Toxic peptides from freshwater cyanobacteria (blue-green algae). I. Isolation, purification and characterization of peptides from Microcystis aeruginosa and Anabaena flos-aquae

Toxic peptides from freshwater cyanobacteria (blue-green algae). I. Isolation, purification and characterization of peptides from Microcystis aeruginosa and Anabaena flos-aquae

Toadrou, Vd. 24, No . 9, pp . s65-a73, 19a6. Printed in Urnt $tltain. OMI-OI01/86 ß .00+ .00 Papoion louma4 l .rd. TOXIC PEPTIDES FROM FRESHWATER CY...

460KB Sizes 0 Downloads 37 Views

Toadrou, Vd. 24, No . 9, pp . s65-a73, 19a6. Printed in Urnt $tltain.

OMI-OI01/86 ß .00+ .00 Papoion louma4 l .rd.

TOXIC PEPTIDES FROM FRESHWATER CYANOBACTERIA (BLUE-GREEN ALGAE). I. ISOLATION, PURIFICATION AND CHARACTERIZATION OF PEPTIDES FROM MICROCYSTIS AERUGINOSA AND ANABAENA FLOSAQUAE T. ICRISHNAMURTHY,'* W. W. CARMICHAEL and E. W. SARVER' 'Chemical Research and Development Center, Aberdeen Proving Ground, MD 21010, and 'Department of Biological Sciences, Wright State University, Dayton, OH 43433, U .S .A . (Accepted for publkation

4

March

1986)

T . ICarstnew~tuar>"tY, W . W . C~aA+rcttwei and E. W . S~AVeR. Toxic peptides from freshwater cyanobacteria (blue-grein algae) . I . Isolation, purification and characterization of peptides from Microcystis arrugina~a and Anabaena Jlos-aqua. Tazirnn 24, 863 - 873, 1986 .-Toxic peptides from two European Mkrocystis amrainaao and one raoa ' " ~ Anabaena Jlas-0quae species of freshwater cyanobacteria (blue-great algae) were purified by high performance liquid chromatography (HPLC) and examined by amino acid analysis and mesa spectrometry . A toxic fraction from a butanol/methanol extract of toxic lyophilized cells was separated by G-23 gel filtration and purified by HPLC using a C-18 semi-preparative Column . A toxic peak with the same elution time was detected for each of the three toxic cyanobacteria. The desalted purified toxins (i.p. t.n, e in mice, 30 pg/kg) caused signs of poisoning identical with previous literature reports of hepatotoxic peptides from Micraystis . On hydroly:;s and amino add analysis all three toxins showed a similar profile, consisting of equimolar amounts of glutamic add, alanine, arginine and leucine. ß-methyl aspartic acid was identified in all of the toxic peptides . The feat atom bombardment mans spectra of the toxins indicated the molecular weight to be 994 for all the peptides . The absence of sequence ions in their wtTesponding fast atom bombardment mass spectra indicated the peptides to be cyclic .

INTRODUCTION HEAVY blooms of blue-green algae/cyanobacteria, are observed regularly in eutrophic natural and man-made water bodies during the summer and fall months . Several of these more common bloom forming species produce toxins (CARMICHAEL, 1981 ; CARMICHAEL and MAHMOOD, 1984; CARMICHAEL et aL, 1985). Consumption of toxin-contaminated water and bloom mass have been implicated in the loss of livestock and wild animals in several countries throughout the world (CARMICHAEL et al., 1985 ; SKULBERG et al., 1984; BEASLEY et al., 1983), as well as human intoxication (BILLINGS, 1982; FALCONER et al., 1983). Increased pollution in urban, recreational and agricultural water sources seems to contribute to the growth of toxic and non-toxic blooms (SKULBERG et al., 1984). The toxins and the algal blooms in municipal and recreational water supplies create human health hazards and increase the costs for water treatment (HOFFMANN, 1976). Of the suspected toxic cyanobacteria genera, blooms of Microcystis aeruginosa continue to be most commonly reported (CARMICHAEL, 1981) . rI'o whom correspondence should be addressed . 863

866

T . IütISHNAMURTHY et al.

The toxins isolated from M. aeruginosa produce distinct hemmorrhagic necrosis of the liver in test animals (CARMICHAEL et al., 1985). These hepatotoxins have been detected from water blooms in Australia, Canada, Great Britain, Japan, Norway, South Africa and the U.S .S.R . (CARMICHAEL et al., 1985). They have been reported to be small peptides with molecular weights in the range of 1000 (CARMICHAEL et al., 1985). The available structural information on these toxins is very limited. Hence, investigations have been initiated recently in several countries to elucidate their chemical structure, as well as to study their chemical and physical properties (CARMICHAEL et al., 1985 ; SKULBERG et al., 1984). This information is essential to develop methods of detection and decontamination . The only known unambiguous structural elucidation of the hepatotoxin from M. aeruginosa is reported by Botes et al., using South African toxic strains (BOTES et al., 1982x, 1982b, 1984 ; SANTIKARN et al., 1983). They have also described the isolation, purification and characterization of four peptides from a single laboratory culture isolated from a water bloom. The amino acid analysis of these peptides indicated that each contained three common and a unique pair of amino acids . The common ones were alanine, glutamic acid and ß-methylaspartic acid. The pairs were combinations of either leucine and arginine or tyrosine and arginine or leucine and alanine or tyrosine and alanine (BOTES et al., 1982b) . The configurations of the asymmetric carbons in the common acids were in the d-form and the pairs in the 1-forms (BOTES et al., 1982x). Based on fast atom bombardment mass spectroscopy, $ANTIKARN et al. (1983) proposed a cyclic structure for a South African variant toxin, BE-4 (mol . wt 909) . This reaffirmed the original observation of BISHOF et al. (1959), who proposed that the toxin of M. aeruginosa NRC-1 was a cyclic polypeptide. The structure of BE-4 was unambiguously defined based on nuclear magnetic resonance, mass spectral data of the intact peptide and its chemical degradation products . We have recently initiated investigations in order to elucidate the total structures of several cyanobacterial toxic peptides . It is felt that these studies will lead to methods of detection and analysis in environmental samples. The hepatotoxins from laboratory cultures of Microcystis aeraginosa strain 7820 (CORD, 1982) and Anabaena flos-aquae strain S-23-g-1 (CARMICHAEL 8nd GORHAM, 1978) and natural blooms of Akers Lake (Akersvatn), Norway, have been investigated . It should be noted that prior to this study A. ,~los-aquae was not observed to produce hcpatotoxic peptides . MATERIALS AND METHODS Toxin purj~rcation and taaidty testing Strain 7820 wax isolated by d . A . Cadd from a toxic water bloom near Dundee, Scotland, in 1976 (Cone end C~ttctust., 1982) . Strain S-23-g-1 was isolated from a lake near Saskattoon, Saskatchewan, Canada, in 1975 (C~ttctuea, and GoattMt, 1978) . The water bloom material was collated from Akeravatn (Akera Lake) near Odo, Norway, in August 1984 b~A, M . Skulberg of the Norwegian Water Raouraa Institute. Toxin from all three sources wax isolated from lyophilized alla aaording to a proadurc modified from that of S~$aet.sttiv et al. (1984) sad shown in Fig . 1 . ICR Swig male mix (18-24 g) were used to monitor all stages of the purification prooedurc . Toxin extracts and aW were irljated by the i .p. route. Signs of poisoning were monitored and characteristic liver hemorrhage, induding the ratio of the liver weight to body weight, wen noted . Amino acid analysis In the first method, peptides (30-100 nmolea) were digested with 6 NHCI oontainlng O.SrFa phenol and 0.39V mercaptoethmol in a sealed ampule under vacuum at 112°C for 24 hr . The rcleaaed amino adds were analyzed using a Liqui-Mat amino add analyzer . The product wax loaded on a standard Mitsubishi iontxchange resin MCI-gel CK 08F column (4 mm x 16 cm) and eluted with wdium dtrate buffer (0.18-0 .20 M) . Ninhydrin

Peptide Toxins of M. atruginosa and A . Jlas-squat

867

PROCEDURE FOR THE EXTRACTION OF HEPATOTOXIC PEPTIDES FROM FRESHWATER CYANOBACTERIA 1.

1 GM CELLS + 200 ML SX BUTANOL-70Z METHANOL-7FX WATER STIR 1-3 HR AT 4~C . CENTRIFUGE 100,000X G - 1 HR AT 4~C . REPEAT 3 TIMES WITH CELL PELLET

2.

COMBINE SUPERNATANTS REDUCE VOLUME TO 300-?~0 ML BY AIR DRYING

3.

SUPERNATANT PASSED THROUGH ANALYTICHEM BOND ELUTE C-'? COLUMN ELUTE TOXIC FRACTION WITII 3-5 ML 100X. MEOH REPEAT PROCESS 3-4 TIMES

4.

DRY COMBINED METHANOL EXTRACT WITH NITROGEN DISSOLVE RESIDUE IN F FIL WATER PASS THROUGH 3.0 MICRON MILLIPORE FILTER

5.

K-26 PHARMACIA COLUMN (26 MM X BO CM) WITH 100 GM SEPHADEX G-25 OL/WATER NgNÎTORNATx24ÔTNM TOXIN IS FIRST LARGE PEAK OFF THE COLUMN

6.

HPLC-ALTEX C-18 9.4 MM X 25 CM 0.01 AFMONIUM ACETATE IN 26t ACETONITRILE/WATER FLOW RATE 3 ML/MIN MONITOR AT 240 NM

7.

LYOPHILIZE TOXIC PEAK DESALT TOXIN BY HPLC (AS IN STEP 6'

8.

STORE TOXIN AT -80~C UNTIL USE

FIa. 1 . EXTRACTION AND PURII7CATION PROCEDURE FOR HEPATOTOXIC PEPTIDES OF Microcystls alILa~A0017 AND AlfaIaOeRa,jIOJ-QQYae.

Purified toxins have an approximate i .p. LD°° in mice of 30 ß/k8, with a survival time of 30-90 min . Each 8ram of Iyophlh'zed ceW contains between 1 and 4 m8 of toxin . Raovery of the toxin is between 70 and 731 .

(0.1 M) in DMSO was used for the port-oohm~n daivatintion in order to detect the amino acids at 410 mn and 370 nm . The dutlon times were calibrated usia8 Q-amino ß-8uanidinopropionic add hydrochloride . In the second method, Peptides (3 ~ were hydrolyzed in 6 N HCI at 106°C for 24 hr . The rdeaaed amino acids were pra~oolumn derivitized with pheayliaothiocyanate (PITC) and the P6eßylthiocarbamyl (PTC) amino acids were analysed usin8 a Water: Pioo Tab HPLC rystem . The daivadvea were loaded on a C-18 (13 cm x 4.6 mm) column sad ehlted ualn8 0.138 M in aqueous sodium acetate trihydrate/acetonitrlle mixture uain8 a 8radknt of 0-60~ aoetonitrile in 8 min . The oohuan flow rate was maintained at 1 .0 ml/min . The doted compounds were detaxed usin8 a u .v . detator at 234 nm . Masr spertrn Maw spectra were obtained from a standard FinniBan-Mat TSQ maw spectrometer with the mass range up to 1700 a .m.u. The fast atom bombardment (FAB) spectra were recorded by placin8 the sampk (30-100 nmok) in a 8lycerol or thio8lyarol matrix on the copper sample àa8e and introducin8 it into the source of 60°C. The sampk was bombarded with 8 kV krypton aroma, +~ht Ia nice the source temperature at 60°C . RESULTS AND DISCUSSION

The toxins were isolated from Mic~rocystis aeruginasa 7820, Anabaena flogs-aquae S-23g-1 and Alceravatn bloom and purified by the procedure described in Fig . 1 . The toxin extracts were separated from the plant pigments using Analytichem i3ond Elute C-18 column and G-2S gel filtration. Purification of this extract was performed by high

868

T . KRISHNAMURTHY et d.

A.

C.

B.

~o

MN

~

Yysct

to

MN

~ ~

to

MN

Fia. 2 . Htax rr~row~twxec uQutn cxieoMwTOanwrx of ruatr~eo rowtvs Faost (A) Microcystis aas~girtaaa strain 7820 (10 ~, (B) wr~asvwnv ww~e Ht.ooM (8 ~ wNn (C) Anabaa+a Jlar-aquae 5-23-x-1 (81~) " Column conditions wen : column, Altex C-18 4 .6 mm x 25 cm; solvent, O.OI M ammonium aortite in 26% acetonitrlle-water ; flow rate, 1 mVmia; wavden~th, 240 nm ; AUFS, 0.1 .

performance liquid chromatography (HPLC) using a semi-preparative reverse-phase C-18 column and eluting with 0.01 M ammonium acetate buffer in 26Wo aqueous acetonitrile. The purified extract was desalted by loading again on the semi-prep column and eluting with 26% aqueous acetonitrile. The protxss was repeated in order to ensure the total removal of salts from the toxin, which is very essential for the FAB mass spectral investigation, in addition to the accurate determination of yield of purified toxin. The purity of the toxin was further ensured by injecting an aliquot onto a C18 analytical (4.6 mm x 2S cm) I3PLC column and eluting with the same solvent. The toxins from Akersvatn (Fig. 2B) and 5-23-g-1 (Fig. 2C) showed a single sharp peak at 10 min, indicating the product to be frce of all contaminants . The chromatogram of the 7820 toxin (Fig. 2A) displayed a major peak at 10 min, followed by a minor one at 11 rain . It was concluded that all three toxins eluting at 10 min were the same, which was later proved by mass spectral data. The lethal potencies, measured as the LDa, values, of the purified toxins and the cellular material were determined . to be approximately SO pg/kg and SO mg/kg, respectively. The results from the amino acid analysis using the Liqui-Mat analyzer are listed in Table 1 . The elution times of the standard materials and their area measurements were utilized in the detection and quantification of the amino acids present in the hydrolysates of the hepatotoxins . Amino acid analysis was also performed by an alternate method

Peptide Taxins of M. aa~u8inaw aad A. JJar-aquar TAHt.E

1.

AMtNO

wciD wN~u.rsts RESUUs 7820

Akertox

Alanine Arginine Olutamic acid Leucine ~Methylaspartic acid

%9 S-23-g-1

Pico Tag (pmolea)

Liqui-Mat (nmolea)

Pico Tag (pmoles)

Liqui-Mat (nmoles)

Pico Tag (pmoles)

Liqui-Mat (nmolea)

66 (1 .0) 57 (0 .9) 92 (1 .4) 7S (1 .1)

86 (1) 80 (0 .9) 87 (1 .0) % (1 .1)

76 (1 .0) 64 (0 .8) 103 (1 .4) 77 (1 .0)

133 (1) 133 (0.9) 162 (1 .0) 160 (1 .0)

202 (1 .0) 222 (1 .1) 229 (LS) 223 (1 .1)

38 (1) 32 (0.9) 61 (1 .1) 61 (1 .1)

44 (0 .7)

40 (0.5)

109 (0.5)

Numbers in parenthesis are molar ratios based on alanine.

0

2

4

8

8

10

12

Mhulea Fta. 3.

trQtnn cfntoMwTOOawrx wNn wMtNO wcm rttoFttB of MldocyatLs aaud~ ws:ERrox (3 ~. Analysis is by Waters Pico Tag precolumn daivitiution with phenylisothiocyanate, to yield phenylthiocarbamyl amino acids. Pioomole amounts of each amino acid are given in Table 1. M. asugtnata 7820 and A. Jlo~r-agtrae 5-23-g-1 toxin gave similar pmfiles. PITC DERIV. unidentified phenylieothiocyate derivitive produced during :emple derivitiution. Ordinate units in millilivolta. HIGH PERFORMANCE

using a Waters Picotag amino acid analyzer (these results are also shown in Table 1). The 1?icotag liquid chromatogram of M. aeruginasa Akersvatn (Akertox) peptide hydrolysate is shown in Fig. 3 . Alanine, arginine, glutamic acid and leucine were identified in all hydrolysates by both methods. The molar ratio of glutamic acid with reference to alanine was 1,4 (Table 1). One of the two unidentified amino acids was found to be ß-methylaspartic acid, by comparison with the elution pattern of an authentic sample of ß-methyl-d, 1aspartic acid (Sigma). Two peaks with the ratio of 7 :1 were observed eluting at 2.2 and 2.7 min (Fig. 4A). This was probably due to the two possible diastereomers of the ß-methyl aspartic acid . The enantiomer of each of these must be co-eluting with the other. One of the diastereomers seemed to co-elute with glutamic acid, which accounts for the observed higher molar ratio of 1 .4 for giutamic acid . This was established by analyzing the spiked amino acid standard mixture (Fig. 4B, G). The toxin from S-23-g-1 was also spiked with ßmethyl aspartic standard (results not shown) . The resulting profile showed an increase both in the glutamic acid peak (2.2 min) and the peak which corresponded with the 2.7

870

T. KRIShIIVAMURTIiY d al.

A. 34.2x2 ,

ß-methyl A3P

f

18.884

2.518

Z x W

0

2

G U t a

4

T8

8

Minutes

B. 82 .386

42283 ~

~30

2.132 Minutes

FIß. 4. (A) HIGH pgRFO4uwur~a LIQUID CHROMATOORAPH OF METHYL ASPAATIC ACID (5 ~, US1N0 WA18RS PICO TAO ANALYSIS, SHOiVINO THE TWO DIASi'BREOMER$ WHICH ELU7E AT 2.2 AND 2.7 M1N . (3LU ~ podtioa where jlutamic acid Would dute if prexnt .

HIGH PSAFOßMANCE LIQUID CF~OMATOORAPH PROFILE OF P~RCE AMINO ACID STANDARDS pmoles eACHI uslNO WATERS Ploo TAO ANALYSIS.

(250

(~ HIOH rERPOaMwNCE LIQUm claeoMwroaawPH PaoFILE of Pu~lecE AMIIiO wclD srwrIDwaDS pmoles BAS uswa WATERS Ptco Two ANALYSIS

(250

(B)

Note appmldmatdy equal lined peaks for a:partic sad slutamic acids.

Sample is spiked with S ~ of ß-methyl a:partic add. Note enhanced 6lutamic acid peak (2 .2 mis) and peak at 2.7 min.

Peptide Toxins of M. aeruginosa and A. Jlas-aquae.

871

DATA : HTK4 +12 MASS SPECTRUM 12/11/8414 :18:00+0:18 CALJ : CAL103101 +1 SAMPLE : AKERTX CONDS. : +/CI - KR/01 MASS SPECTRUM/7KV 12UA

BASE M/Z: 277 RIC: 128000

277 .1

se.eu ++ .7 ,_ I i.?s.z

M/Z

1ISB

us e .7 I2ee

irlz.s

Izse

_ iR~ .s_ ._ I~.s . . _ t1hi4 .4 13â3.1 . I

13SB

IasI,7

I,éO

FIO. S. FASr ATOM 90MHARDM~Nf INA3S SPECCRUM OF M. al'YglIIO~A7 ALEATOX PEP77DH aOLYCEAOL b1ATADL

min peak of the ß-methyl aspartic acid standard. Based on these results, it was concluded that all the peptides contain alanine, arginine, glutamic acid, leucine and ß-methyl- . aspartic acid in an equimolar ratio. In addition, they all have an unidentified amino acid, the elution time of which (8 .5 min, Fig. 3) was the same in all three toxin hydrolysates . It should be notod that all these peptides were found, by two different methods, to contain the same amino acids in the same molar ratio. The fast atom bombardment mass spectra of Akertox in glycerol and thioglycerol were recorded and the molecular weight of the peptide was determined to be 994. The glycerol matrix was observod to be a better suited matrix for obtaining fast atom bombardment mass spectra, and the addition of oxalic acid to the matrix improved the total ion current of the protonated molecular ion considerably . The FAB spectrum of the peptide in glycerol matrix is shown in Fig. S. The lack of sequence information in the FAB spectra indicates the peptide to be cyclic (Tobt et al. , 1984). The FAB spectrum of 7820 toxin in glycerol matrix indicated two quasi-molecular ions m/z 995 and 1013 without any sequence information. These peptides were purified by repeated HPLC separation, and the FAB spectrum of the abundant peptide is given in Fig. 6. The products from acid hydrolysis, trypsin cleavage and the Edman degradation of the 7820 peptide were the same . Only oae compound with a molecular weight of 1012 was detected in all of the reaction mixtures . It was concluded that the rings of the cyclic peptides with a molecular

s7z

T. KRISHNAMURTHY eY al. MA83 SPECTRUM DATA: HTK3 +31 12/11/84 12 :41 :00+0:42 CAU: HCAL7127 +1 SAMPLE : 7820 CONDS: +/CI " KR/01 MA33 SPECTRUM/7KV 12UA aC TEMP: 149 DEß. C

BASE M/Z:1103 RIC: 88192.

2432 :

2432 .

2432 .

2132 .

.e i

M/Z

11

.2

1139

1173.1

1 1

.2 123 .3 1239

1

.4

!3

1

2

1

.6

t33î

FIa. 6 . FAST'ATOM BOMBARDMI~rI' MASS SPBCI'RIJM OF M. aerUglnAtIQ7$ZO PEFrIDE IN THIOOLYCEROL MATRIX .

weight of 994 opened during the process and formed the product with the addition of a water molxule (see Fig. 2A). This mixture with the same molecular weight of 1012 did not undergo any further reactions. All these observations wen further proof of the assumption that these peptides are similar to the Akertox peptide. The FAB spectrum of the 5-23-g-1 peptide indicated the molecule to have a molecular weight of 994. Absence of sequence information in this case indicated that this peptide is cyclic as well. The other common feature noted in all the peptides was the presence of a m/z 135 ion in their FAB spectra. Based on these observations, it is proposed that these peptides, which originated from different geographical areas, are cyclic, with possible similar amino acid sequences and structure. Further investigations in order to elucidate the total structure of these peptides are in progress . Aablowle~daeniuits-The authors thmlt ß. A. Cone and O. M. SruLaeao, respectively, for cliltura and lyophüixed cells of tonic Microcy~rtis. The technical asdstance, in eatracdon, purlficati~ and amino add analysis of tollin:, of J. ESCF~OR is alw ap~edated. This work was supported in part by aUnited States Army Medical ltaearch and Development Command contract and an NIH Biomedical Seed ßra~ to WWC. REFERENCES Beeal.eY, V. R., Corrocx, R. W., SIMM, J., ELY, R., Bucz, W., CORLBY, R. A., CARIBON, D. M. and ßoR1tAM, P. R. (1983) Apparent blue-Been algae polsonine in swl~ subsequent to iaearion of a bloom dominated by Arsabas9ra splroldea. !. Am, ver. road. Ans. 182, 413.

Peptide Toxins of M. aavginos~a and A . Jlos-aquae

873

Btt.Lttvas, W. H. (1982) Water associated human illness in northeast Pennsylvania and its suspected association with bhu-~xn algae blooms . In: The Water Environment : Alltal Toxins and Health, p. 243 (C~~cw~L, W., Ed .) . New York : Plenum . Btsttor, C. T., At~r, E. F. L. J. and Gont~tr+st, P. R. (1939) Isolation and ideatitication of the fast-death factor in Microcystis aauginosa NRC-1. Can. J. Biochan . Physiol. 37, 453. Holes, D. P., VtLtostv, C. C., KRt1aeA, H. Wasssts, P. L. and WtLLtwMS, D. H. (1982x) Configuration assignments of the amino acid residues and the preaeace of N-methyldehydroalanine in toxins from the bluegreen algae Microcystis xrruginasn. Taxkon ?A, 1037 . Bores, D. P., KRttaaR, H. and Vtt.toEN, C. C. (1982b) Isolation and characterisation of four toxins from the blue-green algae, Microcystis aeruglnasa. Toxirnn 20, 943. Bores, D. P., lbtrtnt~tv, A. A., WessEts, P. L., Vtuoart, C. C. and KRUOSR, H. (1984) The structure of cyanoginoain-L . A., a cyclic hepatapeptide toxin from the cyanobacterium Microcystis aauginosa. J. them. Soc. Pa*in Trans. 1, 2311 . Cwtuttctt~eL, W. W., Ed. (1981) The Wata Environment: Alga( Taxies and Health . New York : Plenum Press. CwRattctueL, W. W. and GoRta~t, P. R. (1978) Anatoxina from clones of AnabaataJlas-agerae isolated from lakes of Western Canada . Mltt. int. Varin. Limnol. 21, 283. C~ttcttweL, W. W. and Mwtutoon, N. A. (1984) Toxins from freshwater cyanobacteria . In : Sagjood Toxins, p. 377. (Reason, E., Ed .). Washington, DC: American Chemical Society. CeRMtctieeL, W. W., Josses, C. L. A., Met~tooo, N. A. and Ttmtss, W. C. (1983) Algol toxins aced waterbaaed diseases. CRC crit. Rev . awls. Control 13, 273 . CODD, ß. A. and CwRt~ttctteeL, W. W. (1982) Toxicity of a clonal isolate of the cyanobacterium Microcystis tsauglnasa from ßreat Britain. FEMS Microblol. Lett. 13, 409. Fwt.cotveR, I. R., BeR~oRO, A. M. and RurtrtaoeR, M. T. C. (1983) Evidence of fiver damage by toxin from a bloom of the bloom of the blue-green algae, Microcystis oavginaw. Med. J. Aunt. 1, 511 . Hot~wrt, J. R. H. (197v7 Removal of Mkrocystis toxins in water purification processes. Wat. S. .fir. 2, S8 . SAI~R7KARN, S., Wtt.t.teMS, D. H., $MITN, R. J., Het~nrtorm, S. J., Boss, D. P., TUt1VAMAN, A., Wt~ts, P. L., Vttaoatv, C. C. and KRUaax, H. (1983) A partial structure for the toxin BE-4 from the blue-green algae, Microcystis ~ruginosx. J. them. Soc. Common. 12, 652. $tEOELMAN, A. W., AnetNS, N. H., $TONER, R. D. and SIA77CIP1, D. W. (1984) Toxins of Microcystis amrglnasa and their hematological and histopathological effects. In : Segjood Taxies, p . 407 (Reaat .ts, E., Ed .). Washington, DC : American Chemical Society . Sr~n.sFn O, O. M., CORD, G. A. and CeR~actteeL, W. W. (1984) Toxic blue-green alga( blooms in Europe : a growing problem. Ambto 13, 244. Tot~R, B., CROW, F. W., GROSS, M. L. and KOrrt.sR, C. D. (1984) Fast atom bombardment combined with tandem mass spectrometry for the determination of cyclic peptides. Analyt. Char . 57, 880.