Unless otherwise indicated, pots were irrigated every 3–4 days with sterile-distilled water. The water status of each pot was assessed gravimetrically by weighing the pots before and after watering and draining. Flooded pots were treated in

the same way, except that no holes were placed in the pots; thus, all irrigating water was retained. Control pots without bacteria or with each strain inoculated individually were run in parallel. After 20 days, the strain occupying each nodule was identified with selective antibiotics (López-García et al., 2001). Results were analyzed using the χ2 test. The null hypothesis was that 60% of nodules contained bacteria with the antibiotic marker of the mutant and 40% of nodules contained bacteria with the antibiotic marker of the parental buy Everolimus strain. To obtain the expected values, we multiplied the total number of nodules of each plant by the fraction corresponding to the null hypothesis. With these values and the observed values from each plant, we calculated the χ2 values, which were compared against tabulated χ2 values. The main characteristics of the mutants are summarized in Table 1. Each mutant lacked the desired flagellin, as indicated by its electrophoretic motility, which matched that Pictilisib clinical trial previously identified by Althabegoiti et al. (2008) as FliCI-II or FliC1-4 (Fig. 1). The loss of flagellins led to the loss of corresponding flagellar

filaments (Fig. S2). Phase-contrast microscopy showed that, while LP 5843 and LP5844 (ΔfliC1-4) tumbled more frequently than the wild type, LP6865 and LP 6866 (ΔfliCI-II) swam more straight, while LP6543 and LP6644 (ΔfliCI-IIΔfliC1-4) did not swim, corroborating previous observations by Kanbe et al. (2007). In addition, we recorded the rotation sense of

57 tethered cells. In 16 videos recorded from ΔfliCI-II mutants, we observed clockwise rotation in 18 cells and counterclockwise rotation in another 18 cells (a total of 36 tethered cells of this mutant were observed), suggesting that the thick flagellum rotates in both directions with no bias. In contrast, all 21 Thymidine kinase cells observed in 11 videos from ΔfliC1-4 mutants rotated in the clockwise direction. Because the rotation observed in tethered cells was in the opposite direction to flagellar rotation, these observations indicate that the thin flagellum rotates only in the counterclockwise direction. In agreement with our previous findings, swimming halos produced in Götz 0.3% agar by LP 3008 were wider than those of LP 3004 (Fig. 2). Furthermore, mutants lacking the thick or the thin flagellum produced smaller halos than their respective parental strains. In the background of LP 3004, both mutants lacking one flagellum produced halos of similar size; in contrast, in the background of LP 3008, LP 5844 (ΔfliC1-4) produced wider halos than LP 6866 (ΔfliCI-II).