Isolation and serological analysis of mutant forms of flagellar antigen i of Salmonella typhimurium.

Nine spontaneous mutants with altered forms of flagellar antigen i were obtained by picking more rapidly spreading swarms from growth in semi-solid medium containing enough anti-i serum to retard spreading growth. One mutant was in a line of Salmonella typhi given antigen i by transduction, the rest in S . typhimurium strain L T ~ adeC-7 proA-46. Two mutants of independent origin were serologically identical and presumably arose by a recurrence of the same mutation. Bacteria expressing the mutant phase-1 antigen were normally motile and the LT2 mutants showed normal phase-variation, to give cultures with an apparently unaltered phase-2 antigen, 1,2,3. Flagellate bacteria with flagella of two of the mutant types, i M 6 and iM9, were agglutinated to titres 8-16 by sera from uninoculated rabbits and to titres 50-100 by sera from rabbits immunized with unrelated antigens ; suspensions of flagella of these types, but not of others, caused flocculation of indian ink. The residual activity of anti-i (wild type) sera fully absorbed with mutant antigens showed that each mutant antigen had lost some of the serological specificity of the wild-type antigen; the complex pattern of residual activity on mutant and wild-type antigens of anti-i (wild type) sera absorbed with pairs of mutant antigens indicated the existence of a t least 13 antigenic factors in the wild-type antigen, and that each of the serologically distinct mutant antigens lacked a different combination of these factors. The residual activities on the homologous antigens of antimutant sera fully absorbed with wild-type antigen showed that all the mutant antigens, except perhaps iM 6 and iM 9, had antigenic specificities absent from the wild-type antigen. Each of the 8 serologically different antigens had a unique new specificity, but antigens iM7, iMl0 and iM12 shared some new factors. Attempts to infer the linear order of the presumed sites of amino acid substitution in the polypeptide chain of flagellin from the serological data were unsuccessful ; this probably indicates the incorrectness of an assumption involved : namely, that anti-flagellar antibodies have an absolute affinity for, and only for, all of the amino acid side-chains (or all of a reactive subset of them) in a relevant length of polypeptide chain.