Escherichia coli (a previously conserved strain in the key laboratory for healthy and safe aquaculture at South China Normal University) as prey bacteria was inoculated in sterilized lysogeny broth liquid medium and cultured on a shaker at 37°C under 180 rpm for 24 h. Then, the host cells were collected by centrifugation at 2400 × g, 4°C for 15 min. Then, the resulting bacterial pellet was gently washed twice and re-suspended to reach a final concentration of 10¹⁰ colony-forming units (CFU)/mL with sterile 0.9% saline solution. This bacterial suspension was reserved at 4°C for subsequent experiments.

Water sampling was conducted in a freshwater fish pond at South China Normal University. A total of 500 mL water samples were collected at a depth of 10 cm from the surface of the fish pond using a water harvester. The samples were allowed to settle on ice for 30 min to facilitate the sedimentation of debris and larger particles. Subsequently, the supernatant was centrifuged for 5 min at 4°C and 100 × g for purification. A mixture of 50 mL of the centrifuged supernatant and 1 mL of E. coli (initial concentration 10¹⁰ CFU/mL) was prepared in a sterilized triangular flask and incubated in an oscillator at 30°C and 180 rpm until clarification occurred. The lysate was then placed on ice for 30 min to precipitate larger particles, followed by centrifugation at 400 × g for 10 min at 4°C to eliminate larger bacterial organisms, thereby obtaining the lysate supernatant.

The BALOs was isolated using the double-layer agar plating technique (Stolp and Starr, 1963): 0.5 mL of the treated lytic supernatant and 0.2 mL of E. coli (with an initial concentration of 10¹⁰ CFU/mL) were mixed into 5 mL of 0.6% agar medium containing CaCl₂ (at a final concentration of 3 mM). This mixture was then poured onto pre-solidified 1.5% agar medium. A negative control was established by mixing 0.5 mL of sterile tap water with 0.2 mL of E. coli in 5 mL of the same 0.6% agar medium containing CaCl₂ to validate the specificity of plaque formation. The plates were then incubated in a 30°C environment for 7 days (d) after the top agar solidified at room temperature, with observations made every 12 h.

After the initial plate culture for 48 h, an individual gradually enlarging plaque was carefully transferred into a tube containing 5 mL of sterilized tap water and 0.05 mL of 0.3 M CaCl₂ solution. This solution was then shaken on a shaker at a speed of 180 rpm and a temperature of 30°C for a total of 8 h. This soaking solution was purified using the double-layer agar plates. This same procedure was repeated until a plaque with uniform size and transparency was obtained, indicating that BALOs had been successfully purified.

The fresh BALOs lysate underwent a centrifugation at 400 × g and 4°C for 10 min. A droplet of supernatant was placed on a microscope slide and subjected to Gram staining utilizing the Biosharp kit, for subsequent observation under a standard optical microscope.

To enhance the morphological assessment of BALOs, a supernatant drop was placed on a 300-mesh copper net, sustained for 10, and subsequently sucked dry on filter paper. The copper net was then immersed in a 2% phosphotungstic acid stain for 90 s, clipped, and dried. In addition, a drop of sample supernatant was mixed with a drop of stain, and then, the stain on the copper net was washed off with the mixed solution for four times. The copper net was air-dried and studied under a transmission electron microscope (TEM).

Genomic DNA extraction was performed with the TaKaRa MiniBEST Universal Genomic DNA Extraction Kit Ver.5.0, following the manufacturer’s protocol. The concentration of the genomic DNA was quantified with a NanoDrop ONE spectrophotometer (Thermo Fisher Scientific, America) as well as checked on 1% agarose gel electrophoresis. The extracted DNA is preserved at −20°C.

According to the specific primer sequence designed by Edouard and Robert (2020), the following sequence was synthesized by IGEbio company (Table 1). For phylogenetic analysis, the 16S rRNA gene was selectively amplified by PCR using BALOs-specific primers in a total reaction system of 50 uL: primers (10 μM) 1 μL each, 2 × rTaq premix enzyme 25 μL, DNA template 2 μL, sterile distilled water for volume completion. The optimized thermal cycling parameters were 95°C for 3 min to activate the polymerase, followed by 33 cycles of 98°C for 10 s, 55°C for 30 s, and 72°C for 1 min, and the final extension step at 72°C for 5 min. PCR products (~800 bp) displayed successful amplification via 1% agarose electrophoresis, subsequently purified and dispatched to IGEbio company for sequencing. The measured 16S rRNA partial sequences were compared with those in NCBI database. Multiple sequence alignments were performed using the Cluster X program. A phylogenetic tree was constructed through the Minimum Evolution (p-distance model) using MEGA 11 software with a bootstrap analysis of 1, 000 replicates.

Bdellovibrio bacteriovorus-specific primers were used to ensure genus of the purified strain. PCR was performed at a total volume of 50 uL comprised of 1 uL gDNA, 25 uL 2 × rTaq premix enzyme, 1 uL of each primer, and sterile distilled water for volume completion. The optimized thermal cycling parameters were 95°C for 3 min to activate the polymerase, followed by 33 cycles of 98°C for 10 s, 55°C for 30 s, and 72°C for 30 s, and the final extension step at 72°C for 5 min. PCR products (~400 bp) displayed successful amplification via 1% agarose electrophoresis.

According to the hit gene primer sequence designed by Oyedara et al. (2016), the following sequence was synthesized by IGEbio company (Table 1). The PCR system used was the same as that described above, with a total volume of 50 μL. The optimized thermal cycling parameters were 95°C for 3 min to activate the polymerase, followed by 28 cycles of 98°C for 10 s, 55°C for 30 s, and 72°C for 1 min, and the final extension step at 72°C for 5 min. PCR products (~950 bp) displayed successful amplification via 1% agarose electrophoresis.

Samples were submitted to Shanghai Fuda Detection Technology Group Co., Ltd. for biochemical identification, scrutinizing oxidase, contact enzyme, arginine decarboxylase, gelatin liquefaction, indole test, citrate, nitrate reduction, urease, hydrogen sulfide, and methyl red tests.

Escherichia coli, Aeromonas hydrophila, Vibrio alginolyticus, Vibrio Parahaemolyticus, or Edwardsiella tarda (initial concentration of 10¹⁰ CFU/mL and final concentration of 10⁸ CFU/mL) were co-cultured with bacteria by B. bacteriovorus FWA (initial concentration of 2 × 10⁵ PFU/mL and final concentration of 10⁴ PFU/mL) on the double-layer agar plates at 30°C for 7 d, with plaque observations performed every 24 h. Meanwhile, the previously mentioned pathogenic bacteria were individually cultured on double-agar plates without BALOs as a negative control.

The prepared washed E. coli suspensions (10⁵, 10⁶, 10⁷, 10⁸, or 10⁹ CFU/mL) were, respectively, co-cultured with BALOs (at a final concentration of 2 × 10⁴ PFU/mL) at 30°C for 4 days under 180 rpm, with three replicates per group. The BALOs-free groups on different concentrations of E. coli served as negative control. The BALOs count variations (PFU/mL) were monitored at 24-h intervals.

The assessment of oxygen’s influence on the growth and predation of BALOs was conducted via co-culturing washed E. coli suspensions (at a final concentration of 10⁸ CFU/mL) and BALOs (at a final concentration of 10⁴ PFU/mL) at 30°C under 180 rpm. Vaseline was applied to the anaerobic group to maintain anaerobic conditions, whereas the aerobic group remained untreated. Three replicates were performed for each treatment. The BALOs-free groups consisted solely of the E. coli suspension, which were incubated in various oxygen environments as negative control. BALOs quantitative changes (PFU/mL) were monitored at 24-h intervals, and lysis capacity (OD600) was monitored at 24-h intervals.

The effects of pH on the growth and predation of BALOs were investigated by co-culturing with E. coli (at a final concentration of 10⁸ CFU/mL) and BALOs (at a final concentration of 10⁴ PFU/mL) at 30°C under 180 rpm under diverse pH conditions (pH 5, 6, 6.5, 7, 7.5, 8, 9, or 10). For temperature influence, E. coli and BALOs were co-cultured at pH 7.5 and 180 rpm but varied in temperature (20°C, 25°C, 30°C, 35°C, or 40°C). Each experimental group was replicated three times. The BALOs-free groups consisted solely of the E. coli suspension, which were incubated in different pH or temperature conditions as negative control. BALOs quantitative changes (PFU/mL) were monitored at 24-h intervals, and lysis capacity (OD600) was monitored at 24-h intervals.

Different concentrations of CaCl₂ (0 mM, 3 mM, 5 mM, 8 mM, 10 mM, 15 mM, 20 mM, 25 mM, or 30 mM), E. coli (at a final concentration of 10⁸ CFU/mL), and BALOs (at a final concentration of 10⁴ PFU/mL) were added to a culture system and cultivated at 30°C under 180 rpm to systematically probe the impact of Ca²⁺ on bacterial growth and predation of BALOs. Each experiment was repeated three times. BALOs quantitative changes (PFU/mL) were monitored at 24-h intervals, and lysis capacity (OD600) was monitored at 24-h intervals. The Na⁺ effect was investigated at 30°C under 180 rpm with different concentrations of NaCl (0, 0.5, 1, 1.5%, or 2%) in the same way as mentioned for different concentrations of Ca²⁺ effect. The BALOs-free groups only included the E. coli suspension, which were incubated in varying concentrations of Ca²⁺ or Na⁺ as a negative control.

Twelve 10 L tanks were utilized for the untreated aquaculture wastewater, with each tank containing 6 L of aquaculture effluent. The water temperature was controlled between 27 and 29°C, and air stones were used to provide oxygenation. Three experimental groups were established: the 10³ PFU/mL B. bacteriovorus FWA group, the 10² PFU/mL B. bacteriovorus FWA group, and the 10¹ PFU/mL B. bacteriovorus FWA group, alongside a negative control group lacking B. bacteriovorus FWA. The experiment lasted for 7 days.

To assess the variations in total bacterial counts between the experimental and control groups, water samples were collected at 0, 24, 48, 72, 96, 120, 144, and 168 hpc. The samples were diluted, and the total bacterial number of water samples was calculated by dilution plate counting. To assess variations in ammonia nitrogen and nitrite nitrogen levels in water samples, samples were collected every 12 hpc, spanning a duration of 0 to 168 h. Ammonia nitrogen was quantified using a photometric technique (Okumura et al., 2005), while nitrite nitrogen was analyzed via the N-(1-naphthyl)-ethylenediamine photometric method (Helaleh and Korenaga, 2001). The initial concentrations of ammonia nitrogen and nitrite nitrogen in the water were 0.10 ± 0.02 mg/L and 0.01 ± 0.00 mg/L, respectively.

All experiments were carried out in triplicate, and each test was repeated three times. A p-value of < 0.05 was considered as statistically significant. The results were presented as the mean ± standard error of mean (SEM). Statistical analyses were performed with GraphPad Prism V9 and IBM SPSS Statistics 26.

The water samples from freshwater fish ponds were initially enriched for BALOs in the presence of E. coli to increase predator density to sufficiently large to be observed by subsequent plaque assays. After 3–5 d of incubation on the double-layer agar plates, lytic plaques that were transparent and round with diameters ranging from 0.5 to 0.7 cm were observed in 96 h (Figure 1A). These plaques expanded after longer incubation (after 4 d of incubation), which eventually covered almost the entire lawn. These results suggested that predators were motile within the soft agar, which differentiated them from the types of plaques typically associated with bacteriophages.

The target 16S rRNA fragment and hit genes of BALOs were successfully amplified from isolation using the above primers, with amplicon sizes matching expectations (Figure 1B). Alignment of the 16S rDNA gene sequences isolated from GenBank database indicated that the predator was a member of the Bdellovibrio genus, with 98.89% identity to Bdellovibrio sp. Ec13 (Figure 1C). Therefore, the isolated predatory bacterium was suggested to be a strain of B. bacteriovorus, hereinafter referred to as B. bacteriovorus FWA (NCBI No. OR835679), for Bdellovibrio bacteriovorus strain fishpond water-A.

BALO co-cultures with E. coli were subjected to Gram staining for observation (Figure 2A). In addition to E. coli, which with short rod-shaped and 0.5 mm wide approximately, there were also bar bacteria that 0.4–0.5 μm wide and 0.8–0.9 μm long approximately. Further TEM observations revealed that the bacteria had long single flagella at one end. These results suggested that Bdellovibrio sp. was present in the co-culture (Figures 2B–D).

Analysis by FUDA ANALYTICAL TESTING GROUP (China) showed that the oxidase, citrate, and nitrate reduction tests of B. bacteriovorus FWA were negative, and the contact enzyme, arginine decarboxylase, gelatine, indole, urease, hydrogen sulfide, and methylation tests of B. bacteriovorus FWA were positive (Table 2).

Since Bdellovibrio obtains nutrients from parasitic host bacteria and cleaves the host bacteria to achieve value appreciation, the plaques or OD600 of the co-culture bacteria solution can roughly reflect the change in its titer. In this study, E. coli was used as the host bacteria of B. bacteriovorus FWA and the initial concentration of B. bacteriovorus FWA was 10⁴ CFU/mL. When the bait bacteria concentration was 10⁸ CFU/mL, the number of B. bacteriovorus FWA released progeny was much higher than other bait bacteria concentrations, and the 48 h post-co-culture (hpc) and 96 hpc the number of plaques reached 5.08 ± 0.01 (log, PFU/mL) and 6.54 ± 0.01 (log, PFU/mL), respectively. At 96 hpc, the B. bacteriovorus FWA reproduction efficiency decreased for groups that host concentrations of 10⁷ CFU/mL or 10⁹ CFU/mL, compared with 10⁸ CFU/mL group. For the former two groups, their plaque numbers were 5.13 ± 0.02 (log) and 5.65 ± 0.03 (log), respectively. In the groups with concentrations of 10⁵ CFU/mL and 10⁶ CFU/mL, the reproduction efficiencies of B. bacteriovorus FWA both were relatively low. At 96 hpc, their plaque numbers were 4.55 ± 0.08 (log, PFU/mL) and 2.58 ± 0.03 (log, PFU/mL), respectively. By calculation, the optimal ratio of B. bacteriovorus FWA to host bacteria was approximately 1:10000 (Figure 3A).

Escherichia coli, A. hydrophila, V. alginolyticus, E. tarda, and V. parahaemolyticus were tested for their susceptibility to predation by B. bacteriovorus FWA. In the double-layer agar cultures with E. coli, A. hydrophila, or E. tarda as prey, plaques appeared 36 hours post inoculation, yielding titers of 3.30×10⁴ PFU/mL, 3.31×10⁴ PFU/mL, and 2.06×10⁴ PFU/mL, respectively. Conversely, in the cultures with V. alginolyticus and V. parahaemolyticus, plaques were observed 48 hours after inoculation of B. bacteriovorus FWA, resulting in phage titers of 2.80×10⁴ PFU/mL and 2.11×10⁴ PFU/mL, respectively (Table 3). 96 h post-B. bacteriovorus FWA inoculation, diameter of plaques for double-layer agar culture with E. coli, A. hydrophila, V. alginolyticus, E. tarda, and V. parahaemolyticus prey was 0.35–0.5 cm, 0.25–0.35 cm, 0.3–0.5 cm, 0.25–0.35 cm, and 0.25–0.4 cm, respectively. In addition, for double-layer agar culture with E. coli or A. hydrophila prey, edge of the plaques was clear, and for double-layer agar culture with V. alginolyticus, E. tarda, or V. parahaemolyticus prey, edge of the plaques was foggy (Figure 3B).

OD600 of bacterial solution decreased significantly in the aerobic group during 48–72 hpc, while OD600 bacterial solution in the anaerobic group remained almost unchanged during the experimental period. This result confirmed that oxygen supply has a significant effect on promoting the lysis capacity of B. bacteriovorus FWA. At 72 hpc, the bacterial solution OD600 of aerobic group decreased by 0.57 ± 0.02 (p < 0.05), while the bacterial solution OD600 of anaerobic group decreased by 0.07 ± 0.00 (p < 0.05) (Figure 4A). For the aerobic group, the plaque assay showed that B. bacteriovorus FWA gradually increased during the experiment. Plaque numbers peaked at 72 hpc, which was 4.89 ± 0.07 (log), and the number of B. bacteriovorus FWA decreased from 96 hpc. For the anaerobic group, the number of plaques continued to decrease, and after 24 hpc, no plaques appeared on double-layer agar culture (Figure 4B).

Bdellovibrio bacteriovorus FWA could grow well at pH 5.0–10.0. At pH 7.0 and pH 7.5, the co-culture bacterial solution OD600 began to decrease from 48 hpc, which B. bacteriovorus FWA grew faster than in other pH environments. B. bacteriovorus FWA grew more slowly at pH 5.0 and pH 10.0, and the OD600 bacterial solution began to decline after 72 hpc. At 96 hpc, the bacterial solution OD600 showed the largest reduction at pH 7.0–8.0, with pH 7.5 and 8.0 being 0.60 ± 0.01 and 0.59 ± 0.01, respectively (Figure 5A). Similarly, the plaque assay confirmed that B. bacteriovorus FWA grow well at pH 7.0–8.0. At pH 7.5, the number of plaques at 48 hpc and 96 hpc was 5.22 ± 0.04 (log) and 6.28 ± 0.03 (log), respectively (Figure 5B).

The optimal temperature range for B. bacteriovorus FWA growth is 30°C–35°C. Bacterial solution OD600 of the 30°C group and 35°C culture group began to decrease after 36 hpc. For the 30°C group and 35°C group, at 72 hpc, the bacterial solution OD600 begun to decrease, which was 0.53 ± 0.02 and 0.50 ± 0.02, respectively. For the 25°C group, B. bacteriovorus FWA showed a little weaker grew activity, which lysed host bacteria massively after 84 hpc–96 hpc. For the 20°C group and 40°C group, bacterial solution OD600 was almost unchanged in the duration of the experiment (Figure 6A).

The results of plaques assay showed that at 30°C, B. bacteriovorus FWA had the best growth and reproduction ability. Over the 24 hpc–96 hpc period, the overall trend of plaque numbers was higher than in other temperature groups. For the 30°C group, at 48 hpc, 72 hpc, and 96 hpc, the number of plaques was 3.6466 ± 0.03 (log), 4.84 ± 0.03 (log), and 5.17 ± 0.03 (log), respectively. At 35°C, B. bacteriovorus FWA had similar growth and reproduction ability to that at 30°C, and the number of plaques at 72 hpc and 96 hpc was 4.84 ± 0.03 (log) and 4.09 ± 0.02 (log), respectively. At 20°C and 25°C, B. bacteriovorus FWA growth and reproduction efficiency decreased, and the progeny of B. bacteriovorus FWA was released slowly, and the plaque numbers at 48 hpc and 96 hpc were 2.64 ± 0.03 (log) and 2.72 ± 0.06 (log), respectively (Figure 6B).

At 48–72 hpc, in the difference Ca²⁺ superinduce group, OD600 of the co-culture bacterial solution began to decline, particularly with 10 mM, 15 mM, 20 mM, or 25 mM Ca²⁺ superinduce. For the 0 mM superinduce group, OD600 bacterial solution began to decline since 72–96 hpc. This result indicated that appropriate Ca²⁺ superinduce could promote the lysis and growth activity of B. bacteriovorus FWA. At 72 hpc, the 25 mM Ca²⁺ superinduce group has the greatest decrease in OD600 bacterial solution, which was 0.54 ± 0.01; and the following were the 15 mM Ca²⁺ superinduce group and 20 mM Ca²⁺ superinduce group, which was 0.53 ± 0.05 (p > 0.05) or 0.49 ± 0.04 (p > 0.05), respectively. The 0 mM superinduce group has the lowest decrease in OD600 bacterial solution, which was 0.23 ± 0.03 (p < 0.05; Figure 7A).

In the 25 mM Ca²⁺ superinduce group, B. bacteriovorus FWA’s plaque numbers on double-layer agar medium were fastest. At 48 hpc and 72 h, the number of plaques reached 4.43 ± 0.06 (log, p < 0.05) and 5.55 ± 0.04 (log, p < 0.05), respectively. For the 30 mM Ca²⁺ superinduce group and 20 mM Ca²⁺ superinduce groups, the number of plaques was 4.42 ± 0.03 (log) and 4.22 ± 0.08 (log) at hpc, respectively. In the 0 mM Ca²⁺ superinduce group, the number of plaques increased at the slowest rate, which was 3.00 ± 0.10 (log) at 72 hpc (Figure 7B).

Bdellovibrio bacteriovorus FWA showed the best efficiency in E. coli lysing without additional Na⁺ addition, where the bacterial solution OD600 began to decrease substantially at 24 hpc–48 hpc, and the bacterial solution OD600 at 48 hpc and 72 hpc was 0.29 ± 0.03 (p < 0.01) and 0.50 ± 0.01 (p < 0.01), respectively. Under the 0.5 and 1% Na⁺ superinduce, the co-culture bacterial solution of OD600 decreased significantly at 72 hpc–96 hpc, and the OD600 bacterial solution at 96 hpc was 0.55 ± 0.01 and 0.57 ± 0.03, respectively. Under the 2% Na⁺ superinduce, the bacterial solution of OD600 was almost unchanged during the experiment (Figure 8A).

In the absence of supernumerary Na⁺, B. bacteriovorus FWA has the best growth status. The number of plaques at 24 hpc and 48 hpc was 2.79 ± 0.07 (log, p < 0.01) and 5.33 ± 0.03 (log, p < 0.01), respectively. The number of plaques in the 0.5, 1.0, and 1.5% Na⁺ superinduce groups also increased after 48 hpc, while the increase rates were significantly lower than those of group without additional Na⁺ addition. The number of plaques in the 2% Na⁺ superinduce groups also increased during 72 hpc–96 hpc, and the number of plaques at 96 hpc was 2.59 ± 0.02 (log, p < 0.01, Figure 8B).

For the B. bacteriovorus FWA with a final concentration of 10³ CFU/mL, the total number of bacteria changed most dramatically, and the number of bacteria decreased significantly in the duration of 24–96 hpc. Total bacterial numbers in this group reached the lowest point at 96 hpc, at 0.28 ± 0.01 (log, p < 0.05). Compared with the 0 hpc, the total bacteria number reduces by 2.57 ± 0.04 (log). Conversely, the total number of bacteria in the control group rose by 0.12 ± 0.05 (log) at 96 hpc. For the group B. bacteriovorus FWA with a final concentration of 10² CFU/mL, the total population of bacteria also reached the lowest point at 96 hpc, at 0.83 ± 0.20 (log, p < 0.05). For the group B. bacteriovorus FWA with a final concentration of 10 CFU/mL, the total population of bacteria reached the lowest point at 120 hpc, which was 1.27 ± 0.04 (log, p < 0.01). The total population of bacteria in the control group did not change significantly during the experiment (Figure 9A).

In the first 24 h of the experiment, the ammonia nitrogen content in each experimental group increased, and then, the ammonia nitrogen content in all groups began to decline. Ammonia nitrogen content in the B. bacteriovorus FWA 10³ CFU/mL reached the lowest point at 72 hpc, which was 0.12 ± 0.02 mg/L (p < 0.01). Except for the negative control group, the other groups also showed an effect on ammonia nitrogen reduction after 24 hpc, and the effect was positively correlated with B. bacteriovorus FWA concentration. In the negative control group, the ammonia nitrogen concentrations remained elevated for up to 72 h, peaking at 0.72 ± 0.05 mg/L (Figure 9B). B. bacteriovorus FWA treatment did not significantly change the nitrite nitrogen content of each experimental group and the blank control group (Figure 9C).