have been chemically characterized. Saxitoxin 

 (the toxin of G. catenella) has a molecular 

 weight of about 370 and a purine (nitrogenous) 

 base (Schuett and Rapoport, 1962; Schantz et 

 al., 1966). The toxin obtained from M. aerugi- 

 nosa is a cyclic polypeptide composed of 10 

 amino acid moieties, including D-serine 

 (Bishop, Anet, and Gorham, 1959; Gorham, 

 1960). Prymnesin, an endotoxin from P. par- 

 vum, is a glycolipid with a molecular weight of 

 about 23,000 (Paster, 1968). The lipid portion 

 is composed of four long chain acids (myristic, 

 stearic, palmitic, and oleic) and the polysac- 

 charide portion of glucose, mannose, and galac- 

 tose. The toxin (s) isolated from G. breve is thus 

 one of the few algal toxins to be chemically 

 characterized. 



MATERIALS AND METHODS 



REAGENTS 



Most reagents used in this study were 

 analytical grade, and all solvents were 

 Baker * analyzed reagents with the following 

 exceptions: Silica gel (grade 62, mesh 60-200) 

 used for chromatography was obtained from 

 Grace Davison Chemical, Baltimore, Md. ; bone 

 charcoal was B & A reagent grade; carbon 

 tetrachloride used for spectra was GC-spectro- 

 photometric reagent grade. 



SOURCE OF GYMNODINIUM BREVE 



Organisms were obtained in two ways. Axenic 

 cultures of G. breve were obtained from W. B. 

 Wilson (Wilson, 1965) and were subcultured 

 in B-5 medium (Wilson, 1966). Water samples 

 were collected during an outbreak of G. breve 

 in the fall of 1967 near Big Pass, off Sarasota, 

 Fla. 



PROCEDURE FOR ISOLATION OF TOXINS 



We isolated two substances by using the fol- 

 lowing procedure. Water from the bloom of 

 Gymnodiniicm breve or from laboratory culture 

 media (1X10*' cells/liter) was adjusted to pH 

 4.0 by addition of a small quantity of concen- 

 trated (12 M) hydrochloric acid. Two-liter por- 

 tions of the solution were then extracted in 

 Fernbach flasks with 30 ml. of chloroform. The 



^ References to trade names or suppliers in this publication do 

 not imply endorsement of commercial products. 



mixture was agitated with a mechanical stirrer; 

 the top of the flask was covered with a slotted 

 cardboard to avoid splashing. Each batch was 

 stirred vigorously for 15 minutes and allowed to 

 settle in a modified separatory funnel for 20 

 minutes. Water from the top layer was dis- 

 carded by a side tap, and the chloroform ex- 

 tract (substance II) was taken out from the 

 bottom and stored. The combined chloroform 

 extracts were associated with an upper thick 

 interfacial layer consisting of very small bub- 

 bles and containing substance I. This layer con- 

 taining the bubbles was highly toxic to fish. 

 Extracts from a particular batch of water were 

 always kept in separate containers, stored in a 

 refrigerator. 



PURIFICATION 



The two layers, i.e., the chloroform extract 

 and the interfacial layer, were treated sep- 

 arately. 



The clear chloroform layer (substance II) 

 was separated, concentrated, and then purified 

 by column chromatography. We used a separa- 

 tory funnel to separate the chloroform from the 

 interfacial layer. Generally, an extract contain- 

 ing an appreciable amount of toxin was in- 

 tensely yellow. The chloroform layer (30 to 40 

 ml.) was concentrated in a rotary evaporator to 

 one-tenth of its volume, and then passed 

 through a silica-gel column (4 cm. long, 1 cm. 

 diameter) at a rate of 18 to 20 drops per 

 minute. Yellow material, with a greenish-yellow 

 band above it, was adsorbed on top of the 

 column. The colorless chloroform effluent was 

 evaporated (rotary evaporator) , and the yellow 

 oily residue was nontoxic to fish. The column 

 was next treated with methylene chloride (four 

 10-ml. portions), and the yellowish-green band 

 was eluted, leaving the deep brown band on the 

 silica gel. The effluent was evaporated in cold, 

 and the greenish residue obtained was nontoxic 

 to fish. Finally, the brown band was eluted with 

 two 2- to 3-ml. portions of absolute ethanol. 

 The deep yellow effluent was evaporated slowly 

 in a vacuum desiccator. The residue was toxic. 



The thick yellow residue (substance II) was 

 dissolved in carbon tetrachloride and purified 

 further. The solution was stirred with bone 

 charcoal and filtered. The colorless carbon tetra- 



434 



U.S. FISH AND WILDLIFE SERVICE 



