A Molecular Clutch Disables Flagella in the Bacillus subtilis Biofilm

See allHide authors and affiliations

Science  20 Jun 2008:
Vol. 320, Issue 5883, pp. 1636-1638
DOI: 10.1126/science.1157877


Biofilms are multicellular aggregates of sessile bacteria encased by an extracellular matrix and are important medically as a source of drug-resistant microbes. In Bacillus subtilis, we found that an operon required for biofilm matrix biosynthesis also encoded an inhibitor of motility, EpsE. EpsE arrested flagellar rotation in a manner similar to that of a clutch, by disengaging motor force-generating elements in cells embedded in the biofilm matrix. The clutch is a simple, rapid, and potentially reversible form of motility control.

Many bacteria in the environment live either as motile planktonic individuals or within sessile multicellular groups called biofilms (1). Planktonic cells are motile owing to the presence of rotating flagella: molecular machines assembled from nearly 30 proteins (2). A motor, located at the base of each flagellum, uses the proton motive force to power rotation of an extracellular helical filament and propel the bacterium through the environment (3). In contrast, cells within a biofilm are nonmotile and aggregated by an extracellular matrix composed of secreted macromolecules (4). Motility and matrix synthesis are often coordinately and oppositely regulated.

In Bacillus subtilis, motility and biofilm formation are alternately controlled by the DNA binding transcription factor SinR (5). SinR represses the transcription of genes that encode the structural components of the biofilm including the eps operon that encodes enzymes that synthesize the matrix extracellular polysaccharide (EPS) (6). Cells with mutations in sinR constitutively derepress the eps genes, overproduce the matrix, and grow in sticky aggregates. We investigated why sinR mutants are also nonmotile (Fig. 1A).

Fig. 1.

EpsE is an inhibitor of motility. (A) Circles are top views of swarm agar Petri plates. Plates inoculated with wild-type (3610), sinR (DS859), sinR epsH (DS1674), ΔepsE sinR epsH (DS2946), and ΔepsE (epsE+) sinR epsH (DS3844) were filmed against a black background such that zones of bacterial colonization appear white and uncolonized agar appears black. Motility defects are indicated as small zones of colonization. Bar, 1 cm. (B to E) Wild-type (DS1916), sinR (DS2198), sinR epsH (DS2179), and ΔepsE sinR epsH (DS3843) strains expressing the HagT209C construct were stained for membranes (false-colored red) and flagella (false-colored green). Bar, 2 μm.

SinR has no appreciable effect on the expression of genes required for flagellar motility (7). We considered the possibility that cells with mutations in sinR assembled flagella that were obscured by an excess of EPS and severe cell aggregation. To detect flagella within cell aggregates, we genetically altered the flagellar filament protein, Hag (8), such that the filament could be labeled with a fluorescent probe (HagT209C) (Fig. 1B and fig. S1). Flagella colocalized with the cell membrane in the sinR mutant but were unfettered when EPS biosynthesis and cell aggregation were abolished by introduction of an epsH mutation (Fig. 1, C and D). Nonetheless, the sinR epsH double mutant remained nonmotile (sinR epsH) (Fig. 1, A and D). Thus, the sinR mutant synthesized flagella that were concealed by the biofilm EPS matrix, and the flagella were also nonfunctional.

To ascertain how SinR regulates flagellar function, we genetically selected for second-site sor (suppressor of sinR) mutations that restored motility in the absence of SinR and EpsH. Eighteen spontaneous sor suppressor mutant strains were independently isolated, and nine of the suppressors contained mutations in the gene epsE (formerly yveO), encoding EpsE, a putative family II glycosyltransferase (9) (fig. S2). An artificially constructed in-frame markerless deletion of epsE restored motility to the sinR epsH mutant (ΔepsE sinR epsH) (Fig. 1, A and E), and motility inhibition was complemented when the epsE gene was cloned directly downstream of the native Peps promoter and integrated at an ectopic locus [ΔepsE (epsE+) sinR epsH] (Fig. 1A). The epsE gene is a member of the eps operon, the transcription of which is directly repressed by SinR. Thus, cells with mutations in sinR are nonmotile as a result of constitutive EpsE derepression.

To determine whether EpsE was sufficient to inhibit motility, we fused the epsE gene to an isopropyl β-d-thiogalactopyranoside (IPTG)–inducible Physpank promoter and integrated the fusion at an ectopic locus (amyE) of anotherwise wild-type strain (amyE::Physpank-epsE). In the absence of induction, cells were vigorously motile (movie S1), and fluorescent staining of the HagT209C-modified filament revealed a blur of rapidly rotating flagella (movie S2). After 40 min of induction, EpsE completely inhibited cell motility (movie S3), and paralyzed flagella were clearly visible on the surface of the nonmotile cells (movie S4). Motility was similarly inhibited by an induced allele of EpsE for which a highly conserved glycosyltransferase active-site aspartate residue (10) was mutated to an alanine (amyE::Physpank-epsED94A). Thus, EpsE is sufficient to inhibit motility and does so by arresting flagellar rotation. Furthermore, the mechanism by which EpsE inhibits the flagellum is apparently unrelated to its putative enzymatic activity.

To identify the target with which EpsE interacted to inhibit flagellar rotation, we selected for spontaneous sorE mutations (suppressors of redundant EpsE) in the downstream cellular target that could render motility insusceptible to inhibition by EpsE. To reduce the possibility that rescue of motility mutations would disrupt the epsE gene itself, we expressed two copies of epsE on the chromosome: one at the native site by derepression due to mutation of sinR, and the other at the ectopic amyE site by induction with IPTG (amyE::Physpank-epsE). The epsH gene was mutated to eliminate EPS production that might confound mutant recovery. Twelve motile sorE suppressor strains were isolated from the nonmotile parent, and all 12 suppressors contained missense mutations in fliG.

The fliG gene encodes a protein that is 32% identical to the flagellar motor component FliG of Escherichia coli. FliG subunits polymerize into a wheel-like rotor attached to the flagellar basal body and transduce the energy of proton flux through the MotA-MotB proton channel into the rotational energy of the flagellum (1113). The 12 different suppressors were mutated in four different residues within the C-terminal domain of FliG specifically required for torque generation (fig. S3) (14, 15). Three of the four residues localized to a common surface when mapped onto the three-dimensional structure of FliG from Thermotoga maritima (Fig. 2A). The fourth residue, the C terminus of FliG, fell in an eight–amino acid disordered domain that might occupy a groove in the protein and potentially position the C-terminal residue near the other three suppressor sites. The surface of FliG containing the altered residues is not occluded by known interactions with the flagellum components FliF, MotA, or FliM and is a candidate site for interaction with EpsE (1618).

Fig. 2.

EpsE interacts with the flagellar rotor. (A) The sites of missense mutations generated from sorE alleles (fig. S3) were mapped onto the published structure of FliG from T. maritima (24): Ala264 (blue, corresponding to B. subtilis FliGS267L), Gly267 (red, corresponding to B. subtilis FliGV270G), and Leu300 (green, corresponding to B. subtilis FliGL303S). The final amino acid altered by sorE-class mutations FliGV338A and FliGV338D does not appear on the structure because the most C-terminal residue that is ordered and visible is Arg327 (indicated by an asterisk). The surfaces of FliG predicted to interact with FliF, MotA, and FliM are indicated by curved lines. (B and C) A strain expressing the wild-type allele of FliG (FliGWT, DS2989) or sorE10 allele of FliG (FliGV338A, DS3004) was doubly mutated for sinR and ΔepsE with a complementing EpsE-GFP construct integrated at an ectopic locus. Membranes were stained with FM4-64 and false-colored in red (“Membrane”). GFP signals were false-colored in green (“EpsE-GFP”). Bar, 2 μm. Enlarged images of each panel are shown in fig. S7.

To investigate whether EpsE inhibited motility through interaction with FliG, we determined the subcellular localization of EpsE translationally fused to the green fluorescent protein (GFP) and expressed under the control of the native Peps promoter. In cells with the sinR mutation, the EpsE-GFP construct was fully able to inhibit motility and when visualized by fluorescence microscopy, the EpsE-GFP fusion localized as puncta that were associated with the cell membrane, reminiscent of the sites of flagellar basal bodies (Fig. 2B). EpsE failed to localize as puncta in the presence of the FliG alleles that rendered the flagellum insusceptible to inhibition by EpsE (Fig. 2C and fig. S4). Thus, punctate localization of EpsE was required to inhibit motility, EpsE and FliG interacted in vivo, and EpsE puncta represented sites of flagellar basal bodies.

When EpsE interacts with FliG, it may cause a deflection of the FliG C-terminal domain, alter FliG interaction with MotA, and inhibit flagellar rotation in a manner similar to that of a brake or a clutch (19, 20). As a brake, EpsE would jam the rotor and immobilize the flagellum. As a clutch, EpsE would disengage the rotor from the power source, but the flagellum would still be free to rotate. To distinguish these models, we tethered B. subtilis cells by a single flagellum and measured flagellar rotation by the counter-rotation of the cell body (21). Tethered motile cells rotated at a rate of about one revolution in 5 s (Fig. 3). When EpsE was induced, cell rotation was reduced but not abolished and resembled that of cells failing to express the MotA-MotB proton channel that provides the power for flagellum rotation (Fig. 3). The rotation of cells either induced for EpsE or lacking motA and motB was consistent with a calculated root mean square angular deviation of ∼80° in 90 s for a B. subtilis cell tethered by an unpowered flagellum freely rotating by Brownian motion (supporting online material). In contrast, the angular deviation is much less, ∼3°, for an E. coli cell tethered by an immobilized flagellum (22). Thus, EpsE acted as a clutch; when EpsE was induced, the flagella behaved as though they were unpowered rather than immobilized.

Fig. 3.

When EpsE is expressed, flagella behave as though they are unpowered. Cells were tethered to microscope slides by sheared flagellar stubs, and cell rotation was monitored by video microscopy. The behavior of cells in a particular field was not uniform, and therefore 100 cells were chosen at random for motion analysis. (Left) The histograms indicate the number of cells that rotated through the indicated angle of rotation during a 90-s interval for “EpsE uninduced” (DS3317), “EpsE induced” (DS3317 induced with 2 mM IPTG for 2.5 hours before observation), and “motAmotB” (DS3318) cells. (Right) Each image is a time-lapse composite of a sample field corresponding to the histogram immediately to its left. Parts of the cell that are stationary are black. The range of motion that a cell travels through is depicted as a time-lapse composite and is shown in white. Bar, 10 μm.

The biological function of the clutch appears to be related to the B. subtilis biofilm because epsE is encoded within the 15-gene eps operon that promotes the biosynthesis of the biofilm EPS and is repressed by SinR, the master regulator of biofilm formation. Therefore, control of a single locus ensures that cells become immobilized concomitant with biofilm formation (fig. S5A). In wild-type cells, trapped flagella and puncta of EpsE were observed within biofilm aggregates, and the cells within the aggregates were sessile (fig. S6 and movie S5). Cells expressing the FliGV338A clutch-insusceptible allele formed aggregates, but the cells writhed within the confines of the matrix (movie S6). We hypothesize that the clutch helps to stabilize biofilms in the environment and acts as a fail-safe mechanism to ensure that flagella do not rotate while the cells are bound by EPS.

The bacterial flagellum, powered by a motor that generates 1400 pN-nm of torque, can rotate at a frequency of greater than 100 Hz (23). EpsE disabled this powerful biological motor when associated with a flagellar basal body and, in a manner similar to that of a clutch, disengaged the drive train from the power source (fig. S5B). Clutch control of flagellar function has distinct advantages over transcriptional control of flagellar gene expression for regulating motility. Some bacteria, such as E. coli and B. subtilis, have many flagella per cell. The flagellum is an elaborate, durable, energetically expensive, molecular machine and simply turning off de novo flagellum synthesis does not necessarily arrest motility. Once flagellar gene expression is inactivated, multiple rounds of cell division may be required to segregate preexisting flagella to extinction in daughter cells. In contrast, the clutch requires the synthesis of only a single protein to inhibit motility. Furthermore, if biofilm formation is prematurely aborted, flagella once disabled by the clutch might be reactivated, allowing cells to bypass fresh investment in flagellar synthesis. Whereas flagellum expression and assembly are complex and slow, clutch control is simple, rapid, and potentially reversible.

Supporting Online Material

Materials and Methods

SOM Text

Figs. S1 to S7

Tables S1 to S3


Movies S1 to S6

References and Notes

View Abstract

Stay Connected to Science

Navigate This Article