CwlQ Is Required for Swarming Motility but Not Flagellar Assembly in Bacillus subtilis

Bacteria are surrounded by a wall of peptidoglycan and early work in Bacillus subtilis was the first to suggest that bacteria needed to enzymatically remodel the wall to permit insertion of the flagellum. No PG remodeling enzyme alone or in combination, however, has been found to be essential for flagellar assembly in B. subtilis.

ABSTRACT

Lytic enzymes play an essential role in the remodeling of bacterial peptidoglycan (PG), an extracellular mesh-like structure that retains the membrane in the context of high internal osmotic pressure. Peptidoglycan must be unfailingly stable to preserve cell integrity, but must also be dynamically remodeled for the cell to grow, divide, and insert macromolecular machines. The flagellum is one such macromolecular machine that transits the PG, and flagellar insertion is aided by localized activity of a dedicated PG lyase in Gram-negative bacteria. To date, there is no known dedicated lyase in Gram-positive bacteria for the insertion of flagella. Here, we take a reverse-genetic candidate-gene approach and find that cells mutated for the lytic transglycosylase CwlQ exhibit a severe defect in flagellum-dependent swarming motility. We further show that CwlQ is expressed by the motility sigma factor SigD and is secreted by the type III secretion system housed inside the flagellum. Nonetheless, cells with mutations of CwlQ remain proficient for flagellar biosynthesis even when mutated in combination with four other lyases related to motility (LytC, LytD, LytF, and CwlO). The PG lyase (or lyases) essential for flagellar synthesis in B. subtilis, if any, remains unknown.

IMPORTANCE Bacteria are surrounded by a wall of peptidoglycan and early work in Bacillus subtilis was the first to suggest that bacteria needed to enzymatically remodel the wall to permit insertion of the flagellum. No PG remodeling enzyme alone or in combination, however, has been found to be essential for flagellar assembly in B. subtilis. Here, we take a reverse-genetic candidate-gene approach and find that the PG lytic transglycosylase CwlQ is required for swarming motility. Subsequent characterization determined that while CwlQ was coexpressed with motility genes and is secreted by the flagellar secretion apparatus, it was not required for flagellar synthesis. The PG lyase needed for flagellar assembly in B. subtilis remains unknown.

INTRODUCTION

Most bacteria are surrounded by an extracellular cell wall that prevents catastrophic hyperexpansion of the membrane by the high internal osmotic pressure of the cytoplasm. The wall is a semielastic macromolecular mesh of peptidoglycan (PG) comprised of long polymers of an N-acetylglucosamine-N-acetyl-muramic acid disaccharide that are cross-linked by amino acid side chains (1, 2). While the chemistry of the glycan component is common, bacteria differ in the nature of the peptide side chains and the organization of PG with respect to the overall architecture of the envelope (3). In cells with a Gram-negative envelope, the PG is only 1 to 3 layers thick and lies between the plasma and outer cell membrane, whereas cells with a Gram-positive envelope have a much thicker PG wall and lack an outer membrane (4–7). Regardless of the type of envelope, the semielastic PG network must be both stable and continuous to maintain cell integrity but be dynamically remodeled to allow for cell growth, cell division, and the insertion of transenvelope nanomachines.

One nanomachine that is inserted through the peptidoglycan is the propeller-like flagellum that bacteria rotate to swim in liquid or swarm over solid surfaces. Flagella are constructed from over 30 different proteins that are tightly regulated to ensure stoichiometry and sequential assembly (8, 9). The first architectural unit of the flagellum to be assembled is the basal body that is inserted into the plasma membrane and houses a dedicated type III secretion system (10, 11). Once activated, the type III secretion system secretes the distal components of the flagellum, including the structural units of the axle-like rod that is polymerized until it reaches the outer membrane in Gram-negative bacteria, followed by the flexible universal-joint-like hook (12–17). Hook synthesis terminates when it reaches a particular length, at which point the secretion system transitions to exporting subunits that form the long helical polymer of the filament (18, 19). Thus, flagella are constructed from the inside out and must not only cross all layers of the envelope, but freely rotate within them.

The PG is thought to present a structural barrier to flagellar construction at the level of the rod (20–22). The rod is the part of the flagellum that spans the PG, and the rod’s diameter of 8 to 13 nm (10, 23–26) seems incompatibly wide relative to the estimated PG pore size of 2 to 7 nm (7, 27, 28). The first evidence that PG remodeling was required for flagellar assembly came from the Gram-positive bacterium B. subtilis, in which particular mutants were simultaneously defective in the expression of multiple PG remodeling lyase “autolysins” and flagellar biosynthesis (29, 30). Later, the role of PG remodeling was supported in the Gram-negative bacteria Salmonella enterica, Rhodobacter sphaeroides, and Caulobacter crescentus, when mutants defective in a particular PG lyase were shown to be defective in both motility and flagellar assembly (31–35). In each case, the mutant was defective in FlgJ, a flagellar rod-cap chaperone that is either fused to, or directly interacts with, a PG lyase (31, 33, 34, 36–38). Remarkably, despite the foundational report, the specific autolysin required for flagellar assembly is not known in B. subtilis, and the original mutants that supported the remodeling hypothesis were in fact defective in flagellar synthesis, which in turn inhibited expression of the PG lyases (39–42). Nonetheless, the rod in B. subtilis must penetrate a PG that is approximately 50 nm thicker than that of Gram-negative bacteria and how this is achieved is unknown.

A variety of different enzymes act to cleave various bonds in the PG chemical structure and B. subtilis encodes over 30 annotated lyases from different enzymatic families in its genome (43, 44). Here, we take a reverse-genetic approach to mutate and screen known and putative lyases in an effort to find genes required for flagellar assembly (43, 44). Of the candidates tested, mutation of two PG-degrading enzymes, the vegetative endopeptidase CwlO (45, 46) and the poorly understood lytic-transglycosylase CwlQ (47), exhibited a moderate and severe defect in swarming motility, respectively. Seemingly consistent with being a lyase involved in flagellar assembly, CwlQ required its active site residue for swarming motility, was expressed by the motility sigma factor SigD, and was secreted by the type III secretion system within the flagellum. Inconsistent with being required for flagellar assembly, however, motility was restored to the cwlQ mutant when motility agar concentration was decreased below that of standard swarming conditions. Moreover, cells mutated for CwlQ could both swim and synthesize flagella. Our work suggests that although CwlQ is not required for insertion of the flagella through the PG, it is conditionally required for motility and may play a role in flagella function specifically on harder surface environments. Finally, cells remained proficient for flagellar assembly even in a quintuple mutant that disrupted all known SigD-dependent autolysins and CwlO. The PG lyase required for flagellar assembly in B. subtilis, if any, remains unknown.

RESULTS

CwlQ is conditionally required for swarming motility.

B. subtilis is predicted to encode a variety of peptidoglycan lyases, some of which have been biologically and/or biochemically demonstrated to cleave peptidoglycan, and some of which have a predicted function based on sequence homology (43, 44) (Table 1). A reverse-genetic approach was taken to determine which, if any, of the peptidoglycan lyase candidates were required for flagellar assembly in B. subtilis. To narrow the pool of candidates, lyases and putative lyases were excluded if they had been previously tested for flagellar biogenesis (48), if they were expressed only during sporulation, or if they were encoded within horizontally transferred genetic elements (e.g., prophages). The remaining candidate genes were mutated, and the resulting mutants were tested for flagellum-dependent swarming motility (Table 1). Most of the mutants were wild type for swarming behavior and were discarded from further study (see Fig. S1 in the supplemental material). Cells mutated for either CwlQ (Fig. 1A) or CwlO (Fig. 1B), however, exhibited more severe swarming defects. Moreover, the phenotypes of neither the cwlQ nor cwlO mutants were due to polar effects on neighboring genes, as swarming motility was complemented to wild type when the gene was cloned with 500 bp of upstream DNA (in the case of cwlQ) (Fig. 1A) or expressed from an IPTG-inducible construct (in the case of cwlO) (49) (Fig. 1B) and inserted at an ectopic locus in the respective mutant. We conclude that CwlO and CwlQ are required for swarming motility under standard conditions. CwlO encodes the vegetative endopeptidase required for cell elongation (45, 46, 49), and we thus focused our study on the lytic transglycosylase CwlQ because it conferred a more severe swarming defect and its biological function was poorly understood (47).

TABLE 1.

B. subtilis candidate PG lyase genes

Gene	Product annotation	Swarminga	Excluded (reference[s])
Not tested
blyA	Muramidase	NT	HTE: SPβ prophage (86)
cwlA	Muramidase	NT	HTE: skin element (87)
cwlC	Muramidase	NT	Sporulation (88, 89)
cwlD	Muramidase	NT	Sporulation (σE, σG) (90)
cwlH	Muramidase	NT	Sporulation (σK) (91)
cwlP	Muramidase/endopeptidase	NT	HTE: SPβ prophage (92)
cwlT	Muramidase/endopeptidase	NT	HTE: ICE element (93, 94)
lytH	Endopeptidase	NT	Sporulation (σK) (95)
spoIID	Lytic transglycosidase	NT	Sporulation (σE) (96, 97)
spoIIP	Muramidase/endopeptidase	NT	Sporulation (σE) (97, 98)
xlyA	Muramidase	NT	HTE: PBSX prophage (99)
xlyB	Muramidase (putative)	NT	HTE: PBSX prophage
Wild-type swarming
cwlK	l,d-Endopeptidase	+++	100
lytD	N-Acetylglucosaminidase	+++	Previously tested (48, 101, 102)
lytE	Endopeptidase	+++	103–105
lytF	Endopeptidase	+++	Previously tested (15, 106, 107)
lytG	N-Acetylglucosaminidase	+++	108
cwlS	Endopeptidase	+++	103, 109
yocH	Muramidase	+++	110
yqgA	Wall-associated protein	+++	111
yqgT	Endopeptidase (putative)	+++
yqiI	Muramidase (putative	+++
yrvJ	Muramidase (putative)	+++
Swarming defect
cwlO	Endopeptidase	+	45, 46
cwlQ	Muramidase/transglycosylase	−	47
lytC	Muramidase	+	Previously tested 3, 31, 48

^aNT, not tested; +++, wild-type swarming motility; +, reduced swarming motility; −, severely defective swarming motility.

FIG 1.

CwlQ is required for swarming motility. (A) Quantitative swarm expansion assay of wild type (WT; DK1042), cwlQ (DK1744), cwlQ (cwlQ) (DK3586), and cwlQ (cwlQ^E148A) (DK4127). (B) Quantitative swarm expansion assay of wild type (DK1042), cwlO (DK8462), and cwlO (P_spank-cwlO) in the presence of 1 mM IPTG [cwlO (cwlO); DK8842]. (C) Quantitative swarm expansion assay indicating the swarm radius after 4 h of incubation on medium fortified with the indicated amount of agar for the following strains: wild type (DK1042), cwlQ (DK1744), cwlQ (cwlQ) (DK3586), and cwlQ (cwlQ^E148A) (DK4127). (D) Quantitative swarm expansion assay indicating the swarm radius after 4 h of incubation on medium fortified with the indicated amount of agar for the following strains: wild type (DK1042), swrA (DS2415), and swrD (DS6657). (E) Quantitative swarm expansion assay of wild type (DK1042), smiA (DK5720), cwlQ (DK1744), and cwlQ smiA (DK9007). (F) Quantitative swarm expansion assay of cwlQ mutant ectopically complemented with the indicated versions of cwlQ: wild type (DK3586), Δ36–63 (DK8472), Δ2–63 (DK8470) and Δ2–35 (DK8471). All data points are the average from three replicates.

Swarming motility requires flagella and one way in which CwlQ could promote swarming is by remodeling the peptidoglycan to facilitate flagellar assembly (50). To determine whether the cwlQ mutant exhibited a flagellar assembly defect, the cwlQ gene was mutated in a strain that encoded a variant of the flagellin protein that could be fluorescently labeled with a maleimide dye (Hag^T209C) (51). After staining, the cwlQ mutant was found to be proficient for flagellar filament assembly but appeared to have a qualitative reduction in filament number relative to wild type (Fig. 2A). Precise counting of filaments in B. subtilis is difficult, but filament number can be indirectly assessed by counting the number of flagellar hooks and basal bodies as proxies. Thus, to explore whether the cwlQ mutant exhibited a defect in flagellar number, the cwlQ gene was mutated in a strain that either encoded a variant of the hook protein that could be fluorescently labeled with a maleimide stain (FlgE^T123C) or a green fluorescent protein (GFP) fusion to the flagellar basal body protein FliM (13, 52). In these backgrounds, the flagellar hooks (Fig. 2B) and basal bodies (Fig. 2C) appeared as fluorescent dots, and 3D-structured illumination microscopy was used to count the number of each in both wild type and the cwlQ mutant. Quantitative analysis indicated there was a subtle but statistically significant reduction (Student’s t test, P value < 0.00003) in the number of flagellar hooks and basal bodies in the cwlQ mutant (Fig. 2D; Table S1). A 2-fold increase in flagellar density on surfaces has been shown to be critical for swarming motility in B. subtilis (53) and the inability of the cwlQ mutant to swarm might be related to the slight reduction in flagellar hook number observed in liquid culture.

FIG 2.

Cells mutated for CwlQ have a slight but statistically significant reduction in the number of flagellar hooks and basal bodies. (A) Fluorescence micrographs of cells of the indicated genotype stained for membranes (false-colored red) and flagellar filaments (false-colored green). The following strains were used to generate this panel: DS1919 (WT) and DK1770 (cwlQ). (B) Fluorescence micrographs of cells of the indicated genotype stained for membranes (false-colored red) and flagellar hooks (false-colored green). The following strains were used to generate this panel: DS7673 (WT) and DK1771 (cwlQ). (C) Fluorescence micrographs of cells of the indicated genotype stained for membranes (false-colored red) and flagellar basal bodies (FliM-GFP, false-colored green). The following strains were used to generate this panel: DS8521 (WT) and DK1047 (cwlQ). (D) Scatterplots in which individual wild-type (red) and cwlQ mutant (blue) cells were measured by OMX 3D-SIM for cell length and the number of flagellar hooks (left) and flagellar basal bodies (right) were counted with Imaris software. Thirty cells were measured per experiment and each cell is represented by a different dot on the graph. Averages and standard deviations are colored according to the data set to which they belong. Raw data are included Table S3.

Another way that the absence of CwlQ might give rise to a swarming defect is if the mutant synthesized flagella but those flagella were defective for rotation. To determine whether the flagella of a cwlQ mutant were functional for rotation, cells were centrally inoculated on LB medium fortified with 0.3% agar in which the pores in the agar were sufficiently large to permit swimming motility. As B. subtilis preferentially migrates over surfaces, cells were discouraged from sliding surface migration by using a background that was mutated for both surfactant and extracellular polysaccharide biosynthesis (50, 54–58). The wild-type strain created a large zone of colonization after 12 h of incubation, whereas a mutant defective in the flagellar filament protein Hag grew as a tight central colony (Fig. 3). Cells mutated for CwlQ produced a zone of colonization similar to that of the wild type (Fig. 3). Moreover, cells of the cwlQ mutant were vigorously motile when grown to exponential phase in liquid medium and observed by wet mount microscopy. We conclude that cells mutated for CwlQ are not only proficient in flagellar assembly but are also proficient for swimming.

FIG 3.

Cells with mutated CwlQ are proficient for swimming motility. LB agar petri plates fortified with 0.3% agar were centrally inoculated, incubated at 37°C for 12 h, and filmed against a black background such that zones of colonization appear white and uncolonized agar appears black. Each strain contained the indicated alleles plus mutants in srfAA (or srfAC) and epsH to discourage movement across the surface and force cells to swim through the agar. The following strains were used to generate the figure: DK374 (WT), DK378, (hag), DK2491 (cwlQ), and DK7051 [cwlQ (cwlQ^E148A)].

As the cwlQ mutant exhibited wild-type swimming motility in liquid, wild-type colonization of 0.3% agar, and only a slight reduction in flagellar hook number, we wondered whether the swarming defect was dependent on the hardness of the agar surface. For the wild type, swarm radius was inversely proportional to agar concentration after 4 h of incubation and swarming was fully inhibited on medium solidified with 0.9% agar (Fig. 1C). Swarming of the cwlQ mutant was fully inhibited after 4 h at the standard conditions of 0.7% agar, but swarm radius increased with decreasing agar concentration such that the mutant swarmed like the wild type on medium solidified with 0.5% agar (Fig. 1C). Swarming rescue at substandard agar concentrations appeared to be specific to the cwlQ mutant, as at least two other mutants defective in swarming motility, swrA (defective due to reduced flagellar number) (52, 59, 60) and swrD (defective due to reduced flagellar torque) (61), remained nonswarming at all agar concentrations tested (Fig. 1D). We conclude that the requirement for CwlQ differs from that of other swarming motility mutants, as it is conditional and relieved when agar concentrations are reduced below standard conditions.

To determine how CwlQ might promote swarming, we sought to isolate spontaneous suppressors that restored motility to a cwlQ mutant upon prolonged incubation on a swarm agar plate. Unlike regulatory mutants defective in swarming (15, 48, 57, 61, 62), no spontaneous swarming-proficient suppressor ever emerged as a flare from the nonmotile colony of the cwlQ mutant, even after 48 h of incubation. A lack of spontaneous suppression could indicate either that CwlQ plays a structural/functional role unrelated to flagellar assembly, or that CwlQ might promote swarming in a multifactorial manner such that a single mutation could not restore motility. To directly test the hypothesis that flagellar number was limiting in the cwlQ mutant, a double mutant was generated that was simultaneously defective in cwlQ and smiA encoding SmiA, a specific adaptor protein for the regulatory proteolysis of the master flagellar activator protein, SwrA (53). Flagellar number and swarming increases when smiA is mutated (53), and the cwlQ smiA double mutant exhibited enhanced swarming motility relative to the cwlQ mutant alone (Fig. 1E). We infer that the cwlQ mutant is modestly defective in flagellar number, and swarming motility can be improved either by reducing surface hardness or by increasing flagellar number through mutation of SmiA.

CwlQ is secreted and swarming requires the CwlQ active site.

To further explore the mechanism of CwlQ, the CwlQ primary sequence was analyzed. CwlQ is predicted to have two domains: (i) an N-terminal domain of unknown function and (ii) a C-terminal lytic transglycosylase domain previously shown to require a conserved glutamate for peptidoglycan lyase activity (47, 63, 64) (Fig. 4A). We note that the original publication that characterized CwlQ purified only the C-terminal domain and consequently defined CwlQ based on the truncation, leading to some confusion in gene annotation and nomenclature (47). To determine whether lytic transglycosylase activity was required for swarming, the conserved glutamate active site residue E148 (formerly “E85” [47]) was replaced with an alanine (cwlQ^E148A) in the complementation construct and inserted at an ectopic locus (amyE::P_cwlQ-cwlQ^E148A) in a cwlQ mutant background. Introduction of the active site mutant allele conferred a defect in swarming motility that was more severe than the cwlQ mutant alone (Fig. 1A). Consistent with an enhanced defect, the strain that expressed the cwlQ^E148A allele exhibited reduced swarm expansion relative to the cwlQ null mutant at all agar concentrations tested (Fig. 1C). The enhanced defect appeared to be specific for surface migration however, as the cwlQ^E148A mutant exhibited swimming motility like the wild type (Fig. 3). We conclude that CwlQ requires the lytic transglycosylase active site to promote swarming motility and that the presence of CwlQ may become inhibitory when the active site is abrogated.

FIG 4.

CwlQ is secreted by the flagellar type III secretion system and destroyed by extracellular proteases. (A) Schematic diagram of the 244-amino acid CwlQ protein primary sequence from N terminus (left) to C terminus (right). Important amino acid residue numbers are indicated above the diagram and the domain of unknown function (DUF) and lytic transglycosylase domain are indicated below. (B to E) Western blot analysis of cell pellets and TCA-precipitated supernatants resolved by SDS-PAGE and probed either with anti-CwlQ or anti-SigA antibodies. (B) WT (DK1042), cwlQ (DK1744), cwlQ (cwlQ) (DK3586), and cwlQ (cwlQ^E148A) (DK4127). (C) Δ7 (DS6329), Δ7 cwlQ (cwlQ) (DK8697), and Δ7 cwlQ (cwlQ^E148A) (DK8698). (D) Δ7 fliF (DS6871), Δ7 flgM (DS7160), and Δ7 fliF flgM (DK5150). (E) cwlQ (cwlQ^Δ2-63) (DK8699), cwlQ (cwlQ^Δ2-35) (DK8700), and cwlQ (cwlQ^Δ36-63) (DK8701).

As lytic transglycosylases operate on the extracellular substrate peptidoglycan, we hypothesized that CwlQ was secreted from the cytoplasm. In order to determine if CwlQ was secreted, cell lysates and trichloroacetic acid (TCA)-precipitated supernatants were probed with anti-CwlQ and anti-SigA antibody in Western blot analysis (Fig. 4B). The cytoplasmic housekeeping sigma factor SigA was used as a loading control for the cytoplasmic fraction, and its absence in the supernatant indicated that protein release by spontaneous cell lysis was likely minimal (Fig. 4B). CwlQ was present in cell lysates of wild type, absent in the cwlQ mutant, and was restored in the CwlQ complementation strain (Fig. 4B). Moreover, CwlQ protein was also detected when CwlQ^E148A was expressed in an otherwise cwlQ mutant background, suggesting that the active site mutant was not defective due to inherent protein instability (Fig. 4B). There was no indication of extracellular CwlQ in the strains tested (Fig. 4B).

A failure to detect extracellular CwlQ could either indicate that CwlQ was not secreted and functioned in the cytoplasm, or that it was secreted and subsequently degraded by extracellular proteases, as has been shown for other flagellum-related proteins in B. subtilis (17, 65). To determine whether secreted proteases contributed to extracellular CwlQ degradation, pellets and TCA-precipitated supernatants were harvested, resolved, and subjected to Western blot analysis in a variety of strains deleted for seven extracellular proteases (Δ7) (65). When the seven extracellular proteases were absent, the CwlQ protein was found in the both the pellet and TCA-precipitated supernatant fraction of the otherwise wild type, the cwlQ mutant complemented with the wild-type allele, and the cwlQ mutant complemented with the cwlQ^E148A allele (Fig. 4C). We conclude that CwlQ is a secreted protein that cannot normally be detected in the supernatant due to extracellular degradation by one or more of the proteases secreted by B. subtilis.

The primary sequence of CwlQ does not contain an N-terminal signal sequence consistent with either SEC-dependent or TAT-dependent secretion (66–69). One way in which CwlQ could be secreted in a manner independent of a readily apparent signal sequence is if CwlQ was secreted by the type III secretion system that resides at the core of the flagellum (70–72). To determine whether CwlQ was secreted by the flagellar type III system, cells were mutated for the basal body protein FliF, a protein previously shown to be essential for flagellum-mediated secretion (65). CwlQ protein was detected in the cytoplasm but not the supernatant of the fliF Δ7 mutant (Fig. 4D). The absence of CwlQ from the supernatant could be consistent with a failure of secretion, but many genes related to flagellar motility are under the regulation of the alternative sigma factor SigD, and SigD activity is inhibited by the anti-sigma factor FlgM when fliF is mutated (65, 73–76). Moreover, the total amount of CwlQ protein appeared to be reduced in the fliF Δ7 mutant, perhaps consistent with an expression defect (Fig. 4D). Thus, if cwlQ was expressed as part of the flagellar regulon, it would be difficult to distinguish whether the absence of CwlQ protein in the supernatant in a fliF mutant was either due to a failure of secretion or a failure of expression, or both.

To determine whether the expression of cwlQ was impaired in a fliF mutant, a transcriptional reporter construct was generated in which the cwlQ promoter region (P_cwlQ) was fused to the lacZ gene encoding β-galactosidase and inserted at an ectopic site (amyE::P_cwlQ-lacZ). Mutation of fliF reduced the expression of the P_cwlQ-lacZ reporter 10-fold (Fig. 5). Consistent with the fliF defect, expression of cwlQ was found to be SigD-dependent and cwlQ expression was abolished when SigD was mutated (Fig. 5). Moreover, mutation of flgM increased P_cwlQ expression above that of the wild type, and P_cwlQ expression was restored to the fliF mutant when flgM was also disrupted (Fig. 5). We conclude that cwlQ is a SigD-dependent gene and the lack of extracellular CwlQ in the absence of FliF may have been due, at least in part, to a protein expression failure. To determine whether CwlQ was secreted in the fliF mutant without the confounding expression defect, Western blot analyses were conducted on cells mutated for flgM and a fliF flgM double mutant in the Δ7 background. CwlQ secretion was dramatically reduced in the fliF flgM double mutant relative to the flgM single mutant alone (Fig. 4D). We note that the cytoplasmic control SigA protein also appeared in the supernatant of the fliF flgM double mutant and thus the small amount of CwlQ that appeared to be secreted may have been spuriously released by cellular lysis. We conclude that CwlQ is primarily, and likely exclusively, secreted in a manner that depends on FliF and the flagellar type III secretion system.

FIG 5.

The cwlQ gene is expressed by RNA polymerase and the alternative sigma factor, SigD. β-Galactosidase activities of a transcriptional fusion of the promoter of cwlQ to the lacZ gene encoding β-galactosidase (P_cwlQ-lacZ) in the indicated genetic backgrounds and expressed in Miller units (MU). Each bar is the average of three replicates and standard deviations are provided. The following strains were used to generate the figure: DK2185 (WT), DK8864 (fliF), DK2207 (sigD), DK2347 (flgM), and DK8665 (fliF flgM).

The signal sequence that directs proteins to be secreted by type III secretion system is poorly understood but appears to be contained within the N terminus of a secreted protein (70–72). To determine whether the N-terminal domain of CwlQ was required for secretion, three separate in-frame markerless deletions were generated that separately deleted amino acids 2 to 63 (deleting the entire N terminus), and 2 to 35 and 36 to 63 (deleting the first and second halves of the N terminus, respectively) in the cwlQ complementation construct (Fig. 4A). CwlQ^Δ2–63 and CwlQ^Δ2–35 strains both displayed a defect in swarming motility comparable to that of the cwlQ null mutant when the mutations were introduced to a strain deleted for the native copy of cwlQ, but the CwlQ^Δ36–63 strain swarmed like the wild type albeit with an extended lag period (Fig. 1F). When CwlQ^Δ2–63 was expressed in a Δ7 strain deleted for extracellular proteases, no protein was detected, suggesting that the deletion of the entire N-terminal domain caused severe defects in protein stability (Fig. 4E). CwlQ^Δ2–35 was detected in the cell pellet but not the supernatant, suggesting that the N terminus of CwlQ was required for secretion (Fig. 4E). Finally, CwlQ^Δ36–63 hyperaccumulated in the cytoplasm, with a reduced level of secretion that might account for the delayed rescue of swarming to the cwlQ mutant (Fig. 4E). We conclude that the N-terminal domain of CwlQ is important for its secretion and protein stability.

CwlQ and CwlO are not synergistically required for flagellar assembly.

Cells mutated for CwlQ alone exhibited a conditional defect in swarming motility and were proficient in flagellar assembly. One reason that a peptidoglycan lyase mutant might fail to have a flagellar assembly defect is if other coexpressed peptidoglycan lyases acted redundantly in supporting the activity. Since CwlQ is a fourth secreted peptidoglycan lyase expressed as part of the SigD-regulon (48), a quadruple mutant was generated defective in cwlQ lytC lytD lytF in a background that expressed a version of the flagellar filament that could be fluorescently labeled. The quadruple mutant was still proficient for flagellar synthesis (Fig. 6A). Next, a cwlO mutation was introduced to the cwlQ lytC lytD lytF quadruple mutant and the resulting quintuple mutant showed increased loss of cell integrity (by an increase in frequency of cytoplasmic staining with the maleimide dye), defects in cell shape, and reduced flagellar filament number (Fig. 6B). The cwlO mutant alone did not exhibit a severe reduction in flagellar filament number (Fig. 6C), and neither did a cwlO cwlQ double mutant (Fig. 6D) nor a cwlO lytC double mutant (Fig. 6E). We conclude that neither CwlQ nor CwlO are essential for flagellar assembly and that mutation of as many as five peptidoglycan lyases is required to reduce, but not abolish, flagellar synthesis and/or retention.

FIG 6.

A strain simultaneously mutated for cwlQ, cwlO, and three other PG lyases is proficient for flagellar biosynthesis. Fluorescence micrographs of cells of the indicated genotypes stained for membranes (false-colored red) and flagellar filaments (false-colored green). The following strains were used to generate this figure: DK7570 (cwlQ lytC lytD lytF) (A), DK8815 (cwlQ cwlO lytC lytD lytF) (B), DK8816 (cwlO) (C), DK8783 (cwlQ cwlO) (D), and DK8787 (cwlO lytC) (E).

DISCUSSION

Flagella are elaborate molecular nanomachines, which when assembled span the plasma membrane and all layers of the cell envelope, including the peptidoglycan (PG) wall. The macromolecular structure of PG is porous with pores approximately 2 to 7 nm in diameter (10, 23–26), and peptidoglycan remodeling by dedicated enzymes is thought necessary to allow insertion of the 8 to 13 nm flagellar rod (7, 27, 28). Gram-negative bacteria often encode a PG-degrading enzyme that is either fused to, or associated with, a cap chaperone that permits rod polymerization such that as the rod extends toward the PG, the lyase degrades the wall and permits rod passage (31, 33, 34). Gram-positive bacteria are surrounded by PG many layers thick but, to date, no PG lyase (or associated rod cap chaperone) has been found to be required for flagellar assembly. Here, we took a reverse-genetic, candidate-gene approach to identify PG-remodeling enzymes that might be required for flagellar assembly in B. subtilis and found that mutation of two hydrolases, CwlO and CwlQ, resulted in defects in swarming motility. Neither mutant was defective in flagellar assembly, however, either when mutated singly or when mutated in combination with four other PG-degrading enzymes. To better understand the role of PG lyases in promoting motility, we focused on the less-understood CwlQ.

Cells mutated for CwlQ were defective for swarming motility under standard conditions but the reason for the defect is unclear. CwlQ might play a structural or functional role unrelated to flagellar assembly per se and we note that a related lytic transglycosylase was reported to be required for flagellar function in Helicobacter pylori (77). Perhaps consistent with a structural role, there appeared to be a qualitative reduction in flagellar filaments and we often saw evidence of filaments dissociated from the cell body in fluorescent micrographs, even in liquid-grown culture, of the cwlQ mutant. Perhaps the absence of CwlQ creates a local environment in the peptidoglycan that promotes fracture/instability of the flagellum. Additionally, the cwlQ mutant exhibited a statistically significant reduction in both the number of flagellar basal bodies and hooks, and swarming was somewhat improved by additional mutation of SmiA, a protein that restricts flagellar number. How or why basal body number might be reduced in the absence of CwlQ is unclear, however, as basal body assembly is thought to precede interaction with the peptidoglycan. Finally, swarming could be restored to the cwlQ mutant simply by reducing the agar concentration of the surface. Previous work indicated that the transition to swarming motility in B. subtilis requires that cells exceed a threshold flagellar density (53) and cells mutated for CwlQ, when introduced to a surface, may be below that level for 0.7% agar, but above a reduced level needed for softer environments.

While the mechanism by which CwlQ promotes swarming is unknown, we here make a number of observations that connect the function of CwlQ to flagellar motility. First, a conserved and biochemically determined active site residue for CwlQ lytic transglycosylase activity (47) was required to promote swarming and, when mutated, caused an even more severe defect than deletion of entire gene. Perhaps unproductive binding of the CwlQ active site mutant to PG occludes the action of other enzymes, but we note that even the enhanced inhibition does not prevent flagellar assembly (see Fig. S2 in the supplemental material). Second, CwlQ was found to be part of the flagellar regulon under strict control of the flagellar sigma factor SigD, and was thus coexpressed with, among other proteins, the structural subunits of the distal rod and flagellar filament. Third, CwlQ was secreted in a manner dependent on the type III secretion system housed within the flagellum, and its export was directed by information encoded within the poorly conserved N-terminal domain. Thus, CwlQ promoted swarming motility as a lytic transglycosylase, was coexpressed with flagellar structural subunits and was secreted by the flagellum. Every observation above makes CwlQ seem an ideal candidate to be a PG-remodeling enzyme for flagellar assembly, and yet, flagella were synthesized in its absence.

One reason, often invoked, for why mutation of a PG lyase fails to confer phenotype is the idea of redundancy, that the genome encodes another peptidoglycan lyase with similar activity or that otherwise compensates for the absence. At the most basic level, all PG lyases are redundant as they all operate on the same substrate, peptidoglycan, but what distinguishes them is which bond they cleave in the macromolecular mesh and how their activity is restricted in both space and time. Possible candidates for redundant activity with CwlQ are the LytC, LytD, and LytF lyases with which it is coexpressed and the CwlO PG lyase that promotes cell elongation and is required for full swarming motility. Simultaneous mutation of all five hydrolases reduced, but failed to abolish, flagellar production. The quintuple mutant also showed signs of reduced cellular integrity in the form of frequently misshapen and lysed cells, and thus the effect of the quintuple deletion may be less specific for flagellar synthesis and more an indication of generalized envelope damage. Whatever the case, the identity of one or more specific PG lyases required for flagellar transit of the wall during assembly in B. subtilis remains unknown.

MATERIALS AND METHODS

Strains and growth conditions.

B. subtilis strains were grown in Luria-Bertani (LB) (10 g tryptone, 5 g yeast extract, 5 g NaCl per liter) broth or on LB plates fortified with 1.5% Bacto agar at 37°C. When appropriate, antibiotics were included at the following concentrations: 10 μg/ml tetracycline, 100 μg/ml spectinomycin, 5 μg/ml chloramphenicol, 5 μg/ml kanamycin, and 1 μg/ml erythromycin plus 25 μg/ml lincomycin (mls). Isopropyl-β-d-thiogalactopyranoside (IPTG; Sigma) was added to the medium at the indicated concentration when appropriate.

Strain construction.

All constructs were either first introduced by transformation by natural competence into DK1042 (a competent derivative of strain 3610) (78), or transformed into the domesticated strain PY79 and transduced in 3610 using SPP1-mediated generalized phage transduction (79). All strains used in this study are listed in Table 2. All primers used in this study are listed in Table S2 in the supplemental material. All plasmids used in this study are listed in Table S3.

TABLE 2.

Strains

Strain	Genotype
DK374	srfAC::Tn10 spec epsH::tet
DK378	Δhag srfAC::Tn10 spec epsH::tet
DK1023	cwlS::tet
DK1024	lytE::kan
DK1042	wild type
DK1047	ΔfliM cwlQ::kan amyE::Pfla/che-fliM-GFP spec
DK1128	lytG::kan
DK1744	cwlQ::kan
DK1747	cwlQ::mls
DK1770	Δhag cwlQ::kan amyE::Phag-hagT209C spec
DK1771	ΔflgE cwlQ::kan amyE::Pfla/che-flgET123C cat
DK2185	amyE::PcwlQ-lacZ cat
DK2207	sigD::tet amyE::PcwlQ-lacZ cat
DK2347	flgM::tet amyE::PcwlQ-lacZ cat
DK2491	ΔcwlQ srfAC::Tn10 spec epsH::tet
DK3579	yrvJ::kan
DK3586	cwlQ::kan amyE::PcwlQ-cwlQ spec
DK4127	cwlQ::kan amyE::PcwlQ-cwlQE148A spec
DK4677	yocH::kan
DK4678	yqgT::kan
DK5150	Δ7 ΔfliF ΔflgM
DK5178	cwlK::spec
DK5720	ΔsmiA
DK5824	hagT209C amyE::PcwlQ-cwlQE148A spec cwlQ::kan
DK6432	yqiI::kan
DK6611	yqgA::spec
DK7051	cwlQ::kan srfAA::mls ΔepsH amyE::PcwlQ-cwlQE148A spec
DK7570	ΔcwlQ ΔlytC ΔlytD ΔlytF amyE::Phag-hagT209C spec
DK8462	cwlO::kan
DK8470	cwlQ::kan amyE::PcwlQ-cwlQΔ2-63 spec
DK8471	cwlQ::kan amyE::PcwlQ-cwlQΔ2-35 spec
DK8472	cwlQ::kan amyE::PcwlQ-cwlQΔ36-63 spec
DK8664	ΔfliF amyE::PcwlQ-lacZ cat
DK8665	ΔfliF flgM::tet amyE::PcwlQ-lacZ cat
DK8697	Δ7 cwlQ::kan amyE::PcwlQ-cwlQ spec
DK8698	Δ7 cwlQ::kan amyE::PcwlQ-cwlQE148A spec
DK8699	Δ7 cwlQ::kan amyE::PcwlQ-cwlQΔ2-63 spec
DK8700	Δ7 cwlQ::kan amyE::PcwlQ-cwlQΔ2-35 spec
DK8701	Δ7 cwlQ::kan amyE::PcwlQ-cwlQΔ35-63 spec
DK8783	ΔcwlQ ΔcwlO amyE::Phag-hagT209C spec
DK8787	ΔcwlO ΔlytC amyE::Phag-hagT209C spec
DK8815	ΔcwlQ ΔcwlO ΔlytC ΔlytD ΔlytF amyE::Phag-hagT209C spec
DK8816	ΔcwlO amyE::Phag-hagT209C spec
DK8842	cwlO::kan ycgO::Pspank-5′UTR-cwlO spec
DK9007	ΔsmiA cwlQ::kan
DS1919	Δhag amyE::Phag-hagT209C spec
DS2415	ΔswrA
DS6329	Δ7 (Δmpr ΔaprE ΔnprE Δbpr Δvpr Δepr ΔwprA)
DS6657	ΔswrD
DS6871	Δ7 ΔfliF
DS7160	Δ7 ΔflgM
DS7673	ΔflgE amyE::Pfla/che-flgET123C cat
DS8521	ΔfliM amyE::Pfla/che-fliM-GFP spec

Antibiotic resistance cassette insertion/deletion mutations.

For each gene mutated, a PCR fragment was amplified upstream of the gene and downstream of the gene using the indicated primer pairs. Next, the antibiotic-resistance cassette was amplified from either pDG1515 (for tetracycline resistance), pDG780 (for kanamycin resistance), or pAH52 (for macrolide, lincomycin, streptomycin “mls” resistance) using primer pair 3250/3251 (80, 81). The three products were assembled by Gibson isothermal assembly (ITA) (82) and transformed into DK1042 by natural transformation. Finally, colonies containing the antibiotic resistance marker replacement mutant were determined by PCR amplification over the top of the allele with the far upstream and far downstream primers used to generate the corresponding arms of adjacent DNA. The following primers were used to generate the indicated mutants: cwlQ (3695/3696::3697/3698); cwlS (3699/3700::3701/3702); lytE (3670/3671::3672/3673); lytG (3707/3708::3709/3710); yqgT (5497/5498::5499/5500); yocH (5501/5502::5503/5504); yqgA (5112/5113::5114/5115); yqiL (6425/6426::6427/6428); and yrvJ (4618/4619::4620/4621).

ΔcwlQ in-frame markerless deletion.

To generate the ΔcwlQ in-frame markerless deletion construct, the region upstream of cwlQ was PCR amplified using the primer pair 4118/4120 and the region downstream of cwlQ was PCR amplified using the primer pair 4119/4121. The two fragments were combined with SalI-digested pMiniMAD, which carries a temperature-sensitive origin of replication and an erythromycin-resistance cassette (83), and were assembled by ITA to generate plasmid pSS5. The pSS5 plasmid was passaged though the recA⁺ Escherichia coli strain TG1 before being transformed into DK1042 and selecting for mls resistance at the nonpermissive temperature for plasmid replication, 37°C. To evict the plasmid, the strain was incubated in 3 ml LB broth at a permissive temperature for plasmid replication (22°C) for 14 h, diluted 30-fold in fresh LB broth, then serially diluted and plated on LB agar at 37°C. Individual colonies were patched onto LB plates and LB plates containing mls to identify mls-sensitive colonies that had evicted the plasmid. Chromosomal DNA from colonies that had excised the plasmid was purified and screened by PCR using primers 4118/4121 to determine which isolate had retained the ΔcwlQ allele.

ΔcwlO markerless deletion.

A cwlO::kan mutant allele generated from high-throughput directed mutagenesis (84) was requested from the Bacillus Genetic Stock Center (The Ohio State University, Columbus OH). The kanamycin resistance cassette is flanked by lox recombination sites and was excised by transformation with plasmid pDR244 encoding the cre recombinase and a spectinomcyin resistance cassette by plating on LB containing spectinomycin at 30°C. Colonies were restreaked on LB, grown at 37°C, and deletion of cwlO was determined by PCR product length polymorphism using primers 4741/4744.

cwlQ complementation constructs.

To generate the amyE::P_cwlQ-cwlQ complementation construct (pSS9), a PCR product containing the cwlQ coding region plus 393 bp of upstream sequence was amplified from B. subtilis 3610 chromosomal DNA using the primer pair 4373/4374, digested with BamHI and EcoRI, and cloned into the BamHI and EcoRI sites of pAH25 containing a polylinker and spectinomycin resistance cassette between two arms of the amyE gene (generous gift of Amy Camp, Mount Hoyloke College).

The active site mutant amyE::P_cwlQ-cwlQ^E148A allele construct was generated using a modified ITA protocol. Briefly, the region upstream of the complementation construct of cwlQ (DK3586) was PCR amplified using the primer pair 953/4888 and the region downstream of the complementation strain was PCR amplified using the primer pair 4887/954. The two fragments were assembled by isothermal assembly and retransformed into B. subtilis, selecting for spectinomycin resistance. The N-terminal cwlQ deletion constructs were built using a similar approach with the indicated primer pair sets: amyE::P_cwlQ-cwlQ^Δ2-63 (953/7304::7303/954), amyE::P_cwlQ-cwlQ^Δ2-35 (953/7389::7388/954), and amyE::P_cwlQ-cwlQ^Δ36-63 (953/7387::7386/954).

P_cwlQ-lacZ reporter construct.

To generate the P_cwlQ-lacZ reporter construct pSS2, the P_cwlQ promoter was amplified from B. subtilis 3610 chromosomal DNA using the primers 4114/4115, digested with EcoRI and BamHI, and cloned into the EcoRI and BamHI sites of plasmid pDG268, which carries a chloramphenicol resistance marker and a polylinker upstream of the lacZ gene between two arms of the amyE gene (85).

CwlQ-His expression construct and CwlQ purification.

To generate the CwlQ-6×His expression plasmid pSS14, the cwlQ gene was PCR amplified from B. subtilis 3610 chromosomal DNA with primers 4757/4758, digested with EcoRI and HindIII, and cloned into the EcoRI and HindIII sites of pET21a (Novagen). Next, pSS14 was transformed into Rosetta gami E. coli, grown to an optical density at 600 nm (OD₆₀₀) of 0.6 in 1 liter of LB broth, induced with 1 mM IPTG, and grown for 16 h at 16°C. Cells were pelleted and resuspended at room temperature in lysis buffer (50 mM Tris [pH 8.0], 300 mM NaCl, 10% glycerol) and treated with lysozyme and DNase I, and the lysis was carried out using a pressurized cell homogenizer. The lysed cells were centrifuged at 15,000 rpm for 30 min. The cleared supernatants were combined with Ni-NTA resin (Novagen) and immediately poured onto a 1-cm separation column (Bio-Rad); the resin was allowed to pack and was washed with lysis buffer. CwlQ-6×His bound to the resin was then eluted with elution buffer (50 mM Tris [pH 8.0], 300 mM NaCl, 10% glycerol, 100 mM imidazole). The elution fractions were then run on SDS-PAGE gels and appropriate fractions were pooled and concentrated to 2 ml. The final purification of CwlQ-6×His protein was conducted via size exclusion chromatography on a Superdex 75 16/60 (GE Healthcare) column using gel filtration buffer (25 mM Tris [pH 8.0], 300 mM NaCl, 10% glycerol) and submitted to Cocalico (Stephens, PA) for injection into rabbits and polyclonal antibody generation.

Isothermal assembly reaction.

First, a 5× ITA stock mixture was generated (500 mM Tris-HCl [pH 7.5], 50 mM MgCl₂, 50 mM dithiothreitol [DTT; Bio-Rad], 31.25 mM PEG-8000 [Fisher Scientific], 5.02 mM NAD [Sigma-Aldrich], and 1 mM each dNTP [New England BioLabs]), aliquoted, and stored at −80°C. An assembly master mixture was made by combining prepared 5× isothermal assembly reaction buffer (131 mM Tris-HCl, 13.1 mM MgCl₂, 13.1 mM DTT, 8.21 mM polyethylene glycol 8000, 1.32 mM NAD, and 0.26 mM each dNTP) with Phusion DNA polymerase (New England BioLabs) (0.033 unit/μl), T5 exonuclease diluted 1:5 with 5× reaction buffer (New England BioLabs) (0.01 unit/μl), Taq DNA ligase (New England BioLabs) (5,328 units/μl), and additional dNTPs (267 μM). The master mix was aliquoted to 15 μl and stored at −80°C. DNA fragments were combined at equimolar amounts to a total volume of 5 μl and added to a 15-μl aliquot of prepared master mix. The reaction mixture was incubated for 60 min at 50°C.

SPP1 phage transduction.

To 0.1 ml of dense culture grown in TY broth (LB broth supplemented after autoclaving with 10 mM MgSO₄ and 100 μM MnSO₄), serial dilutions of SPP1 phage stock were added and statically incubated for 15 min at 37°C. To each mixture, 3 ml TYSA (molten TY supplemented with 0.5% agar) was added, poured atop fresh TY plates, and incubated at 37°C overnight. Top agar from the plate containing near-confluent plaques was harvested by scraping into a 50-ml conical tube, vortexed, and centrifuged at 5,000 × g for 10 min. The supernatant was treated with 25 μg/ml DNase final concentration before being passed through a 0.45-μm syringe filter and stored at 4°C. Recipient cells were grown to stationary phase in 2 ml TY broth at 37°C. An aliquot of 0.9 ml of cells was mixed with 5 μl of SPP1 donor phage stock. TY broth (9 ml) was added to the mixture and allowed to stand at 37°C for 30 min. The transduction mixture was then centrifuged at 5,000 × g for 10 min, the supernatant was discarded, and the pellet was resuspended in the remaining volume. Cell suspension (100 μl) was then plated on TY fortified with 1.5% agar, the appropriate antibiotic, and 10 mM sodium citrate.

Motility assays.

For the swarm expansion assay, cells were grown to mid-log phase at 37°C in LB broth and resuspended to 10 OD₆₀₀ in pH 8.0 phosphate-buffered saline (PBS) (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/plate) (for percent agar concentration assay, the freshly prepared LB contained between 0.5% to 0.9%) was dried for 10 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. For consistency, an axis was drawn on the back of the plate and swarm radii measurements were taken along this transect.

For swim assays, cells were grown to mid-log phase at 37°C in LB broth 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) and 10 μl of culture was inoculated into the agar. Freshly prepared LB containing 0.3% Bacto agar (25 ml/plate) was dried for 10 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. Plates were visualized with a Bio-Rad Geldoc system and digitally captured using Bio-Rad Quantity One software.

Western blotting.

B. subtilis strains were grown in LB broth to an OD₆₀₀ of ∼0.5 and then 10 ml was harvested by centrifugation and resuspended to 100 OD₆₀₀ in lysis buffer (20 mM Tris [pH 7.0], 10 mM EDTA, 1 mg/ml lysozyme, 10 μg/ml DNase I, 100 μg/ml RNase I, and 1 mM phenylmethylsulfonyl fluoride [PMSF]) and incubated 30 min at 37°C. Each lysate was then mixed with the appropriate amount of 6× SDS loading dye to dilute the loading dye to 1× concentration. Samples were separated by 12% sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE). The proteins were electroblotted onto nitrocellulose and developed with a 1:1,000 dilution (anti-CwlQ) or 1:80,000 dilution (anti-SigA) of primary antibody and a 1:10,000 dilution of secondary antibody (horseradish peroxidase-conjugated goat anti-rabbit immunoglobulin G). The immunoblots were developed using the Immun-Star HRP developer kit (Bio-Rad).

For experiments involving trichloroacetic acid (TCA) precipitation, the 10 ml of supernatant was saved during the pelleting step, combined with 1 ml of 0.015% sodium deoxycholate, vortexed, and incubated 10 min at room temperature. Next, 500 μl of ice cold TCA was added, the mixture was vortexed and incubated on ice for 2 h. The supernatant was precipitated by centrifugation at 30,000 × g for 10 min at 4°C. The pellet was resuspended in 1 ml ice-cold acetone and repelleted in a tabletop centrifuge. Finally, the pellet was resuspended in the same amount of 1× protein sample buffer as the corresponding pellet.

β-Galactosidase assays.

Cells were grown to approximately OD₆₀₀ 0.7 to 1.3 in LB medium in triplicate. An aliquot of 1 ml of each sample was harvested and resuspended in 1 ml Z-Buffer (40 mM HaH₂PO₄, 60 mM Na₂HPO₄, 1 mM MgSO₄, 10 mM KCl, and 38 mM β-mercaptoethanol). Lysozyme was added to each sample to a final concentration of 0.2 mg/ml, and samples were incubated for 15 min at 30°C and thereafter kept on ice. Each sample was diluted appropriately to a final volume of 500 μl in Z buffer, and the reaction was started with 100 μl of 4 mg/ml O-nitrophenyl-β-d-galactopyranoside (in Z buffer) and stopped with 250 μl of 1 M Na₂CO₃. The OD₄₂₀ of the reaction mixtures was recorded and the β-galactosidase activity was calculated with the following formula: {OD₄₂₀/[time (min) × OD₆₀₀]} × dilution factor × 1,000. All reactions were stopped prior to saturation of yellow color (A₄₂₀ < 1.2). For those reactions with low to no β-galactosidase activity, the reaction was run for a maximum of 1 h before stopping.

Microscopy.

Fluorescence microscopy was performed with a Nikon 80i microscope along with a phase contrast objective Nikon Plan Apo 100× and an Excite 120 metal halide lamp. Alexa Fluor 594 C₅ maleimide fluorescent signals were visualized with a C-FL HYQ Texas Red filter cube (excitation filter 532 to 587 nm, barrier filter >590 nm). Green fluorescent protein (GFP) was visualized using a C-FL HYQ FITC filter cube (fluorescein isothiocyanate [FITC], excitation filter 460 to 500 nm, barrier filter 515 to 550 nm). Yellow fluorescent protein (YFP) was visualized using a C_FL HYQ YFP filter cube (excitation filter 490 to 510 nm, barrier filter 515 to 550 nm). TMA-DPH fluorescent signal was visualized using a UB-2E/C DAPI filter cube (excitation filter 340 to 380 nm, barrier filter 435 to 485 nm). Images were captured with a Photometrics Coolsnap HQ² camera in black and white, false-colored, and superimposed using Metamorph image software.

For superresolution microscopy using structured illumination, the OMX 3D-SIM Super Resolution system at Indiana University Bloomington Light Microscopy Imaging Center was used. Superresolution microscopy was performed using a 1.4-numerical-aperture Olympus 100× oil objective. FM4-64 was visualized using laser line 561 nm and emission filter 609 to 654 nm, and Alexa Fluor 488 was visualized using laser line 488 nm and emission filter 500 to 550 nm. Images were captured by Photometrics Cascade II EMCCD camera and processed by SoftWoRx imaging software (Applied Precision). For counting hooks, images reconstructed with SoftWoRx were used in Imaris (Bitplane) to determine the number of FlgE^T123C foci on the surface of each cell. The spots feature labeled each FlgE^T123C foci by the search parameter of identifying spots of 1 μM in the 488 wavelength and we verified by eye that the spots labeling identified bona fide foci on the cell surface. The hook count data sets were plotted against cell length (micrometers) using IMARIS software.