When E. coli K235 is grown in the presence of mitomycin C it elaborates part of its colicin as a protein, which has been purified by gel filtration and electrofocusing to yield a product that is serologically homogeneous. The colicin K of E. coli K235 is chemically and serologically identical with that of Proteus mirabilis. Like the colicin of the latter, the E. coli bacteriocin occurs in multiple forms that exhibit slight differences in mobility upon electrophoresis in a polyacrylamide gel.The exact chemical nature of the colicins has been established only in the past several years, an achievement made possible by the addition of inducing agents to the culture media in which the colicinogenic microorganisms are grown (1). When we began our investigations of these potent bactericidal agents in 1953 not much was known of them. Our early work (2) revealed that both colicins K (3) and V (4), obtained from the culture medium of the respective microorganisms, were invariably associated with the 0 antigen of the bacterium in question-an observation that was subsequently corroborated by others (5). We showed, furthermore, that when type K colicinogeny was transferred from a strain of Escherichia coli to an unrelated dysentery bacillus by mating, the bacteriocin and 0 antigen of the latter were inseparable (6). Despite this, we have always been alert to the possibility that the two might well be separate entities. Indeed, we showed that when the 0 antigen-bacteriocin complex was dissociated into its components, the lipopolysaccharide and protein, it was the protein that bore the bacteriocin (3). Thus, the protein nature of these agents was at that time firmly established.The first investigator to obtain a bacteriocin from cultures of a colicinogenic micororganism by induction with ultraviolet light was Reeves (7), who isolated colicin F (now colicin E2). He showed it to be a protein, albeit his substance was still contaminated with carbohydrate. Colicin K from a strain of Proteus mirabilis was the first of the bacteriocins to be obtained in a state of purity and to be unequivocally characterized as a protein (8). Its molecular weight, aminoacid composition, and immunological properties were all thoroughly described (9). Shortly thereafter, this same bacteriocin was obtained from a strain of E. coli K12 (10), and several years later a brief description of its isolation from E. coli K235 appeared (11). More recently, the isolation of colicins A, E, Q, and K, presumably in states of purity, was described (12). Some months after the characterization of colicin K from P. mirabilis (8) In this communication we shall describe the isolation of colicin K from a culture of E. coli K235 induced with mitomycin C, and we shall show that the bacteriocin apparently exists in two forms-one a protein identical in all respects with the colicin K from P. mirabilis, the other the bacteriocin associated with the 0 antigen of the bacillus (and will be described in a later communication).
EXPERIMENTALPreparation of Colicin K. The col...