Beruflich Dokumente
Kultur Dokumente
com
Journal of Biotechnology 132 (2007) 251258
Large-scale production and efcient recovery of PHB with desirable
material properties, from the newly characterised
Bacillus cereus SPV
S.P. Valappil
a,b
, S.K. Misra
b
, A.R. Boccaccini
b
, T. Keshavarz
a
, C. Bucke
a
, I. Roy
a,
a
Department of Molecular and Applied Biosciences, School of Biosciences, University of Westminster,
115 New Cavendish Street, London W1W 6UW, UK
b
Department of Materials, Imperial College London, London SW7 2BP, UK
Received 26 September 2006; received in revised form 13 March 2007; accepted 15 March 2007
Abstract
A newly characterised Bacillus strain, Bacillus cereus SPV was found to produce PHB at a concentration of 38% of its dry cell weight in shaken
ask cultures, using glucose as the main carbon source. Polymer production was then scaled up to 20 L batch fermentations where 29% dry cell
weight of PHB was obtained within 48 h. Following this, a simple glucose feeding strategy was developed and the cells accumulated 38% dry
cell weight of PHB, an increase in the overall volumetric yield by 31% compared to the batch fermentation. Sporulation is the cause of low PHB
productivity fromthe genus Bacillus [Wu, Q., Huang, H., Hu, G.H., Chen, J., Ho, K.P., Chen, G.Q., 2001. Production of poly-3-hydroxybutyrate by
Bacillus sp. JMa5 cultivated in molasses media. Antonie van Leeuwenhoek 80, 111118]. However, in this study, acidic pH conditions (4.55.8)
completely suppress sporulation, in accordance with Kominek and Halvorson [Kominek, L.A., Halvorson, H.O., 1965. Metabolism of poly--
hydroxybutyrate and acetoin in Bacillus cereus. J. Bacteriol. 90, 12511259], and result in an increase in the yield of PHB production. This
observation emphasises the potential of the use of Bacillus in the commercial production of PHB and other PHAs. The recovery of the PHB
produced was optimised and the isolated polymer characterised to identify its material properties. The polymer extracted, was found to have similar
molecular weight, polydispersity index and lower crystallinity index than others reported in literature. Also, the extracted polymer was found to
have desirable material properties for potential tissue engineering applications.
2007 Elsevier B.V. All rights reserved.
Keywords: Bacillus cereus SPV; Polyhydroxybutyrate; Fermentation; Polymer extraction; Polydispersity index; Crystallinity index
1. Introduction
Polyhydroxyalkanoates (PHAs) are polyesters of 3-, 4-, 5-
and 6-hydroxyalkanoic acids, produced by a variety of bacterial
species and are classied as both biodegradable and biocompat-
ible polyesters (Chen and Wu, 2005). They are synthesised by
bacteria as storage compounds for energy and carbon, normally
in the presence of excess carbon with at least one nutrient essen-
tial for growth, such as nitrogen, phosphorus, sulphur or oxygen
present in limiting concentration (Anderson and Dawes, 1990).
The material properties of these polymers vary depending on
C after growth on
nutrient agar (containing (g/L): Lab-Lemco Powder, 1.0;
yeast extract, 2.0; peptone, 5.0; sodium chloride, 5.0; agar,
15.0).
2.3. Growth media and culture conditions
2.3.1. Shaken asks culture
A semi-dened PHB production medium, Kannan and
Rehacek medium (Kannan and Rehacek, 1970), was used with
modication, by removing the calcium carbonate from its origi-
nal composition, for the production of PHB in both shaken ask
and 20 L fermentor studies. The pH of the medium was adjusted
to 6.8 before autoclaving. The modied medium contained in
g/L, glucose, 20; Difco yeast extract, 2.5; potassium chloride, 3;
ammonium sulphate, 5.0; 100 mL of defatted soybean dialysate
(prepared from 10 g of defatted soybean meal in 1000 mL of
distilled water for 24 h at 4
C (data not
shown). In order to maintain reproducibility and comparable
conditions of aeration after sample removal, six asks for each
set (ask sets I, II and III in duplicates) were simultaneously
inoculated with the same inoculum. Samples were taken from
the ask set I immediately after inoculation, however, the ask
sets II and III were allowed to grow until 48 and 96 h, respec-
tively, before removing samples. A maximum of 50 mL culture
broth was removed aseptically from each ask for the analy-
sis of dry cell weight, PHB accumulation, glucose consumption
and pH variation, during the growth, so as to maintain simi-
lar media volume to head space ratio, crucial for aeration, in
each case.
2.3.2. Fermentor culture
The production of PHB using B. cereus SPV was carried out
in 20 L fermentors with a 14 L working volume. The fermentors
were sterilized while containing only the salts of the modied
Kannan and Rehacek medium (Kannan and Rehacek, 1970).
Separately sterilized glucose and soybean dialysate were added
to the fermentors aseptically before inoculation with 1.4 L of
a 24 h inoculum. The pH of the medium was checked before
inoculation, and was adjusted to 6.8 using IN NaOH. The dis-
solved oxygen tension (DOT) was set at 100% air saturation
at the beginning of the run. The pH, DOT, glucose concentra-
tion and cell optical density were monitored throughout the run.
The impeller speed was kept at 250 rpm, airow rate at 1.0 vvm
and the temperature at 30
Cfor 12 h, centrifuged at
10,000 g, and the residue was washed twice each with water,
acetone, ethanol and diethyl ether (Ramsay et al., 1994; Manna
et al., 1999). Finally, the residue was dried and subjected to
soxhlet extraction for 24 h using chloroform as the solvent. The
chloroform containing the PHB was then concentrated by vac-
uum rotary evaporation and precipitated using 10 volumes of
ice-cold methanol. The precipitate obtained was centrifuged and
air-dried.
2.6.2. Chloroform extraction
The polymer was extractedfromcells bystirring1 gof freeze-
dried cell in 100 mL chloroform for 48 h at 37
C and puried
by re-precipitation with 10 volumes of ice-cold methanol (Hahn
et al., 1995).
2.6.3. Chloroformhypochlorite dispersion extraction
PHB was extracted by treating 1 g of freeze-dried cells with
a dispersion containing 50 mL of chloroform and 50 mL of
a diluted (30%) sodium hypochlorite solution in water, in an
orbital shaker at 100 rpm. The cell powder was treated at 38
C
for 1 h. The mixture obtained was then centrifuged at 4000 g
for 10 min, which resulted in three separate phases. PHB was
recovered from the bottom phase, i.e. that of chloroform by pre-
cipitation using 10 volumes of ice-cold methanol (Hahn et al.,
1995).
2.7. Characterisation and quantication of PHB
2.7.1. Solid-state
13
C NMR analysis
13
C NMR cross-polarization with magic-angle sample spin-
ning (CP/MAS) experiments were performed at 75.4 MHz on a
VarianUnityInova spectrometer usinga 5 mmmagic-angle spin-
ning (MAS) probe with the following acquisition and processing
parameters: spinning speed 5.11 kHz, number of acquisitions
316, recycle delay 5 s and contact time 1 ms. Selective spectra of
the quaternary and methyl carbons were obtained using the dipo-
lar dephasing (non-quaternary suppression) pulse sequence
(Opella andFrey, 1979). This studywas carriedout at the Depart-
ment of Chemistry, University of Durham, UK.
2.7.2. Gas chromatography (GC)
For the quantication of the PHB, following slight modi-
cation of the gas chromatographic method of Huijberts et al.,
1994, was employed. One milligramper millilitre of sample and
1 mg/mL methyl benzoate in chloroformwas added to a mixture
of 2 mL of 15% sulphuric acid in methanol (ratio 1:1) at 100
C
for 5 h in a reux. After the reaction, the tubes were cooled on ice
for 5 min, 1.0 mL distilled water was added and the tubes were
vortexed for 1 min. After phase separation the organic phase was
collected and dried over anhydrous sodium sulphate. The analy-
sis was carried out using a Sigma 3Bgas chromatograph (Perkin
Elmer, USA) equipped with a BP21 (SGE Europe Ltd., Kiln
Farm, Milton Keynes, UK) column (50 mlength, 0.25 mminter-
nal diameter and 0.25 mlmthickness). The sample (1 L), in
chloroform, was injected with helium (1 mLmin
1
) as the car-
rier gas. The injector temperature was 225
C at 20
C/min and
held at the nal temperature for 10 min.
2.7.3. Fourier transform infrared spectroscopy (FTIR)
Freeze-dried, precipitated PHB from B. cereus SPV was
used to prepare KBr discs (sample:KBr, 1:100) (Kemp, 1989).
An FTIR spectrum 1720X spectrometer (Perkin Elmer, USA)
was used under the following conditions: spectral range,
4000400 cm
1
; window material, CsI; 16 scans; resolution
4 cm
1
; the detector was a temperature-stabilized, coated FR-
DTGS detector.
2.7.4. Molecular mass measurements using gel permeation
chromatography (GPC)
Molecular mass analysis was conducted with puried PHB,
which was dissolved in chloroform (1 mg/mL PHB) and intro-
duced into a GPC system. The GPC system was equipped with
a PLgel guard plus 2 mixed bed-B column (30 cm10 m).
The eluted polymer was detected with a differential refractome-
ter. The data were collected and analysed using Viscotek Trisec
2000 and Trisec 3.0 software. This work was carried out at
RAPRA, UK.
2.7.5. Differential scanning calorimetry
Thermal analysis was performed in a Perkin Elmer DSC 7
(Pyris software 3.81). Samples were heated from25 to 200
Cat
a heating rate of 10
C
254 S.P. Valappil et al. / Journal of Biotechnology 132 (2007) 251258
for 2 min, the samples were rapidly quenched to 50
C. They
were then heated again at a rate of 10
C/min to 200
C (Run 2).
DSC curves were recorded at Run 2.
3. Results and discussion
3.1. PHB production in shaken asks
To optimise the chance of a successful fermentation at a larger
scale, the process parameters need to be set at the shaken ask
level and translated into a set of standard parameters in large-
scale fermentation.
In shaken ask cultures, the cell mass increased steadily, lead-
ing to a maximum cell density within 24 h of cultivation after
whichthere was a gradual decrease (Fig. 1). The pHof the culture
medium decreased during the growth, from its initial value of 7
to a minimumof 4.5. Cessation of logarithmic growth coincided
with the approach of the pH minimum and rapid consump-
tion of glucose. PHB accumulated rapidly during the stationary
phase and reached a maximum concentration of 38% of dry cell
weight (dcw) at 60 h of growth. We found modied Kannan and
Rehacek medium to be the best for PHB production using the
newly characterised B. cereus SPV (Valappil et al., 2007b) as
compared to other reported minimal media such as potassium
decient production medium (Wakisaka et al., 1982) and phos-
phate decient production medium(Lopez et al., 1986) (data not
shown). Another unique and important observation in this study
is that once maximal PHBconcentration was achieved, the PHB
concentration remained almost constant for a period of 64 h.
Thus, it seems that the PHB produced was not degraded in order
to be utilized for sporulation, in contrast to the report of Wu et
al., 2001. The extent of sporulation was tested at each time point
in this study using the SchaefferFulton method and the results
obtained were always negative. The unbuffered medium used in
this study led to low values of pH for the media (4.55.8) during
the stationary phase, resulting in prevention of PHB degrada-
tion. This observation was consistent to a previous report on
a Bacillus sp. where low pH conditions have been reported to
Fig. 1. Growth and PHB accumulation of the newly identied Bacillus cereus
SPV in shaken asks. Changes in the pH (), dry cell weight (), glucose
() and PHB () when grown in the Kannan and Rehacek medium. All the
measurements were duplicatedandthe relative standarddeviations were 12%.
Fig. 2. Growth and PHB accumulation of the newly characterised Bacillus
cereus SPV in a 20 L fermentor. Changes in the pH (), dry cell weight (),
dissolved oxygen tension (DOT, % air saturation) () and PHB () when the
newly identied Bacillus cereus SPV was grown in the Kannan and Rehacek
medium.
inhibit utilization of the polymer as well as spore formation in
B. cereus T (Kominek and Halvorson, 1965). This observation
in the present study is crucial for the production of PHB using
B. cereus SPV and other PHB producing Bacillus spp.
3.2. PHB production in fermentors (batch)
The successful production of PHBin shaken ask cultures led
to the large-scale production of PHB in batch cultures (Fig. 2).
The two main medium parameters that are known to affect the
PHB yield are the pH of the medium and the dissolved oxy-
gen tension (DOT) (Wu et al., 2001). The pH of the medium
was not controlled leading to a low pH of 5.2. As in the case
of the shaken ask cultures, PHB accumulation was found to
begin after the initial logarithmic growth, after the pHminimum
of the culture was achieved. As expected, PHB production was
accompanied by a simultaneous decrease in glucose concentra-
tion. The organism reached stationary phase at 23 h of growth
and PHB accumulation reached a maximum (29%dcw) at 48 h
of growth after which the concentration remained constant. The
extent of sporulation was also monitored throughout the fermen-
tation using the SchaefferFulton method. No sporulation was
observed; hence, as in the shaken ask cultures, there was no
degradation of the PHB, possibly due to the lack of sporulation.
However, the yield of PHBproduction was not very high leading
to the trial of fed-batch fermentation with an aim to increase the
PHB yield.
3.3. PHB production in fermentors (fed-batch)
The fermentation prole obtained from the batch grown B.
cereus SPV in 20 L fermentors along with the shaken ask
growth analysis was used to develop a simple nutrient feeding
strategy. As seen in Fig. 3, the cell mass increased steadily and
attained its maximum value within 24 h of cultivation. The cul-
ture reached stationary phase at 24 h, which coincided with an
increase in dissolved oxygen concentration. During the growth,
the pH of the culture medium decreased from its initial value of
7 to a minimum of 5.2 (Fig. 3).
S.P. Valappil et al. / Journal of Biotechnology 132 (2007) 251258 255
Fig. 3. Growth and PHB accumulation of the newly identied Bacillus cereus
SPV in 20 L fed-batch fermentation. Changes in pH (), dry cell weight (),
dissolved oxygen tension (DOT, %air saturation) () and PHB() when grown
in the Kannan and Rehacek mediumwith feeding of glucose in order to maintain
nal glucose concentration above 10 g/L.
The two major signicant differences between the batch and
fed-batch fermentations were the time required for maximal
accumulation of PHB, which showed a large shift from 48 to
32 h and the nal PHB yield, which increased from 29%dcw to
a much higher value of 38%dcw, an increase of 31%in the over-
all volumetric yield (Table 1). These two changes would lead to
a much higher level of PHB accumulation within a shorter time
interval leading to a useful decrease in the resources required
for scaling up the production of PHB. Hence, in the fed-batch
fermentation, feeding glucose to maintain a nal concentration
above 10 g/L has successfully improved the large-scale pro-
duction of PHB. PHB concentration has remained constant at
38%dcw once the maximum concentration has been achieved
even after 32 h of growth, consistent with observations made
in shaken ask and batch fermentations. We postulate that low
pH values have resulted in the inhibition of intracellular PHB
depolymerisation, hence preventing the wasteful degradation of
PHB after it has been synthesised.
Also, cell growth and PHAaccumulation need to be balanced
to avoid incomplete accumulation of PHA or premature termi-
nation of PHA synthesis at low cell concentration. The highest
residual cell concentration, practically achievable, with a high
Table 1
Comparison of the PHB yield (% dcw) obtained in the batch and fed-batch
fermentation of newly identied Bacillus cereus SPVin the Kannan and Rehacek
medium
Time (h) Dry cell weight (g/L) PHA yield (% dry cell weight)
Batch Fed-batch Batch Fed-batch
20 2.10 2.40 16.00 17.00
24 2.50 3.00 20.00 26.86
28 2.50 3.00 23.00 30.00
32 2.49 3.00 24.00 38.00
36 2.49 2.90 25.00 38.00
40 2.49 2.80 26.00 38.00
44 2.50 2.75 28.00 38.00
48 2.50 2.73 29.00 38.00
52 2.49 2.68 29.00 38.00
56 2.49 2.84 29.00 37.00
60 2.48 3.00 29.00 38.00
PHA content will give the best results. In the present study,
the highest cell concentration achieved using the production
medium was 2.5 g/L in the batch and 3 g/L in the fed-batch fer-
mentation. To achieve this, the fermentation should be stopped at
48 h in the batch culture (Y
P/X
=0.29 g/g; Y
P/S
=0.071g/g) and at
32 h in the fed-batch culture (Y
P/X
=0.38 g/g; Y
P/S
=0.114 g/g)
since the productivity is highest at these times. As the PHA
content remains stable from this point onwards prolonged
cultivation to achieve higher PHA concentration will be coun-
terproductive.
3.4. Extraction of PHB
After the successful scaling up of PHB production,
the efciency of polymer extraction using hypochlorite
and organic solvents were evaluated. Three different tech-
niques namely chloroform extraction (Ramsay et al., 1994),
chloroformhypochlorite dispersion extraction (Hahn et al.,
1993, 1994, 1995) and soxhlet extraction (Ramsay et al., 1994;
Manna et al., 1999) were tested for the rst time, for PHBextrac-
tion from B. cereus SPV. The soxhlet extraction involves the
extraction of the polymer fromdesiccated cells into chloroform.
The resulting solution is ltered to remove debris, concentrated,
and the polymer precipitated using methanol or ethanol, leav-
ing low molecular weight lipids in solution. In both the soxhlet
and chloroform hypochlorite dispersion techniques, the anionic
surfactant hypochlorite is used to remove cellular debris before
the intracellular polymer is dissolved in chloroform. Hypochlo-
rite extraction is known to cause severe degradation of PHB
leading to a signicant reduction of the polymer chain length
(Berger et al., 1989). However, the extent of this degradation
varies considerably between organisms. Within Gram-negative
organisms, the amount of PHB degraded to a lower molecular
weight compound in A. eutrophus (75% reduction in the num-
ber average molecular weight, M
N
) during the recovery process,
was signicantly higher compared to that in recombinant E. coli
(15%reduction in M
N
). The chloroformextraction method relies
solely on the solubility of the PHB in chloroform and does not
involve any treatment with sodium hypochlorite. This method
is widely used to recover PHB as it results in less polymer
degradation.
In this work, the chloroform extraction method gave the
highest crude yield of the polymer (31%dcw), followed by the
dispersion method (30%dcw) and the lowest yield (27%dcw)
was obtained using soxhlet extraction (Table 2). The level
of PHB purity was then calculated from the total amount
of PHB in the isolated powder as determined by GC com-
pared to the crude weight of polymer powder. The results
were veried by comparing it with standard PHB obtained
from SigmaAldrich Company Ltd. (Dorset, England). It was
found that the purity of PHB extracted was highest for soxh-
let extraction (99%), followed by the chloroformhypochlorite
dispersion extraction (95%) and nally the chloroform extrac-
tion method (92%). This could be attributed to the rigorous
washing with solvents and the physical separation of the
extract from the cellular debris in the soxhlet extraction
method.
256 S.P. Valappil et al. / Journal of Biotechnology 132 (2007) 251258
Table 2
Thermal and molecular characterisation of PHB samples extracted from the Bacillus cereus SPV cells using three different techniques
Sample Yield (%) Purity (%) M
w
PDI T
m
(
C) T
g
(
C) H
f
(J/g) X
c
(%) CI
Soxhlet 27 99 1,100,00 1.75 169.71 2.04 86.140 57.66 0.76
Chloroform 31 92 882,000 2.6 160.83 2.45 81.289 54.42 0.74
ChloroformHOCl dispersion 30 95 885,000 3.1 171.71 2.72 95.72 64.08 0.78
M
w
, molecular weight; PDI, polydispersity index; T
m
, melting temperature; T
g
, glass transition temperature; H
f
, heat of fusion; X
c
, crystallinity; CI, crystallinity
index.
3.5. Characterisation of the extracted PHB
3.5.1.
13
C NMR CP/MAS analysis
The NMR analysis was used to determine the structure of the
isolatedpolymer fromB. cereus SPVgrowninthe modiedKan-
nan and Rehacek medium. Four narrowlines appeared with very
strong intensities which were identical to the CP/MAS
13
CNMR
spectra of PHBreported previously (Doi et al., 1989). These four
peaks were assignable to the methyl (CH
3
; 21.2 ppm), methy-
lene (CH
2
; 42.7 ppm), methine (CH; 68.5 ppm) and carbonyl
(C=O; 169.7 ppm) carbon resonance of PHB (Doi et al., 1986,
1989). This analysis thus conrmed the molecular composition
of the polymer to be PHB.
3.5.2. Differential scanning calorimetry (DSC)
To corroborate the results described above and to exam-
ine possible morphological differences in the PHB obtained by
different extraction techniques, the melting temperatures and
enthalpies of fusion of the PHB samples obtained were deter-
mined using DSC. The thermal properties of the polymer such
as the glass transition temperature (T
g
) and the melting tem-
perature (T
m
) are crucial for polymer processing. The melting
temperature obtained for PHBextracted using the three different
methods were slightly lower compared to what is reported in the
literature, between 173 and 180
C, respectively) as compared
to a relatively lower value for the sample obtained using the
chloroform extraction technique (160.83