CONFUSION exists about the chemical and physico-chemical characteristics of enterotoxins produced in culture supernates by strains of Escherichia coli pathogenic for pigs and reported upon by Smith and Halls (1967), Kohler (1968), Larivi6re (1971), andBywater (1972). We considered that some of the apparent anomalies might be attributed to the presence of peptides and polypeptides in the media used for bacterial growth, namely " syncase " medium (Finkelstein, Norris and Dutta, 1964;Kohler, 1968) and semisolid nutrient agar (Smith and Halls, 1967). This paper describes the development of a defined medium based on mineral salts, glucose and vitamins, and compares the characteristics of the enterotoxic supernates from growth in this medium with those yielded by growth in " syncase " medium.
MATERIALS AND METHODS
Strains of Escherichia coli.Strains P6 (serotype O?; K?; H16), P99 (0141; K85; K88a, b; H4) and P16 (09; K9) were obtained from Dr H. W. Smith (Houghton Poultry Research Station, Houghton, Huntingdonshire) and strain K12 from the National Collection of Industrial Bacteria (no. NCIB9482). Strains P99 and P16 produce enterotoxin and cause diarrhoea in pigs (Smith and Halls, 1967), whereas the porcine strain P6 and the laboratory strain K12 do not.Media. A defined medium, which will be described under Results, and various modifications of it, were used in studies of the optimum conditions for enterotoxin production. For comparison cultures were grown in " syncase " medium (Finkelstein et al., 1964;Kohler, 1968) and tested similarly. The composition of " syncase " medium (percentage w/v) was Na2HP04 0.5, K2HP04 0.5, sucrose 0.5, NH4Cl 0.1 18, Na2SO4 0.0089, MgCL 6H2O 0-0042, MnCl2 4H20 O.OOO4, FeC12 -6H2O 0.0005, acid-hydrolysed casein 1 -0.
Assay of enterotoxin.The activity of supernates was measured in ligated segments (loops) of porcine intestine by the method of Smith and Gyles (1970), derived from those of De and Chatterjee (1953) and Moon, Sorensen and Sautter (1966). The loop volume (ml of fluid per cm length) for the test sample was divided by the loop volume for a "standard toxin" prepared by %hour growth of strain P16 in " syncase " medium. Thus a comparison could be made of samples in different pigs. Samples in inactive regions of intestine were disregarded, but even so the results in this assay varied by as much as &25% for the same sample. All samples were tested once in at least two and usually four different pigs. Pigs were 21 days old when used and had been weaned on the previous day and starved for 24 hours.Gel filtration. Supernates of cultures in " syncase " medium and in the defined medium that had similar enterotoxic activities in ligated intestine were freezedried and reconstituted in distilled water to give 15-fold concentration. These were then passed through Sephadex-G50 columns with 0.85 % NaCl eluant. Sephadex-G5O was obtained from Pharmacia Fine Chemicals, Uppsala, Sweden, and the results were expressed in KAV units calculated as follows.Ve-Vo KAV= ~
vtv owhere Ve=elution volume, Vt=be...