First report of microcystins in Taiwan

FIRST REPORT OF MICROCYSTINS IN TAIWAN

TZONG-HUEI LEE,

_{YIH-MIN CHEN}

_{and HONG-NONG CHOU}

_*

1_{Department of Zoology, National Taiwan University, Taipei, 10617, Taiwan R.O.C.; and} 2_{Institute of Fisheries Science, National Taiwan University, Taipei, 10617, Taiwan R.O.C.}

(Received 3 June 1997; accepted 13 September 1997)

Tzong-Huei Lee, Yih-Min Chen and Hong-Nong Chou. First report of

microcystins in Taiwan. Toxicon 36, 247±255, 1998.ÐThis is the ®rst report

on microcystins from Microcystis aeruginosa KuÈtzing in Taiwan. A total of

nine strains of cyanobacteria have been isolated from eutrophic aquaculture

ponds and water reservoirs. By mouse toxicity assay, six of the nine strains

had

in the range of 25±100 mg per kg mouse for dried bacterial mass.

Microcystin-LR and -RR were found in all toxic strains and their contents

ranged from 0.11±10.06 mg and 0.08±2.21 mg per mg of dried bacteria,

re-spectively. Microcystin-RA, a minor component found only in M. TN-2 and

M. CY-1 strains, was identi®ed as a new microcystin. All three toxins were

isolated by a serial separation on an LH-20 column, Si-¯ash column

chroma-tography and reverse phase HPLC. Toxins were further identi®ed by

com-paring their FABMS,

_{H and}

_H±

_{H COSY NMR spectra with the}

authentic microcystin-LR. Several other microcystin-like compounds were

also found in the cultured strains and their structures are being determined.

# 1998 Elsevier Science Ltd. All rights reserved

Keywords: Cyanobacteria, Microcystis aeruginosa, Hepatotoxin, Microcystins

INTRODUCTION

Microcystis has been known to be the major genus among the cyanobacteria to cause

blooms in fresh waters (Carmichael, 1992). Most species belonging to this genus have

been reported to produce a family of over 50 hepatotoxic cyclic peptides which are

termed microcystins (MCYST). Cyclo (-

-Ala-X-

-Me-iso-Asp-Z-Adda-

-iso-Glu-Mdha-) was developed in the literature to describe their cyclic peptide structure, where

Adda represented (2S, 3S, 8S,

9S)-3-amino-9-methoxy-2,6,8-trimethyl-10-phenyldeca-4,6-dienoic acid, a special structural feature responsible for their toxicity (Saito et al.,

1994), MeAsp represented the

-erythro-b-methylaspartic acid, and Mdha,

N-methylde-hydroalanine (Carmichael, 1992). X and Z in the structure denote the variable

-form

amino acids at the 2nd and 4th position, respectively, that were known to include

argi-nine, leucine, phenylalaargi-nine, tryptophan, . . ., etc. to form the various cyclic

heptapep-tides (Fig. 1).

# 1998 Elsevier Science Ltd. All rights reserved Printed in Great Britain 0041-0101/98 $19.00 + 0.00

PII: S0041-0101(97)00128-1

* Author to whom correspondence should be addressed. 247

Like many other areas in the world, Microcystis blooms frequently occur in the

eutrophic waters of Taiwan, especially in the aquaculture ponds for eel, tilapia and

carp. Toxicities and toxins of these Microcystis are of interest due to their threat to

humans and animals. There are also reports describing a possible link of human liver

disease due to trace microcystins in drinking waters from reservoirs with Microcystis

blooms (Falconer et al., 1983; Matsushima et al., 1992; Carmichael, 1994). In Taiwan a

relatively comprehensive survey for toxic Microcystis and their toxins was carried out

during the past four years.

MATERIALS AND METHODS Cyanobacterial cultures

Eight strains of Microcystis aeruginosa KuÈtzing and one Coelosphaerium kuetzingianum Naegeli strain were isolated from fresh water ponds and reservoirs at various locations in Taiwan. These species were collected from the water blooms and then cloned for laboratory cultures (Table 1). They were identi®ed according to the species described in the Plankton Algae of Reservoirs in Taiwan (Moriwaka and Chyi, 1996) and in Mizuno (1980). All clones were cultured in modi®ed Fitzgerald media (Hughes et al., 1958) at 23218C and illuminated with ¯uorescent light of 26.4 mmol quanta/m2_{/s for 12 h a day. Cyanobacterial cells were collected}

in their late exponential phase of growth and concentrated by continuous centrifugation, followed by lyophili-zation and storage. A 1.2 ton mass culture of M. TN-2 was carried out for toxin preparation.

Mouse assay

Dried cell mass (150 mg) of each strain of cultured Microcystis and Coelosphaerium was vortexed three times with 10 ml MeOH and the combined MeOH extract was dried in vacuum. Dried extract from each cya-nobacterial strain was redissolved in 1 ml saline solution for mouse toxicity assay using three mice for each dose level. Mice of 20 g, male, ICR strain, were injected intraperitoneally with the bacterial extracts and they were observed for at least 4 h for the lethal result (Aune and Berg, 1986). A toxicity threshold dosage was de®ned in this experiment to show the relative toxicity among strains. The toxicity threshold dose, expressed as the dry weight of bacterial mass per kg of mouse, is the least quantity of extract that kills all the triplicates of the mice. Six dose levels of the extract equivalent to 1/32, 1/8, 1/2, 2, 8, and 32 mg dried cell mass of each

strain were tested for the toxicity threshold. Dead mice were dissected to observe the swollen blood-engorged liver that is generally recognized as an indicator of microcystin poisoning (Azevedo et al., 1994).

A ®eld-collected sample, consists of M. aeruginosa cells from the eel pond scum, was studied in a similar way to compare its toxicity with the cultured strains. One gram of dried cells was extracted three times with 30 ml MeOH. The dried extract was then diluted with various amounts of saline solution to give 5 di€erent doses equivalent to 1, 5, 10, 50, and 100 mg of dried cells per ml. Six mice were used for each dose level as duplications.

Sample preparation and high performance liquid chromatography

Lyophilized cells (50 mg) of each cyanobacterial strain were extracted 3 times with 3 ml MeOH and then centrifuged at 1500 g for 10min to remove cell debris. Supernatant was vacuum-dried and redissolved in trace amount of ethylacetate: isopropanol (4:3, v/v) solution for solid phase extraction. Accubond silica SPE col-umns (6 ml, 0.5 g silica gel) (Fison, England) were preconditioned with 6 ml of the same solution and then washed with 10 ml of the solution after the load of samples. The toxin fraction was eluted with 4 ml of ethyla-cetate: isopropanol: H2O (4:2.5:1.5, v/v) through the column and evaporated to remove the solvent. The eluent

volume had been determined through a preliminary experiment using the samples of toxin standards (MCYST-LR and MCYST-RR) mixed with crude extracts. Toxin fractions of each bacterial sample were dried and redissolved in 250 ml MeOH, using 2 ml for each HPLC analysis. The high performance liquid chro-matography (HPLC) used a 4.6  250 mm Cosmosil 5C18-AR column (Nacalai Tesque, Japan) and an iso-cratic solution of 0.05% aq. tri¯uoroacetate: MeOH (48:52, v/v) as mobile phase, ¯ow rate 1 ml/min.

Toxin extraction and isolation

Dried cells (24 g) from M. aeruginosa strain M. TN-2 were extracted three times with 400 ml MeOH for 30 min each. The MeOH extract was adjusted to 85% in aqueous solution for hexane partition. The aqueous layer was then vacuum-evaporated to dryness (1.8 g) and redissolved in 90 ml MeOH for nine di€erent batches of chromatographic separations.

The ®rst separation step was carried out using gel ®ltration chromatography on a Sepadex LH-20 (Pharmacia Biotech. Sweden) column (3  55 cm) for each 10 ml sample solution. A ¯ow of 13 ml/min of MeOH was used to elute the toxin fractions. Each fraction was compared with an authentic sample of micro-cystin-LR (Sigma, U.S.A.) by thin layer chromatography (TLC) using plates of Silica gel 60, PF254, 200 mm thickness (Merck, U.S.A.) and a solution of ethylacetate: isopropanol: H2O (4:2.5:2, v/v/v) for development.

Vanillin±sulfuric acid charring to form dark blue spots in addition to UV absorption was used to detect the toxic components. Subsequently the toxin fractions were combined and chromatographed on a 3 cm i.d. ¯ash column (42.5 g, Baker's silica gel for ¯ash column, U.S.A.) using the same solution in TLC as the eluent, ¯ow rate 36 ml/min. The toxin fractions were combined into two major portions based on the result of TLC separ-ation and vanillin±sulfuric acid charring. Microcystin (MCYST)-LR, -RR, and -RA were further puri®ed from the above toxin fractions by repetitive HPLC separations on an Econosil C18 5U analytical column (Alltech, U.S.A.) using 0.1 M ammonium acetate: methanol (43:57, v/v) as mobile phase, ¯ow rate 1 ml/min. Puri®ed toxins were vacuum-dried and stored in ÿ208C freezer for further chemical structure analysis. Table 1. Collection date and site of Microcystis aeruginosa and Coelosphaerium kuetzingianum cultured in this

experiment and their threshold dosage in the mouse toxicity assay

Cyanobacterial strain Collection date Sampling site Threshold dosage (mg drybacteria/kg mouse) Microcystis aeruginosa

M. TY-1 Sep. 1992 Gongshi, Tauryuan 25

M. TY-2a _{? 1992 ? Tauryuan 100}

M. CY-1 Aug. 1993 Dongshyr, Chiayi 100

M. TN-1 Aug. 1992 Dahliao, Tainan ÿb

M. TN-2 Jul. 1993 Duujia, Tainan 25

M. TN-3 Jul. 1993 Duujia, Tainan 100

M. TN-4 Sep. 1994 Duujia, Tainan 25

M. KS-1 Aug. 1989 Cherngching Lake ÿb

Coelosphaerium kuetzingianum

C. TN-1 Aug. 1992 Dahliao, Tainan ÿb

a_{Strain was a gift from Dr. J. T. Wu, Institute of Botany, Academia Sinica, R.O.C.} b_{Classi®ed as nontoxic due to the threshold dose higher than 1.6 g dry bacteria/kg mouse.}

Structure identi®cation

UV spectra of isolated microcystins were measured on a U-2000 Spectrophotometer (Hitachi, Japan). Molecular weight of each toxin was determined by FABMS on a Jeol SX102A Spectrometer (Jeol, Japan) using glycerol (m/z = 92) as the matrix. Structures of the isolated toxins were elucidated by comparing their

1_{H nuclear magnetic resonance (NMR) spectra with that of authentic MCYST-LR (Sigma, U.S.A.).}

Identi®cation of MCYST-LR was also supported by comparing its chromatographic characteristics with those of authentic. All the NMR spectra of MCYST analogues were measured on a Bruker AM-400 FT-NMR Spectrometer (Bruker, Germany), using the solvent MeOH-d4(peaked at d4.9) for chemical shift calibration.

Amino acid analysis

Each toxin component was hydrolyzed by gas-phase hydrolysis in 500 ml of 7 M HCl and 10% tri¯uoroace-tic acid containing 0.1% phenol at 1588C for 30 min (Chang and Liu, 1988). Released amino acids were then reacted with 4-dimethylaminoazobenzene-4'-sulfonyl chloride (DABS-Cl) to form color derivatives prior to the HPLC analysis. HPLC separation used an Alltima C18 Column (4.6  250 mm, Alltech, U.S.A.) with a gradi-ent elution of 15% to 40% of acetonitrile in 35 mM sodium acetate solution containing 4% N,N-dimethylfor-mamide for the ®rst 20 min, then a gradient of 40% to 70% of acetonitrile for the following 12 min and an isocratic wash of 70% acetonitrile for the last 2 min, at a ¯ow rate of 1 ml/min for the entire elution period. DABS-derivatives of amino acids were detected by 436 nm absorption.

RESULTS

Eight M. aeruginosa strains and one C. kuetzingianum strain were tested for their

mouse toxicities by injecting intraperitoneally various amounts of their MeOH extracts.

The toxicity threshold dosages were de®ned here as the minimal dose of the MeOH

extract of dried cells to kill all the triplicates of injected mice within four hours. The

threshold doses of each strain of cyanobacteria were presented in Table 1 to show the

relative toxicity among these cultured strains. Among the cyanobacteria, M. TY-1 was

the most toxic because a dose level (6.25 mg dry cell mass/kg mouse) lower than its

threshold dosage (25 mg) showed lethal toxicity to two thirds of the tested mice. M.

TN-2 and M. TN-4 strains of M. aeruginosa having the same threshold toxicity as M.

TY-1 but without any toxicity at a lower dose level were ranked the second. M. TN-1

and M. KS-1 of M. aeruginosa and C. TN-1 of C. kuetzingianum were classi®ed as

non-toxic because none of the injected mice died even at a dose level of 1.6 g dry cell mass/

kg mouse, the highest dose level in this experiment. Six of eight M. aeruginosa strains

were found to be toxic but variable in their toxicity. All dead mice from injection of

Microcystis extracts showed swollen blood-engorged livers which comprised about 10%

of the body weight (compared to 6% found in the control mice). In addition to the

cul-tured strains of M. aeruginosa, cells of the same species from a natural bloom in an eel

pond were also extracted for mouse toxicity assay. All ®ve dose levels except the extract

of 1 mg dried bacteria killed mice. The toxicity threshold dose for this natural

popu-lation of Microcystis was estimated to range between 50±250 mg dry cell mass/kg mouse

and thus less toxic than the cultured strains. It was also noted that at least 40 min, a

minimal reaction time, was required to observe the death response of mouse after an

intraperitoneal (i.p.) administration of a lethal dose of the cyanobacterial extract.

Toxin pro®les of cultured Microcystis and Coelosphaerium strains were studied by a

reversed phase HPLC. HPLC analysis showed that all six toxic Microcystis strains

con-tain MCYST-LR and MCYST-RR. Identi®cation of MCYST-LR was con®rmed by

comparing its NMR spectrum with that of an authentic standard and by co-injecting

the authentic sample with the cyanobacterial extract in the HPLC analysis.

MCYST-RR was also identi®ed by comparing its NMR spectrum with that of MCYST-LR.

Content of MCYST-LR and MCYST-RR in cultured strains of Microcystis and

Coelosphaerium are listed in Table 2. All the toxic strains studied in this experiment

contained both MCYST-LR and -RR, although they varied in their relative contents.

Generally the toxic strains contained about 1.5±2.9 times more MCYST-RR than

MCYST-LR, except M. TY-1 strain which contained mostly MCYST-LR at higher

than 10 mg per mg dry cell mass, much higher than the average 0.54 mg in the rest of the

toxic Microcystis spp. studied.

Microcystins-LR, -RR, and -RA, reported in this paper co-eluted in a very close

frac-tion by the gel permeafrac-tion chromatographic separafrac-tion. They were the major

microcys-tins in M. TN-2 cells and appeared in the fractions shown by the ®rst and second peaks

if the separation was monitored with UV

absorption. R

values of MCYST-LR,

MCYST-RR, and MCYST-RA shown on the silica gel plate of TLC separation were

0.28, 0.16, and 0.32 respectively. A TLC analytical system using silica gel 60 plates with

ethylacetate: isopropanol: H

O (4:2.5:2, v/v/v) was always used as a performance check

for preparative separations. Further separation by the silica gel ¯ash column

chroma-tography yielded two toxin fractions, one contained MCYST-LR and MCYST-RA and

another contained MCYST-RR. From these two fractions toxin components were

further puri®ed by a reversed phase high performance liquid chromatography. From

24 g dried cells of M. TN-2, 2.0 mg of MCYST-LR, 0.8 mg of MCYST-RR, and 1.4 mg

of MCYST-RA were obtained. The recovery of MCYST-RR was apparently much

lower than that of MCYST-LR during preparative separation of microcystins. Toxin

analysis of the natural Microcystis population showed less MCYST-LR and -RR but

with several other di€erent microcystins and their derivatives in trace amounts. These

microcystins are currently under investigation.

All three toxin components were collected until a sucient amount was obtained for

NMR 1D and 2D spectroscopic studies. Data of the

_{H-NMR and their assignments of}

these three toxins were compared with those of MCYST-LR given by Namikoshi et al.

(1990). It was found that proton chemical shifts of MCYST-LR isolated in this research

coincided very well with the data given by Namikoshi et al. except some irresolvable

coupling constants. Two dimensional NMR of

_H±

_{H homoCOSY gave further}

evi-dences showing the correlation of protons within each amino acid of these cyclic

hepta-peptides. FABMS spectrometry provided its molecular ions (M + H)

_{at 995 for}

MCYST-LR, 1038 for MCYST-RR, and 953 for MCYST-RA that also matched to the

postulated toxin compositions which were further con®rmed by the amino acid analysis.

Amino acid composition of these three toxins was analyzed by an HPLC separation on

the DABS-derivatives of their acid hydrolysate.

-iso-Glu,

-Ala, and

-Arg of the

tox-Table 2. Contents of microcystins, MCYST-LR and MCYST-RR in the toxic strains of Microcystis aeruginosa. Figures of the content were calcu-lated from the linear relationship (y = 500.55x ÿ 4845.40, r2_{=0.9952 for}

MCYST-LR; y = 828.01x ÿ 23945.01, r2_{=0.9951 for MCYST-RR) of the}

peak area (y) and the injected authentics (x, in ng) obtained by an HPLC analysis using a Cosmosil 5C18-AR (4.6  250 mm) column and MeOH:

0.05% aq. TFA (52:48) as mobile phase, ¯ow rate 1 ml/min microcystin-LR microcystin-RR

Strain (mg/mg dry bacteria)

M. TY-1 10.06 0.08 M. TY-2 0.54 1.58 M. CY-1 0.11 0.18 M. TN-2 0.33 0.72 M. TN-3 0.23 0.63 M. TN-4 1.49 2.21

ins gave separated peaks which were identi®ed by the authentic amino acid individually

(Fig. 2). Amino acid isomers can not be distinguished by the method reported here.

-Me-iso-Asp derivatives from the toxins gave two separated peaks of Asp and MeAsp

adducts with their peak areas in a ratio about 2:1 (Fig. 2). A common peak of

methyl-amine derivative shown in all three toxin chromatograms (Fig. 2) was generated by

Mdha of microcystins. Peak derived from

-Arg of MCYST-RR showed peak areas in

twofold of that of MCYST-LR and MCYST-RA. MCYST-RA also gave a higher

ala-nine peak than MCYST-LR or MCYST-RR in the amino acid separation. An extra

peak which was only observed in MCYST-LR chromatogram was identi®ed as

DABS-derivative of

-leucine by comparing the retention time of the authentic (Fig. 2).

MCYST-RA, the least polar component among the toxins identi®ed in this research,

was postulated a new toxin, even its molecular weight was found to be the same as the

reported MCYST-AR (Namikoshi et al., 1992). The location of

-arginine (at the 2nd

position) and

-alanine (at the 4th position) in the cyclopeptidic MCYST-RA was

deter-mined by comparing their chemical shifts on NMR spectrum with those of MCYST-LR

and MCYST-RR. The same proton chemical shifts of [Arg]

of MCYST-LR and

MCYST-RR were not found, instead the proton signals of [Arg]

as assigned for

MCYST-RR were observed on MCYST-RA spectrum.

Fig. 2. HPLC analysis of amino acid composition of microcystin (MCYST)-LR, -RR, and -RA using a column of Alltima C18 (4.6  250 mm) eluted with a gradient solution of 15% to 70% of acetonitrile in 35 mM sodium acetate solution containing 4% N,N-dimethylformamide. Samples

DISCUSSION

While calculating the data in Table 1, the toxicity threshold dose, and Table 2, the

contents of MCYST-LR and -RR in cyanobacteria, we came to an assumption that the

speci®c toxicity of MCYST-LR and MCYST-RR in a form of

(per kg of mouse)

ranges, respectively, 252±8 mg and 221±2 mg. Comparing with the

(50 mg/kg mouse

for MCYST-LR and 600 mg/kg mouse for MCYST-RR) given by Rinehart et al. (1994),

our result suggested that there were other toxins beside the identi®ed MCYSTLR and

-RR in the toxic Microcystis strains. From one of our toxic strains, M. TN-2, cultured

cells gave a bundle of microcystin analogues in the preliminary screen and their

deriva-tives to be resolved. In this report we demonstrated a new toxin, MCYST-RA, in

ad-dition to the popular ones, MCYST-LR and -RR from M. TN-2. However there are

still many other microcystins in M. TN-2 that remain to be determined. MCYST-LR

and MCYST-RR were not only in M. TN-2 but also in all the toxic Microcystis strains

in this study. This result follows the conclusion made by Rinehart et al. (1994) that

MCYST-LR and -RR were the most common components among the known

microcys-tins in the studied Microcystis strains.

Mouse assay of the toxic Microcystis strain showed the toxicity close to that

(

=31 mg/kg mouse by i.p. injection) obtained by Azevedo et al. (1994). Our

mouse assay also gave a result to show that a minimum of 40 min was needed to

ob-serve the lethal e€ect on mouse after an i.p. injection of cyanobacterial extract at a dose

higher than the threshold dosage. It has been reported that the time required for

maxi-mal toxin accumulation in liver varied from 1 min (Brooks and Codd, 1987) to 60 min

(Robinson et al., 1989) depending on the amount of MCYST-LR administered.

Histological data of Hermansky et al. (1993) also showed that the most striking

ultra-structural changes within the hepatic parenchyma appeared at 45 min after the i.p.

injec-tion of MCYST-LR. These studies demonstrated that the liver damage caused by toxin

accumulation in hepatocytes is a time-consuming process. Moreover, the transportation

of injected microcystins through hepatocyte membrane may require longer time. A

newly-found inhibition activity of microcystins on the protein phosphatase 1 and 2A

(Eriksson et al., 1990); Nishiwaki-Matsushima et al., 1992) was an assumed mechanism

by which microcystins exert their hepatotoxicity. Due to the enzyme inhibitions,

hepato-cytes shrink and cause liver damage, followed by internal hemorrhage (Runnegar and

Falconer, 1986). The death of a 20 g mouse due to the hemorrhage shock takes more

than 40 min in this case. It is expected that the time needed to show the death of

exper-iment animals will vary with the di€erent species, and di€erent weights and ages of the

same species. There are also other toxicological mechanisms on di€erent group of

ani-mals. Recently Bury et al. (1996) and Zambrano and Canelo (1996) reported that

micro-cystins exert their lethal toxicity on fresh water ®sh by disrupting the ion homeostasis

of the internal medium via blockage of gill function.

According to Carmichael (1994), there are 47 microcystins reported to have a

com-mon feature of a cyclic heptapeptide structure (Fig. 1). Acom-mong these amino acid

resi-dues, the 2nd and 4th

-form amino acids are usually substituted with di€erent amino

acids to form the microcystin variants. Rinehart et al. (1994) have summarized all these

analogues and made a list showing the structural di€erences. It was found that only

two microcystins contained

-alanine at the 4th position while almost 80% of the

known microcystins had

-arginine at this position. The postulated MCYST-RA in this

paper having an alanine at the 4th amino acid position is a new and distinctive

com-pound.

Blooms of Microcystis in the lakes and reservoirs have been known to cause

poison-ing for livestock and wildlife (Gorham and Carmichael, 1988). Recent studies showed

that toxins of Microcystis may also cause the hepatotoxicosis, gastroenteritis, and

aller-gic reactions in human (Turner et al., 1990; Soong et al., 1992) and the adverse e€ects

on aquatic lives (Codd, 1984; Philips et al., 1985). Microcystins and their toxicology

have attracted the worldwide attentions for more than twenty years. In Taiwan

Microcystis blooms are very common in aquaculture ponds of both fresh and brackish

water. Some blooms may contain microcystins and remain blooming for the year

round. However, so far we have never observed any serious adverse e€ects of

Microcystis bloom on aquatic lives or aquaculture product consumers. Further toxin

research and ecological or toxicological studies are still needed to better understand

chronic e€ects of microcystins on human health and the formation and decay of

micro-cystins in the environment.

AcknowledgementsÐPart of this work was grant funded by National Science Council (Grant No.: NSC86-2113-M002-014), Taiwan, Republic of China. Portions of this manuscript were preliminarily presented during the 7th International Conference on Toxic Phytoplankton, Sendai, Japan, 1995.

REFERENCES

Aune, T. and Berg, K. (1986) Use of freshly prepared rat hepatocytes to study toxicity of blooms of the blue± green algae Microcystis aeruginosa and Oscillatoria agardhii. J. Toxicol. Environ. Health 19, 325±336. Azevedo, S. M. F. O., Evans, W. R., Carmichael, W. W. and Namikoshi, M. (1994) First report of

microcys-tins from a Brazilian isolate of the cyanobacterium Microcystis aeruginosa. J. Appl. Phycol. 6, 261±265. Brooks, W. P. and Codd, G. A. (1987) Distribution of Microcystis aeruginosa peptide toxin and interactions

with hepatic microsomes in mice. Pharmacol. Toxicol. 60, 187±191.

Bury, N. R., Flik, G., Eddy, F. B. and Codd, G. A. (1996) The e€ects of cyanobacteria and the cyanobacterial toxin microcystin-LR on Ca2+_{transport and Na}+_/K+_{-ATPase in tilapia gills. J. Exp. Biol. 199, 1319±1326.}

Carmichael, W. W. (1992) Cyanobacteria secondary metabolites Ð the cyanotoxins. J. Appl. Bact. 72, 445± 459.

Carmichael, W. W. (1994) The Toxins of cyanobacteria. Sci. Am. 270, 64±72.

Chang, C. S. and Liu, C. S. (1988) A picomole-level amino acid analysis by means of gas-phase hydrolysis and DABS-Cl/HPLC method. J. Chinese Biochem. Soc. 17, 12±19.

Codd, G. A. (1984) Toxins of freshwater cyanobacteria. Microbiol. Sci. 1, 48±52.

Eriksson, J. E., Toivola, D., Meriluoto, J. A. O., Karaki, H., Han, Y.-G. and Hartshorne, D. (1990) Hepatocyte deformation induced by cyanobacterial toxins re¯ects inhibition of protein phosphatases. Biochem. Biophys. Res. Commun. 173, 1347±1353.

Falconer, I. R., Beresford, A. M. and Runnegar, M. T. C. (1983) Evidence of liver damage by toxin from a bloom of the blue±green alga, Microcystis aeruginosa. Med. J. Aust. 1, 511±514.

Gorham, P. R. and Carmichael, W. W. (1988) Hazards of freshwater blue±green algae (cyanobacteria). In Algae and Human A€airs, eds. C. A. Lembi, J. R. Waaland, pp. 403±431. Cambridge University Press, Cambridge.

Hermansky, S. J., Markin, R. S., Flower, E. H. and Stohs, S. J. (1993) Hepatic ultrastructural changes induced by the toxin microcystin-LR (MCLR) in mice. J. Environ. Pathol. Toxicol. and Oncol. 12, 101±106. Hughes, E. O., Gorham, P. R. and Zehnder, A. (1958) Toxicity of a unialgal culture of Microcystis aeruginosa.

Can. J. Microbiol. 4, 225±236.

Matsushima, R. N., Ohta, T., Nishiwaki, S., Suganuma, M., Kohyama, K., Ishikawa, T., Carmichael, W. W. and Fujiki, H. (1992) Liver tumor promotion by the cyanobacterial cyclic peptide toxin microcystin-LR. J. Cancer Res. Clin. Oncol. 116, 609±614.

Mizuno, T. (1980) Illustrations of the Freshwater Plankton of Japan. Hoikusha Publishing, Osaka.

Moriwaka, M. and Chyi, J. (1996) Plankton Algae of Reservoirs in Taiwan. Natl. Inst. Environ. Anal., EPA, Taipei, R.O.C.

Namikoshi, M., Rinehart, K. L., Sakai, R., Sivonen, K. and Carmichael, W. W. (1990) Structures of three new cyclic heptapeptide hepatotoxins produced by the cyanobacterium (blue±green alga) Nostoc SP. strain 152. J. Org. Chem. 55, 6135±6139.

Namikoshi, M., Rinehart, K. L., Sakai, R., Stotts, R., Dahlem, A. M., Beasley, V. R., Carmichael, W. W. and Evans, W. R. (1992) Identi®cation of 12 hepatotoxins from a Homer Lake bloom of the cyanobacteria Microcystis aeruginosa, Microcystis viridis, and Microcystis wesenbergii: nine new microcystins. J. Org. Chem. 57, 866±872.

Nishiwaki-Matsushima, R., Ohta, T., Nishiwaki, S., Suganuma, M., Kohyama, K., Ishikawa, T., Carmichael, W. W. and Fujiki, H. (1992) Liver tumor promotion by the cyanobacterial cyclic peptide toxin microcystin-LR. J. Cancer Res. Clin. Oncol. 118, 420±424.

Philips, M. J., Roberts, R. J. and Stewart, J. A. (1985) The toxicity of the cyanobacterium Microcystis aerugi-nosa to rainbow trout, Salmo gairneri Richardson. J. Fish Dis. 8, 339±343.

Rinehart, K. L., Namikoshi, M. and Choi, B. W. (1994) Structure and biosynthesis of toxins from blue±green algae (cyanobacteria). J. Appl. Phycol. 6, 159±176.

Robinson, N. A., Miura, G. A., Matson, C. F., Dinterman, R. E. and Pace, J. G. (1989) Characterization of chemically tritiated microcystin-LR and its distribution in mice. Toxicon 27, 1035±1042.

Runnegar, M. T. C. and Falconer, I. R. (1986) E€ect of toxin from tha cyanobacterium Microcystis aerugi-nosa on ultrastructural morphology and actin polymerization in isolated hepatocytes. Toxicon 24, 109±115. Saito, S., Nakano, Y., Kushida, K., Shiral, M., Harada, K.-I. and Nakano, M. (1994) Cross-reactivity and

neutralizing ability of monoclonal antibodies against microcystins. Microbiol. Immunol. 38, 389±392. Soong, F. S., Maynard, E., Hollinrake, K. and Luke, C. (1992) Illness associated with blue±green algae. Med.

J. Aust. 156, 67.

Turner, P. C., Gammie, A. J., Hollinarake, K. and Codd, G. A. (1990) Pneumonia associated with contact with cyanobacteria. Br. Med. J. 300, 1440±1441.

Zambrano, F. and Canelo, E. (1996) E€ects of microcystin-LR on the partial reactions of the Na+_±K+_pump