Adaptor-mediated Lon proteolysis restricts Bacillus subtilis hyperflagellation

Significance

Bacteria are thought to change physiology when in contact with a solid surface, but the mechanism of surface-contact signal transduction and the output physiological changes are often poorly understood. Here, we show that Bacillus subtilis controls flagellar density by regulatory proteolysis of the master flagellar activator protein SwrA. We further show that the broadly conserved AAA+ protease LonA degrades SwrA only in the presence of swarming motility inhibitor A, the first substrate-specific adaptor protein reported for the Lon family. We propose that surface contact inhibits proteolytic turnover such that SwrA accumulates and the cells synthesize flagella in excess of a critical threshold required for swarming migration.

Abstract

The Lon AAA+ protease is a highly conserved intracellular protease that is considered an anticancer target in eukaryotic cells and a crucial virulence regulator in bacteria. Lon degrades both damaged, misfolded proteins and specific native regulators, but how Lon discriminates among a large pool of candidate targets remains unclear. Here we report that Bacillus subtilis LonA specifically degrades the master regulator of flagellar biosynthesis SwrA governed by the adaptor protein swarming motility inhibitor A (SmiA). SmiA-dependent LonA proteolysis is abrogated upon microbe-substrate contact causing SwrA protein levels to increase and elevate flagellar density above a critical threshold for swarming motility atop solid surfaces. Surface contact-dependent cellular differentiation in bacteria is rapid, and regulated proteolysis may be a general mechanism of transducing surface stimuli.

Protein degradation and turnover serves multiple purposes. At one level, protein degradation is a form of quality control that prevents accumulation of damaged proteins and recycles amino acids (1, 2). At another level, proteolysis is a regulatory strategy to reshape the cellular proteome and govern cellular differentiation (3–5). Regulatory proteolysis bypasses transcription and translation to allow rapid adaptation but it is also irreversible and therefore demands strict discrimination of specific targets from all other proteins in the cell. In bacteria, proteolytic targeting is determined by a combination of intrinsic substrate preference of the protease together with substrate-specific adaptor proteins that noncovalently associate with their targets for presentation to proteolytic enzymes (6, 7). Adaptors often work as tethers, increasing local concentration of substrates to drive degradation, or as obligate activators, required for protease assembly itself (4, 7–9).

Lon (La) was the first intracellular protease discovered and arguably remains one of the least understood with respect to substrate recognition (10, 11). Lon is thought to degrade misfolded proteins by recognizing hydrophobic peptide sequences known as degrons that are normally buried in the hydrophobic core of folded proteins and become solvent exposed upon protein misfolding (12, 13). Lon is also known to conduct regulatory proteolysis by highly specific degradation of key cellular regulators but to date, little has been reported regarding Lon-specific adaptor proteins (10, 14–17). Thus, if and how Lon distinguishes specific regulatory targets from general misfolded proteins is an open question. Further, despite highly pleiotropic effects, Lon is often not essential but is commonly associated with alterations in bacterial motile behavior (18–22).

Bacteria exhibit different motile behaviors depending on whether they are in a liquid environment or in contact with a solid surface (23, 24). In liquid media, some bacteria make flagella and swim as individuals in three dimensions but when presented with a solid surface they initiate rapid swarming motility in two dimensions over the substrate (25, 26). Further, swimmer and swarmer cells are physiologically distinct as transitioning to surfaces includes a lag period of immobility that may be a time for cellular differentiation in response to surface contact. Importantly, differentiation to a swarmer cell has been associated with enhanced virulence, elevated antimicrobial resistance, and increased flagellar density (25, 27–32). The mechanism of surface contact recognition and subsequent signal transduction is unknown save that the transition may be controlled by master regulators of flagellar biosynthesis and the Lon protease (33, 34).

Here we demonstrate that the LonA protease of Bacillus subtilis inhibits the swarmer cell state by proteolytically restricting accumulation of SwrA, the master activator of flagellar biosynthesis, in liquid environments. When presented with a solid surface, LonA restriction is relieved and SwrA accumulation results in increased flagellar synthesis. We further show that an increase in flagellar density is an obligate requirement to swarm and that B. subtilis swarmer cells require an excess of seven flagella per micrometer of cell length. Importantly, SwrA was not degraded by LonA either in vivo or in vitro unless the protein swarming motility inhibitor A (SmiA) was also present, making SmiA the first substrate-specific adaptor reported for Lon. Finally, the importance of surface contact-dependent responses is becoming increasingly recognized in bacteria and LonA/SmiA appears to be an integral part of the B. subtilis surface signal transduction machinery as mutation of either LonA or SmiA resulted in predifferentiated swarmer cells in liquid. Thus, restraining proteolysis of a master regulator upon surface contact, perhaps by disrupting an adaptor, ensures strict on-demand governance of energetically costly flagellar biosynthesis.

Results

SmiA Mediates LonA Proteolysis of SwrA.

Cells mutated for the LonA AAA+ protease swarm immediately upon surface inoculation, whereas liquid-grown wild-type cells lag for a period of ∼1.5 h before initiating swarming motility (Fig. 1A and SI Appendix, Fig. S1A) (20). Thus, LonA appeared to inhibit swarming but the mechanism of LonA-mediated swarming inhibition was unknown. To determine whether the LonA proteolytic activity was required for swarming inhibition, ectopic complementation constructs were generated in which the native lonA promoter expressed either the wild-type lonA gene or alleles of lonA mutated for the conserved active site residues S677 and K720 (35). Whereas complementation of a lonA mutant with the native lonA gene restored the wild-type swarm lag duration, complementation with either active site mutant phenocopied the null allele (Fig. 1B and SI Appendix, Fig. S1 B–D). Furthermore, lagless swarming was specific to mutation of LonA as loss of ClpP or ClpQ protease activity by mutation of the ClpC, ClpX, and ClpY unfoldase subunits had the opposite effect and impaired swarming (SI Appendix, Fig. S1E) (6). We conclude that inhibition of swarming motility is specific to LonA proteolytic activity.

Fig. 1.

LonA protease and SmiA inhibit swarming motility. Quantitative swarm expansion assays for strains are shown. (A) WT (open circles, 3610), lonA (black diamonds, DS5286), and lonA swrA (gray diamonds, DS6242); (B) lonA (lonA^WT) complement (black diamonds, DK1533), lonA (lonA^S677A) (gray diamonds, DK1534), and lonA (lonA^K720Q) (open diamonds, DK1535); (C) WT (open circles, 3610), swrA (gray squares, DS2415), and swrA P_IPTG-swrA + 1 mM IPTG (black squares, DS526). (D) smiA (black triangles, DS4987), smiA P_IPTG-smiA + 1 mM IPTG (open triangles, DS5035), and smiA swrA (DS6243). Each point is the average of three replicates.

One way in which LonA might inhibit swarming is by proteolyzing SwrA, the master activator of swarming motility in B. subtilis (36, 37). SwrA appeared to be genetically downstream of LonA as a lonA swrA double mutant was nonswarming (Fig. 1A and SI Appendix, Fig. S1F). Further, artificial IPTG (isopropyl β-d-1 thiogalactopyranoside)-dependent overexpression of SwrA abolished the swarming lag period phenocopying the absence of LonA (Fig. 1C and SI Appendix, Fig. S1G). To determine whether SwrA was proteolytically degraded, wild-type B. subtilis cells were treated with the translation inhibitor chloramphenicol and SwrA protein levels were assayed at various time points by Western analysis. SwrA was robustly depleted in the wild type, whereas a control protein, the SigA vegetative sigma factor, persisted for up to 4 h (Fig. 2A and SI Appendix, Fig. S2A). Mutation of LonA increased SwrA protein stability and complementation with wild type, but not active site LonA mutants, restored proteolytic turnover (Fig. 2 B–E and SI Appendix, Fig. S2 A and B). By contrast, cells mutated for ClpC, ClpX, or ClpY did not result in stable SwrA protein (SI Appendix, Fig. S2 C–F). We conclude that SwrA is proteolyzed in vivo by the LonA protease.

Fig. 2.

SwrA is degraded in vivo and stabilized in the absence of either LonA or SmiA. Cells were grown to midexponential phase in lysogeny broth and the translation inhibitor chloramphenicol was added to a final concentration of 10 µg/mL at time T = 0 min. Cell samples were withdrawn immediately after addition of chloramphenicol T = 0, and at T = 30, 60, 120, 180, and 240 min, lysed, separated by SDS/PAGE, and separately probed with anti-SwrA and anti-SigA antibody. Strains used were as follows: (A) WT (3610), (B) lonA (DS5286), (C) lonA (lonA^WT) (DK1533), (D) lonA (lonA^S677A) (DK1534), (E) lonA (lonA^K720Q) (DK1535), (F) smiA (DS4987), and (G) smiA P_hyspank-smiA + 1 mM IPTG (DS5035).

To determine whether LonA proteolyzed SwrA directly, an in vitro proteolysis assay was conducted. Purified LonA and SwrA proteins were incubated together over a time period similar to that used in the in vivo chloramphenicol time course experiment. Unlike the in vivo experiment, however, SwrA levels did not decrease over the period of 4 h in the presence of LonA (Fig. 3A). Importantly, the failure to degrade SwrA was not due to proteolytically inactive LonA as purified LonA protein rapidly degraded misfolded α-casein (SI Appendix, Fig. S3A). We conclude that if LonA directly proteolyzes SwrA, then the LonA protease alone is insufficient.

Fig. 3.

LonA protease degrades SwrA in vitro in the presence of SmiA. (A) LonA alone fails to proteolyze SwrA. ^GSTSwrA (10 µM) was incubated with LonA (1.4 µM) in the presence or absence of ATP. Creatine kinase (CK) is required for ATP regeneration. In gel headers (in A and B), + indicates the presence of ATP, − indicates the absence of ATP, and min is the duration of the reaction in minutes. (B) LonA proteolyzes SwrA when SmiA is present. ^GSTSwrA (5 µM) was incubated with LonA (0.7 µM) and SmiA (5 µM) in the presence or absence of ATP. SmiA was originally purified as an N-terminal 6His-SUMO fusion and when cleaved some 6His-SUMO (SUMO) remained present in the reaction mixture. We note that LonA did not degrade SwrA in the presence of purified SUMO alone (SI Appendix, Fig. S3F). (C–G) Densitometry scan averages and SDs of protein bands in proteolytic assays from three reaction replicates. (C) Points indicate the levels of SwrA protein (5 µM initial) over time in the presence of 0.7 µM LonA alone (open circles) and in the presence of 0.7 µM LonA and 5 µM SmiA (closed circles). (D) Points indicate the levels of SwrA protein (5 µM initial) over time in the presence of 5 µM SmiA (closed circles). (E) Points indicate the levels of SwrA protein (5 µM initial) (closed circles) and α-casein (20 µM initial) over time in the presence of 0.7 µM LonA alone. Thirty minutes after the start of the experiment, 5 µM SmiA was added to the reaction mixtures. (F) Points indicate the levels of α-casein (5 µM initial) over time in the presence of 0.7 µM LonA alone (open squares) and in the presence of 0.7 µM LonA and 5 µM SmiA (closed squares). (G) Points indicate the levels of SwrA protein (5 µM initial) over time in the presence of 1.4 µM Lon^Ec and 5 µM SmiA (closed circles). Representative raw data are presented in SI Appendix, Fig. S3.

LonA may have failed to proteolyze SwrA in vitro due to the absence of a critical cofactor. A candidate cofactor is a second protein, SmiA, that was identified in a genetic screen that also identified LonA as an inhibitor of swarming motility (20). Like cells mutated for LonA, cells mutated for SmiA exhibited a lagless swarming phenotype and mutation of SwrA was epistatic to mutation of SmiA as the swrA smiA double mutant failed to swarm (Fig. 1D and SI Appendix, Fig. S1H). Further, mutation of SmiA abolished proteolytic turnover of SwrA in vivo and turnover was restored when SmiA was expressed from an artificially induced promoter (Fig. 2 F and G and SI Appendix, Fig. S2G). To determine whether SmiA participated in LonA-dependent proteolysis of SwrA, all three proteins were purified and combined in an in vitro proteolysis assay. Unlike the addition of LonA alone, the addition of both SmiA and LonA caused SwrA levels to decrease (Fig. 3 B and C). SmiA did not have proteolytic activity on its own as SwrA protein levels remained high in reactions lacking LonA (Fig. 3D and SI Appendix, Fig. S3B). We conclude that SmiA is required for LonA-dependent degradation of SwrA both in vivo and in vitro.

SmiA could enhance SwrA degradation either as an adaptor for specific SwrA proteolysis or as an allosteric activator of generalized protease activity (7, 13, 16). Unfolded proteins have been proposed to be both substrates and allosteric Lon activators but the presence and degradation of α-casein did not promote proteolysis of SwrA (Fig. 3E and SI Appendix, Fig. S3C). By contrast, addition of SmiA at later time points triggered SwrA degradation after α-casein had been depleted (Fig. 3E and SI Appendix, Fig. S3C). Furthermore, SmiA was specific for SwrA and was not a general activator of Lon proteolysis, as the presence of SmiA did not accelerate degradation of α-casein or another synthetic misfolded Lon substrate, Titin I27^CM-β20 (Fig. 3F and SI Appendix, Fig. S3 A and D). Finally, SmiA did not promote SwrA proteolysis by the Escherichia coli Lon ortholog (Lon^Ec), indicating that SmiA was specific for its cognate substrate-protease pair (Fig. 3G and SI Appendix, Fig. S3E). We conclude that SmiA acts as an adaptor, and not an allosteric activator, to target SwrA for proteolysis and reduce SwrA steady-state levels in the cell.

SwrA Accumulation Governs Swarming Motility by Hyperflagellation.

We hypothesize that the lag period that precedes swarming is a time required for SwrA accumulation, as reducing SwrA turnover by mutation of LonA/SmiA caused cells to swarm immediately upon contact with a solid surface. To determine whether SwrA protein level varied in wild type, cell lysates were generated from populations swimming in liquid or swarming on solid surfaces. SwrA Western analysis of SDS/PAGE resolved cell lysates indicated that SwrA levels appeared to be greater in swarmer cells (Fig. 4A). To determine the fold change, lysates were serially diluted and dot blotted followed by Western analysis. Consistent with SDS/PAGE, SwrA levels were 3- to 10-fold higher in wild-type swarmer cells than swimmer cells (Fig. 4B). SwrA levels were elevated posttranslationally, as swrA gene expression measured by β-galactosidase reporters fused transcriptionally or translationally to the swrA promoter and swrA ORF, respectively, showed no difference between swimmers and swarmers (SI Appendix, Fig. S4 A and B). Finally, the absence of LonA or SmiA resulted in constitutively elevated levels of SwrA (Fig. 4 A and B and SI Appendix, Fig. S4 A and B). We conclude that SwrA levels are higher in wild-type swarmer cells than swimmer cells and that LonA/SmiA posttranslationally restricts SwrA levels in swimmers.

Fig. 4.

Absence of LonA or SmiA results in accumulation of SwrA protein in swimmer cells. (A) Western blot analysis of B. subtilis cell lysates of WT (3610), lonA (DS5286), and smiA (DS4987) harvested either from liquid cultures or from the surface of an active swarm, and separately probed with anti-SwrA and anti-SigA primary antibodies. (B) Dot blot analysis of serially diluted B. subtilis cell lysates of WT (3610) and lonA (DS5286), harvested either from liquid cultures or from the surface of an active swarm separately probed with anti-SwrA and anti-SigA primary antibodies. Open carets represent the first dilutions of the WT swimmer and swarmer samples that leave saturation and represent the two points being qualitatively compared for relative SwrA protein levels in the text.

We hypothesize that the lag period that precedes swarming is a time required for increased flagellar biosynthesis because increased SwrA levels have been correlated with an increase in flagellar number per cell (38). To determine whether wild-type swarmer cells had an increase in flagellar number, flagella were counted by fluorescently labeling flagellar hooks and high-resolution 3D structured-illumination microscopy (3D-SIM) (SI Appendix, Fig. S5A) (38, 39). Surface-grown swarming wild-type cells had an average of 9 ± 2 hooks per micrometer of cell length, and liquid-grown wild-type swimmer cells had an average of 4 ± 1 hooks per micrometer of cell length (Fig. 5A and SI Appendix, Fig. S5C). Importantly, the swarmer and swimmer datasets did not overlap and appeared to be separated by a critical threshold flagellar density (SI Appendix, Fig. S5C). Consistent with LonA inhibition of SwrA protein levels in liquid culture, a lonA mutant had an average of 10 ± 2 hooks per micrometer of cell length, above the theoretical threshold, and phenocopied swarmer cells of the wild type (Fig. 5A and SI Appendix, Fig. S5C). We conclude that swarmer cells are hyperflagellate, having twice the flagellar density of swimmer cells and that LonA inhibits hyperflagellation in liquid environments.

Fig. 5.

Hyperflagellation is required for swarming. (A and B) Graphical representation of flagellar number as determined by counting the number of stained FlgE^T123C puncta per cell length using high-resolution structured illumination microscopy (3D-SIM) and Imaris image analysis software. FlgE is the structural component of the flagellar hook and introduction of a cysteine at position 123 enables fluorescent labeling with a cysteine-reactive maleimide stain without impairing function (39). Each cell is presented as a dot on the scatterplot. Average of cell length and spot number is indicated as a larger open circle with SDs that emerge as horizontal and vertical lines. Hypothetical threshold flagellar density required to swarm is indicated as a diagonal dashed line (black). Fifty cells of each type were analyzed (raw data presented in SI Appendix, Table S4). For A, green closed circles indicate individual swimmer cells of lonA (DK611), blue closed circles indicate individual swimmer cells of wild type (DS7673), and red closed circles indicate individual swarmer cells of wild type (DS7673). For B, individual cell data from flagellar hook count at various IPTG concentrations, added to liquid cultures of the strain DK30 (ΔflgE amyE:P_fla/che-flgE^T123C cat P_fla/cheΩP_IPTG-fla/che operon kan), are presented as a scatterplot and color coded to match the IPTG concentration indicated on the graph. (C) Quantitative swarm expansion assays for strain DK30 (ΔflgE amyE:P_fla/che-flgE^T123C cat P_fla/che ΩP_IPTG-fla/che operon) induced with increasing concentrations of IPTG (swarm graph color coded to match the IPTG concentration). Each point is the average of three replicates.

Swarming motility has been correlated with an increase in flagellar density, but the requirement for hyperflagellation is controversial, as it has not been determined whether hyperflagellation is a consequence of, or prerequisite for, swarming motility (25, 40, 41). To test whether hyperflagellation is a strict requirement for swarming, a strain was generated that replaced the SwrA controlled promoter of the fla/che operon with an artificial IPTG-inducible promoter (P_IPTG-fla/che) (37, 38). Cells of the P_IPTG-fla/che strain were separately incubated in liquid broth (for counting fluorescently labeled hooks) and on solid surfaces (for assaying swarming motility) in the presence of various concentrations of IPTG. Hook number increased with increasing IPTG concentration and swarming motility was abruptly restored once flagellar density reached an average of 10 ± 2 hooks per micrometer of cell length (Fig. 5 B and C). Importantly the flagellar density required to swarm over solid surfaces in the artificial system was similar to that observed in both wild-type swarmer cells and swimmer cells mutated for LonA (SI Appendix, Fig. S5C). We conclude that an increase in flagellar density above a critical threshold is required for swarming motility. We further conclude that surface contact induces hyperflagellation by relief of SmiA-mediated Lon proteolysis of the SwrA master regulator.

Discussion

Bacteria appear to sense contact with a solid surface and transduce surface contact information to alter their physiology (34). Surface sensing is poorly understood but seems to integrate information from cell surface appendages such as the adhesion of pili or the impedance of flagellar rotation (28, 30, 40, 42, 43). Transduction of the surface signal is also poorly understood and the signal transduction components that have been discovered may be species specific (34, 43, 45, 46). In B. subtilis, the mechanism of surface recognition is unknown, but here we demonstrate that signal transduction is mediated by Lon-dependent proteolysis of a master regulator and the physiological output is an increase in flagellar density. Lon protease is highly conserved, has been implicated in the regulation of surface-contact behaviors, and in particular, has been shown to restrict hyperflagellation in phylogenetically diverse bacteria (18–22). Thus, the involvement of Lon-dependent regulatory proteolysis in surface contact signal transduction may be generalizable.

Bacterial regulatory proteolysis often involves adaptor proteins, defined here as one protein that is essential for the proteolysis of another, and to the best of our knowledge, SmiA is the first adaptor protein reported for the Lon family of proteases (6, 7). Consistent with an adaptor function, SmiA facilitates Lon-dependent proteolysis of SwrA both in vivo and in vitro. Unlike known allosteric activators of Lon, such as unfolded protein substrates, SmiA does not simply enhance general proteolytic activity, as the presence of SmiA did not accelerate the LonA-dependent degradation of α-casein or titin (13, 16, 44). Further, SmiA did not function as a generalized unfolding chaperone, as SmiA did not promote SwrA proteolysis by the heterologous Lon protease from E. coli. Combined, we infer that SmiA acts as an adaptor to specifically bridge the folded SwrA protein with the LonA protease from B. subtilis. Consistent with a functional relationship, SmiA and SwrA co-occur in the same set of closely related genomes from the genus Bacillus (SI Appendix, Fig. S6). Due to their inherent substrate specificity, adaptors differ widely in both structure and sequence, and we infer that other phylogenetically unrelated adaptors of Lon protease await discovery (11).

SwrA is the master regulator of flagellar biosynthesis in B. subtilis and is required both for swarming motility and for increasing flagellar number (36–38). Here we show that the two activities of SwrA are explicitly related. We demonstrate that SmiA adaptor-mediated Lon proteolysis restricts SwrA levels in liquid environments and keeps liquid grown cells below a critical flagellar density of seven flagella per micrometer of cell length that is required to swarm under standard laboratory conditions (Fig. 6). When cells contact a solid surface, we infer the SwrA proteolysis is abrogated, SwrA accumulates, and flagella are synthesized in excess of the threshold density (Fig. 6). Thus, the lag period represents a time for increased flagellar biosynthesis, and we note that the swarm lag period is on par with the duration necessary to synthesize new flagella (38). The mechanism by which surface contact inhibits the LonA/SmiA proteolytic module is unknown but could perhaps be related to mechanosensing mediated by impedance of flagellar rotation or strain on the peptidoglycan cell wall. Proteolysis is an expedient and efficient means of governing a rapid change in response to environmental conditions where transduction of the surface contact signal into adaptor activation could occur by a variety of mechanisms (such as release of a sequestered antiadaptor for SmiA).

Fig. 6.

Model for SmiA-dependent LonA proteolysis of SwrA. SmiA adaptor targets SwrA, the master regulator of flagellar biosynthesis, for degradation by LonA protease in WT swimmer cells resulting in low levels of SwrA protein and low flagellar density. Upon contact with a surface, however, SmiA-dependent LonA proteolysis of SwrA is inhibited and SwrA protein levels accumulate, resulting in hyperflagellation and swarming motility. T-bar indicates inhibition. SwrA substrate is red hexagon, SmiA adaptor is violet circle, LonA protease is orange cylinder, cell is light pink, and flagella are green.

The Lon protease is conserved in all domains of life (11, 15). Eukaryotic Lon (37% identical to LonA of B. subtilis) (SI Appendix, Fig. S7) is involved in both the general degradation of misfolded proteins and in the specific remodeling of the mitochondrial proteins cytochrome C oxidase COX4-1, transcription factor A (TFAM), and the cholesterol regulator StAR (47–50). It has been suggested that Lon could be a viable anticancer drug target as drugs that inhibit Lon activity lead to cancer cell apoptosis in vitro either due to uncontrolled accumulation of misfolded proteins in general or due to the failure to control a specific regulator (48, 51, 52). General inhibitors of Lon activity, however, could cause potentially problematic pleiotropic side effects, and Lon defects have been associated with poor mitochondrial health and multiple human diseases (47, 53, 54). Thus, specific inhibitors that block proteolysis of key oncogenic substrates would be preferable and the SmiA precedent offers the potential for pathway-specific antiadaptor therapeutics.

Experimental Procedures

Strain construction, growth conditions, and detailed experimental procedures are described in SI Appendix, Extended Experimental Procedures. Strains, plasmids, and primers used in this study are listed in SI Appendix, Tables S1–S3, respectively.

Swarm Expansion Assay.

Cells were grown to midlog phase at 37 °C in lysogeny broth (LB) and resuspended to 10 OD₆₀₀ in pH 8.0 PBS buffer (137 mM NaCl, 2.7 mM KCl, 10 mM Na₂HPO₄, and 2 mM KH₂PO₄) containing 0.5% India ink (Higgins). Freshly prepared LB containing 0.7% bacto agar (25 mL per plate) was dried for 20 min in a laminar flow hood, centrally inoculated with 10 µL of the cell suspension, dried for another 10 min, and incubated at 37 °C. The India ink demarks the origin of the colony and the swarm radius was measured relative to the origin every 30 min. For consistency, an axis was drawn on the back of the plate and swarm radii measurements were taken along this transect. For experiments including IPTG, cells were propagated in broth in the presence of IPTG, and IPTG was included in the swarm agar plates.

Microscopy.

For counting flagellar number and determining cell length, the OMX 3D-SIM Super-Resolution system at Indiana University Bloomington Light Microscopy Imaging Center was used. Images were captured by Photometrics Cascade II EMCCD camera, and processed by SoftWoRx imaging software (Applied Precision). SI Appendix, Extended Experimental Procedures provides details of hook labeling, microscopy, and image analyses.

In Vivo Protein Turnover Assay.

The 25-mL cultures of B. subtilis WT and mutant strains were grown at 37 °C with shaking until they reached midexponential phase (OD₆₀₀ 0.5–0.8) and chloramphenicol was added to a final concentration of 10 µg/mL. One milliliter was immediately harvested as T = 0, and subsequent 1-mL samples were harvested at 30, 60, 120, 180, and 240 min. Samples were pelleted by centrifugation at 18,000 × g, the supernatant was removed, and pellets were frozen at −20 °C. Next, the samples were analyzed by Western blot. SI Appendix, Extended Experimental Procedures provides details of Western blotting and anti-SwrA antibody preparation.

In Vitro Proteolysis Assay.

SwrA proteolytic degradation was assayed at 37 °C in LonA degradation buffer [25 mM Tris (pH 8.0), 100 mM KCl, 10 mM MgCl2, and 1 mM DTT] with 75 μg/mL creatine kinase, 15 mM creatine phosphate, and 4 mM ATP and concentrations of LonA as monomer, GST-SwrA, SmiA, and α-casein as indicated in the figure legends. Samples were withdrawn at appropriate time points, quenched with 6× SDS loading dye, separated by 15% SDS/PAGE, and proteins were detected by Coomassie Brilliant Blue staining. SI Appendix, Extended Experimental Procedures provides details of GST-SwrA, His-SUMO-SmiA, and LonA protein purifications.