Microcystin-LR Biodegradation by Bacillus sp.: Reaction Rates and Possible Genes Involved in the Degradation

Harmful cyanobacteria blooms may deteriorate freshwater environments, leading to bad water quality that can adversely affect the health of humans, animals, and aquatic life. Many cyanobacteria can produce toxic metabolites, with Microcystin-LR (MC-LR) being the most commonly detected cyanotoxin in fresh water bodies. In this study, a MC-LR degrading Bacillus sp. strain was isolated from Hulupi Lake (HLPL), Taiwan and tested for its degradability of the cyanotoxin. The results showed that the degradation of Microcystin-LR by the isolated Bacillus sp. was temperature-dependent with an optimum MC-LR removal at 37 ◦C and a first order degradation constant rate for 0.22 day−1. The degradation rate was also found to increase with decreasing MC-LR concentrations and increasing Bacillus sp. concentrations. Biomolecular monitoring of three types of genes (mlrA, CAAX, and GST) involved in the degradation indicated that mlrA, and CAAX genes were present in the indigenous bacteria in HLPL water samples. However, for the isolated Bacillus sp. strain, only CAAX genes were detected. The absence of the mlrA gene in the isolated Bacillus sp. strain shows that the degradation of MC-LR does not necessarily follow the pathways with mlrA, and can also follow the pathways involved with CAAX type II amino-terminal protease.


Introduction
Lakes and reservoirs are the main sources of drinking and recreational water in Taiwan and many other countries. Understanding the quality of these waters is therefore an important issue to safeguard public health. Cyanobacteria are an important group of microorganisms present in lakes and reservoirs, and some produce toxins [1,2] and taste and odor compounds [3,4]. Microcystins (MCs) [cyclo-(D-Ala1-X2-D-MeAsp3-Z4-Adda5-D-Glu6-Mdha7)], a group of heptapeptide hepatotoxins, are produced by several cyanobacteria genus, such as Microcystis, Anabaena, Nostoc, and Planktothrix [5], and they are toxic to animals and humans [6,7]. Two of the amino acids at positions two and four in MCs are changeable, leading to more than 100 congeners of microcystins [8]. Among these, Microcystin-LR (MC-LR) with Leucine and Arginine at positions two and four, respectively, is known to date to be the most toxic [9], the most abundant, and the most studied microcystin congener [10]. The presence of microcystins in the public water supply has been extensively documented [2,[11][12][13][14]. In 2014, the MC-LR concentration in the finished water of Toledo, OH, USA, was found to be higher than the World Health Organization (WHO) drinking water guideline value of 1.0 µg·L −1 [15], which led to a state of emergency and left more than half a million people without potable water for three days [16]. After 55 days of culture, Microcystis cells (~3 × 10 6 cells/mL) were concentrated from 5 L of the cultured solution using centrifugation at 3000× g for 1.5 min with 50 mL tubes (Labcon, Super Clear™, Petaluma, CA, USA) in a centrifuge (Hermle Labortechnik GmbH, Model Z206A, Wehingen, Germany). After centrifugation, the supernatant was removed from the sample tube, with 1 mL of cyanobacteria pellet laden solution remaining. The concentrated solution was put into 50 mL tubes and was frozen at −80 °C for 1 day, and dried using a freeze dryer (FD3-12P, Kingmech, Taiwan) for ~2 days. The dried M. aeruginosa powder was then used for the extraction of MC-LR following the procedures modified from Hu et al. [43]. In brief, 0.4 g of the cyanobacteria powder were put into a 100 mL amber glass vial, 10 mL of water-methanol solution (20:80 by volume) was added, the solution was sonicated for 30 min (110 V-43 KHz, Model D200H, DELTA ® , Sacramento, CA USA), and then centrifuged at 12,000 rpm for 10 min at 4 °C (0.7 KW, Refrigerated Micro Centrifuge -Smart R17, Hanil BioMed Inc., Gwangju, Korea). The supernatant was then put in a 500 mL glass beaker, and incubated at 37 °C (low temperature incubator, Cheng Sang, Taiwan) for two days for drying. The dried extract was diluted with 10 mL of deionized water (produced by Milli-Q ® RiOs™, Darmstadt, Germany), and adjusted to pH 3.0 with 0.1% acetic acid (CH3COOH) (99.8%, Sigma-Aldrich Chemie, Steinheim, Germany). The ensemble was centrifuged again at 12,000 rpm for 10 min at 4 °C. The new supernatant was then filtered with a 0.2 μm pore size cellulose membrane (Sartorius Stedim Biotech Gmbh, Goettingen, Germany), was adjusted to pH 7.0 with a 10% ammonia solution (NH3) (25%, Merck KGaA 64271, Darmstadt, Germany), and was then stored at 4 °C for experimental use. Since this is a crude extract, cyanobacterial metabolites other than MC-LR are expected to be present in the MC-LR laden solution. As stipulated earlier, M. aeruginosa cells were harvested after 55 days as the MC-LR concentration was higher during that period of time ( Figure S1). MC-LR extraction gave a 10 mg•L −1 concentration in the solution, and the cell quota of MC-LR for M. aeruginosa PCC7820 was 0.2 pg•cell −1 (Figure S2), similar to that found in a previous MC-LR monitoring analysis [44].

MC-LR and TOC Concentration Measurements
MC-LR concentration was determined with an Enzyme-Linked Immuno-Sorbent Assay (ELISA) with specificity to microcystins, using a commercial kit (Prod. No. ALX-850-319-KI01, Enzo Life Sciences Inc., Farmingdale, NY, USA). Following the protocol described by the manufacturer [45], the concentration was measured at 450 nm using a Microplate Spectrophotometer Reader (Thermo Scientific™ Multiskan™ GO Microplate Spectrophotometer type 357, Thermo Scientific, Helsinki, Finland), with a microcystin detection limit of 0.1 μg•L −1 . After 55 days of culture, Microcystis cells (~3 × 10 6 cells/mL) were concentrated from 5 L of the cultured solution using centrifugation at 3000× g for 1.5 min with 50 mL tubes (Labcon, Super Clear™, Petaluma, CA, USA) in a centrifuge (Hermle Labortechnik GmbH, Model Z206A, Wehingen, Germany). After centrifugation, the supernatant was removed from the sample tube, with 1 mL of cyanobacteria pellet laden solution remaining. The concentrated solution was put into 50 mL tubes and was frozen at −80 • C for 1 day, and dried using a freeze dryer (FD3-12P, Kingmech, Taiwan) for~2 days. The dried M. aeruginosa powder was then used for the extraction of MC-LR following the procedures modified from Hu et al. [43]. In brief, 0.4 g of the cyanobacteria powder were put into a 100 mL amber glass vial, 10 mL of water-methanol solution (20:80 by volume) was added, the solution was sonicated for 30 min (110 V-43 KHz, Model D200H, DELTA ® , Sacramento, CA USA), and then centrifuged at 12,000 rpm for 10 min at 4 • C (0.7 KW, Refrigerated Micro Centrifuge -Smart R17, Hanil BioMed Inc., Gwangju, Korea). The supernatant was then put in a 500 mL glass beaker, and incubated at 37 • C (low temperature incubator, Cheng Sang, Taiwan) for two days for drying. The dried extract was diluted with 10 mL of deionized water (produced by Milli-Q ® RiOs™, Darmstadt, Germany), and adjusted to pH 3.0 with 0.1% acetic acid (CH 3 COOH) (99.8%, Sigma-Aldrich Chemie, Steinheim, Germany). The ensemble was centrifuged again at 12,000 rpm for 10 min at 4 • C. The new supernatant was then filtered with a 0.2 µm pore size cellulose membrane (Sartorius Stedim Biotech Gmbh, Goettingen, Germany), was adjusted to pH 7.0 with a 10% ammonia solution (NH 3 ) (25%, Merck KGaA 64271, Darmstadt, Germany), and was then stored at 4 • C for experimental use. Since this is a crude extract, cyanobacterial metabolites other than MC-LR are expected to be present in the MC-LR laden solution. As stipulated earlier, M. aeruginosa cells were harvested after 55 days as the MC-LR concentration was higher during that period of time ( Figure S1). MC-LR extraction gave a 10 mg·L −1 concentration in the solution, and the cell quota of MC-LR for M. aeruginosa PCC7820 was 0.2 pg·cell −1 ( Figure S2), similar to that found in a previous MC-LR monitoring analysis [44].

MC-LR and TOC Concentration Measurements
MC-LR concentration was determined with an Enzyme-Linked Immuno-Sorbent Assay (ELISA) with specificity to microcystins, using a commercial kit (Prod. No. ALX-850-319-KI01, Enzo Life Sciences Inc., Farmingdale, NY, USA). Following the protocol described by the manufacturer [45], Total organic carbon (TOC) of HLPL water was analyzed via the NPOC (non-purgeable organic carbon) measurement. 20 mL of each sample was filtered using a 0.45 µm syringe (Whatman ® GD/X syringe filters sterile, Sigma-Aldrich, GmbH, Essen, Germany). The filtrate was transferred into a 40 mL TOC vial (Zodiac Life Sciences, Telangana, India) for analysis. The samples were acidified with high purity HCl (Hydrochloric acid ≥ 37%, Sigma-Aldrich, GmbH, Essen, Germany) to pH 2, and the TOC was immediately measured using a TOC analyzer (TOC-500, Shimadzu, Singapore).

Isolation of Bacterial Strain
Water collected from HLPL was used for bacterial isolation. HLPL water was first filtered through a 0.7 µm glass microfiber filter (Z242519 ALDRICH Whatman ® glass microfiber filters, Germany) to remove large particles. 10 mL of the filtered HLPL water was then placed into a 150 mL crystal grade polystyrene cell culture flask, and 90 mL of mineral salts medium (MSM) was added, the composition of which is listed in the Supplementary Materials (S6) [46], and spiked with crude MC-LR solution to allow for a final concentration of MC-LR = 3 mg·L −1 . The samples were then incubated for 12 days at 30 • C in the dark under shaking conditions (200 rev·min −1 ) in a low-temperature orbital shaker incubator (Model LM-570R, Yihder Technology, Taipei, Taiwan). 2 mL of the incubated solution was then mixed with 38 mL of fresh MSM in a new 150 mL polystyrene flask for another incubation under the same conditions for 12 days. The initial MC-LR concentration in the solution was 5 mg·L −1 .
To monitor the degradation of MC, the remaining MC-LR in the solution was measured with the ELISA test. Serial dilutions were streaked onto LB agar plates (Section S6) and incubated at 30 • C. The colonies were streaked onto new LB agar plates four times for further purification. Finally, the colonies were used for the experiments of biomolecular tests and biodegradation.

MC-LR Biodegradation in HLPL Water
Biodegradation of MC-LR in HLPL water containing the natural bacterial community was conducted for different MC-LR concentrations. The experiments were conducted in 40 L portable water plastic tank (polyethylene resins) reactors, filled with HLPL water (for biodegradation) or with filtered and autoclaved (autoclave TM-320, TOMIN, Hsinchu City, Taiwan) HLPL water (for control). Three experimental runs were conducted; one with the original MC-LR concentration in the lake water at 0.55 µg·L −1 , another one spiked with MC-LR to reach a concentration of 9.2 µg·L −1 , and the last one spiked with MC-LR to reach 9.2 µg·L −1 into the autoclaved HLPL water. In addition, a set of experiments was also conducted, in the same conditions, by spiking M. aeruginosa in original HLPL water to allow for a final M. aeruginosa concentration of 7 × 10 5 cells/mL).
The reactors were all incubated at 25 • C and all runs were conducted in triplicate. Samples were collected from the reactors and analyzed for MC-LR and/or bacteria concentrations, in which the former followed those described in Section 2.3 and the latter was measured with heterotrophic plate counts (HPC) according to the method described by Bartram et al. [47]. In addition, the quantitative polymerase chain reaction (qPCR) method was also used to analyze the abundance of mlrA and/or EUB genes in the samples, following the procedure discussed in Section 2.8.
The degradation rates of MC-LR in the samples were calculated according to a first order reaction (Equation (1)).
where C is the MC-LR concentration, C 0 is the MC-LR concentration at time = 0, k is the reaction rate constant, and t is the time of degradation.

MC-LR Biodegradation by the Isolated Bacteria
The isolated bacteria were tested for their degradation of MC-LR under different temperatures, cell concentrations, and initial concentrations. The batch reactors, same as those for degradation in HLPL water (Section 2.5), was used for the experiments. In conducting the experiments, the isolated bacteria were first prepared. Five colonies of isolated bacteria were picked using a sterile pipette tip, and spiked in 9 mL of sterile LB broth (Section S6) using a 50 mL plastic vial (Labcon, SuperClear™, Petaluma, CA, USA) for 72 h in the dark at 25 • C and under constant shaking for 150 rpm. After reaching exponential growth stage, the bacteria laden samples were centrifuged at 5000× g for 5 min, washed with distilled water, and centrifuged again at 5000× g for 5 min, with the washing and centrifuging processes being repeated three times. The centrifuged bacterial laden pellets were then incubated with MSM medium and MC-LR to reach predetermined bacterial concentrations.
One To test the degradation ability of the isolated bacteria for MC-LR, another set of experiments was performed with two different isolated bacteria concentrations, 8.3 × 10 6 and 8.1 × 10 10 CFU/mL. Each concentration of the isolated bacteria was tested against four concentrations of MC-LR; 0.022, 0.1, 0.35, and 0.51 mg·L −1 , and in triplicate at 25 • C. The reactors for the degradation experiments were incubated at 25 • C for 12 days under 12 h/12 h day/light conditions and the residual MC-LR was measured according to the protocol described previously in Section 2.3. Similar to those for degradation in HLPL water, the samples in the current experiments were also analyzed for the abundance of mlrA and EUB genes, using the qPCR method.

DNA Extraction for Cyanobacteria and Bacteria
A commercial kit (Plant Genomic DNA Extraction Mini Kit, FAVOR PREP™, Pingtung, Taiwan) was used for the extraction of DNA from bacteria and cyanobacteria, according to the manufacturer's protocol. In the extraction of laboratory cultures, 1 mL of cultured M. aeruginosa was first centrifuged at 5000× g for 10 min, and the centrifuged pellets were used for the DNA extraction. For the lake water samples, 1 L of the lake water was filtered with a 47 mm diameter cellulose acetate 0.2 µm-pore-size filter (Sartorius Stedim Biotech GMBH 37070, Goettingen, Germany), and the membrane was used as the matrix for the DNA extraction of the bacteria. The extracted DNA was then stored at −20 • C until utilization. For extraction of the DNA from the agar plate, 10 colonies of isolated bacteria were picked using a sterile pipette tip, inoculated in 90 mL of sterile LB broth, and left to grow at 37 • C for 72 h. 1 mL of the bacterial solution was then centrifuged at 5000× g for 5 min; and the laden bacterial pellets were utilized for the isolated bacterial DNA extraction (Plant Genomic DNA Extraction Mini Kit, FAVOR TREP™), according to the manufacturer's instructions.

Polymerase Chain Reaction (PCR) and Quantitative PCR (q-PCR)
The experiments were conducted according to the Minimum Information for Publication of Quantitative Real-Time PCR Experiments (MIQE) in its guidelines of 2009 [48], for a good quality in both PCR (C1000™ Thermal Cycler, Bio-Rad Laboratories, Foster City, CA, USA), and q-PCR (Smart Cycler ® II, Cepheid, q-PCR device, Foster City, CA, USA) results. The primers and their probes/dyes used in this study are listed in Table 1. Only CAAX type II amino-protease gene amplification primers were developed in this study, with the PCR conditions as follows: an initial denaturation of 95 • C for 120 s, followed by 45 cycles of a typical denaturation of 95 • C for 5 s, an annealing at 55 • C for 20 s, and a typical extension at 72 • C for 20 s. Standard curves for mlrA and EUB genes are represented respectively in Figures S3 and S4.
Plasmid insert vector M13 (Isolated bacteria DNA sequencing) Bacterial DNA cloning was carried out following the method described by Knoche and Kephart [56]. The 16S rRNA sequencing was conducted at National Cheng Kung University Hospital (NCKUH) in Taiwan, and the obtained 16S rRNA sequence was used to specify the isolated bacterium via a neighbor-joining phylogenetic tree. The phylogenetic tree was constructed using the NCBI website [57] and Mega6 program [58].

Statistical Analyses
Statistical analyses were performed using a one-way ANOVA (p < 0.05) and a Normality Test (Shapiro-Wilk) (p ≥ 0.05 normal, and p < 0.05 not normal) (Sigma Plot v.11, Systat Software Inc. (SSI), San Jose, CA, USA) to know if there was a significant difference and/or a normal distribution in the MC-LR biodegradation by the isolated bacterial strain under different initial MC-LR concentrations, bacterial concentrations, and different temperatures, as well as for the biomolecular aspect of the degradation.

Degradation of MC-LR in HLPL Water
For the control sample, Figure 2 shows that no obvious change of MC-LR concentration was found, as the concentrations remained almost constant throughout the experimental period (MC-LR concentrations were read against the ELISA calibration curve Figure S5 of the Supplementary Materials). In the batch experiment for the HLPL water sample, all the MC-LR was degraded after six days of incubation, with a half-life time of 2 days and a k at 0.4 day −1 . HLPL water heterotrophic bacterial count (HPC) at the sampling time (September 2015) was 5 × 10 5 CFU/mL, with high similarity to the HPC of riverine samples in South Africa for 10 5 -10 6 CFU/mL [59]. However, the same HLPL water spiked with MC-LR for a final concentration of 9.2 µg·L −1 shows a removal rate of 70% after 12 days of incubation with a MC-LR half-life time of 7 days and a first-order rate constant k of 0.1 day −1 . There was not a statistically significant difference in MC-LR degradation by HLPL bacterioplankton at p < 0.05. In addition, for 9.2 µg·L −1 , data matches the pattern expected from a population with a normal distribution (p = 0.52).

DNA Cloning, Sequencing, and Phylogenetic Tree Analysis for Bacteria
Bacterial DNA cloning was carried out following the method described by Knoche and Kephart [56]. The 16S rRNA sequencing was conducted at National Cheng Kung University Hospital (NCKUH) in Taiwan, and the obtained 16S rRNA sequence was used to specify the isolated bacterium via a neighbor-joining phylogenetic tree. The phylogenetic tree was constructed using the NCBI website [57] and Mega6 program [58].

Statistical Analyses
Statistical analyses were performed using a one-way ANOVA (p < 0.05) and a Normality Test (Shapiro-Wilk) (p ≥ 0.05 normal, and p < 0.05 not normal) (Sigma Plot v.11, Systat Software Inc. (SSI), San Jose, CA, USA) to know if there was a significant difference and/or a normal distribution in the MC-LR biodegradation by the isolated bacterial strain under different initial MC-LR concentrations, bacterial concentrations, and different temperatures, as well as for the biomolecular aspect of the degradation.

Degradation of MC-LR in HLPL Water
For the control sample, Figure 2 shows that no obvious change of MC-LR concentration was found, as the concentrations remained almost constant throughout the experimental period (MC-LR concentrations were read against the ELISA calibration curve Figure S5 of the Supplementary Materials). In the batch experiment for the HLPL water sample, all the MC-LR was degraded after six days of incubation, with a half-life time of 2 days and a k at 0.4 day −1 . HLPL water heterotrophic bacterial count (HPC) at the sampling time (September 2015) was 5 × 10 5 CFU/mL, with high similarity to the HPC of riverine samples in South Africa for 10 5 -10 6 CFU/mL [59]. However, the same HLPL water spiked with MC-LR for a final concentration of 9.2 μg•L −1 shows a removal rate of 70% after 12 days of incubation with a MC-LR half-life time of 7 days and a first-order rate constant k of 0.1 day −1 . There was not a statistically significant difference in MC-LR degradation by HLPL bacterioplankton at p < 0.05. In addition, for 9.2 μg•L −1 , data matches the pattern expected from a population with a normal distribution (p = 0.52).   The stable cyclic structure of MC-LR is a real challenge for its removal via abiotic methods, which are generally used in conventional drinking water treatment plants [60]. In addition, MC-LR resists boiling and chemical hydrolysis at neutral pH [61]. Compared to the control which was autoclaved to remove all kind of biological organisms, the water from HLPL presented a MC-LR degradation, leading us to conclude that MC-LR was degraded by indigenous bacteria since no other degrader was brought into the batch culture [62]. The biodegradability shows that bacteria exist within HLPL water, and the degradation relevant characteristics of the probable bacteria in the lake water are discussed in a later section. Figure 3 shows the total bacteria concentrations in the experimental system, with almost stable concentrations within twelve days, although the MC-LR was degraded. In addition, the total organic carbon (TOC) concentrations did not change much during the experimental period. Farhadkhani et al. [63] observed that TOC did not significantly impact the microbial community growth during MC-LR degradation, which is similar to our observation in the current study. Mou et al. [38] reported that in a biodegradation study of MC-LR, the degrading bacteria did not use MC-LR as a carbon source and MC-LR might be removed via a xenobiotic method. This may be the reason why the degradation of MC-LR did not lead to an increase in total bacteria concentration in this study. However, there was a statistically significant difference between the mean of the sampled population and the hypothesized population mean (p < 0.01) for EUB gene evolution. In addition, both TOC and EUB gene evolution data matched the pattern expected if the data were drawn from a population with a normal distribution. The stable cyclic structure of MC-LR is a real challenge for its removal via abiotic methods, which are generally used in conventional drinking water treatment plants [60]. In addition, MC-LR resists boiling and chemical hydrolysis at neutral pH [61]. Compared to the control which was autoclaved to remove all kind of biological organisms, the water from HLPL presented a MC-LR degradation, leading us to conclude that MC-LR was degraded by indigenous bacteria since no other degrader was brought into the batch culture [62]. The biodegradability shows that bacteria exist within HLPL water, and the degradation relevant characteristics of the probable bacteria in the lake water are discussed in a later section. Figure 3 shows the total bacteria concentrations in the experimental system, with almost stable concentrations within twelve days, although the MC-LR was degraded. In addition, the total organic carbon (TOC) concentrations did not change much during the experimental period. Farhadkhani et al. [63] observed that TOC did not significantly impact the microbial community growth during MC-LR degradation, which is similar to our observation in the current study. Mou et al. [38] reported that in a biodegradation study of MC-LR, the degrading bacteria did not use MC-LR as a carbon source and MC-LR might be removed via a xenobiotic method. This may be the reason why the degradation of MC-LR did not lead to an increase in total bacteria concentration in this study. However, there was a statistically significant difference between the mean of the sampled population and the hypothesized population mean (p < 0.01) for EUB gene evolution. In addition, both TOC and EUB gene evolution data matched the pattern expected if the data were drawn from a population with a normal distribution.

Isolated MC-LR Degrading Bacteria
Isolation was conducted to obtain a bacterial strain with the capability to degrade MC-LR. A strain was isolated from HLPL, and was able to degrade MC-LR. The neighbor-joining tree (Figure 4) obtained by blasting the 16S rRNA sequence against the sequences in the NCBI software [57] gave a 93% similarity with Bacillus sp. Figure S6 (Supplementary Materials) shows a rod-shaped bacteria, as observed via a scanning electron microscope (SEM), which concords with the general shape of Bacillus sp. [64].
The isolated Bacillus sp. strain was further tested for its degradability of MC-LR under different temperatures, and the degradation kinetics will be discussed in Section 3.3.

Isolated MC-LR Degrading Bacteria
Isolation was conducted to obtain a bacterial strain with the capability to degrade MC-LR. A strain was isolated from HLPL, and was able to degrade MC-LR. The neighbor-joining tree (Figure 4) obtained by blasting the 16S rRNA sequence against the sequences in the NCBI software [57] gave a 93% similarity with Bacillus sp. Figure S6 (Supplementary Materials) shows a rod-shaped bacteria, as observed via a scanning electron microscope (SEM), which concords with the general shape of Bacillus sp. [64].
The isolated Bacillus sp. strain was further tested for its degradability of MC-LR under different temperatures, and the degradation kinetics will be discussed in Section 3.3.

Effect of Temperature on MC-LR Degradation
As the bacterial degradation rate depends on the incubation temperature [30], the degradation of MC-LR (for an initial concentration of 0.22 mg•L −1 ) was conducted for four different temperatures (15 °C, 25 °C, 30 °C, and 37 °C) and under the same bacterial concentration (8.4 × 10 6 CFU/mL). It was observed that the degradation kinetics follow a first order reaction and differ at the studied temperatures. MC-LR was degraded rapidly at 37 °C (rate constant k = 0.22 day −1 ) with 74% of MC-LR being removed within 12 days. The rate constant at 37 °C is 27.5 times higher than that at 15 °C (k = 0.008 day −1 ) ( Figure 5). Based on the experimental results, the degradation rates of MC-LR were 0.06 day −1 and 0.04 day −1 respectively at 30 °C and 25 °C. Sumaiya et al. [26] reported that at 18 °C, no MC-LR was degraded with Bacillus cereus (12 GK strain), however a pronounced MC-LR degradation was observed at 32 °C with complete MC-LR removal after six days of incubation. Somdee et al. [30] reported a slow MC-LR degradation with 20% of MC-LR (initial concentration of 25 μg•L −1 ) removal after one day of incubation at 30 °C in a system spiked with 7.9 × 10 6 CFU/mL of NV-3 bacterial concentration. They also noticed that the degradation rates increased with increasing incubation temperature from 15 °C to 30 °C. It is clear that the temperature plays an important role in MC-LR degradation by bacteria, and the MC-LR degradation rate strongly depends on the temperature [26,65]. The statistical analyses showed that the degradation of MC-LR was quite dependent on the temperature, and the degradation rate constants were significant for p = 0.027 for 37 °C, 30 °C, and 15 °C with a normal distribution, except at 25 °C (with p = 0.02).  Figure 4. Neighbor-joining tree of the isolated bacteria. The unknown strain is the isolated bacteria with 93% similarity to Bacillus sp.

Effect of Temperature on MC-LR Degradation
As the bacterial degradation rate depends on the incubation temperature [30], the degradation of MC-LR (for an initial concentration of 0.22 mg·L −1 ) was conducted for four different temperatures (15 • C, 25 • C, 30 • C, and 37 • C) and under the same bacterial concentration (8.4 × 10 6 CFU/mL). It was observed that the degradation kinetics follow a first order reaction and differ at the studied temperatures. MC-LR was degraded rapidly at 37 • C (rate constant k = 0.22 day −1 ) with 74% of MC-LR being removed within 12 days. The rate constant at 37 • C is 27.5 times higher than that at 15 • C (k = 0.008 day −1 ) ( Figure 5). Based on the experimental results, the degradation rates of MC-LR were 0.06 day −1 and 0.04 day −1 respectively at 30 • C and 25 • C. Sumaiya et al. [26] reported that at 18 • C, no MC-LR was degraded with Bacillus cereus (12 GK strain), however a pronounced MC-LR degradation was observed at 32 • C with complete MC-LR removal after six days of incubation. Somdee et al. [30] reported a slow MC-LR degradation with 20% of MC-LR (initial concentration of 25 µg·L −1 ) removal after one day of incubation at 30 • C in a system spiked with 7.9 × 10 6 CFU/mL of NV-3 bacterial concentration. They also noticed that the degradation rates increased with increasing incubation temperature from 15 • C to 30 • C. It is clear that the temperature plays an important role in MC-LR degradation by bacteria, and the MC-LR degradation rate strongly depends on the temperature [26,65]. The statistical analyses showed that the degradation of MC-LR was quite dependent on the temperature, and the degradation rate constants were significant for p = 0.027 for 37 • C, 30 • C, and 15 • C with a normal distribution, except at 25 • C (with p = 0.02).

Effect of Temperature on MC-LR Degradation
As the bacterial degradation rate depends on the incubation temperature [30], the degradation of MC-LR (for an initial concentration of 0.22 mg•L −1 ) was conducted for four different temperatures (15 °C, 25 °C, 30 °C, and 37 °C) and under the same bacterial concentration (8.4 × 10 6 CFU/mL). It was observed that the degradation kinetics follow a first order reaction and differ at the studied temperatures. MC-LR was degraded rapidly at 37 °C (rate constant k = 0.22 day −1 ) with 74% of MC-LR being removed within 12 days. The rate constant at 37 °C is 27.5 times higher than that at 15 °C (k = 0.008 day −1 ) ( Figure 5). Based on the experimental results, the degradation rates of MC-LR were 0.06 day −1 and 0.04 day −1 respectively at 30 °C and 25 °C. Sumaiya et al. [26] reported that at 18 °C, no MC-LR was degraded with Bacillus cereus (12 GK strain), however a pronounced MC-LR degradation was observed at 32 °C with complete MC-LR removal after six days of incubation. Somdee et al. [30] reported a slow MC-LR degradation with 20% of MC-LR (initial concentration of 25 μg•L −1 ) removal after one day of incubation at 30 °C in a system spiked with 7.9 × 10 6 CFU/mL of NV-3 bacterial concentration. They also noticed that the degradation rates increased with increasing incubation temperature from 15 °C to 30 °C. It is clear that the temperature plays an important role in MC-LR degradation by bacteria, and the MC-LR degradation rate strongly depends on the temperature [26,65]. The statistical analyses showed that the degradation of MC-LR was quite dependent on the temperature, and the degradation rate constants were significant for p = 0.027 for 37 °C, 30 °C, and 15 °C with a normal distribution, except at 25 °C (with p = 0.02).   Figure 6a shows the degradation rates of MC-LR by 8.3 × 10 6 CFU/mL of Bacillus sp. under different initial MC-LR concentrations at 25 • C. The degradation rates were 0.065 day −1 , 0.026 day −1 , 0.016 day −1 , and 0.022 day −1 for the cases of MC-LR concentrations of 0.022 mg·L −1 , 0.1 mg·L −1 , 0.35 mg·L −1 , and 0.51 mg·L −1 , respectively. As stipulated earlier in Section 3.3, MC-LR degradation is temperature-dependent. The MC-LR concentration did not significantly impact the kinetics of the toxin's removal at 25 • C under 8.3 × 10 6 CFU/mL of Bacillus sp. For a 25 times higher MC-LR concentration level (0.022 mg·L −1 and 0.51 mg·L −1 ), the degradation rate under 8.3 × 10 6 CFU/mL of Bacillus sp. at 25 • C was observed to be only 5.41 times different (respectively k = 0.065 day −1 and k = 0.012 day −1 ). Yang et al. [39] reported that MC-LR degradation rates were pH-, temperature-, and initial MC-LR concentration-dependent for Stenotrophomonas acidaminiphila MC-LTH2, with the highest degradation rate of 3.0 mg·L −1 ·day −1 standing for a rate constant k = 0.14 day −1 at 30 • C under an initial MC-LR concentration of 21.2 mg·L −1 using a 48 h bacterial culture, although their experiments were conducted separately in order to know the importance of the environmental parameters with regards to the degradation kinetics.   [39] reported that MC-LR degradation rates were pH-, temperature-, and initial MC-LR concentrationdependent for Stenotrophomonas acidaminiphila MC-LTH2, with the highest degradation rate of 3.0 mg•L −1 •day −1 standing for a rate constant k = 0.14 day −1 at 30 °C under an initial MC-LR concentration of 21.2 mg•L −1 using a 48 h bacterial culture, although their experiments were conducted separately in order to know the importance of the environmental parameters with regards to the degradation kinetics.  Another batch experiment was conducted for MC-LR degradability for higher Bacillus sp. concentrations and under 25 • C. Figure 6b shows the degradation rate constants of MC-LR by 8.1 × 10 10 CFU/mL of Bacillus sp. under different initial MC-LR concentrations. We observed that for 8.1 × 10 10 CFU/mL Bacillus sp. incubated with 0.022 mg·L −1 , 0.1 mg·L −1 , 0.35 mg·L −1 , and 0.51 mg·L −1 of MC-LR, the degradation rate constants k were respectively 0.71 day −1 , 1.01 day −1 , 0.035 day −1 , and 0.04 day −1 (Figure 6b). Therefore, at higher bacterial concentration the MC-LR degradation kinetics were proportionally higher, although low degradation rate constants were observed for 0.35 mg·L −1 and 0.51 mg·L −1 of MC-LR. The degradation rate constant under 8.1 × 10 10 CFU/mL of Bacillus sp. at 25 • C for a MC-LR concentration ratio of 25 times (0.022 mg·L −1 and 0.51 mg·L −1 ) was 18 times different (respectively k = 0.71 day −1 and k = 0.04 day −1 ). In the same experimental conditions, Somdee et al. [30], with an incubation temperature of 30 • C (known to be the optimal temperature for the isolated NV-3 bacterial strain), observed that with a higher NV-3 bacterial concentration of 1.0 × 10 8 CFU/mL, only 20% of the initial MC-LR was remaining after one day of incubation. In addition, they reported that at low concentrations of 1 µg·L −1 and 10 µg·L −1 , the MC-LR was completely degraded within one day of incubation, while at a higher concentration of 50 µg·L −1 , a period of 6 days was necessary to reach a complete MC-LR removal with a degradation rate constant k = 0.17 day −1 .

Effect of MC-LR and Bacteria Concentration on the Degradation Rates
Under

Biomolecular Aspect of MC-LR Degradation in HLPL Water
The mlrA gene is well known to be directly involved in the degradation of MC-LR [32] by cleaving the ring structure of the heptapeptide between (all-S, all-E)-3-Amino-9-methoxy-2,6,8-trimethyl-10-phenyldeca-4,6-dienoic acid (Adda) and Arginine amino acid, leading to a significant decrease of toxicity for MC-LR [27]. A decrease in mlrA gene concentration was observed during the batch experiments for MC-LR degradation, from 5.4 × 10 5 copies/L at day 0 to 1.6 × 10 3 copies/L at day 12 ( Figure 7). Li et al. [66] monitored mlrA gene concentrations during MC-LR degradation in lake water samples spiked with 1000 mg·L −1 glucose for the enrichment of degradation bacteria for MCs. They observed a decrease in mlrA gene concentration during the degradation experiments, from 3.2 × 10 6 copies/L initially to 3.9 × 10 5 copies/L at day 7, which is similar to our observation of a decrease in mlrA gene concentrations in the degradation experiments ( Figure 7). The decrease in mlrA gene concentrations observed by Li et al. [66] and in the current study might be due to the presence of bacteria with other degradation genes involved in the MCs' degradation process [33,38]. In addition, there was no statistically significant difference (p = 0.290) in mlrA gene evolution with or without the cyanobacteria enrichment, with a normal distribution for p = 0.49 and p = 0.21, respectively with and without the enrichment in M. aeruginosa.
Although the isolated strain Bacillus sp. has the capability to degrade MC-LR, in the current study the PCR and q-PCR runs show that no mlrA gene was detected in the samples for the bacteria in the degradation experiments. The same observations were reported in Manage et al. [33], where the bacterial strains Arthrobacter spp., Brevibacterium sp., and Rhodococcus sp. were shown to have the ability to degrade MC-LR, although no mlr genes were detected. In another study, Stenotrophomonas acidaminiphila MC-LTH2 was reported to be able to degrade MC-LR with a degradation rate of 3.0 mg·L −1 ·day −1 at 30 • C, and no mlrA gene was detected in the bacterial strain [39].
HLPL water, and MC-LR degradation probably does not follow a GST xenobiotic pathway for its degradation. However, the assertion should be further confirmed since the primers utilized were more specific to Sphingomonas sp. [51]. The absence of the entire genomic sequence of the mlrA gene from a MC-LR degrading bacterial strain raises the issue of whether the microcystin's degradation is encoded within a mobile gene [37,40]. A mlrA active site was thus proposed, with a probability of it being a zinc-binding motif (HEXXH) found in metalloproteases [37,72]. CAAX type II amino-terminal protease belongs to the CPBP (CAAX Proteases and Bacteriocin Processing enzymes) enzymes family, and might encode for a microcystinase function to that endoprotease [36,40,73]. CAAX type II amino-terminal protease enzymes are 27%-35% similar to mlrA [35,36,57].
The CAAX type II amino-terminal protease gene in the isolated bacterial strain was monitored via PCR reactions. The experimental results show that the isolated Bacillus sp. was found to contain only the CAAX type II amino-terminal protease gene with a 230 bp DNA band shown in Figure 8 of an electrophoresis gel.  To date, members of the mlr gene cluster, and particularly mlrA, are known to be highly involved in the MC-LR degradation process in microorganisms [27,32,67]. It has been proposed that the mlrA protein might be periplasmic [32] or found within the plasma membrane [24], however the emplacement of the mlrA gene is still unclear. Beside the mlr gene cluster, MC-LR degradation in the environment by bacteria has been reported to follow two other pathways, the Gluthathione-S-Tranferase (GST) found in many bacteria [68][69][70], and the CAAX type II amino-terminal protease found in Bacillus spp. [35,37,40]. Mou et al. [38] reported a xenobiotic metabolism in the MC-LR degradation pathway under a bacterioplankton microcosm with an important MC-LR removal of up to 75% within 24 h at room temperature. They observed an overrepresentation of GST and cytochrome P450 oxidase, which are proposed to catalyze the synthetic metabolism of MC-LR to cysteine (Cys) and glutathione (GSH) conjugates in animals [71]. The GST gene was not detected in HLPL water, and MC-LR degradation probably does not follow a GST xenobiotic pathway for its degradation. However, the assertion should be further confirmed since the primers utilized were more specific to Sphingomonas sp. [51]. The absence of the entire genomic sequence of the mlrA gene from a MC-LR degrading bacterial strain raises the issue of whether the microcystin's degradation is encoded within a mobile gene [37,40]. A mlrA active site was thus proposed, with a probability of it being a zinc-binding motif (HEXXH) found in metalloproteases [37,72].
The CAAX type II amino-terminal protease gene in the isolated bacterial strain was monitored via PCR reactions. The experimental results show that the isolated Bacillus sp. was found to contain only the CAAX type II amino-terminal protease gene with a 230 bp DNA band shown in Figure 8 of an electrophoresis gel.
The CAAX type II amino-terminal protease gene in the isolated bacterial strain was monitored via PCR reactions. The experimental results show that the isolated Bacillus sp. was found to contain only the CAAX type II amino-terminal protease gene with a 230 bp DNA band shown in Figure 8 of an electrophoresis gel.   Table 2 summarizes the results of the microcystin degradation genes and the degradation rates of MC-LR observed in this study and reported in the literature. MC-LR degradation in HLPL water shows a degradation rate constant k of 0.4 day −1 at 25 • C for an initial MC-LR concentration of 0.55 µg·L −1 (Figure 2). MlrA gene was quantified via q-PCR for an initial concentration of 5.4 × 10 5 copies/L (Figure 7) measured at the time the water was sampled from HLPL. Sphingomonas sp. Y2 (AB084247) [65] and Sphingomonas isolate NV-3 [30] are also reported to have the mlrA gene with MC-LR degradation rate constants of 0.27 day −1 and 0.33 day −1 , respectively, both at the incubation temperature of 30 • C. Although no mlrA gene was detected in the degrading bacteria Stenotrophomonas acidaminiphila MC-LTH2 [39] and a mixed culture of Arthrobacter spp., Brevibacterium sp., and Rhodococcus sp. [33], MC-LR degradation was clearly observed, with rate constants of 0.14 day −1 and 0.33 day −1 , respectively. The results suggest that MC-LR may be degraded without the assistance of the mlrA gene. As shown in Table 1 for the microcosm study of Lake Erie water, MC-LR was observed to be degraded. At the same time, the GST gene was the only one among the three microcystin degradation genes to be detected. The presence of the GST gene in MC-LR laden water was also observed in the HLPL water sample, as reported in Table 1, further suggesting that the GST gene might be associated with MC-LR degradation. In the current study, besides mlrA and GST, another gene, CAAX type II amino-terminal protease, was also found in both HLPL water samples and the isolated Bacillus sp., with MC-LR degradation rate constants of 0.1 day −1 (Figure 2) and 0.22 day −1 (Figure 5), respectively. This is the first time that the CAAX type II amino-terminal protease gene was detected in a water system with MC-LR degradation, and provides evidence of the CAAX type II amino-terminal protease acting like a microcystinase, as proposed in previous studies [35,37]. More studies should thus be performed to deepen the overall knowledge on the genes involved in MC-LR degradation by bacteria in the natural environment. Notes: +: Detected (PCR); -: Not detected (PCR); #: Quantified (q-PCR); ∞: Not studied; *: Calculated from data reported in the reference.

Conclusions
In this study, we isolated a MC-LR degrading Bacillus sp. from HLPL water. The bacterium was demonstrated to have good capabilities in biodegrading MC-LR by up to 74% of its original concentration (0.22 mg·L −1 ) within only twelve days for a low bacterial concentration (8.3 × 10 6 CFU/mL). The bacterium was tested for its biodegradability of MC-LR under different temperatures, cell densities, and initial MC-LR concentrations. The results show that the degradation rates increased with increasing temperature and cell densities but were not influenced by initial MC-LR concentrations. Neither mlrA nor GST genes were detected in the isolated bacterial strain. However, the isolated Bacillus sp. presented an mlrA gene homologue, the CAAX type II amino-terminal protease enzyme, which is less than 30% similar to the mlrA gene. To date, this is the first time that the CAAX type II amino-terminal protease gene has been monitored and qualitatively determined in a MC-LR degrading bacterial strain. However, the CAAX type II amino-terminal protease's role is still not well known as a microcystinase. In the MC-LR degradation system, mlrA, GST, and CAAX type II amino-terminal protease genes may be detected, and thus have the potential to be used as indicators in MC-LR degradation pathways. Further studies should be conducted to elucidate the implications of these genes in MC-LR degradation in real water environments.
Supplementary Materials: The following are available online at www.mdpi.com/2073-4441/8/11/508/s1, Figure S1: Growth curve of M. aeruginosa and the change of MC-LR concentration, Figure S2: Correlation between MC-LR (mg·L −1 ) and M. aeruginosa PCC7820 cells/mL, Figure S3: MlrA gene standard curve, Figure S4: EUB gene standard curve for total bacteria, Figure S5: MC-LR measurement standard curve via ELISA, Figure S6: SEM images of the isolated bacteria Bacillus sp., and Figure S7: Bacillus sp. growth inoculated with and without MC-LR.