Bioprocess Biosyst Eng

DOI 10.1007/s00449-015-1428-1

ORIGINAL PAPER

Effect of medium components and culture conditions in Bacillus

subtilis EA-CB0575 spore production
Luisa F. Posada-Uribe1,2 • Magally Romero-Tabarez3 • Valeska Villegas-Escobar1

Received: 13 November 2014 / Accepted: 12 June 2015

Ó Springer-Verlag Berlin Heidelberg 2015

Abstract Bacillus subtilis spores have important compared to that of PB design. These results indicate the
biotechnological applications; however, achieving both, potential of B. subtilis EA-CB0575 to produce both, high
high spore cell densities and sporulation efficiencies in spore cell densities and sporulation efficiencies, with very
fermentation, is poorly reported. In this study, medium low nutrient requirements and short incubation period
components and culture conditions were optimized with which can represent savings of process production.
different statistical methods to increase spore production of
the plant growth promoting rhizobacteria B. subtilis EA-
Keywords Bacillus subtilis  Bacterial culture
CB0575. Key medium components were determined with
optimization  Spore production  Sporulation efficiency
Plackett–Burman (PB) design, and the optimum concen-
tration levels of two components (glucose, MgSO47H2O)
were optimized with a full factorial and central composite Introduction
design, achieving 1.37 9 109 CFU/mL of spore cell den-
sity and 93.5 % of sporulation efficiency in shake flask. Aerobic Endospore-Forming Bacteria (AEFB), especially
The optimized medium was used to determine the effect of some species of Bacillus spp., have received growing
culture conditions on spore production at bioreactor level, attention because of their industrial potential in the
finding that maintaining pH control did not affect signifi- development of biopesticides and biofertilizers. In the
cantly spore production, while the interaction of agitation agricultural industry, spores are used as alternatives to
and aeration rates had a significant effect on spore cell agrochemicals as Plant Growth Promoting Rhizobacteria
density. The overall optimization generated a 17.2-fold (PGPR) or as Biological Control Agents (BCA) [1–4]. In
increase in spore cell density (8.78 9 109 CFU/mL) and the market economy, Bacillus-based products represent
1.9-fold increase in sporulation efficiency (94.2 %) about 85 % of the commercially available bacterial BCA
[5] but to be able to commercialize these bio-products,
industrial exploitation of spores requires high cell density
Electronic supplementary material The online version of this bio-reaction and good sporulation efficiency; for this rea-
article (doi:10.1007/s00449-015-1428-1) contains supplementary son, spore production is a key step in bio-products devel-
material, which is available to authorized users.
opment when AEFB are the active ingredient [6].
& Valeska Villegas-Escobar Bacillus spp. uses a system of Quorum-Sensing to ini-
[email protected] tiate its sporulation process that occurs in high cell density
1
populations and only if the cell is in a stressful condition
Grupo de Investigación en Ciencias Biológicas y Bioprocesos
(CIBIOP), Department of Process Engineering, Universidad
[7]. Furthermore, during the stationary growth phase in a
EAFIT, Cra. 49 # 7 Sur-50, Medellı́n, Antioquia, Colombia fermentation process, when nutrients are exhausted, the
2
Centro de Investigaciones del Banano-AUGURA, Carepa,
culture initiates sporulation at a cell density of about
Colombia 108 cells/mL [8]. Under ideal conditions, typical cell den-
3
Universidad Nacional de Colombia, Calle 59A # 63-20,
sities and sporulation efficiencies for B. subtilis will be in
Medellı́n, Antioquia, Colombia the range of 1.00 9 108–1.52 9 1010 CFU/mL [9–12] and

123
 Bioprocess Biosyst Eng

30–80 %, respectively using serial dilution and plate count 0.59 g/L, KH2PO4 6.0 g/L, meat extract 5.0 g/L, special
methods [11–14]. Other studies have reported higher peptone 3.0 g/L, NaCl 0.01 g/L and stock salt solution
sporulation efficiencies or higher spore cell densities for (1.136 mL/L of FeSO47H2O 0.1 M, 300 lL/L of
other Bacillus species different from B. subtilis [3, 6, 15, ZnSO47H2O 0.1 M, 9.9 mL/L of CaCl2 0.1 M, 30 mL/L
16], but to our knowledge obtaining high values of both of MnCl2 0.1 M). Modified SBM medium was composed
variables has not been reported yet. of the same components but we replaced meat extract by
Different optimum culture media and culture conditions yeast extract in the same concentration.
for AEFB sporulation have been reported, where each
particular strain has its own requirements and optimum Shake flask culture conditions
conditions [10, 12, 17, 18]. For example, the presence of
glutamate in the medium benefits the growth of B. cereus For shake flask inoculum preparation, a colony of the strain
while for other species of Bacillus this compound inhibits grown in TSA was transferred to 250 mL Erlenmeyer flaks
growth [19]. It has also been reported that adding salts to containing 50 mL of TSB, and incubated at 30 °C and
the media improves sporulation, but high concentrations of 150 rpm for 24 h. The culture was then centrifuged and the
heavy metals contained in these salts could affect growth pellet suspended in the respective culture medium with a
[10]. Elements such as Ca, Mn, Mg, Fe and Zn in appro- final optical density (OD600) of 1.0 (equivalent to
priate concentrations are essential in sporulation since they 1.0 9 108 CFU/mL). Forty mL of each culture medium
are present in the spore layers and enable it to resist high and 5 mL of a previously filtrated stock salt solution was
temperatures [11] and the last two elements accelerate the added to a 250 mL Erlenmeyer flask, inoculated with 5 mL
sporulation process [3, 20]. Regarding the culture condi- of the inoculum culture, incubated at 30 °C at 150 rpm for
tions, several studies have evaluated the effect of different 96 h. Total cells and spore cell densities were determined
effective volumes (1–70 L), aeration rates (0.5–2.0 vvm), by the spread plate technique.
agitation rates (200–500 rpm), pH (5.0–9.0) and inoculum
size (1–3 %) in order to obtain large amounts of biomass Bioreactor culture conditions
and secondary metabolites [3, 9, 12, 21, 22], concluding
that optimum culture conditions are very specific for each The inoculum used for the bioreactor fermentations was
strain. Therefore, optimization of culture medium and prepared by transferring two colonies of B. subtilis EA-
culture conditions of new strains can lead to industrially CB0575 grown in TSA to a 2000 mL Erlenmeyer flask
viable processes. containing 240 mL of modified SBM and incubated at
In this study we designed and optimized a culture 30 °C and 150 rpm for 24 h. Three percent (v/v) of
medium to increase spore cell density and sporulation inoculum with a final optical density of 2.0 and 330 mL of
efficiency of B. subtilis EA-CB0575, a PGPR strain, at sterile stock salts solution was added to 7.43 L of sterile
shake flask level. The optimized medium was used to modified SBM into a 14 L Bioreactor (BioFlo110, New
optimize the culture conditions pH, agitation and aeration Brunswick Scientific Co) equipped with a diffuser ring type,
rates in a 14 L bioreactor. Finally kinetics of cell growth two Rushton turbine, 6 impellers of flat blades and 4 baffles.
and substrate uptake are described in optimized conditions.
Design and formulation of culture medium

Materials and methods Plackett and Burman design 2(7 9 3/32) with 3 central points
was used to determine which of eight medium components
Microorganism and culture mediums significantly affected B. subtilis EA-CB0575 total cell
density, spore cell density and sporulation efficiency in
The PGPR, B. subtilis EA-CB0575, was isolated from the shake flask. The components included glucose, MgSO47-
rhizosphere of banana plants in Uraba (Northeast Colom- H2O, MnCl24H2O, KH2PO4, yeast extract, meat extract,
bia, 1,352,081 N, 1,044,577 E) in September 2009 and peptone, and (NH4)2SO4 which were evaluated in three
identified by analysis of 16S rDNA gene sequencing concentrations, high (?1), medium (0) and low (-1) in
(Genbank Accession Number KC170988). The bacteria fifteen experimental trials (Table 1). The experiment was
was stored in TSB (Tripticase Soy Broth, Merck) plus independently repeated three times.
20 % v/v glycerol at -80 °C (Humboldt Institute Collec-
tion No. 191) and activated in TSA (Merck) for 24 h at Medium optimization
30 °C before any experimental use.
Optimized medium SBM (Sporulation Bacillus Med- Full factorial and central composite (CC) designs were
ium) was composed of: glucose 1.04 g/L, MgSO47H2O employed to optimize the component concentrations in

123
 Table 1 PBD design for total cell and spore cell densities of B. subtilis EA-CB0575
Bioprocess Biosyst Eng

Factors (g/L) Response variables

Treatment Glucose MgSO47H2O MnCl24H2O KH2PO4 Yeast Meat Peptone (NH4)2SO4 Total cell density Spore cell density Sporulation
(X1) (X2) (X3) (X4) extract extract (X7) (X8) (109 CFU/mL) (109 CFU/mL) efficiency (%)
(X5) (X6)

M1 2 0 0 6 5 5 0 4 1.12 ± 1.15 0.61 ± 0.30 54.2 ± 22.0

M2 2 0 0.5 6 5 0 3 4 0.00 ± 0.00 0.00 0.0
M3 20 0 0.5 0 0 0 3 4 0.09 ± 0.00 0.00 0.0
M4 2 0.5 0.5 0 5 0 0 0 0.78 ± 0.23 0.51 ± 0.02 65.5 ± 13.6
M5 20 0 0 0 5 5 3 0 1.55 ± 1.10 0.00 0.0
M6 11 0.25 0.25 3 2.5 2.5 1.5 2 0.29 ± 0.13 0.00 0.0
M7 20 0.5 0 6 5 0 3 0 0.12 ± 0.00 0.00 0.0
M8 11 0.25 0.25 3 2.5 2.5 1.5 2 0.19 ± 0.01 0.00 0.0
M9 20 0.5 0.5 0 5 5 0 4 1.07 ± 0.01 0.00 0.0
M10 11 0.25 0.25 3 2.5 2.5 1.5 2 0.12 ± 0.00 0.00 0.0
M11 2 0.5 0.5 6 0 5 3 0 2.01 ± 0.00 1.87 ± 0.05 93.2 ± 0.3
M12 2 0 0 0 0 0 0 0 0.00 ± 0.02 0.00 0.0
M13 20 0.5 0 6 0 0 0 4 0.08 ± 0.00 0.00 0.0
M14 20 0 0.5 6 0 5 0 0 0.06 ± 0.00 0.00 0.0
M15 2 0.5 0 0 0 5 3 4 1.03 ± 0.01 0.52 ± 0.15 50.7 ± 16.0
Cell density 0.570 0.041* 0.301 0.599 0.083 0.020* 0.954 0.212 Significance a = 0.05
P value
Spore cell density 0.001* 0.081 0.991 0.993 0.992 0.964 0.986 0.951
P value
Sporulation 0.022* 0.096 0.036 0.492 0.989 0.231 0.650 0.231
efficiency
P value
Bold values are statistically significant (P \ 0.05)

123
 Bioprocess Biosyst Eng

order to get optimum spore cell densities and sporulation of the variable yðxi Þ in the multiple optimization. The
efficiencies of B. subtilis EA-CB0575 in shake flask. Two weights used were equal to 1 for each of the two responses,
components (glucose and MgSO47H2O) that significantly spore cell density and sporulation efficiency. With the di
affected the response variables were optimized by a 32 full functions defined, they were combined to obtain a global
factorial design. The concentration levels of glucose were desirability function (D) represented by Eq. (4).
0.8, 1.4 and 2.0 g/L and the same levels for MgSO47H2O. 1
!P1
P
n n
The other medium components were kept constant at the   ri Y
n ri

concentration they had in the best assay of the Plackett and D ¼ d1r1  d2r2 :::  dnrn i¼1 ¼ diri i¼1 ð4Þ
i¼1
Burman design (peptone 3.0 g/L, meat extract 5.0 g/L,
KH2PO4 6.0 g/L, MnCl24H2O 0.5 g/L, and stock salt where D is the value of the global desirability function, (d1 )
solution as stated before). A series of nine experiments and (d2 ) are the partial desirability functions computed for
were carried out in duplicate and the spore cell density and the spore cell density and sporulation efficiency, n is the
sporulation efficiency were determined 96 h after culture. number of variables, and ri is the relative importance
This experiment was independently repeated three times assigned to each response.
and the average of the three was statistically analyzed with This experiment was independently repeated three
Design Expert 8.0.7.1. The model employed to correlate times. The optimized medium was named SBM medium
the response variable to the independent variables is (Sporulation Bacillus Medium).
showed in Eq. (1):
X
n Validation of the optimum medium and evaluation
y ð x i Þ ¼ b0 þ bi  xi þ e ð1Þ of modified media at flask level
i¼1

where y(xi ) represents the response variable, xi is the To validate the prediction of the mathematical optimiza-
independent variable (medium components), bo is the tion, spore cell density and sporulation efficiency of B.
interception coefficient, bi the coefficient of the linear subtilis EA-CB0575 were determined through an inde-
effect and e is the random error [23]. pendent trial with SBM medium in Erlenmeyer flasks.
Afterwards a CC design 22?star with 2 center points was Furthermore, the spore production in SBM medium and
employed with glucose and MgSO47H2O as study factors. A modified SBM was compared with a unifactorial design
total of 10 experiments were carried out in triplicate and the with three replicates per treatment. This experiment was
spore cell density and sporulation efficiency were deter- independently repeated twice.
mined 96 h after culture. The design was statistically ana-
lyzed with a second-order polynomial model presented in Effect of pH on B. subtilis EA-CB0575 spore
Eq. (2) through the Response Surface Methodology RSM. production
X
n X
n n X
X n
y ð x i Þ ¼ b0 þ bi x i þ bii x2i þ bij xi xj þ e ð2Þ The effect of pH on total cell density, spore cell density
i¼1 i¼1 i¼1 j¼1 and sporulation efficiency of B. subtilis EA-CB0575 was
where y(xi ) represents the response variable, xi is the evaluated using a single factorial design in a 14 L
independent variable (medium component), bo is the bioreactor. The levels evaluated were pH 6.5, 7.0, and
interception coefficient, bi the coefficient of the linear uncontrolled pH (initial pH = 5.5) with two replicates
effect, bii is the coefficient for the quadratic effect, bij per treatment and the response variables were determinated
corresponds to the coefficient for the interaction effect and at 48 h of fermentation. Batch fermentation of B. subtilis
e is the random error. EA-CB0575 was carried out in a 14 L bioreactor contain-
The responses predicted by the second-order models ing 8 L medium, at 30 °C, 300 rpm and 8 L/min of aera-
were used to generate a partial desirability function for tion rate.
each response (di ) that were defined as in Eq. (3).
8 Effect of agitation and aeration rate on B. subtilis
>
> 0; if yðxi Þ\Li
< yðx Þ  L s EA-CB0575 spore production
i i
diðyðxi ÞÞ ¼ ; if Li  yðxi Þ  Ui l ð3Þ
>
> Ui  Li
: The effect of agitation and aeration rates on B. subtilis EA-
1; if yðxi Þ  Ui
CB0575 spore cell density and sporulation efficiency were
where yðxi Þ is the predicted response using the fitted model evaluated using a CC design 22?star with two central points
(Eq. 2), Li and Ui are the highest and the lowest values in a 14 L bioreactor. A total of ten experiments were car-
obtained for the response i, respectively, and s is the weight ried out in duplicate and the spore cell density and

123
Bioprocess Biosyst Eng

sporulation efficiency were determined 48 h after culture. Results and discussion

The design was statistically analyzed with a second-order
polynomial model presented in Eq. (2) through the RSM. Selection of medium components for spore
Additionally, the responses predicted by the second-order production
models were used to generate a partial desirability function
(di ) for each response that were defined as in Eq. (3). With Among the fifteen media tested by the Plackett and Burman
the di functions defined, they were combined to obtain a design, only four (M1, M4, M11 and M15) produced spores
global desirability function (D) represented by Eq. (4). The after 96 h of evaluation (Table 1). Media M1, M4, M11,
weights used were equal to 1 for each of the two responses, and M15, all of which had low glucose content in common
spore cell density and sporulation efficiency. (2.0 g/L), had spore cell densities and sporulation effi-
ciencies between 0.51 9 109 and 1.87 9 109 CFU/mL,
Kinetics of cell growth and substrate uptake and 50.7 and 93.2 %, respectively. Medium M11 had the
highest spore cell density (1.87 9 109) and the highest
Kinetics of cell growth and substrate uptake of B. subtilis sporulation efficiency (93.2 %). The media with high
EA-CB0575 were determined in a 14 L bioreactor with 8 L quantities of glucose ([11 g/L) did not produce spores and
of modified SBM medium at the optimized conditions only two of the media with low quantities did not produce
(432 rpm, 12 L/min, and uncontrolled pH). One milliliter them either, such as medium M12 which lacked a source of
samples were taken every 2 h for the first 30 h and after- nitrogen. Media with and without MnCl2 produce spores,
wards every 6 h until 50 h for the determination of total but those that have this salt at its highest evaluation level
cell and spore cell densities (CFU/mL), sporulation effi- (0.5 g/L) have the highest sporulation efficiencies.
ciency (%), dry weight biomass (g/L), optical density (OD) For total cell density, a significant positive effect of
and sugar uptake (g/L). MgSO47H2O (X2) (P value = 0.041) and meat extract (X6)
(P value = 0.020) factors was seen (Table 1). These
Analytical methods results indicate that in order to increase biomass produced
by B. subtilis EA-CB0575, it is necessary to raise the levels
Dry weight biomass (g/L) and glucose concentration (g/L) of these two factors. Regarding the production of spores
in the culture medium were determined by the dry weight and sporulation efficiencies, glucose (X1) is the only factor
methodology and DNS methodology [24]. that had a significant negative effect on these variables
Total cells and spore cell densities were determined by (P value = 0.001 and 0.022, respectively). Therefore, the
the spread plate technique. Briefly, samples were serially concentration of glucose in the culture medium should be
diluted and 100 lL of each dilution was plated in TSA reduced to increase the spore cell density and the sporu-
for total cell density determination in CFU/mL. For spore lation efficiency. The negative effect of glucose on
cell density, serially diluted samples were subject to heat sporulation has also been determined by different studies
shock (80 °C, 20 min) and then plated in TSA. Response [12, 25] and it suggests that an excess of glucose inhibits
variables (total cell and spore cell density) were estimated sporulation by repressing the transcription of the spo0A
after 48 h of incubation at 30 °C for those dilutions gen. This gene is responsible of encoding Spo0A, a
containing 30–300 CFU/mL. Sporulation efficiency was response regulator activated by phosphorylation in
determined as the ratio between spore cell and total cell response to several internal and external stimuli and is the
densities. master regulator to entry into sporulation [8, 26]. The
importance of compounds such as KH2PO4, yeast extract,
Statistical analysis peptone and (NH4)2SO4 had been shown to significantly
stimulate sporulation [6, 27]. Although, in this study these
The statistical software Design ExpertÒ 8.0.7.1. (Stat-Ease, factors didńt affect sporulation possible because the mayor
Inc., Minneapolis, USA) was used to perform the experi- role of glucose on sporulation, which is sensitive to cata-
mental design, the regression and the graphical analysis of bolic repression [28].
data obtained. Analysis of variance (ANOVA) and LSD According to the previous results, the factors that were
multiple comparison test were carried out with the data selected for the next step of optimization, were glucose
obtained. The confidence level used for ANOVA analyses (range 0.2–1.2 g/L) and MgSO47H2O (range 0.5–1.5 g/L).
was 95 % and data of total cell and spore cell density Meat extract concentration (5.0 g/L), MnCl24H2O (0.5 g/
(CFU/mL) were transformed using a logarithmic scale L), KH2PO4 (6.0 g/L) and peptone (3.0 g/L) were left
(Log10). unchanged. Although, the meat extract had significant

123
 Bioprocess Biosyst Eng

effects on the total cell density, it was left unaltered at the model predicted a maximum spore cell density and
higher level because of its cost and because it had no effect sporulation efficiency of 1.0 9 109 CFU/mL and 133 %,
on sporulation. Furthermore, the yeast extract and respectively, in the medium containing (g/L): 2.0, glucose;
(NH4)2SO4 were eliminated from the culture medium and 0.5, MgSO47H2O. A sporulation efficiency of 133 %,
the concentrations of the previously established salts were which has no biological meaning, can be due to the low
left unchanged. determination coefficient of the model.
 
CFU
Medium optimization for B. subtilis EA-CB0575 Log ¼ 9:41 þ 0:086X1
mL
spore production
 0:89X2  0:11X1 X2 þ 0:46X22 ð5Þ
A full factorial design was used to find the optimization Sporulation efficiency ð%Þ ¼ 7:64 þ 133:996X1 þ 1:81X2
area for the two significant components, glucose (X1) and  43:46X12  17:58X1 X2
MgSO47H2O (X2). A total of 18 experiments were per-
ð6Þ
formed, obtaining spore cell densities between 0.74 9 109
and 1.58 9 109 CFU/mL, and sporulation efficiencies Different authors have selected factors that have a sig-
between 50.9 and 95.8 % (Table 2). The ANOVA analysis nificant effect by means of factorial designs or fractional
of the optimization study indicated that among the two factorials [10, 29] but go directly to an optimization design
significant variables selected by the Plackett and Burman or to a response surface method (RSM). However, in this
design experiment, MgSO47H2O (X2) was found to have a study, a full factorial design was used to define the opti-
significant effect on spore cell density (P value = 0.001) mization zone and, once in this zone, a CC design was used
and sporulation efficiency (P value = 0.022), contrary to to find the optimum point for each variable.
glucose concentration. Furthermore, the interaction In the CC design, the spore cell density and sporulation
between (MgSO47H2O) 9 (MgSO47H2O) (X2 9 X2) and efficiency were between 0.91 9 109 and 1.45 9 109 CFU/
glucose 9 glucose (X1 9 X1) were significant on the spore mL and 66.3–100.0 %, respectively (Table 3), which rep-
cell density (P value = 0.002) and sporulation efficiency resents an increase with respect to the full factorial design.
(P value = 0.032), respectively. Although, glucose did not The ANOVA analysis of the CC design indicated that
have a significant effect on the response variables, con- glucose (X1), MgSO47H2O (X2) and MgSO47H2-
centrations between 0.2 and 1.2 g/L were evaluated for the O 9 MgSO47H2O (X2 9 X2) had a significant effect on
next assay. spore cell density (P value = 0.010, 0.052, 0.041, respec-
The P values and the R2 for the model of spore cell tively); and glucose (X1), glucose 9 MgSO47H2O
density (P value = 0.025, R2 = 0.820, R2 adj = 0.748, (X1 9 X2), MgSO47H2O 9 MgSO47H2O (X2 9 X2) and
Eq. 5) and sporulation efficiency (P value = 0.022, glucose 9 glucose (X1 9 X1) had a significant effect on
R2 = 0.698, R2 adj = 0.557, Eq. 6) suggested that the sporulation efficiency (P value = 0.055, 0.045, 0.024 and
experimental data fit well with the model. These values 0.011, respectively). The P values for the model and the
indicated that less than 18.0 and 30.2 % of the total vari- lack of fit test of spore cell density (0.027, 0.343, Eq. 7)
ation was not explained by the model, respectively. The and sporulation efficiency (0.022, 0.133, Eq. 8), indicated

Table 2 Full factorial design for glucose and MgSO4 factors respect to spore cell density and sporulation efficiency of B. subtilis EA-CB0575
RUN Glucose (X1) MgSO47H2O Spore cell density (109 CFU/mL) Sporulation efficiency (%)
(g/L) (X2) (g/L)
Observed Predicted Observed Predicted

1 0.8 0.5 1.30 ± 0.05a 1.26 85.1 ± 5.8 80.9

2 1.4 0.5 1.16 ± 0.04 1.31 72.0 ± 2.3 98.7
3 2 0.5 1.58 ± 0.06 1.36 95.8 ± 4.4 85.2
4 0.8 1 0.88 ± 0.11 0.89 72.5 ± 8.7 74.8
5 1.4 1 0.96 ± 0.09 0.86 88.8 ± 0.5 87.3
6 2 1 0.74 ± 0.13 0.82 64.9 ± 1.8 68.5
7 0.8 1.5 0.99 ± 0.03 1.06 66.7 ± 1.3 68.6
8 1.4 1.5 1.06 ± 0.03 0.96 80.0 ± 8.6 75.9
9 2 1.5 0.83 ± 0.02 0.85 50.9 ± 0.4 51.8
a
Intervals represent standard errors of the mean

123
Bioprocess Biosyst Eng

Table 3 Central composite design for glucose and MgSO4 factors respect to spore cell density and sporulation efficiency of B. subtilis EA-
CB0575
RUN Glucose (X1) (g/L) MgSO47H2O (X2) (g/L) Spore cell density (109 CFU/mL) Sporulation efficiency (%)
Observed Predicted Observed Predicted

1 0.7 0.7 1.15 ± 0.04a 1.30 100.0 ± 7.4 99.4

2 0.2 0.2 0.91 ± 0.04 1.03 66.3 ± 1.7 70.2
3 1.2 0.2 1.10 ± 0.10 1.35 96.4 ± 0.8 97.2
4 0.2 1.2 1.02 ± 0.05 1.04 89.5 ± 10.2 87.1
5 1.2 1.2 1.35 ± 0.02 1.2 76.6 ± 6.8 81.2
6 0 0.7 1.15 ± 0.04 1.05 84.6 ± 3.5 78.6
7 1.41 0.7 1.45 ± 0.05 1.38 95.1 ± 2.7 93.0
8 0.7 0 0.97 ± 0.05 1.15 83.5 ± 5.7 82.3
9 0.7 1.41 1.07 ± 0.13 1.05 89.3 ± 1.4 83.4
10 0.7 0.7 1.20 ± 0.08 1.30 98.7 ± 4.1 99.4
a
Intervals represent standard errors of the mean

that the model fit appropriately with the experimental data. a concentration of 1.52 9 1010 CFU/mL for the B. subtilis
The second-order polynomial equations (Eqs. 7, 8), WHK-Z12 strain, but to our knowledge, the sporulation
obtained from the ANOVA analysis, showed that the value efficiency reached in this work (93.5 ± 4.0 %) is the
of R2 was 0.917 (R2 adj = 0.815) for spore cell density and highest reported to date for B. subtilis strains that are in the
0.927 (R2 adj = 0.835) for sporulation efficiency, indicat- range of 30–80 % [11, 12, 14].
ing that less than 8.3 and 7.3 % of the total variation was
not explained by the model, respectively.
  Model verification and evaluation of modified
CFU
Log ¼ 8:94 þ 0:23X1 þ 0:22X2  0:07X12 medium
mL
 0:06X1 X2  0:15X22 ð7Þ
In order to verify the optimization results, an experiment
Sporulation efficiency ð%Þ ¼ 45:31 þ 71:95X1 þ 71:56X2 was performed with the optimized levels of nutrients pre-
 27:38X12  33:0X1 X2 dicted by the model. Spore cell density and sporulation
efficiencies of (1.37 ± 0.15) 9 109 CFU/mL and
 34:32X22 ð8Þ
93.5 ± 4.0 %, respectively, were obtained, suggesting that
Multiple response regression was evaluated using spore experimental and predicted values were in good agreement
cells density and sporulation efficiency responses and with percentage errors of 6.8 and 2.8 %, respectively,
glucose and MgSO47H2O were used as factors. Desir- validating the model. Additionally, the medium optimiza-
ability was the dependent variable for this regression and tion resulted in a 2.7- and 1.9-fold increase in spore cell
the optimum value for both responses was determined. density and sporulation efficiency, respectively, as com-
Desirability measures the overlap for the optimum value of pared to that of the Plackett and Burman treatment that
the response variables. In this case, the desirability value showed lower spore cell density (5.10 9 108 CFU/mL). In
for multiple response regression was 0.875 with 1.2 g/L of this optimization process glucose concentration was also
glucose and 0.66 g/L of MgSO47H2O, which would reach reduced form 2.0 g/L used in the Plackett and Burman
1.33 9 109 CFU/mL of spore cell density and 97.7 % of design to 1.2 g/L used in the SBM medium, representing
sporulation efficiency. Supplementary material Figure S1 savings of 40 %.
presents the multivariable optimization and the stated The effect of changing the nutrient meat extract by a
values of those compounds that make it possible to get the nutrient of lower cost (yeast extract) in the optimized SBM
highest desirable function. medium on the production of spores was also determined.
The optimization of the medium by the CC design Spore cell density and sporulation efficiency had no sig-
resulted in a 1.5-fold increase in spore cell density com- nificant differences between the SBM and modified SBM
pared to that of the full factorial design prediction medium (P value = 0.232 and 0.249, respectively), sug-
(1.0 9 109 CFU/mL). These values do not surpass the gesting that spore production is not affected by changing
spore cell density reached by Chen et al. [9], who reported the complex nitrogen source.

123
 Bioprocess Biosyst Eng

Influence of pH, agitation and aeration rate on B. agitation rate (X1), agitation 9 agitation (X1 9 X1) and
subtilis EA-CB0575 spore production agitation 9 aeration (X1 9 X2) had a significant effect on
spore cell density (P value = 0.053, 0.021, 0.024, respec-
To determine the effect of pH on B. subtilis EA-CB00575 tively), while none of the two variables had significant
spore production at bioreactor level a single factorial effects on sporulation efficiency. The P values for the
design was conducted. The ANOVA analysis indicated that model and the lack of fit test of spore cell density (0.027,
maintaining the pH at a constant value (6.5 or 7.0) or under 0.339, Eq. 9) indicated that the model fit appropriately with
noncontrolled pH bioreaction did not affect significantly the experimental data. The second-order polynomial
total cell density (P value = 0.209), spore cell density equation (Eq. 9), obtained from the ANOVA analysis,
(P value = 0.285) and sporulation efficiency showed that the value of R2 was 0.900 (R2 adj = 0.754),
(P value = 0.895). Under noncontrolled conditions, pH indicating that less than 10 % of the total variation was not
variation was between 5.5 and 7.0 during 48 h of fer- explained by the model.
mentation, and spore cell densities and sporulation effi-  
ciencies of (2.27 ± 0.56) 9 109 CFU/mL and 92.9 ± CFU
Log ¼ 2:21 þ 0:02X1 þ 4:1X2  2:5  105 X12
5.1 % were achieved. Contrary to our results, Monteiro mL
et al. [12], reported that maintaining the pH at a constant  2:9  105 X1 X2  1:32X22 ð9Þ
value during the fermentation increases spore production,
although in the ranges of 6–9 the sporulation efficiency did CC design showed higher total cell and spore cell den-
not depend on the pH value (11). Therefore our results sities (CFU/mL) around the central points of agitation and
could indicate that the low variation on pH during the aeration ranges (400 rpm and 12 L/min). The highest value
fermentation process did not affect the sporulation for spore cell density was 9.33 9 109 CFU/mL, reporting a
achieved. sporulation efficiency of 91.2 % at 400 rpm and 12 L/min
To determine the effect of the agitation and aeration rate of air (Table 4). The response surface curve for simple and
on B. subtilis EA-CB0575 spore production under non- multivariate optimization shown in Supplementary mate-
controlled pH bioreaction, operating conditions between rial Figure S2 shows an optimum value for both response
300 and 500 rpm of agitation rate and 8–16 L/min of variables at 432 rpm and 12 L/min, with a total cell and
aeration rate were evaluated. In the CC design, the spore spore cell densities prediction of 1.02 9 1010 CFU/mL and
cell density and sporulation efficiency were between 9.59 9 109 CFU/mL, respectively, and 94.0 % of spore
1.24 9 109 and 9.33 9 109 CFU/mL and 78.9–96.2 %, efficiency. These results are in accordance with several
respectively (Table 4), which represents a 6.8 increase in reports were an increase on agitation and aeration rates
spore cell density with respect to the shake flask conditions. generate higher biomass and metabolites production on
The ANOVA analysis of the CC design indicated that different Bacillus species [22, 30–32].

Table 4 Central composite design for agitation and aeration factors respect to spore cell density and sporulation efficiency of B. subtilis EA-
CB0575 at bioreactor level
RUN Agitation (X1) (rpm) Aeration (X2) L/min (vvm) Spore cell density (109 CFU/mL) Sporulation efficiency (%)
Observed Predicted Observed
a
1 400 12 (1.5) 9.33 ± 0.0 8.81 91.2 ± 1.1
2 300 8 (1.0) 1.56 ± 0.2 1.59 96.2 ± 3.8
3 500 8 (1.0) 5.20 ± 1.0 3.55 83.9 ± 4.8
4 300 16 (2.0) 1.24 ± 0.0 1.56 91.9 ± 7.8
5 500 16 (2.0) 3.60 ± 0.7 3.18 81.8 ± 9.0
6 259 12 (1.5) 2.14 ± 0.3 1.70 79.8 ± 6.0
7 541 12 (1.5) 3.30 ± 0.3 4.70 97.1 ± 1.4
8 400 6.5 (0.8) 1.53 ± 0.5 1.96 78.9 ± 1.3
9 400 18 (2.2) 2.17 ± 0.2 1.89 95.2 ± 2.6
10 400 12 (1.5) 8.72 ± 0.1 8.81 95.5 ± 4.3
a
Intervals represent standard errors of the mean

123
Bioprocess Biosyst Eng

Fig. 1 Growth, substrate and

oxygen uptake of B. subtilis EA-
CB0575 in optimized conditions
at bioreactor level

Kinetics of cell growth and substrate uptake of exponential phase (10 h), which can represent savings of
B. subtilis EA-CB0575 culture medium.

The kinetics of cell growth and substrate uptake of

B. subtilis EA-CB0575 in modified SBM medium were
determined at the optimal conditions in a batch culture at
Conclusions
bioreactor level (Fig. 1). Biomass concentration (g/L)
High cell density and sporulation efficiency are important
increased during 10 h in the exponential phase reaching
issues for the profitability of the production of Bacillus
3.5 g/L, consuming 96.9 % of reducing sugars and 92.0 %
used as bioproducts. In literature, spore cell densities and
of dissolve oxygen. After 10 h of growth, cells entered a
sporulation efficiencies have been reported up to
stationary phase and after 40 h of culture total biomass
1.52 9 1010 CFU/mL (Chen et al. [9]) and 80 % (Monteiro
decline and dissolve oxygen reached 91.6 % suggesting the
et al. [12]), respectively, although obtaining high values of
formation of spores [12, 15].
both variables is poorly reported. In the present study, the
Additionally in order to verify the optimization results at
improvement of both spore cell density and sporulation
bioreactor level, spore cell density and sporulation effi-
efficiency of B. subtilis EA-CB0575 was achieved by the
ciencies were determined at 50 h of culture obtaining
optimization of medium components and culture condi-
8.78 9 109 CFU/mL and 94.2 %, respectively. These
tions. In general, spore production of B. subtilis EA-
results suggest that experimental and predicted values were
CB0575 was affected by the medium components glucose
in good agreement with percentage errors of 7.9 and 1.1 %,
and MgSO47H2O and by the culture conditions agitation
respectively, validating the model. The optimization of the
and aeration rate. The culture conditions stablished in this
culture conditions resulted in a 6.4-fold increase in spore
study, permitted to reach a high final spore cell density of
cell density compared to that obtained at shake flask
8.78 9 109 CFU/mL and a high sporulation efficiency of
(1.37 9 109 CFU/mL) and the overall optimization gen-
94.2 % at bioreactor level, being the sporulation efficiency
erated a 17.2-fold increase in spore cell density and 1.9-
the highest value reported to date for B. subtilis in batch
fold increase in sporulation efficiency compared to that of
culture. Finally, with this culture approach the quantity of
PB treatment that showed the lower values (5.10 9 108
nutrients needed for the fermentation were reduced, the
CFU/mL and 50.7 %). Our results did not surpass the spore
costs of the reagents used for pH control were avoid, and
cell density (1.56 9 1010 spores/mL) reported previously
the high spore production was obtained in a short incuba-
for B. subtilis WHK-Z12 in a 30 L bioreactor with 16.18 g/
tion period, which could represent savings on bioprocess
L of corn steep liquor, 17.53 g/L of soybean flour and
costs.
8.14 g/L of yeast extract, although we achieved a high
spore cell density and sporulation efficiency with very low Acknowledgments This research was made possible by funding
carbon and nitrogen requirements (1.04 g/L glucose and from the Universidad EAFIT, the Association of Banana Producers of
5 g/L of yeast extract), with no lag phase and a short Colombia (Asociación de Bananeros de Colombia, AUGURA), the

123
 Bioprocess Biosyst Eng

Administrative Department of Science, Technology and Innovation of 15. Yezza A, Tyagi RD, Valero JR, Surampalli RY (2006) Biocon-
Colombia (Departamento Administrativo de Ciencia, Tecnologı́a e version of industrial wastewater and wastewater sludge into
Innovación de Colombia-COLCIENCIAS, Contract number Bacillus thuringiensis based biopesticides in pilot fermentor.
0130-2012); and by the Subscribe Contract Number REG00116 with Bioresour Technol 97:1850–1857
the Ministry (Ministerio de Medio Ambiente y Desarrollo Territoral) 16. Liu BL, Tzeng YM (1998) Optimization of growth medium for
in the category ‘‘Contrato de Acceso a Recursos Genéticos para la the production of spores from Bacillus thuringiensis using
Investigación Cientı́fica sin Interés Comercial’’. response surface methodology. Bioprocess Eng 18:413–418
17. Flores ER, Pérez F, De la Torre M (1997) Scale-up of Bacillus
thuringiensis fermentation based on oxygen transfer. J Ferment
Bioeng 83:561–564
18. Gu XB, Zheng ZM, Yu HQ, Wang J, Liang FL, Liu RL (2005)
References Optimization of medium constituents for a novel lipopeptide
production by Bacillus subtilis MO-01 by a response surface
1. Adams TT, Eiteman MA, Adang MJ (1999) Bacillus method. Process Biochem 40:3196–3201
thuringiensis subsp. kurstaki spore production in batch culture 19. Buhr TL, McPherson DC, Gutting BW (2008) Analysis of broth-
using broiler litter extracts as complex media. Bioresour Technol cultured Bacillus atrophaeus and Bacillus cereus spores. J Appl
67:83–87 Microbiol 105:1604–1613
2. Omer AM (2010) Bioformulations of Bacillus Spores for using as 20. Farrera RR, Perez-Guevara F, de la Torre M (1998) Carbon :
Biofertilizer. Life Sci J Acta Zhengzhou Univ Overseas Ed nitrogen ratio interacts with initial concentration of total solids on
7:124–131 insecticidal crystal protein and spore production in Bacillus
3. Tzeng YM, Rao YK, Tsay KJ, Wu WS (2008) Effect of culti- thuringiensis HD-73. Appl Microbiol Biotechnol 49:758–765
vation conditions on spore production from Bacillus amyloliq- 21. Liu S, Fang Y, Lv M, Wang S, Chen L (2010) Optimization of
uefaciens B128 and its antagonism to Botrytis elliptica. J Appl the production of organic solvent-stable protease by Bacillus
Microbiol 104:1275–1282 sphaericus DS11 with response surface methodology. Bioresour
4. Soe KT, De Costa DM (2012) Development of a spore-based Technol 101:7924–7929
formulation of microbial pesticides for control of rice sheath 22. Saran S, Isar J, Saxena R (2007) Statistical optimization of
blight. Biocontrol Sci Tech 22:633–657 conditions for protease production from Bacillus sp. and its scale-
5. Ongena M, Jacques P (2008) Bacillus lipopeptides: versatile up in a bioreactor. Appl Biochem Biotechnol 141:229–239
weapons for plant disease biocontrol. Trends Microbiol 23. Angelopoulos P, Evangelaras H, Koukouvinos C (2009) Small,
16:115–125 balanced, efficient and near rotatable central composite designs.
6. Rao YK, Tsay K-J, Wu W-S, Tzeng Y-M (2007) Medium opti- J Stat Plan Inference 139:2010–2013
mization of carbon and nitrogen sources for the production of 24. Miller GL (1959) Use of dinitrosalicylic acid reagent for deter-
spores from Bacillus amyloliquefaciens B128 using response mination of reducing sugar. Anal Chem 31:426–428
surface methodology. Process Biochem 42:535–541 25. Chibanzakura T, Kawamura F, Takahashi H (1991) Differential
7. Schleifer KH (2001) Phylum XIII. Firmicutes Gibbons and regulation of spo0A transcription in Bacillus subtilis: glucose
Murray 1978, 5 (Firmicutes [sic] Gibbons and Murray 1978,5). repress promoter switching at initiation of sporulation. J Bacteriol
In: Garrity GM (ed) Bergeys. Springer, Berlin 173:2625–2634
8. Sonenshein AL (2000) Control of sporulation initiation in 26. Grimshaw CE, Huang S, Hanstein CG, Strauch MA, Burbulys D,
Bacillus subtilis. Curr Opin Microbiol 3:561–566 Wang L, Hoch JA, Whiteley JM (1998) Synergistic cinetic
9. Chen ZM, Li Q, Liu HM, Yu N, Xie TJ, Yang MY, Shen P, Chen interactions between components of the phosphorelay controlling
XD (2010) Greater enhancement of Bacillus subtilis spore yields sporulation in Bacillus subtilis. Biochemistry 37:1365–1375
in submerged cultures by optimization of medium composition 27. Shi F, Zhu Y (2007) Application of statistically-based experi-
through statistical experimental designs. Appl Microbiol mental designs in medium optimization for spore production of
Biotechnol 85:1353–1360 Bacillus subtilis from distillery effluent. Biocontrol 52:845–853
10. Cho JH, Kim YB, Kim EK (2009) Optimization of culture media 28. Chevanet C, Besson F, Michel G (1986) Effect of various growth
for Bacillus species by statistical experimental design methods. conditions on spore formation and bacillomycin L production in
Korean J Chem Eng 26:754–759 Bacillus subtilis. Can J Microbiol 32:254–258
11. Hageman J, Shankweiler G, Wall P, Franich K, McCowan G, 29. Jacques P, Hbid C, Destain J, Razafindralambo H, Paquot M, De
Cauble S, Grajeda J, Quinones G (1984) Single, chemically Pawn E, Thonart P (1999) Optimization of biosurfactant
defined sporulation medium for Bacillus subtilis: growth, sporu- lipopeptide production from Bacillus subtilis S499 by Plackett-
lation and extracellular protease production. J Bacteriol Burman design. Appl Biochem Biotechnol 77–79:223–233
160:438–441 30. Yeh M-S, Wei Y-H, Chang J-S (2006) Bioreactor design for
12. Monteiro SM, Clemente JJ, Henriques AO, Gomes RJ, Carrondo enhanced carrier-assisted surfactin production with Bacillus
MJ, Cunha AE (2005) A procedure for high-yield spore pro- subtilis. Process Biochem 41:1799–1805
duction by Bacillus subtilis. Biotechnol Prog 21:1026–1031 31. Potumarthi R, Ch S, Jetty A (2007) Alkaline protease production
13. Donnellan JE, Nags EH, Levinson HS (1964) Chemically defined by submerged fermentation in stirred tank reactor using Bacillus
synthetic media for sporulatgion and for germination and growth licheniformis NCIM-2042: effect of aeration and agitation
of Bacillus subtilis. J Bacteriol 87:332–336 regimes. Biochem Eng J 34:185–192
14. Tavares M, Souza R, Luiz W, Cavalcante RM, Casaroli C, 32. Feng Y, He Z, Ong SL, Hu J, Zhang Z, Ng WJ (2003) Opti-
Martins E, Ferreira RC, Ferreira LS (2013) Bacillus subtilis mization of agitation, aeration, and temperature conditions for
endospores at high purity and recovery yields: optimization of maximum b-mannanase production. Enzyme Microbial Technol
growth conditions and purification method. Curr Microbiol 32:282–289
66:279–285

123