Degradation of Microcystins Using Immobilized Microorganism Isolated in an Eutrophic Lake

Chemosphere 65 (2006) 117–124 www.elsevier.com/locate/chemosphere

Degradation of microcystins using immobilized microorganism isolated in an eutrophic lake

Kiyomi Tsuji a,*, Miki Asakawa a, Yojiro Anzai b, Tatsuo Sumino c, Ken-ichi Harada d

a Kanagawa Prefectural Institute of Public Health (Formerly, Kanagawa Prefectural Public Health Laboratory), 1-3-1 Shimomachiya, Chigasaki, Kanagawa 253-0087, Japan b Faculty of Pharmaceutical Sciences, Toho University, Miyama, Funabashi, Chiba 274-8510, Japan c Research Division, Hitachi Plant Engineering and Construction Corporation, Kamihongo 537, Matsudo, Chiba 271-0064, Japan d Graduate School of Environmental and Human Sciences and Faculty of Pharmacy, Meijo University, Tempaku, Nagoya 468-8503, Japan

Received 28 September 2005; received in revised form 14 February 2006; accepted 14 February 2006 Available online 24 March 2006

Abstract

The final purpose of our series of studies is to establish a biological removal method of cyanobacteria and their toxic products using immobilized microorganisms that can lyse cyanobacteria and decompose microcystins. To establish the biological removal method in non-point areas and water purification plants, as the first step, we explored bacteria active against the cyanobacterial hepatotoxin microcystin in the present study. Eleven active bacteria were isolated from samples taken from Lakes Tsukui and Sagami, Japan. Among 3 strains (B-9 to B-11) with degradative activity, strain B-9 exhibited the strongest activity. The 16S rDNA sequence of the strain B-9 showed the highest similarity to that of Sphingomonas sp. Y2 (AB084247, 99% similarity). Microcystins-RR and -LR were completely degraded by strain B-9 (SC16) within 1 d, which led to an immobilized microorganism with a polyester resin. The degradation of microcystin-RR in a bioreactor using the immobilized strain B-9 was observed and microcystin-RR (>90%) was completely degraded after 24 h. Microcystin-RR was added to the lake water at regular intervals and the degradation after 24 h was observed in the bioreactor over a 72-d period. An over 80% removal efficiency continued for 2 months, showing that the life of the immobilized B-9 in terms of activity was at least 2 months under the optimized conditions. From these results, this immobilized B-9 is feasible for the practical treatment of microcystins in non-point areas and water purification plants. 2006 Elsevier Ltd. All rights reserved.

Keywords: Microcystin; Biodegradation; Immobilization; Cyanobacteria

1. Introduction tins, the cyclic heptapeptide toxins produced by cyanobacteria, such as Microcystis, show a potent hepatotoxicity Cyanobacteria (blue-green algae) commonly occur in a and tumor-promoting activity by inhibition of the protein variety of water types throughout the world. A variable phosphatases 1 and 2A (Kuiper-Goodman et al., 1999; but high proportion of the cyanobacterial blooms and Sivonen and Jones, 1999). Although animal poisoning scums, which grow annually in lakes, reservoirs, canals and human health problems associated with the ingestion and slow-ﬂowing rivers, contain potent toxins. Microcys- of or contact with cyanobacterial scums have long been recognized, a toxic incident involving the death of 50 peo- ple occurred in Brazil in 1996 due to microcystins in the * Corresponding author. Tel.: +81 467 83 4400; fax: +81 467 83 4457. water used for hemodialysis (Jochimsen et al., 1998; Pouria E-mail address: [email protected] (K. Tsuji). et al., 1998). Because toxic cyanobacteria containing

0045-6535/$ - see front matter 2006 Elsevier Ltd. All rights reserved. doi:10.1016/j.chemosphere.2006.02.018 118 K. Tsuji et al. / Chemosphere 65 (2006) 117–124 microcystins threaten human health and life, we must The MA medium was prepared as follows: a mixture establish an effective method for the regulation of the of bicine, 500 (mg); Ca(NO3)2 Æ4H2O, 50; KNO3, 100; occurrence of cyanobacteria and their toxic metabolites. NaNO3, 50; Na2SO4, 40; MgClÆ6H2O, 50; b-Na2 glycero- Although many methods, such as biomanipulation and phosphate, 100, metal mixture composed of Na2EDTA algicides, have been tried for the elimination of cyanobac- (5 mg), FeCl3 Æ6H2O (0.5 mg), MnCl34H2O (5 mg), ZnCl2 teria in a lake, no suitable method has been developed, and (0.5 mg), CoCl2 Æ6H2O (5 mg), Na2MoO4 Æ2H2O (0.8 mg) it would be difficult to regulate the occurrence of cyanobac- and H3BO3 (20 mg) were dissolved in 1 l of purified water teria by conventional methods (Chorus and Mur, 1999). and the resulting solution was adjusted to pH 8.6. Furthermore, no effective method has been proposed for degrading microcystins in the natural environment. Our 2.3. HPLC laboratory is developing a biological control system using microorganisms co-existing in the same ecosystem to HPLC was carried out under the following conditions: decrease the outbreak of cyanobacteria (Sigee et al., pump, Shimadzu LC-9A (Kyoto, Japan); photodiode array 1999) and the decomposition of microcystins (Jones detector, Shimadzu SPD-M10A, CLASS-LC10 integrating et al., 1994; Bourne et al., 1996). Immobilized microorgan- system; Inertsil ODS-3 column (150 · 4.6 mm ID, GL Sci- isms have been assessed for water purification applications ence Inc, Tokyo, Japan), COSMOSIL 5C18 AR-II column (Sumino et al., 1985, 1991; Kokufuta et al., 1986; Hashi- (150 · 4.6 mm ID, Nacalai Tesque Inc, Kyoto, Japan); moto and Furukawa, 1987). In our system, suitable micro- mobile phase, methanol: 0.05 M NaH2PO4 (pH 3) = organisms are explored from eutrophic lakes, and they are 58:42; flow rate, 1.0 ml min1; detection, UV 238 nm; col- improved for the purposes mentioned above. Additionally, umn temperature, 40 C. these microorganisms are immobilized with appropriate resins and then applied to an eutrophic lake. After this 2.4. Isolation of microorganism with anticyanobacterial operation, the applied immobilized microorganism should activity and degradative activity be recovered from the lakes to avoid secondary pollution due to the microorganism used. Lake waters, sediments and soils were collected from In this study, lake waters, sediments and soils were col- Lakes Sagami and Tsukui in Kanagawa Prefecture, Japan, lected from Lakes Sagami and Tsukui in Kanagawa Prefec- from July 1997 to July 1998. The soil samples were sus- ture, Japan, from July 1997 to July 1998 in order to find pended in sterilized water (10 ml), and the supernatants suitable microorganisms. From these samples, the isolation and lake water were subjected to the soft-agar overlayer of microorganisms, which have a degradative activity method (Uchida et al., 1998). After incubation for several toward microcystins, was tried. To develop an immobiliza- d, the desired microorganisms were obtained from the tion method having a high degradative activity, the effect of resulting plaques and they were then transferred to Sakurai purified bacteria and the immobilization method on the (0.2% peptone, 0.1% yeast extract, 0.05% glucose, 1.5% degradative activity of the microcystin were investigated. agar) or ISP No. 2 medium. To isolate the degradative bac- For optimizing the immobilizing conditions, a continuous terium, the supernatants of the soil samples and lake water bioreactor using immobilized microcystin-degradative bac- were inoculated onto Sakurai medium. Single colonies teria was used for the practical treatment of lake water con- from these plates were transferred to a solution of microcy- taining the microcystins. stin-RR in distilled water (2 mg l1), and the microcystin- RR degradation was monitored by HPLC. During this 2. Materials and methods operation, 23 bacteria, of which 12 bacteria were actinomy- cete, were collected. Among them, three bacteria showed 2.1. Toxins and other reagents microcystin-degradative activity. The OD value at 660 nm was used as index to the viable bacteria count of strain Microcystins-RR and -LR were purchased from Wako B-9. Pure Chemical Industries (Osaka, Japan). Microcystins- RR and -LR were also isolated and purified from the sur- 2.5. Identification of the microorganisms face blooms collected from Lake Suwa in Japan according to the method described by Harada et al. (1988). All One isolated single strain (B-9), which showed the high- reagents used were of analytical grade or HPLC grade. est degradative activity among the 3 strains, was selected for subsequent studies. Strain B-9 was routinely main- 2.2. Cyanobacteria tained on peptone–yeast extract agar. Strain B-9 was identified using morphological observations, chemotaxo- NIES-102 (toxic, Microcystis viridis) was obtained from nomical analyses, and 16S rRNA sequencing. The total the National Institute for Environmental Studies, DNA of strain B-9 was extracted from cells cultured on Tsukuba, Japan. This strain was cultivated in 500 ml Erlen- an agar plate by the benzyl chloride method according to meyer flasks containing 200 ml each of the modified MA Zhu et al. (1993). Amplification of the 16S rRNA coding medium at 25 C for 8 d under continuous illumination. region of the DNA and sequencing of the rDNA using K. Tsuji et al. / Chemosphere 65 (2006) 117–124 119 the direct sequencing method was performed as described 2.8. Degradation of microcystins-LR and -RR in a by Anzai et al. (2000). The 16S rRNA gene sequence data cyanobacterium using immobilized adhesively B-9 of the related strains were obtained from the GenBank/ EMBL/DDBJ databases for comparison. The genetic dis- Thirty-five pieces of polyester (Fabios) were added to 6 1 tances between the sequences were estimated using the Knuc the B-9 culture (8 · 10 cells ml ) and the suspension values (Kimura, 1980). A phylogenetic tree was then con- was shaken at 80 rpm and 27 C for 30 min. After filtra- structed by the neighbor-joining method (Saitou and Nei, tion, the resulting resin was placed in a flask, in which a 1987), and the evaluation of the tree was carried out by cyanobacterium, NIES 102, was cultivated in lake water the bootstrap method using the Clustal W program and a (300 ml) for 6 d. The suspension was incubated at 25 C total of 1000 bootstrapped trees were generated (Felsen- and 120 rpm. An aliquot (10–50 ml) of the sample was stein, 1985; Thompson et al., 1994). Deleted and unknown taken at regular intervals and analyzed according to the positions were eliminated for the comparison of the method described previously (Tsuji et al., 1994) and a spec- sequences. Positions (in Escherichia coli numbering system) trophotometer at 660 nm. The same experiment was car- 70–100, 181–219, 447–487, 1004–1036, 1133–1141, and ried out without B-9 as the control. 1446–1456 were eliminated from the comparison because the secondary structures of these regions differed between 2.9. Entrapped method for immobilization strains. The 16S rRNA gene sequence, which we determined, has been deposited in the DDBJ, and the sequence The entrapping method for obtaining the immobilized is available from GenBank, EMBL, and DDBJ under the B-9 was carried out by entrapping in a polyethylene glycol accession number AB159609. (PEG) gel (Sumino et al., 1992). The same microorganisms were used as described for the adhesive immobilized 2.6. Selection and strain improvement of degradative method. The culture was suspended in PEG prepolymer microorganism solution containing a promoter (N,N,N0,N0-tetramethyl- ethylene-diamine) and mixed with an initiator (potassium The strain was diluted by a physiological salt solution, persulfate) at pH 7. This mixture was immediately poly- and streaked onto Sakurai medium. All cultures were sub- merized into a gel, which was cut into 3 mm cubic pellets. sequently maintained at 27 C. Fifty to eighty single colo- The composition of the immobilizing material was 10% nies from these plates were transferred to slants with (w v1), PEG prepolymer, 0.5% (w v1) promoter, 0.25% Sakurai medium. These colonies were inoculated into the (w v1) initiator, active carbon 3% (w v1), together with Sakurai liquid medium and incubated for 3 d. The culture the B-9 culture. solution was inoculated into a solution of microcystin-RR or -LR in distilled water (2 mg l1) and the microcystin-RR 2.10. Degradation of microcystin using a bioreactor or -LR degradation was monitored by HPLC. Degradation of the microcystins-RR and -LR by strain B-9 was exam- This experiment was performed using a bioreactor con- ined at the initial concentration of 1–20 mg l1. Strain B- sisting of an aeration tank (5 l), and a cylindrical basket 9 was inoculated at a density of 2 · 105–2 · 106 cells ml1. with two blades at the top and bottom (Fig. 1). Three hundred pieces of polyester (Fabios) and 3 l of Sakurai 2.7. Adhesive immobilization medium were added to the bioreactor. The strain B-9 culture was inoculated into a tank of the bioreactor, and it The B-9 strain was inoculated into Sakurai medium in a flask, and the flask was shaken at 200 rpm and 27 C for 3 d. The B-9 cultures (50 ml) were then transferred to another flask (200 ml) and each carrier (1-2 g) was added to the flask. After the flask was shaken at 77 rpm and 27 C for 1 h, the carrier was filtered through a glass filter. The carrier was added to 100 ml of Sakurai medium. After the flask was shaken at 200 rpm and 27 C for 2 d, the immobilized B-9 was obtained by filtering the B-9 culture, which was washed with 20 ml of sterilized water. This resin was added to a 100 ml solution of microcystin-RR in phosphate buffer (pH 7.4) and the solution was incubated at 25 C and 77 rpm. An aliquot sample was taken at regular intervals and analyzed by HPLC and a spectropho- tometer at 660 nm. The following carriers were used; cellulose (BM-AQU-ACEL BT-H, Biomaterial, Fukui, Japan) and polyester (Marimo and Fabios, Unitika, Fig. 1. Bioreactor diagram using immobilized B-9 for removal of Osaka, Japan). microcystin from lake water. 120 K. Tsuji et al. / Chemosphere 65 (2006) 117–124 was cultivated at 27 C and 150 rpm for 3 d. The B-9 cul- stin-RR after 16 h. Because microcystins-RR and -LR were ture (7.9 · 106 cells ml1) was transferred into the tank, completely degraded by strain B-9 (SC16) within 1 d and then 3 l of phosphate buffer (pH 7.4) was poured. (Fig. 2), this strain was selected for immobilization. The The reactor was rotated at 150 rpm and 25 C during the degradative behavior of microcystin-LR by stain B-9 was aeration. The degradation experiment was commenced by observed by HPLC. As shown in Fig. 3, the three degrada- the addition of microcystin-RR (600 lg) to the bioreactor. tion products were designated peaks A, B and C, and these Samples were taken at regular intervals and the filtered peaks were identified as a tetrapeptide, intact Adda samples were analyzed according to the method described previously (Tsuji et al., 1994). After degradation, phosphate buffer was removed from the tank. The toxin in 100

buﬀer was added to the bioreactor as the second trial. n i t s

From the third trial, phosphate buﬀer was replaced by lake y c water and this operation was repeated seven times. At o 75 r c regular intervals microcystin-RR was added to the tank i m

RR and the degradation was observed in this bioreactor f o 50 LR experiment. n o i t RR-control 3. Results LR-control 25 Degrada 3.1. Collection of microorganisms from the lakes % 0 In order to obtain suitable microorganisms to decrease 0 51015202530 the outbreak of cyanobacteria and the degradation of Time (h) microcystins, water, soil and sediment were collected from Fig. 2. Degradation of microcystin-RR and -LR by strain B-9. The initial Lakes Tsukui and Sagami, and the Sagami River. These concentration of microcystin was adjusted to 2 mg l1 of water at 25 C. samples were screened according to the method described Initial cell density of B-9 strain was 6.4 · 105 cells ml1. in the Materials and Method section. Selected microorganisms were classified into two groups, actinomycetes belong- ing to Streptomyces and bacteria. Because it is well known that some Streptomyces produce taste and odor com- pounds (off-flavor), such as geosmine and 2-methylisobor- neol (Falconer et al., 1999), bacteria were used in the subsequent experiments. The obtained 11 bacteria were preliminarily identified as shown in Table 1 and classified into two groups, B-1 to B-8 and B-9 to B-11, according to their activities that can lyse cyanobacteria and can decompose microcystins, respectively. Among the 3 strains with degradative activity, B-9 showed the highest activity, which was further improved by the single colony (SC) screening. A single colony, SC16, was selected from 80 strains, showing the highest degradation rate of microcy-

Table 1 Isolated bacteria from lakes Tsukui and Sagami No. Date Place Sample Genus B-1 July 9, 1997 Lake Sagami Sediment Brevibacillus B-2 July 9, 1997 Lake Sagami Water Bacillus B-3 August 1, 1997 Lake Tsukui Water Bacillus B-4 September 3, 1997 Lake Sagami Water Micrococcus B-5 November 5, 1997 Lake Tsukui Sediment B-6 January 29, 1998 Sagami River Water Bacillus Fig. 3. HPLC analysis of the degradation processes of microcystin-LR by B-7 February 1, 1998 Lake Tsukui Soil Bacillus strain B-9. The initial concentration of microcystin-LR was adjusted to B-8 February 1, 1998 Lake Tsukui Soil 20 mg l1 of Sakurai medium at 27 C. Initial cell density of B-9 strain was 4.7 · 106 cells ml1. Peak A, tetrapeptide; Peak B, Adda; Peak C, B-9 September 3, 1997 Lake Tsukui Water linearized microcystin-LR. HPLC conditions: column, COSMOSIL 5C18 B-10 November 5, 1997 Lake Sagami Water Bacillus AR-II (4.6 · 150 mm); mobile phase, MeOH:0.05 M phosphate buﬀer (pH B-11 November 5, 1997 Lake Sagami Water Sphingomonas 3.0) = 58:42; ﬂow rate; 1.0 ml min1, detection; UV 238 nm. K. Tsuji et al. / Chemosphere 65 (2006) 117–124 121

((2S,3S,8S,9S)-3-amino-9-methoxy-2,6,8-trimethyl-10-phe- 3.2. Identiﬁcation of strain B-9 nyldeca-4(E),6(E)-dienoic acid), and linearized microcystin-LR, respectively, by LC/MS (Harada et al., 2004; Strain B-9 was aerobic, chemo-organotrophic, and Imanishi et al., 2005). Gram-negative. The cell was rod-shaped and produced

Fig. 4. Phylogenetic tree indicating the position of strain B-9 within the radiation of related species of the genus Sphingomonas and other representative species of the a-Proteobacteria. The value on each branch is the estimated confidence limit (expressed as a percentage) for the position of the branch as determined by a bootstrap analysis. E. coli (V00348) is used as the root organism. 122 K. Tsuji et al. / Chemosphere 65 (2006) 117–124 yellow-colored colonies on agar media. The activity of cat- ues was the lowest, indicating low leakage of the microor- alase was positive and ubiquinones were contained in the ganisms from this carrier. In order to increase the number cell. The 16S rRNA sequence of the strain was determined of B-9 cells and to decrease the leakage of B-9, an and compared with the sequences available from the Gen- entrapped method using PEG with active carbon was tried. Bank/EMBL/DDBJ databases to obtain more definitive The result of the immobilized B-9 using the entrapped information on the taxonomic and phylogenetic position method is also shown in Table 2. Although the degradation of strain B-9. The BLAST search available on the World of the microcystin-RR was found to be 100% in the first Wide Web showed that the 16S rDNA sequence of strain trial, the repeated use caused a gradual decrease in the effi- B-9 was most similar to the sequence of Sphingomonas ciency that decreased to 65% in the fourth trial. This was sp. Y2 (AB084247, 99% similarity). The Sphingomonas considered to be caused by the contamination of other sp. Y2 that showed the degradative activity for microcystin microorganisms and/or by the decrease in the concentra- was also isolated from Lake Suwa, Japan, in 1995 (Park tion of B-9 by the leakage of active carbon. These experi- et al., 2001). ments showed that polyester (Fabios) was the most The sequence of strain B-9 was compared with the suitable carrier for the adhesive immobilization of the B- published sequences of related species of the genus 9 strain. Sphingomonas and several representative species of the a- Degradation of microcystins-LR and -RR in a cyano- Proteobacteria. The evolutionary distances were calculated bacterium using adhesively immobilized B-9 was investi- for 44 sequence data including the data of E. coli (V00348) gated. A cyanobacterium, NIES-102, liberated constantly as the root organism. A rooted phylogenetic tree based on a small amount of microcystins-LR and -RR into the sur- the distance matrix data thus obtained is given in Fig. 4. rounding medium during growth. As shown in Fig. 5, most The total number of compared nucleotides was 1152. The of microcystins-LR and -RR were not decomposed using transfer of several species from the genus Sphingomonas only the control carrier soaked in Sakurai medium, to three new genera, Sphingobium, Novosphingobium, and whereas both microcystins were completely decomposed Sphingopyxis, was proposed by Takeuchi et al. (2001). using the polyester (Fabios) immobilized B-9 within 48 h. Strain B-9 was included in the cluster of the former genus These experiments indicated that microcystin-LR and Sphingomonas. However, strain B-9 constituted an inde- -RR liberated from NIES-102 can be definitely degraded pendent cluster with Sphingomonas sp. Y2, and the boot- the using immobilized B-9. strap value for the branch of the cluster was high enough (100%). Park et al. (2001) described that it was more appropriate to classify Sphingomonas sp. Y2 as a new genus and species than to include it as a member of the genus

) 0.20 1 - Sphingomonas. l g m (

n RR i

3.3. Immobilization and bioreactor experiment t s

y RR- control c

o 0.10

r LR c

Three kinds of carriers (1–2 g) for the adhesive immobi- i LR- control lization were soaked in B-9 culture solution (8.4 · M 106 cells ml1) and the carriers were shaken at 200 rpm for 2 d. After filtration, decomposition of microcystin 1 0.00 (2 mg l ) using the resulting immobilized B-9 was 0 24 6 observed in phosphate buffer (pH 7.4). Table 2 shows the Time (d) degradation efficiencies obtained from the decreased level Fig. 5. Degradation of microcystins-LR and -RR in NIES-102 culture of microcystin-RR by HPLC and the OD (660 nm) values with strain B-9 immobilized adhesively with a polyester (Fabios) carrier. of the reaction mixtures. The degradation efficiency for Control indicates degradation of microcystins with a polyester (Fabios) polyester (Fabios) reached 100% after 16 h and the OD val- carrier without bacteria.

Table 2 Degradation of microcystin-RR with immobilizing carriers of strain B-9 Carrier Type % degradation of microcystin-RR OD (660 nm) After 16 h After 24 h After 24 h Adhensive immobilization Cellulose (BM-AQU-ACEL BT-H) Cube 56 73 0.008 Polyester (Marimo) Sphere 90 96 0.012 Polyester (Fabios) Pillar-shaped 100 100 0.003 Entrapped immobilization Polyethylene glycol gel Cube 82 95 0.004 K. Tsuji et al. / Chemosphere 65 (2006) 117–124 123

0.3 was much less that that estimated in the cells of cyanobacteria. In our previous studies, it was found that some co-existing microorganisms can eﬀectively decompose ) 1 - l 0.2 microcystins in comparison with other conditions such as g m

( sunlight, heat and adsorption under natural conditions

R (Harada and Tsuji, 1998). Therefore, we explored a suitable

R 0.1 microorganism with a microcystin degradation activity. In 1994, Jones et al. (1994) isolated first a microcystin- 0 degrading bacterium, MJ-PV from Australian water, which 0 20406080was identified to be Sphingomonas sp. based on the 16 S Time (d) rRNA gene sequence. Later, Bourne et al. (2001) carried out cloning and gene library screening of the Sphingomonas Fig. 6. Degradation of microcystin-RR by strain B-9 immobilized adhesively with a polyester (Fabios) carrier in a bioreactor. #: Addition strain and detected the microcystin-degrading gene cluster, of microcystin-RR (600 lg). The reactor was rotated at 150 rpm during mlr, AB, C and D. Three groups then successively reported the aeration at 25 C. the occurrence of microcystin-degrading bacteria, strain MD-1 from Lake Kasumigaura, Japan (Saito et al., 2003a), strain Y2 from Lake Suwa, Japan (Park et al., The carriers (polyester (Fabios), 300 pieces, about 2001) and strain 7CY from Lake Suwa, Japan (Ishii 100 ml), immobilized B-9 by rotating in B-9 culture et al., 2004). Saito et al. (2003b) detected and sequenced (7.9 · 106 cells ml1) were placed inside a cylindrical basket the microcystin-LR degrading gene, mlr A, from three bac- of a bioreactor and phosphate buffer (3 l) was added to the teria, MJ-PV, MD-1 and Y2. As mentioned in the results tank with aeration. Microcystin-RR (600 lg) was added to section, strain B-9 is genetically quite similar to strain this suspension and the cylindrical basket was rotated at Y2. Therefore, an extensive study comparing strain B-9 150 rpm. The degradation was observed at 25 C and micr- with Sphingomonas sp. Y2 and the genera Sphingomonas, ocystin-RR was almost totally degraded after 16 h. After Sphingobium, Novosphingobium,andSphingopyxis should 24 h, 3 l of solution in the tank was then transferred to be required for a definite taxonomic conclusion. another phosphate buffer (pH 7.4). In order to confirm According to the study by Bourne et al. (1996, 2001) the removal efficiency of the immobilized B-9, microcy- mentioned above, microcystinase (MlrA) catalyzes the ini- stin-RR was added to phosphate buffer (twice) and the lake tial ring opening of MCLR at the Adda-Arg peptide bond water (7 times), filtered through a glass GF/C microfiber to produce the linearized microcystin-LR, which is further filter at regular intervals and the degradation after 24 h degraded to a tetrapeptide by the second enzyme, MlrB. was observed in this bioreactor experiment for 72 d. As The enzyme (MlrC) hydrolyzes the tetrapeptide into smaller shown in Fig. 6, the removal efficiency of the immobilized and constituent amino acids. In a previous study, we iso- B-9 continued almost for 2 months. The initial absorbance lated Adda by the microbial degradation of microcystin- of the medium including immobilized microorganism was LR using B-9 strain and determined its structure by spectral estimated as 0.022 and the reaction medium showed very analysis (Harada et al., 2004). It was found that these deg- faint yellow-color, but the subsequent absorbencies contin- radation products, the linearized microcystin-LR, the tetra- ued to be less than 0.006 and no yellow-color was observed peptide and Adda, are essentially non-toxic, strongly in the reaction medium. Therefore, the degradation was indicating that the microbial degradation using microcy- not caused by free microorganism in the medium, but the stin-degrading bacteria is quite effective for the detoxifi- toxins could be degraded by B-9 immobilized. From these cation of microcystins. Very recently we have shown that results, it was possible to use immobilized B-9 for the prac- B-9 strain degrades not only microcystin related com- tical treatment of microcystin in an eutrophic lake. pounds but also nodularin (Imanishi et al., 2005). Although microcystins are structurally stable as men- 4. Discussion tioned above, it is known that UV, chlorine and titanium oxide are effective for their decomposition under appropri- Microcystins are cyclic heptapeptides containing a ate conditions (Tsuji et al., 1995, 1997; Robertson et al., unique b-amino acid, Adda, and are chemically very stable. 1997). However, no systematic method using these condi- Our preliminary experiment on the stability of toxins tions has been used as a practical method for the regulation showed that the half-life was estimated to be 3 weeks in of microcystins in the natural environment. In the present solution even at pH 1 and 40 C. To degrade completely study, we tried to use microorganisms to decompose micro- them, it is necessary for them to be treated under strong cystins and successfully immobilized such microorganisms. acidic conditions such as 6 M hydrochloric acid and triflu- The removal efficiency of the immobilized B-9 continued oroacetic acid while refluxing. The toxins are also resistant for more than 2 months. This methodology may rely on to enzymatic hydrolysis by the usual enzymes, such as tryp- the usual phenomenon in the freshwater ecosystem. To sin (Harada, 1996). However, the amount of microcystins our knowledge, no regulation method for microcystins detected in lake water was at most a few lgl1, which using immobilized microorganisms has been reported. 124 K. Tsuji et al. / Chemosphere 65 (2006) 117–124

Our established regulation method will be available for Kuiper-Goodman, T., Falconer, I., Fitzgerald, J., 1999. Human health practical regulation of released microcystins after the lysis aspects. In: Chorus, I., Bartram, J. (Eds.), Toxic Cyanobacteria in of toxic cyanobacteria by immobilized microorganisms Water—A Guide to Their Public Health Consequences, Monitoring and Management. E&FN Spon, London, pp. 113–153. with lytic activity against cyanobacteria, which will be Park, H.D., Sasaki, Y., Maruyama, T., Yanagisawa, E., Hiraishi, A., reported elsewhere. Kato, K., 2001. Degradation of the cyanobacterial hepatotoxin microcystin by a new bacterium isolated from a hypertrophic lake. Environ. Toxicol. 16, 337–343. References Pouria, S., de Andrade, A., Cavalcanti, R.L., Barreto, V.T., Ward, C.J., Preiser, W., Poon, G.K., Neild, G.H., Codd, G.A., 1998. Fatal Anzai, Y., Kim, K., Park, J.-Y., Wakabayashi, H., Oyaizu, H., 2000. microcystin intoxication in haemodialysis unit in Caruaru, Brazil. Phylogenetic affiliation of the Pseudomonads based on 16S rRNA Lancet 352, 21–26. sequence. Int. J. Syst. Evol. Microbiol. 50, 1563–1589. Robertson, P.K.J., Lawton, L.A., Munch, B., Rouzade, J., 1997. Bourne, D.G., Jones, G.J., Blakeley, R.L., Jones, A., Negri, A.P., Riddles, Destruction of cyanobacterial toxins by semiconductor photocatalysis. P., 1996. Enzymatic pathway for the bacterial degradation of the Chem. Commun. 4, 393–394. cyanobacterial cyclic peptide toxin microcystin-LR. Appl. Environ. Saito, T., Sugiura, N., Itayama, T., Inamori, Y., Matsuura, M.J., 2003a. Microbiol. 62, 4086–4094. Degradation characteristics of microcystins by isolated bacteria from Bourne, D.G., Riddles, P., Jones, G.J., Smith, W., Blakeley, R.L., 2001. Lake Kasumigaura. Water SRT-Aqua. 52, 13–18. Characterization of a gene cluster involved in bacterial degradation of Saito, T., Okano, K., Park, H.D., Itayama, T., Inamori, Y., Neilan, B.A., the cyanobacterial toxin microcystin-LR. Environ. Toxicol. 16, 523–534. Burns, B.P., Sugiura, N., 2003b. Detection and sequencing of the Chorus, I., Mur, L., 1999. Preventive measure. In: Chorus, I., Bartram, J. microcystin LR-degrading gene, mlr A, from new bacteria isolated (Eds.), Toxic Cyanobacteria in Water—A Guide to Their Public from Japanese lakes. FEMS Microbiol. Lett. 229, 271–276. Health Consequences, Monitoring and Management. E&FN Spon, Saitou, N., Nei, M., 1987. The neighbor-joining method: a new method for London, pp. 235–273. reconstructing phylogenetic trees. Mol. Biol. Evol. 4, 406–425. Falconer, I., Bartram, J., Chorus, I., Kuiper-Goodman, T., Utkilen, H., Sigee, D.C., Glenn, R., Andrews, M.J., Bellinger, E.G., Butler, R.D., Burch, M., Codd, G.A., 1999. Safe levels and practices. In: Chorus, I., Epton, H.A.S., Hendry, R.D., 1999. Biological control of cyanobac- Bartram, J. (Eds.), Toxic Cyanobacteria in Water—A Guide to Their teria: principle and possibilities. Hydrobiologia 395/396, 161–172. Public Health Consequences, Monitoring and Management. E&FN Sivonen, K., Jones, G., 1999. Cyanobacterial toxins. In: Chorus, I., Spon, London, pp. 155–178. Bartram, J. (Eds.), Toxic Cyanobacteria in Water—A Guide to Their Felsenstein, J., 1985. Confidence limits on phylogenies: an approach using Public Health Consequences, Monitoring and Management. E&FN the bootstrap. J. Evol. 39, 783–791. Spon, London, pp. 41–111. Harada, K.-I., 1996. Chemistry and detection of microcystins. In: Sumino, T., Kon, M., Mori, N., Nakajima, K., 1985. Development of Watanabe, M.F., Harada, K.-I., Carmichael, W.W., Fujiki, H. wastewater treatment technique by immobilized microorganisms. J. (Eds.), Toxic Microcystis. CRC Press, Boca Raton, FL, pp. 103–148. Water Waste 27, 1024–1029. Harada, K.-I., Tsuji, K., 1998. Persistence and decomposition of hepa- Sumino, T., Nakamura, H., Mori, N., 1991. Immobilization of activated totoxic microcystins produced by cyanobacteria in natural environ- sludge by the acrylamide method. J. Ferment. Bioeng. 72, 140–142. ment. J. Toxicol. Toxin Rev. 61, 385–403. Sumino, T., Nakamura, H., Mori, N., Kawaguchi, Y.J., 1992. Immobi- Harada, K.-I., Matsuura, K., Suzuki, M., Oka, H., Watanabe, M.F., lization of nitrifying bacteria by polyethylene glycol prepolymer. J. Oishi, S., Dahlem, A.M., Beasley, V.R., Carmichael, W.W., 1988. Ferment. Bioeng. 73, 37–42. Analysis and purification of toxic peptides from cyanobacteria by Takeuchi, M., Hamana, K., Hiraishi, A., 2001. Proposal of the genus reversed-phase high performance liquid chromatography. J. Chroma- Sphingomonas sensu stricto and three new genera, Sphingobium, togr. 448, 275–283. Novosphingobium and Sphingopyxis, on the basis of phylogenetic and Harada, K.-I., Imanishi, S., Kato, H., Mizuno, M., Ito, E., Tsuji, K., chemotaxonomic analyses. Int. J. Syst. Evol. Microbiol. 51, 1405– 2004. Isolation of Adda from microcystin-LR by microbial degrada- 1417. tion. Toxicon 44, 107–109. Thompson, J.D., Higgins, D.G., Gibson, T.J., 1994. CLUSTAL W: Hashimoto, S., Furukawa, K., 1987. Immobilization of activated sludge improving the sensitivity of progressive multiple sequence alignment by PVA-boric acid method. Biotechnol. Bioeng. 30, 52–89. through sequence weighting, position-specific gap penalties and weight Imanishi, S., Kato, H., Mizuno, M., Tsuji, K., Harada, K.-I., 2005. matrix choice. J. Nucl. Acids Res. 22, 4673–4680. Bacterial degradation of microcystins and nodularin. Chem. Res. Tsuji, K., Naito, S., Kondo, F., Watanabe, M.F., Suzuki, S., Nakazawa, Toxicol. 18, 591–598. H., Suzuki, M., Shimada, T., Harada, K.-I., 1994. A clean-up method Ishii, H., Nishijima, M., Abe, T., 2004. Characterization of degradation for analysis of trace amount of microcystin in lake water. Toxicon 32, process of cyanobacterial hepatotoxin by a Gram-negative aerobic 1251–1259. bacterium. Water Res. 38, 2667–2676. Tsuji, K., Watanuki, T., Kondo, F., Watanabe, M.F., Suzuki, S., Jochimsen, E.M., Carmichael, W.W., An, J.S., Cardo, D.M., Cookson, Nakazawa, H., Suzuki, M., Uchida, H., Harada, K.-I., 1995. Stability S.T., Holmes, C.E.M., Antunes, M.B., de Melo Filho, D.A., Lyra, of microcystins from cyanobacteria—II. Effect of UV light on T.M., Barreto, V.S.T., Azevedo, S.M.F.O., Jarvis, W.R., 1998. Liver decomposition and isomerization. Toxicon 33, 1619–1631. failure and death after exposure to microcystins at a hemodialysis in Tsuji, K., Watanuki, T., Kondo, F., Watanabe, M.F., Nakazawa, H., Brazil. New Engl. J. Med. 338, 873–878. Suzuki, M., Uchida, H., Harada, K.-I., 1997. Stability of microcystins Jones, G.J., Bourne, D., Blakeley, R.L., Doelle, H., 1994. Degradation of from cyanobacteria—IV. Effect of chlorination on decomposition. the cyanobacterial hepatotoxin microcystin by aquatic bacteria. Nat. Toxicon 35, 1033–1041. Toxins 2, 228–235. Uchida, H., Kouchiwa, T., Watanabe, K., Kawasaki, A., Hodoki, Y., Kimura, M., 1980. A simple method for estimating evolutionary rates of Otani, I., Yamamoto, Y., Suzuki, M., Harada, K.-I., 1998. A coupled base substitutions through comparative studies of nucleotide assay system for the lysis of cyanobacteria. Jpn. J. Water Treat. Biol. sequences. J. Mol. Evol. 16, 111–120. 34, 67–75. Kokufuta, E., Shimohashi, M., Nakamura, I., 1986. Immobilization of Zhu, H., Qu, F., Zhu, L.H., 1993. Isolation of genomic DNA from plants, Paraciccus denitrificans cells with polyelectrolyte complex and denitri- fungi and bacteria using benzyl chloride. Nucl. Acids Res. 21, 5279– fying activity of the immobilized cells. J. Ferment. Technol. 64, 533–538. 5280.