The Use of Response Surface Methodology as a Statistical Tool for Media Optimization in Lipase Production from the Dairy Effluent Isolate Fusarium solani

The optimization of extracellular lipase production by Fusarium isolani strain SKWF7 isolated from dairy wastewater was carried out in this study. Initially, the physicochemical factors significantly influencing enzyme production were studied by varying one-factor-at-a-time (OFAT). A mesophilic temperature of 40°C, alkaline pH of 8, and incubation period of 72 hours were found to be the optimal conditions for lipase production. Among the media components, the disaccharide sucrose acted as the best carbon source; palm oil as the best inducing lipid substrate; casein and (NH4)2SO4 as the best organic and inorganic nitrogen sources; Ca2+ ion as the best trace element. In the next phase of work, statistical optimization of medium components was performed by employing the Box-Behnken design of Response Surface Methodology (RSM). The optimum concentrations of three significant factors, namely, palm oil, (NH4)2SO4, and CaCO3 were determined by this method to be 5% (v/v), 5.5 g/L, and 0.1 g/L, respectively. RSM-guided design of experiments resulted in a maximum lipase production of 73.3 U/ml, which is a 1.7-fold increase in comparison with that obtained in the unoptimized medium. These results point towards the success of the model in developing a process for the production of lipase, an enzyme of enormous industrial significance.


Introduction
Lipases are enzymes that belong to the class of hydrolases and are involved in catalyzing the hydrolysis of triglycerides to fatty acids and glycerol, this reaction occurs at the oil-water interface [1]. Besides, they are also capable of catalyzing the reverse reaction, that is, ester synthesis, in water-restricted environments [2]. Transesteri�cation and resolution of racemic mixtures are also reactions that could be facilitated by lipases. Such a versatile nature has paved the way for their application in diversi�ed industries including food, dairy, cosmetic, detergent, and pharmaceuticals [3]. e environmental applications of the enzyme are also innumerable, where they break down fat, oil, and greasy material in the wastewater [4]. ese constituents could pose problems in the sewers as well as the treatment plant, impeding oxygen transfer in aerobic biological treatment systems. ese are some of the underlying reasons for a sustained interest in the enzyme.
New organisms continue to be screened, with hopes of stumbling upon isolates with extremophilic properties that can survive harsh conditions and produce enzymes with uni�ue characteristics desirable for speci�c industrial applications. Aer this initial step of isolation and screening, optimization of enzyme production by the isolate is warranted. pH, temperature, aeration rate, incubation period, inoculum level and age, carbon, nitrogen sources and trace elements in the medium, and so forth are some of the 2 ISRN Biotechnology physicochemical factors that in�uence enzyme production and these have to be optimized [14].
Medium optimization has conventionally been done by varying one factor at a time (OFAT), and it is workable as long as the production process is in�uenced by a limited number of variables. Even then, it fails to assess the interaction effects of those variables and becomes quite cumbersome, entailing a large number of trials when several variables are to be considered. Hence, researchers seek the aid of statistical tools such as Placket Burman and Response Surface Methodology (RSM) for effective optimization of their production media [15,16]. is applies to lipase production as well [17].
In our study, fungal strains isolated from dairy wastewater were screened for lipase activity; physicochemical parameters having signi�cant impact on lipase production were identi�ed through the classical OFAT approach, and later RSM (Box-Behnken Design) was applied in determining the optimal concentrations as well as interaction effects of those parameters.

Materials and Methods
2.1. Chemicals. p-nitrophenyl palmitate (p-NPP) was purchased from Sigma Aldrich, USA. All other chemicals and reagents used were of analytical grade and purchased from Hi-Media, India.

Sample Collection.
Wastewater sample was collected from "Central Ghee" dairy industry, Gobichettipalayam, Tamil Nadu, India. It was stored at 4 ∘ C and used as a source of possible lipolytic fungal strains.

Isolation and
Screening. e effluent sample was subjected to serial dilution and then plated on potato dextrose agar (PDA) plates for fungal isolation. Colonies on the plate that showed morphological difference were subcultured on PDA slants and maintained as pure culture. Such pure cultures were inoculated into glucose-yeast extract-peptone (GYP) broth supplemented with 2% olive oil and containing trace amounts of mineral salts (glucose 20 g/L, yeast extract 10 g/L, peptone 10 g/L, CH 3 COONa⋅3H 2 O 10 g/L, MgSO 4 0.3 g/L, MnSO 4 0.1 g/L, CuSO 4 ⋅5H 2 O 1.5 mg/L, KCl 0.5 g/L). ey were cultured at room temperature for 48 hours in a rotary shaker (90 rpm) aer which they were centrifuged at 10,000 g for 10 minutes at 4 ∘ C. e supernatant was used as the crude enzyme extract, whose activity was assayed by using p-NPP as substrate (ELICO UV-VIS Spectrophotometer) [18].
For the enzyme assay, 30 mg of p-NPP was dissolved in 10 mL of isopropanol (solution A); 0.1 g of gum arabic and 0.4 mL Triton X-100 were added to 90 mL of 50 mM Tris-HCl buffer, pH 8 (solution B). ey were mixed together to get the substrate solution. To 9 mL of this solution, 1 mL of enzyme was added and the mixture incubated at 32 ± 2 ∘ C for 15 min. Its absorbance was measured at 410 nm, the absorption coefficient of p-nitrophenol at 410 nm is 0.0148. One unit of lipase activity (U) was expressed as mol of pnitrophenol released per minute under the assay conditions. e isolate that exhibited maximum activity was selected for further optimization studies.

Optimization of Production Medium
2.4.1. OFAT Method. Growth conditions and chemical components were varied one at a time in the GYP medium and their impacts on lipase production were studied. e incubation time of the culture was varied in the range of 12-96 hours; the incubation temperature in the range of 30-60 ∘ C; initial pH in the range of 4.0-12.0. e chemical composition of the growth medium was then altered in the following ways: glucose was substituted with starch, sucrose, or maltose (2%); nitrogen source was replaced by other organic (1%) or inorganic (0.5%) sources, namely, soy peptone, casein, sodium nitrate, ammonium sulfate, or ammonium chloride; olive oil was replaced with gingelly, sun�ower, groundnut, palm, or tributyrin oil (2%); �nally, the basic GYP medium was augmented with trace amounts of different inorganic ions (Ca 2+ , Fe 3+ , Zn 2+ , Ba 2+ , or Mg 2+ ).

Response Surface Methodology. e interactive effects of three signi�cant factors
(oil source), (nitrogen source), and (inorganic ion) on the response, namely, lipase production were determined statistically using RSM. A Box-Behnken design developed by the Design Expert soware, version 8.0.7.1 (Stat Ease Inc. Minneapolis, USA, trial version) was adopted for this purpose.
Each one of the above independent variables , , and was taken at a central coded value considered as zero and studied at three different levels. e values of these variables in actual and coded forms are presented in Table  1. A matrix consisting of 17 experiments with 5 replicates at the centre point generated by the soware was applied for maximizing the lipase production. Production was carried out in 250 mL Erlenmeyer �asks containing 100 mL of the production medium (pH 8.0). e �asks were sterilized by autoclaving at 120 ∘ C for 20 min, inoculated with the culture under aseptic conditions, and incubated at 40 ∘ C for 72 h, in an orbital shaker set at 90 rpm. At the end of the incubation period, lipase assay was performed using the cellfree supernatant and activity was recorded as the response (dependent variable). All experiments were carried out in triplicate and the data represent the mean. e following second-order polynomial equation describes the relationship between the dependent and independent variables: where, is the predicted response, 0 is the intercept, 1 2 ,

Isolation and Screening.
Several fungal colonies grew on the PDA plates that were inoculated with the serially diluted wastewater sample. 19 morphologically different colonies were subcultured and the resulting pure cultures were preserved on PDA slants overlaid with liquid paraffin. p-NPP assay of these isolates aer cultivation in the production medium revealed appreciable activities for 10 ( Figure 1) and negligible activities for the rest (data not shown). e isolate SKWF7, which showed a maximum activity of 36.9 U/mL was identi�ed to be Fusarium solani by the Indian Agricultural Research Institute (IARI), New Delhi. Further experiments on medium optimization were performed with this isolate.

Optimization of Production Medium: OFAT Method.
When the physicochemical parameters were varied one at a time and their impact on lipase production was investigated, the following results were obtained. Enzyme production steadily increased with increasing incubation time of the culture and reached a plateau at 72 hours ( Figure 2). Similar or longer incubation periods have been widely reported for fungi [19]. In the case of Mucor sp., highest production was obtained only aer 6    days of cultivation [20]. Highest enzyme production was recorded with an incubation temperature of 40 ∘ C (Figure 3). Upon altering the culture medium pH, it was witnessed that there was good production in the range of 4.0-8.0, with the production peaking at 8 and rapidly falling thereaer ( Figure  4). A strain of Aspergillus niger showed maximum enzyme production aer 96 h of growth in a medium with an initial pH of 7.0 [21].      e disaccharide sucrose was found to be the best carbon source ( Figure 5). e inducible nature of lipases necessitates the presence of a lipid substrate in the medium and in this case, induction of enzyme synthesis was best favoured by the presence of palm oil ( Figure 6). A combination of 2% olive oil and 2% glucose was demonstrated to be conducive for lipase production by A. niger [22]. e organic nitrogen source casein and inorganic source ammonium sulphate resulted in maximum lipase production ( Figure 7). Fusarium oxysporum showed optimal production with the use of peptone and ammonium dihydrogen phosphate [23].
Ca 2+ best stimulated the enzyme production by up to 20%, in comparison to a medium devoid of trace elements (Figure 8). Several reports vouch for the stimulatory role of this element [24]. e presence of Mn 2+ had no effect, while Ba 2+ had a negative impact on enzyme production.

Response Surface
Methodology. RSM incorporates the interaction effects of variables and aids us in simultaneously optimizing several process parameters within a minimal number of experimental runs. Such statistically assisted experimental designs can lead to signi�cantly enhanced production. In our investigation, palm oil ( ), (NH 4 ) 2 SO 4 ( ) and CaCO 3 ( ), which were inferred to be signi�cant media constituents based on OFAT studies were selected for optimization through the Box-Behnken design. e experimental design set up is shown in Table 1.
e ANOVA results and diagnostic case studies are given in Tables 2 and 3, respectively. e value serves as a tool for checking the signi�cance of each of the coefficients and is indicative of the interaction strength of each independent variable. Low values of 0.05 indicate high signi�cance of the corresponding coefficients. In general, larger t, and smaller values indicate that the corresponding coefficient terms are signi�cant [25]. Subjecting our model to ANOVA showed an value of 1704.58 and a value 0.0001 which indicate that the model is highly signi�cant. All three linear coefficients, squared coefficients, and one interaction coefficient (AB) are signi�cant, as evidenced from low and high values.
2 value gives a measure of how much variability in the observed response can be explained by the experimental parameters and their interactions [26]. e 2 value for this model is 0.9995. When expressed as a percentage, it implies that a total variation of 99.95% in enzyme activity can be attributed to the independent variables and only 0.05% cannot be ascribed to them. e predicted 2 of 0.9927 is in acceptable agreement with the adjusted 2 of 0.9989. e adjusted arranges the values for the sample size and for the number of variables in the model.
Adequate precision measures signal-to-noise ratio and detects which experimental parameters generate signals that are large in comparison to the noise. A ratio greater than 4 is desirable. In this case, the value obtained is 327.78 and is thus indicative of an ample signal.
e regression equation is expressed graphically in the form of contour plots which depict the interactions among the independent variables and their in�uence on enzyme production. e contour plots might be elliptical mounds, saddle points, or rising ridges [27]. Here, contour plots were generated by varying the levels of two factors while keeping the third one constant. e plot of oil concentration versus calcium carbonate concentration (Figure 9(a)) is elliptical in nature indicating signi�cant interactions between the two variables. e plot between (NH 4 ) 2 SO 4 and oil concentrations Figure 9(b) and the one between (NH 4 ) 2 SO 4 and CaCO 3 concentrations Figure 9(c) are not elliptical in nature, but saddle points, indicating that there are fewer interactions between them. RSM design yielded a maximum lipase activity of 73.3 U/mL. is was obtained in trial no. 13 and the levels of the independent variables were 5% oil (+1), 5.5 g/L (NH 4 ) 2 SO 4 (0), and 0.1 g/L CaCO 3 (−1). is is a 1.7-fold increase in comparison with the preoptimization maximum of 42.3 U. is goes to prove that the developed model has resulted in appreciable enhancement of enzyme production, thereby improving the feasibility of the process. Earlier, central composite design has been applied to optimize lipase production by the yeast Candida cylindracea using two different carbon sources and a yield of 17.30 U/mL using glucose as carbon source and 47.25 U/mL using olive oil as carbon source was obtained [27]. Lipase production by Candida rugosa has been optimized using RSM-based design and cheese whey was used as a substrate for this. From the experimental design it was concluded that a combination of brewery product, yeast and malt extract, Tween 80, and olive oil with cheese whey resulted in a 287% increase in enzymatic activity [28]. A medium comprised of soyabean meal 0.77% (w/v), (NH 4 ) 2 SO 4 0.1 M, KH 2 PO 4 0.05 M, rice bran oil 2% (v/v) CaCl 2 0.05 M, PEG 6000-0.05% (w/v), NaCl 1% (w/v), inoculum 1% (v/v), pH 3, incubation temperature 35 ∘ C, and incubation period of 5 days has been identi�ed to be the optimal conditions for lipase production by Aspergillus awamori [29]. In another study, a medium containing 0.45% (w/v) peptone, 0.65% (v/v) Tween-80, and 2.2% (v/v) inoculum has been recommended for a maximal lipase production of 20.6 U/mL by Candida cylindracea [30].
�.�.�. Model �eri�cation and Con�rmation. e objective of the validation study is to demonstrate that the polynomial expression could correctly predict and describe the response function. e validation report generated by the design expert soware is given in Table 4, which predicts a maximum lipase activity of 73.65 U/mL.

Conclusion
is work has culminated in the isolation of several fungal strains from dairy effluent, of which Fusarium solani SKWF7 exhibited pronounced extracellular lipolytic activity. e physical and chemical parameters in�uencing enzyme production by this isolate were varied one at a time and their impact was probed, revealing the ones that lead to a heightened enzyme production. e feasibility of using statistically based experimental design to simultaneously screen multiple variables and optimize fermentation conditions for increasing the enzyme yield was also demonstrated. Based on this design, 5% (v/v), 5.5 g/L, and 0.1 g/L were found to be the optimal concentrations of palm oil, (NH 4 ) 2 SO 4 , and CaCO 3 , resulting in a 1.7-fold increase in enzyme production. e model was statistically signi�cant as is evident from low and high values. us, it can be concluded that the model can potentially be applied for lipase production by Fusarium solani. Based on the promising results obtained in this shake-�ask study, production in a bioreactor and the utilization of cheap substrates derived from waste materials are to be taken up in future studies, with an objective of exploring scaleup feasibility, bringing down production costs and thereby augmenting the economic viability of the process.