Optimization studies for enhanced bacteriocin production by Pediococcus pentosaceus KC692718 using response surface methodology

Abstract

Bacteriocins have been produced by various Lactic acid bacteria (LAB) strains isolated from dairy and fermented vegetable sources. In the current study we have isolated a novel bacteriocin producing strain Pediococcus pentosaceus KC692718 from mixed vegetable pickles (India). A 2 step process optimization for enhancing production of bacteriocin from the isolates was carried out with One-factor-at-a-time (OFAT) and Response Surface Methodology (RSM) methods. A 2.5 fold (AU/mL) increase of bacteriocin was observed for sucrose (2.4 %) as carbon source and 4.7 fold (AU/mL) increased bacteriocin was observed in the presence of soyatone (1.03 %) as nitrogen source in the OFAT experiments. In order to increase bacteriocin production RSM tool was performed with optimized chemical and physical sources using Design expert 8.0.7.1. Soyatone (1.03 %), sucrose (2.4 %), pH (5.5) and temperature (34.5 ºC) condition yielded 25,600.34 AU/mL of bacteriocin from P. pentosaceus KC692718. This is the first report which has produced 20 fold increase of bacteriocin for Pediococcus pentosaceus KC692718 from that of MRS medium with 1 280 AU/mL

Introduction

Bacteriocins are defined as ribosomally synthesized antimicrobial peptides or proteins, produced by bacteria, with antagonistic activity against bacteria genetically closely related to the producer strain (De Vuyst and Vandamme 1992). However, an exception to the rule i.e. activity against a range of Gram-negative bacteria has been reported (Jamuna and Jeevaratnam 2004).Bacteriocins produced by Lactic Acid Bacteria have presented a potential use either in food industries as biopreservatives or in the prevention and treatment of some infectious diseases, having medical and veterinary applications. In recent years, greater interest is being drawn towards application of bacteriocins as food biopreservative because they are innocuous due to their proteolytic degradation in gastrointestinal tract (Dominguez et al. 2007). The pediocin-like bacteriocins fall into class IIa and are defined as small heat-stable Listeria-active peptides (Klaenhammer 1993; Horn et al. 1998).

The most important aspects in the bacteriocin study are production and purification. Most often, production of bacteriocin is very low and the complex growth media commonly used for bacteriocin production interfere in its purification (Mackay et al. 1997). The preservation may be achieved either by using a bacteriocin-producing starter culture or by applying the bacteriocin itself as a food additive. Both the methods necessarily require optimization of production that can be dependent on multiple factors, and are usually strain specific (Carolissen-Mackay et al. 1997). Moreover all the ingredients present in the complex medium are not used by the organism for the production of bacteriocins and for antagonistic activity against food borne pathogens especially Listeria monocytogenes. So, bacteriocin production requires optimized process, complex media and well controlled physical condition including pH and temperature (Daba et al. 1993; Leroy and De Vuyst 1999;Mukesh et al. 2012). On the other hand, optimization of a medium by the classical method involves changing one independent variable at a time while keeping others at a fixed level. This is extremely time-consuming and expensive for many variables (Adinarayana et al. 2003) and can also result in unacceptable conclusions (Oh et al. 1995).

The optimization of bacteriocin production and enhancement of its activity are economically important to reduce the production cost (Mahrous et al. 2013). The response surface method (RSM) is a statistical method that involves individual and interaction effects to account for curvature, to improve optimal process settings, and to troubleshoot process problems and weak points (Oh et al. 1995) and to build models, evaluating the effects of factors for desirable response(Li et al. 2001; Puri et al. 2001). It has been successfully used in many areas of biotechnology, including some recent studies on bacteriocin production (Cassia and Adriano 2010). Hence, statistically based experimental designs are preferred to evaluate the influence of medium components in batch fermentations for the production of industrially important enzymes and proteins (Karthikeyan and Jayaraman 2011).

The main aim of this study is to simplify commonly used complex MRS medium, for bacteriocin production by Lactic acid bacteria (LAB) with two step optimization process, namely One-factor-at-a-time method (OFAT) and Response surface methodology (RSM). This study also envisages the use of cheap sources for production of bacteriocin by Pediococcus pentosaceus KC692718 under optimized condition.

Materials and methods

Bacterial strains

Pediococcus pentosaceus KC692718 the bacteriocin producer strain was isolated from mixed vegetable pickle. The strain was identified by sequencing the 1 344 bases of 16S rRNA gene followed by BLAST homology search. The nucleotide sequences have been deposited with NCBI database under accession number KC692718.

Indicator strain and culture media

The indicator strain Listeria monocytogenes MTCC657 was cultured in Tryptic soy broth with yeast extract (TSB, Difco) at 37 ºC. Strain were stored at–20 ºC in TSB medium with 20 % (v/v) glycerol.

Bacteriocin activity assay

Bacteriocin activity was determined using the agar-well diffusion method. Arbitrary units (AU) of bacteriocin activity were calculated according to Yamamoto et al. (2003). Pediococcus pentosaceus KC692718 were grown in MRS broth at 37 ºC for 24 h. Culture broth was centrifuged at 10,000 rpm for 20 min; at 4 ºC. Cell free supernatant was serially diluted two fold with phosphate buffer (10 mM, pH 6.5). Solidified TSB with YE agar plates were swabbed with the log phase culture of Listeria monocytogenes. After 30 min of drying, wells were made and 10 μl of each cell free supernatant sample was added. The plates were refrigerated for 4 h to allow the radial diffusion of the compound and later incubated at 24 h at 37 ºC. Thus, bacteriocin activity was expressed as the reciprocal of the highest dilution (2 N) showing a distinct zone of inhibition to obtain the arbitrary units per milliliter (AU/mL) was calculated as.

$\raisebox{1ex}{$2^{\mathrm{N}}$}\!\left/ \!\raisebox{-1ex}{$10\mathrm{\mu l}$}\right.\mathrm{X}1000$ μl, where N = dilution number with smallest zone of inhibition.

Primary optimization stage: One-factor-at-a-time method study

Various cheap carbon and nitrogen sources were selected for one factor-at-a-time method (OFAT) study using Pediococcus pentosaceus KC692718 for bacteriocin production. To test the effect of carbon sources, each carbon source (molasses were purchased from Thiru Arooran Sugars Ltd, Virudhachalam, TN, India, fructose, malt extract, molasses, sucrose and galactose purchased from Himedia, Mumbai) were added at 2 % (w/v) level to MRS medium, by replacing 2 % glucose and the influence of nitrogen sources in a modified MRS medium was used by replacing protease peptone with each nitrogen source such as (black gram husk purchased from local market, Vellore, TN, India, soybean meal, enzymatic digest of casein, soyatone and urea procured from Himedia, Mumbai) at 2 % level. Cultivation was done for 24 h of incubation at 37 ºC with pH 6.5. Antimicrobial activity against Listeria monocytogenes MTCC657 was also examined using agar-well diffusion, as described previously. Cell free supernatant was serially diluted two fold. Sterile phosphate buffer (10 mM, pH 6.5) was used as diluents and each diluted solution of 10 μl was loaded on Solidified TSB with YE agar lawn containing Listeria monocytogenes MTCC657. The bacteriocin activity was expressed as AU/mL.

Experimental design of RSM

Bacteriocin production was optimized using one-factor-at-a time approach. The primary optimization experiments revealed that the optimization yield was observed in the presence of 2 % sucrose as carbon source and 2 % soyatone as nitrogen source, with the above conditions 2 more variables was chosen namely physical parameters pH and temperature.

Central composite design (CCD)

RSM was carried out using CCD design, optimize for further process to identify the interactions between the significant factors obtained from OFAT. The 4 variables chosen in this experiment were (C source) sucrose, (N source) soyatone, pH and temperature with 5 coded levels (−α,-1, 0, +1, + α) were used for their combined influence on bacteriocin production. 30 experimental trials were carried out with 16 factorial points, 8 axial points with α = 2 and 6 replication of central points.

In developing regression equation, the test factors were coded according to the Eq. (1)

$${\mathrm{X}}_{\mathit{i}}=\left({\mathrm{X}}_{\mathit{i}}\right.-\left.{\mathrm{X}}_0\right)/{\mathrm{\delta X}}_{\mathit{i}}$$ 1

Where x_i is the dimensionless coded value of the i^th independent variable; X_i the natural value of the i^th independent variable; X₀ the natural value of the i^th independent variable at the centre point and δX_i is the difference in effect.

The response data obtained from the design were fitted with a second order polynomial. The general polynomial equation is as follows in Eq. (2)

$$\mathrm{Y}={\mathit{\beta }}_0+{\mathit{\beta }}_1{\mathrm{X}}_1+{\mathit{\beta }}_2{\mathrm{X}}_2+{\mathit{\beta }}_3{\mathrm{X}}_3+{\mathit{\beta }}_4{\mathrm{X}}_4+{\mathit{\beta }}_{11}{\mathrm{x}}_{12}{+}_{\mathit{\beta }22}{\mathrm{x}}_{2^2+\mathit{\beta }33}{\mathrm{x}}_{3^2+\mathit{\beta }44}{\mathrm{x}}_{4^2+\mathit{\beta }12}{\mathrm{x}}_1{\mathrm{x}}_{2+\mathit{\beta }13}{\mathrm{x}}_1{\mathrm{x}}_{3+\mathit{\beta }14}{\mathrm{x}}_1{\mathrm{x}}_{4+\mathit{\beta }23}{\mathrm{x}}_2{\mathrm{x}}_{3+\mathit{\beta }24}{\mathrm{x}}_2{\mathrm{x}}_{4+}{\mathit{\beta }}_{34}{\mathrm{x}}_3{\mathrm{x}}_4$$ 2

Where, Y is the predicted response, β₀ is the model constant, β₁, β₂, β₃, β₄ the linear co-efficients, β₁₁, β₂₂, β₃₃, β₄₄ the squared co-efficients and β₁₂, β₁₃, β₁₄, β₂₃, β₂₄, β₃₄ the interaction co-efficients.

Data analysis

Design expert 8.0.7.1(Stat-Ease, Inc., Minneapolis, USA) was used for the regression analysis of the experimental data obtained. The quality of fit of the polynomial model equation was expressed by the coefficient of determination R² and analysis of variance (ANOVA). The significance of the model, an optimum value of parameters was assessed by the determination coefficient, correlation coefficient and statistical testing of the model was made by Fisher’s test (Myers and Montgomery 2002).

Results and discussion

Primary optimization stage: One-factor-at-a-time method study

The maximum bacteriocin activity was determined as 3 200 AU/mL, 1 600 AU/mL for sucrose and fructose respectively with control in MRS medium observed to be 1 280 AU/mL. It was evident that sucrose was found to be the best carbon source for the enhanced bacteriocin activity when compared to fructose and other sugars tested (Fig. 1). Similar results were reported by (Biswas et al. 1991; Ray 1995) where sucrose has been recommended as a suitable carbon source for pediocin AcH production on the liquid media even with terminal pH below 4.0. When compared to the other nitrogen sources soyatone enhanced the bacteriocin activity up to 6 400 AU/mL given in Fig. 2. Ours is the first study on enhanced bacteriocin production using soyatone as simple source in modified MRS media. Hence sucrose and soyatone was selected as the potential carbon and nitrogen source for production of bacteriocin from Pediococcus pentosaceus KC692718.

Fig. 1.

Effect of different C sources on bacteriocin production by P.pentosaceus KC692718 (n = 5)

Fig. 2.

Effect of different N sources on bacteriocin production by P. pentosaceus KC692718 (n = 5)

Experimental design of RSM: central composite design (CCD)

Based on the result of OFAT approach optimum levels of significant factors and the effect of their interactions on bacteriocin production was determined by CCD experiments. The following variables sucrose, soyatone was selected based on result of the OFAT method further pH and temperature were added for the optimization by RSM. The experimental design was carried out to determine the parameter ranges together with coded and actual values of the 4 independent variables for bacteriocin production, is presented in Table 1.

Table 1.

Range of value for response surface method

Factors	Independent variables	Coded levels
		-α	–1	0	+α	+1
A	Soyatone %	0	0.5	1	1.5	2.0
B	Sucrose %	0	1	2	3	4
C	pH	3.5	4.5	5.5	6.5	7.5
D	Temperature (degree)	29.5	32	34.5	37	39.5

The results of 30 run from CCD experiments for studying the effects of 4 independent variables on bacteriocin production are represented in Table 2.

Table 2.

Coded experimental design and results for the response surface of maximum bacteriocin production of P. pentosaceus KC692718 as a function of Soyatone, Sucrose, pH, Temperature

Runs	Coded Values
Number	Order	A	B	C	D	Bacteriocin Activity AU/mL
27	1	1	2	5.5	34.5	25600.34
20	2	1	4	5.5	34.5	6400.19
23	3	1	2	5.5	29.5	200.05
2	4	1.5	1	4.5	32	800.05
22	5	1	2	7.5	34.5	1600.24
7	6	0.5	3	6.5	32	3200.18
4	7	1.5	3	4.5	32	1600.12
3	8	0.5	3	4.5	32	800.06
12	9	1.5	3	4.5	37	3200.10
5	10	0.5	1	6.5	32	800.08
29	11	1	2	5.5	34.5	25600.34
26	12	1	2	5.5	34.5	25600.34
8	13	1.5	3	6.5	32	3200.05
19	14	1	0	5.5	34.5	200.02
14	15	1.5	1	6.5	37	1600.18
17	16	0	2	5.5	34.5	200.06
15	17	0.5	3	6.5	37	1600.32
30	18	1	2	5.5	34.5	25600.34
24	19	1	2	5.5	39.5	200.01
1	20	0.5	1	4.5	32	1600.10
25	21	1	2	5.5	34.5	25600.34
10	22	1.5	1	4.5	37	800.09
6	23	1.5	1	6.5	32	800.03
11	24	0.5	3	4.5	37	1600.23
13	25	0.5	1	6.5	37	800.12
21	26	1	2	3.5	34.5	400.16
16	27	1.5	3	6.5	37	6400.01
18	28	2	2	5.5	34.5	1600.12
28	29	1	2	5.5	34.5	25600.34
9	30	0.5	1	4.5	37	400.05

The maximum experimental value for bacteriocin production was 25600.34 AU/mL based on RSM. The regression analysis data were fitted to a quadratic model and the second order regression equation obtained was full actual model on bacteriocin production is shown in Eq. (3)

$$\mathrm{Y}=+25600.34+{433.32}^{*}\mathrm{A}+{1100.03}^{*}\mathrm{B}+{416.68}^{*}\mathrm{C}+{150.01}^{*}\mathrm{D}-{6095.90}^{*}{\mathrm{A}}^2-{5495.89}^{*}{\mathrm{B}}^2-{6070.87}^{*}{\mathrm{C}}^2-{6070.91}^{*}{\mathrm{D}}^2+{424.97}^{*}{\mathrm{A}}^{*}\mathrm{B}+{224.98}^{*}{\mathrm{A}}^{*}\mathrm{C}+{474.99}^{*}{\mathrm{A}}^{*}\mathrm{D}+{425.00}^{*}{\mathrm{B}}^{*}\mathrm{C}+{275.00}^{*}{\mathrm{B}}^{*}\mathrm{D}+{75.01}^{*}{\mathrm{C}}^{*}\mathrm{D}$$ 3

Where Y is bacteriocin activity AU/mL, A is soyatone (g/L), B is sucrose (g/L), C is pH and D is temperature (ºC). The statistical significance of Eq.3 was checked by an F-test; and the analysis of variance (ANOVA) for response surface quadratic model is given in Table 3. ANOVA for bacteriocin production from the Table 3 that the F value of 398.27 implies model is highly significant, it is evident from the model F value and a low probability value at P > F value was <0.0001. Thus 0.01 % chance that a "Model F-Value" this may occur large owing to noise. The goodness of the model can be made by the detremination coeffecient (R²) and the correlation coeffecient (R) (Sunitha et al. 1999; Osorio et al. 2001).

Table 3.

Analysis of variance (ANOVA) for the quadratic modal of bacteriocin production obtained from the experimental results

Source	Sum of squares	df	Mean squares	F value	Probability > F
Model	2.810E + 009	14	2.007E + 008	98.27	< 0.0001*
A-Soyatone	4.506E + 006	1	4.506E + 006	8.94	0.0092*
B-Sucrose	2.904E + 007	1	2.904E + 007	57.62	< 0.0001*
C-pH	4.167E + 006	1	4.167E + 006	8.27	0.0116*
D-Temperature	5.401E + 005	1	5.401E + 005	1.07	0.3170
AB	2.890E + 006	1	2.890E + 006	5.73	0.0302*
AC	8.098E + 005	1	8.098E + 005	1.61	0.2243
AD	3.610E + 006	1	3.610E + 006	7.16	0.0173*
BC	2.890E + 006	1	2.890E + 006	5.73	0.0301*
BD	1.210E + 006	1	1.210E + 006	2.40	0.1421
CD	90022.50	1	90022.50	0.18	0.6786
A2	1.019E + 009	1	1.019E + 009	2022.32	< 0.0001*
B2	8.285E + 008	1	8.285E + 008	1643.81	< 0.0001*
C2	.011E + 009	1	1.011E + 009	2005.75	< 0.0001*
D2	1.079E + 009	1	1.079E + 009	2140.11	< 0.0001*
Residual	7.560E + 006	15	5.040E + 005
Lack of Fit	7.560E + 006	10	560E + 005
Pure Error	0.000	5	5
Cor-Total	2.818E + 009	29

*Significant

The R² value of 0.9973 gives 99.7 % variability in the production of bacteriocin, and about 2.8 % total variation cannot explained by the model. The close the value of R (R = multiple correlation coeffecient) to 1, the better the correlation between the experimental and predicted values (Pujari and Chandra 2000; Liu et al. 2003). The adequate precision is used to measure the ratio of signal to noise, it is believed to be desirable greater 4. Here the value 51.395 shows that the polynomial quadratic model is of an adequate signal, and can be used to navigate the design space. The “Pred R-Squared” 0.9845 were in reasonable agreement with the “Adj R-Squared” 0.9948 presented in Table 4.

Table 4.

R-Squared, Adj R-Squared, Pred R-Squared, and Adeq Precision value of the model

Std. Dev	709.93	R-Squared	0.9973
Mean	6453.49	Adj R-Squared	0.9948
C.V. %	11.00	Pred R-Squared	0.9845
PRESS	4.355 E + 007	Adeq Precision	1.395

The coefficient estimates of Eq.(3), along with the corresponding P values are given in Table 3.

The P values implied the significance of each coefficient and it is important in turn to indicate the pattern of mutual interaction between the coefficients. The smaller the value of P, the more significant is the corresponding coefficient (Rao et al. 2000). It can be seen from the Table 3 that the 3 linear coefficient (A, B and C), all quadratic coefficients and three of interaction coefficients i.e. AB, AD, BC are highly significant. The insignificant coefficients were not omitted, since it is a hierarchical model.

Figure 3a-3b. Response surface representation provides a method to show the relation between the response and experimental levels of each variable. These figures are useful in understanding the kind of interaction among test variables to deduce the optimum conditions. This technique has been widely adopted for optimizing the process of enzymes and peptides (Dey et al. 2001; Vohra and Satyanarayana 2002). Figure 3a, b, c depicts the contour plot showing the effects of soyatone and sucrose to have a significant effect on bacteriocin production, and it is positively interacted with the other 2 factors.

Fig. 3.

Three- dimensional curves showing the effects of a soyatone and sucrose, b soyatone and pH, c soyatone and temperature, d sucrose and pH, e sucrose and temperature, f pH and temperature.

This is in evidence with our result from the one-factor-at-a time approach; increase in soyatone concentration and interaction with sucrose at maximum level of 2.88 %, pH at the range of 4.50-5.56, and temperature at the range of 32–34.52 ºC increased the bacteriocin activity. But increase in sucrose, pH and temperature concentration decreases the activity of bacteriocin. As seen in Fig.3d and e, the optimum activity was obtained at higher sucrose concentration with pH and temperature at the same range as mentioned earlier. Further, Fig. 3f shows that pH at the range of 4.50-5.56, temperature at the range of 32–34.52 ºC increases the bacteriocin production. Therefore the results apparently prove that soyatone and sucrose was found to be the important component interacted with pH and temperature for the enhanced bacteriocin activity. The effect of variation in level of all 4 independent variables on bacteriocin production has been shown in the perturbation graph Fig. 4. From the graph it can be concluded that the sucrose plays an important role in enhanced bacteriocin activity and production, followed by soyatone with pH and temperature. Normal probability versus residuals was plotted in a graph Fig. 5 which showed that data were very close to the straight line and situated at both sides indicating the model is fairly good.

Fig. 4.

The perturbation graph showing the effect of all independent variables on bacteriocin production by P pentosaceus KC692718

Fig. 5.

Plot between expected normal values versus residuals

Validation of the optimum condition

In order to verify the optimization results and to determine the accuracy of the model, experiment was conducted in duplicate with the optimized media containing 1.03 % soyatone, 2.4 % sucrose, 5.5pH and 34.5 ºC. On experimentation, observed response of bacteriocin yield from the Pediococcus pentosaceus KC692718 was 25600.34 AU/mL strongly proves the suitability of the developed model.

Conclusion

Lactic acid bacteria (LAB) are well known microorganism in fermented foods like dairy products and processed vegetables. The present research was carried out to enhance bacteriocin produced by the LAB isolate Pediococcus pentosaceus KC692718 which was isolated from fermented vegetable pickle. In this study 2 step media optimization of bacteriocin production was carried out using OFAT and RSM for various physical and chemical factors. It was observed that bacteriocin produced by Pediococcus pentosaceus KC692718 was increased by 2.5 fold (AU/mL) for sucrose as carbon source and 4.7 fold (AU/mL) increase in the presence of soyatone which acted as nitrogen source when compared to its normal bacteriocin produced in the MRS medium (1 280 AU/mL). Further studies were carried out to increase bacteriocin production using RSM tool with optimum carbon and nitrogen sources as chemical factors and pH and temperature as physical factors in the modified media. A total of 4 fold increase in bacteriocin production (25600.34 AU/mL) was observed from OFAT experiments and a total of 20 fold increase of bacteriocin production was observed from modified MRS media. This is the first report on 20 fold increase of bacteriocin only by Pediococcus pentosaceus KC692718 isolated from fermented vegetable pickle.