Response Surface Methodology as an Approach for Optimization of Vinegar Fermentation Conditions Using Three Different Thermotolerant Acetic Acid Bacteria ()
1. Introduction
Vinegar is one of the most valuable food products due to its adaptability and range of usages, and the global vinegar market has reached a value of US$ 2.27 Billion in 2021 and is expected to increase by 2.6% up to 2027 [1] . Vinegar is a well-known natural food product derived from alcoholic and subsequently acetous fermentation of carbohydrate-rich foods. Vinegar is widely used in the food industry; domestically for pickling vegetables and fruits, as an ingredient in condiments like ketchup, and mayonnaise; and traditionally as a food seasoning and preservative [2] . It is a two-step biochemical process, in which the first step involves the transformation of sugar into ethanol by the action of yeast usually Saccharomyces species, followed by the oxidation of ethanol into acetic acid under aerobic conditions by acetic acid bacteria (AAB) [3] [4] [5] . Fermentative oxidation of ethanol to acetic acid by AAB is dependent on two sequential reactions of membrane-bound pyrroloquinoline quinone-dependent alcohol dehydrogenase (PQQ-ADH) and aldehyde dehydrogenase (ALDH), both of which are localized on the periplasmic side of the inner membrane [6] . These strictly aerobic microorganisms, used in several biotechnological processes, have the particular ability to transform alcohols and sugar alcohols into the corresponding organic acids [7] .
A major challenge in vinegar production is that AAB are extremely sensitive to environmental conditions [8] ; these limiting factors affect their growth and production capacities through parameters such as temperature, pH, oxygen, ethanol concentration or acetic acid concentration in the culture medium [9] [10] [11] . These inhibitory parameters are difficult to avoid since they are often substrates, targeted products, or physicochemical conditions.
In general, optimization is the process of bringing together and applying the process by considering the interactions of independent variables, as well as the effects of the independent variables on the response in accordance with the determined objectives [12] . Response surface methodology (RSM) is an efficient experimental strategy to determine optimal conditions for a multivariate system rather than the conventional method of one factor at a time (OFAT) which is incapable of determining the true optimum [13] . It is a mathematical and statistical technique for designing experiments, model building, and assessing the effect produced by several factors [14] .
In this study, the Response Surface Methodology was used to optimize the conditions of vinegar fermentation of 3 AAB strains (VMA1, VMA7 and VMAO). The aim of this study was to optimize the effect of 3 independent process parameters: fermentation temperature, original alcohol concentration and original acetic acid concentration using RSM coupled with BBD in order to obtain maximum acetic acid yield.
2. Materials and Method
2.1. Bacterial Strains
Three (3) AAB strains VMA1, VMA7, and VMAO previously isolated from fermented mango alcohol were tested in this study [15] . These strains were selected for their capacity to produce acetic acid, and their thermo-alcohol-acid tolerance capacities [16] .
2.2. Statistical Experimental Design
Response surface methodology is an empirical modeling technique, and it is used to estimate the relationship between the set of controllable experimental factors and the observed results [17] . This study used RSM to determine the optimum conditions affecting the acetic acid fermentation process of three AAB strains VMA1, VMA7 and VMAO. The experiments are performed by using a Box-Behnken-Design. A BBD is a type of response surface design that do not have axial points, thus, we can be sure that all design points fall within your safe operating zone. BBD also ensure that all factors are not set at their high levels at the same time [18] [19] . The independent variables of acetic acid fermentation were X1, X2 and X3 respectively fermentation temperature (FT), original alcohol concentration (OAC) and original acetic acid concentration (OAAC). Each of variables to be optimized was coded at 3 levels: −1, 0, and 1. Table 1 shows the variables, their symbols and levels. This gives a range of these variables in the acetic acid fermentation process. Experimental design and statistical analysis of the data were performed using STATISTICA 10 for Windows™.
Variables were combined in 15 experiments repeated in triplicate. The experimental values were fitted in Equation (1) as a second order polynomial equation, including the linear and cross effect of the variables:
where: Y represents the predicted response, i and j are linear and quadratic coefficients, respectively, β is the regression coefficient, and n is the number of variables studied in the experiments.
2.3. Preparation of Microorganisms and Inoculum
First, the strains were cultured in 100 mL of YPG broth (Yeast-extract, Peptone, Glucose), containing respectively 3% (w/v) yeast, 1% (w/v) Peptone, 3% (w/v) glucose. Then, the medium was incubated at 37˚C until the optical density (OD600 nm) of the suspension reaches 0.4.
Table 1. Levels of factors and variables chosen for Box–Behnken experimental design.
2.4. Fermentation Conditions
Acetic acid fermentations were carried out in YGKM fermentation medium containing 5% (w/v) Yeast, 5% (w/v) Glucose, 3.3% (w/v) K2HPO4, and 1.1% (w/v) MgSO4. Sterile alcohol and acetic acid were added to the fermentation media supplemented at the specific concentrations listed in Table 2. The fermentation media were in each case inoculated with 10% (v/v) of pre-culture of the microorganism concerned. The flasks were then incubated at the temperature of each experiment (30˚C, 37.5˚C or 45˚C) with shaking at 150 rpm for a total of 21 days.
2.5. Titratable Acidity
In order to determine the amount of acid in the vinegar, the acetic acid will be titrated with a solution of 1N of sodium hydroxide (NaOH) using phenolphthalein as an indicator. All measurements were conducted in triplicate. The amount of acetic acid produced in g per 100 mL was calculated in using the following formula:
Vb: average volume of sodium hydroxide needed to reach the end point;
Cb: concentration of NaOH solution;
pm: molar mass of acetic acid;
Va: Volume of vinegar sample;
mv: volumic mass of acetic acid.
3. Results and Discussion
This study used RSM to determine the optimum conditions affecting the fermentation process of vinegar production by three AAB strains VMA1, VMA7 and VMAO. The three independent variables, fermentation temperature, original alcohol concentration and original acetic acid concentration and its influence on acetic acid percent as an important response, were studied using Response Surface Methodology coupled with a 3 factors 3 levels Box-Behnken experimental Design. RSM is widely used for experimental data analysis; it creates a link between responses and control variables and guesses the response values of the control variables within a specific range [20] . This model anticipates experimental changes such as changing operating conditions and various processing steps, and define not only the influences of independent variables on the responses, but also the interaction between parameters to achieve best system performance [21] [22] . RSM has been used worldwide for the optimization of numerous industrial processes [23] - [35] .
In this study, 45 runs of vinegar production were carried out according to the experimental design provided, using three AAB strains VMA1, VMA7, and VMAO. Analysis of the experimental data produced the following mathematical models presented in Equations (1)-(3):
(1)
(2)
(3)
3.1. Parameters Effects and Parameters Interactions
Equations (1)-(3) represent the relationship between response and fermentation process variables on acetic acid production respectively by VMA1, VMA7 and VMAO. Positive regression of the models indicated a synergic effect whereas negative coefficients indicate an antagonist effect on the dependent response variable [13] . A positive effect of a factor indicates that the corresponding response increases when the factor changes from low to high level, whereas a negative effect indicates that the corresponding response decreases when the factor varies from low to high level [36] .
3.1.1. For VMA1
The results showed that the linear FT, the linear and quadratic OAC and the linear OAAC have a positive effect on the production of acetic acid. The interaction FT-OAC, the interaction FT-OAAC and the interaction OAC-OAAC also have a positive effect on the acetic acid production. While the quadratic FT and the quadratic OAAC have a negative effect on acetic acid production (Table 3).
3.1.2. For VMA7 and VMAO
The results exhibited that the linear FT, the linear OAC, the linear OAAC, the interaction FT-OAC, the interaction FT-OAAC and the interaction between
Table 3. The data of parameter effects and parameter interactions on the production of vinegar by strain VMA1.
OAC-OAAC have a positive effect on the acetic acid yield. However, the quadratic FT, the quadratic OAC and the quadratic OAAC showed a negative effect on acetic acid production (Table 4 and Table 5).
3.2. Suitability and Adequacy of Models
The suitability of model and data can be seen from the value of R2. The coefficient of determination (R-Squared) represents the proportion of variability in the data explained or accounted by the model [37] . The determination coefficient (R2) value indicates the extent to which the independent variables can explain the observed response values. R2 values range from 0 to 1. The closer to the value of 1, the equation of the model approaches the experimental data [38] [39] . A good fitted model should have a minimum R2 of 80% an R2-value that is greater than 0.9 proves a very high degree of correlation [40] . The results of data analysis showed that the R2 values of the models are 0.96, 0.94, and 0.98 respectively for VMA1, VMA7 and VMAO. These R2 values indicate an excellent fit between predicted and experimental values. Meaning for VMA1, VMA7 and VMAO respectively 96%, 94% and 98% of the variability of the response can be explained by the model. The adjusted R2 in this study were found to be respectively 0.95, 0.92 and 0.98 which are high and clarified the significance of the model, implying that the predicted and experimental values of the acetic acid production are quite similar for all three strains.
The adequacy of models is also supported by the F values of the models as shown in Tables 6-8. Values of “Prob F” < 0.0500 indicate models are significant
Table 4. The data of parameter effects and parameter interactions on the production of vinegar by strain VMA7.
Table 5. The data of parameter effects and parameter interactions on the production of vinegar by strain VMAO.
[41] . The Models F-values of 22.34, 5.83 and 5.33 indicated the statistical significance of the experimental models and were statistically significant (p < 0.05).
3.3. Significance and Magnitude of the Factors
It is important to determine the most influential variable in the acetic acid
Table 6. Examination of F value model for VMA1.
Table 7. Examination of F value model for VMA7.
Table 8. Examination of F value model for VMAO.
production for each AAB strain. Pareto chart provide information about the most influential process variables. It reveals significance and magnitude of the factors that influence the process, depending on the significance level [42] [43] .
Figures 1-3 present the Pareto chart of the acetic acid fermentation process using respectively strains VMA1, VMA7 and VMAO. It shows the order and significance of the factors.
For strain VMA1, fermentation temperature (L) is the most influential variable in the acetic acid production, and then followed by the original alcohol concentration (L) and the original acetic acid concentration (L).
For strain VMA7, original acetic acid concentration (L) is the most influential variable in the acetic acid production, and then followed by fermentation temperature (L) and the original alcohol concentration (L).
For strain VMAO, original alcohol concentration (L) is the most influential variable in the acetic acid production, and then followed by fermentation temperature (L) and the original acetic acid concentration (L).
3.4. Three-Dimensional and Two-Dimensional Response Surface Plots
Figures 4-6 illustrate the 3D and 2D response surface plots of the acetic acid fermentation using respectively strains VMA1, VMA7 and VMAO. The surface plots were generated to determine the best levels and interactions between the parameters in order to determine the optimal conditions for maximum acetic acid production. Surface plot shows a functional relationship between a designated dependent variable (Y), and two independent variables (X and Z) [44] [45] . The plots showed the pair-wise combinations of the three process variables, (FT-OAC, OAC-OAAC, FT-OAAC); the interaction of two independent variables and their effect in the acetic acid production was assessed while fixing the
Figure 1. Pareto chart of the standardized effect estimate (Absolute Value) using strain VMA1.
Figure 2. Pareto chart of the standardized effect estimate (Absolute Value) using strain VMA7.
value of the third variable.
3.4.1. For VMA1
Figures 4(a)-(c) show the combined effect of FT-OAC, OAC-OAAC and FT-OAAC on acetic acid production by AAB strain VMA1 while OAAC, FT and OAC values were respectively kept constant at 1% (w/v), 37.5˚C and 12.5% (v/v).
It was observed that acetic acid yield increased as the two independent parameters were increasing. Low values of FT (≤34˚C) associated with low values of OAC (≤6% (v/v)) led to decrease of acetic acid production. In the same way, whatever the OAC value, lower FT (≤30˚C) and low OAAC values (≤0.2) have a negative effect on acetic acid yield. Likewise, low OACs (≤10% (v/v)) have a negative effect on acetic acid yield whatever the OAAC values.
3.4.2. For VMA7
Figures 5(a)-(c) present the combined effect of FT-OAC, OAC-OAAC and
Figure 3. Pareto chart of the standardized effect estimate (absolute value) using strain VMAO.
FT-OAAC on acetic acid production by AAB strain VMA7 while OAA concentration, FT and OAC values were respectively fixed at 1% (w/v), 37.5˚C and 12.5% (v/v). Low FT [30˚C - 35˚C] decreased the medium yield when associated with OAC values between [6% - 13.5% (v/v)] and OAAC ≤ 0.4% (w/v). Low values of OAAC (≤0.3% (w/v)) have a negative effect in the AA production whatever the OAC. In the contrary, high OAAC ranging [1.7 - 2] associated with fermentation temperatures between [33˚C; 39˚C] give good responses.
3.4.3. For VMAO
Figures 6(a)-(c) present the combined effect of FT-OAC, OAC-OAAC and FT-OAAC on acetic acid production by AAB strain VMAO while OAA concentration, FT and OAC values were respectively fixed at 1% (w/v), 37.5˚C and 12.5% (v/v).
It was observed that low values of OAC and FT ≤ 34.5˚C led to a decrease of the fermentation yield whatever the OAAC. Likewise, low FT even associated with high OAC ([18% - 20% (v/v)], as well as OAAC ≤ 0.6% (w/v) associated with OAC ≤ 13.5% (w/v) decreases the production of AA. Low OAC (≤10% (v/v)) and low values of OAAC (≤0.4% (w/v)) have a negative effect on AA production whatever the fermentation temperature value.
These results allowed us to define the following optimization ranges for each of the strains used in this study.
· VMA1
FT: [37.5˚C - 45˚C]
OAC: [16% - 20% (v/v)]
OAAC: [1.5% - 2% (w/v)]
· VMA7
FT: [40˚C - 45˚C]
Figure 4. 3D and 2D surface plot for acetic acid yield using AAB strain VMA1. (a): Interaction Fermentation Temperature-Original alcohol concentration; (b): Interaction original alcohol concentration-Original acetic acid concentration; (c) Interaction Fermentation temperature-Original acetic acid concentration.
Figure 5. 3D and 2D surface plot for acetic acid yield using AAB strain VMA7. (a) Interaction Fermentation Temperature-Original alcohol concentration; (b) Interaction original alcohol concentration-Original acetic acid concentration; (c) Interaction Fermentation temperature -Original acetic acid concentration.
Figure 6. 3D and 2D surface plot for acetic acid yield using AAB strain VMAO. (a): Interaction Fermentation Temperature-Original alcohol concentration; (b) Interaction original alcohol concentration-Original acetic acid concentration; (c) Interaction Fermentation temperature-Original acetic acid concentration.
OAC: [14.5% - 20% (v/v)
OAAC: [1.7% - 2% (w/v)]
· VMAO
FT: [42˚C - 45˚C]
OAC: [17% - 20% (v/v)]
OAAC: [1.5% - 2% (w/v)]
These results confirm the resistant thermo-ethanol-acido character of the strains used in this study and previously described by Kourouma [16] . The use of these acetic strains in the production of vinegar could seriously lead to a decrease or even an ablation of the costs related to the cooling of bioreactors especially in warm and hot countries where the ambient temperature can easily exceed 37˚C, in a context of global warming. This would probably lead to a reduction in the total production cost, thus impacting the final price of the vinegar marketed, in addition to having a bank of acetic strains able to withstand such drastic conditions.
4. Conclusion
The oxidation of ethanol to acetic acid in order to produce vinegar is one of the best well-known uses of AAB in industries. However, during the acetification process, AAB are affected by many variables. This study used RSM coupled with BBD using 3 process parameters: fermentation temperature, original alcohol concentration and original acetic acid concentration to enhance the acetic acid production of 3 AAB strains, namely VMA1, VMA7 and VMAO. RSM optimization was successfully achieved. Results indicate an excellent fit between predicted and experimental values; the R2 values are respectively 0.96, 0.94, and 0.98 which means 96%, 94% and 98% of the variability of the response can be explained by the model; the Model F-values of 22.34, 5.83 and 5.33 indicated that experimental models were statistically significant, with p < 0.05. The most influencing variables for AA production using VMA1, VMA7 and VMAO were respectively fermentation temperature, original acetic acid concentration and original alcohol concentration. The maximum acidity was obtained respectively at fermentation temperatures, original alcohol concentrations, original acetic acid concentrations ranging [37.5˚C - 45˚C], [16% - 20% (v/v)], [1.5% - 2% (w/v)] for VMA1; [40˚C - 45˚C], [14.5% - 20% (v/v)], [1.7% - 2% (w/v)] for VMA7; and [42˚C - 45˚C], [17% - 20% (v/v)], [1.5% - 2% (w/v)] for VMAO.