Microcystis Sp. Co-Producing Microcystin and Saxitoxin from Songkhla Lake Basin, Thailand

The Songkhla Lake Basin (SLB) located in Southern Thailand, has been increasingly polluted by urban and industrial wastewater, while the lake water has been intensively used. Here, we aimed to investigate cyanobacteria and cyanotoxins in the SLB. Ten cyanobacteria isolates were identified as Microcystis genus based on16S rDNA analysis. All isolates harbored microcystin genes, while five of them carried saxitoxin genes. On day 15 of culturing, the specific growth rate and Chl-a content were 0.2–0.3 per day and 4 µg/mL. The total extracellular polymeric substances (EPS) content was 0.37–0.49 µg/mL. The concentration of soluble EPS (sEPS) was 2 times higher than that of bound EPS (bEPS). The protein proportion in both sEPS and bEPS was higher than the carbohydrate proportion. The average of intracellular microcystins (IMCs) was 0.47 pg/cell on day 15 of culturing, while extracellular microcystins (EMCs) were undetectable. The IMCs were dramatically produced at the exponential phase, followed by EMCs release at the late exponential phase. On day 30, the total microcystins (MCs) production reached 2.67 pg/cell. Based on liquid chromatograph-quadrupole time-of-flight mass spectrometry, three new MCs variants were proposed. This study is the first report of both decarbamoylsaxitoxin (dcSTX) and new MCs congeners synthesized by Microcystis.


Introduction
The overgrowth of cyanobacteria leads to critical environmental problems and ecological events, including hypoxia, reducing water clarity, and imbalance of food webs. Among bloom-forming cyanobacteria, Microcystis sp. are the most abundantly reported [1]. The organisms are capable of producing several toxins, including hepatotoxic microcystins (MCs) [2] and neurotoxins (anatoxin-a, cyanopeptolin, and β-N-methylamino-L-alanine) [3]. MCs cause liver failure in humans and animals and increase the risks of primary liver cancer in humans [4]. Co-production of MCs and other toxins has been reported in certain Microcystis spp. Microcystis producing both MCs and saxitoxins (STXs) were isolated from a Brazilian reservoir [5] and freshwaters in Scotland and Brazil [6]. Saxitoxins, comprising mand (BOD) and chemical oxygen demand (COD) were 1. 58-5.4 mg/L and 28-50 mg/L, respectively (Table 1). Total phosphorous and total nitrogen were 0.08-0.24 mg/L and 0.58-1.15 mg/L, respectively. Other water quality characteristics included dissolved oxygen (DO) of 1.82-7.02 mg/L, pH of 6.7-8. 34, and salinity of 0.1-0.3 g/L. The lowest DO level (1.82 mg/L) that appeared at SG might have resulted from restricted water exchange, wind calm, and high tidewater. Salinity (0.3 g/L) indicated that two water samples from Songkhla Lake, Songkhla, were brackish. The onsite temperature of the samples ranged from 29 to 32 • C. from Hat Yai municipal wastewater treatment plant (HMWTP), which was effluent water drained to SLB ( Figure 1). All samples were collected in January, the end of the rainy season (September-December). Nutrients easily load and accumulate in the lake, favoring cyanobacterial proliferation [28]. For all samples, biochemical oxygen demand (BOD) and chemical oxygen demand (COD) were 1. 58-5.4 mg/L and 28-50 mg/L, respectively (Table  1). Total phosphorous and total nitrogen were 0.08-0.24 mg/L and 0.58-1.15 mg/L, respectively. Other water quality characteristics included dissolved oxygen (DO) of 1.82-7.02 mg/L, pH of 6.7-8. 34, and salinity of 0.1-0.3 g/L. The lowest DO level (1.82 mg/L) that appeared at SG might have resulted from restricted water exchange, wind calm, and high tidewater. Salinity (0.3 g/L) indicated that two water samples from Songkhla Lake, Songkhla, were brackish. The onsite temperature of the samples ranged from 29 to 32 °C.    Cyanobacteria were isolated from the water samples. Ten green colonies were selected and observed under an optical microscope. The cells ranged from 1 to 5 µm in diameter with a round shape and green color. The 16S rDNA analysis indicated that all strains fell into Microcystaceae family and Microcystis genus. The phylogeny based on 16S rDNA sequences was generated by the neighbor-joining tree method; compared with M. aeruginosa, M. panniformis, M. ichthyoblabe, M. wesenbergii, M. viridis, and Gloeothece membranacea; and deposited in the NCBI database ( Figure 2). Most Microcystis isolates were closely related to M. aeruginosa strains NIES-843 (NR_074314.1), except WIN01. SG12, SG03, and SG11 isolated from the exact location were in the same phylogenic group. SH13, SH12, and SF11 also showed similar 16S rDNA gene sequences. Three Microcystis strains isolated from Songkhla Lake (SG10), Thale Noi (PB07), and HMWTP (WOT01) fell into the same cluster. Our results reveal that despite the absence of cyanobacterial bloom, Microcystis was detected. Cyanoblooom is facilitated by other environmental factors, including nutrients, carbon dioxide, weather conditions, water bodies, salinity, heavy metals, and sunlight [29,30].
Cyanobacteria were isolated from the water samples. Ten green colonies were se lected and observed under an optical microscope. The cells ranged from 1 to 5 µm in di ameter with a round shape and green color. The 16S rDNA analysis indicated that al strains fell into Microcystaceae family and Microcystis genus. The phylogeny based on 16S rDNA sequences was generated by the neighbor-joining tree method; compared with M aeruginosa, M. panniformis, M. ichthyoblabe, M. wesenbergii, M. viridis, and Gloeothece mem branacea; and deposited in the NCBI database ( Figure 2). Most Microcystis isolates were closely related to M. aeruginosa strains NIES-843 (NR_074314.1), except WIN01. SG12 SG03, and SG11 isolated from the exact location were in the same phylogenic group. SH13 SH12, and SF11 also showed similar 16S rDNA gene sequences. Three Microcystis strains isolated from Songkhla Lake (SG10), Thale Noi (PB07), and HMWTP (WOT01) fell into the same cluster. Our results reveal that despite the absence of cyanobacterial bloom, Mi crocystis was detected. Cyanoblooom is facilitated by other environmental factors, includ ing nutrients, carbon dioxide, weather conditions, water bodies, salinity, heavy metals and sunlight [29,30].

Microcystin Genes and a Saxitoxin Gene Analysis
Microcystin genes and a saxitoxin gene were investigated in all 10 strains. The MC gene cluster is arranged in two divergently transcribed operons (ORFs), mcy A-C and mcy D-J. The mcy A-C operon is engaged in synthesizing amino acids attached to Adda through a subsequent modification or tailoring by mcyJ, mcyE, and mcyF genes [31]. The mcyA genes are widely employed to detect the toxigenic Microcystis genus [32][33][34]. Ou results show that all isolates carried mcyA ( Table 2). The mcyA a primers detected the targe gene in 5 out of 10 strains, while the mcyA gene was detected in 9 of 10 strains by mcyA primers. The mcyA b primers were created based on nucleotides of toxigenic Microcysti and Planktothrix strains [35]. However, certain non-toxic strains harbor the mcy genes

Microcystin Genes and a Saxitoxin Gene Analysis
Microcystin genes and a saxitoxin gene were investigated in all 10 strains. The MC gene cluster is arranged in two divergently transcribed operons (ORFs), mcy A-C and mcy D-J. The mcy A-C operon is engaged in synthesizing amino acids attached to Adda through a subsequent modification or tailoring by mcyJ, mcyE, and mcyF genes [31]. The mcyA genes are widely employed to detect the toxigenic Microcystis genus [32][33][34]. Our results show that all isolates carried mcyA ( Table 2). The mcyA a primers detected the target gene in 5 out of 10 strains, while the mcyA gene was detected in 9 of 10 strains by mcyA b primers. The mcyA b primers were created based on nucleotides of toxigenic Microcystis and Planktothrix strains [35]. However, certain non-toxic strains harbor the mcy genes [36,37]. Our study suggests that the mcyA primers might be restricted by a small nucleotide database of the toxic strains and that mcyA sequences are divergent among Microcystis isolates [32]. A report on MCs detection in Greece revealed that despite the presence of the toxin, the mcyA gene was not detected in freshwater samples [38]. Therefore, MCs prediction strictly based on molecular assay might be insufficient. Combining molecular techniques (PCR) and other techniques such as chemical methods (HPLC) and immunoassay will assure the MC presence. Our results show that 5 of 10 strains were positive for sxtA and also carried mcyA ( Table 2). The STXs biosynthesis initiates via PKSs-like enzymes encoded by sxtA [39]. The sxtA is a core gene among the gene clusters for the synthesis of STX and analogs [40,41]. This gene has been widely used to detect STX producers [42].

Growth and EPS Composition
The Microcystis density and Chl-a content were measured. The specific growth rate of all isolates was 0.2-0.3 day −1 , as shown in Table 2. As a general metric for algal biomass, Chl-a was employed as a reference for cyanoblooms [43]. Our results show that Chl-a production was consistent with bacterial growth. Chl-a was 0.1 to 4 µg/mL from day 0 to day 15.
Extracellular polymeric substance (EPS) of Microcystis is produced as part of the metabolism process. It forms and surrounds the cell surface as a protective barrier between the cell and the external environment [12,44,45]. The content, types, and composition of EPS also facilitate colony formation, impacting the distribution and persistence of cyanobacteria [46][47][48]. A study by Holland et al. suggested that Microcystis with increased colony size contained higher MCs [49]. Therefore, correlations between MCs and the content and composition of EPS are needed to be revealed.
Here, we determined EPS content and composition in 10 Microcystis isolates ( Figure 3). The total EPS varied from 0.37 to 0.49 µg/mL. The highest total EPS concentration was found in SG10. For all isolates, the average soluble EPS (sEPS) concentration ( Figure 3b) was roughly twice higher than the average bound EPS (bEPS) concentration ( Figure 3a) (0.27 vs. 0.14 µg/mL). Our results confirm the finding by Xiao et al. (2018), that Microcystis are commonly unicellular cells under an axenic condition [12]. High sEPS content was found in unicellular cells, while increasing bEPS content could be concluded for colony formation [46]. The proportion between protein and carbohydrate in sEPS in 10 isolates was similar. In contrast, the protein proportion in bEPS was higher, ranging from 2.014 to 3.056 at an average value of 2.369, which was in agreement with a previous study [50]. However, the proportions of bEPS in this study were inconsistent with Xu et al.  [51].
During cyanobloom, large aggregated colonies form scum floating on a water body's surface [52]. This upward cell migration is enhanced by photosynthetic activity resulting in O 2 saturation and nucleation into bubbles [13,53] and gas vesicles in Microcystis cells [47]. The O 2 bubble is trapped inside EPS, indicating that scum formation is highly correlated with the amount of EPS production [13,54]. Here, the scum formation was enhanced Toxins 2021, 13, 631 6 of 17 by culturing in high concentration media and was found in SG03, SG11, and SH13 after culturing for 2 to 7 days, in agreement with previous studies [13,53]. The remaining isolates did not generate scums over the incubation period. Future work should be conducted to determine the effects of EPSs composition on scum formation and large aggregated colonies.  During cyanobloom, large aggregated colonies form scum floating on a water body's surface [52]. This upward cell migration is enhanced by photosynthetic activity resulting in O2 saturation and nucleation into bubbles [13,53] and gas vesicles in Microcystis cells [47]. The O2 bubble is trapped inside EPS, indicating that scum formation is highly correlated with the amount of EPS production [13,54]. Here, the scum formation was enhanced by culturing in high concentration media and was found in SG03, SG11, and SH13 after culturing for 2 to 7 days, in agreement with previous studies [13,53]. The remaining isolates did not generate scums over the incubation period. Future work should be conducted to determine the effects of EPSs composition on scum formation and large aggregated colonies.

Microcystins Production of Microcystis Sp.
After 15 days of culturing intracellular microcystins (IMCs) and extracellular microcystins (EMCs), production in all Microcystis isolates was determined by ELISA. This technique is recommended to quantify several MCs variants in cyanobacterial cells and to track MCs production in water bodies [55,56]. The test kit is designed to detect Adda, which is a non-protein amino acid at position 5 in MCs, with cross-reactivity between different MCs variants [57]. Our results show that high IMCs production ranging from 210 to 314 µg/mL was observed in Microcystis isolated from Songkhla Lake. Microcystis sp. isolated from HMWTP was approximately 280 µg/mL of MCs ( Table 2). The EMCs were not detected in any isolates. The absence of EMCs might be the result of low cell density after 15 days of culturing (10 7 CFU/mL) and the kit's sensitivity. More prolonged cultivation than 15 days resulted in EMCs production ( Figure 4b). SG03 produced IMCs at the highest concentration of 1.3 pg/cell, while SH13 was the lowest in IMCs production at 0.14 pg/cell. Microcystis SG03 was observed under SEM. As seen in Figure 4a, with round shape, cell diameter ranged from 3 to 5 µm.

Microcystins Production of Microcystis Sp.
After 15 days of culturing intracellular microcystins (IMCs) and extracellular microcystins (EMCs), production in all Microcystis isolates was determined by ELISA. This technique is recommended to quantify several MCs variants in cyanobacterial cells and to track MCs production in water bodies [55,56]. The test kit is designed to detect Adda, which is a non-protein amino acid at position 5 in MCs, with cross-reactivity between different MCs variants [57]. Our results show that high IMCs production ranging from 210 to 314 µg/mL was observed in Microcystis isolated from Songkhla Lake. Microcystis sp. isolated from HMWTP was approximately 280 µg/mL of MCs ( Table 2). The EMCs were not detected in any isolates. The absence of EMCs might be the result of low cell density after 15 days of culturing (10 7 CFU/mL) and the kit's sensitivity. More prolonged cultivation than 15 days resulted in EMCs production ( Figure 4b). SG03 produced IMCs at the highest concentration of 1.3 pg/cell, while SH13 was the lowest in IMCs production at 0.14 pg/cell. Microcystis SG03 was observed under SEM. As seen in Figure 4a, with round shape, cell diameter ranged from 3 to 5 µm.

Kinetics of MCs Production Based on Microcystis Growth
Growth and MCs kinetics of SG03 were investigated from day 0 to day 30 (Figure 4b). The initial IMCs content at 0.03 pg/cell on day 0 was detected in the inoculum (6 × 10 6 CFU/mL). On day 5, the specific growth rate (µ) based on cell number was 0.522 day −1 , indicating that the exponential growth phase followed the late exponential growth phase from day 5 to 15 (µ = 0.186 day −1 ). The stationary growth phase spanned from day 20 to day 30 (µ = 0.002 day −1 ). MCs production was determined along the growth curve. The IMCs were dramatically produced at 1.94 pg/cell during the exponential growth phase (day 10), and then the production was steady when cyanobacteria entered the stationary growth phase (Figure 4b). The EMC production (0.06 pg/cell) was measured in the late exponential growth phase and dramatically rose on day 20 (0.66 pg/cell) during stationary growth, indicating cell death. Our results suggest that the MCs were the primary metabolite produced intracellularly at the log phase. Our finding is well supported by previous work in which the maximum MCs production of Microcystis was found in the log phase and persisted for months [43,58,59]. Studies have shown that MCs production rescues the cells from damage by photosynthesis, reactive oxygen species [60], and imbalance of C:N by redirecting primary metabolism in cells [61].

Kinetics of MCs Production Based on Microcystis Growth
Growth and MCs kinetics of SG03 were investigated from day 0 to day 30 (F 4b). The initial IMCs content at 0.03 pg/cell on day 0 was detected in the inoculum ( CFU/mL). On day 5, the specific growth rate (µ) based on cell number was 0.522 indicating that the exponential growth phase followed the late exponential growth from day 5 to 15 (µ = 0.186 day −1 ). The stationary growth phase spanned from day day 30 (µ = 0.002 day −1 ). MCs production was determined along the growth curv IMCs were dramatically produced at 1.94 pg/cell during the exponential growth (day 10), and then the production was steady when cyanobacteria entered the stati growth phase (Figure 4b). The EMC production (0.06 pg/cell) was measured in th exponential growth phase and dramatically rose on day 20 (0.66 pg/cell) during stati growth, indicating cell death. Our results suggest that the MCs were the primary m

Discussion
Microcystis blooms cause ecosystem loss as well as threats to public health and animals. Microcystis possesses a well-known hepatotoxin, MCs; a few strains also produce neurotoxin (β-N-methylamino L-alanine and paralytic shellfish poison) [63]. However, no STX biosynthesis genes have been identified in Microcystis.
Various factors facilitate the abundance of cyanobacteria in aquatic environments. EPSs prolong cyanobacterial persistence in water and protect cells from the environment. Cyanobacterial aggregation and scum formation can be promoted by higher EPS content

Discussion
Microcystis blooms cause ecosystem loss as well as threats to public health and animals. Microcystis possesses a well-known hepatotoxin, MCs; a few strains also produce neurotoxin (β-N-methylamino L-alanine and paralytic shellfish poison) [63]. However, no STX biosynthesis genes have been identified in Microcystis. Various factors facilitate the abundance of cyanobacteria in aquatic environments. EPSs prolong cyanobacterial persistence in water and protect cells from the environment. Cyanobacterial aggregation and scum formation can be promoted by higher EPS content [13,47,64]. The EPS composition involves colony size and colony formation [12,65]. Microcystis with increased colony size was reported to produce higher MCs [49]. Our results show that the average sEPS concentration was higher than the average bEPS concentration as a unicellular cell under an axenic condition. There was no statistical difference between EPS content and composition among all isolates. Therefore, correlations between MCs and EPS content and composition cannot be defined in this study. Associations of MCs and EPS types, content, and composition still need to be discovered.
The MCs can contaminate different environmental compartments, including sediment, animals, and aquatic and terrestrial plants [66]. MCs produced by Microcystis could persist for months in axenic cultures [67]. In our study, IMC production of Microcystis SG03 remained constant after 20 days of incubation. The averaged MCs production was 1.91 pg/cell on day 30 of culturing, with stable cell density. Under low nutrient/starvation conditions with sufficient light, the SG03 cells entered the stationary phase without changing MCs production. However, other studies reported increasing MCs production under nutrient starvation [68,69]. During stress conditions, MCs synthesis is enhanced by phycobilisomes degradation and by the toxins combined with phycobilins, stabilizing cell integrity [68,70]. Furthermore, MCs might protect the cells from oxidative stress under insufficient nutrients [71]. Our study indicates that Microcystis can produce MCs and survive for a long time, suggesting a need for a vigilant surveillance program on toxigenic Microcystis and MCs in aquatic areas.
Here, Glu(OMe) was found in compounds 2 and 4. The amino acid residue in position 7 is variable with commonly found N-methyldehydroalanine (Mdha). The Mdhb existed in compound 2, while compound 4 possessed tentative Mala. Even though several MC congeners have been presented in the environment and laboratories, their toxicity has been inadequately assessed.
The SLB has a unique ecosystem as fresh, brackish, and saline water lead to plenty of bioresources and high biodiversity. This area has benefited aquaculture, agriculture, factories, tourist resorts, and housing developments. Here, we isolated toxigenic Microcystis, which produced both MCs and STXs, from SLB with no observed blooms. Christensen et al. (2019) reported the production of both toxins before visible blooms [88]. When season changes combine with other environmental factors, harmful cyan blooms occur. In Thailand, MC accumulation was found in prawn farms [89,90], fish ponds [90][91][92], and recreational reservoirs [93,94], especially in the temperate region of the country with high N and P concentrations. In contrast, the reports of STX from Thailand were associated with only freshwater puffers [95,96]. The observation of MCs-and STXs-producing cyanobacteria in our study calls for more research on harmful cyanobacterial communities and toxin distribution in recreation areas and aquaculture farms in Songkhla Lake.

Conclusions
This is the first report of Microcystis in Songkhla Lake, Thailand. In axenic culture, Microcystis produced sEPS 2 times more than bEPS. The ratio of protein and carbohydrate was equal in sEPS, while the protein fraction was higher than carbohydrate in bEPS. All isolates produced MCs. One isolate (SG03) presented both MCs and dsSTX. SG03 produced three new tentative MC variants as undescribed congeners. The MCs accumulation in SG03 was observed until 30 days of incubation, and the toxin was released extracellularly on day 15. The IMCs were synthesized mainly in the middle exponential growth phase, while the EMCs were initially produced at the stationary growth phase. The rising of total MCs synthesis was observed during the exponential growth phase and was constantly produced during the stationary growth phase. This study indicates that harmful cyanobacteria could present in water bodies without visible blooms. The unique complexity of SLB provided groups of Microcystis producing novel toxic compounds. We exhibited that Microcystis extended some distance into brackish and/or marine water bodies. Further studies should focus on the effects of water quality parameters, including N and P on Microcystis aggregation, scum formation, and toxin production, thereby preventing harmful effects of cyanobacteria and their toxins.

Water Quality Characterization
Four lake water samples from Songkhla Lake Basin (SLB), Thailand (7687 km 2 of land area and 1042 km 2 of the Lake surface), were collected on 25 January 2019, without visible blooms (Figure 1). These locations were located close to urban settlements. Effluent water drained to the SLB was collected from the Hat Yai municipal wastewater treatment plant (HMWTP) on 25 January 2019 (Figure 1). The 5 samples containing 1 L with duplicates were taken at a 50 cm depth and kept on ice during transportation. The water quality, including biochemical oxygen demand (BOD), chemical oxygen demand (COD), total phosphorus (TP), total Kjeldahl nitrogen (TKN), temperature, pH, and salinity was characterized. All samples were used for Microcystis isolation in further experiments.

Microcystis Spp. Isolation
Ten milliliters of water samples and the duplicates were enriched in 10 ml of 2 × BG-11 broth (pH 7.5) [97]. The samples were incubated under a 12:12 h light/dark cycle (25 µmol photon m −2 /s, cool fluorescent light) at 25 • C for 14 days and hand-shaken three times a day. The enriched samples were then streaked onto BG-11 agar plates supplemented with imipenem (10 µg/mL) [97], followed by incubation for 7-14 days under the above conditions. The isolated green bacterial colonies were obtained, and their morphology was documented. Characteristics of the cyanobacterial cells were observed under an optical microscope.

Cyanobacterial DNA Extraction
All cyanobacterial isolates were cultivated in 50 mL BG-11 under the above conditions for 10 days. Then, DNA from each isolate was extracted. The DNA extraction method was modified from Hisbergues et al. (2003) and Rogers and Bendich (1989). Briefly, the culture was centrifuged at 10,000× g for 5 min, and the pellet was resuspended in 300 µL suspension buffer (50 mM Tris-HCL (pH 8), 5 mM EDTA (pH 8), and 50 mM NaCl), followed by the addition of lysozyme (2 mg/L) and proteinase K (2 mg/mL). The cell suspension was incubated at 37 • C for 30 min and at 60 • C for 10 min. The mixture was added with an equal volume of lysis buffer (2% CTAB, 2% SDS, 1.4 M NaCl, 20 mM EDTA (pH 8), 100 mM Tris-HCl (pH 8), 2% 2-mercaptoethanol) and incubated at 70 • C for 1 h. The DNA was separated by adding an equal volume of phenol/chloroform/isoamyl alcohol (25:24:1, v/v), followed by centrifugation at 10,000× g for 15 min. The supernatant was collected and mixed with an equal volume of chloroform/isoamyl alcohol (24:1, v/v). Then, the mixture was centrifuged at 10,000× g for 15 min. The supernatant containing DNA was precipitated with 0.3 volume of 3 M sodium acetate and 1 volume of isopropanol and kept at −20 • C for 2 h. After centrifugation at 10,000× g for 20 min, the DNA pellet was obtained and washed with 1 mL of 70% ethanol. The pellet was air-dried and dissolved with 30 µL sterilized deionized water and kept at −20 • C.
Cyanobacterial identification was based on 16S rDNA sequencing. Briefly, the 16S rDNA was amplified using the primer 27F and 809R [99]. The PCR product was purified using a GenepHlow TM Gel/PCR kit and sequenced (Novogene co. Ltd., Cambridge, UK). Sequences were aligned using ClustalW followed by comparing their similarity with the other sequences in the Genbank database. A phylogenetic tree was constructed via a neighbor-joining (NJ) method by MEGA 7.0 software with 1000 replicates of bootstrap.

Measurements of Specific Growth Rate and Chlorophyll a (Chl-a)
The cyanobacteria isolate was transferred into 100 mL of BG-11 medium and adjusted to 10 6 CFU/mL. The culture was incubated under a 12:12 h light/dark cycle (25 µmol photon m −2 /s) at 25 • C for 15 days. Eleven milliliters were collected at three-day intervals.
To measure the cyanobacterial growth, 10 µL of each sample was dropped on a hemocytometer, and cell numbers were counted under a microscope. The specific growth rate (µ) was calculated from where N 1 and N 2 are the cell numbers at times t 1 and t 2 , respectively. To evaluate chl-a content, 10 mL of the culture was centrifuged at 3600× g for 10 min. The pellets were resuspended in 400 µL of acetone (90%, v/v) and kept at 4 • C for 24 h in darkness. After centrifugation at 3600× g for 30 min, the supernatant was measured for optical density (OD) using a UV-vis spectrophotometer (LUMIstar ® Omega, Ortenberg, Germany) at 630, 645, 663, and 750 nm, and 90% acetone solution was used as a blank [100]. The concentration of Chl-a was calculated as Chl-a (µg/mL) = [11.64 × (OD 663 − OD 750 )) − 2.16 × (OD 645 − OD 750 ) + 10 × (OD 630 − OD 750 )] × 25 (2)

Microcystins Production
To measure the MCs, 10 Microcystis isolates were grown in 30 mL of BG-11 medium for 15 days, as described above. The culture was centrifuged at 6000× g for 10 min. The cell pellet and supernatant were collected to measure intracellular MCs (IMCs) and extracellular MCs (EMCs), respectively. The pellet was resuspended in 10 mL 50% methanol-water solution in order to extract IMCs [101]. Then cell suspension in the methanol solution was sonicated for 10 min, followed by incubation in 50 • C water for 20 min [102]. After centrifugation at 10,000× g for 5 min, the supernatant was filtered through a 0.22 µm pore-size polyethersulfone filter (Sartorius, Surrey, UK), and the filtrate was used for IMCs determination. An Abraxis Microcystin Kit (520011) was used to determine IMCs and EMCs concentrations following the manufacturer's instructions. The medium was used as blank. All measurements for MCs were done in duplicate, according to the instructions.

Extracellular Polymeric Substances Determination
Soluble EPS (sEPS) and bound EPS (bEPS) concentrations were determined for protein and carbohydrate content, according to the method described elsewhere [103,104]. Briefly, 30 mL of cyanobacteria was cultured as described above for 15 days. After centrifugation at 11,550× g for 15 min, sEPS and bEPS were determined in the supernatant and pellet, respectively. To obtain bEPS, the pellet was resuspended in 5 mL distilled water and incubated at 45 • C for 4 h. The mixture was centrifuged at 11,550× g for 15 min, and the supernatant containing bEPS was collected. The protein and carbohydrate concentrations of the sEPS and bEPS were quantified by Lowry [105] and phenolsulfuric [106] assays, respectively. The determination was done in triplicate.
Data obtained were analyzed by ANOVA with the post hoc Tukey test using SPSS statistical software version 17.0 for Windows EDU to investigate the significance of EPS content and composition at p ≤ 0.05.

Scum Formation
Scum formation was determined by observing cell migration to the water surface. Scum formation pictures were illustrated elsewhere [13,53]. Cyanobacteria culture was prepared for 7 days, as previously described. After centrifugation at 10,000× g for 5 min, the pellet was resuspended with 5 mL of 8× concentrated BG-11 medium. The OD 580 of the culture was adjusted to 1, followed by incubation under the constant fluorescent light of 20 µmol photons m −2 /s. Scum formation was observed daily for 7 days with the naked eye.

Dynamic of MCs Production and Growth
Microcystis SG03 was cultured in 500 mL BG-11 medium under the above conditions for 30 days. Thirty-one milliliters was collected every 5 days during the cultivation. To measure cyanobacterial growth, cell numbers from 10 µL of each sample were counted under a microscope using a hemocytometer. IMCs and EMCs were determined, as previously described. The experiment was conducted in independent duplicate.

Analyses of Cyanotoxins
Microcystis SG03 was cultured in 300 mL BG-11 medium for 10 days, as described above. After centrifugation at × g for 5 min, the pellet was resuspended in 5 mL 50% methanol-water solution. The IMCs were extracted using the same method, as previously described. Cyanotoxins were analyzed by a liquid chromatograph-quadrupole time-of-flight mass spectrometer (LC-QTOF MS) (G6545A, Agilent Technologies, CA, USA) [107] equipped with a Zorbax Eclipse plus C18 Rapid Resolution column HD (150 mm length × 2.1 mm inner diameter and particle size 1.8 µm). The column was held at 40 • C with gradient elution: high-purity water containing 0.1% formic acid (solvent A) and 11.5% acetonitrile solution (solvent B). The elution steps were: (1) 70% A and 30% B for 2 min; (2) 5% A and 95% B for 14 min; (3) 70% A and 30% B for 14.20 min; and (4) 70% A and 30% B for 20 min at a flow rate of 0.2 mL/min. The UHPLC system was connected with a quadrupole time-of-flight mass spectrometer, equipped with a Dual Agilent Jet Stream Electrospray Ionization (Dual AJS ESI). The positive electrospray ionization parameters were as follows: gas temp, 325 • C; gas flow, 13 L/min; nebulizer, 35 psig; sheath gas temp.,