Healthy samples of hybrid grouper (E. fuscoguttatus♀ × E. lanceolatus♂) without symptoms of infection [i.e., gross examination of hemorrhage, edema, lethargic, lesions, and detachment of scales (43)] with an average weight of 85 ± 2.77 g were obtained from a local farm situated close to the South China sea (Zhanjiang, Guangdong Province, China) and kept in aerated tanks that were later transported to the laboratory for the immediate commencement of the experiment.

Biochemical characterization tests are conducted to examine the agreement of genetic analysis and phylogenetic studies of the isolates. As illustrated in Table 2, the selected probiotic strains' biochemical characterization was performed using commercial kits procured from Guangdong Huankai Microbial Sci. and Tech. Co., Ltd. (Guangdong, China) following the company's instruction. With the help of a Bacillus cereus identification bar (HBIG07-1) purchased from the Qingdao Hope Bio-Tech. Co., Ltd. (Qingdao, China), the biochemical tests conducted were confirmed.

A selected single colony of probiotic bacteria strain from LB agar (pH 7.2) was first incubated overnight in 5 ml LB broth (37°C). One (1) milliliter of the overnight culture was inoculated into 100 ml LB broth (pH 7.2) in 500-ml Erlenmeyer flasks of which were incubated in a shaken incubator (150 rpm/min at 37°C for 24 h). During the culturing period, the bacteria growth was measured at 2-h intervals until reaching the 24-h stipulated time using a spectrophotometer (Evolution™ 220 UV-Visible Spectrophotometer, Thermo Scientific™, USA) at an absorbance optical density (OD) of 600 nm (46). Finally, the growth curve of the different strains was plotted. All the experiments were repeated in triplicate with the reading of the profiles being averaged.

This experiment was conducted to ascertain the possible harmful effects of the selected probiotic strains in healthy hybrid grouper. Healthy hybrid grouper fish (E. fuscoguttatus♀ × E. lanceolatus♂) weighing 90–100 g were purchased from a local fish farm (Zhanjiang, China). They were thus maintained in aerated cement pools [4.5 m (l) × 3.45 m (w) × 1.8 m (h)] for an acclimatization period of 2 weeks and were daily hand-fed twice (08:00 and 16:30) at 5% of their body weight with commercial diets (procured from Zhanjiang Aohua Feed Co. Ltd., Guangdong, China). After adaptation, a total of 90 fish (average weight 97 ± 1.34 g) after 24-h starvation were weighed and randomly distributed into nine fiberglass tanks (0.3 m³) at 10 fish-density per tank (i.e., divided into three groups for the three isolated probiotic strains (GPSAK2, GPSAK9, and GPSAK4 groups) with three replications each. The biosafety experiment, which lasted 21 days, was conducted in an indoor facility of the Marine Biological Research Base of Guangdong Ocean University (situated close to the South China Sea) under a photoperiod of natural 12-h light/12-h dark regime with a 2-day interval of 50% water exchange. Single airstones provided aeration, and the water quality of dissolved oxygen, temperature, pH, and salinity maintained as ≥6 mg L⁻¹, 28–30°C, 7.8–8.2, and 28.5–32%, respectively (YSI 556 multiprobe system, YSI Inc., USA). The isolates' suspension was prepared as stated above (see section Growth Pattern of the Selected Strains in Luria–Bertani Broth). Afterward, 0.1 ml [10⁸ colony-forming unit (CFU)/ml] of each isolated bacteria was injected intraperitoneally into their respective grouped fish. An additional 30 fish were kept in three different fiberglass tanks (10 fish/tank) of which were injected with the same volume of sterile phosphate-buffered saline (PBS) (pH 7.2) to serve as the control group. The steps used in acquiring the isolated bacteria's concentration (i.e., 10⁸ CFU/ml) followed a previously described process (20). Fish were then monitored daily to ascertain whether there were any clinical signs, and their mortality was recorded until the 21st day.

The susceptibility of the isolated Bacillus strains was assessed by the disc diffusion method against 24 antibiotics including ampicillin (10 μg), penicillin (10 μg), kanamycin (30 μg), ceftriaxone (30 μg), chloramphenicol (30 μg), erythromycin (15 μg), clindamycin (2 μg), furazolidone (300 μg), polymyxin B (300 μg), vancomycin (30 μg), sulfamethoxazole (27.5 μg), amikacin (30 μg), minocycline (30 μg), ofloxacin (5 μg), norfloxacin (10 μg), doxycycline (30 μg), neomycin (30 μg), gentamicin (10 μg), tetracycline (30 μg), carbenicillin (100 μg), midecamycin (30 μg), ciprofloxacin (5 μg), piperacillin (100 μg), and cefoperazone (75 μg) (Hangzhou Microbial Reagent Co., Ltd., Hangzhou, China). Briefly, 0.1 ml (at an OD adjusted to No. 0.5 McFarland standard 1.5 × 10⁸ CFU/ml) of cultured isolated probiotics was spread on 20 ml Mueller–Hinton agar (Beijing Land Bridge Tech. Co., Ltd.) plates. Subsequently, the antibiotic plates were carefully placed on the surface of the agar plates and incubated at 37°C for 24 h. After incubation, the probiotics' susceptibility was analyzed by measuring (mm) the zone of inhibition as described previously (47).

The bile salt tolerance was determined following the modified methods of Argyri et al. (48). Tolerance was examined by checking bacterial growth. In brief, bacteria cells from overnight culture after harvesting (9,000 × g for 5 min at 4°C) were washed twice with PBS (pH 7.4) and resuspended in 10 ml PBS (pH 7.4) containing 0.5% (w/v) of bile salts (Sangong Biotech Co., Ltd., Shanghai, China). Additional overnight cultures of isolates after harvesting and washing that served as the control were resuspended in PBS void of bile salt. Subsequently, 0.1 ml of the control and the exposed (after 0.5% bile salt exposure) bacteria at different time points (1, 2, 3, and 4 h) were taken and spread onto LB agar plates which was later incubated for 12 h. Viable colonies of the isolates that survived with or without 0.5% bile salt exposure were counted, and the percentage survival of isolates was calculated using the following formula: survival (%) = (B_t/B₀) × 100, where B_t represented the viable counts or the number of survived cells after incubation when bacterial isolates were exposed to PBS with 0.5% bile salt for 1, 2, 3, or 4 h, and B₀ is the viable counts or number of survived cells obtained after incubation when bacterial isolates were exposed to PBS with no bile salt (control). Evaluation of the isolated strains' tolerance to bile salt was conducted in triplicate.

Following Guo et al. (29) methodology with slight modification, the isolated bacterial strains' resistive capacity to different temperatures was evaluated. Since the processing of fish feed and other animal feeds at times requires high temperature, it is essential to know the isolated bacteria's resistive capacity to understand their survival in such harsh conditions. In doing this, the isolates' overnight cultures, after washing twice with 40 ml PBS (pH 7.4), were then exposed to 80, 90, and 100°C temperature using the Med-L-Hh 6 Electrothermal Thermostatic water-bath heater (Guangzhou Med Equipment Co. Ltd.) for 2, 5, and 10 min. Subsequently, an equal volume of LB broth was added to the heat-treated isolates to determine their ability to grow after treatment with heat. Growth was observed by measuring the absorbance at 600 nm after 12 h of incubation (37°C) with continuous shaking at 150 rpm/min. The high temperature-resistant assay was performed in triplicate.

Following the methodologies of Rajyalakshmi et al. (49), the test of compatibility of the three isolates was conducted. Briefly, the three probiotic isolates were vertically streaked on an LB agar plate 5 mm apart, followed by a perpendicular streaking of 10 mm apart from each other. The plates were afterward incubated (24 h at 37°C), and the compatibility was assessed by observing the zone of inhibition among the isolates.

The selected Bacillus strains were evaluated for antimicrobial activity against four pathogenic bacteria, namely; Streptococcus agalactiae, S. iniae, Vibrio harveyi, and V. alginolyticus, which were provided by the Guangdong Key Laboratory of Control for Diseases of Aquatic Economic Animals, Zhanjiang, China. The culturing process and the concentration of the disease bacteria used followed our previously described procedure (20). Briefly, 0.1 ml of each disease bacterium was added to 100 ml of LB in a 250-ml Erlenmeyer flask and cultured in a shaken incubator (180 rpm for 16 h at 37°C). After obtaining the cell pellets through centrifugation (4,930 × g at 4°C for 10 min), they were washed twice using PBS, and their concentration was adjusted at 600 nm wavelength. Through the serial dilution and spread plate technique, the supernatants were resuspended in PBS to get graded doses (10⁶, 10⁷, 10⁸, and 10⁹ CFU/ml). A prior cross streaking and agar well-diffusion experiment was conducted using the graded doses (three repeats each) with the isolated strains to determine which of the concentrations was best for the experiment, and 1 × 10⁸ CFU/ml was chosen. Finally, following the cross streaking and agar well-diffusion method of (50) using the chosen concentration, the pathogens were tested against the selected Bacillus strains isolated. The antimicrobial experiment was conducted in triplicate.

Auto-aggregation assays were performed as defined by Shin et al. (51) with slight modification. Briefly, Bacillus strains were grown overnight at 37°C in LB broth. Their bacterial cells were centrifuged (9,000 × g, 3 min), washed twice with sterile PBS (pH, 7.2), resuspended in the supernatant, and then vortexed for 30 s. The absorbance was measured at different times (0, 1, 2, 3, and 24 h) using a UV/visible spectrophotometer at a wavelength of 600 nm. The auto-aggregation percentage was calculated using the following formula: auto-aggregation (%) = (1–[A_t/A₀]) × 100, where A_t represented the absorbance at time t = 1, 2, 3, or 4, and A₀ represented the absorbance at t = 0 h.

The degree of cell surface hydrophobicity of the carefully chosen Bacillus strains was determined by employing the method described by Lee et al. (52) based on adhesion of cells to organic solvents with slight modification. Briefly, bacteria from a 24-h culture were harvested by centrifugation (9,400 × g, 3 min), washed twice with PBS (pH 7.2), and resuspended in 5 ml of the same buffer. The absorbance of the cell suspension was measured at 600 nm and used as the value A₀ to determine hydrophobicity (%). The cell suspension was then mixed with the same volume of solvent and mixed thoroughly by vortexing for 5 min, wherein ethyl acetate (a basic solvent), chloroform (an acidic solvent), and xylene (a non-polar solvent) were used. The suspension was incubated at room temperature for 30 min to allow two-phase separation. The aqueous phase removed was read at an absorbance wavelength of 600 nm and was subsequently labeled A_t. The percentage of bacterial cell adhesion to solvent was calculated using the following formula: hydrophobicity (%) = 1– (A_t/A₀) × 100, where A_t represented the absorbance of the aqueous phase after the two-phase separation, whereas that of the A₀ represented the absorbance before mixing with solvent.

The three isolated Bacillus species were subjected to a hemolytic test by streaking them onto agar plates supplemented with 7% sheep blood. The hemolytic zones were observed after incubation of plates at 37°C for 48 h. The isolates were then classified as α-, β-, and γ-hemolysis. The isolates having a green zone around the colony were recorded as α-hemolysis, while those with a clear zone were designated as β-hemolysis. Also, those that did not produce any zone around the colony were referred to as no- or γ-hemolysis (52, 53).

The optimal growth and pH were evaluated in conformity to Kavitha et al. (33). Briefly, the three bacterial isolates' fresh overnight cultures were inoculated in LB broth with varying pH levels (1–10), which was adjusted with acetic acid (99%) and 5 N NaOH. Subsequently, the inoculated broths were incubated at 37°C for 24 h, and growth was monitored with a spectrophotometer (Evolution™ 220 UV-Visible Spectrophotometer, Thermo Scientific™, USA) at 600 nm wavelength against uninoculated broth.

The production of biofilm was analyzed according to previously described methods (33). In brief, the isolated probiotic strains were streaked on Mueller–Hinton agar supplemented with 0.8 g/L of Congo red dye (Shanghai Macklin Biochemical Co., Ltd.) and incubated later on at 37°C for 48 h. Black colony presence with dry crystalline consistency showed biofilm production, whereas those with red colonies showed non-biofilm-producing strains.

All the experiments were conducted in triplicate, and the results obtained were subjected to a one-way analysis of variance (ANOVA) using the Statistical Package for Social Sciences (SPSS) software for Windows (IBM SPSS v20.0, Inc., 2010, Chicago, USA). Differences between means were tested by Tukey's honestly significant difference (HSD) test. A difference was considered to be statistically significant (P < 0.05), and the results are presented as mean ± standard error (SE).

The three isolated probiotic strains were designated as GPSAK2, GPSAK9, and GPSAK4. These probiotic strains were selected based on their morphological and biochemical characterization as described (see section Morphological and Biochemical Characterization of Isolates). The three isolated strains, GPSAK2, GPSAK4, and GPSAK9, showed close sequence homology (98–99%) to Bacillus tequilensis, Bacillus velezensis, and Bacillus subtilis, respectively. The phylogenetic tree generated by the NJ method using NCBI Distance tree online tool, as illustrated in Figure 1, established that the isolates GPSAK2, GPSAK4, and GPSAK9 were closest to B. tequilensis, B. velezensis, and B. subtilis, respectively. The nucleotide sequences obtained for these three strains have been deposited in the NCBI GenBank database under accession numbers MW548630 (B. tequilensis GPSAK2), MW548635 (B. velezensis GPSAK4), and MW548634 (B. subtilis GPSAK9).

The morphological and biochemical identification results are summarized in Tables 1, 2, respectively. It showed that all the isolates had an almost similar biochemical characteristics and thus were positive for Gram staining, rhamnose, inositol, sorbitol, adonitol, Simon's citrate, Vorges Proskauer (VP), arginine dihydrolase, spore formation, gelatin liquefaction; capable of producing catalase; aiding in the reduction of nitrate; able to metabolize lactose, starch, and glucose; and able to grow in lysozyme broth. All isolates were noted to reveal negative results for methyl red, urease, hippuric acid, and biofilm production test. GPSAK2 and GPSAK9 showed negative results for mannitol in contrast to the results observed for the GPSAK4 strain. GPSAK4 strain was thereby noted to be halophilic, since it could grow in lysozyme broth and was mannitol positive (Table 2). Figure 2 L1 shows the different morphological characteristics of the isolated strains.

The bacterial growth pattern of the three isolated strains (GPSAK2, GPSAK4, and GPSAK9) is presented in Figure 3. Although the log phase of strains GPSAK4 and GPSAK9 started approximately at 2 h, that of the GPSAK2 strain started at ~ 6 h after incubation (37°C) with continuous shaking (150 rpm) (Figure 3). In the culture process, the count of the vegetative cells for the GPSAK4 strain was higher within the 24-h culture period than the GPSAK2 and GPSAK9 strains. GPSAK2 strain attained its stationary phase earlier than the others.

Regarding the in vivo biosafety test, no pathological symptoms (i.e., gross examination of hemorrhage, edema, lethargic, lesions, and detachment of scales) were observed in the control and experimental fish. Furthermore, there were no recordings of mortalities confirming the non-pathogenic property of the three isolates.

Table 3 illustrates the results of the antibiotic susceptibility of the selected isolates. Twenty-four (24) antibiotics were tested in this current study. It was observed that all the isolates were highly susceptible (S) to most (19) of the antibiotics. GPSAK9 strain showed resistance (R) to clindamycin, furazolidone, vancomycin, and ampicillin and again showed moderate susceptibility (MS) to sulfamethoxazole and medecamycin antibiotics, whereas that of GPSAK2 showed MS to only furazolidone.

The three isolated strains' tolerance and survivability to 0.5% bile salt were monitored by counting the number of CFUs after 1, 2, 3, and 4 h of exposure, which was then expressed in percentage. The results demonstrated that more than 50% of the isolates survived after 4 h of exposure (Figure 4). It was observed that increasing the culture hours during bile salt exposure revealed a decreasing trend concerning the survival percentages of the isolated strains' CFU count. There were significant reductions (P < 0.05) in percentage survival of all isolates after 1 h bile salt exposure. While no significant differences (P > 0.05) were observed between the hours (1, 2, 3, and 4 h) exposed to bile salt in the GPSAK4 strain, that of the GPSAK2 and GPSAK9 strains revealed significant reductions (P < 0.05) in the percentage survival of the CFU counts.

Figure 5 illustrates the results obtained after isolates were exposed to different temperature conditions (80, 90, and 100°C) at different time points (2, 5, and 10 min) of which the three isolates gave promising results. After the trial, higher OD growth signifying higher growth of vegetative cell counts was observed in all the isolates exposed to the varying temperatures compared to the control (isolates without exposure to the higher temperatures). There were significant differences (P < 0.05) at the different times of exposure among the different temperatures (Figure 5). In the GPSAK2 strain, there were significantly high (P < 0.05) OD values witnessed when heated at (i) 80°C for 2 min compared to when heated at 90 and 100°C; (ii) 100°C for 5 min compared to when heated at 80 and 90°C; and (iii) both 90 and 100°C for 10 min compared to when heated at 80°C. For the GPSAK4 strain, a significant increase (P < 0.05) in the OD values was observed when heated at (i) both 90 and 100°C for 2 min compared to when heated at 80°C; (ii) 90°C for 5 min compared to when heated at 80 and 100°C; and (iii) 90°C for 10 min compared to when heated at 80 and 100°C. The GPSAK9 strain, on the other hand, showed significantly high (P < 0.05) OD values when heated at i) 90°C for 2 min compared to when heated at 80 and 100°C; (ii) 100°C for 5 min compared to when heated at 80 and 90°C; and (iii) 80°C for 10 min compared to when heated at 90 and 100°C.

In the current study, when the isolated strains were characterized for their compatibility, no definite sign of suppression of the three bacterial isolates was observed on each other, suggesting that they were compatible.

In the current study, the three isolated bacteria were assessed for their antimicrobial properties against the four fish pathogens, viz. S. agalactiae, S. iniae, V. harveyi, and V. alginolyticus. There were promising results for the isolates in the cross streaking and agar well-diffusion method (Figure 6). It was revealed at the end of the trial that the three isolates showed inhibition against the four pathogenic strains (Table 4). The GPSAK9 strain exhibited higher antimicrobial activity compared with the GPSAK2 and GPSAK4 strains.

Figure 7A shows the auto-aggregation ability of the three isolates. Auto-aggregation ability assays strongly correlate with cell adhesions to the digestive tract. The results revealed that all the isolates (GPSAK2, GPSAK4, and GPSAK9) witnessed low cell adhesion ability (< 40%) at the first 3 h. Nonetheless, after 24 h, the cell adhesion of the GPSAK2, GPSAK4, and GPSAK9 strains increased significantly (P < 0.05) to 83.7, 90.8, and 83.5%, respectively.

The adhesion of the isolated strains (GPSAK2, GPSAK4, and GPSAK9) to ethyl acetate, chloroform, and xylene solvents was again tested to determine adhesion capabilities of bacteria to cell surfaces of which the results are illustrated in Figure 7B. It was observed that the cell surface hydrophobicity of the GPSAK2 and GPSAK4 isolates to ethyl acetate and xylene was significantly lower (P < 0.05) than that of chloroform. The GPSAK2 strain showed much higher hydrophobicity (90.8 and 98%) with ethyl acetate and chloroform, illustrating that its bacterial adhesion to hydrocarbon compound is better than that of GPSAK4 (77.4 and 92.9%) and GPSAK9 (88.4 and 93.4%). However, concerning the results of xylene, the GPSAK9 strain revealed the highest hydrophobicity percentage (87.5%) compared to the GPSAK2 (79.6%) and the GPSAK4 (73.4%) strains. No significant difference (P > 0.05) was witnessed in the solvents of the GPSAK9 strains.

Concerning the hemolytic activities, GPSAK2 and GPSAK9 exhibited γ-hemolysis, whereas that of the GPSAK4 strain exhibited α-hemolysis (see Table 2 in section Morphological and Biochemical Characterization of Isolates).

All isolates, after different pH exposures, gave promising tolerance results. Although there was an irregular growth pattern, there was a gradual increase in vegetative cell growth within the pH range of 1.0–7.0 (GPSAK2 and GPSAK4 strains, optimum growth found to be at 7.0 pH) and 1.0–8.0 (GPSAK9 strain, optimum growth found between 6.0 and 8.0 pH). No significant difference (P > 0.05) in vegetative cell growth of strain GPSAK9 was observed between pH 6.0 and 8.0. Decreased growth was observed as pH increased from 7.0 to 10.0 (for both GPSAK2 and GPSAK4 strains) and 8.0 to 10.0 (for GPSAK9 strain), implying that the isolates could survive in extreme alkaline and acidic conditions. There were significant differences (P < 0.05) displayed in the isolates at different pH levels (Figure 8).

Biofilm formation detection was conducted using the Congo red agar method. All the isolates indicated negative biofilm production at the end of the test, as none of them formed black colonies [Table 2 and Figure 2 (L2)].