Swarm plates were inoculated by placing a drop of the cell culture on one side and then placing three antibiotic disks on the other side of the plates. The plate cultures were all grown at 22 °C for 3 weeks. The experiment was performed twice in triplicate. Swarming motility was observed for R. leguminosarum at agar concentrations ranging from 0.5% to 1% (Fig. 1). At a lower concentration of agar (0.5%), a mixture of swimming (i.e. penetrating

into the soft agar) and swarming cells was observed. Swarming motility was inhibited at 1.3% agar concentration. Optimal swarming motility was observed using 0.7% agar; hence, this concentration was adopted for the subsequent experiments. The effect of inoculum size was determined using cell cultures with OD600 nm values of Birinapant ic50 0.005, 0.01, 0.05, and 1.8. Swarming migration was observed for all the cell densities used. However, the onset of swarming from the point of inoculation was slower with fewer cells.

Initiation of swarming migration was faster as the cell density was increased. This trend was observed for both VF39SM and 3841 strains. Therefore, in subsequent swarming experiments, cell suspensions with OD600 nm values between 1.2 and 1.8 were used to obtain a full swarming phenotype in 2–3 weeks. Swarming motility was observed when the swarm plates were IWR-1 supplier incubated at 22 °C, and was inhibited at the normal incubation temperature (30 °C) for R. leguminosarum (data not shown). Although there are slight differences in the swarming pattern, all of the carbon sources (glycerol, mannitol, rhamnose, and erythitol) supported swarming motility (Supporting Information, Fig. S1). To determine whether sugar metabolism is important for swarming motility, we performed swarm assays using strains LRS39301 and 3841c−, both of which are cured of the c plasmid (pRleVF39c/pRL9) that contains the genes Ketotifen for glycerol utilization (Yost et al., 2006). We also determined the swarming motility of strain LRS39601 cured of the f plasmid (pRleVF39f), which is needed for erythritol uptake and catabolism (Yost

et al., 2006). Swarm media containing either glycerol or erythritol as the carbon source were used for strains LRS39301/3841c− and LRS39601, respectively. The plasmid-cured strains were unable to swarm under these conditions and the colonies appeared dry (Fig. 2). The wild-type strains swarmed without a supplementary carbon source, but the swarming was significantly reduced (Fig. 2). The two R. leguminosarum strains exhibited different swarming patterns, with VF39SM swarming better than 3841 (Fig. 2b and f). Faster initiation of surface migration was also observed for VF39SM and this strain was able to colonize almost the entire surface of the medium. The description of swarming motility that we present here is for VF39SM. The development of the swarm colony was observed by time-lapse photography (Video S1).