cell Disruption: A Key Unit
Operation in the Recovery of Intracellular Products. Biotech-
nol. Adv. 1991, 9, 217-240.
(10) Hahn, S. K.; Chang, Y. K.; Kim, B. S.; Lee, K. M.; Chang,
H. N. The Recovery of Poly(R-hydroxybutyrate) by Using
Dispersions of Sodium Hypochlorite Solution and Chloroform.
Biotechnol. Tech. 1993, 7, 209-212.
(11) Hahn, S. K.; Chang, Y. K.; Kim, B. S.; Chang, H. N.
Optimization of Microbial Poly(R-hydroxybutyrate) Recovery
Using Dispersions of Sodium Hypochlorite Solution and
Chloroform. Biotechnol. Bioeng. 1994, 44, 256-261.
(12) Holmes, P. A.; Lim, G. B. Separation Process. U.S. Patent
4,910,145, 1990.
(13) Harrison, S. T. L.; Dennis, J. S.; Chase, H. A. Combined
Chemical and Mechanical Processes for the Disruption of
Bacteria. Bioseparation 1991, 2, 95-105.
(14) Tamer, I. M.; Chisti, Y.; Moo-Young, M. Disruption of
Alcaligenesis latus for Recovery of Poly(β-hydroxybutyric
acid): Comparison of High-pressure Homogenization, Bead
Milling and Chemically Induced Lysis. Ind. Eng. Chem. Res.
1998, 37, 1807-1814.
(15) Tamer, I. M.; Moo-Young, M.; Chisti, Y. Optimization of
Poly(β-hydroxybutyric acid) Recovery from Alcaligenesis la-
tus: Combined Mechanical and Chemical Treatments. Bio-
process Eng. 1998, 19, 459-468.
(16) Schutte, H.; Kroner, K. H.; Hustedt, H.; Kula, M. R.
Experiences With a 20 L Industrial Bead Mill for the
Disruption of Microorganisms. Enzyme Microb. Technol. 1983.
5, 143-148.
(17) Dillow, A. K.; Dehghani, F.; Foster, N. R.; Hrkach, J.;
Langer, R. S. Inactivation of Various Microorganisms Using
Carbon Dioxide at High Pressure. Proc. Natl. Acad. Sci.
U.S.A. 1999, 96, 10344-10348.
(18) Enomoto, A.; Nakamura, K.; Nagai, K.; Hashimoto, T.;
Hakoda, M. Inactivation of Food Microorganisms by the High-
pressure Carbon Dioxide Treatment With or Without Explo-
sive Decompression. Biosci. Biotechnol. Biochem. 1997, 61,
1133-1137.
(19) Spilimbergo, S.; Bertucco, A. Nonthermal Bacterial Inac-
tivation With Dense CO2. Biotechnol. Bioeng. 2003, 84, 627-
638.
(20) Hong, S. I.; Pyun, Y. R. Membrane Damage and Enzyme
Inactivation of Lactobacillus plantarum by High-Pressure
CO2 Treatment. Int. J. Food Microbiol. 2001, 63, 19-28.
(21) Aaltonen, O.; Rantakyla, M. Biocatalysis in Supercritical
CO2. Chem. Tech. 1991, 21, 240-248.
(22) Mesiano, A. J.; Beckman, E. J.; Russell, A. J. Supercritical
Biocatalysis. Chem. Rev. 1999, 99, 623-633.
Biotechnol. Prog., 2004, Vol. 20, No. 6
(23) Bruno, T. J.; Castro, C.; Hamel, J. P.; Palavra, A. M. F.
Supercritical Fluid Extraction of Biological Products. In
Recovery Processes for Biological Materials; Kennedy, J. F.,
Cabral, J., Eds.; John Wiley and Sons: New York, 1993; pp
303-354.
(24) Hampson, J. W.; Ashby, R. D. Extraction of Lipid-grown
Bacterial Cells by Supercritical Fluid and Organic Solvent
to Obtain Pure Medium Chain-length Polyhydroxyalkanoates.
J. Am. Oil Chem. Soc. 1999, 76, 1371-1374.
(25) Isenschmid, A.; Marison, I. W.; Stockar, U. V. The Influence
of Pressure and Temperature of Compressed CO2 on the
Survival of Yeast Cells. J. Biotechnol. 1995, 39, 229-237.
(26) Lin, H. M.; Yang, Z.; Chen, L. F. An Improved Method for
Disruption of Microbial Cells With Pressurized Carbon
Dioxide. Biotechnol. Prog. 1992, 8, 165-166.
(27) Juhasz, T.; Szekely, E.; Simandi, B.; Szengyel, Zs.; Reczey,
K. Recovery of a Recombinant Thermostable Endoglucanase
from E. Coli Using Supercritical Carbon Dioxide Cell Disrup-
tion. Chem. Biochem. Eng. Q. 2003, 17, 131-134.
(28) Kamat, S.; Beckman, E. J.; Russell, A. J. Enzyme Activity
in Supercritical Fluids. Crit. Rev. Biotechnol. 1995, 15, 41-
71.
(29) Zheng, Y.; Tsao, G. T. Avicel Hydrolysis by Cellulase
Enzyme in Supercritical CO2. Biotechnol. Lett. 1996, 18, 451-
454.
(30) Goodarznia, I.; Eikani, M. H. Supercritical Carbon Dioxide
Extraction of Essential Oils: Modeling and Simulation. Chem.
Eng. Sci. 1998, 53, 1387-1395.
(31) Wells, S. L.; DeSimone, J. CO2 Technology Platform: an
Important Tool for Environmental Problems Solving. Angew.
Chem., Int. Ed. 2001, 40, 518-527.
(32) Castor, T. P.; Hong, G. T. Supercritical Fluid Disruption
of and Extraction from Microbial Cells. U.S. Patent 5,380,-
826, 1995.
(33) Lin, H. M.; Chen, L. F. Method for Recovery of Intracellular
Material by Disruption of Microbial Cells with Carbon
Dioxide under Pressure. U.S Patent 5,306,637, 1994.
(34) Lin, H. M.; Chan, E. C.; Chen, C.; Chen, L. F. Disintegra-
tion of Yeasts Cells by Pressurized Carbon Dioxide. Biotech-
nol. Prog. 1991, 7, 201-204.
(35) Nakamura, K.; Enomoto, A.; Fukushima, H.; Nagai, K.;
Hakoda, M. Disruption of Microbial Cells by the Flash
Discharge of High-Pressure Carbon Dioxide. Biosci., Biotech-
nol., Biochem. 1994, 58, 1297-1301.
(36) Foster, J. W.; Cown, R. M.; Mag, T. A. Rupture of Bacteria
by Explosive Decompression. J. Bacteriol. 1962, 83, 330-334.
(37) Fraser, D. Bursting Bacteria by Release of Gas Pressure.
Nature 1951, 167, 33-34.
(38) Hejazi, P.; Vasheghani-Farahani, E.; Yamini, Y. Super-
critical Fluid Disruption of Ralstonia eutropha for Poly(β-
hydroxybutyrate) Recovery. Biotechnol. Prog. 2003, 19, 1519-
1523.
(39) Montgomery, D. C. Design and Analysis of Experiments;
John Wiley and Sons: New York, 1997; pp 197-249.
(40) Roy, R. K. A Primer on the Taguchi Method; Van Nostrand
Reinhold: New York, 1990; pp 40-99.
(41) Bradford, M. M. A Rapid and Sensitive Method for the
Quantitation of Microgram Quantities of Protein Utilizing the
Principle of Protein-dye Binding. Anal. Biochem. 1976, 72,
248-254.
(42) Braunegg, A.; Sonnleitrer, B.; Lafferty, R. M. A Rapid Gas
Chromatogram Method for the Determination of Poly(R-
hydroxybutyric acid) in Microbial Biomass. Eur. J. Appl.
Microbiol. Biotechnol. 1978, 6, 29-37.
(43) Berger, E.; Ramsay, B. A.; Ramsay, J. A.; Chavarie, C. Poly-
(R-hydroxbutyric acid) Recovery by Hypochlorite Digestion
of Non-PHB Biomass. Biotechnol. Tech. 1989, 3, 227-232.
(44) Suzuki, T.; Deguchi, H.; Yamane, T.; Shimuzu, S.; Gekko,
K. Control of Molecular Weight of Poly(β-hydroxybutyric acid)
Produced in Fed-batch of Protomonas extorquens. Appl.
Microbiol. Bioeng. 1988, 27, 487-491.
Biotechnol. Prog., 2004, Vol. 20, No. 6
(45) Chen, Y.; Xu, Q.; Yang, H.; Gu, G. Effect of Cell Fermenta-
tion Time and Biomass Drying Strategies on the Recovery of
Poly(R-hydroxyalkanoates) from Alcaligenesis eutrophus Us-
ing a Surfactant Chelate Aqueous System. Process Biochem,
2001, 36, 773-779.
(46) Wase, D. A.; Patel, Y. R. Effect of Cell Volume on
Disintegration by Ultrasonic. J. Chem. Technol. Biotechnol.
1985, 35, 165-173.
(47) Resch, S.; Gruber, K.; Wanner, G.; Slater, S.; Dennis, D.;
Lubitz, W. Aqueous Release and Purification of Poly(R-
hydroxybutyrate) from Escherichia coli. J. Biotechnol. 1998,
65, 173-182.
(48) Witte, A.; Wanner, G.; Sulzner, M.; Lubitz, W. Dynamics
of PhiX174 Protein E-mediated Lysis of Escherichia coli. Arch.
Microbiol. 1992, 157, 381-388.
(49) Davies, L. Factorial Design at Two Levels. In Efficiency
in Research, Development and Production: The Statistical
1765
Design and Analysis of Chemical Experiments; Davies, L., Ed.;
The Royal Society of Chemistry: Cambridge, 1993; pp 72-
89.
(50) McHugh, M. A.; Krukonis, V. J. Supercritical Fluid Extrac-
tion: Principles and Practice; Butterworth: Boston, 1994.
(51) Kashket, E. R. Bioenergetics of Lactic Acid Bacteria:
Cytoplasmic pH and Osmotolerance. FEMS Microbiol. Rev.
1987, 46, 223-244.
(52) Mackey, B. M.; Forestiere, K.; Isaacs, N. Factors Affecting
the Resistance of Listeria monocytogenes to High Hydrostatic
Pressure. Food Biotechnol. 1995, 9, 1-11.
Accepted for publication September 15, 2004.
BP0498037