Simultaneous Removal of the Freshwater Bloom-Forming Cyanobacterium Microcystis and Cyanotoxin Microcystins via Combined Use of Algicidal Bacterial Filtrate and the Microcystin-Degrading Enzymatic Agent, MlrA

Freshwater cyanobacterial blooms (e.g., Microcystis blooms) constitute a major global environmental problem because of their risks to public health and aquatic ecological systems. Current physicochemical treatments of toxic cyanobacteria cause the significant release of cyanotoxin microcystins from damaged cells. Biological control is a promising eco-friendly technology to manage harmful cyanobacteria and cyanotoxins. Here, we demonstrated an efficient biological control strategy at the laboratory scale to simultaneously remove Microcystis and microcystins via the combined use of the algicidal bacterial filtrate and the microcystin-degrading enzymatic agent. The algicidal indigenous bacterium Paenibacillus sp. SJ-73 was isolated from the sediment of northern Lake Taihu, China, and the microcystin-degrading enzymatic agent (MlrA) was prepared via the heterologous expression of the mlrA gene in the indigenous microcystin-degrading bacterium Sphingopyxis sp. HW isolated from Lake Taihu. The single use of a fermentation filtrate (5%, v/v) of Paenibacillus sp. SJ-73 for seven days removed the unicellular Microcystis aeruginosa PCC 7806 and the native colonial Microcystis strain TH1701 in Lake Taihu by 84% and 92%, respectively, whereas the single use of MlrA removed 85% of microcystins. Used in combination, the fermentation filtrate and MlrA removed Microcystis TH1701 and microcystins by 92% and 79%, respectively. The present biological control thus provides an important technical basis for the further development of safe, efficient, and effective measures to manage Microcystis blooms and microcystins in natural waterbodies.


Introduction
Cyanobacterial blooms can have long-term negative impacts on water-quality and ecosystem health, such as increasing water turbidity, diminishing dissolved oxygen, smothering fish, producing unpleasant odor compounds, and excreting various cyanotoxins [1,2]. The control of harmful cyanobacterial blooms might become more challenging in the future, because it is widely confirmed that cyanobacterial blooms are gradually increasing in frequency, magnitude, and duration on a global scale owing to progressive eutrophication, rising atmospheric CO 2 levels, and global warming [3,4]. Increasing attention has been cess cyanobacteria and MCs simultaneously. The algae-lysing performance of the strain Paenibacillus sp. SJ-73 and the desirable conditions for heterologously expressing MlrA will be characterized. Moreover, the efficiency of the co-application of the algicidal bacteria and MlrA will be evaluated in terms of the cyanobacteria biomass removal rate and the MC degradation rate.

Isolation and Characterization of Algicidal Bacteria 2.2.1. Screening for Algicidal Bacteria Strains
One gram of sediment sampled from Meiliang Bay, Lake Taihu, was suspended in 10 mL of sterilized deionized water (dH 2 O), spread onto LB plates, and incubated at 30 • C for 24 h. A single colony was randomly picked to screen potential algicidal bacteria by observing their effects on the growth of M. aeruginosa PCC 7806 in 24-well plates. Thereafter, cultures of potential algicidal bacteria (5%, v/v) were added into an axenic culture of M. aeruginosa PCC 7806 to measure their algicidal efficiency. The algicidal efficiency was calculated using the formula: Algicidal efficiency (%) = [(C 0 − C)/C 0 ] × 100% where C 0 was the chlorophyll a (Chl-a) concentration at the initial stage, while C was the Chl-a concentration at the end of the experiment. The concentration of Chl-a was measured according to standard methods [36]. Bacterial cultures of Pseudomonas aeruginosa (CCTCC-AB91095) were used as the negative control. The strain with the strongest algicidal efficiency was isolated and prepared for subsequent experiments.

Morphology and Molecular Identification of Algicidal Bacterial Strain SJ-73
Strain SJ-73 was identified by its morphological characteristics and 16S rDNA sequence. Cells of strain SJ-73 were collected by centrifugation at 4000× g at 25 • C for 10 min, and washed with 0.1 M phosphate-buffered saline (PBS) twice. The washed cells were added into 2.5% (v/v) glutaraldehyde and fixed at 4 • C overnight. The cells were then dehydrated through a graded series of ethanol (30%, 50%, 70%, 90% and 100%, v/v) for 10 min and then freeze-dried. Finally, the cells were coated with a 20 nm-thick layer of gold and palladium, and their morphological features were observed under a scanning electron microscope (SEM) (Hitachi-S4800, Hitachi, Tokyo, Japan) [37].
The 16S rDNA sequence of strain SJ-73 was sequenced and compared using BLAST with similar 16S rDNA sequences extracted from GenBank. A phylogenetic tree was constructed by the neighbor-joining method using the software molecular evolutionary genetics analysis version 7.0 (MEGA 7.0) [38]. The bacterial concentrations throughout the experiments were determined using the colony-forming units procedure [39] and measured using the optical density at 600 nm (OD 600 ). The bacterial culture was centrifuged for 10 min at 4000× g to collect two different fractions, including the cells and the cell-free supernatant. The filtrate was collected for further experiments by filtering the cell-free supernatant through a 0.22 µm fast flow and low binding Millipore express PES membrane (Merck Millipore Ltd., Dublin, Ireland). To determine the direct or indirect algicidal pathway of SJ-73 against M. aeruginosa, the filtrate and the isolated cells were separately added into the algal culture of M. aeruginosa PCC 7806 at 5% (v/v) to assess their algicidal efficiency by measuring Chl-a.
To investigate the algicidal host specificity against the different bloom-forming cyanobacteria species, M. aeruginosa (PCC 7806), Ana. sp. (PCC 7120) and Aph. flos-aquae (FACHB 1168) were co-cultivated with SJ-73 filtrate for 7 days. In addition, two green algae, C. sorokiniana (FACHB 26) and T. obliquus (FACHB 416) were tested as well. The algicidal efficiency against these species were measured on the first and seventh day based on the Chl-a contents.

Algicidal Effects on the Growth of Unicellular and Colonial M. aeruginosa
The strain SJ-73 and Pseudomonas aeruginosa AB91095 were inoculated into a liquid LB medium at 30 • C, with shaking at 200 rpm for 24 h. Then, the bacterial cultures were centrifuged for 10 min at 4000× g and filtered through a 0.22 µm membrane to collect filtrates. For unicellular PCC 7806, 6 mL of SJ-73 or AB91095 filtrates were added into 120 mL of M. aeruginosa PCC 7806 in a conical flask. For colonial M. aeruginosa TH1701, 6 mL of SJ-73 filtrate or Gause's medium were mixed with 120 mL colonial M. aeruginosa TH1701 cultures. The algicidal effects of the strain SJ-73 on M. aeruginosa PCC 7806 and colonial M. aeruginosa TH1701 were evaluated using the algicidal efficiency and photosynthetic activity.
The Chl-a concentrations were measured every two days and the photosynthetic activities parameter (Fv/Fm) was tested every day using a WATER-PAM (WATER-ED, Heinz WALZ, Germany) [40].

Effects on Cell Morphology
To further investigate the process of algal cell lysis under treatment by strain SJ-73, the cells were observed under an SEM and a transmission electron microscope (TEM) (Hitachi-7700, Hitachi, Tokyo, Japan). M. aeruginosa PCC 7806 cells were harvested by centrifugation at 4000× g for 10 min after treatment with the strain SJ-73 filtrate for 0, 3 and 7 days. The collected cells were fixed overnight under 4 • C in 0.1 M PBS (pH 7.4) and dehydrated through a graded series of ethanol (30%, 50%, 70%, 90% and 100%, v/v) for 10 min. The fixed samples were subjected to SEM and TEM observations according to the standard methods [37].

Cloning and Heterologous Expression of MlrA
Sphingopyxis sp. HW was cultured in an LB medium with 5 ng/L MC-LR at 30 • C for 48 h, and then genomic DNA was extracted using an MP FastDNA SpinDown Kit (MP Biomedicals, LLC, Illkirch-Graffenstaden, France). The mlrA gene sequence was acquired by whole genome sequencing (data not published), and the mlrA gene was amplified from the whole genomic DNA of Sphingopyxis sp. HW using primers mlrA-F (5 -AGGAGAATACCATGCGGGAGTTTGTCAAACAGCGA-3 ) and mlrA-R (5 -TGGTGGTGG TGATGATGCGCGTTCGCGCCGGAC-3 ). PCR was performed in a total volume of 50 µL, including 25 µL KAPA HiFi Hot Start Ready Mix (KAPA Biosystems. Co., Ltd., Wilmington, MA, USA), 0.8 µL forward and reverse primer, 2 µL DNA samples, and 21.4 µL ddH 2 O. The PCR reaction included initial denaturation at 95 • C for 4 min; followed by 30 cycles of 20 s at 95 • C, 55 • C for 15 s, and 72 • C for 100 s, with a final extension step for 5 min at 72 • C.
The target band was excised from a 1.2% (m/v) agarose gel and purified using a TIANGEN DNA purification Kit (TIANGEN Biotech. Co., Ltd., Beijing, China) according to the manufacturer's instructions. Subsequently, the amplicon was ligated into vector pET28H using a ClonExpress ® Ultra One Step Cloning Kit (Vazyme Biotech. Co., Ltd., Jiangsu, China) to obtain the plasmid, which was transformed into E. coli BL21 (DE3) competent cells.
The expression of recombinant MlrA was according to reference [32], with minor modifications. The bacteria and negative control were grown in LB media with 50 µg/mL kanamycin and with shaking at 180 rpm under 30 • C. When the OD 600 of the cultures reached 0.6, the expression was induced with 0.2 mM isopropyl-β-d-thiogalactoside (IPTG, sigma, USA) at 18 • C overnight. The cultures were centrifuged at 4000× g and 25 • C for 10 min and re-suspended in PBS (50 mM, pH 7.4) at an equal volume. Then, the samples were treated using a low temperature ultra-high-pressure continuous flow cell disrupter (750 W, HN-mini Juneng Biotech. Co., Ltd., Guangzhou, China) to lyse the cells, and centrifuged at 8000× g, 4 • C for 20 min to collect the cell-free extract, which contained MlrA.  [36]. All experiments were conducted in triplicate.
The degradation product sample was used to determine the molecular weight (MW) of the parent [D-Asp 3 ]MC-LR and its degradation products using ultraperformance liquid chromatography-tandem mass spectrometry (UPLC-MS/MS) (LC-Q-TOF 1290II/6545B, Agilent, Santa Clara, CA, USA). The drying gas temperature and flow rate were set at 320 • C and 8 L min −1 , respectively. The mobile phase consisted of a gradient of water (solvent A) and acetonitrile (solvent B), both containing 0.05% formic acid (Sigma, St. Louis, MO, USA). The samples were run with the following linear gradient program: 20% B to 80% B over 8 min, then to 80% B over 2 min, where it was held for 2 min.

Preparation of MlrA
Cell cultures were harvested by centrifugation (4000× g at 4 • C for 25 min), and washed twice with PBS (50 mM, pH 7.4). The cells were disrupted using a cell disrupter (750 W for 20 min) in an ice bath. After centrifugation, the supernatant was collected as a cell-free extract (CE) containing crude protein. The total protein contents were determined using an enhanced BCA protein assay kit (Beyotime Biotech. Co., Ltd., Jiangsu, China) and the concentration of the MlrA enzyme was detected using a His-tag enzyme-linked immunosorbent assay (ELISA) detection kit (Genscript Biotech. Co., Ltd., Piscataway, NJ, USA).

The Degradation Efficiency and Characteristics of MlrA
To analyze the characteristic of MlrA, expression assays were carried out using three parallel experiments. The CE, containing 0.8 mg/L MlrA protein, was added into the test tubes with ddH 2

Efficiency of Immobilized and Free MlrA
Sodium alginate was dissolved in sterile water to a final concentration of 3.0%. An equal volume of MlrA was mixed with an equal amount of sodium alginate solution at a 1:1 ratio. Then, the mixture was added into 4% (m/v) calcium chloride (Sigma, St. Louis, MO, USA) solution and gently stirred at room temperature for 30 min to form immobilized beads. The immobilized beads were washed with sterilized water 3-5 times to remove the calcium chloride on the surface. The immobilization efficiency was calculated according to the following formula: where m 0 was the amount of protein at initial stage; C was concentration of remaining protein in solution; and V was the volume of the solution. The immobilized beads were stored at 4 • C and 25 • C, separately. To explore the stability of free MlrA and immobilized MlrA over time, samples were taken out of storage and incubated with 5 µg/mL [D-Asp 3 ]MC-LR for 30 min at 25 • C every day.

Combined Treatment of Algicidal Bacteria and MlrA on M. aeruginosa and MCs
The strain SJ-73 filtrate was collected and filtered through a 0.22 µm fast flow and low binding Millipore express PES membrane. Then, the filtrate (5%, v/v) was poured into M. aeruginosa PCC 7806 and colonial M. aeruginosa TH1701 cultures, respectively. At the beginning of experiments, the cell densities of the PCC 7806 culture reached 4.4 × 10 7 cells/mL and the Chl-a concentrations of TH1701 were 500 µg/L. Both were higher than those of cyanobacteria during the harmful cyanobacterial blooms of Lake Taihu.
For the axenic unicellular strain M. aeruginosa PCC 7806, 6 mL of 0.8 mg/L MlrA or PBS (50 mM, pH 7.4) was added into 60 mL of M. aeruginosa PCC 7806 cultures treated with strain SJ-73 for 6 days. For the non-axenic colonial strain M. aeruginosa TH1701, 6 mL of 0.8 mg/L MlrA was mixed with the SJ-73 filtrate and added to the algal culture at the beginning of the experiment. Then, 0.8 mg/L MlrA was added again on the third day. The concentrations of MCs in all samples were determined using a microcystins ELISA detection kit (Institute of Hydrobiology, Wuhan, China) according to the manufacturer's instructions. All experiments were conducted in triplicate.

Statistical Analysis
All experiments were performed in triplicate and presented as means ± standard deviation of the replicates using standard parametric analyses. The significant differences among all the data were evaluated using ANOVA in SPSS 19.0 for windows (IBM Corp., Armonk, NY, USA), with p < 0.05 (*) and p < 0.01 (**).

Isolation and Identification of Indigenous Algicidal Bacteria
Bacterium strain SJ-73, which causes the lysis of M. aeruginosa, was isolated from the sediment sample of Lake Taihu during a cyanobacterial bloom period. It was a Grampositive and rod-shaped bacterium (0.3-0.5 µm × 1-3 µm) (Figure 1), which formed white colonies with a smooth surface on Gause's and LB plates. Phylogenetic analysis based on 16S rDNA sequence revealed that strain SJ-73 was affiliated with the genus Paenibacillus, sharing the highest similarity (99.93%) with the Paenibacillus amylolyticus strain HSE1 (accession number: MK764982.1) ( Figure S1).

Isolation and Identification of Indigenous Algicidal Bacteria
Bacterium strain SJ-73, which causes the lysis of M. aeruginosa, was isolated from the sediment sample of Lake Taihu during a cyanobacterial bloom period. It was a Grampositive and rod-shaped bacterium (0.3-0.5 μm × 1-3 μm) (Figure 1), which formed white colonies with a smooth surface on Gause's and LB plates. Phylogenetic analysis based on 16S rDNA sequence revealed that strain SJ-73 was affiliated with the genus Paenibacillus, sharing the highest similarity (99.93%) with the Paenibacillus amylolyticus strain HSE1 (accession number: MK764982.1) ( Figure S1).

Algicidal Model and Spectrum
The algicidal mode of strain SJ-73 was determined by comparing the lytic activity among the bacterial culture, cell-free filtrate, and washed cells against the Microcystis aeruginosa. The bacterial culture of strain SJ-73 and the cell-free filtrate exhibited significant inhibition against Microcystis aeruginosa, with algicidal rates of 90.48% and 89.51%, respectively (p < 0.01), whereas the washed cells of strain SJ-73 exhibited no inhibitory effect ( Figure 2). This suggested that strain SJ-73 exerted its algicidal activity mainly by secreting active compound(s).  The algicidal mode of strain SJ-73 was determined by comparing the lytic activity among the bacterial culture, cell-free filtrate, and washed cells against the Microcystis aeruginosa. The bacterial culture of strain SJ-73 and the cell-free filtrate exhibited significant inhibition against Microcystis aeruginosa, with algicidal rates of 90.48% and 89.51%, respectively (p < 0.01), whereas the washed cells of strain SJ-73 exhibited no inhibitory effect ( Figure 2). This suggested that strain SJ-73 exerted its algicidal activity mainly by secreting active compound(s).

Isolation and Identification of Indigenous Algicidal Bacteria
Bacterium strain SJ-73, which causes the lysis of M. aeruginosa, was isolated from the sediment sample of Lake Taihu during a cyanobacterial bloom period. It was a Grampositive and rod-shaped bacterium (0.3-0.5 μm × 1-3 μm) (Figure 1), which formed white colonies with a smooth surface on Gause's and LB plates. Phylogenetic analysis based on 16S rDNA sequence revealed that strain SJ-73 was affiliated with the genus Paenibacillus, sharing the highest similarity (99.93%) with the Paenibacillus amylolyticus strain HSE1 (accession number: MK764982.1) ( Figure S1).

Algicidal Model and Spectrum
The algicidal mode of strain SJ-73 was determined by comparing the lytic activity among the bacterial culture, cell-free filtrate, and washed cells against the Microcystis aeruginosa. The bacterial culture of strain SJ-73 and the cell-free filtrate exhibited significant inhibition against Microcystis aeruginosa, with algicidal rates of 90.48% and 89.51%, respectively (p < 0.01), whereas the washed cells of strain SJ-73 exhibited no inhibitory effect ( Figure 2). This suggested that strain SJ-73 exerted its algicidal activity mainly by secreting active compound(s). Strain SJ-73 had a relatively broad algicidal spectrum, the common bloom-forming cyanobacteria species of Microcystis, Anabaena and Aphanizomenon were all susceptible to this bacterium, with algicidal rates of 83.97%, 71.35% and 71.04%, respectively, after 7 days' treatment with 5% (v/v) fermentation filtrate of strain SJ-73. By contrast, SJ-73 showed weak inhibitory activity toward unicellular green algae Chlorella (3.67%) and Tetradesmus (8.03%) ( Table 1). As shown in Figure 3, upon the addition of a cell-free filtrate of SJ-73, for both strains (PCC 7806 and TH1701), the cell growth and photosynthetic activities were significantly suppressed. Meanwhile, a serious release of MCs was observed following cell rupture in the treated groups.
For the axenic unicellular strain M. aeruginosa PCC 7806, when 5% (v/v) cell-free filtrate was added, the obvious decrease in the content of Chl-a occurred at day 3, and the algicidal rate of strain SJ-73 against M. aeruginosa PCC 7806 reached 95.87% on day 9, while the Chl-a content of the control increased by 7.85-fold compared with the initial content ( Figure 3A). The maximum quantum yield of PSII (Fv/Fm) of M. aeruginosa PCC7806 increased on the first day and then dramatically declined from 0.52 to 0 at day 4, whereas the Fv/Fm in the control remained around 0.494 ( Figure 3B). The addition of filtrate of SJ-73 caused marked MC release, the concentration of extracellular MCs in the treated group steadily increased from 8.31 µg/L to 36.27 µg/L after co-cultivation for 2 days, and the maximum release of MCs (140.40 µg/L) occurred on day 8, while the concentration of MCs in the control group was only half (64.65 µg/L) ( Figure 3C).
For the non-axenic colonial strain M. aeruginosa TH1701, when two concentrations (5% and 10%, v/v) of cell-free filtrate were added, the Chl-a content of treated M. aeruginosa TH1701 decreased dramatically from day 0. The algicidal rates for 5% (v/v) and 10% (v/v) treated groups were 47.66% and 59.58%, respectively, at day 1. And reached 92.10% and 94.38% at day 7 ( Figure 3D). The Fv/Fm of M. aeruginosa TH1701 declined from 0.29 to 0.11 and 0.04 within 1 day, and were not detected after day 3 ( Figure 3E). The concentration of extracellular MCs in the 5% (v/v) and 10% (v/v) treated groups increased markedly; the maximum concentration of MCs reached 122.77 µg/L and 152.84 µg/L at day 7, respectively ( Figure 3F). Overall, the algicidal rate was slightly higher in the 10% (v/v) than in the 5% (v/v) treated group; however, the 10% (v/v) treated group displayed more severe MC release during the late period. Microorganisms 2021, 9, x FOR PEER REVIEW 9 of 18

Changes in Cell Morphology
The changes of the cell surface and ultrastructure of treated M. aeruginosa PCC7806 cells were recorded using an SEM and TEM, and samples were collected at day 0, day 3, and day 7. On day 0, the intensive and intact microsphere cells were observed: the cells were plump and rounded with smooth surfaces ( Figure 4A); thylakoids and gas vesicles were stacked in an orderly and tight manner in the cytoplasm, and the cell wall was distinct ( Figure 4B). On day 3, part of the cells had shrunk, and deformed cells with rough surfaces were observed ( Figure 4C); the thylakoids and gas vesicles were disrupted, the cytoplasm was disordered, and the cell wall was obscure, but still maintained the basic shape ( Figure 4D). On day 7, the integrity of the cells was severely damaged, (Figure 4E),

Changes in Cell Morphology
The changes of the cell surface and ultrastructure of treated M. aeruginosa PCC7806 cells were recorded using an SEM and TEM, and samples were collected at day 0, day 3, and day 7. On day 0, the intensive and intact microsphere cells were observed: the cells were plump and rounded with smooth surfaces ( Figure 4A); thylakoids and gas vesicles were stacked in an orderly and tight manner in the cytoplasm, and the cell wall was distinct ( Figure 4B). On day 3, part of the cells had shrunk, and deformed cells with rough surfaces were observed ( Figure 4C); the thylakoids and gas vesicles were disrupted, the cytoplasm was disordered, and the cell wall was obscure, but still maintained the basic shape ( Figure 4D). On day 7, the integrity of the cells was severely damaged, ( Figure 4E), and the cell wall appeared disrupted; thylakoids and gas vesicles could not be distinguished; and cytoplasmic vacuolation and cytoclasis were evident ( Figure 4F). These results suggested that the algicidal compound secreted by SJ-73 could destroy the cell membrane system and cell wall.
Microorganisms 2021, 9, x FOR PEER REVIEW 10 of 18 and the cell wall appeared disrupted; thylakoids and gas vesicles could not be distinguished; and cytoplasmic vacuolation and cytoclasis were evident ( Figure 4F). These results suggested that the algicidal compound secreted by SJ-73 could destroy the cell membrane system and cell wall.

Expression of MlrA
The degradation activity of the crude MlrA enzyme from recombinant E. coli cells was monitored using HPLC-PDA. As shown in Figure 5A The degradation activity of the crude MlrA enzyme from recombinant E. coli cells was monitored using HPLC-PDA. As shown in Figure 5A Figure 5C). The pseudo-first-order rate constant was up to 9.36 h −1 . The minimum degradable MC concentration was below 0.8 μg/L ( Figure 5D) as assessed using UPLC-MS detection.

Degradation Efficiencies of MlrA under Different pH and Temperature Conditions
Single factor experiments were performed (results shown in Figure S2), and MlrA degradation efficiencies fluctuated in the pH range of 6.0-9.0. The optimal pH of the reaction was 8.0 and the degradation ratio was 80.21%, while the ratio reduced to 64.59% when the pH decreased to 6.0 ( Figure S2A). The optimal temperature was 30 • C, at which degradation efficiency reached 93.43%; however, the degradation efficiency of MlrA decreased to below 55.01% at 10 • C, while it remained above 88.0% when the temperature increased to 35 • C ( Figure S2B).  To acquire the optimal application conditions of MlrA to eliminate MCs caused by algicidal bacteria lysis, a preliminary experiment was performed to test the MlrA activity in SJ-73 treated unicellular M. aeruginosa cultures. MlrA at 0.8 mg/L was added to an M. aeruginosa PCC7806 culture that had been treated with 5% (v/v) SJ-73 filtrate for 3 days. As shown in Figure 6, the MC content decreased sharply from 125.89 µg/L to 3.39 µg/L after the addition of MlrA for 0.5 h, a decline of almost 97.30%, suggesting the quick elimination of MCs. The MC content remained consistently low until 48 h, before it increased slightly to 39.16 µg/L, indicating that the MlrA activity had decreased and was insufficient to degrade the persistently released MCs from incompletely cracked cells. Thus, a second dose of MlrA supplementation was necessary as the bacterial lysis of algae was relatively mild and the MlrA could not remain stable under 25 • C for a long period of time.
degradation efficiencies fluctuated in the pH range of 6.0-9.0. The optimal pH of the reaction was 8.0 and the degradation ratio was 80.21%, while the ratio reduced to 64.59% when the pH decreased to 6.0 ( Figure S2A). The optimal temperature was 30 °C, at which degradation efficiency reached 93.43%; however, the degradation efficiency of MlrA decreased to below 55.01% at 10 °C, while it remained above 88.0% when the temperature increased to 35 °C ( Figure S2B).  To acquire the optimal application conditions of MlrA to eliminate MCs caused by algicidal bacteria lysis, a preliminary experiment was performed to test the MlrA activity in SJ-73 treated unicellular M. aeruginosa cultures. MlrA at 0.8 mg/L was added to an M. aeruginosa PCC7806 culture that had been treated with 5% (v/v) SJ-73 filtrate for 3 days. As shown in Figure 6, the MC content decreased sharply from 125.89 μg/L to 3.39 μg/L after the addition of MlrA for 0.5 h, a decline of almost 97.30%, suggesting the quick elimination of MCs. The MC content remained consistently low until 48 h, before it increased slightly to 39.16 μg/L, indicating that the MlrA activity had decreased and was insufficient to degrade the persistently released MCs from incompletely cracked cells. Thus, a second dose of MlrA supplementation was necessary as the bacterial lysis of algae was relatively mild and the MlrA could not remain stable under 25 °C for a long period of time.  The efficiency of supplementation of MlrA following the treatment of colonial M. aeruginosa TH1701 culture with SJ-73 filtrate (5% and 10%, v/v) was tested. The MC concentration increased drastically from 5.61 µg/L to 118.49 µg/L and 152.841µg/L, respectively, after treatment with 5% and 10% (v/v) SJ-73 filtrate for one day. The MCs content were trending higher in the following days, achieving 21.16 and 27.29-fold increases compared with the initial content at day 7 ( Figure 7A,B). As shown in Figure 7A, MlrA was only added once at day 0, after which the MC level decreased immediately to below 15.00 µg/L, and then increased slightly at day 3, indicating that a second MlrA supplementation was needed. As shown in Figure 7B, when MlrA was added twice, on day 0 and day 3, the MCs remained at a low concentration, ranging from 2.55 µg/L to 30.25 µg/L, with the peak value being observed on day 3 in both MlrA treatment groups. After the second MlrA addition at day 3, the MCs immediately decreased again. The MCs removal rates were 79.11% and 94.2% in the 5% and 10% (v/v) SJ-73 filtrate treatment groups, respectively, at day 7, suggesting that the introduction of two supplementations of MlrA enzyme and a higher dose of algicidal bacteria (10%, v/v) produced a better performance in removing MCs.

Sustainability of Free
The efficiency of supplementation of MlrA following the treatment of colonial M. aeruginosa TH1701 culture with SJ-73 filtrate (5% and 10%, v/v) was tested. The MC concentration increased drastically from 5.61 μg/L to 118.49 μg/L and 152.841μg/L, respectively, after treatment with 5% and 10% (v/v) SJ-73 filtrate for one day. The MCs content were trending higher in the following days, achieving 21.16 and 27.29-fold increases compared with the initial content at day 7 ( Figure 7A, 7B). As shown in Figure 7A, MlrA was only added once at day 0, after which the MC level decreased immediately to below 15.00 μg/L, and then increased slightly at day 3, indicating that a second MlrA supplementation was needed. As shown in Figure 7B, when MlrA was added twice, on day 0 and day 3, the MCs remained at a low concentration, ranging from 2.55 μg/L to 30.25 μg/L, with the peak value being observed on day 3 in both MlrA treatment groups. After the second MlrA addition at day 3, the MCs immediately decreased again. The MCs removal rates were 79.11% and 94.2% in the 5% and 10% (v/v) SJ-73 filtrate treatment groups, respectively, at day 7, suggesting that the introduction of two supplementations of MlrA enzyme and a higher dose of algicidal bacteria (10%, v/v) produced a better performance in removing MCs.

Discussion
Algae-lysing bacteria are considered to be an effective and eco-friendly method to control the prevalence of cyanobacterial blooms. The selection of algae-lysing bacteria should consider the source and prey spectrum, to avoid exotic species invasion, the introduction of exogenous hazards, and detrimental effects on non-targeted organisms [41]. Isolation of the indigenous algicidal bacteria is the best option in a practical water body bioremediation process. In the present study, an indigenous algicidal bacterium, Paenibacillus amylolyticus strain SJ-73, was isolated from the sediments sampled from Taihu Lake. The highest algicidal rate of SJ-73 reached 93.88%. Microcystis blooms generally occur in the colony form and often coexist with several other cyanobacterial species under natural conditions; however, previous laboratory research mainly focused on the effect of water treatment processes on single Microcystis cells only. In this study, strong algicidal activity of Paenibacillus sp. SJ-73 against the major bloom-forming species, including colonial Microcystis and filamentous Aphanizomenon and Anabaena, was observed. Moreover, SJ-73 was also demonstrated to have no adverse impact on green algae. These results

Discussion
Algae-lysing bacteria are considered to be an effective and eco-friendly method to control the prevalence of cyanobacterial blooms. The selection of algae-lysing bacteria should consider the source and prey spectrum, to avoid exotic species invasion, the introduction of exogenous hazards, and detrimental effects on non-targeted organisms [41]. Isolation of the indigenous algicidal bacteria is the best option in a practical water body bioremediation process. In the present study, an indigenous algicidal bacterium, Paenibacillus amylolyticus strain SJ-73, was isolated from the sediments sampled from Taihu Lake. The highest algicidal rate of SJ-73 reached 93.88%. Microcystis blooms generally occur in the colony form and often coexist with several other cyanobacterial species under natural conditions; however, previous laboratory research mainly focused on the effect of water treatment processes on single Microcystis cells only. In this study, strong algicidal activity of Paenibacillus sp. SJ-73 against the major bloom-forming species, including colonial Microcystis and filamentous Aphanizomenon and Anabaena, was observed. Moreover, SJ-73 was also demonstrated to have no adverse impact on green algae. These results showed the promising potential of strain SJ-73 in controlling multi-species cyanobacterial blooms with a negligible influence on green algae.
Our study further confirmed that the algicidal mechanism of strain SJ-73 was via secreting extracellular algicidal compounds that could cause an inhibition of photosynthesis and the disruption of cell membranes. Bioactive compounds, e.g., fusaricidin and polymyxin produced by Paenibacillus spp., were documented to have algicidal properties [42,43]. Although the algicidal substances were not characterized in this study, the three gene clusters involved in the synthesis of the antibiotics, polymyxin, bottromycin, and xenocoumacin, were predicted in the whole genome of Paenibacillus sp. SJ-73 using antiSMASH (https://antismash.secondarymetabolites.org) (data not shown). Moreover, a group of glycoside hydrolase (GH) family genes, including those encoding galactose hydrolase, xylosidase, α-mannasidase, and glucosidase, were also found in the SJ-73 genome based on the carbohydrate-active enzymes database (CAZyme) [44], which might play a role in decomposing the exopolysaccharides of colonial Microcystis (data not shown). This information provides a hint regarding the algicidal mechanisms of strain SJ-73.
The MC degradation process mediated by wild bacteria always show low efficiency and requires an adaptation period [45,46]. By contrast, heterologous expression of MlrA could greatly enhance the yields and the biodegradation abilities of MCs, as the MlrA activity of the cell-free extract from heterologous expression was 6800 times greater than that from the natural expression [32]. In our study, the degradation rate of [D-Asp 3 ]MC-LR by crude Sphingopyxis sp. HW MlrA (cell-free extract from recombinant E. coli) was 17.82 mg L −1 h −1 , whereas the value of MC-LR degradation by native Sphingopyxis sp. HW was 0.34 mg L −1 h −1 [47]. Further improvement of MlrA activity could be achieved by substituting a more efficient expression strain, e.g., E. coli C4 (DE3), as the host [48], or using an N-terminal maltose binding protein (MBP) tag, to increase enzyme solubility, stability, and overexpression [49].
The pH and temperature of environmental water bodies fluctuate frequently during the cyanobacterial bloom period; however, these two parameters are crucial for MlrA activity [32,46,48,49]. In this study, the degradation activities of heterologously expressed MlrA were tested under different pH and temperature conditions, revealing the optimal conditions for MC degradation as 30 • C and pH 8.0, which were slightly different to those determined by Dziga [48] (i.e., pH 7.5 and 20 • C). The degradative activities of MlrA were above 60% in the pH range of 6-9 and were maintained above 55% in the temperature range from 10 to 35 • C. These features will enhance the application potential for MlrA in field conditions. Except for pH and temperature, metalloprotease inhibitors (e.g., ophenantroline and EDTA) should also be considered because MlrA is a metalloprotease [32].
Immobilized enzymes have advantages in recycling use and maintaining enzyme activity for a longer time in comparison to free enzymes, and the enzymatic activity of immobilized enzyme is strongly linked to the carrier materials. Excellent MC degrading activity of recombinant cells (BL21_MlrA), immobilized in alginate beads, has been reported [22,50]. Wu [51] prepared CysGO-immobilized MlrA and found that the degradation rate of CysGO-MlrA was slightly lower than that of the free enzyme, and that the degradation efficiency of the enzyme was maintained at >81% of the initial efficiency after 7 cycles of repeated usage. The immobilization of MlrA in alginate beads was attempted in the present study; however, the degradation efficiency of the immobilized MlrA was lower than that of free MlrA. Screening of more immobilization carriers is needed to improve the MlrA utilization efficiency in the future.
Normally, the treatment for toxic cyanobacteria is accompanied by the release of cyanotoxins, increasing the exposure risks for aquatic organisms and humans. A laboratory experiment conducted by Dziga [52] first proposed the combined treatment of M. aeruginosa with hydrogen peroxide (H 2 O 2 ) and MCs biodegradation agents (Sphingomonas sp. or heterologously expressed MlrA enzyme), which eliminated both Microcystis aeruginosa and MCs simultaneously. In addition, the degradation activity of Sphingomonas sp. or the MlrA enzyme was not affected by H 2 O 2 at a working concentration of 50 mg/L. Soon afterwards, this strategy was extended to an in situ mesocosm experiment in a eutrophic shallow lake (Lake Ludoš, Serbia) [48]. However, H 2 O 2 , as a safe algaecide, remains questionable because H 2 O 2 was proved to upregulate MC production at concentrations of 2-3 mg/L [48]. In this study, we first proposed the co-application of algicidal bacteria and heterologous expressed MlrA enzyme, which could remove 92.07% of the Microcystis biomass and 79.11% of the MCs by treatment with 5% (v/v) fermentation filtrate for 7 days. In a preliminary experiment testing the activity on the axenic strain M. aeruginosa PCC 7806, after the supplementation with MlrA, the MC content in the treatment group decreased dramatically and remained at low concentrations. However, after 48 h the MC content began to increase slightly, suggesting the persistent releasing of MCs from incompletely cracked cells and the inactivity of MlrA after 2 days, thus a second supplementation was necessary. In the colonial M. aeruginosa TH1701 group, the introduction of a higher dose of algicidal bacteria (10%, v/v), and the supplementation with MlrA twice, achieved higher removal efficiencies of both algal biomass and MCs. Notably, although the minimum degradable MC concentration was below 0.8 µg/L, the MCs could not be completely removed, which might have been caused by the pH variation in the culture. As shown in Figure S4, in the 7806-MlrA group, the pH value increased from 8.6 to 9.5, while the pH value of TH1701-MlrA group remarkably decreased from 8.5 to 6.5; however, both shifts in pH could result in a decrease of MlrA activity ( Figure S2).

Conclusions
Microcystis aeruginosa blooms and MCs are causing serious environmental and ecological issues. The co-application of algicidal bacteria and MlrA enzymes to remove M. aeruginosa and MCs simultaneously was performed for the first time in this study, and the following conclusion might be formulated: (a) algicidal strain SJ-73 exhibited strong algicidal activities against colonial Microcystis as well as some other cyanobacterial bloom species; (b) the heterologously expressed MlrA enzyme showed high MC degradation efficiency, with a minimum degradable MC concentration below the safety threshold, and MlrA possessed a wide tolerance to pH (6.0-9.0) and temperature (10-35 • C); and (c) the marked release of MCs following cell rupture could be removed by supplementation with the MlrA enzyme; and a dose of 10% algicidal bacteria and two supplementations of MlrA enzyme exhibited the best performance. The outcome will provide valuable information for the design of strategies to deal with algae-contaminated water.
Supplementary Materials: The following are available online at https://www.mdpi.com/article/ 10.3390/microorganisms9081594/s1, Figure S1: Phylogenetic tree based on bacterial 16S rDNA sequence of the isolated algicidal bacterium SJ-73 and closely related members. Figure S2: Influences of reaction pH (A) and temperature (B) on the activity of MlrA. Figure S3: The relative activity of the immobilized and free enzyme at different storage temperature (4 • C and 25 • C). Figure S4