The red N. scintillans strain was originally isolated from the coastal waters of Lianyungang, China (October 2020). Briefly, N. scintillans single cell was isolated with a large-diameter pipette based on the morphology and maintained in sterilized seawater (salinity of 30‰), which was prepared by filtration (0.22 μm filter) and autoclaving of in situ seawater. The culture was incubated at 20°C under light (100 μE m^–2 s^–1, 14: 10 h light–dark cycle) and with Tetraselmis subcordiformis as the prey (Wu et al., 1994; Luo et al., 2022).

To inhibit the growth of excess bacteria in the algae, N. scintillans cultures of 100 mL were pretreated with an antibiotic mixture of 100 μL ampicillin (200 mg/mL), kanamycin (100 mg/mL), and streptomycin (100 mg/mL) before the experiment. In preliminary experiments, this antibiotic mixture did not adversely affect the growth of N. scintillans.

We collected surface seawater from a phytoplankton bloom in the coastal waters of Lianyungang, China (34°46’57"N, 119°27’8"E) (Zhang et al., 2022). The water samples were serially diluted 10-fold with sterilized seawater, and 0.2 mL of each 10-fold serially diluted dilution was spread onto 2216E (10 g L^–1 of tryptone, 2 g L^–1 of yeast extract, 0.2 g L^–1 of ferrous sulfate, and 1.0 L seawater) agar plates and incubated at 28°C for 48 h. The single colonies with significant differences in morphology were selected to obtain pure strains.

In order to screen the strains with algicidal activity against N. scintillans, we conducted pre-experiments in 24-well plates. All the purified strains were inoculated in 100 mL 2216E medium and cultured at 28°C, 180 rpm to their exponential growth period (OD₆₀₀: 0.6–0.8). The bacterial culture was inoculated into 1 mL (10 cells mL^–1) of algae cultures at 2 and 5% volume ratio, and an equal volume of sterilized 2216E medium and natural seawater was added to the algal cultures as control and blank groups, respectively. Bioassays were performed in triplicate, and all experiments were conducted during the light phase of the light cycle (20°C). The algal cells were counted at 4, 8, and 12 h to calculate the algicidal activity as follows:

where Nc and Nt represent the cell concentrations of N. scintillans measured in the control and experimental groups, respectively.

The cell density of N. scintillans was counted using a macroscopic observation because the cells are large enough (∼200–2,000 μm in diameter) to be seen by the naked eye (Zhang et al., 2020b). Gently shake the algal culture and collect 1 mL of culture for cell counting. Counting was performed three times in triplicate.

The morphological changes of algae cells were also examined and recorded under an inverted microscope to confirm the algicidal effect. The most effective and stable algicidal strain was selected for follow-up experiments.

Individual colonies of the screened algicidal bacterium were picked in 50 μL ddH₂O (double distilled water), incubated at 100°C for 10 min, and then quickly transferred to ice. The bacterial lysate was taken as a template, and its 16S rDNA was amplified by PCR using the universal primers 16S-27F (5′-AGAGTTTGATCMTGGCTCAG-3′) and 16S-1492R (5′-TACGGYTACCTTGTTACGACTT-3′) for bacterial identification (DeLong, 1992). PCR amplification was performed in a 25 μL amplification volume, which contained 1 μL of each primer (10 μM), 2 μL of template, 12.5 μL of 2 × Accurate Taq Master Mix (Accurate Biology, China), and 8.5 μL ddH₂O. An initial denaturation period of 95°C for 5 min, followed by 30 cycles of 94°C for 1 min, 55°C for 1 min, 72°C for 2 min, and the final extension at 72°C. Then the amplicons were sent to MAP (MAP Biotech, Shanghai, China) for sequencing. The obtained sequences were compared with those from the database for homology analysis using BLAST from the National Center for Biotechnology Information (Altschul et al., 1990). Multiple 16S rDNA sequences were aligned using ClustalW (Thompson et al., 2003). After removing the variable regions at both ends, 1,216 nucleotides positions were selected for phylogenetic analysis using the maximum-likelihood method with the Tamura-Nei model (1,000 bootstrap replications) in the MEGA 7.0 program (Kumar et al., 2016).

Among 45 isolates, strain LD-B6 was chosen for further analysis. The LD-B6 strain was inoculated in 100 mL 2216E medium and cultured at 28°C, 180 rpm to their exponential growth period to obtain the LD-B6 cultures (Supplementary Figure 1). In order to assess the algicidal activity of different final concentrations of LD-B6 against N. scintillans, LD-B6 cultures were added to the pretreated (antibiotic-treated) algal cultures with different volume ratios (0.5, 1, 2, and 5%). To elaborate, 0.25, 0.5, 1, and 2.5 mL of bacterial cultures were added to 50 mL (10 cells mL^–1) of the pretreated algal cultures. The same volume of sterilized seawater and 2216E medium was added to 50 mL of N. scintillans cultures as blank and control, respectively. All treatments were performed in triplicate and conducted during the light phase of the light cycle (20°C). The algal cells were counted after 4, 8, and 12 h of treatment, and the algicidal activity was calculated as described above.

To investigate the algicidal effect of LD-B6 on different densities of algal cultures, a 2% volume of cell-free supernatant of LD-B6 was added to 50 mL of the pretreated N. scintillans algal cultures with initial densities of 10, 30, and 50 cells mL^–1, respectively. Also, a 2% volume of sterilized 2216E medium was used as a control. All experiments were performed in triplicate and conducted during the light phase of the light cycle (20°C). The algal cell concentration was counted at 12 h, and the algicidal activity was calculated as described above.

To study the algicidal mode of LD-B6 against N. scintillans, the bacterial culture was treated in different ways: first, strain LD-B6 was inoculated in 100 mL of sterilized 2216E medium and cultured at 28°C, 180 rpm to their exponential growth phase to obtain the bacterial cultures, and then 10 mL bacterial cultures were collected and centrifuged at 12,000 rpm for 10 min. The collected supernatant was filtered with a 0.22-μm membrane to obtain the cell-free supernatant. The remaining cells were washed thrice with sterilized seawater and resuspended in 10 mL of sterilized seawater to obtain bacterial cells. The bacterial culture, cell-free supernatant, and bacterial cells were added to 50 mL (10 cells mL^–1) of the pretreated algal cultures at the optimum concentration of 2% (v/v) to investigate the algicidal activity of these fractions. In addition, the same amounts of sterilized 2216E medium were added to the algal culture as the experimental control. All experiments were performed in triplicate and conducted during the light phase of the light cycle (20°C). The algal cells were counted after 4, 8, and 12 h of treatment, and the algicidal activity was calculated as described earlier.

To examine the stability of the cell-free supernatant of strain LD-B6, the effect of different treatments on algicidal activity was investigated. The LD-B6 cell-free supernatant was incubated at different temperatures of −80, −20, 0, 15, 30, 60, and 100°C for 2 h, and then thawed or cooled to room temperature. The LD-B6 cell-free supernatant was adjusted to pH 3, 8, and 12 using NaOH or HCl and then adjusted back to the initial pH of 8 after 2 h. The supernatants treated in these two ways were inoculated into 50 mL (10 cells mL^–1) of the pretreated algal cultures at 2% (v/v). All treatments were performed in triplicate and conducted during the light phase of the light cycle (20°C). The algicidal activity was calculated at 12 h to test the heat and acid-base tolerance of the cell-free supernatant. The algal cell count and the algicidal activity were computed as described above.

LD-B6 cell-free supernatant tubes were then put in liquid nitrogen for 5 min. They were taken it out and thawed in a thermostat water bath at 65°C, and this freeze-thaw step was repeated three times. The supernatant was cooled to room temperature and inoculated into 50 mL (10 cells mL^–1) of the pretreated algal cultures at 2% (v/v). All experiments were performed in triplicate and conducted during the light phase of the light cycle (20°C). The algal cells were counted at 4, 8, and 12 h to calculate the algicidal activity. The algal cell counting and algicidal activity computations were as described above.

The algicidal efficiency of strain LD-B6 against typical HAB species was tested, including dinoflagellates (N. scintillans CCMA-LYG006, Prorocentrum micans CCMA-NT010, Gymnodinium impudicum CCMA-LYG009, and Heterocapsa steinii CCMA-LYG002), raphidophytes (Heterosigma akashiwo CCMA-LYG001), and diatoms (Skeletonema costatum CCMA277). Among them, P. micans was isolated from Nantong, China. S. costatum was provided by the Center for Collection of Marine Algae at Xiamen University. The rest strains were isolated from Lianyungang, China. Among them, S. costatum was cultured in the F/2 medium, and the other algae were cultured in the F/2-Si medium (Guillard, 1975). All algal cultures were incubated at 20°C under a 14: 10 h light–dark cycle with a photon flux of 100 μE m^–2 s^–1.

Flasks containing 50 mL of each species in an exponential growth period were inoculated with LD-B6 cultures at a final volume ratio of 2% (v/v). The same amounts of sterilized 2216E medium were used as controls. All treatments were performed in triplicate and incubated at 20°C. Algal cell concentrations were counted after 24 h. The cell concentration of N. scintillans was determined using macroscopic observations, and the algal cell density of other cultures was counted using a Sedgwick-Rafter chamber under a microscope.

Here, sawdust was selected as a carrier for LD-B6 immobilization. Briefly, sawdust was sterilized twice, dried, and used as the carrier. The carrier was added to the bacterial culture in the stable growth phase of LD-B6 at a mass fraction of 15% and then placed at 28°C and 180 rpm for co-incubation for 6 h to facilitate the sufficient adsorption of the sawdust. The un-adsorbed bacterial culture was removed by a screen mesh, and the adsorbed sawdust was placed on a glass petri dish and freeze-dried for 24 h. The lyophilized powder was stored at 4°C.

To assess the effectiveness of algicidal powder, 0.1 g of powder was added to 50 mL (10 cells mL^–1) of the pretreated algal cultures, and the same volume of sterilized sawdust was added to 50 mL of N. scintillans cultures as control. All treatments were performed in triplicate and conducted during the light phase of the light cycle (20°C). The algal cells were counted at 12 h, and the algicidal activity was calculated as described above.

All experiments were performed in triplicate, and data were presented as mean ± standard deviations. Statistical analyses were performed using SPSS16.0 software, and the significant differences among treatments in this study were analyzed by one-way ANOVA (p < 0.05).

After isolation and purification, a total of 45 strains were initially obtained. Among them, 11 strains showed algicidal activity against N. scintillans, and the LD-B6 strain had the strongest algicidal activity in the experiments with 2 and 5% volumes (Supplementary Figure 2). At the same time, microscopic imaging confirmed the effective algicidal activity of strain LD-B6. As shown in Figure 1 and Supplementary Video 1, the cell wall of N. scintillans began to break (6 h), and the contents flowed out after a period of co-culture of the bacterial culture and the algal culture, and then the rupture expanded continuously. Thus, the strain LD-B6 was selected for further study.

The 16S rDNA gene of LD-B6 was amplified and sequenced (GenBank accession number: OP010006.1). The 16S rDNA sequence of this strain showed the highest similarity (100%) to that of Pseudoalteromonas sp. CF1 (accession number KX570621). Meanwhile, in phylogenetic analysis, strain LD-B6 formed a monophyletic group with Pseudoalteromonas sp., corresponding to the Gammaproteobacteria subclass (Figure 2). This result indicated that strain LD-B6 belonged to Pseudoalteromonas sp., and it was designated as Pseudoalteromonas sp. LD-B6.

The algicidal activity of LD-B6 against N. scintillans at different concentrations was examined to determine the optimum final volume of the bacterial culture for further experiments. As shown in Figure 3B, LD-B6 at 1, 2, and 5% final volume activity have a certain algicidal effect against N. scintillans, and its algicidal activity increased in a concentration- and time-dependent manner, reaching 37.9 ± 1.8, 90.5 ± 3.2, and 100% at 12 h, respectively. However, the 0.5% treatment group only inhibited the growth of N. scintillans and did not kill algal cells (Figure 3A). Thus, the algicidal effect of strain LD-B6 against N. scintillans was concentration-dependent. As the algicidal activity in the 2% treatment group was higher than 90% at 12 h, this dose was chosen for further studies.

The cell-free supernatant of LD-B6 was added to N. scintillans cultures at different initial densities to examine its algicidal activity. As shown in Figure 4, the algicidal effect of the cell-free supernatant was the highest when the initial density was 10 cells mL^–1, followed by 30 and 50 cells mL^–1at algicidal activity of 93.8 ± 3.1, 78.2 ± 1.7, and 61.9 ± 1.7%, respectively. Generally, the algicidal activity of LD-B6 against N. scintillans decreased with the initial density of the algal culture increasing.

To initially investigate the algicidal mechanisms of strain LD-B6, we added different fractions of bacterial cultures to N. scintillans cultures. The strain culture and cell-free supernatant showed an algicidal activity of 89.5 ± 1.8 and 89.5 ± 4.8%, respectively, after treatment for 12 h, while the algicidal effect of the bacterial cells was only 14.7 ± 3.2% and not obvious (Figure 5). The algicidal effect of the supernatant was significantly better than that of the bacterial cells themselves. It is presumed that strain LD-B6 lyses algal cells by secreting active compounds into the cell-free supernatant, and therefore strain LD-B6 lysed N. scintillans indirectly.

Under different temperature treatment conditions, the algicidal effect of strain LD-B6 compounds on N. scintillans is shown in Figure 6A. Although the algicidal activity was slightly different among treatments, there was no significant difference. It shows that the algicidal compounds of strain LD-B6 have definite thermal stability. Incubation at different pH values for 2 h had an effect on the algicidal activity of the cell-free supernatant of strain LD-B6. The algicidal compounds of strain LD-B6 were sensitive to acid, and their algicidal activities were significantly decreased under acidic conditions. In contrast, the algicidal effect of the cell-free supernatant was stable under alkaline conditions (Figure 6B). Repeated freezing and thawing had almost no effect on the algicidal activity of the LD-B6 cell-free supernatant, and there was no significant difference in the algicidal effect between the control and the experiments (Figure 6C).

The algicidal activity of LD-B6 culture against six different HABs species is shown in Figure 7. Within 24 h of incubation, the LD-B6 culture showed the strongest algicidal effect on N. scintillans and H. akashiwo, and the algicidal activity reached 100%. It also had a strong algicidal activity on H. steinii and G. impudicum, with the algicidal activity of 73.3 ± 1.3 and 63.7 ± 9.2%, respectively. A weak effect on P. micans (26.7 ± 0.68%) was observed. In contrast, the growth of S. costatum was slightly promoted by strain LD-B6.

Algicidal powder product status is shown in Supplementary Figure 3. The algicidal powder of LD-B6 was added to N. scintillans cultures to assess its algicidal effect. As shown in Supplementary Figure 4, the algicidal activity of 100% at 12 h with 0.2% final volume of algicidal powder. This result indicated that the developed algicidal powder has a strong algicidal effect and that the production method is feasible.