Optimization of cultivation medium composition for production of bioactive compounds effective against Penicillium sp.

SUMMARY Biological control is one of the best alternatives to pesticides as it avoids their weak points in plant disease control. In this study, the composition of cultivation medium for production of bioactive compounds by Bacillus subtilis ATCC 6633 was optimized. The produced bioactive compounds were tested against a phytopathogenic Penicillium sp. known for infesting different agricultural products and causing substantial crop losses. Antimicrobial activity assaying was carried out using the diffusion-disc method, and inhibition zone diameters were measured as direct indicators of antifungal activity. The response surface methodology (RSM) was used to evaluate the effects of different contents of initial nutrients (glycerol, NaNO 2 and K 2 HPO 4 ) in cultivation medium on inhibition zone diameter. Optimization was carried out using the desirability function method in order to maximize bioactive compounds yield and to minimize residual nutrients contents. The optimized concentrations of the selected nutrients in cultivation medium for production of bioactive compounds were: glycerol 20 g/l, NaNO 2 1 g/l and K 2 HPO 4 15 g/l.

Optimization of cultivation medium composition for production of bioactive compounds effective against Penicillium sp.

INTRODUCTION
Different phytopathogens infect fruit and vegetables during their growth, harvest, transport or storage. Phytopathogens cause economic losses in terms of yield decrease, and raise human health concerns due to harmful phytopathogen metabolites that remain in fresh fruit and vegetables, as well as in different products obtained by fruits and vegetables processing. One of common fruit and vegetable phytopathogens is the genus Penicillium, known for infesting citrus fruits (Nunes et al., 2009), apples (Quagliaa et al., 2011), tomato (Kalyoncu et al., 2005) and other agricultural crops. Crop losses due to diseases caused by Penicillium spp. reach up to 50% of crop yield (Mari et al., 2002). Synthetic pesticides are still the most common agents for treatment or prevention of fruit and vegetables diseases (Spadaro & Lodovica Gullino, 2004). Commercial pesticide formulations used for controlling Penicillium plant diseases contain active substances such as thiabenzadole, thiophanate-methyl, pyrimethanil and iprodione (Quagliaa et al., 2011). Along with their high cost, environmental pollution and negative effects on autochthonous organisms in soil, concerns have recently emerged about their insufficiently examined effects on human health (Janisiewicz & Korsten, 2002;Grahovac et al., 2009). Phytopathogens can also develop resistance to these synthetic agents after a period of repeated applications (Joshi et al., 2008), creating a need either for higher concentrations of pesticides or a different strategy of phytopathogen control. All of these listed reasons suggest a need for new methods to be found for fruit and vegetable disease control. One of such methods, avoiding the disadvantages of pesticides, is biological control, which consists of using different antagonistic microorganisms and their metabolites for plant disease control (Droby et al., 2009). The mechanisms by which microbial antagonists suppress plant diseases can be different, and the most common mechanisms include the production of bioactive antimicrobial compounds and competition for nutrients and space (Sharma et al., 2009). Generally, the market share of biopesticides is approximately $2-3 billion, compared to the synthetic pesticides market share of $56 billion, and its projected annual growth rate is more than 15%. Microbial biopesticides account for approximately 85% of biopesticides market. There are several commercial products that can be used for biological control of Penicillium spp.: Biosave™ 10LP and Biosave™ 11LP (JET Harvest Solution, Longwood, FL, USA), Serenade™ (AgraQuest, Davis, CA, USA), YeldPlus™ (Anchor Yeast, Cape Town, South Africa), Shemer™ (Bayer CropScience, AG), etc. (Quagliaa et al., 2011).
Soil bacteria of the Bacillus genus are well known and have been widely studied as possible agents for biological control of different plant diseases caused by microbial pathogens. Bacillus subtilis is one of the most commonly used biocontrol agents with proven antagonistic effect against various phytopathogens (Gisi et al., 2009). Some of these antagonistic microorganisms are naturally present on fruits and vegetables infected by phytopathogens or in soil, but these microorganisms are not normally able to produce sufficient amounts of antimicrobial compounds to suppress pathogen growth. Therefore, after isolating these antagonistic microorganisms they can be used for large-scale bioactive compounds biosynthesis by employing an appropriate production medium under defined conditions. Cultivation medium composition has great impact on the biomass growth and type and yield of synthesized metabolites (Ibrahim & Elkhidir, 2011), as well as overall process cost. Consequently, optimization of medium composition according to specific productive microorganism nutrition requirements is a critical factor for economically effective production of functional bioactive formulations (Managamuri et al., 2016). Optimization of medium composition, i.e. appropriate selection of nutrient sources (mostly of carbon, nitrogen and phosphorus) and precise defining of their concentrations, is the main method of directing metabolic activity of productive microorganisms towards biomass growth or synthesis of metabolites with antimicrobial activity (Sanchez & Demain, 2002). It is also important to optimize medium composition in terms of nutrient quantities that remain in cultivation broth after the biosynthesis in order to reduce the cost of effluents treatment and environmental pollution (Rončević et al., 2014).
According to literature data, glycerol is a very good carbon source for the biosynthesis of antimicrobial compounds by B. subtilis (El-Bana, 2005). Furthermore, as a consequence of increased biodiesel production, a request has emerged in recent years for investigating possible applications of waste glycerol as a carbon source for different microbial bioconversions (Li et al., 2013). Waste glycerol utilization in bioprocesses that result in obtaining value-added products, e.g. antimicrobial compounds, is a good way to reduce production costs and prevent waste glycerol disposal in the environment (Yang et al., 2012). Regarding nitrogen and phosphorus sources, nitrites and phosphates have been shown as appropriate for biosynthesis of antimicrobial compounds by B. subtilis (El-Banna & Quddoumi, 2007).
The aim of this study was to optimize the cultivation medium composition regarding glycerol, sodium nitrite and phosphate contents for the production of bioactive compounds with antifungal activity against Penicillium sp., using the response surface methodology (RSM) and desirability function method. Biosynthesis of bioactive compounds was carried out by Bacillus subtilis ATCC 6633.

Microorganisms
In this study, B. subtilis ATCC 6633 was used as a productive microorganism for the biosynthesis of antifungal compounds which were tested against a Penicillium sp. isolated from the environment. Both microorganisms were stored at 4ºC and subcultured at four-weeks interval.

Inoculum preparation and biosynthesis conditions
Inoculation was performed by adding 10% (v/v) of inoculum, prepared under aerobic conditions at 28ºC over 48 h by mixing on a laboratory shaker (Ika® Werke IKA® KS 4000i control, Germany) at 150 rpm. The production of bioactive compounds with antifungal activity was performed in Erlenmeyer flasks (300 ml) containing 100 ml of appropriate medium according to the experimental design. The biosynthesis of antifungal compounds was carried out under aerobic conditions at the temperature of 28ºC and agitation rate of 150 rpm on the laboratory shaker for 96 h.

In vitro assaying for antifungal activity check
Production of bioactive compounds was estimated in vitro by the diffusion-disc method (Bauer et al., 1966) and expressed as antifungal activity against the test microorganism presented by inhibition zone diameter (mm). Cultivation broth samples used in each experiment were concentrated by evaporation on a rotary vacuum evaporator (MRC ROVA-100, Israel) to one tenth of their initial mass and then their antifungal activities were tested against the test microorganism. Penicillium sp. was grown on a commercial medium (Sabouraud maltose agar, Himedia, India) at 28°C and inhibition zone diameters were measured after 48 h.

Determination of residual nutrients contents in cultivation media samples
After the end of biosynthesis, samples of cultivation media were centrifuged at 10000 rpm for 15 min (Eppendorf Centrifuge 5804, Germany). Only the liquid phase of cultivation media was used in further examination. The obtained supernatants were filtered through a 0.45 μm nylon membrane (Agilent Technologies, Germany) and filtrates were analyzed by the HPLC (Thermo Scientific Dionex UltiMate 3000 series, California, USA) to determine residual glycerol content. The HPLC instrument was equipped with an HPG-3200SD/RS pump, WPS-3000(T)SL autosampler (10 μl injection loop), ZORBAX NH 2 (250 mm x 4.6 mm, 5 μm) column (Agilent Technologies, Germany), and a refractive index detector (ERC RefractoMax520, Germany). Acetonitrile (70%, v/v) was used as eluent at a flow rate of 1.0 ml/min and elution time of 20 min at the column temperature of 30ºC. The Kjeldahl method (Herlich, 1990) was used for determining the total nitrogen residual content, while the residual content of total phosphorus was determined by spectrophotometric analysis (Gales et al., 1966).

Experimental design and optimization by RSM
Experiments were carried out according to the Box-Behnken experimental design with three factors at three levels and three repetitions at the central point, as presented in Table 1. The examined factors and their values (g/l) were: X 1 -glycerol content (20-50), X 2 -NaNO 2 content (1-3) and X 3 -K 2 HPO 4 content (5-15). Experimental results were fitted into the polynomial models of second degree that describe selected responses [Y 1 -inhibition zone diameter (mm), Y 2 -residual glycerol content (g/l), Y 3 -residual total nitrogen content (g/l) and Y 4 -residual total phosphorus content (g/l)]: where b 0 represents the intercept, b i represents the linear, b ii 2 quadratic and b ij interaction regression coefficients. Statistical analyses of the experimental results were performed using Statistica software v. 12.0. The same software was used for generating response surface plots, drawn for a constant value of one of the factors, while the remaining two factors were varied. Optimization of the examined factors according to the selected optimization aims was performed using the desirability function method (Design-Expert 8.1 software).

RESULTS AND DISCUSSION
Biosynthesis of bioactive compounds was carried out using B. subtilis ATCC 6633, and cultivation media prepared according to the Box-Behnken experimental design, under previously defined cultivation conditions. After biosynthesis, the samples of cultivation media were analysed and antifungal activity of each cultivation broth was examined against a Penicillium sp. In order to investigate the effects of chosen factors (glycerol, NaNO 2 and K 2 HPO 4 content) on appropriately selected responses (inhibition zone diameter, residual glycerol content, residual total nitrogen and residual total phosphorus contents), four regression equations were established based on the experimental results. The significance of the obtained models and regression coefficients was

Statistical analysis of experimental results
According to the Box-Behnken experimental plan, the following responses were examined: Y 1 -inhibition zone diameter (mm), Y 2 -residual glycerol content (g/L), Y 3 -residual total nitrogen content (g/l), Y 4 -residual total phosphorus content (g/l). Regression coefficients of mathematical models for each response obtained by regression analysis and their p-values are given in Table  2. Regression coefficient p-values below 0.05 indicate that the regression coefficient is statistically significant at the confidence level of 95%. Significant regression coefficients are bolded in Table 2.
The results of an analysis of variance for the selected responses are presented in Table 3. High values of the coefficients of determination for each response indicate good fit of experimental data to the obtained regression equations. The obtained F-values shown in Table 3 and p-values less than 0.05 for each response indicate that the models for selected responses are statistically significant.

Model for inhibition zone diameter response
Inhibition zone diameter is a major indicator of the amount of synthesized bioactive compounds, and it is therefore the most significant response for antifungal compounds production. The ranges of the selected nutrients contents were chosen according to literature data. Various data about carbon source concentrations in cultivation media could be found, ranging from approximately 8 g/l (Cheng et al., 2011) to 50 g/l (de Faria et al., 2011). The range examined in this study (20-50 g/l) was set around the central point of 35 g/l of glycerol, in conformity with a study conducted by El-Banna (2005). Concentration range of the inorganic nitrogen source, i.e. NaNO 2 (1-3 g/l), was selected in the same way (El-Banna & Quddoumi, 2007). Regarding phosphorus source concentration, literature data range from very low (Mnif et al., 2012) to very high concentrations (El-Bana, 2005), depending on bioproduct type. Therefore, phosphorus content range in this study was set to 5-15 g/l.
In order to better understand the effects of the examined factors and their interactions on inhibition zone diameter, response surface plots were generated (Figures 1, 2, 3). The response surface plots show the effects of two examined factors on the selected response, while the third factor remains constant and has the value of the central point of experimental design. Figure 1 presents the effects of initial glycerol and NaNO 2 contents on inhibition zone diameter, while the initial content of K 2 HPO 4 was 10 g/l. The results presented in the figure show that, with the initial glycerol content remaining constant, increase in initial NaNO 2 content leads to inhibition zone diameter increase. Reversely, when the initial NaNO 2 content is constant, the initial glycerol content increase of over 40 g/l leads to inhibition zone diameter decrease. As the results presented in Figure 1 show, inhibition zone diameter was at a maximum when the initial content of NaNO 2 in cultivation medium was maximal (2.7-3.0 g/l), and initial glycerol content ranged 30-40 g/l. The presented results indicate that a large amount of nitrogen source is essential for biomass growth or bioactive metabolite(s) synthesis under given cultivation conditions. Regarding glycerol content, the presented results are in accordance with the initial carbon source concentration reported by El-Banna (2005) and Issazadeh et al. (2012). The effects of initial glycerol and initial K 2 HPO 4 contents on inhibition zone diameter under constant initial NaNO 2 content of 2 g/l are presented in Figure 2. Under constant initial glycerol content, initial K 2 HPO 4 content increase leads to inhibition zone diameter increase. As previously, an increase in initial glycerol content of over 35 g/l leads to inhibition zone diameter decrease when the initial K 2 HPO 4 content is constant and near its maximum value. Maximum inhibition zone diameter was obtained at maximum K 2 HPO 4 content and glycerol content within a range of 20-35 g/l. An analysis of the inhibition zone diameter model using response surface plots presented in Figures 1 and 2 indicates that initial glycerol contents exceeding approximately 35 g/l inhibit the biosynthesis of bioactive compounds effective against Penicillium sp., leading to inhibition zone diameter decrease. The results also indicate that a large amount of phosphorus source is essential for the biosynthesis of antifungal compounds. As limiting phosphorus source content in cultivation medium is one of common ways to direct the metabolism of productive microorganisms towards synthesis of secondary metabolites (Martin & Demain, 1980), the large amount of phosphorus source that B. subtilis ATCC 6633 required in this study can indicate that the bioactive agent is not a product of secondary metabolism. Similar initial K 2 HPO 4 contents in cultivation media for B. subtilis have been used by different researchers (de Carvalho et al., 2010;de Sousa et al., 2014). Figure 3 shows the effects of initial NaNO 2 and initial K 2 HPO 4 contents on inhibition zone diameter under constant initial glycerol content of 35 g/l. The results presented in the figure show that when initial NaNO 2 content is constant, increase in initial K 2 HPO 4 content leads to an inhibition zone diameter increase. Also, increase in initial NaNO 2 content under the constant initial content of K 2 HPO 4 leads to an inhibition zone diameter increase. Therefore, simultaneous increase in initial NaNO 2 and initial K 2 HPO 4 contents leads to maximum inhibition zone diameter. The fact that phosphorus and nitrogen sources are required in large amounts for bioactive compounds biosynthesis support an assumption that either the biomass of B. subtilis ATCC 6633 or the primary metabolite(s) were the antifungal agent(s) since phosphorus and nitrogen are the key nutrients required by the productive microorganism during the exponential growth phase (Martin & Demain, 1980;Sanchez & Demain 2002).

Optimization of cultivation medium composition for biosynthesis of antifungal compounds
The main aim of optimization of cultivation medium composition is to obtain an economically viable medium that completely satisfies the nutritive requirements of a productive microorganism for biosynthesis of large amounts of a desired product along with maximum nutrients utilization and minimum environmental pollution. The main goal of biosynthesis of antifungal compounds effective against Penicillium sp. is to maximize the desired product yield. Since the main indicator of the amount of synthesized bioactive compounds is inhibition zone diameter, the first optimization set had maximal inhibition zone diameter as the goal. Optimization results, optimal values of the examined factors and predicted values of analyzed responses are presented in Table 4.
In the first optimization set, maximum inhibition zone diameter of 32.11 mm was predicted at the initial glycerol content of 26.74 g/l and maximum initial NaNO 2 and K 2 HPO 4 contents (3 g/l and 15 g/l, respectively). The desirability function value of 0.96 shows that the obtained results are in a very good agreement with the assigned optimization aim. On the other hand, these initial nutrients contents in cultivation media resulted in very high predicted residual contents of nutrients after biosynthesis, i.e. 18.50 g/l for glycerol, 0.47 g/l for residual total nitrogen and 0.74 g/l for residual total phosphorus.
Another objective of optimization of cultivation medium composition is to formulate a medium with optimal contents of carbon, nitrogen and phosphorus sources, and to achieve maximal utilization of nutrients by the productive microorganism. Minimal amounts of unutilized nutrients reduce the process cost, as well as the cost of necessary effluent treatment and environmental pollution. Another optimization set was made in order to minimize residual contents of the main nutrients (glycerol, NaNO 2 and K 2 HPO 4 ) after biosynthesis.
The second set of optimization results shows that the optimum values of the examined factors were glycerol 20 g/l, NaNO 2 1 g/l and K 2 HPO 4 15 g/l ( Table 4). The second optimization data set shows that minimization of initial glycerol content from 26.74 g/l to 20 g/l, as well as initial NaNO 2 content to its minimum examined value, led to a reduction in predicted residual glycerol, total nitrogen and total phosphorus contents of 98.59%, 44.68% and 71.62%, respectively, and to an inhibition zone diameter decrease of 9.16%, compared to the first optimization set. The results show that optimization of cultivation medium composition can significantly contribute to a reduction in bioprocess costs. Optimal values of the initial nutrients contents near minimum values of the examined range reduce the cost of cultivation medium, while the cost of necessary effluent treatment could be decreased by minimizing the residual nutrients contents that remain in cultivation broth after cultivation. The desirability function value of 0.88 in the second optimization set indicates that the obtained results fulfill the assigned optimization aims.

Validation experiment
Cultivation media with optimal contents of the selected factors from the second optimization set (glycerol 20 g/l, NaNO 2 1 g/l and K 2 HPO 4 15 g/l) were prepared and biosynthesis of antifungal compounds by B. subtilis ATCC 6633, as well as antifungal activity assaying against Penicillium sp. and analysis of cultivation broth samples after biosynthesis, were performed under the same conditions as in the optimization experiments. The experiment was conducted in triplicate tests and the mean value of each response was calculated along with standard deviation determination. The observed inhibition zone diameter was 28.83±1.04 mm, which is a very good agreement with the predicted value of 29.17 mm. Residual glycerol, total nitrogen and total phosphorus contents of 0.29±0.04 g/l, 0.28±0.03 g/l and 0.21±0.01 g/l, respectively, also showed very good agreement with the predicted values of selected responses. The experimental results presented above show that the established models for selected responses are statistically significant and could be used for further development of the bioprocess for production of antifungal compounds effective against Penicillium sp. by B. subtilis ATCC 6633.

CONCLUSIONS
The results of this study indicate that the RSM can be successfully employed for optimization of cultivation medium composition aimed for the production of bioactive compounds by B. subtilis with antifungal activity against Penicillium sp. Glycerol (carbon source) content is optimized to the minimum value of a selected range, which indicates that, in case of biomass synthesis, the optimal ratio between carbon and nitrogen source contents would be 20:1, and optimal ratio between carbon and phosphorus source contents 1.33:1. The results of this study indicate a need for further examination of mechanisms by which bioactive compounds of B. subtilis affect Penicillium sp. Also, this study presents a basis for further development of commercial biocontrol products based on microbial antagonists, and a need for determining the optimal concentration of the antagonistic microorganism, as well as an appropriate formulation of the final product.