Treponema phagedenis is an anaerobic, motile spirochete with several periplasmic flagella (PFs) at each cell end. This study provides the first genetic evidence that multiple protein species are associated with the PFs. In addition, these proteins were found to reside together on a given PF. Nonmotile mutants which lacked the PFs were isolated, and spontaneous revertantsWto motility regained the PFs. These results suggest that the PFs are involved in the motility of T. phagedenis. Isolated PFs had two major protein bands with molecular weights of 33,000 and 39,800, as revealed by sodium dodecyl sulfate-polyacrylamide, gel electrophoresis. Western blots with monoclonal and polyclonal antibodies indicated that both proteins were absent in the PF mutants but present in the revertants. Immunoelectron microscopy revealed that the 39,800-molecular-weight protein was distributed along the entire PF. Immunoprecipitation analysis suggested that the 39,800-and 33,000-molecular-weight proteins were closely associated in situ.Treponema phagedenis is an anaerobic spirochete that is cultivable on enriched media (36). The cells are approximately 0.2 ,um wide and consist of a right-handed helical cell cylinder (23) surrounded by an outer membrane sheath (9, 17-21, 31, 36, 38). Inserted subterminally at each end of the cell, between the helical cell cylinder and the outer membrane sheath, are several periplasmic flagella (PFs), also referred to as axial filaments, endoflagella, and periplasmic fibrils (9, 17-20, 31, 36, 38).The PFs of various spirochetes have been analyzed in some detail. Whereas most bacterial flagella consist of a single major protein, the major proteins isolated from PFs vary from one to six (5,6,16,22,28,30,32). In these studies, the major criterion used for the identification of PF proteins was enrichment during purification. In this paper, we present the first genetic evidence that multiple protein species are indeed associated with PFs.The finding of multiple protein species associated with PFs is consistent with three possibilities. Firsi, there could be two types of PFs, with each type consisting of a different protein. Both types of PFs could be found on each cell. This situation is analogous to the morphological flagellar types found in Vibrio parahaemolyticus (33). Second, phase variation could be occurring in the cell population. Each cell could be synthesizing only one or the other PF type. Salmonella species undergo such a phase variation in their flagella (34). Third, there could be one type of PF, with each PF consisting of multiple protein species. Both Bacillus pumilis (24, 37) and Caulobacter crescentus (26, 41) have multiple protein species associated on a given flagellum. We present evidence in this paper in support of the third explanation for T. phagedenis, i.e., there is more than one protein species residing on a given PF in T. phagedenis.( ). Cells were plated on PYG-RS plus 0.5% Bacto-Agar (Difco Laboratories, Detroit, Mich.) by using overlays made with 0.5% SeaPlaque agarose (FMC C...