Optimizing Carbon, Nitrogen sources, and pH for P. paraflexa ONG1 cell density production
Bacterial growth is intricately linked to the availability and utilization of both carbon and nitrogen sources. Diverse carbon and nitrogen substrates exhibit varying degrees of suitability for biomass production.^24,25 Various carbon and nitrogen sources were evaluated to determine the most significant factors for maximizing high cell density production in P. paraflexa ONG1. Carbohydrate utilization tests indicated that probiotic bacteria could ferment glucose, sucrose, lactose, and starch. As a result, the selected carbon sources were the sole focus of the subsequent optimization. Lactose yielded the highest cell density (OD₆₀₀ = 0.40 ± 0.005) among the four carbon sources tested. Starch, sucrose, and glucose resulted in lower densities of 0.35 ± 0.028, 0.32 ± 0.003, and 0.25 ± 0.011, respectively. Statistical analysis confirmed that lactose was the preferred carbon source for P. paraflexa ONG1 (p < 0.05). The observed preference for lactose aligns with the findings of Niekerk and Pott,²⁶ who demonstrated the superior biomass production of Bacillus circulans DSM-11 on lactose compared with glucose, achieving a 1.7-fold increase in biomass yield. Interestingly, lactose supplementation in a growth medium for Priestia flexa N7 significantly enhanced hyaluronic acid (HA) production, with concentrations reaching 601.6 mg/L in a recent study.²⁷ Figure 1A and 1B illustrate how various carbon and nitrogen sources influence cell density.

Tryptone was the most effective nitrogen source, leading to the highest growth (0.58 ± 0.017) at 600 nm, followed by ammonium sulfate (0.54 ± 0.023), peptone (0.53 ± 0.011), and urea (0.52 ± 0.040). Peptone and yeast extract are the favored nitrogen sources for Priestia species. Remarkably, P. filamentosa and P. aryabhattai demonstrated superior growth capabilities when provided with organic nitrogen sources compared to inorganic nitrogen sources.²⁸ Our findings concur with this study, likely due to the use of tryptone as an organic nitrogen source. Supplementation of growth medium with urea led to a slight reduction in biomass production by P. paraflexa ONG1. Conversely, 0.1% urea emerged as the optimal nitrogen source for maximizing curdlan production by Priestia megaterium, whereas the strain demonstrated a limited preference for peptone.²⁹ Probiotic strain P. paraflexa ONG1 exhibits the metabolic versatility to utilize a variety of nitrogen sources, including both organic and inorganic compounds. This adaptability enables them to successfully colonize and persist in environments with limited nutrient availability.

Optimizing probiotic growth requires careful consideration of not only the specific media composition but also physicochemical properties, such as pH. The optimal cell growth (0.59 ± 0.00) occurred at pH 7.6, with decreasing values observed on either side of this point (Figure 1C). Minimal growth (0.48 ± 0.02) was recorded at pH 6.0. Notably, the organism exhibited minimal variation in cell density between pH 6.4 and 8.0 suggesting that the probiotic can thrive across a broad pH range. Our results align with a previous study on P. megaterium gasm, an elastase-producing bacterium from processed meat, which found that pH 8 is optimal for elastase production by this strain.³⁰ Concurrently, prior research corroborates the observation that Priestia species exhibit a remarkable capacity to thrive within a diverse pH range, encompassing slightly acidic to alkaline environments.^31-33

Plackett-Burman design for identifying key factors
PBD was implemented to determine the influence of each individual medium component on achieving high cell densities of P. paraflexa ONG1. Seven variables (lactose, tryptone, yeast extract, NaCl, KH₂PO₄, MgSO₄, and MnSO₄) were examined at two levels (-1, 1) (Table 1), leading to 12 experimental runs. Three replicates were conducted for each experimental condition to improve the reliability of the data. The resulting growth yields, measured as optical density at 600 nm, varied from 0.42 to 0.63 (Table 2). Table 3 presents the effects of the seven variables. The data revealed the following effects for each component: Lactose 0.020, tryptone 0.010, yeast extract 0.063, NaCl -0.008, KH₂PO₄ 0.021, MgSO₄ 0.047, and MnSO₄ 0.073. Among the seven variables examined, NaCl exhibited a negative correlation with probiotic growth (Figure 2). Elevating the NaCl concentrations within the tested range of 2.5 to 25 g/L resulted in a decline in cell density production. The standardized impact of each variable is visually represented in the Pareto chart (Figure 3). The vertical line on the chart indicates the statistical significance. Variables with columns extending to the right of this line have a substantial influence on bacterial growth. Statistical analysis revealed that yeast extract, MgSO₄, and MnSO₄ significantly influenced high cell density production (p < 0.05). Additionally, as indicated by the R² value of 0.963, the model effectively accounts for 96.3% of the response variability. The high adjusted R² value of 0.898 (indicating model significance) supports the suitability of the PBD. Thus yeast extract, MgSO_4, and MnSO₄ were chosen for subsequent RSM analysis. Shi et al.³⁴ highlighted the significant impact of magnesium sulfate on the colony formation of Bacillus velezensis BHZ-29, alongside other contributing factors. Another investigation, employing the Plackett-Burman experimental design, demonstrated that yeast extract was a significant factor in promoting spore production of B. coagulans X26.³⁵ Manganese, supplied as MnSO₄, is an essential micronutrient for Bacillus growth and serves as a key inducer of sporulation. Prior investigations have indicated the potential for manganese supplementation to augment spore formation efficiency,³⁶ enzyme production³⁷ and bacterial biomass production.³⁸

Response surface analysis
The approximate optimal region was identified using a response surface design. BBD was specifically utilized to discover the optimal concentrations of yeast extract (1-10 g/L), MgSO₄ (0.15-0.6 g/L), and MnSO₄ (0.13-0.5 g/L). The coded levels and their corresponding concentrations for the three variables used in the experimental investigation are summarized in Table 4. Table 5 presents the experimental design matrix, comprising 15 runs, including three center point replicates, and their corresponding responses. A maximum cell density 1.19 OD₆₀₀ was observed in run 15, corresponding to a combination of 10 g/L yeast extract, 0.375 g/L MgSO₄, and 0.5 g/L MnSO₄. Multiple regression analysis was used to fit a quadratic polynomial model to the experimental data for predicting P. paraflexa ONG1 cell density. The fitted equation is:

Y = 0.7764 + 0.05577 X₁ + 0.079 X₂ + 0.419 X₃ – 0.003807 X₁² + 0.156 X₂² – 0.207 X₃² – 0.01235 X₁ X₂ + 0.02703 X₁ X₃ – 0.300 X₂ X₃

In the equation, Y denotes the cell density (OD₆₀₀), while X_1, X_2, and X₃ correspond to yeast extract, MgSO_4, and MnSO₄, respectively.
Table (5):
Experimental design matrix for BBD

X₁: Yeast extract; X₂: MgSO₄; X₃: MnSO₄
Table (6):
Analysis of variance for Box-Behnken design (BBD) experiments in the optimization of P. paraflexa ONG1 cell density

_{X1: Yeast extract; X2: MgSO4; X3: MnSO4; DF: Degree of freedom; SS: Sum of squares; MS: Mean square; *represents p < 0.05; **represents p < 0.01}

Statistical significance was examined through the application of Analysis of variance (ANOVA), with subsequent utilization of an F-test. The calculated F-value of 43.25, as presented in Table 6, indicates a highly significant model (p-value 0.000). The developed regression model, exhibiting an R² of 98.73% and an adjusted R² of 96.45%, effectively predicted the cell density of P. paraflexa ONG1, accounting for over 95% of the variability in the data. Furthermore, a statistically non-significant lack-of-fit (p-value 0.072) indicated that the model provided a good fit to the experimental data. The magnitude of the p-value is inversely related to the term’s significance, with lower values indicating stronger evidence against the null hypothesis and potentially elucidating variable interactions.¹⁴ Analysis of Table 6 revealed that X₁, X₃, and X₁² were the most influential factors in this model, each demonstrating a statistically significant association (p < 0.01). The insignificance of the interaction terms is further supported by the elevated p-value
(p > 0.05), with the exception of the interaction between X₁ and X₃ (p-value 0.042). To gain a visual understanding of the interactive effects between the two variables and to identify the optimal concentration combinations for maximizing cell density, 3D response surface and contour plots were constructed using Minitab statistical software (Figure 4). Graphical representations based on the model equation were generated, which displayed convex response surfaces. Each graph illustrates the interactive effects of the two independent variables, whereas the remaining variable is fixed at a specific level. Cell density exhibited a positive correlation with yeast extract concentration within the intervals of 4.8 g/L to 10 g/L and 7.5 g/L to 8.82 g/L, as shown in Figures 4A, 4B, 4C and 4D. However, a further increase in yeast extract concentration beyond this point resulted in a decrease in cell density. The concentration-dependent response was investigated for MgSO₄ and MnSO₄, while maintaining the yeast extract concentration at the central level. Likewise, MgSO₄ demonstrated a similar pattern, with maximum cell density production at 0.20 g/L, followed by a reduction. Concurrently, varying concentrations of MgSO₄, as shown in Figure 4A and 4B, did not yield a statistically significant enhancement in cell density. The data indicates that elevated levels of the independent variables (yeast extract and MgSO₄) negatively impacted cell density. In contrast, for MnSO₄, maximum growth was observed at a higher concentration (0.5 g/L) (Figure 4E and 4F). The model predicted that a combination of yeast extract ( 8.82 g/L), MgSO₄ (0.15 g/L), and MnSO₄ (0.5 g/L) would maximize probiotic growth, resulting in a predicted cell density of 1.22 Optical Density at 600 nm.

Following cultivation in the optimized medium and under the optimized conditions, P. paraflexa ONG1 exhibited a maximum cell density of 6.5 × 10⁹ CFU/ml. This represents a 28.3-fold enhancement compared to the cell density obtained in nutrient broth medium, which was 2.3 × 10⁸ CFU/ml. The significant increase in cell density observed in the optimized medium strongly suggests that it provides an environment conducive to the growth and metabolism of P. paraflexa ONG1. This indicates that the optimized medium likely contained a well-balanced combination of nutrients that effectively meet the specific requirements of this organism. Using the Box-Behnken experimental approach, Eyahmalay et al.³⁹ investigated a simple growth medium formulated with lactose (76 g/L), soybean meal (72 g/L), yeast extract (2 g/L), and MgSO₄ (0.7 g/L). This optimization process resulted in a significant increase in Lactobacillus casei cell biomass, reaching a 164.6% increase. Shi et al.³⁴ utilized PBD and BBD to optimize the growth medium for Bacillus velezensis BHZ-29, resulting in an increased viable cell count from 7.83 × 10⁹ to 3.39 × 10¹⁰ CFU/ml. Information on the optimal growth conditions for Priestia species is scarce. To the extent of our current understanding, this investigation signifies the first attempt to optimize the growth of P. paraflexa, a promising aquaculture probiotic, utilizing Response Surface Methodology (RSM).