Bacillus subtilis strain NCIM 2063, obtained from the National Collection of Industrial Microorganisms (NCIM, Pune, India), was used for microencapsulation using spray drying. The bacterial strain was stored at -80 °C in a cryovial and at 4 °C on agar plates. A pre-inoculum was formed from a frozen stock by transferring a single loop to nutrient agar medium (Torlak, Serbia), comprising 0.5% peptone, 0.3% beef extract, 1.5% agar, 0.5% NaCl, and incubating at 37 °C. Erlenmeyer flask containing 300 mL of nutrient broth (Torlak, Serbia) comprising 0.5% peptone, 0.3% beef extract, 0.5% NaCl and 0.03% KH₂PO₄, was inoculated with a single colony and grown for 24 h in a rotary shaker at 37 °C and 150 rpm to form inoculum, 1% of which was then used for bioreactor inoculation.

NCIM 2063 isolate was cultivated in a 2.5 L bioreactor (KLFM, BioEngineering, Wald, Switzerland) containing sterilized DSM medium: nutrient agar, 10% (w/v) KCl, 1.2% (w/v) MgSO₄ with addition of 1 mL of filter-sterilized solutions: 1 M Ca(NO₃)₂, 0.01 M MnCl₂ and 1 mM FeSO₄. The pH was adjusted by adding 0.01 M NaOH. The culture was grown for 24 h at constant conditions: 33 °C, agitation rate of 440 rpm, airflow rate of 0.3 vvm.

After 48 h of cultivation, fermentation broth containing B. subtilis NCIM 2063 culture was exposed to a thermal shock for 1 h at 54 °C. Fermentated broth samples were mixed with different amounts of maltodextrin (according to the Box-Benkhen experimental design) and incubated for 10 min to form a homogenous suspension. The initial concentration of viable cells was confirmed by plating on nutrient agar. Each suspension was spray dried in a laboratory scale spray drier Büchi mini B-290 (Flavil, Switzerland). The airflow rate and atomization pressure had constant values of 600 L/h and 0.55 bar, respectively. The inlet temperature and the feed flow rate were set according to the Box-Benkhen experimental design.

he optimization of the spray drying procedure was based on the Box-Benkhen experimental design coupled with response surface technology and Deringer’s desirability function (Derringer & Suich, 1980). Seventeen experiments were conducted in total, with 3 factors at 3 levels. Three independent process variables: maltodextrin concentration (10-50 g/L), inlet air temperature (110-140 °C), and feed flow rate (6-10 mL/min), were determined based on a literature search and preliminary experiments. Dependent variables: moisture content (%), product yield (%), and number of viable cells (CFU/g), were determined in triplicate; mean values were used for the regression analyses using Design Expert software package (Trial version 7.0.0, STAT-EASE Inc., Minneapolis, MN, USA). Experimental data were fitted to the model proposed by the software. Obtained model equation was used to calculate the values predicted by the model. Model adequacy, statistical significance and deviation from the experimental results were evaluated by ANOVA. Optimization of spray drying parameters was performed using Derringer’s desirability function, which is used in complex systems requiring simultaneous optimization of several factors. Based on a optimization criteria it transforms all the system responses on a scale from 0 to 1. By combining individual functions of desirable responses, the total function of preferred responses (D) is calculated with a maximum value ranging from 0 (which is an unwanted answer) to 1 (desired response), and then combined into a mutual response that should have a maximum value under previously defined optimization criteria. As the value of D is approaching 1, the system is being closer to the global optimum value (Derringer & Suich, 1980).

The number of viable microencapsulated cells was determined immediately after spray drying, after six months and one year of storage at room temperature, respectively. Powder formulation (0.1 g) was rehydrated in 0.9% saline solution (NaCl, distilled water), shaken vigorously, and allowed to rehydrate for 12 h. Viable cell concentration was determined using the spread plate method.

To determine the moisture content, 2 g of the powder formulation obtained after spray drying was transferred to an aluminum dish and dried at 105 °C to a constant weight. The residual moisture content was calculated according to the following equation:

where w_f and w_i are weighed mass of the formulation prior to and after drying at 105 ºC, respectively.

The yield of the powder formulation obtained after spray drying was calculated using the following equation:

where W_m is the weight of the recovered powder and Wp is the dry weight of the initial suspension in addition to maltodextrin.

Encapsulation efficacy was calculated using the following equation:

where N_r is log cfu/mL before spray drying and N_f is log cfu/mL after spray drying.

Powder formulation obtained under optimum conditions was analyzed for hygroscopicity, solubility dissolution time and morphology of the microparticles. Hygroscopicity and solubility were determined according to the method described earlier (Bakar et al., 2013), while the dissolution time was calculated as the time required for total dissolution of 1 g of powder in 50 mL distilled water using a magnetic stirrer at 829 rpm (Bhagwat et al., 2020). The morphology of the formulation obtained at optimum conditions was investigated using a JEOL JSM-6610LV scanning electron microscope (30 kV accelerating voltage) with an energy-dispersive X-ray spectrometer (SEM/EDS; XMax Large Area Analytical Silicon Drift connected with INCAEnergy 350 Microanalysis System). Dried powder was coated with a thin gold film at 20 kV voltage to obtain a higher quality secondary electron image for further SEM examination.

Microencapsulated B. subtilis NCIM 2063 formulation was evaluated for promoting growth of green pepper (Capsicum annuum). The experiments were conducted in growth chamber in Leskovac (Serbia) during March-April, 2019. Pepper seeds were disinfected by immersion in a mixture (1:1) of 30% hydrogen peroxide and 70% ethanol for 10 minutes, subsequently washed by immersion in distilled water several times. Seed inoculation was performed by immersion in a resuspended culture of B. subtilis NCIM 2063 with a cell density of 9.1 LOG (CFU/g), after which sowing was performed in previously sterilized plastic pots containing 200 g of dry sterilized fertile chernozem soil. Treatment was performed in 5 replications, 10 pots per replication. The seeds immersed in sterilized water served as a control. The plants were grown for 8 weeks in an incubation chamber at a constant temperature (22 °C), average relative humidity 50-60%, exposed to light of 140 µmol/m²·s wiht a 16 h photoperiod. Vegetative tissue was analyzed for: stem height, root length, number of leaves, leaves and root dry weight and leaf chlorophyll content. Dry weights were determined after drying plant tissues at 65 °C for 24 h, and then at 110 °C until constant weight (Garcia et al., 2011).

The leaf chlorophyll content was determined using the Hiscox & Israelstam´s (1979) method. Briefly, 100 mg of leaves were immersed in 7 mL of DMSO and incubated at 65 °C for 30 min. After adding DMSO to a total volume of 10 mL, 1 mL was transferred to a cuvette and the absorbance was measured at 645 and 663 nm (DMSO blank) using a UV/VIS spectrophotometer (UV/Vis Spectrophotometer - Pye Unicam Ltd, Cambridge England). The chlorophyll content was calculated using Arnon’s equations:

All experiments were performed in three parallel replications, and the results were presented as mean value of three repetitions ± standard deviation. The programs Origin 6.0, Excel 2013 and Expert Design 7.0 were used for statistical processing, modeling and graphical analysis of experimental data. Multicriteria optimization was performed by applying the Box-Benkhen experimental design, Response surface methodology and Deringer’s desirability function. The adequacy of the response surface model was assessed using the analysis of variance (ANOVA).

Various factors showed a significant effect on product viability, and thus the efficacy of encapsulation. The experimental data log values were subjected to a nonlinear regression analysis and fitted to a second-order polynomial model function that predicts the viability of cells in spray dried powder:

The results of B. subtilis viability affected by spray drying inlet temperature, feed flow rate and maltodextrin concentration are visually represented in Fig. 1 in the form of three-dimensional (3D) response surfaces. It can be seen that an increase in concentration of maltodextrin had a positive effect on the viability of microorganisms in the final product, regardless of inlet temperature. Although in most cases an increase in temperature reduced the cell viability, Fig. 1a clearely shows that increasing the feed flow rate can reduce the negative impact of temperature. Encapsulation efficacy undergoes the same dependancy pattern as viability. A maximum number of viable cells within the tested range (1.99·10⁹ CFU/g) and a maximum encapsulation efficacy (94.38%) were achieved at the lowest temperature (110° C), medium flow rate (8 mL/min) and the maximum concentration of maltodextrin (50 g/L).

The yield of dry powder after spray drying varied from 39.2 to 96.8%. The influence of selected variables on product yield was determined by the response surface method. Regression analysis of experimental data presents the response equation:

Based on response surface analysis (Fig. 2), it can be concluded that the maximum yield was achieved by a simultaneous increase of maltodextrin concentration and temperature. The influence of maltodextrin concentration was most pronounced at the highest tested temperatures, while at lower temperatures, a decrease in feed flow rate weakened the influence of maltodextrin. Increasing the feed flow rate generally reduced the yield of the product, but as the temperature increased, the influence of the flow rate on the yield decreased.

Quality formulation should contain less than 5% moisture after spray drying (Peighambardoust et al., 2011), although some authors indicate that up to 12% moisture is also acceptable for bacterial formulation. Low moisture content prevents contamination of the product and provides longer shelf life (Keswani et al., 2016). Hence, the combination of factors that reduce the moisture content was considered most desirable. The following model equation was proposed:

where Y is moisture content (%).

An increase in temperature and decrease in feed flow rate decreased the moisture content (Fig 3). However, the influence of maltodextrin concentration is insignificant. Although maltodextrin concentration dose did not affect moisture content significantly, the influence of maltodextrin concentration became more pronounced at a lower feed flow rate in the examined range. Hence, when maltodextrin concentration was increased along with the low feed flow rate, the moisture content gradually decreased. Still, at higher feed flow rates, maltodextrin did not play a crucial role for this system response.

Spray drying variables were optimized using Derringer’s desirability function in order to obtain a microbial formulation of maximum viability and yield with a minimum residual moisture.

According to desirability criteria (Table 3) the following conditions were proposed for spray drying of B. subtilis NCIM 2063 with D=0.87: temperature 133 oC, maltodextrin concentration 50 g/L and feed flow rate 6.5 mL/min.

The optimized solution (Fig. 4) proposed in this study was validated in laboratory conditions (Table 4), which confirmed the minimum deviation of experimental and software predicted values, as well as the high survival rate of encapsulated bacteria after being stored at room temperature for six months and one year.

The morphology of microcapsules (Fig. 5) can be described as amorphous glassy structure, varying from flat and deflated to round spherical particles with porous surface. The diameter of spherical particles obtained in this study ranged from 5 to 300 µm. The dimensions of the particles are not of crucial importance as they can be modifed subsequently, by granulation adjustment methods. The outer surfaces of the microcapsules were characterized by the presence of indentation, with a few cracks on the surface. On the other hand, variables such as hygroscopicity, dissolution time and solubility (Table 4) showed that the powder has enhanced stability and a potential to be stored long-term, which was also proven by high number of surviving cells after one-year period.

The effect of pepper seeds treatment with a spray-dried microbial formulation of B. subtilis NCIM 2063 on growth variables is shown in Table 5. The application of formulation resulted in a significant increase in the leaf dry weight and plant height, while root weight and leaf number did not differ significantly from the control. Particularly, plants treated with the formulation were about 40% higher and had almost 3 times larger weight of the leaves, than the control.

It is known that chlorophyll plays a key role in the process of photosynthesis and that its content is an indicator of the capacity of photosynthesis and plant health. Table 5 shows a significant positive effect of B. subtilis NCIM 2069 isolate on chlorophyll content in treated pepper seedlings. Compared to the control, the application of the formulation significantly increased the content of chlorophyll a (by 76.77%) and chlorophyll b (by 52.5%). This in turn contributed to an increase in the total chlorophyll content in peppers by 57.5% compared to the control.

Quadriatic model equation was found to be adequate and statistically significant for all three of the tested variables. Apart from F, p and lack of fit values, goodness of fit was also confirmed by an adequate mean relative percent deviation (MRPD) value. More than 94% of variation can be explained by the models and their adequacy was confirmed by the compliance of the predicted R² (Pred R²) and Adj R², which differ in less than 0.2. Adequate values of the coefficient of variation (CV) confirmed that the models are reproducible.

The linear effect of inlet temperature was found to be significant for all response variables at 99% confidence interval. This factor also demonstrated a strong interaction effect with the feed flow rate affecting viability and yield, along with maltodextrin concentration affecting moisture content. Among other interaction effects, the interaction of maltodextrin concentration with feed flow rate was found to be significant at a 95% confidence level for moisture content. Maltodextrin concentration also exhibited a positive linear effect on viability and yield, while the linear effect of the feed flow rate was found to be significant for the yield and the moisture. The positive linear effect of a dependent variable reveals that increase in its value also causes increase in the response. Although feed flow rate did not show statistical significance for viability as an individual linear term, squared and interacting with temperature, this factor nevertheless shows statistical significance at a 95% confidence level, which denotes that the optimum level is within the experimental region, and not at the extremes (Table 2) (Baş & Boyacı, 2007). Similarly, in the case of spray dried Bacillus thuringiensis, Adjallé et al. (2011) found that linear terms of temperature and feed flow rate had a statistically significant positive effect on the viability, their quadratic terms had a negative effect, while interaction terms were not statistically significant. Wang et al. 2018) also found that inlet air temperature and feed flow rate had the greatest effect on the survival of spray-dried Sphingomonas sp. in the formulation. The same factors, particularly their linear, quadratic and interaction effects were also found statistically significant by Seth et al. (2017) for the moisture content in the spray drying process of Streptococcus thermophilus and Lactobacillus delbrueckii bulgaricus.

Processes of mass multiplication and formulating procedures are the most important steps for the effectiveness of the obtained formulation. Therefore, to obtain a commercially competent, stable and effective product, the influence of crucial drying varibles on product properties should be carefully analyzed and explained (Vassilev et al., 2015). Product yield, the number of viable cells per gram of the product and mositure content are the three variables that denote the quality of the product: its potential to colonize the plant’s root, durability and economical profitability. Hence, spray drying should be conducted with a combination of variables that will enable high yield powder formulation with good cell viability and minimum moisture content.

Although in most cases an increase in temperature affects the reduction of cell viability, the research has shown that increasing the feed flow rate can reduce the negative impact of temperature. Namely, increasing the feed flow rate leads to a decrease in the temperature on the surface of the droplets, which causes changes in heat transfer thus reducing the physical damage to the bacteria cell membranes (Behboudi-Jobbehdar et al., 2013). In the present work, we noticed that the feed flow rate influence was more pronounced at lower and medium temperatures within the examined range. Nevertheless, feed flow rate values in the range 6-8 mL/min had practically no impact on the survival of microorganisms in the formulation when applied in combination with higher temperatures. Maxiumum number of viable cells achived in this study was 1.99·10⁹ CFU/g, which is consistent with previous results of spray drying B. subtilis CPA-8 when the viability of 3.3·10⁹ CFU/g was achieved at feed flow rate of 8 mL/min with the addition of MgSO₄ (Yánez-Mendizábal et al., 2012). The difference in the number of surviving cells can be attributed to the difference in carriers and the cultivation conditions applied before spray drying (Peighambardoust et al., 2011).

The results of this study indicate that the maximum yield was achieved by a simultaneous increase of maltodextrin concentration and temperature. Such effect is explained by a fact that a higher concentration of the carrier in the feed increases the viscosity, thus reducing the radial velocity of the particles. This in turn reduces the possibility of collision with the walls of the drying chamber and increases the number of particles reaching the collecting vessel which directly increseas the yield (Schuck, 2009). Decrease in feed flow rate also increases the yield of the product. The same effect was noticed for the moisture content of the formulation. High temperatures combined with low feed flow rate guarantee a low moisture content. An increase in inlet air temperature creates a larger temperature gradient between atomized particles and hot air, which enables a stronger driving force to remove moisture. That in turn reduces the likelihood of agglomeration of microparticles and the deposit formation on the walls of the drying chamber, which directly affects the increase in yield (Amiet-Charpentier et al., 1998; Adhikari et al., 2005). On the other hand, an increase in feed flow rate affects the formation of larger droplets, which makes evaporation more difficult, resulting in higher residual moisture and lower yield (Seth et al., 2017). This conclusion is consistent with previous spray drying results of Lactobacillus acidophilus (Behboudi-Jobbehdar et al., 2013) and Pseudomonas putida (Amiet-Charpentier et al., 1998) which state that the yield was increased significantly after a simultaneous increase in air inlet temperature and feed flow rate. Seth et al. (2017) also concluded that in the spray drying process of Streptococcus thermophilus and Lactobacillus delbrueckii bulgaricus, temperature and flow rate had the greatest influence on the moisture content, whose linear, quadratic and interaction effects were statistically significant. Other authors examined the influence of different carriers (skimmed milk and MgSO₄) on the moisture content of the B. subtilis CPA 8 formulation obtained by spray drying. The lowest moisture content (6%) was achieved at 150 °C and 8 mL/min by applying 10% skimmed milk powder and 10% MgSO₄ (Yánez-Mendizábal et al., 2012). Given that the lowest moisture content achieved in the present work was 1.4%, it is concluded that maltodextrin in combination with the high inlet air temperature achieves a significantly better effect.

B. subtilis NCNIM 2063 powder formulation had the best viability when spray dried on low temperatures and high feed-flow rates in the tested intervals. Contrary to that, high temperature and low feed flow rate valus are needed to obtain the highest yield and the lowest moisture content of the same formulation. To obtain products in satisfactory yield, viability and moisture content at the same time, multicriteria optimization was performed. The results of multicriteria optimization revealed that the ideal spray drying temperature value is 133 °C, while maltodextrin concentration and feed flow rate should be set to 50 g/L and 6.5 mL/min, respectively. A much higher temperature (180 °C) was used by Adjallé et al. (2011) for microencapsulation of B. thuringiensis by spray drying at a feed flow rate of 0.29 g/min, which resulted in 2.2·10⁸ viable cells in the dry formulation.

Behboudi-Jobbehdar et al. (2013) achieved the maximum desirability of Deringer’s function for spray drying of L. acidophilus at 133.54 °C at a feed flow rate of 7.14 mL/min, using maltodextrin in combination with glucose and whey protein as a carrier; under these optimal conditions, the viability of 8.59 CFU/g and a water activity of 0.171 were achieved. The differences in optimized values of spray drying are mainly attributed to differences in the used strains and their properties, which is why it is necessary to perform optimization for each microbial species separately. As far as Bacillus is concerned, one-factor optimization was performed by Ma et al. (2015) for spray drying of B. subtilis with maltodextrin and gum arabic; under the proposed optimal conditions (temperature 140 ºC and feed flow rate 9.1 mL/min), 4.83-5.35·10⁸ CFU/g was reached. Barcelos et al. (2014) dealt with spray drying of biosurfactants obtained from B. subtilis LBBMA RI4914, used as a carrier maltodextrin at a concentration of 250 g/L. Our study showed that it is possible to use lower inlet air temperatures in combination with a lower feed flow rate to achieve better viability of the bacterial culture. Significant savings can also be achieved from the aspect of consumed material, considering that significantly lower concentrations of maltodextrin are sufficient for the achievement of satisfactory results. Morphology of the particles is also in accordance with previous literature data. While the presence of indentation is usual for polysaccharide carries, cracks can be attributed to a long storage time (Campos et al., 2014). Maltodextrin as a carrier can particularly influence the morphology of the particles (Behboudi-Jobbehdar et al., 2013).

lity to adhere to the pepper seed surface and positively influence the growth of the seedlings in the early period. The positive effect of the isolate is a result of its ability to produce growth hormones and siderophores, affect nitrogen fixation, and phosphorus solubilization (Ben Khedher et al., 2021). Similarly, B. subtilis SL-13 isolate increased pepper leaf area (Tao et al., 2019), B. subtilis CAS 15 contributed to an increase in yield, height and weight of pepper (Yu et al., 2011), while B. subtilis V26 promoted the growth of potato seedlings (Ben Khedher et al., 2021). Yu et al. (2011) explained that microbial siderophores, which are produced by this microbe, stimulte plant growth as they increase the availability of the iron in the soil. Additionaly, B. subtilis produces indole acetic acid, which stimulates plant cell division and allows better adsorption of minerals and water, which in turn results in better growth parameters (Swain et al., 2007).

The increase in leaf dry weight observed in this study, is in a correlation with the increased content of chlorophyll a and b and the total chlorophyll in the leaves. The contents of chlorophyll a and b achieved in this study was almost 3.5 times higher than the one achieved by the isolate B. subtilis SL-13 (Tao et al., 2019). Multiple increases in chlorophyll content is explained by the ability of this isolate to modulate endogenous signaling molecules for glucose and abscisic acid, which have a regulatory role in the process of photosynthesis (Zhang et al., 2008). Very often, in addition to chlorophyll, an increase in content of other pigments, such as α-tocopherol and β-carotene can be observed (Sonbarse et al., 2020). A mechanistic study confirmed that applying B. subtilis to pepper seeds improves the efficiency of photosystem II and positively affects photosynthesis by increasing the amount of chlorophyll and the rate of carbon dioxide assimilation. Such an effect is explained by the increase in the electron transfer rate in the thylakoid membrane due to the action of the bacterium (Samaniego-Gámez et al., 2016).

In conclusion, this research proved that B. subtilis NCIM 2063 can be succesffuly microencapsulated with maltodextrin using spray drying procedure at following conditions: temperature 133 °C, maltodextrin concentration 50 g/L and feed flow rate 6.5 mL/min. Additionaly, the influence of spray drying process variables on the most important formulation characteristics is determined and explained, which greatly facilitates the selection of drying conditions with other plant growth-promoting microorgansms and encourages future research on this subject. Obtained formulation was stable over a 1-yer period and had several positive plant growth-promoting traits, facilitating the growth of green pepper seedlings in controled conditions. The next step of the research shall be the confirmation of the phytostimulatory effect in field conditions, after which commercialization of the product can be started.