Optimization of antifungal activity by Bacillus subtilis isolate CCASU 2021-4 using Response Surface Methodology

Fungal infections represent an enormous load on the public especially with the development of resistance to the most currently used antifungal drugs in practice. In the present work, a bacterial isolate coded A3 was recovered from soil and was shown to express antagonistic activity against Candida (C.) albicans ATCC 10231. This isolate was identified as Bacillus (B.) subtilis subsp. spizizenii isolate CCASU 2021-4 using 16S ribosomal RNA sequencing. D-optimal design from response surface methodology (RSM) was used to optimize the environmental variables affecting the antifungal activity of the respective isolate. The optimum conditions were a temperature of 30 °C, a pH of 8, and an inoculum size of 5 x 10 7 cfu/mL, resulting in an enhancement in the antifungal activity by 1.2 fold. This is the first report, to the best of our knowledge, on an antifungal activity from B. subtilis subsp. spizizenii culture broth against human fungal pathogens along with its optimization through RSM.


INTRODUCTION
Fungal diseases have greatly risen with the increase in patients with malignancies, immunocompromised patients, as well as those suffering severe illnesses. Candida species are responsible for a lot of these infections [1,2]. About 20% of intensive care unit infections were attributed to invasive candidiasis [3] while opportunistic infection by C. albicans represents about 95% of oropharyngeal candidiasis in human immunodeficiency virus (HIV)-positive patients [4]. The great increase in the appearance of drug-resistant Candida isolates represents a serious threat to humans [5, 6]. Recently, pharmaceutical practices consider combating fungi/yeast as a challenge facing common drugs owing to the adverse effects of these drugs as well as regimes of antifungals that are becoming inefficient over time [7]. Therefore, it is a must to discover new antifungal drugs for the treatment of fungal infections in humans.
The microbial communities in soil confer a diverse and highly complicated system [8]. There are complex interactions between bacteria and fungi in any environment which supports the development of mixed bacterial and fungal flora. Bacterial secondary metabolites which have antifungal activity donate bacteria an ecological advantage in these environments. A variety of in vitro methods have been used to confirm this activity and represent the basis for the improvement of numerous drugs with antifungal activity [9]. For example, amphotericin B was obtained from Streptomyces nodosus by purification of the metabolites [10]. Likewise, nystatin was obtained from Streptomyces noursei as well as pyrrolnitrin which is obtained from Pseudomonas pyrrocinia [11]. Moreover, alterations of these bacterial secondary metabolites chemically aided in the development of semi-synthetic antifungals with upgraded activity [12].
In addition to Pseudomonas and Streptomyces, Bacillus sp. has also proved to be efficient against fungal pathogens. B. subtilis has been reported to produce more than 70 different antibiotics [13]. According to several authors, secondary metabolites with antifungal activity against phytopathogenic microorganisms are produced by certain species of Bacillus. Some authors have suggested that the use of some of those species, or their metabolites, may be an alternative to the protection of plants chemically [14]. endospore-forming Bacillus sp. have properties that make them suitable for use as biocontrol agents, such as good stress resistance and producing fungal toxic compounds of low molecular weight. For example, subtilin, a peptide antibiotic, has been isolated from B. subtilis ATCC 6633 [15].
Therefore, our study aimed to screen soil bacterial isolates for a promising antifungal producer in addition to optimization of this antifungal activity using RSM.

Collection of soil samples, recovery, and maintenance of isolates
Ten soil samples were randomly obtained from different areas in Cairo taken from a depth of 10-20 cm. A gram of each sample was transferred into 100 mL saline, agitated for 30 min then 100 μL of the supernatant was spread on trypticase soy agar (TSA) and incubated at 37 °C for 2 days. The isolated colonies were then selected, grown on the same culture media, and preserved at 4 °C. After screening, the selected isolate which showed maximum antifungal activity was maintained in Trypticase soy broth (TSB) (Lab M, Topley house, England) with 20% glycerol at -80 °C.

Screening of the obtained isolates for antifungal activity
The activity of the recovered isolates against standard strain C. albicans ATCC 10231 was tested using cross streak assay using Sabouraud`s dextrose agar (SDA) plates and incubated at 30 °C for 24 h to allow the production of metabolite(s) [16]. Subsequently, C. albicans was inoculated perpendicular to the isolates' growth. Any inhibition zone (IZ) between the tested isolates and C. albicans ATCC 10231 after incubation at 28

Antifungal metabolite(s) production
Seed culture was prepared by inoculating isolate A3 into 50 mL TSB and incubating for 15-18 h at 200 rpm and 30 °C. About 1 mL from this culture was then centrifuged then the obtained cells, after washing with normal saline, were resuspended in the production media to obtain a final count of 1 x 10 7 cfu/mL. The medium used for production was prepared according to Sayed et al. [22] and Singh et al [23] and consisted of : 5 g/L starch, 6 g/L Na 2 HPO4, 3g/L K 2 HPO4, 0.2 g/L CaCl2, 0.5 g/L NaCl, 5 g/L NH 4 Cl, 0.12 g/L MgSO 4 and distilled H 2 O to 1 L. The pH was adjusted using KOH pellets to pH 7. From the seed culture, one milliliter was used to inoculate each 30 mL of the medium of then incubated at 200 rpm and 28 °C.

Estimation of the antifungal concentration
The equation of a standard calibration curve of standard terbinafine (Lamisil®) designed in our previous study [22] was used to calculate antifungal concentration as follows: Y= 0.0169 X + 5.0022, Where X is the concentration of the antifungal metabolite(s) (µg/mL) and Y is IZ diameter (mm).

Antifungal production time course
To determine the required time for greatest antifungal production, we prepared seven flasks (250 mL) as mentioned above and incubated them at 28 °C and 200 rpm. A single flask was sampled daily for 5 days to obtain the CFS. Briefly, 1 ml of the culture was centrifuged and the filter-sterilized CFS (200 µL/well) was evaluated by the agar well diffusion technique. After 24 h of incubation at 28 °C, IZ against C. albicans ATCC 10231 was measured.

RSM for Optimization of antifungal production
RSM, specifically D-optimal design was used to optimize 3 variables including temperature (A), pH (B), and inoculum size (C). Three levels were chosen for each factor, as illustrated in Table 1 and a sum of 13 runs was constructed ( Table 2). At the end of each experiment, the culture was centrifuged then the antifungal activity of the CFS was detected. A single response value, IZ diameter (mm), was obtained after 3 days of incubation. An equation, including all significant terms, was obtained from the program. Design Expert ® v. 7.0 (Design Expert ® Software, Stat-Ease Inc., MN, USA) was used to perform the design of experiments.

Experimental Validation of RSM results
The optimal environmental conditions (which were attempted experimentally) were obtained using the numerical optimization function in the software. Production using these conditions was then compared to the production using initial conditions.

Statistical and graphical analysis
The values recorded for all experiments were the average of 3 experiments whereas the data standard deviation was indicated by error bars. Design Expert v. 7.0 was used to obtain the design of experiments, response surfaces, model diagnostic plots, and ANOVA.

Recovery of the bacterial isolates and screening for antifungal activity
A sum of 54 bacterial isolates was obtained from the 10 gathered soil samples. Only, 4 (coded A2, A3, D5, and G4) out of 54 isolates (7.4%) displayed IZ against C. albicans ATCC 10231 suggesting their antifungal activities. The antifungal activities of these 4 isolates were confirmed using the agar well diffusion method. Isolate A3 showed the largest inhibition zone of 20 mm when compared to the rest and therefore, it was chosen for subsequent experiments (Fig.  1).

Time course of antifungal production in production media
The highest production, represented by an IZ= 24 mm, was attained after an incubation period of 3 days (Fig. 3). Consequently, the best incubation time for isolate A3 in ensuing trials was 3 days.

RSM for optimizing the antifungal activity
The D-optimal design was undertaken to determine the best conditions for antifungal production by isolate A3. The tested variables, the runs, and the practical results for each experiment are presented in Table 2.
A fitted equation was proposed by the program using the results attained from the 13 runs as follows: The suitability of the model was determined using ANOVA and the P-value, which denotes the factors' significance on the metabolite(s) production ( Table 3).
The proposed model was found to be significant, as suggested by the F-value and Pvalue (F value= 16.22 and P-value= 0.001). The model terms A, C, AC, and C 2 are significant model terms since their P-value was less than 0.5. The coefficient of determination R 2 suggested that 92% of changeability in response may well be clarified by our model. Furthermore, a good agreement was noticed among the predicted R 2 (0.77) and the adjusted R 2 (0.86). Last of all, the adequate precision ratio, which was 9.82, confirmed an adequate signal. Therefore, our model was suitable to navigate the design space.
The three-dimensional (3D) response surface plots (Fig. 4), taken with the numerical optimization from the program, suggested the best conditions for greatest antifungal activity which were: a temperature of 30 °C, a pH 8, and an inoculum size 5 x 10 7 cfu/mL.

Model diagnostics
The normal probability plot of residuals (Fig.  5 A) suggested no signs of error verified by linear patterns.
The residuals versus Run number plot revealed that the points were distributed around zero (Fig. 5 B) indicating that our model fits the data.
The Box-Cox plot is useful in determining the most suitable power transformation. Here, the software recommended a transformation to the log and therefore this was carried out as shown in

Validation experiment
An IZ of 27 mm was attained for B. subtilis isolate CCASU 2021-4 by using the proposed best levels of the conditions tested. This result was very near to that estimated by the model (27.5 mm), demonstrating an effective model. This IZ corresponded to a concentration of 1,360.82 µg/mL of the antifungal metabolite(s). Explicitly, optimization resulted in a 1.2-fold rise in antifungal production by the B. subtilis isolate CCASU 2021-4 when compared to yield using the unoptimized conditions (1124.13 µg/mL).

DISCUSSION
Invasive, deadly fungal infections, especially in immunodeficient patients are a critical reason for human morbidity and mortality [24]. The presently offered antifungals have numerous adverse effects which lessen their usage and safety profile. So searching for antifungal agents from natural sources is a necessity. In this study, a total of 10 soil samples were obtained from diverse areas in Cairo and were screened for the isolation of bacterial strains having antagonistic activity against the clinically relevant human fungal pathogen, C. albicans. Isolate A3, which depicted maximum antifungal activity, was Gram-positive and was identified as B. subtilis isolate CCASU 2021-4 (NCBI GenBank accession code, NR 112686.1). Microbial secondary metabolites production is a complex process impacted by even minor variations in the media or conditions. Hence, our study targeted achieving the best conditions for enhanced production of the antifungal metabolite(s) from the tested isolate. Carbon and nitrogen sources are vital components in the medium for bacterial growth and metabolite production [35]. The media optimized in our previous study was selected as production media for the efficient antifungal metabolite(s) production from the tested isolate [22]. Earlier studies showed an improved secondary metabolite production exploiting complex carbon sources like starch [36]. Ammonium chloride was also established to be the appropriate nitrogen source in former studies [36].

B. subtilis is categorized into three subspecies:
The traditional methods employed for optimizing fermentation conditions are sluggish, monotonous, and costly; also, they disregard the mutual interactions between diverse factors [37]. The power of the model to expect a response can be stated by the predicted R 2 which should not vary from adjusted R 2 by more than 0.2 [45]. Thus, our data presented a reasonable agreement between the adjusted and predicted R 2 values (0.86 and 0.77, respectively). The signal-to-noise ratio, expressed by the adequate precision was 9.82, demonstrating a pleasing model discernment being greater than 4 [48].
Additionally, each factor's significance was described by P-value. In our study, the model terms A (temperature), C (pH) were significant (P<0.05). The AC interaction was also significant as was shown by the P-value obtained (P<0.05) while the AB and CB interactions were not (P>0.05). A significant interaction indicates that one factor's effect relies on the level of the other [49]. The response surface plots were also generated, which help to understand the interactions between factors and therefore, to identify their optimum levels. From the 3D plots and numerical optimization function, the suggested ideal conditions for best antifungal yield were pinpointed and tried practically. The best concentration produced by B. subtilis A3 was 1,360.82 µg/mL which corresponds to a 1.2fold enhancement in production as compared to that produced by using the un-optimized conditions (1124.13 µg/mL). Finally, the model diagnostic plots also proved the reliability of the model created.

Conclusion
In this study, we report for the first time a promising antifungal activity from B. subtilis subsp. spizizenii against C. albicans. We tested the influence of several factors on this activity and pinpointed the ideal levels required for the highest production. D optimal design proved to be very efficient in determining the interaction between the variables and brought about a 1.2fold rise in yield. Therefore, these findings may function as baseline data to elucidate the metabolites' nature and for scaling up its production by the respective isolate.

Declarations
Ethics approval and consent to participate not applicable

Availability of data and materials
All are included in the main manuscript

Competing interests
Authors declare no competing interests

Funding Statement
No funding source was received

Authors' contributions
All authors participated in the practical work. The design of the work and analysis was carried out by GSH. All authors read and approved the final manuscript.