Flagellar stator genes control a trophic shift from obligate to facultative predation and biofilm formation in a bacterial predator

Abstract

The bacterial predator Bdellovibrio bacteriovorus is considered to be obligatorily prey (host)-dependent (H-D), and thus unable to form biofilms. However, spontaneous host-independent (H-I) variants grow axenically and can form robust biofilms. A screen of 350 H-I mutants revealed that single mutations in stator genes fliL or motA were sufficient to generate flagellar motility-defective H-I strains able to adhere to surfaces but unable to develop biofilms. The variants showed large transcriptional shifts in genes related to flagella, prey-invasion, and cyclic-di-GMP (CdG), as well as large changes in CdG cellular concentration relative to the H-D parent. The introduction of the parental fliL allele resulted in a full reversion to the H-D phenotype, but we propose that specific interactions between stator proteins prevented functional complementation by fliL paralogs. In contrast, specific mutations in a pilus-associated protein (Bd0108) mutant background were necessary for biofilm formation, including secretion of extracellular DNA (eDNA), proteins, and polysaccharides matrix components. Remarkably, fliL disruption strongly reduced biofilm development. All H-I variants grew similarly without prey, showed a strain-specific reduction in predatory ability in prey suspensions, but maintained similar high efficiency in prey biofilms. Population-wide allele sequencing suggested additional routes to host independence. Thus, stator and invasion pole-dependent signaling control the H-D and the H-I biofilm-forming phenotypes, with single mutations overriding prey requirements, and enabling shifts from obligate to facultative predation, with potential consequences on community dynamics. Our findings on the facility and variety of changes leading to facultative predation also challenge the concept of Bdellovibrio and like organisms being obligate predators.

Importance: The ability of bacteria to form biofilms is a central research theme in biology, medicine, and the environment. We show that cultures of the obligate (host-dependent) "solitary" predatory bacterium Bdellovibrio bacteriovorus, which cannot replicate without prey, can use various genetic routes to spontaneously yield host-independent (H-I) variants that grow axenically (as a single species, in the absence of prey) and exhibit various surface attachment phenotypes, including biofilm formation. These routes include single mutations in flagellar stator genes that affect biofilm formation, provoke motor instability and large motility defects, and disrupt cyclic-di-GMP intracellular signaling. H-I strains also exhibit reduced predatory efficiency in suspension but high efficiency in prey biofilms. These changes override the requirements for prey, enabling a shift from obligate to facultative predation, with potential consequences on community dynamics.

Keywords: Bdellovibrios; cyclic di-GMP; flagellar motility; predation; stator.

Fig 1

Biofilm formation by H-I strains BFF MHI39 and MHI179; SAD MHI153, MHI353, and MHI182; and SAS MHI154 and MHI167 and by the parental H-D100Sm strain, under various conditions. Biofilms were measured by crystal violet (CV) adsorption (A) and using the XTT cell-respiration assay (B). CV measurements of biofilms at different temperatures (C), pH (D), and concentrations of divalent (Ca²⁺, Mg²⁺) cations (E). Bars with different letters above show categories significantly different by Tukey’s test at P < 0.05. Error bars represent standard error of the mean. Microscopic examination of biofilms formed by H-I strains: strains SAS MHI154-tdT (F), SAD MHI153-tdT (G); and BFF MHI179-tdT (H) biofilms viewed by red fluorescence (RF)P epifluorescence. Scanning electron micrographs of biofilms formed by strains SAS MHI154 (I), SAD MHI153 (J), and BFF MHI179 (K). Insets are higher magnifications of a smaller field.

Fig 2

Confocal microscopy-based acquisition of BFF strain MHI179-tdT biofilms grown in static culture for 2 days (A) and under continuous culture for 7 days (B). Biofilm thickness is color-coded based on scale. The bottom part of panels A and B is an x-y axis view, rendering biofilm thickness. Effect of lytic enzymes on biofilm EPS of BFF strain MHI179, and differential staining of matrix components. The percentage of biofilm removal by proteinase K, bromelain, DNase I, α-amylase, and lyticase in comparison to their heat-inactivated counterparts (C). The results are averages with stardard errors of three or more experiments. The percentage of biofilm reduction by different enzymes was statistically analyzed by unpaired t test comparing the mean of each treatment with the mean of their heat-inactivated counterparts. BFF strain MHI179-tdT biofilm EPS proteins and eDNA stained by FITC isomer I (D, E, and F) and DiTO-1 (G, H, and I), respectively, as merged (D and G), stain only (E and H), and tdT (F and I) signals.

Fig 3

Predatory growth on E. coli ML35 of H-I vs H-D100 strain in suspension and temperature effects. (A) Planktonic growth at 28°C of strains BFF MHI179-Tdt, SAD MHI153-Tdt, SAS MHI154-Tdt, and of the parental H-D100Sm-Tdt, measured by plaque forming units (PFU/mL). (B) Prey population reduction, measured by remaining prey counts colony forming units (CFU/mL). The associated growth and prey decay curves showing predation dynamics, maximal growth rates, and inflexion points are presented in Fig. S9 and S10. (C) The effect of temperature on predation by the same strains was measured by remaining prey counts. Error bars indicate starndard error. E. coli ML35 without predatory interaction was used as control. The statistical significance of mean values was analyzed by RM (repeated measure) one-way ANOVA with Geisser-Greenhouse correction and P-values. ns = P > 0.05, * = P ≤ 0.05, ** = P ≤ 0.01, *** = P ≤ 0.001, and **** = P ≤ 0.0001.

Fig 4

Biofilm formation by P. fluorescens and remaining biofilms (A and B), predation within the biofilms (C and D), and growth of the predators (E and F), using strains SAS MHI154, SAS MHI167, and SAD MHI153 (A, C, and E), and BFF MHI179 (B, D, and F), and by the parental H-D100Sm (A to F). (A and B) Crystal violet biofilm staining. (C and D) Plate counts (CFU/mL) of remaining prey in the biofilms. (E and F) Plaque counts (PFU/mL) of the predators in the biofilms. Error bars indicate standard error. The statistical significance of mean values was analyzed by RM (repeated measure) one-way ANOVA with Geisser-Greenhouse correction and P-values. ns = P > 0.05, * = P ≤ 0.05, ** = P ≤ 0.01, *** = P ≤ 0.001, and **** = P ≤ 0.0001.

Fig 5

Phenotypes of Δbd1076-deleted biofilm-forming (BFF) MHI39 and MHI179 strains, and native bd1076-complemented strains, and of surface-associated (SAS) MHI154, in comparison to the parental H-D100Sm strain. Percentage of biofilm reduction in clonal lineages (identified by numbers) of (A) Δbd1076MHI39 and (B) Δbd1076MHI179; compared to their respective parental strain (100%). (C) Growth of Δbd1076H-D100Sm strains in PYE medium. (D) Predation on E. coli ML35 by Δbd1076H-D100Sm strains. (E) CV measurement of biofilm formation by Δbd1076H-D100Sm strains. (F) Growth of bd1076-complemented bd1076; (G) of bd1076-complemented bd1076:MHI167, in PYE medium. The results are averages and SEs of three or more experiments (except SE values that are omitted from predation and growth curves). Bars with different letters above show categories significantly different by Tukey’s test at P < 0.05.

Fig 6

Swimming and swarming motility of H-I and parental H-D100Sm strains. Swimming speed (μm/s) of AP cells in amHEPES (A) and in PVP-supplemented amHEPES (B). Plaque diameter (mm) of H-D100Sm, SAS MHI154, SAD MHI153, and BFF MHI179 after 2 days (C), 4 days (D), and 7 days (E) of incubation. Values are the mean along with the SEs. In each line and box plot, different letters mark significant differences by Tukey’s test at P < 0.05.

Fig 7

Conceptual schematic model of possible signaling cascades leading to different H-I phenotypes emerging from mutations in stator genes or in bd0108. (A) Parental strain: upon prey binding, CdG fluxes change between the invasion and flagellar poles, bringing upon changes in gene expression (29). Prey invasion leads to the dismantling of the flagellar machinery (77), along with further changes in CdG fluxes, and induction of the GP. (B) In stator gene mutants, the stator is destabilized, altering CdG fluxes in the AP and the action of GP control proteins. The mutants are Type II, and poorly associate (surface-associated, SAS) or adhere (surface-adhering, SAD) to surfaces but do not form biofilms (left side). bd0108 mutations form Type I H-I strains unable to form biofilms. Additional mutations result in Type II strains, which adhere (SAD) or form biofilms on surfaces (biofilm formers [BFF]). Stator genes are necessary for strong biofilm growth (right side, with biofilm intensity indicated by darker colors).