Valorization of de-oiled cakes for xylanase production and optimization using central composite design by Trichoderma koeningi isolate

Aim: Evaluation of the medium components and conditions for the optimization of xylanase production in submerged fermentation by Trichoderma koeningi isolate (KT006533). Methods: Trichoderma koeningi was identified by the 18s rRNA. NaOH pre-treated corn cobs were used in order to reduce the crystallinity of corn cobs. De-oiled cakes proximate composition was analyzed according to AOAC. Plackett- Burman design (PBD) was used to screen the 19 media variables that affect xylanase production and optimized the media by central composite design (CCD). Results: The isolate was identified as Trichoderma koeningi . Among the 19 variables more effective variables on xylanase production was mustard cake, xylose, FeSO 4 ·7H 2 O, ZnSO 4 ·7H 2 O and pH. The optimization of xylanase and protein were quantified as 74.84 U/mL and 2.94 mg/mL, respectively, owing to the influence of the above components at g/L 3.5, 6.0, 0.06, 0.04 at levels respectively and pH 6.5, by using a CCD. The coefficient of determination (R 2 ) is 0.92%. The correlations between the actual and coded results are found to be good. Conclusion: NaOH pre-treated corn cobs were used with mustard cake powder as fermentation medium constitu-ent to induce the xylanase production. From the results we conclude that mustard cake induced the xylanase production.


Introduction
In general, the production of xylanase based on microbial fermentation and biosynthesis renders its industrial application more feasible and economical. The use of submerged fermentation (SmF) for induction of cellulolytic and hemicellulolytic enzymes provides a mean to use of various agro-industrial byproducts, such as wheat bran, wheat straw, corn cobs, sugar beet pulp, apple pomace, and cassava waste [1]. Filamentous fungi have been widely used to produce hydrolytic enzymes for industrial application, including xylanase, whose levels in fungi are generally much higher than in yeast and bacteria. The most common industrial xylanase producing strains are the Aspergillus sp. and Trichoderma sp., as well as bacterial strains of the species Bacillus sp. [2]. Xylanase from filamentous fungi are extra-cellular, inducible enzymes. The nutrient medium compositions and culture factors strongly influence the xylanase production.
The optimization of fermentation conditions and screening of important nutritional factors are of essential importance to determine the optimal variables for efficient production and to avoid the wastage of nutrients in the production. As a result, the production cost for xylanase should be significantly reduced. However, optimization of all factors and establishment the best possible conditions by considering all interaction of parameters, numerous experiments have to be carried out, which is not economical and practical. For this type of cases, design of experiments (DOE) and statistical tools help to gain more information about the optimization conditions within a few trials [3]. There are two ways to solve this problem: conventional and statistical methods of DOE [4]. Plackett-Burman design (PBD) is a powerful statistical technique for screening medium components and has been widely used in optimization [5]. Response surface methodology (RSM) is one such technique based on the fundamental principles of statistics, such as randomization, replication and duplication, which simplifies the optimization by studying the mutual interactions among the variables over a range of values in a statistically valid manner [6]. The application of statistical experimental design in fermentation improves the product yield, and reduces the process variability, development time and overall costs [7].
The present objective on ability of Trichoderma koeningi isolate to utilize inexpensive substrate and an attempt was made to evaluate the medium components with PBD and statistically optimize the medium under submerged fermentation by using the central composite design (CCD).

Microorganism and identification
The T. koeningi was isolate from the sorghum dumping soil using potato dextrose agar medium (PDA). The extraction of total DNA and its purification was followed by Purohit et al. [8]. Amplification of 581 base pair fragment within the gene coding for the small ribosomal subunit 18S rRNA of fungi was performed in a thermal cycler gene AMP cycler system using the fungal specific primers 18R5185′-GCATTTGCCAAGGATGTTTT-3′ and 18F5305′-TTCGTGCCAGCAGCCGCGG-3′. The purified and amplified 18S rDNA PCR products were sequenced by Xleris services (Ahmadabad, Gujarat, India). The resulted sequences were aligned using the Clustal W and then the sequence was compared with those from Gene Bank using BLAST [9]. The phylo-genetic tree was constructed by using joining bootstrap method [10]. The 18S rRNA partial sequences were submitted to gene bank database.

Substrate preparation
Corn cobs were collected from agro-fields and local market of Tirupati (India) and pretreated with 2% of NaOH for 90°C at 90 min [11]. Mustard, soybean and ground nut cakes procured from a local small scale oil extracting units, was used as the supplementation for xylanase production. They were dried at 60°C for 72 h to reduce the moisture content and ground to a desired particle size (1 mm).

Cultivation of fungal spores and inoculum preparation
Fungal spores were obtained by growing the T. koeningi on PDA (Himedia, India) slants at room temperature. They were harvested after 5 days of cultivation with saline containing 0.1% (w/v) Tween 80. The spore suspension at 1 × 10 8 spores/mL was used for inoculum preparation for SmF process.

Chemical composition of de-oiled cakes
It was analyzed according to method of AOAC as follows: Lipids content of the samples were determined by extracting the residue with petroleum ether (40-60°C) for 7 h in a Soxhlet apparatus. Crude fiber was determined as loss on dried lipid-free residue after digestion with 1.25% H 2 SO 4 and ash was determined by ignition at 550°C in a muffle furnace to a constant weight. Moisture was determined by drying is a oven at 105°C for 8 h. Total carbohydrate content was determined according to Dubois et al. [12].

Estimation of minerals in de-oiled cakes by ICP-OES
Elemental analysis was carried out on an optima 2100 DV inductively coupled plasma-optical emission spectrometer (Dual view, Perkin Elmer life and analyti cal sciences). 1 g of sample (mustard, soybean and ground nut cakes) were taken and digested with 1 mL of 0.2% (v/v) HNO 3 and centrifuged for 20 min at 2000 rpm (Hettich Universal 30F, Tuttlingen, Germany). All sample vials, sample cups and glass ware were cleaned by soaking in 10% (v/v) HNO 3 and rinsed with de-ionized water prior to use. The appropriate standards for each el ement were made within the concentration range of the elements in the samples. The results were obtained from triplicate measurements.

Submerged fermentation (SmF) for xylanase production
It was carried out in Erlenmeyer conical flasks (250 mL) that contained corn cobs (5 g) and the medium components of varying concentrations were added into the flask based on the PB design matrix basal medium. The contents were thoroughly mixed and autoclaved at 121°C for 20 min. The medium was inoculated with 5% (v/w) inoculum and incubated at 30°C for 7 days. The culture was incubated at 30°C on a rotary shaker at 200 rpm. The samples were withdrawn at regular intervals. The mycelium was removed by cooling centrifugation at 7000 rpm at 4°C for 15 min to obtain a clear supernatant. This supernatant was used for measurement of enzyme activities.

Optimization of selected variables for xylanase production by CCD
CCD with quadratic model was employed to study the combined effect of RSM [13], which was employed to optimize the selected five significant independent variables and three levels each with four concentric point combinations was used for the enhancement of dependent variable such as xylanase (U/mL) production. Actual and coded levels of variables along with response variables for CCD are presented in (Table 1). The overall second order polynomial mathematical relationship of the response Y1, Y2 and the five variables, can be approximated by the quadratic Eq. (1).
where Y is the predicted responses, X 1 , X 2 , X 3 , X 4 and X 5 are independent variables, b 0 is the offset term,  (5-15) were taken for optimization and for measuring the enzymatic activities. The design was applied for selection range of each variable was minimum and maximum. Total 50 experiments were designed by the model and performed in triplicate. The obtained results in the performed assays, executed in terms of xylanase production have been recorded. The model was validated by running the experiments within the experimental range. The data validation runs were also statistically analyzed to find out the correlation between observed actual and predicted values.

Protein estimation
The protein content in the culture filtrate was carried out by Folin-Ciocalteu reagent using bovine serum albumin (BSA) as standard [14].

Enzyme assays
Xylanase activity was quantitatively assayed in cell free supernatant [15]. The reaction mixture (2.0 mL) containing 1.0 mL of 1.0% (w/v) oat spelt xylan in 0.1 M citrate buffer, pH 5.0; 0.9 mL citrate buffer and 0.1 mL of a suitably diluted enzyme solution was incubated at 50°C for 5 min. The reaction was stopped by adding 3 mL of 1.0% dinitro-salicylic acid reagent. The reaction mixture was then kept on boiling water bath for 10 min. Amount of reducing sugars liberated was determined by measuring absorbance of the resulting color at 540 nm in a UV-vis spectrophotometer using xylose as standard [16]. One unit (IU) of xylanase activity was defined as the amount of enzyme that produced 1 μmol of xylose equivalent per min of reaction and per mL of enzyme solution under the assay condition. The controls without either enzyme or substrate were run simultaneously.

Fungal genome identification
It was carried out by 18s rRNA sequencing, and the sequence was obtained, followed to BLAST against the total fungal genome database in National Center for Biotechnology Information (NCBI). The BLAST result showed 98% identity with existence fungal species. According to Dhiman et al. [17] the maximum identity 1 or 2% difference between isolate and existed strains showed that isolate was considered as a novel organism and thus was found to be a new strain (named BROV1). Hence, this sequence was deposited in the GenBank and was provided with a new accession number (KT006533).

Evolutionary relationship
It was inferred using phylo-genetic neighbor-joining program, using bootstrap consensus test with 500 in MEGA 5 and the branch lengths are in the same as those of the evolutionary distances used to infer the phylo-genetic tree. The newly isolated strain was closely related to T. koeningi with 98% similarity. Based on this similarity the isolated new fungal strain was identified as a T. koeningi BROV1 (Figure 1).

Selection of media components by PBD
Conventional single dimensional search involves changing one independent variable at a time while fixing the A, B, C, D and E represents the process parameters denoted as X 1 , X 2 , X 3 , X 4 and X 5 , respectively, for regression equation.
others at a constant level, which gives unreliable results, inaccurate conclusion, and even frequent interactions of two or more factors. PBD is a powerful statistical technique for screening medium components in a shake flask and has been widely used in optimization of fermentation [2,5]. From the Pareto chart ( Figure 2), the nutrients, namely mustard cake, xylose, FeSO 4 ·7H 2 O, ZnSO 4 ·7H 2 O and pH were found to be significant for the production of xylanase by T. koeningi using corn cobs. Hence, these nutrients were selected for further optimization using RSM to maximize the production of xylanase. The results obtained in this study, the highest and lowest xylanase production were found to be 46.69 U/mL and 12.85 U/mL as observed in runs 4 and 20, respectively (Data not shown). The Penicillium sp.WX-Z1 achieved the maximum xylanase production of 29.72 U/mL [18], when compared to present results to earlier result highest production was observed in this study using PBD with different variables.

Optimization of xylanase production by CCD
The optimization (selected variables) was attempted to improve the enzymatic yield under SmF on corn cobs. SmF was carried out according to the experimental design model for 5 days. The influences of independent variables on xylanase and protein results were summarized in ( Table 2). The model was analyzed by multiple regression analysis and a second order polynomial equation (Eq. 1) was derived to represent the xylanase production as a function of the independent variables tested.   Statistical testing of the CCD model was performed with the Fisher's statistical test for analysis of variance (ANOVA) using design expert software and the results of ANOVA for xylanase activity and protein is shown in ( Table 3). The F-value is the ratio of the mean square due to regression to the mean square and indicates the influence of each controlled factor on the tested model. ANOVA of the quadratic regression model suggest that the model is significant with a computed F-value of 15.84 and P > F lowers than 0.05. The model determination coefficient R 2 (0.9269) suggested that the fitted model could explain 92.69% of the total variation and unable to explain only 7.31% of the total variation. This implies a satisfactory representation of the process by the model. The "Pred R-Squared" of 0.7076 is in reasonable agreement with the "Adj R-Squared" of 0.8684. The determination coefficient value is closer to 1.0, the better is the correlation between the observed and predicted values and the R 2 value obtained indicated a better correlation. A lower value  (Design-expert, 2012). The natural logarithm (ln) of the residual SS (sum of square) against λ is one, dip suddenly with a minimum in the region of the best optimum value 0.95, 0.62 and 0.48. The data do not require a transformation, as current value of confidence interval it contains (λ) very close and near to the optimum value. The 3D response surface plots ( Figure 3) were obtained by plotting the response (enzyme activity) on Z-axis against any two variables while keeping other variable at its '0' level. The iso-response contour and surface plots for the optimization of conditions for enzyme activity was follows.

Effect of mustard cake and xylose on xylanase production
The data on xylanase production as influenced by mustard cake and xylose levels along with the variables such as FeSO 4 ·H 2 O, ZnSO 4 ·H 2 O and pH are presented in (Figure 3A). Increased xylanase production was observed at FeSO 4 ·H 2 O 0.06, ZnSO 4 ·H 2 O 0.04 and pH 6.5, respectively. The production of xylanase significantly influenced by the mustard cake (p < 0.005) and the mustard cake (p < 0.005) in interaction terms. The mustard cake (p < 0.005) and xylose were (p < 0.005) significant in terms of squared terms. The negative interactive coefficient (Eq. 2) suggests that decreasing the level increases the enzyme activity while mustard cake at 3.96 g/L.

Effect of mustard cake and FeSO 4 ·7H 2 O on xylanase production
The interaction effects plotted for mustard cake and FeSO 4 ·H 2 O show that there are no significant interactions between these variables that affect xylanase yield ( Figure 3B). The enzyme production was influenced by the KH 2 PO 4 in linear term (p > 0.05) and MgSO 4 (p > 0.05) though their individual p-value is not significant. However this confirmed that optimal mustard cake range lied between 2 and 3.5 g/L and the optimal FeSO 4 ·H 2 O was between 0.04 and 0.06 g/L. The ( Figure 3B and Table 2) show that amount of xylanase production under these conditions were 17.83 U/mL. The remaining variables at their optimal level support the enzyme production.

Effect of mustard cake and ZnSO 4 ·7H 2 O on xylanase production
The enzyme production was influenced by the mustard cake in linear term (p > 0.05) and ZnSO 4 ·7H 2 O (p > 0.05) though their individual p value is significant and also both KH 2 PO 4 and MgSO 4 (p < 0.005) are having significant squared terms and they showed significant interaction between these two variables that should influence the xylanase production ( Table 2). The positive interactive coefficient (Eq. 2) of these two variables suggested that the enzyme activity was significantly influenced by these variables by fixing other variables at their fixed levels. The Figure 3C shows that optimal xylanase activity could be obtained when the mustard cake was in the range of 2-3.5 g/L, and the ZnSO 4 ·7H 2 O levels between 0.02 and 0.04 g/L, at the optimal level. The maximum yield of enzyme activity was 50.64 U/mL.

Effect of mustard cake and pH on xylanase production
The Figure 3D shows interactive effect of mustard cake and pH level on xylanase yield. Increase in mustard cake level up to 3.5 g/L and pH up to 6.5, improved the xylanase yield at the optimum variables were xylose 6 g/L, FeSO 4 ·7H 2 O 0.06 g/L and ZnSO 4 ·7H 2 O 0.04 g/L. The linear terms and squared terms of mustard cake and pH are positive and also the positive interactive coefficient (Eq. 2) of these variables suggested the significant increase in the xylanase yield. The Table 2 shows the actual yield was 44.83 U/mL, with the predicted xylanase yield of 40.82 U/mL, at optimum level.

Effect of xylose and FeSO 4 ·7H 2 O on xylanase production
Regardless of FeSO 4 ·7H 2 O, the maximum xylanase yield was obtained in between 0.04 and 0.06 g/L of xylose and concentration variations in FeSO 4 ·7H 2 O did not affect the xylose optima between 4 and 6 g/L, It confirms the lack of interaction between these parameters. The negative interactive coefficient (Eq. 2) suggests that decreasing the level of enzyme yield while the FeSO 4 ·7H 2 O in negative form and also squared terms of these variables. The Figure 3E shows the maximum actual yield of xylanase was 11.94 U/mL and predicted yield of xylanase was 18.82 U/mL, at other variables at their optimum conditions.

Effect of xylose and ZnSO 4 ·7H 2 O on xylanase production
The enzyme production was affected by xylose in linear term (p > 0.05) and ZnSO 4 ·7H 2 O (p > 0.05) though their individual p value was significant. However, both xylose and ZnSO 4 ·7H 2 O are having not significant in squared terms and they showed significant interaction between these two variables should influence on the xylanase production. The positive interactive coefficient (Eq. 2) of these two variables suggested that the enzyme activity was significantly influenced by these variables by fixing other variables at their fixed levels. The Figure 3F shows that optimal xylanase activity could be obtained when the xylose was in the range of 4-6 g/L, and the ZnSO 4 ·7H 2 O levels between 0.04 and 0.06 g/L, at the optimal level. The maximum actual yield of enzyme was 74.84 U/mL and predicted yield of xylanase was 79.91 U/mL. The other variables tested were at their optimal levels. Xylose could induce xylanase production by Trichosporon cutaneum SL409 and Aspergillus nidulans, respectively [19,20]. The results are observed in the present work was high when compared to T. longibrachitum, and also with Penicillium sp. [21].

Experimental validation
The adequacy of the model was validated by performing a total five verification experiments within the experimental range is given in (Table 4). The data of the validation runs were also statistically analyzed to find out the correlation between actual observed and predicted values. The determination of coefficient (R 2 ) of experimental and predicted values was found to be 0.9651 indicating that the group of experimental values is in good agreement with that of predicted, showing thereby the accuracy of the model.

Composition of de-oiled cakes
The proximate composition of different de-oiled cakes used in the present study was depicted in the (Table 5).
From the data it is observed that soybean, mustard and ground nut cake contained crude protein 35.56 ± 0.05, 41.03 ± 0.06 and 21.64 ± 0.02%, respectively, whereas crude fiber 7.18 ± 0.02, 10.63 ± 0.05 and 21.64 ± 0.02%, respectively. Similarly the crude fat present in the soybean, mustard and ground nut cake is 21.5 ± 0.04, 18.39 ± 0.06 and 17.63 ± 0.02%, respectively, whereas total carbohydrates are 29.72 ± 0.06, 32.73 ± 0.02 and 27.31 ± 0.04%, respectively. Crude protein was high in mustard cake than the others, fat and fiber values are lower in ground nut cake, followed by soybean cake and mustard cake. The total carbohydrate content is higher in mustard cake compared to soybean and ground nut due to the higher cellulose content of mustard cake. Similarly the nitrogenous substance i.e. crude protein also high in mustard cake, this resembles that mustard cake increases the xylanase production. There are several reports describing production of various enzymes using oilseed cakes as supplements in the production medium. Oil cakes are ideally suited nutrient support in SmF rendering both carbon and nitrogen sources, and reported to be good substrate for enzyme production using fungal species. This is due to the fact that oilseed cakes are among the agricultural residues commonly used for xylanase production, since they contain some residual nutrients that can serve as carbon, nitrogen inducer sources, and have been reported to be good substrates for microbial enzyme production [22]. Canola oil seed cake (CaOC) has been used as a substrate for xylanase production by Trichoderma reesei. The results suggested that xylanase yields were better in CaOC than from Solka-Xoc, xylan or glucose. The enzyme system produced using CaOC also contained a higher proportion  of acetyl-xylan esterase, cellulase, and xylosidase activities [23]. These results indicate that xylanase production was enhanced by some oil cakes. Generally the carbohydrates stimulated the growth of fungi [24]. In the present investigation the de-oiled mustard cake contained higher amount of total carbohydrates (32.73 ± 0.02% w/v) than the other cakes that served as growth promoters and supported the production of xylanase in the medium along with the corn cobs. The mineral content of the de-oiled cakes is shown in Figure 4. The data showed the presence of nitrogen, calcium, magnesium, manganese, zinc, and iron (1.84, 1.09, 0.69, 0.71, 0.96, and 1.52 mg/100 mg, respectively) in mustard cake. The phosphorus (1.01 mg/100 mg) content was high in ground nut cake and copper (0.64 mg/100 mg) was high in soybean cake. Although ions of Mg, Ca, Fe, Mn and Zn were chosen in the PBD, the result showed that only Fe 2 + and Zn 2 + had a significant impact on xylanase production. Some researchers have indicated the importance of Mg 2 + as a trace element in xylanase production by Aspergillus fischeri Fxn 1 [25]. Zn 2 + , Mg 2 + and Fe 2 + could also stimulate the yield of xylanase by Streptomyces thermodiasticus [26]. In addition to these elements, the mustard cake having the higher amounts of Zn and Fe appear to support xylanase production indirectly.

Conclusions
Utilization of cheaply available lignocellulosic substrates such as corn cobs for fermentation process would eventually reduce the environmental problems, which may be caused by their accumulation. The use of by-products of oil-seed processing industry such as cakes in fermentation medium will also reduce the cost of microbial products. The effect of cakes (mustard, soybean and groundnut) in SmF was evaluated, and optimization to maximize the xylanase production was carried out. Optimization of these parameters had a significant effect on the enzyme production. The actual optimum experimental value of xylanase activity of 74.84 U/mL is a good fit with the predicted value of 79.91 U/mL, which showed the model accuracy.