parahaemolyticus is c. 30 nm and 15 nm for the lateral filament (McCarter, 2004). In contrast, type IV pili are much thinner and show a diameter that ranges between 50 and 80 Å (Craig et al., 2004). We also analyzed the ion preference for the rotation of both flagella. This was achieved by including amiloride in 0.3% or 0.5% soft agar plates. At 0.3% agar, motility is mediated by the polar ERK inhibitor solubility dmso flagellum and it is drastically reduced by amiloride, indicating that the polar flagellum is driven by Na+ ions. In contrast at 0.5% agar, motility in the presence of

amiloride was slightly reduced, suggesting that at this agar concentration, V. shilonii swarms using mainly lateral flagella. Hence, presumably, protons drive lateral flagella, given that swarming is insensitive to the presence of amiloride. As mentioned, the presence of lateral flagella correlates with an increase in density at an agar concentration of 0.5%; however, the alternative use of Na+ and H+ gradients for cell motility in V. shilonii is an issue that remains to be further explored. In this work, we also analyzed the subunit composition of the isolated HBB

complex of the polar flagellum of this bacterium. The internal sequences of eight flagellar proteins were obtained by MS. These correspond to three different flagellins (FlaA, FlaB and FlaC), the hook protein (FlgE), the PARP activation L-ring protein (FlgH), the MS-ring protein (FliF), a rod protein (FlgG) and the Na+-driven motor component (MotY). The genes encoding these proteins were identified in the complete genome of V. shilonii. We determined

that six of these sequences are encoded by genes located in what we have named flagellar region I. FlgG is encoded in flagellar region III and MotY is encoded by a gene in an unlinked region. The finding that the polar flagellum contains an FlgG from a different Nintedanib (BIBF 1120) flagellar locus was unexpected, given that flagellar region I also includes an flgG gene. Furthermore, the FlgG protein encoded in region I shows 95% similarity to FlgG from the polar flagellum of V. parahaemolyticus, whereas FlgG encoded in region III shows a lower similarity (66%). It remains to be elucidated whether other components of the polar flagellum could be encoded in region III. In this regard, it should be noted that flagellar region I does not include genes homologous to pomA and pomB. The motor proteins of the polar flagellum may correspond to those encoded in the flagellar region III or may be encoded by a bicistronic operon, which is unlinked to the flagellar regions described above and spans from positions 4 290 113 to 4 291 852 (see Fig. S1). According to our sequence analysis, the flagellar genes located in region II are highly similar to lateral flagellar genes that have been characterized previously in other Vibrio species. Hence, the lateral flagellum of V. shilonii would presumably be encoded by flagellar genes located in region II (2 985 403–3 021 130) (Fig. S1).