Interaction of the cyclic-di-GMP binding protein FimX and the Type 4 pilus assembly ATPase promotes pilus assembly

Abstract

Type IVa pili (T4P) are bacterial surface structures that enable motility, adhesion, biofilm formation and virulence. T4P are assembled by nanomachines that span the bacterial cell envelope. Cycles of T4P assembly and retraction, powered by the ATPases PilB and PilT, allow bacteria to attach to and pull themselves along surfaces, so-called "twitching motility". These opposing ATPase activities must be coordinated and T4P assembly limited to one pole for bacteria to show directional movement. How this occurs is still incompletely understood. Herein, we show that the c-di-GMP binding protein FimX, which is required for T4P assembly in Pseudomonas aeruginosa, localizes to the leading pole of twitching bacteria. Polar FimX localization requires both the presence of T4P assembly machine proteins and the assembly ATPase PilB. PilB itself loses its polar localization pattern when FimX is absent. We use two different approaches to confirm that FimX and PilB interact in vivo and in vitro, and further show that point mutant alleles of FimX that do not bind c-di-GMP also do not interact with PilB. Lastly, we demonstrate that FimX positively regulates T4P assembly and twitching motility by promoting the activity of the PilB ATPase, and not by stabilizing assembled pili or by preventing PilT-mediated retraction. Mutated alleles of FimX that no longer bind c-di-GMP do not allow rapid T4P assembly in these assays. We propose that by virtue of its high-affinity for c-di-GMP, FimX can promote T4P assembly when intracellular levels of this cyclic nucleotide are low. As P. aeruginosa PilB is not itself a high-affinity c-di-GMP receptor, unlike many other assembly ATPases, FimX may play a key role in coupling T4P mediated motility and adhesion to levels of this second messenger.

Fig 1. FimX localizes to the leading pole in twitching bacteria.

PA103ΔfimX ptdimer2-FimX twitching at an agar-glass interface was imaged every 20s by spinning disk confocal epifluorescence microscopy. (A) Location of tdimer2-FimX was scored in 94 moving bacteria. Fourteen of 17 cells with bipolar FimX were in close proximity to moving cells in which FimX localized to the leading pole. (B) Montage of ΔfimX ptdimer2-FimX bacteria moving singly or as a group; time lapse interval (s) is indicated for each frame. (C) tdimer2-FimX is predominantly bipolar in stationary cells (n = 52). Lower panel shows a representative time lapse series (every 20s) of cells with bipolar tdimer2-FimX. (D) tdimer2-FimX localization is dynamic and correlated with bacterial movement. A cell (marked with a yellow arrowhead) pauses during imaging (120-240s) before twitching again (280-320s). Fluorescence intensity profile of this cell (E) shows signal at the leading (green) and lagging (red) poles over time of imaging. The blue line indicates a moving (1 or -1) or stationary (0) cell. (F) Bacteria with bipolar FimX appear to be “dragged” by cells with unipolar FimX signal. The fluorescence intensity profile for the cell being pulled (yellow asterisk) is shown in Panel G.

Fig 2. The EAL domain is required for polar localization of FimX.

(A) Distribution of tdimer2-FimXΔEAL in PA103ΔfimX. Bacteria were imaged on at least 2–3 independent days over 5–6 different fields. The figure shows representative examples. (B) Intensity profiles of tdimer2-FimX, -FimX(AAA), and -FimXΔEAL expressed in PA103 (n = 30/strain). Cell length is shown in pixels.

Fig 3. FimX unipolar localization requires structural components of the pilus assembly machinery.

(A) tdimer2-FimX localization in live PA14 fimX, pilQ, pilF or pilC mutants. Plate grown bacteria were spotted onto 1% agar pads. Upper panel, phase contrast; lower panel, tdimer2 epifluorescence. (B) Distribution of tdimer2-FimX in fimX, pilQ, pilF and pilC mutants. The number of cells scored for each strain is noted above each bar. (C) Fluorescence intensity profile of tdimer2-FimX for fimX and pilQ mutants over the length of the cell in pixels (n = 50). (D) Distribution of endogenous FimX between membrane (M) and cytosolic (C) fractions differs in PA14 vs. pilQ::Tn bacteria. Samples were immunoblotted for FimX, the cytosolic protein Hfq, and the outer membrane protein PilQ.

Fig 4. Localization of tdimer2-FimX is altered by the absence of minor pilins and PilY1.

(A) The location of tdimer2-FimX in PA14 pilA (major pilin), pilY1(regulator/adhesin) or pilV,W,X,E (minor pilins) mutants was determined by epifluoresence imaging of live bacteria harvested from agar plates. (B) Distribution of tdimer2-FimX expressed in bacteria with transposon insertions in the indicated genes; the number of cells scored for each mutant appears above the bars.

Fig 5. FimX distribution is altered in mutants lacking T4P regulatory proteins.

tdimer2-FimX distribution in bacteria lacking (A) the PilS/PilR TCS and (B) Pil/Chp chemotaxis cluster proteins was determined by epifluoresence imaging of live bacteria. (C) Localization of tdimer2-FimX in PA14 pilH::Tn and WT pilH complemented strain. The pilH gene was introduced at the attB site on the chromosome and expressed from its endogenous promoter.

Fig 6. FimX and PilB localization depend on each other’s presence.

(A) tdimer2-FimX localization in ATPase mutants was determined by epifluoresence imaging of live bacteria. (B) Distribution of tdimer2-FimX localization is shown for each mutant; numbers above each bar indicate how many cells were scored. (C) Localization of YFP-PilT in PA103, PA103ΔfimX and PA103ΔfimX attB::FimX(AAA). The number of cells scored for each strain is shown above the corresponding bar. (D) Fluorescence images of PA103 bacteria expressing plasmid borne YFP-PilT (left) or single copy attB:: P_pilT-GFP-PilT. (E) Fluorescence images of PA103 and PA103ΔfimX bacteria expressing single-copy attB::P_pilB-GFP-PilB.

Fig 7. FimX promotes pilus assembly rather than inhibiting pilus retraction.

(A) Western blot with anti-PilA antiserum of whole cell extract and sheared surface-associated pilin from PA103 and isogenic deletion mutants after overnight culture. Samples were normalized to the amount of protein in the total cell extract. (B) Bacterial cells were vortexed (to shear off surface pili), pelleted and resuspended in fresh media. Cell-associated pili (Surface Pili) and spent media (Media) were collected for indicated strains immediately after resuspension (t = 0h), 3h after resuspension, and after overnight (o/n) culture. Surface pili and TCA-precipitated spent medium were separated by 15% SDS-PAGE and immunoblotted with anti-PilA antiserum. (C) Bacterial cells were vortexed to shear off surface pili, resuspended in fresh medium, and collected at specified intervals after resuspension (0.5h, 1h, 1.5h, 2h, 3h, 4h and overnight (o/n)). Surface pili were then sheared from bacteria, separated by 15% SDS-PAGE and immunoblotted for PilA protein.

Fig 8. PilB and FimX interactions are seen in a split luciferase complementation assay.

PA103 co-expressing Nluc- and Cluc-fusion protein constructs as indicated (or empty vector “V”) were grown overnight on LB plates containing appropriate antibiotics and 0.05% arabinose. Bacteria were harvested by scraping, resuspended in PBS and normalized by OD₆₀₀. RLU (luciferase units/OD₆₀₀) was measured on addition of 1 μM coelenterazine substrate to the bacterial cells. Bars represent the mean ± SD of two technical replicates from a representative experiment. Values for FimX-FimX are set to 100%.

Fig 9. The elution profile of FimX, but not FimX(AAA), is altered by co-incubation with PilB.

PilB, FimX and FimX(AAA) were analyzed by size exclusion chromatography individually and after co-incubation, as indicated. Collected fractions (Fraction 17–25) were separated by SDS-PAGE and visualized by staining. The positions at which the molecular weight standards thyroglobulin (670 kDa) and γ-globulin (158 kDa) eluted are marked by arrows.