Colicine typing of Shigella sonnei. I. Principle, technique, selection of indicator strains, and foundations for a typing scheme.

Colicine typing of Shigella sonnei using the methods of Naito et al. and Abbott and Shannon were performed in parallel on a large number of epidemic strains, the indicator strains used in both methods, and on stock strains used to test the accuracy of the indicators. The following results were obtained after typing 462 epidemic strains: by Naito's method, 18 strains were in A1, 87 in A2, 33 in B1, 253 in B2, 38 in C1, 3 in C2, 2 in D, and 3 in E, and 25 strains were unclassifiable; while by Abbott and Shannon's method, using the present authors' simplified designation, 189 strains were in type 6/11, 124 in type 4/14, 85 in type O, and 41 in type 8, and 23 strains were either in other types or found unclassifiable. Naito's type A was in large part, equivalent to Abbott and Shannon's type O, but included some type 6/11 and type 4/14 strains. Naito's type B consisted of types 6/11, 4/14, 3A and 13, and type E consisted of 1A, IB, 3, 5, 9, and 10. Type C coincided with type 8 and type D with type 12. In addition, there were strains determined to be Abbott and Shannon's types 6/11, 4/14, 2, 7 and 9. The Abbott and Shannon's method revealed the possibilities of these strains to be classified in further detail. This is considered to be attributable not merely to the lack of indicators in the Naito's method corresponding to those of Abbott and Shannon's, but also to insufficient production of certain colicines in the Naito's method. Because of the type distribution found in Japan at present, further investigation should be done so that types 6 and 11, and types 4 and 14 can be separated.