Abstract

Bacteriocins produced by various lactic acid bacteria (LAB) have found enormous use in the food industries as biopreservatives. This study evaluated the effect of varied culture conditions (temperature, pH, and NaCl concentration) on bacteriocin production by Lactococcus lactis MT186647 isolated from fermenting African oil bean seeds (Pentaclethra macrophylla Benth) using response surface methodology. A three-factor central composite design (CCD) was adopted with the interest of estimating the optimal conditions for its production using the response surface regression model, which evaluated the linear, squared, and interactive relationship between the response variables. The analysis of variance (ANOVA) using Minitab statistical software version 14.13 showed an R² = 0.869 variations in the response variable for culture conditions. It was accounted for by the predictors suggesting that the model was adequate. The optimal culture condition for bacteriocin production by L. lactis was estimated at 30.5C, pH 5.9, 1.94% NaCl concentration at Y = 12.31 mm where Y represents the response (zone of inhibition) against Staphylococcus aureus ATCC 19095 using the agar well diffusion assay method. Validation of optimal values according to the regression model, produced an inhibition zone at Y = 13.33 ± 0.29 mm. There was a 19.33% increase in bacteriocin activity compared to an OFAT optimized medium from a previous study.

1. Introduction

Bacteriocins are antimicrobial proteins that are ribosomally synthesized by bacteria and possess a wide range of antagonistic activity towards genetically related bacterial species ^{1, 2}. However, there have been reports of bacteriocin activity against some Gram-negative bacteria, especially those from lactic acid bacteria ^{3, 4}. Bacteriocins from lactic acid bacteria (LAB) have presented substantial potential applications in the prevention and treatment of many diseases of the gastrointestinal tract ^{5, 6}, as well as in food industries as bio-preservatives ^{7, 8}.

Bacteriocin production by LAB largely depends on bacterial growth during optimum metabolic activity associated with maximum cellular growth ^{9, 10}. However, certain environmental factors, such as pH, temperature, and media composition, can influence bacteriocin production ^{11, 12, 13, 14}. Various components, such as salts, carbohydrates, surfactants included in the fermenting medium, can interfere with the bacteriocin production by LAB ^{15, 16, 17}.

The successful application of surface response methodology (RSM) in many fields, especially in biotechnology, has facilitated the study of distinct factors that influence bacteriocin production ^{18, 19, 20}. Numerous studies have employed RSM in evaluating the effect of environmental conditions and medium components on the antibacterial activity of bacteriocins of Pediococcus acidilactici BA28 ²¹, Lactobacillus acidophilus ²², Lactobacillus brevis DF01 ²³, and Lactobacillus plantarum ²⁴.

This study aimed to optimize bacteriocin production by Lactococcus lactis MT186647 in MRS broth containing different supplements, pH levels, temperatures, and salt concentrations using RSM.

2. Materials and Methods

Lactococcus lactis LB10 was isolated from fermenting African oil bean seeds (AOBS). The isolate was identified by sequencing the 16S rRNA gene followed by BLAST and nucleotide sequence deposited with the NCBI under the accession number MT186647.

The indicator strain Staphylococcus aureus ATCC 19095 obtained from the Federal Institute of Industrial Research, Oshodi (FIRO), Lagos, Nigeria, was cultured on Mannitol salt agar (HiMedia) at 37°C. The strain was then sub-cultured on Nutrient broth and stored at -20°C with 20% (v/v) glycerol.

Lactococcus lactis MT186647 was grown in MRS broth (HiMedia) at 35°C for 24 h. The culture broth was centrifuged at 10 000 rpm for 15 mins; at 4C. The pH of the cell-free supernatant (CFS) was adjusted to 6.5 using 1 mol/l NaOH then partially purified according to ²⁵. Bacteriocin precipitation was achieved by adding ammonium sulfate and then centrifuged at 10 000 rpm for 20 mins. The residue was then re-dissolved in double-distilled water and filtered through a 0.22 µm syringe filter. Bacteriocin activity was determined using the agar well diffusion method, according to ²⁶.

The optimization of bacteriocin production by L. lactis MT186647 was done using RSM. Culture condition variables (temperature, pH, and sodium chloride concentration) were optimized for enhanced antibacterial activity using the central composite design (CCD). A 2-level, 3-factor (2³) response surface methodology was adopted to estimate the optimal culture conditions and supplementation on bacteriocin production by the isolate. Bacteriocin production was estimated using a response surface regression model to estimate the linear, squared, and interactive relationship between the response variable and the predictors. Optimization comprises of 20 experimental trials (8 factorial points with six axial and six central points). Fermentations were in random order involving three variables (pH = X₁, temperature = X₂, and NaCl concentration = X₃) with five coded levels (-1.68, -α, 0, α and 1.68) to determine the effect of independent variables on response variable (antibacterial activity) according to ²⁷.

In developing the regression equation to determine the relationship between the independent and dependent variables for the RSM, Minitab statistical software 14.13 was used to determine the effects of linear, squared, and interactive terms of the independent variables.

Given the values of the minimum and maximum level of temperature, pH, and Sodium Chloride concentration to be 28 and 32°C, 5.5 and 6.5, and 1.5 and 2.5%, respectively, we obtain the uncoded units for the design.

where:

where X is the absolute value, X_min = the minimum level value, X_max is the maximum level value.

The response factor (Y_i) was modeled into the regression analysis as a mathematical function of a few continuous factors. The response was then expressed using a second-order polynomial equation.

(1)

Where Y is the predicted response (bacteriocin activity in terms of zone of inhibition), β₀ the intercept, β₁, β₂ and β₃ are the linear coefficients, β₁₁ β₁₁ β₃₃ are the squared coefficients, β₁₂ β₁₃ β₂₃ are interaction coefficients, X₁ the temperature, X₂ the pH and X₃ sodium chloride concentration.

A three-dimensional surface plot was drawn to show the interactive effect of the independent variables on bacteriocin production. The optimal values were obtained by solving the regression equation and analyzing the surface contour plots. The coefficient of determination (R²) was employed for the goodness of fit of the model equation ²⁸. All experiments were conducted in triplicates. Obtained values were analyzed using Fisher’s test, and differences were considered statistically significant at p<0.05.

To verify the optimization results and accuracy of the model, experiments were carried out in duplicates using the optimal culture conditions obtained from the RSM and compared to results obtained using one factor at a time OFAT, according to ¹⁴.

3. Results

The results of 20 runs from CCD experiments for studying the effect of culture conditions (pH, temperature, and NaCl concentration) on bacteriocin production by Lactococcus lactis MT186647 from fermenting African oil bean seeds (AOBS) is as shown in Table 1. The maximum experimental value for bacteriocin production was 12.87 mm based on RSM.

Equation 2 shows the regression analysis data obtained were fitted into a quadratic model to generate a second-order regression equation for the full actual model on bacteriocin production. Where Y is bacteriocin activity (zone of inhibition in mm), A is the temperature (C), B is pH, and C is NaCl concentration (%).

An F-test checked the statistical significance of Eq. 2, and Table 2 shows the ANOVA table testing the model, linear, squared, and interaction effects of variables. The model produced an F-value (3.75), P-value (0.031), and an insignificant lack of fit (P_model > F) at 0.102.

The regression equation generated after ANOVA indicated an R² value of 0.7894. All linear and interactive effects were not significant at (P>0.05), while the squared effects of temperature (A) and NaCl concentration (C) were significant at (P<0.05).

(2)

Figure 1 shows the validation (graphical) of the regression model using residual plots. The probability plot showed linearity, residual versus fitted were randomly scattered; whereas, the histogram of frequency versus residual had a dumbbell shape, and residual versus observation order had a symmetrical pattern.

The 2D contour and 3D response surface plots of antibacterial activity of bacteriocin from Lactococcus lactis MT186647 against S. aureus are shown in Figure 2 and Figures 3a, b, and c respectively.

Figure 4 displays the predicted optimal level based on the regression model. An optimal response variable of 12.31 mm was obtained when temperature, pH, and NaCl concentration were fixed at 30.5°C, 5.9, and 1.94%, respectively.

The correlation between actual and predicted bacteriocin activity is presented in the parity plot in Figure 5. Actual and predicted values are closely aligned.

4. Discussion

The ANOVA result shows the statistical analysis for the level of significance of all variables. The model adequacy was analyzed using the R² coefficient, correlation, and model significance (F-value). The P-value >F of less than 0.005 shows that the model is statistically significant and means that the model can predict experimental results ^{23, 29, 30}. All axial points with zero coded values were repeated six times to estimate the pure error for lack of fit as predictions can be desirably made with an insignificant lack of fit. The model produced a negligible lack of fit (P _model > F) at 0.102. The regression equation generated after ANOVA indicated an R² value of 0.7894 (a value of R² > 0.75 indicates aptness of the model). These coefficient values ensured a fine adjustment of the quadratic model to experimental data and explained 78.94% of the variability in the response ²⁹.

The linear coefficient values of all variables showed no impact on bacteriocin production; whereas, the quadratic effect of variables were all insignificant. The squared effect coefficient of temperature (p=0.004) and NaCl concentration (p=0.007) were significant, while others were not significant, as bacteriocin production decreased as values of variables increased. Alteration of temperature and NaCl concentration had a strong influence on bacteriocin production. Similar results were observed by ³¹ as the two factors significantly affected bacteriocin production by Lactobacillus brevis OG1.

A comparison of actual (experimental) and predicted values detailed a good correspondence as there was proximity indicating an adequate description of the relationship between variable and response owing to the empirical model derived from the RSM ²⁹.

The residual plot analyzes the error pattern in data gotten from experiments. Linearity in the error pattern of the probability plot indicates little or no issues in data and verifies the regression model. The representative graphical validation of the regression model with the help of residual plots showed that the residual is symmetry (normally distributed) and contains a constant variance. This continuous variation indicates the results from this model can be used for inferential purposes following the central limit theorem ²⁷. The standard probability plot was linear, residual versus fitted values scattered around zero (0) point. There was a dumbbell shape in the frequency versus residual plot, while there was a symmetrical pattern observed in the residual versus order plot.

The interaction of factors shown graphically with the 2-D and 3-D response surface plots showed that culture conditions should not exceed specified values. The Maximum bacteriocin production was achieved at 30.5°C, pH 5.9, and 1.94% NaCl concentration, producing a zone of inhibition of 12.31 mm against Staphylococcus aureus ATCC 19095, as shown in the optimal plot. The statistical optimization of culture conditions showed that temperature and NaCl concentration significantly affected bacteriocin production. Different temperatures affect the ability of LAB to produce bacteriocins ^{28, 32}. One factor at a time optimization of temperature, pH, and NaCl concentration for bacteriocin production by Lactobacillus species showed peak values of 11.17 ± 0.33 mm, 10.50 ± 0.29 10.70 ± 0.17 mm, respectively ¹⁴. The increase in bacteriocin activity indicates the effect of optimized conditions for fermentation by LAB as a 10.20% increase in activity was observed when CCD was employed.

Validation of obtained results was carried out applying the optimal culture variable conditions as follows: temperature 31°C, pH 6.0, and 1.9% NaCl concentration in a shaker incubator. The maximum antibacterial activity obtained experimentally was 13.33 ± 0.29 mm diameter of inhibition against 12.31 mm predicted from RSM and agreed with the model prediction. Therefore, the developed model is considered reliable for predicting enhanced bacteriocin production in Lactococcus lactis MT1186647. More so, comparing the OFAT optimized medium, according to ²³, showed a 19.33% increase in bacteriocin activity, which means RSM is a useful tool for optimization.

5. Conclusion

The enhancement of bacteriocin production by Lactococcus lactis MT1186647, using CCD by optimizing some culture conditions (temperature, pH, and NaCl concentration), resulted in the antibacterial activity of 13.33 ± 0.29 mm. The statistical optimization resulted in antifungal activity of 13.33 ± 0.29 mm against S. aureus, which was 19.33% higher than that of the OFAT optimized medium (11.17 ± 0.17 mm). The experimental results were in correlation with predicted values, hence, validating the experiment and confirming the accuracy of the RSM model.

Acknowledgements

The authors wish to thank the staff of the National Agency for Federal Institute of Industrial Research, Oshodi (FIRO), Lagos, Nigeria and the National Agency for Food and Drug Administration and Control (NAFDAC), Agulu, Anambara State, Nigeria.

References