99

［Japanese Journal of Water Treatment Biology Vol.43 No.2 99 111 2007］

Degradation of the Cyanobacterial Hepatotoxin

Microcystin by Bacteria Isolated
from a Monoxenic Culture
of the Flagellate Monas guttula
NAOSHI FUJIMOTO1, NAOKI OHNO1, KUNIHIRO TANAKA1, ITARU NARAHARA1,
AKIHIRO OHNISHI1, MASAHARU SUZUKI1, NORIO IWAMI2,
MOTOYUKI MIZUOCHI3, and YUHEI INAMORI3
1
Faculty of Applied Bioscience, Tokyo University of Agriculture
/1 1 1 Sakuragaoka, Setagaya-ku, Tokyo 156 8502, Japan
2
Department of Environmental Systems, Meisei University
/2 1 1 Hodokubo, Hino, Tokyo 191 8506, Japan
3
Independent Administrative Institution National Institute for Environmental Studies
/16 2 Onogawa, Tsukuba-shi, 305 8506, Japan

Abstract
Microcystins are a group of cyclic heptapeptide hepatotoxins produced by the
cyanobacteria. Two microcystin-degrading strains of bacteria (MG and MG ) were
isolated from a monoxenic culture of Monas guttula preying on Microcystis viridis. Both
strains were related to Sphingopyxis and Novosphingobium groups based on phylogenetic
analysis of S rDNA. Degradation of microcystin by these strains was tested under the
following conditions: temperature from to ℃, pH from . to . , three different
microcystin analogs and initial microcystin concentrations from to μg･l– . The rate
of microcystin RR degradation by strain MG significantly increased at ℃, whereas
the effect of temperature on the degradation rate by strain MG was small. Both strains
could degrade microcystin under alkaline pH. Both strains could degrade microcystin RR,
YR and LR. The degradation rate of microcystin RR by the two strains was faster than
the degradation rates of microcystin YR and LR. Both strains could degrade microcystin
LR at all initial concentrations.

Key words: Microcystis, Monas guttula, predation, microcystin-degrading bacteria,

16S rDNA

center in Caruaru, Brazil that resulted from

INTRODUCTION
the use of microcystin-contaminated water
Microcystins are a group of cyclic for dialysis treatment3). The microcystins
heptapeptide hepatotoxins produced by the have also been implicated as relating to
cyanobacteria Microcystis, Anabaena, cancer through animal experiments and a
Planktothrix and Nostoc1 – 2). The mass retrospective cohort study4 – 5). In 1998, the
occurrence of toxic cyanobacteria in eutrophic World Health Organization recommended a
lakes and reservoirs worldwide causes the provisional guideline value of 1 μg l–1 for
death of animals and poses a health hazard microcystin LR6). Globally, many people
for humans. Many patients died from an might be exposed to toxins in areas where
outbreak of acute liver failure at a dialysis there is no effective system for treating water
100 Japanese J. Wat. Treat. Biol. Vol.43 No.2

contaminated with cyanobacterial toxins. To honeycomb tube medium in a water treatment

avoid exposure to these toxins, technologies facility16). Xie et al. demonstrated that a bio-
for removing cyanobacteria and their toxins ceramic filter and moving-bed biofilm reactor
from water are necessary. could be used to effectively remove
Several bacterial strains with microcystin- microcystins from Yellow River water17).
degrading activity have been isolated. In a biological process for treating lake
Bacterial degradation of microcystin LR was and reservoir water, it is believed that
confirmed in surface water samples, and the removal of microcystin-producing cyano-
microcystin-degrading strain Sphingomonas bacteria may depend on several processes
MJ PV was subsequently isolated7 – 8). The such as biofilm adsorption, predation by
microcystin-degrading bacterium MD 1 strain protozoa and metazoa and sedimentation of
was isolated from Lake Kasumigaura, Japan sloughed biofilm. Microcystins might be
and was proved to be related to Sphingomonas released by lysis of cyanobacteria or excreted
on the basis of 16S rDNA sequence analysis9). intact by protozoa and metazoa, followed by
The strain MD 1 could degrade microcystin degradation of dissolved microcystin by
LR, YR and RR. A bacterium isolated from a microcystin-degrading bacteria in biofilms.
hypertrophic lake (strain Y2) was shown to However, the mechanism of microcystin
be phylogenetically distinct from any removal in a microbial ecosystem consisting
established species of Sphingomonas based of cyanobacteria, their predators and bacteria
on 16S rDNA sequence analysis10). The Y2 has yet to be determined.
strain was characterized using chemo- The flagellate Monas guttula, existing in
taxonomic and phenotypic analysis and was biofilms at lake water treatment plants, was
proposed to be a new genus and species, isolated from the water of Lake Kasumigaura,
Sphingosinicella microcystinivorans11). Japan. M. guttula has been cultured in a
The enzymatic pathway for bacterial suspension of Microcystis viridis because M.
degradation of microcystin LR has also been guttula preys on M. viridis. The M. guttula
determined8). Bourne et al. cloned and culture also contains co-existing bacteria as
screened the gene library of Sphingomonas they are not removed during the isolation
MJ PV and detected the microcystin process. A simultaneous decrease in the
degrading gene clusters, mlrA, mlrB, mlrC concentration of microcystins has been
and mlrD12). The enzyme encoded by mlrA observed when M. guttula preys on
can cleave the ADDA Arg peptide bond in Microcystis18). Moreover, when purified micro-
microcystin LR and open the cyclic structure. cystin is added to the suspension in a
Linearized microcystin LR is degraded to predation experiment, the microcystin con-
individual amino acides by the peptidases centration decreases prior to the growth of
encoded by mlrB and mlrC. The gene mlrD M. guttula, suggesting that the co-existing
encodes the transporter protein that allows bacteria contribute to the degradation of
microcystin uptake into the cell. Saito et al. microcystin. In the present study, microcystin-
discovered the existence of a gene homologous degrading bacteria were isolated from a
to mlrA in the microcystin-degrading strains monoxenic culture of M. guttula, and the
MD 1 and Y213). characteristics of microcystin degradation
In the biological treatment process for were examined to clarify the mechanism of
drinking water, some studies demonstrate microcystin removal and the role of
degradation of microcystin. Slow sand microcystin-degrading bacteria in cultures of
filtration has been shown to be an effective M. guttula.
treatment for eliminating microcystins from
MATERIALS AND METHODS
drinking water14). Ho et al. reported that
biological sand filters could degrade Cyanobacteria and protozoa used An
microcystin LR, and a microcystin-degrading isolate of M. viridis from Lake Kasumigaura
gene was detected in the biofilm15). was obtained from the National Institute for
Degradation of microcystin has been Environmental Studies, the Independent
demonstrated using a biofilm attached to the Administrative Institution (culture collection
 Microcystin-Degrading Bacteria Isolated from Culture of Protozoa 101

No. NIES 102). This strain is axenic and of the filtrate containing the co-existing
known to produce three microcystins, bacteria were then inoculated into 8 ml
microcystin RR, YR and LR18). liquid medium9) (5 mg Ca(NO3)2･4H2O, 10 mg
M. viridis, prey of M. guttula, was KNO3, 5 mg NaNO3, 5 mg Na2SO4, 5 mg
cultivated in M 11 medium consisting of 100 MgCl2･6H2O, 0.5 mg Na2EDTA･2H2O, 0.05
mg of NaNO3, 75 mg of MgSO4･7H2O, 40 mg mg FeCl3･6H2O, 0.5 mg MnCl2･4H2O, 0.05
of CaCl2 2H2O, 20 mg of Na2CO3, 1 mg of mg ZnCl2, 0.5 mg CoCl2･6H2O, 0.5 mg
FeSO4･7H2O and 1 mg of Na2･EDTA･2H2O Na2MoO4･2H2O and 2mg H3BO3 in 1 liter of
in 1 liter of Milli Q water, for two weeks at Milli Q water) in a test tube. Subsequently,
30℃. The pH was adjusted to 8. Light was 1 ml of 10 mg･l–1 microcystin LR (Wako Pure
supplied continuously with fluorescent bulbs Chemical Industries, Ltd.) was added to the
at a photon fluence rate of 0.39×1016 quanta･ test tube. The bacteria in the liquid medium
cm–2･s–1 (2,000 lx). that contained microcystin were cultured at
Protozoa used in this study were isolated 30℃ with shaking (130 rpm, Monosin, Taitec).
from surface water of Lake Kasumigaura in The concentration of microcystin LR was
autumn. Droplets of lake water and sterilized determined by HPLC. When the microcystin
water were placed on a glass slide to isolate concentration decreased to less than 10% of
protozoa under the microscope. An individual the initial concentration, the culture was
protozoan cell in lake water on a glass slide diluted and streaked onto a solidified GYP
was drawn into a capillary pipette and medium plate, which consisted of 5 g of
transferred to a droplet of sterilized water. peptone, 2.5 g of yeast extract, 1 g of glucose
The individual protozoan cell in a droplet of and 15 g of agar in 1 liter of Milli Q water,
sterilized water was again drawn into a in order to isolate single colonies. To assess
capillary pipette and transferred to another the ability of the isolated bacterial strains to
droplet of sterilized water. This washing degrade microcystin, each colony was then
process was repeated a few times, and inoculated in liquid medium containing 1
ultimately the individual protozoan cell was mg･l–1 of microcystin LR, cultured at 30℃
transferred to a suspension of M. viridis in a and the microcystin LR concentration was
test tube. This culture was incubated at 30℃ monitored.
in total darkness. A clone culture in which
the protozoa grew from an individual cell Phylogenetic analysis PCR was per-
was subcultured periodically by transferring formed directly on bacterial colonies (colony
a small amount of culture at the stationary PCR). PCR amplification of the 16S rDNA
phase to a new suspension of M. viridis. gene was conducted using primers 20F (5’
Characteristics of the isolated protozoan were AGTTTGATCATGGCTCA 3’, positions 10 26)
consistent with those of M. guttula: spherical and 1540R (5’ AAGGAGGTGATCCAACGCA
to ovoid, 14–16 μm long, free-swimming or 3’, positions 1541 1522) (Escherichia coli
attached, longer flagellum about one to two numbering system20)), designed to generate
times the body length19). Thus, the isolated an amplification product of approximately
protozoan was identified as M. guttula. The 1500 base pairs21). For colony PCR, isolated
clone culture of M. guttula was used to colonies were picked from plates with a
isolate microcystin-degrading bacteria. bamboo stick and suspended in 20μl of PCR
reaction solution. A 20 μl reaction volume
Isolation of microcystin-degrading bacteria contained 1.6 μl of a 2.5 mM of dNTP, 2 μl
One milliliter of M. guttula culture was buffer, 2 μl (20 pmol) of each primer solution,
added to a 10 ml suspension of M. viridis 12.35 μl H2O, and 0.05 μl (0.25U) of TaKaRa
and was cultured at 30℃ in total darkness Ex Taq polymerase (Takara Bio, Inc.).
for ten days (to stationary-phase). The Thermal cycling was performed at 95℃ for 3
stationary-phase M. guttula culture was then min followed by 30 cycles of 95℃ for 30 s,
filtered through a membrane filter (pore size: 55℃ for 30 s, and 72℃ for 1 min (PTC 200,
5 μm, Minisart, Sartorius) to remove the M. MJ Research). The amplified products were
guttula from the suspension. Two milliliters purified using 20% polyethyleneglycol and
102 Japanese J. Wat. Treat. Biol. Vol.43 No.2

separated by 1% agarose gel electrophoresis in PY medium (pH 7.0), which consisted of 5

(Mupid 2, Advance Co.). After staining with g of peptone and 2.5 g of yeast extract in 1
ethidium bromide, amplification was liter of Milli Q water, at 30℃ for 24 h.
confirmed by UV illumination (Printgraph, Bacteria in the late logarithmic growth phase
Atto). were centrifuged at 6400 ×g for 20 min
DNA was sequenced with an ABI (Hitachi Koki Co., CR22GII). The supernatant
PRISM310 (Applied Biosystems) using the was decanted, and the pelleted bacteria were
BigDye Terminator Cycle Sequencing Ready resuspended in 0.05 M PBS buffer (pH 7).
Reaction kit (Applied Biosystems). For 16S This washing procedure was repeated three
rDNA gene sequencing, primers 20F, 1540R, times with centrifugation as above. Washed
350F (5’ CCTACGGGCAGCAGT 3’, positions bacteria were inoculated in M 11 medium or
341 358), 800F (5’ GTAGTCCACGCCGTAA glycine NaOH buffer at an optical density of
ACGA 3’, positions 800 819), and 900R (5’ 0.3 at 660 nm. The glycine NaOH buffer was
CGGCCGTACTCCCCAGGCGG 3’, positions used for pH 8.8 and 10. Bacterial suspension
898 879) were used. The sequencing reactions (9 ml) was mixed with 1 ml of 10 mg･l–1
(20 μl) contained 8μl of Premix (Applied microcystin to establish an initial concen-
Biosystems, diluted 1:10 with sequencing tration of 1000 μg･l–1 9). The culture volume
buffer), 3.2 μl of primer, 7.8 μl of H2O, and 1 was consequently set at 10 ml.
μl of template DNA. The template DNA was To investigate the effect of temperature on
prepared by diluting PCR products to 2–5 the microcystin degradation rate, the test
ng･μl–1. Thermal cycling was performed at tubes containing the above bacterial sus-
96℃ for 2 min followed by 90 cycles of 96℃ pension and microcystin RR were shaken
for 10 s, 50℃ for 5 s, and 60℃ for 4 min (130 rpm, Monosin, Taitec) in water baths
(PTC 200, MJ Research). After sequencing, a and kept at a constant temperature of 20, 25
composite sequence consisting of approxi- or 30℃. The pH was 7.6 in these samples.
mately 1500 base pairs was determined using To investigate the effect of pH on
AutoAssembler (Applied Biosystems). Related microcystin degradation, four pH values
sequences were obtained using the FASTA between 6.8 and 10 were evaluated. To
sequence search program22) of the DNA Data achieve pH 6.8 and 7.6, the pH of the M 11
Bank of Japan (DDBJ). medium was adjusted by adding 0.1 M HCl
The 16S rDNA sequences determined were and 0.1 M NaOH. To achieve pH 8.8 and 10,
aligned with 16S rDNA gene sequences of 50 mM glycine NaOH buffer was used. Test
other strains using CLUSTAL X software23). tubes containing the bacterial suspension
Some positions, including gaps and alignment and microcystin RR were shaken (130 rpm)
uncertainties, were omitted from the analysis. in a water bath kept at a constant
Evolutionary distance matrices were temperature of 25℃.
calculated using the algorithm of Jukes & The effect of glycine on microcystin RR
Cantor24) with the DNADIST program within degradation by the two strains was evaluated
the PHYLIP package25). A phylogenetic tree to assess the effect of glycine NaOH buffer
was constructed by the neighbor-joining on microcystin RR degradation in the pH
method26) as implemented within the experiment. Washed bacteria were inoculated
NEIGHBOR program of the PHYLIP package. in M 11 medium with and without 50mM
The stability of relationships was assessed glycine (pH 7.6), and microcystin RR was
by a bootstrap analysis of 1000 data sets added. As above, the bacterial suspension
using the programs SEQBOOT, DNADIST, was shaken and degradation of microcystin
NEIGHBOR and CONSENSE of the PHYLIP RR was evaluated.
package. To investigate the degradation of
microcystin RR, YR and LR, the test tubes
Effect of temperature, pH and initial containing bacterial suspension and each
microcystin concentration on the rate of microcystin were shaken (130 rpm) in water
microcystin degradation by the isolated baths kept at a constant temperature of
bacteria The isolated bacteria were grown 25℃. The pH was 7.6 in these samples.
 Microcystin-Degrading Bacteria Isolated from Culture of Protozoa 103

The above degradation experiments that Microcystins in the filtrates were analyzed
assessed the effects of temperature, pH, and by HPLC and a UV/VIS detector (239 nm,
microcystin analogs were performed in Waters 2417). A Symmetry C18 5μm column
triplicates. (Waters Co., 4.6 by 150 mm) was used. The
To investigate microcystin degradation toxins were identified and measured by
under various initial microcystin LR comparing their UV spectra and peak areas
concentrations, the bacterial suspension and with those of samples spiked with purified
microcystin LR solution were mixed to microcystin YR, RR and LR standards (Wako
achieve an initial microcystin concentration Pure Chemical Industries, Ltd.).
of 20 to 890 μg･l–1 for strain MG 15 and 17
RESULTS
to 917 μg･l–1 for strain MG 22. The test
tubes were shaken (130 rpm) in a water Isolation of microcystin-degrading bacteria
bath kept at a constant temperature of 25℃. from a M. guttula culture and phylogenetic
This experiment was repeated twice at each analysis of isolated bacteria Twenty-five
microcystin concentration. The pH was 7.6 in strains of bacteria were isolated from a M.
these samples. guttula culture. Colonies were randomly
In all degradation experiments, the selected from the media. These strains were
bacterial suspensions were removed from the named MG 1 through MG 25, and their
test tubes periodically and the residual ability to degrade microcystin LR was tested;
microcystin concentration was measured two strains, MG 15 and MG 22, degraded
using HPLC. Initial and final bacterial approximately 70% of the available micro-
numbers were determined by DAPI staining cystin LR in 7days. The other 23 strains of
and counting27) using fluorescence microscopy bacteria did not degrade microcystin LR in 7
(BZ 8000, Keyence). The bacterial suspension days.
was diluted with sterilized Milli Q water, The 16S rDNA gene sequences of strains
and DAPI stock solution (Sigma Co.; 10 MG 15 and MG 22 that we determined were
μg･ml–1 in sterilized Milli Q water) was continuous stretches of 1433 and 1402 bp,
added (final concentration, 1μg･ml–1). DAPI respectively. Searches for similarity with the
stained bacteria were trapped by vacuum FASTA sequence search program indicated
onto a black membrane filter (pore size: 0.2 that the closest relative of MG 15 and MG
μm, Advantec). Filters were then air-dried 22 was alpha proteobacterium F0813 (100.0%
and mounted on glass microscope slides with and 99.7% identity, respectively, GenBank
non-fluorescence immersion oil (Olympus accession number AF235997).
Co.). DAPI stained bacteria were enumerated The 16S rDNA similarity values of strains
under UV excitation. The rate of microcystin MG 15 and MG 22 with the type species of
degradation was calculated by dividing the the representative genera of the family
amount of microcystin degraded in the first Sphingomonadaceae were 93.2 and 93.3 with
30 min or 1 h by the initial cell count. Sphingobium chlorophenolicum ATCC 33790T
Statistically significant differences were (X87161), 94.6 and 94.3 with Novosphingobium
determined by one-way analysis of variance subarcticum KF1T (X94102), 94.4 and 94.1
(ANOVA) and Tukey’s honestly significant with Sphingopyxis terrae IFO 15098T
difference (HSD) post hoc test based on the (D13727), and 93.6 and 93.3 with
ANOVA using Kaleida Graph version 3.6 Sphingopyxis macrogoltabida IFO 15033T
(Synergy Software). (D13723), respectively. The 16S rDNA
similarity values of strains MG 15 and MG
Microcystin analysis Microcystins were 22 with the microcystin-degrading bacteria
analyzed using HPLC. The mobile phase were 93.2 and 92.9 with Sphingosinicella
consisted of 60% / 40% methanol / 0.05M microcystinivorans Y2 (AB084247), 93.6 and
phosphate buffer (pH 3.0). Water samples 93.4 with Sphingomonas sp. MD 1
obtained from the microcystin degradation (AB110635), and 92.4 and 92.3 with
test were filtered through a membrane filter Sphingomonas sp. MJ PV ACM 3962
(pore size: 0.2 μm, DISMIC 25cs, Advantec). (AF411072), respectively.
104 Japanese J. Wat. Treat. Biol. Vol.43 No.2

The phylogenic tree based on the sequences minutes of incubation, whereas the decrease
of the bacteria isolated in this study, the of microcystin RR by strain MG 22 was
microcystin-degrading bacteria isolated linear during 1 h of incubation. The initial
previously, and related strains, is shown in degradation rate of microcystin RR by strain
Fig. 1. The two strains MG 15 and MG 22 MG 15 increased as temperature increased
were nested in the clusters of Sphingopyxis (Table 1). An optical density of 0.3 corre-
species and Novosphingobium species. sponded to cell density of 6.5 × 108 cells･ml–1
for both strains. The degradation rates of
Effect of temperature on microcystin microcystin in all experiments were calculated
degradation by isolated bacteria Micro- by dividing the microcystin degradation rate
cystin RR degradation by the novel bacterial by the cell density. The rate of degradation
strains MG 15 and MG 22 was evaluated at 30℃ by strain MG 15 was significantly
between 20 and 30℃. Microcystin RR different from that at 20 and 25℃ (p＜0.05).
decreased after the start of the experiment The initial degradation rate of microcystin
at each temperature assayed (Fig.2). Strain RR by strain MG 22 increased slightly with
MG 15 degraded microcystin RR faster than increasing temperature, but this apparent
strain MG 22. The decrease of microcystin increase was not statistically significant.
RR by strain MG 15 was linear during 30

Sphingomonas sp. IFO 15915 (AB033949)

778
Sphingopyxis chilensis S37 (AF367204)
804

976 Sphingopyxis macrogoltabida IFO 15033T (D13723)

993 Sphingomonas sp. IFO 15917 (AB033950)

Sphingopyxis terrae IFO 15098T (D13727)

359

MG-15
1000
MG-22
501

Sphingomonas sp. MJ-PV ACM-3962 (AF411072)

1000
Sphingomonas sp. MD-1(AB110635)
913

615
Novosphingobium subarcticum KF1T (X94102)

Sphingomonas agrestis HV3 (Y12803)

468
Sphingobium japonicum UT26 (AF039168 )
1000

1000 Sphingobium chlorophenolicum ATCC 33790T (X87161)

298
Sphingobium yanoikuyae EC-S01 (AB120764)

Sphingomonas parapaucimobilis IFO 15100T (D13724)

1000
Sphingomonas sanguinis IFO 13937T (D13726)
998

892 Sphingomonas sp. IFO 15496 (AB033946)

Sphingomonas echinoides ATCC 14820T (AB021370 )

Sphingosinicella microcystinivorans Y2 (AB084247)

Rhodospirillum rubrum ATCC 11170T (D30778)

0.01

Fig. 1 Phylogenetic relationships between isolated bacterial strains MG 15 and MG 22, currently known microcystin-
degrading bacteria7)9)11) and closely related strains. The tree is based on a distance matrix analysis of the 16S
rDNA sequences (accession numbers given in parentheses). Numbers at the nodes indicate bootstrap values,
derived from 1000 samples. Scale represents one substitution per 100 nucleotide positions. Strains in bold form
are microcystin-degrading bacteria.
 Microcystin-Degrading Bacteria Isolated from Culture of Protozoa 105

100 Effect of pH on microcystin degradation by

isolated bacteria Microcystin RR degra-
Microcystin RR remaining (%)

80 dation by MG 15 and MG 22 was evaluated

between pH 6.8 and 10. Microcystin RR
60 decreased after the start of the experiment
at each pH (Fig. 3). The microcystin RR
40 concentration decreased to less than 20%
after 9 h in the presence of strain MG 15 at
each pH, but more than 20% of the microcystin
20
remained after 9 h of incubation with strain
MG 22 at pH 6.8 and 10. No change in pH
0
0 2 4 6 8 10 was observed in samples containing glycine
Time (h) NaOH buffer after 9 h. The pH changed
a) MG-15 slightly in the samples containing M 11
medium initiated at pH 6.8 or 7.6 after 9 h
100 (samples initiated at pH 7.6 changed to 7.3
for MG 15, and to 7.5 for MG 22; samples
Microcystin RR remaining (%)

80
initiated at pH 6.8 changed to 7.2 for both
strains). The initial degradation rate of
microcystin RR was calculated (Table 2). The
60
degradation rate was the highest at pH 7.6
and 8.8 for strain MG 15 and at pH 8.8 for
40
strain MG 22, and the rate was the lowest
at pH 10 for both strains. ANOVA showed
20 that the pH affected the degradation rate by
both strains (p＜0.0001). The degradation
0 rate by strain MG 15 was significantly
0 2 4 6 8 10
Time (h)
different at each pH (p＜0.05, Table 2) except
for between the degradation rates at pH 7.6
b) MG-22
and 8.8. In the case of strain MG 22, the
Fig. 2 Degradation of microcystin RR from an initial degradation rate was significantly different
concentration of 1000 μg l–1 over a 9 h period at each pH (p＜0.001) except for between the
by strains MG 15 and MG 22 as a function of
degradation rates at pH 6.8 and 10.
temperature (pH 7.6). Error bars indicate
standard deviations (n=3). Addition of 50 mM glycine did not affect
○ 20℃ ● 25℃ ◇ 30℃ the degradation rate of microcystin RR for
strain MG 15 (without glycine: 6.27 ± 0.58
(×10–4pg･cell–1･h–1), with glycine: 5.85±0.33
Table 1 Initial degradation ratea of microcystin RR by (×10–4pg･cell–1･h–1), but glycine significantly
strains MG 15 and MG 22 as a function of enhanced the degradation rate for MG 22
temperature.
(p＜0.0001) (without glycine: 5.0±0.18
20℃ 25℃ 30℃ (×10–4pg･cell–1･h–1), with glycine: 8.12±0.53
(×10–4pg･cell–1･h–1)).
MG 15 9.81 ± 0.73 12.2 ± 0.83 26.4 ± 8.08b
MG 22 3.90 ± 1.14 5.58 ± 0.43 5.98 ± 2.64 Degradation of microcystin RR, microcystin
a
Rates are expressed as ×10–4 pg microcystin･cell–1･h–1 YR and microcystin LR This experiment
measured at pH 7.6 and represent the mean±standard examined the degradation rate of three
deviation (n=3).
b
Value is significantly different (p＜0.05) from the other microcystin analogs by strains MG 15 and
values of MG 15. MG 22. Microcystin RR was degraded faster
than microcystin YR and microcystin LR by
both strains (Fig. 4). In the test for strain
MG 15, the degradation rates of microcystin
YR and microcystin LR were the same (Table
106 Japanese J. Wat. Treat. Biol. Vol.43 No.2

100 100
Microcystin RR remaining (%)

Microcystin remaining (%)

80 80

60 60

40 40

20 20

0 0
0 2 4 6 8 10 0 2 4 6 8 10
Time (h) Time (h)
a) MG-15
a) MG-15

100
100
Microcystin RR remaining (%)

Microcystin remaining (%)

80
80

60
60

40
40

20
20

0
0 0 2 4 6 8 10
0 2 4 6 8 10
Time (h)
Time (h)
b) MG-22
b) MG-22

Fig. 3 Degradation of microcystin RR from an initial Fig. 4 Degradation of microcystin RR, YR and LR from
concentration of 1000 μg l–1 over a 9 h period an initial concentration of 1000 μg l–1 over a 9 h
by strains MG 15 and MG 22 as a function of period by strains MG 15 and MG 22 (25℃, pH
pH (25℃). Error bars indicate standard 7.6). Error bars indicate standard deviations
deviations (n=3). (n=3).
○ pH 6.8 ● pH 7.6 ○ microcystin RR ● microcystin YR
◇ pH 8.8 △ pH 10 ◇ microcystin LR

Table 2 Initial degradation ratea of microcystin RR by Table 3 Initial degradation ratea of microcystin RR, YR
strains MG 15 and MG 22 as a function of and LR by strains MG-15 and MG-22.
pH.
MC RR MC YR MC LR
pH 6.8 pH 7.6 pH 8.8 pH 10
MG-15 12.2 ± 0.83 4.06 ± 1.07 4.06 ± 0.15
MG 15 6.99 ± 0.15b 8.75 ± 0.28c 8.19 ± 0.83c 4.43 ± 0.33d
MG-22 5.58 ± 0.43 3.27 ± 0.67 2.60 ± 0.24
MG 22 2.80 ± 0.25b 5.07 ± 0.19c 7.54 ± 0.18d 2.20 ± 0.5b aRates are expressed as ×10–4 pg microcystin･cell–1･h–1
aRates are expressed as ×10–4 pg microcystin･cell–1･h–1 measured at 25℃, pH 7.6 and represent the mean±standard
measured at 25℃ and represent the mean±standard deviation (n=3).
deviation (n=3).
b, c, d There is a significant difference (p＜0.05) between

values marked with different superscript letters.

 Microcystin-Degrading Bacteria Isolated from Culture of Protozoa 107

3). The degradation rate of microcystin RR l–1 for strain MG 15 and 17 to 917 μg･l–1 for
by MG 15 was three times higher than that strain MG 22. In all of the experiments, the
of microcystin YR and microcystin LR. In the microcystin LR concentration decreased (Fig.
test for strain MG 22, microcystin LR had 5). Microcystin LR degradation over time was
the lowest degradation rate. calculated as the average of results from two
experiments. The initial degradation rate of
Degradation of microcystin LR as a function microcystin LR was calculated based on
of initial microcystin concentration Micro- degradation during 1 h of incubation and
cystin LR degradation by the novel bacterial was plotted against the initial concentration
strains MG 15 and MG 22 was evaluated at (Fig. 6). The degradation rate of microcystin
initial concentrations between 20 to 890 μg･ LR increased as initial microcystin LR
concentration increased. The difference in
the degradation rate between strains MG 15
1000 and MG 22 also increased with concentration.
The microcystin LR degradation rate did not
800 plateau at the highest microcystin LR
Microcystin LR (µg L–1)

concentration for either MG 15 or MG 22.

600
DISCUSSION
400 Isolation and phylogenetic analysis of
microcystin-degrading bacteria isolated from a
culture of M. guttula Phylogenetic analysis
200
of the 16S rDNA gene sequences and
parsimony methods indicated that the
0
0 2 4 6 8 10 currently known species of the genus
Time (h) Sphingomonas can be divided into four
a) MG-15 clusters, and three new genera
Novosphingobium, Sphingopyxis and
1000
Sphingobium have been proposed28). According
to the phylogenetic tree, microcystin-
800
Microcystin LR (µg L–1)

4
Degradation rate of microcystin LR

600
(×10–4 pg cell–1 h–1)

400 3

200
2
0
0 2 4 6 8 10
Time (h)
1
b) MG-22

Fig. 5 Degradation of microcystin LR from varying

initial concentrations over a 9 h period by 0
strains MG 15 and MG 22 (25℃, pH 7.6). 0 200 400 600 800 1000
a) MG 15
○ 20 μg l–1 ● 41 μg l–1 Initial microcystin LR concentration (µg L–1)
△ 77 μg l–1 ■ 165 μg l–1
◇ 337 μg l–1
b) MG 22
◫ 890 μg l–1 Fig. 6 Initial degradation rate of microcystin LR (×10–4
pg cell–1 h–1) by strains MG 15 and MG 22 as
○ 17 μg l–1 ● 40 μg l–1 a function of initial microcystin LR concentration
△ 81 μg l–1 ■ 173 μg l–1 (μg l–1) (25℃, pH 7.6).
◇ 345 μg l–1 ◫ 917 μg l–1 ○ MG 15 ◆ MG 22
108 Japanese J. Wat. Treat. Biol. Vol.43 No.2

degrading strains Sphingomonas sp. MJ PV7) installed to remove algae and dissolved
and Sphingomonas sp. MD 19) were associated organic matter. In such biological processes,
with the Novosphingobium group. Strains a biofilm consisting of bacteria, protozoa and
MG 15 and MG 22 exhibited low similarity metazoa becomes attached to the media
values of less than 95% with the Sphingopyxis made of sand, plastic and ceramic and
and Novosphingobium groups. In the thereby purifies water. Saito et al. reported
phylogenetic tree, strains MG 15 and MG 22 that when the biofilm was sampled and then
were separated from the Sphingopyxis cluster added to M. viridis suspension, Monas spp.
with a low bootstrap value of 359. These low increased with a concominant decrease of M.
similarity values and low bootstrap value viridis; microcystin also decreased16). Thus,
could not specify the phylogenetic group, and Monas spp. and microcystin-degrading
it was shown that strains MG 15 and MG 22 bacteria might exist in the biofilm and
were related to the Sphingopyxis and contribute to the purification. Ho et al.
Novosphingobium groups. examined removal of 2 methylisoborneol,
Two of the 25 strains isolated were shown geosmin and microcystin LR by biological
to degrade microcystin, suggesting that many sand filtration. They reported that a homolog
bacterial species may be present in the M. of the gene mlrA was detected in the biofilm
guttula culture and that microcystin- forming on the sand15). These findings suggest
degrading bacteria such as strains MG 15 that when using biological processes to treat
and MG 22 may play an important role in lake and reservoir water contaminated with
microcystin degradation. Microcystis, protozoa such as M. guttula can
We previously studied growth of the prey upon Microcystis and that excreted
protozoan ciliate Trithigmostoma cucullulus microcystin can be degraded by microcystin-
preying on Planktothrix agardhii (PCC 7821, degrading bacteria.
obtained from Institut Pasteur) which
produces microcystin RR in cell. With growth Characteristics of microcystin degradation
of T. cucullulus, P. agardhii decreased due by the isolated bacteria Initial cell density
to predation. Simultaneously dissolved for each strain was 6.5 × 108 cells･ml–1.
microcystin RR increased (unpublished data). After 9 h, cell density tended to stay constant
Consequently, we surmised that T. cucullulus or decrease (data not shown). During
ingested trichomes of P. agardhii and that degradation experiments over 9 h, 1000 μg･l–1
microcystin RR was excreted intact by T. of microcystin may not have been sufficient
cucullulus after P. agardhii had been to support the bacteria because no growth
consumed. In the case of M. guttula used in was observed.
this study, it was presumed that M. guttula The effect of temperature on microcystin
prey on M. viridis and that microcystin was degradation differed for the two strains. The
excreted intact as was seen for T. cucullulus. rate of microcystin RR degradation by strain
M. guttula has been maintained for more MG 15 increased at 30℃, whereas no
than ten years in a microcystin-producing significant difference was observed in the
culture of M. viridis. In the M. guttula rate of degradation by strain MG 22 at each
culture, microcystins might be released from temperature. The difference in the effect of
M. viridis cells by predation and autolysis, temperature on microcystin degradation rate
so it is presumed that microcystin-degrading between the two strains might be due to
bacteria either survived or had accumulated activity of compounds involved in degradation
over a long period by utilizing microcystin. in cells. The degradation rate of microcystin
This hypothesis might be supported by the RR by strain Y2 increased with an increase
finding of Park et al. that the microcystin- in temperature from 5 to 30℃10). Generally,
degrading strain Y2 can use microcystin as a algal blooms occur at temperatures above
carbon and energy source10). 20℃. Indeed, both of the microcystin-
In water purification plants, biological degrading strains MG 15 and MG 22 could
treatments such as biological filtration and degraded microcystin at temperatures at
submerged filter bed processes have been which algal blooms occur.
 Microcystin-Degrading Bacteria Isolated from Culture of Protozoa 109

Strains MG 15 and MG 22 could also MG 15 and MG 22, whereas the bio-

degrade microcystin at pH levels ranging degradability of microcystin YR and RR
from 6.8 to 10. The initial degradation rate tended to be different between species. Jones
of microcystin RR peaked at pH 7.6 and 8.8 et al. reported that a cell-free extract of
for strain MG 15 and at pH 8.8 for strain Sphingomonas sp. MJ PV could degrade both
MG 22. Enhancement of the degradation microcystin RR and microcystin LR at similar
rate at pH 8.8 might be due to the effect of rates7). Further investigation of the
glycine for strain MG 22. The precise trend degradation rate of each microcystin analog
along the pH gradient could not be evaluated using cell-free extracts of strains MG 15 and
for strain MG 22; however it was shown MG 22 is necessary to establish bio-
that the degradation rate for the two strains degradability of each microcystin analog.
was high at pH 7.6 and 8.8. Saitou et al. In the experiment that examined the
reported the characteristics of the microcystin- degradation rate as a function of initial
degrading bacteria Sphingomonas sp. MD 1, microcystin LR concentration, the rate of
for which the percentage of microcystin LR degradation by both strains did not reach a
degraded peaked at pH 7.2 (more than 98%) saturation point. Park et al. investigated the
and decreased as pH increased9). At pH 9, effect of initial concentration of microcystin
the percentage degraded was approximately RR (4, 7, 18 and 37 mg･l–1) and LR (3, 5, 10
less than 30%. Strains MG 15 and MG 22 and 20 mg･l–1) on the degradation rate by
degraded 90% and more than 70% of strain Y210). For both microcystins, the
microcystin RR, respectively, at pH 10. This degradation rate tends to be saturated at
suggests that compounds involved in approximately 20 mg･l–1.
degradation of microcystin are active under At the initial microcystin LR concentrations
alkaline condition for strains MG 15 and of less than 400 μg･l–1, the degradation rate
MG 22. In eutrophic lakes in summer, the of microcystin LR by strain MG 15 was
pH increases up to 9 or 9.5 because of slightly higher than that of MG 22. At
photosynthesis29). Further investigation of the approximately 900 μg･l–1, however, the
growth as a function of pH is necessary to degradation rate by strain MG 15 was
surmise microcystin degradation under the substantially higher (Fig. 6). This suggested
alkaline conditions present in eutrophic that high activity or a high level of
lakes. degradation-associated compound was present
The initial degradation rate of microcystin in the MG 15 strain. Park et al. reported
RR was the highest of the three microcystins that the microcystin concentration during
for both bacterial strains. The rates of the warm season in Lake Suwa, Japan,
microcystin YR and microcystin LR deg- ranged 0.11 to 184 μg･l–1 30). Microcystin LR
radation were same for strain MG 15, decreased to the detection limit within 9 h
whereas the rate of microcystin YR from an initial concentration of 20, 41, 77
degradation was higher than that of and 165 μg･l–1for strain MG 15, and from
microcystin LR for strain MG 22. The 17, 40, 81 and 173 μg･l–1 for strain MG 22.
degradation observed for strain MG 22 These results indicated that each of the
toward the three microcystins was similar to bacterial strains identified in this study could
that of activity of Sphingomonas sp. MD 19). degrade microcystin at the actual
Park et al. reported that the rate of concentration that occurs in lakes and
microcystin RR degradation by strain Y2 is reservoirs.
about twice as high as that of microcystin
LR at an initial concentration 3 to 20 mg･l–1
CONCLUSIONS
and that the rate of microcystin YR In the research of the isolation of
degradation is about 10 times higher than microcystin-degrading bacteria from a
that of microcystin LR at an initial monoxenic culture of M. guttula, and the
concentration of 22 mg･l–1 10). The microcystin characteristics of microcystin degradation,
LR degradation rate was the lowest among the following results were obtained.
the three analogs for strains MD 1, Y2, (1) Two microcystin-degrading strains of
110 Japanese J. Wat. Treat. Biol. Vol.43 No.2

bacteria (MG 15 and MG 22) were isolated and colorectal cancer, Biomed. Environ.
from a monoxenic culture of Monas guttula Sci., 15, 166 171 (2002)
preying on Microcystis viridis. 6 ）World Health Organization: Guidelines for
(2) Strains MG 15 and MG 22 were related drinking-water quality, 2nd ed. Addendum
to Sphingopyxis and Novosphingobium groups to Volume 2, Health criteria and other
based on phylogenetic analysis of 16S rDNA. supporting information, Geneva, Switzer-
(3) The rate of microcystin RR degradation land: WHO (1998)
by strain MG 15 significantly increased at 7 ）Jones, G. J., Bourne, D. G., Blakeley, R. L.,
30℃, whereas the effect of temperature on and Doelle, H.: Degradation of the cyano-
the degradation rate by strain MG 22 was bacterial hepatotoxin microcystin by
small. aquatic bacteria, Nat. Toxins, 2, 228 235
(4) Strains MG 15 and MG 22 could degrade (1994)
microcystin RR under alkaline pH. 8 ）Bourne, D. G., Jones, G. J., Blakeley, R. L.,
(5) Strains MG 15 and MG 22 could degrade Jones, A., Negri, A. P., and Riddles, P.:
microcystin RR, YR and LR. The degradation Enzymatic pathway for the bacterial
rate of microcystin RR by the two strains degradation of the cyanobacterial cyclic
was faster than the degradation rates of peptide toxin microcystin LR, Appl.
microcystin YR and LR. Environ. Microbiol., 62, 4086 4094 (1996)
9 ）Saitou, T., Sugiura, N., Itayama, T., Inamori,
Y., and Matsumura, M.: Degradation
REFERENCES characteristics of microcystins by isolated
1 ）Carmichael, W. W., Beasley, V. R., Bunner, D. bacteria from Lake Kasumigaura, Journal
L., Eloff, J. N., Falconer, I. R., Gorham, P., of Water Supply: Research and Technology
Harada, K., Krishnamurthy, T., Yu, M. J., AQUA, 52, 13 18 (2003)
Moore, R. E., Rinehart, K., Runnegar, M., 10）Park, H. D., Sasaki, Y., Maruyama, T.,
Skullberg, O. M., and Watanabe, M. F.: Yanagizawa, E., Hiraishi, A., and Kato, K.:
Naming of cyclic heptapeptide toxins of Degradation of the cyanobacterial
cyanobacteria (blue-green algae), Toxicon, hepatotoxin microcystin by a new
26, 971 973 (1988) bacterium isolated from a hypertrophic
2 ）Sivonen, K., Namikoshi, M., Evans, W. R., lake, Environ. Toxicol. 16, 337 343
Carmichael, W. W., Sun, F., Rouhiainen, L., (2001)
Luukkainen, R. R., and Rinehart, K. L.: 11）Maruyama, T., Park, H. D., Ozawa, K.,
Isolation and characterization of a variety Tanaka, Y., Sumino, T., Hamana, K., Hiraishi,
of microcystins from seven strains of the A., and Kato, K.: Sphingosinicella
cyanobacteria genus Anabaena, Appl. microcystinivorans gen. nov., sp. nov., a
Environ. Microbiol., 58, 2495 2500 (1992) microcystin-degrading bacterium, Int. J.
3 ）Carmichael, W. W., Azevedo, S. M. F. O., An, Syst. Evol. Microbiol., 56, 85 89 (2006)
J. S., Moloca, R. J. R., Jochimsen, E. M., Lau, 12）Bourne, D. G., Riddles, P., Jones, G. J.,
S., Rinehart, K. L., Shaw, G. R., and Smith, W., and Blakeley, R. L.: Charac-
Eaglesham, G. K.: Human fatalities from terisation of a gene cluster involved in
cyanobacteria: chemical and biological bacterial degradation of the cyanobacterial
evidence for cyanotoxins, Environ. Health toxin microcystin LR, Environ. Toxicol.
Perspect., 109, 663 668 (2001) 16, 523 534 (2001)
4 ）Nishiwaki-Matsushima, R., Ohta, T., 13）Saito, T., Okano, K., Park, H. D., Itayama, T.,
Nishiwaki, S., Suganuma, M., Kohyama, K., Inamori, Y., Neilan, B. A., Burns, B. P., and
Ishikawa, T., Carmichael, W. W., and Fujiki, Sugiura, N.: Detection and sequencing of
H.: Liver tumor promotion by the the microcystin LR degrading gene, mlrA,
cyanobacterial cyclic peptide toxin micro- from new bacteria isolated from Japanese
cystin LR, J. Cancer Res. Clin. Oncol., lakes, FEMS Microbiol. Lett., 229, 271
118, 420 424 (1992) 276 (2003)
5 ）Zhou, L., Yu, H., and Chen, K.: Relationship 14）Grutzmacher, G., Bottcher, G., Chorus, I.,
between microcystin in drinking water and Bartel, H.: Removal of microcystins by
slow sand filtration, Environ. Toxicol. 17,
 Microcystin-Degrading Bacteria Isolated from Culture of Protozoa 111

386 394 (2002) Jeanmougin, F., and Higgins, D. G.: The

15）Ho, L., Hoefell, D., Aunkofer, W., Meyn, T., ClustalX windows interface: flexible
Keegan, A., Brookes, J., Saint, C., and strategies for multiple sequence alignment
Newcombe, G.: Biological filtration for the aided by quality analysis tools, Nucleic
removal of algal metabolites from drinking Acids Res., 24, 4876 4882 (1997)
water, Water Sci. Technol.: Water Supply, 24）Jukes, T. H. and Cantor, C. R.: Evolution of
6, 153 159 (2006) protein molecules, In: Munro, H. N.,
16）Saito, T., Sugiura, N., Itayama, T., Inamori, editor, Mammalian Protein Metabolism,
Y., and Matsumura, M.: Biodegradation of Vol. 3, New York: Academic Press, p 21
Microcystis and microcystins by indigenous 132 (1969)
nanoflagellates on biofilm in a practical 25）Felsenstein, J.: PHYLIP. Phylogeny in-
treatment facility, Environ. Technol., 24, ference package, Cladistics, 5, 164 166
143 151 (2003) (1989)
17）Xie, S. G., Tang, X. Y., Wu, W. Z., Wen, D. H., 26）Saitou, N. and Nei, M.: The neighbor-joining
and Wang, Z. S.: Biological pretreatment of method: a new method for reconstructing
Yellow River water, J. Environ. Sci., 17, phylogenetic trees, Mol. Biol. Evol., 4,
557 561 (2005) 406 425 (1987)
18）Iwami, N., Inamori, Y., Sugiura, N., and 27）Yamaguchi, N., Kenzaka, T., and Nasu, M.:
Matsumura, M.: Degradation characteris- Rapid in situ enumeration of physiologically
tics of toxic Microcystis viridis by active bacteria in river waters using
mastigophora, Monas guttula using aquat- fluorescent probes, Microbes Environ., 12,
ic large scale microcosm, J. Jpn. Soc. 1 8 (1997)
Water Environ.,19, 140 146 (1996) (in 28）Takeuchi, M., Hamana, K., and Hiraishi, A.:
Japanese) Proposal of the genus Sphingomonas
19）Kudo, R. R.: Protozoology, Springfield sensu stricto and three new genera,
Illinois USA: Charles C. Thomas Publisher, Sphingobium, Novosphingobium and
314 p. (1977) Sphingopyxis, on the basis of phylogenetic
20）Brosius, J., Palmer, J. L., Kennedy, J. P., and and chemotaxonomic analysis, Int. J. Syst.
Noller, H. F.: Complete nucleotide sequence Evol. Microbiol., 51, 1405 1417 (2001)
of a 16S ribosomal RNA gene from 29）Shitara, K., Nomura, M., Chiba, N., Fujimoto,
Escherichia coli, Proc. Natl. Acad. Sci. N., Nakano, K., and Nishimura, O.: Char-
USA, 75, 4801 4805 (1978) acteristics of water quality stratification
21）Ishikawa, M., Nakajima, K., Yanagi, M., related to the water-bloom occurrence in
Yamamoto, Y., and Yamasato, K.: Urushizawa dam lake, Environ. Eng. Res.,
Marinilactibacillus psychrotolerans gen. 42, 487 493 (2005) (in Japanese)
nov., sp. nov., a halophilic and alkaliphilic 30）Park, H. D., Iwami, C., Watanabe, M. F.,
marine lactic acid bacterium isolated from Harada, K., Okino, T., and Hayashi, H.:
marine organisms in temperate and Temporal variabilities of the concentrations
subtropical areas of Japan, Int. J. Syst. of intra- and extracellular microcystin
Evol. Microbiol. 53, 711 720 (2003) and toxic Microcystis species in a
22）Pearson, W. R., and Lipman, D. J.: Improved hypertrophic lake, Lake Suwa, Japan
tools for biological sequence analysis, (1991 1994), Environ. Toxicol. Water
Proc. Natl. Acad. Sci. USA, 85, 2444 2448 Qual., 13, 61 72 (1998)
(1988) (Submitted 2007. 1. 11)
23）Thompson, J. D., Gibson, T. J., Plewniak, F., (Accepted 2007. 3. 22)