Salmonella flagellar ifiaments are polymers of a highly antigenic protein, termed flageilin. Eight main subfactors have been identified in the Salmonella phase-i g ... series flagellar antigen. To determine the molecular basis for expression of the epitopes by which the g ... family subfactors are distinguished, 10 members of this series were selected and their fliC (the structural gene for phase-i flagellin) genes were sequenced. Comparative analyses of the inferred primary structures of these flagellins did not allow the identification of linear epitopes responsible for the antigen subfactors. This suggests that conformational aspects are involved in determining the antigenic specificity in these cases. A phylogenetic analysis of the flagellin sequences showed that members of the g.. . series do not form a single coherent unit. (13) were located within region IV. These properties led others (27,28,43) to suggest that flagella might be useful in vaccine development if part of the hypervariable region could be substituted with a medically important epitope-specifying oligonucleotide, in the correct orientation and reading frame, resulting in exposure of the epitope at the surface of the flagella. To date, such substitutions have been limited to a unique restriction site in the structural gene for the flagellin antigen i of S. typhimurium (27) and to a 48-bp segment in regi6n IV (41) of the structural gene for the flagellar antigen d of S. muenchen (2,20,24,28,29,38,39,43), which is readily excised by restriction enzyme EcoRV. These attempts have been only partially successful. The primary problem seems to be that substitutions in flagellin result in an inability of the molecules * Corresponding author.to polymerize into flagellar filaments. It occurred to one of us (T.M.J.) that an examination of other flagellins may reveal more suitable molecules for interchange with known epitopes. The d and i flagellins investigated to date are both defined (4, 15, 17) by their production of a single major flagellar antigen. Other flagellar antigens have been divided into subfactors (4,15,17), and it would seem a priori that the existence of such natural variation might reveal a site at which manipulation could be attempted with less effect on the properties of the molecule. In the Kauffmann-White scheme (15), eight major subfactors (f,g, m,p, q, s, t, and u) of the Salmonella phase-i g antigen were described, and it has been shown that the factor g itself is a complex composed of two or more of at least five subfactors, gl togS (44). On the basis of amino acid composition, McDonough (22) determined that members of the g.. . series (selected serovars with flagellar antigens g,m; gp; and g,s,t) showed greater variability than single-factor flagellins. For this reason, we selected flagellins from the g... series of Salmonella flagellar antigens in the expectation that the subfactors that separate the different serovars would be located in a common region that would make an efficient site for directed substitutions. Here we rep...