FlhF Is Required for Swimming and Swarming in Pseudomonas aeruginosa

Abstract

FlhF is a signal recognition particle-like protein present in monotrichous bacteria. The loss of FlhF in various bacteria results in decreased transcription of class II, III, or IV flagellar genes, leads to diminished or absent motility, and results in the assembly of flagella at nonpolar locations on the cell surface. In this work, we demonstrate that the loss of FlhF results in defective swimming and swarming motility of Pseudomonas aeruginosa. The FlhF protein localizes to the flagellar pole; in the absence of FlhF, flagellar assembly occurs but is no longer restricted to the pole. ΔflhF bacteria swim at lower velocities than wild-type bacteria in liquid media and can no longer swarm when assayed under standard swarming conditions (0.5% agar). However, ΔflhF bacteria regain swarming behavior when plated on 0.3% agar. ΔflhF organisms show decreased transcription and expression of flagellin (FliC) both in liquid media and on swarming plates compared to wild-type bacteria. However, changes in flagellin expression do not explain the different motility patterns observed for ΔflhF bacteria. Instead, the aberrant placement of flagella in ΔflhF bacteria may reduce their ability to move this rod-shaped organism effectively.

Pseudomonas aeruginosa is a motile, opportunistic pathogen responsible for numerous acute and chronic infections in humans (22). During log-phase growth, P. aeruginosa has a single, polar flagellum that drives swimming motility. P. aeruginosa also exhibit swarming motility on soft agar in the presence of specific carbon and nitrogen sources (17, 26). Swarming is a community behavior that promotes flagellum-dependent motility on surfaces. In many bacteria, including those that use polar flagella during swimming, swarming is associated with the production of multiple, lateral flagella (2, 13). However, swarmer cells of P. aeruginosa appear to possess on average two polar flagella and no lateral flagella (17, 26). P. aeruginosa can also use type IV pili to move along solid surfaces by a process called twitching motility (23). However type IV pili, unlike flagella, do not appear to be required for swarming motility (26).

Transcription of flagellar genes in P. aeruginosa proceeds in a highly regulated fashion (10). FleQ is the highest-level positive regulator identified in P. aeruginosa; its synthesis is σ⁷⁰ dependent and is inhibited by Vfr, a homolog of Escherichia coli CRP (8). FleQ activity is posttranslationally inhibited by a direct interaction with FleN (9). Loss of FleN results in increased numbers of polar flagella in P. aeruginosa (7). Class II genes, which include those encoding the motor, basal body, and export apparatus, require both FleQ and RpoN for activation. Expression of class III genes, which encode hook proteins, requires FleR, FleQ, and RpoN, while class IV genes, which include fliC (encoding flagellin), are transcribed only after the activation of FliA.

flhF is a class II gene whose product is required for polar flagellar placement in Pseudomonas putida (25). Overexpression of FlhF in P. putida results in increased numbers of polar flagella, the same phenotype as that observed with the loss of FleN (7, 25). A current model for flagellar biogenesis in P. aeruginosa proposes that FlhF is initially present at the cell pole and directs flagellar assembly to that location (10). The regulation of flagellar expression during P. aeruginosa swarming motility is poorly understood. Because swarmer cells increase the number of polar flagella that they possess from one to two, we speculated that FlhF, FleN, and/or FleQ might play a role in this upregulation. In this study, we investigated the role of FlhF in P. aeruginosa swimming and swarming motility.

MATERIALS AND METHODS

Bacterial strains, media, and culture conditions.

The strains and plasmids used in this study are listed in Table 1. The DNA primers used for PCR amplification and strain construction are listed in Table 2. These sequences were derived from the available online PAO1 genome sequence database (http://www.pseudomonas.com) (31). Bacteria were cultured and propagated in Luria broth (LB) (10.0 g of tryptone, 5.0 g of yeast extract, 10.0 g of NaCl per liter), in M8 media with 0.2% glucose and 0.05% sodium glutamate, or on Vogel-Bonner minimal medium agar plates (17, 27, 35). PA103 and PAK strains were maintained as stocks in LB plus 15% glycerol (vol/vol) at −80°C. All P. aeruginosa strains used in this study were constructed from the wild-type parental strain PAK. Antibiotics were added to liquid and solid media as appropriate at the following concentrations: E. coli, 100 μg/ml ampicillin, 15 μg/ml gentamicin, and 20 μg/ml tetracycline; P. aeruginosa, 200 μg/ml carbenicillin, 100 μg/ml gentamicin, and 100 μg/ml tetracycline.

TABLE 1.

Bacterial strains and plasmids used in this study

Strain or plasmid	Descriptiona	Source or reference
E. coli strains
XL1-Blue	recA1 endA1 gyrA96 thi-1 hsdR17 supE44 relA1 lac [F′ proAB lacIqZΔM15 Tn10 (Tcr)]	Stratagene
XL2-Blue	recA1 endA1 gyrA96 thi-1 hsdR17 supE44 relA1 lac [F′ proAB lacIqZΔM15 Tn10 (Tcr) Amy Cmr]	Stratagene
S17-1	Used for mating constructs into P. aeruginosa; thi pro hsdR recA RP4-2 (Tc::Mu) (Km::Tn7)	30
P. aeruginosa strains
PA103	Virulent lung isolate of P. aeruginosa	21
PAK	Wild-type isolate	J. Mattick
PAK ΔflhF	In-frame deletion of amino acids 7-430 of FlhF	This work
PAK pFlhF-tdimer2	FlhF-tdimer2 expressed from the lac promoter of the high-copy pUCP-SK plasmid; Cbr	This work
PAK ptdimer2	Tdimer2 expressed from the lac promoter of the high-copy pUCP-SK plasmid; Cbr	This work
PAK fliC-lux	Transcriptional fliC-lux fusion integrated at the attB site	This work
ΔflhF fliC-lux strain	Transcriptional fliC-lux fusion integrated at the attB site in ΔflhF background	This work
ΔfliC strain	Tetracycline cassette disrupting fliC; Tcr	J. Mattick
PAK zwf-lux	Transcriptional zwf-lux fusion integrated at the attB site; Tcr	This work
ΔflhF zwf-lux strain	Transcriptional zwf-lux fusion integrated at the attB site in ΔflhF background; Tcr	This work
Plasmids
pGEM-T	Cloning vector; Apr	Promega
pGEM-T FliCN	Cloning vector containing DNA fragment with fliC promoter region; Apr	This work
pEX18 Gm	Allelic replacement suicide plasmid; GmrsacB oriT	14
pEX18 flhF	Plasmid with DNA fragments upstream and downstream of flhF used to generate gene deletion	This work
pUCP-SK	P. aeruginosa expression vector; Apr (Cbr)	38
pBS-tdimer2	Cloning vector with tdimer2; Apr	C. Jacobs-Wagner
ptdimer2	Tdimer2 expressed from the lac promoter of pUCP-SK	This work
pFlhF-tdimer2	FlhF-timer2 fusion protein expressed from the lac promoter of pUCP-SK; Apr (Cbr)	This work
mini-CTX-lux	Vector used for construction of transcriptional fusions with the luxCDABE operon from Xenorhabdus luminescens; contains attP site for integration at the attB site of P. aeruginosa chromosome; Tcr	3
mini-CTX-fliC-lux	Vector for introduction of fliC-lux into the attB site; Tcr	This work
mini CTX zwf-lux	Vector for introduction of zwf-lux into the attB site; Tcr	20
pFLP2	Source of inducible Flp recombinase; Apr (Cbr)	14

^aAp, ampicillin; Cb, carbenicillin; Gm, gentamicin; Tc, tetracycline.

TABLE 2.

Primers used in this study

Primer	Sequence (5′ to 3′)a
FlhF N1	GCGAATTCCCAGGTATTCGGCAGCCTCAACGGC
FlhF N2	GACGAGCTCGAAGCGTTTGACTTGCATTGCCG
FlhF C1	GCGAGCTCGTGACGGAAGCGAAGGTGGCTTCC
FlhF C2	GCAAGCTTGTTGGACCAGTCGTTCGACGAAGAACTCC
FliC N1	CAATTGGTGGACTGGGTGTTCTTCG
FliC N2	GGATCCGCGGCCGCGGTCAGACGCTGCAACGAGG
FlhF N3	GACGAATTCATGCAAGTCAAACGCTTCTTCGCC
FlhF C3	GACGAATTCGCCGGCACGCCGCGCCGGTTGCTGGTAGAG
RplU-RT-PCR-1	CGCAGTGATTGTTACCGGTG
RplU-RT-PCR-2	AGGCCTGAATGCCGGTGATC
FliA-RT-PCR-1	CGCGACGCTAAAGATCACGAAGTT
FliA-RT-PCR-2	TGGTGTCTTCCGGCAATCCATGTT

^aRestriction sites are underlined.

Molecular cloning and strain construction.

Plasmid purification, restriction endonuclease analysis, ligation, and transformation were all carried out using standard techniques (27). E. coli strains XL1 and XL2 (Stratagene) were used to propagate plasmids.

An unmarked, in-frame deletion of flhF was constructed in PAK via allelic exchange (14). DNA immediately upstream and downstream of flhF was amplified from PA103 genomic DNA with Taq polymerase (Invitrogen), using primers FlhF N1 and FlhF N2 (upstream) and FlhF C1 and FlhF C2 (downstream). Amplification products were cloned in tandem into the sacB-containing pEX18-Gm^R plasmid (14) generating pEX18flhF. This plasmid was transformed into E. coli S17-1 and was mobilized into P. aeruginosa by mating. Selection of double recombinants was carried out on sucrose plates as previously described, and mutant constructs were confirmed by PCR and Southern blot analysis (20).

The promoter region of fliC was amplified from PAO1 genomic DNA, using Taq polymerase with primers FliC N1 and FliC N2. The product was subcloned from pGEM into mini-CTX-lux (15) to generate a transcriptional reporter and mobilized into P. aeruginosa strains by mating. After integration into the attB site, vector backbone sequences were excised by Flp recombinase as previously described (14).

Full-length FlhF lacking its stop codon was amplified from PA103 genomic DNA, using Pfu Turbo polymerase and primers FlhF N3 and FlhF C3. The amplification product was cloned into pBS-KS tdimer2 (kindly provided by C. Jacobs-Wagner) to construct a fusion protein consisting of FlhF in frame with a tandem of two dimeric variants of DsRed (tdimer2) (5). The FlhF-tdimer2 fusion sequence was subcloned into pUCP-SK, allowing for constitutive expression in P. aeruginosa from the plasmid lac promoter. Tdimer2 was also subcloned into pUCP-SK under the control of the lac promoter; this construct served as a negative control for subcellular localization studies.

Isolation of RNA, preparation of cDNA, and quantitative real-time PCR.

Total RNA was prepared from exponential-phase cultures (optical density at 600 nm [OD₆₀₀] of 1.0) of PAK and ΔflhF bacteria and reverse transcribed into cDNA as previously described (19). Quantitative real-time PCR was carried out as detailed by Laskowski et al. (19): one microliter of each cDNA sample was amplified with 250 nM each of primer sets fliA-RT-PCR-1 and fliA-RT-PCR-2 and rplU-RT-PCR-1 and rplU-RT-PCR-2. The latter primer set amplifies the 50S ribosomal protein RplU and serves as a control for the amount of total cDNA present in each reaction. Reactions were carried out in a DNA Engine Opticon2 continuous fluorescence detection system (Bio-Rad) and analyzed using Opticon Monitor 3.0 software (Bio-Rad). The absence of DNA contamination of RNA samples was confirmed by the absence of a PCR amplification product when RNA, but not cDNA, was used as a template in these reactions (data not shown).

Motility assays.

Swimming assays were done by stabbing 0.3% LB agar plates with a single colony of PAK or ΔflhF bacteria. The swimming zone was measured after overnight incubation at 30°C. Swarming assays were done as described previously, using M8 medium plates supplemented with 0.2% glucose and 0.05% sodium glutamate and containing either 0.3% or 0.5% agar (17). Swarming plates were inoculated with a single colony picked from an LB plate, incubated at either 37°C (0.5%) or 30°C (0.3%) overnight, and then incubated at room temperature for an additional 48 h. Plates were photographed with an Image Station 2000R (Kodak).

Measurement of fliC transcription.

Wild-type and ΔflhF strains with transcriptional fliC-lux fusions were monitored for luciferase activity. One-milliliter samples from bacterial cultures grown in LB liquid media at 37°C with aeration were harvested at 30- to 60-min intervals. Luminescence (relative light units) was measured with a TD 20/20 luminometer (Turner Designs) and normalized to culture density (OD₆₀₀). To determine fliC activity during swarming, the above strains were plated onto M8 swarming plates. Luminescence was measured at 48 h using an Image Station 2000R (Kodak) and quantified with Kodak 1D software (v3.6.2). At least four areas at the swarming edge were measured for each colony. To control for the effects of cell density on luminescence measurements from swarming plates, luminescence from a zwf-lux transcriptional reporter was also measured directly from bacteria growing on swarming plates (20). zwf encodes the housekeeping protein glucose-6-phosphate dehydrogenase. Each experiment was performed in triplicate.

Measurement of FliC expression.

Bacteria were harvested at either log phase (OD₆₀₀ of 0.7) or stationary phase (OD₆₀₀ of 2.0) from shaking liquid cultures or picked directly from 0.3% or 0.5% M8 swarming plates at 48 to 72 h postinoculation. In the latter instances, several 20-μl drops of distilled water were placed at the colony edge and the bacteria were gently harvested. Cells were lysed in 4% sodium dodecyl sulfate (SDS), and the protein concentrations of whole-cell lysates were determined by a bicinchoninic acid assay (Pierce). Ten micrograms of total protein per sample was separated by SDS-polyacrylamide gel electrophoresis (PAGE) and transferred to a polyvinylidene difluoride membrane (Immobilon). Membranes were probed with anti-FliC polyclonal antiserum (1:750; kindly provided by R. Ramphal), followed by horseradish peroxidase-conjugated goat anti-rabbit immunoglobulin G (1:2,000; Bio-Rad). Membrane blocking, washes, and visualization of bound antibody by enhanced chemiluminescence were carried out as previously described (4). Signals were detected using an Image Station 2000R (Kodak) and quantified with Kodak 1D software (v 3.6.2).

Transmission electron microscopy.

Bacteria were grown overnight in 3 ml LB media and then subcultured (1:30) into fresh LB media for 4 h. Bacterial suspensions were placed directly on glow-discharged carbon-coated grids, stained with 2.0% uranyl acetate for 2 min, washed twice with water, and air dried. Bacteria were visualized using a Tecnai 12 Biotwin electron microscope (Center for Cell Imaging, Yale University).

Epifluorescence microscopy.

P. aeruginosa strains expressing either tdimer2 or FlhF-tdimer2 were grown overnight at 37°C in LB, diluted 1:20 to fresh media the next morning, and harvested at an OD₆₀₀ of 0.4. Cells were pelleted by centrifugation (2,500 × g for 5 min), fixed in 4% paraformaldehyde for 30 min at room temperature, washed with phosphate-buffered saline (PBS), and then blocked with PBS plus 0.7% (wt/vol) fish scale gelatin (PBS-FSG) (Sigma) for 30 min at room temperature. Flagella were stained with anti-FliC polyclonal antiserum (1:100 in PBS-FSG; kindly provided by R. Ramphal) and detected with Alexa 488-conjugated goat anti-rabbit immunoglobulin G (1:1,000 in PBS-FSG). Stained bacteria were visualized using a Nikon Eclipse TS100 microscope equipped with fluorescein isothiocyanate and Texas Red filters and a monochrome Spot camera (Diagnostic Instruments) running Spot 4.0.1 software. Images were pseudocolored and compiled using Photoshop 7.0 (Adobe). The placement of flagella (polar versus lateral) was determined by examining 50 to 100 bacteria for each strain in which the point of flagellar attachment to the cell body was unambiguous. The poles were defined as the areas at the cell ends over which cell wall curvature could be appreciated and corresponded to ca. 25% of the total cell length. Samples were scored independently by both authors, and interrater agreement appeared to be very good, as indicated by kappa coefficients of 0.702 to 0.832.

Video microscopy.

The swimming behaviors of wild-type and ΔflhF bacteria were examined in both M8 and LB liquid media containing polyvinylpyrrolidone (PVP; Sigma) at final concentrations of 0 to 6% (wt/vol). Bacteria were observed by brightfield microscopy with a Nikon Eclipse TS100 microscope using a ×60 oil immersion objective and 1- to 2-s video clips obtained with a Kodak EasyShare DX6340 3.1 megapixel camera. These QuickTime videos were analyzed with ImageJ 1.36b software (NIH) using the manual tracking plug-in. For each strain/condition, the movement of 25 bacteria (tracked on the center of the cell) was followed for 30 consecutive images (1 second). Mean velocity was calculated for each cell.

Measurement of surface flagella and rhamnolipid production.

Surface flagella were harvested as previously described (33). P. aeruginosa strains were grown at 37°C in 1 liter LB, harvested by centrifugation, resuspended in 50 mM sodium phosphate (pH 7.0) and 10 mM magnesium chloride buffer, and blended for 20 seconds in a Waring blender. The samples were centrifuged at 10,000 × g for 30 min to pellet cell bodies. The supernatants containing sheared flagella were centrifuged for 1 h at 100,000 × g to pellet flagella. These were resuspended in 200 μl of the above buffer and visualized by Coomassie staining of a 10% gel after SDS-PAGE. Densitometry of gel bands was carried out using an Image Station 2000R (Kodak) equipped with Kodak ID software (v. 3.6.2).

Rhamnolipid production was qualitatively estimated on M8 swarming plates by visualizing the zone of wetting material around the colony. The diameter was measured for multiple colonies after overnight incubation of the swarming plates at 37°C.

Statistical analysis.

Statistical analysis and graphing were carried out using Prism 4.0 (GraphPad) and Excel OS X (Microsoft) softwares.

RESULTS

FlhF is required for normal flagellum-dependent motility.

FlhF is required for normal swimming motility in bacteria with polar flagella, including P. putida. We constructed an in-frame, unmarked deletion of flhF in P. aeruginosa strain PAK which removes amino acids 7 to 430 (inclusive) of this protein. ΔflhF bacteria exhibited no growth defect compared with PAK bacteria in either rich (LB) or minimal (M8) media (data not shown). In assays of swimming motility, PAK strains consistently showed a larger swimming zone than ΔflhF bacteria (Fig. 1A and B); however, the residual swimming motility of ΔflhF bacteria was significantly greater than that of ΔfliC bacteria, which do not express flagellin (Fig. 1D). The swimming motility defect of ΔflhF bacteria could partially be complemented in this assay by expressing the FlhF-tdimer2 fusion protein (Fig. 1C). We confirmed that the deletion of ΔflhF had no apparent polar effects on downstream gene transcription by measuring fliA mRNA levels in wild-type and ΔflhF cultures sampled at mid-exponential phase. Quantitative real-time PCR detected 1.20 ± 0.13 copies of fliA mRNA/100 copies of rplU mRNA in the wild-type strain; this was not significantly different from the 0.91 ± 0.16 copies of fliA mRNA/100 copies of rplU mRNA measured in the ΔflhF sample (means ± standard deviations; n = 5).

FIG. 1.

FlhF is required for optimal swimming motility. Swimming zones were measured 16 h after inoculating 0.3% LB agar plates; zone diameters (means ± standard deviations; n = 4) are listed for each strain. (A) PAK (40.8 ± 1.9 mm), (B) ΔflhF (15.3 ± 1.3 mm), (C) ΔflhF pFlhF-tdimer2 (26.8 ± 1.5 mm), and (D) ΔfliC (0 mm) strains are shown.

We also examined the motilities of wild-type and ΔflhF bacteria in liquid by brightfield video microscopy. Whereas wild-type bacteria were seen to have typical straight swimming behavior interrupted by occasional tumbling, ΔflhF bacteria often appeared to rotate or “twirl” around a fixed point and were more frequently nonmotile (Table 3) (see Videos S1 and S2 in the supplemental material). We next traced the movement of individual wild-type and ΔflhF bacteria for one second and calculated the mean velocity (n = 25). Representative tracings are shown in Fig. 2A and B, and the distribution of velocities is plotted in Fig. 3A. The majority of wild-type bacteria had a velocity of 30 to 50 μm/s (35.4 ± 13.6 μm/s), which is in agreement with previous observations from the literature for PAK (12). The mean velocity of ΔflhF bacteria was 14.7 ± 8.6 μm/s, with no bacteria observed to have a velocity of >30 μm/s. The swimming defect of ΔflhF organisms became more pronounced as we increased the viscosity of the media by adding PVP. In 3% PVP, the velocity of all ΔflhF organisms was ≤10 μm/s, while wild-type bacteria attained velocities as high as 30 μm/s (Fig. 3B). In 6% PVP (Fig. 2C and D), most ΔflhF bacteria showed minimal movement, while many wild-type bacteria still swam at speeds of up to 15 μm/s (Fig. 3C).

TABLE 3.

FlhF is required for linear swimming

Strain	No. of strains with indicated behavior/no. studied (%)
	Linear swimming	Twirling	Nonmotility
PAK	93/105 (88.6)	2/105 (1.9)	10/105 (9.5)
ΔflhF strain	6/101 (6)	73/101 (73)	22/101 (22)
ΔflhfF pFlhF-tdimer2 strain	61/100 (61)	14/100 (14)	25/100 (25)

FIG. 2.

FlhF is required for linear swimming. The movement of individual PAK (A and C) and ΔflhF (B and D) bacteria was tracked for 30 frames (1 s) in M8 media (A and B) or M8 media plus 6% (wt/vol) PVP (C and D), using the manual tracking plug-in of ImageJ (NIH). White lines show the paths taken by individual cells during 1 s under these conditions. Bars = 10 μm.

FIG. 3.

ΔflhF bacteria swim at decreased velocity. PAK and ΔflhF bacteria were visualized by brightfield video microscopy as described in Materials and Methods. The distributions of velocities calculated for 25 individual cells in M8 media (A), M8 plus 3% (wt/vol) PVP (B), or M8 plus 6% (wt/vol) PVP (C) are shown.

ΔflhF bacteria assemble lateral flagella.

FlhF is required for polar placement of flagella in P. putida. We examined the placement of flagella in wild-type and ΔflhF P. aeruginosa by indirect immunofluorescence microscopy, using a polyclonal antiserum that recognizes FliC to label flagella. The proportions of bacteria lacking flagella were similar in both wild-type (6/102 or 5.9%) and ΔflhF (10/103 or 9.7%) samples. We next examined the origin of the flagellum in ca. 100 bacteria per strain in which the point of flagellar attachment was clearly seen. Wild-type bacteria consistently had a single polar flagellum (98%); bacteria which appeared to be late in cell division often showed a single flagellum at each old pole (Fig. 4A). In contrast, 80% (74/93) of the flagella observed on ΔflhF bacteria were not polar (Fig. 4B). Polar versus lateral wall points of origin for flagella could also be seen by transmission electron microscopy for wild-type (Fig. 5A) and ΔflhF (Fig. 5B to C) bacteria, respectively.

FIG. 4.

FlhF localizes to the bacterial pole and is required for the assembly of polar flagella. PAK (A) and ΔflhF (B) bacteria expressing tdimer2 (to allow for visualization of the cell body) were fixed and labeled with anti-FliC antiserum as described in Materials and Methods to visualize flagella. The localization of FlhF-tdimer2 to the cell pole was seen by live cell imaging of PAK pFlhF-tdimer2 (C). In contrast, diffuse fluorescence was observed for PAK expressing tdimer2 alone (D). ΔflhF pFlhF-tdimer2 (E) and PAK pFlhF-tdimer2 (F) bacteria were fixed and stained with anti-FliC antiserum, allowing us to see colocalization of FlhF-tdimer2 and flagella at the bacterial pole.

FIG. 5.

Transmission electron microscopy shows that ΔflhF bacteria lack polar flagella. Wild-type PAK (A) and ΔflhF (B and C) bacteria were processed for transmission electron microscopy as described in Materials and Methods. Panel C shows a higher-magnification view of the flagellar origin seen in panel B.

FlhF is localized to the same cell pole as the flagellum.

The requirement of FlhF for polar placement of flagella has led to the hypothesis that FlhF resides at the bacterial pole and regulates flagellar assembly at that site (10). We constructed an FlhF-tdimer2 fusion protein to allow us to visualize the subcellular distribution of FlhF. This fusion protein partially complemented the swimming defect of ΔflhF bacteria, as shown in Fig. 1C. Video microscopy of ΔflhF pFlhF-tdimer2 bacteria swimming in liquid LB demonstrated that the number of bacteria showing wild-type (straight) swimming increased vis-à-vis that of ΔflhF bacteria but not to the level observed for wild-type bacteria (Table 3) (see Video S3 in the supplemental material). FlhF-tdimer2 appeared to be concentrated at the cell pole both in live cells (Fig. 4C) and in fixed bacteria (Fig. 4E and F), in contrast to the uniform red fluorescence seen for bacteria expressing tdimer2 (Fig. 4D). Bacteria which appeared to be close to cell division, as judged by their increased length and nascent septum formation, frequently showed FlhF-tdimer2 localization at both old poles (Fig. 4C). Labeling of flagella with anti-FliC antiserum in cells expressing FlhF-tdimer2 allowed us to observe that flagella originated from the pole at which FlhF-tdimer2 was concentrated (Fig. 4E and F). Most (59/67 or 88%) ΔflhF pFlhF-tdimer2 cells possessed a polar flagellum; however, we did not see cells with more than one flagellum at the pole, unlike previous observations for P. putida strains overexpressing FlhF (25).

fliC transcription and flagellin expression are reduced in the absence of FlhF.

flhF is transcribed early in the flagellar biosynthesis pathway and affects transcription of downstream flagellar genes in other bacteria (6, 10, 24). We tested whether this was also the case for P. aeruginosa by measuring the transcription of fliC, a class IV gene encoding flagellin, which is transcribed late in flagellar biosynthesis. PAK and ΔflhF bacteria harboring an fliC-lux transcriptional reporter integrated into the chromosome at the attB site were grown in LB liquid media. As seen in Fig. 6A, luciferase activity was consistently decreased in ΔflhF bacteria, indicating that fliC transcription is diminished in the absence of FlhF.

FIG. 6.

fliC transcription and flagellin expression are decreased in the absence of FlhF. (A) fliC transcription. PAK fliC-lux and ΔflhF fliC-lux cultures were sampled over 6 h; culture density (OD₆₀₀) and luminescence (relative light units) were determined at each time point as described in Materials and Methods. The results shown are representative of three independent experiments. (B) Flagellin expression in whole-cell lysates. Lysates were prepared from PAK and ΔflhF bacteria harvested from log-phase (OD₆₀₀ of 0.7) or stationary-phase (OD₆₀₀ of 2.0) cultures. Samples were normalized to total protein, separated by SDS-PAGE, and transferred to a polyvinylidene difluoride membrane. Flagellin was detected by immunoblotting with anti-FliC antiserum. (C) Surface flagellin expression. PAK and ΔflhF bacteria were harvested from log-phase (OD₆₀₀ of 0.7) or stationary-phase (OD₆₀₀ of 2.0) cultures. Surface flagella were isolated as described in Materials and Methods, separated by SDS-PAGE, and visualized by Coomassie blue staining. The amount of protein loaded in each pair of lanes represents equal numbers of bacteria, as normalized by an OD₆₀₀.

We also compared flagellin expression levels in PAK and ΔflhF bacteria during growth in liquid media. Flagellin levels were determined both in whole-cell lysates (total flagellin) (Fig. 6B) and in preparations of sheared flagella (surface flagellin) (Fig. 6C). Compared with wild-type bacteria, decreased amounts of both total and surface-expressed flagellin were present in ΔflhF bacteria harvested from log-phase or stationary-phase liquid cultures. Quantitative analysis of multiple Western blots of whole-cell lysates probed with anti-FliC antiserum revealed that during log-phase growth, ΔflhF cells had 84.5% ± 8.4% of wild-type FliC levels (n = 4). The difference in flagellin expression was more pronounced during stationary phase, with ΔflhF bacteria producing 72.5% ± 7.1% of wild-type levels (n = 4).

FlhF is required for swarming.

The nonpolar flagella observed in ΔflhF bacteria appear to be less effective than polar flagella in supporting swimming motility through liquid media of increasing viscosity. As swarming is another flagellum-dependent form of motility (11, 26), we compared the behaviors of wild-type and ΔflhF bacteria on swarming plates. On 0.5% agar M8 plates, wild-type bacteria showed characteristic swarming motility (Fig. 7B), while ΔflhF bacteria did not swarm (Fig. 7D). However, when the ΔflhF strain was inoculated on 0.3% agar M8 plates (Fig. 7C), it demonstrated swarming behavior similar to that of the wild-type strain on 0.5% agar M8 plates. Of note, wild-type bacteria swam when inoculated to 0.3% agar M8 plates (Fig. 7A).

FIG. 7.

ΔflhF bacteria have altered requirements for swarming motility. M8 medium swarming plates containing standard (0.5%) or reduced (0.3%) concentrations of agar were inoculated with PAK or ΔflhF bacteria and photographed after 72 h of incubation. Swarming of ΔflhF is observed only under conditions that promote swimming in the wild-type strain. (A) PAK (0.3% agar), (B) PAK (0.5% agar), (C) ΔflhF (0.3% agar), and (D) ΔflhF (0.5% agar) strains are shown.

Differential swarming of ΔflhF bacteria on 0.3% versus 0.5% agar is not due to changes in FliC expression.

P. aeruginosa swarming is accompanied by an increase in the number of polar flagella, from one to two (17, 26). A doubling of the number of flagella is also observed when Salmonella enterica serovar Typhimurium transitions from swimming to swarming and is accompanied by a roughly twofold increase in fliC mRNA levels as measured by microarray analysis (36). As a recent paper suggests that transcription of late flagellar genes in S. enterica serovar Typhimurium may be regulated by flagellar sensing of hydration of the external milieu (37), we tested whether ΔflhF bacteria swarmed on 0.3% and not 0.5% agar M8 plates because of a difference in the expression of flagellin under these conditions. Wild-type and ΔflhF strains carrying the fliC-lux transcriptional fusion were inoculated to 0.5% and 0.3% agar M8 swarming plates, and luminescence was measured over multiple areas at the edge of the swarmer colony. As previously observed in liquid culture, ΔflhF bacteria showed decreased fliC transcription compared with wild-type PAK bacteria when grown on 0.5% agar M8 plates (Fig. 8A, B, and D). As a control, we measured transcription of the housekeeping gene encoding glucose-6-phosphate dehydrogenase (zwf). As seen in Fig. 8D, activations of the zwf-lux transcriptional reporter were identical in PAK and ΔflhF bacteria under these conditions. We also prepared lysates from bacteria harvested from the edges of wild-type or ΔflhF swarmer colonies grown on 0.5% agar M8 plates. Consistent with decreased fliC transcription, flagellin protein expression was also decreased in ΔflhF bacteria compared with that in wild-type bacteria (Fig. 8E).

FIG. 8.

FliC transcription and protein expression during swarming. PAK fliC-lux inoculated to 0.5% agar M8 swarming media (A) and ΔflhF fliC-lux inoculated to 0.5% agar (B) or 0.3% agar (C) were imaged at 48 h postinoculation with an Image Station 2000R, allowing luminescence to be measured from the swarmer colony. Luminescence intensity was measured in at least four areas at the swarmer colony periphery by using 1D software (Kodak). Circles (indicated by arrows) show the area over which luminescence was measured. (D) Luminescence produced by fliC-lux and zwf-lux transcriptional reporters, measured from the swarmer colony periphery for PAK and ΔflhF strains under the indicated conditions after 48 h. Bars represent means ± standard deviations of three independent experiments. (E) The total flagellin present in whole-cell lysates of bacteria harvested from swarming plates was detected by immunoblotting with anti-FliC antiserum. Samples were normalized to total protein prior to SDS-PAGE. The amount of FliC present in each sample was estimated by measuring chemiluminescence signal intensity; values are the averages of two independent experiments and are normalized to that of PAK (0.5%).

Despite the observation that ΔflhF bacteria swarm on 0.3% agar M8 plates but not on 0.5% agar M8 plates, we found no difference in either fliC transcription or FliC expression under these two conditions in the mutant strain (Fig. 8A to E). Thus, these observations do not support the hypothesis that ΔflhF bacteria increase flagellin synthesis, albeit under altered conditions of hydration or viscosity compared to those of the wild type, in order to swarm on 0.3% M8 agar plates. Instead, these results argue that the flagella assembled by ΔflhF mutants propel P. aeruginosa effectively only when resistance is low, i.e., under conditions of decreased viscosity or increased wetness. Indeed, when we harvested swarming ΔflhF bacteria from 0.3% agar M8 plates and inoculated them directly onto 0.5% agar M8 plates, swarming motility ceased (data not shown). This confirmed that FlhF is required for swarming on higher-percentage agar plates even if swarming has been initiated at a lower agar concentration.

Lastly, the production of rhamnolipid wetting material, which forms a visible ring surrounding the swarmer cells, is also required for P. aeruginosa swarming (4, 11). Both wild-type (18.75 ± 1.25 mm; n = 4) and ΔflhF (18.5 ± 0.6 mm; n = 4) bacteria produced rhamnolipid zones of similar sizes on 0.5% agar M8 plates. Thus, a defect in rhamnolipid production does not account for the failure of ΔflhF bacteria to swarm.

DISCUSSION

FlhF is concentrated at the cell pole and is required for polar placement of the flagellum in P. aeruginosa.

The role of FlhF in flagellar biosynthesis is still unclear. The loss of FlhF in either Vibrio cholerae or Helicobacter pylori results in nonmotile, aflagellate organisms (6, 24). In P. putida and Vibrio alginolyticus, FlhF is required for polar flagellar placement and a ΔflhF mutant shows reduced, but not absent, motility (18, 25). Our work demonstrates that this is also true for P. aeruginosa. As FlhF is required for polar flagellum placement, it has been hypothesized to direct the assembly of flagellar components at this site (10). Such a model predicts that FlhF should be present at the cell pole. Our results confirm this prediction, as we see that FlhF is concentrated at the pole from which the single flagellum originates. However, the signal required for polar localization of FlhF remains unknown.

The fact that flagellar assembly does occur in the absence of FlhF suggests that the signal recognition particle (SRP) function of FlhF can be assumed by another SRP protein, of which there are two in P. aeruginosa, i.e., FtsY (PAO373) and Ffh (PA3746). Indeed, FlhF is dispensable for motility in Bacillus subtilis, which also expresses the SRP proteins FtsY and Ffh, suggesting that its role in flagellar assembly can be assumed by another protein in this organism (39). Our observation that ΔflhF flagella are less efficient at moving bacteria in both liquid and semisolid media might be due to their placement at locations other than the pole. However, it is possible that a flagellum assembled in the absence of FlhF differs structurally from the wild-type organelle. To our knowledge, the fine structure and protein composition of flagella assembled in the absence of FlhF have not been examined.

FlhF is required for wild-type levels of fliC transcription under swimming and swarming conditions.

In many organisms, FlhF is required for transcription of downstream flagellar genes at wild-type levels. In H. pylori, the loss of FlhF results in a decrease in transcription of level II and III flagellar genes as well as flagellin (24). The down-regulation of class III, but not class II, gene transcription in H. pylori ΔflhF strains is dependent on the anti-sigma factor FlgM. In V. cholerae, FlhF is also required for wild-type expression of flagellin (6). Our data demonstrate a modest decrease in transcription of the class IV gene fliC during log-phase growth in P. aeruginosa when FlhF is absent, as well as a corresponding decrease in flagellin expression. Decreased surface flagellin expression could reflect the assembly of shorter flagella on all cells versus the assembly of full-length flagella on a subpopulation of cells. Although we have not explicitly examined the length of flagella assembled by ΔflhF cells, we did not observe a marked difference in the numbers of aflagellate wild-type and ΔflhF bacteria by indirect immunofluorescence.

flhF is an early stage II gene transcribed in an operon with fleN.

Transcription of both flhF and fleN is dependent on RpoN and FleQ (10). Many studies have demonstrated the importance of FleQ in determining the number of flagella in P. aeruginosa (7, 10, 32). Deletion of FleN, an antiactivator of FleQ, also results in increased numbers of flagella, likely through upregulation of FleQ-dependent flagellar genes (7). During swarming motility, P. aeruginosa increases the number of polar flagella it assembles from one to two. It is not clear how this change in expression is regulated, although it is likely that FleQ or FleQ-dependent genes are involved. A role for FlhF in regulating the number of flagella is suggested by some observations from the literature. An insertional mutation in flhF in the plant pathogen Xanthomonas oryzae results in aflagellate bacteria during growth in liquid media. The same mutant, however, expresses flagella when harvested from 0.3% agar swarmer plates, though the kinetics of swarming are delayed compared to those of the wild type (28). However, this phenotype may result from polar effects of the mutation on fleN (28). Overexpression of FlhF in P. putida results in the assembly of multiple polar flagella (25). We did not observe this with our FlhF-tdimer2 construct, which partially complemented the swimming defect of the ΔflhF strain, with a myc-FlhF fusion protein expressed from this same high-copy-number vector, or with a wild-type (untagged) allele of FlhF expressed from a low-copy-number vector under the control of the inducible tac promoter (our unpublished results). Thus, expression of FlhF from an exogenous promoter may disrupt the normal balance of gene expression for optimal flagellar assembly.

FlhF is required for normal swarming motility of P. aeruginosa.

Bacteria adapt to movement on surfaces in different ways. Peritrichous bacteria have been observed to upregulate the number of flagella and produce surfactants to facilitate swarming behavior on surfaces (reviewed in reference 13). Alternatively, bacteria which use a polar flagellum for swimming, such as V. alginolyticus and Vibrio parahaemolyticus, can express an alternate flagellar system to facilitate movement in viscous environments and on surfaces, resulting in bacteria which express both polar and lateral flagella under these circumstances (1, 29). P. aeruginosa strains make a modest switch from one polar flagellum to two polar flagella during the transition from swimming to swarming motility. However, two sets of flagellar motors exist in P. aeruginosa; although either can power swimming motility, only one (encoded by PA1460-1461) is capable of supporting swarming (12, 34). A similar situation appears to exist in B. subtilis, which possesses two flagellar motors that differ in their abilities to support swarming motility (16). Little is known about the regulation of these alternate stators and their association with flagella under conditions that favor swimming versus swarming motility. However, the swarming motility defect that we have observed for ΔflhF bacteria is similar to that described for mutants lacking the PA1460-PA1461 stator (12, 34). This raises the possibility that flagella assembled by ΔflhF bacteria are incapable of associating with this flagellar motor, either because they are assembled without a necessary adaptor or because of their nonpolar localization.

We were able to rule out several reasons ΔflhF bacteria might swarm only on 0.3% agar and not on 0.5% agar. Rhamnolipid production normally occurs in the ΔflhF mutant on 0.5% agar, which suggests that an absence of surfactant does not account for the lack of swarming. Furthermore, the production of rhamnolipids under these conditions may indicate that environmental “sensing” necessary for the swarmer transition can occur in the ΔflhF mutant. A direct test of this hypothesis, however, requires that the events marking a transition from swimming to swarming be described and measured for P. aeruginosa, as has been done for other organisms (36, 37). Although the ΔflhF mutant produces less flagellin than the wild-type strain, we detected no difference in the amounts of flagellin produced by the mutant on 0.3% and 0.5% agar. Thus, a change in flagellin transcription in response to this difference in agar concentration does not explain the change in bacterial motility.

Compared to wild-type organisms, ΔflhF bacteria show decreased swimming speeds in liquid media of various viscosities and smaller swimming zones on semisolid agar. Swimming speed depends on motor torque, cell shape, and the length, number, and location (polar or peritrichous) of flagellar filaments (1). As discussed above, it is possible that flagella assembled in the absence of FlhF subtly differ in their composition from wild-type flagella or show different patterns of association with the two sets of flagellar stators expressed by P. aeruginosa. This may lead to changes in the torque generated by these filaments. However, the aberrant positioning of the ΔflhF flagella may also influence their ability to move P. aeruginosa. While it has been shown that the multiple lateral flagella of V. alginolyticus allow for more-rapid movement of this organism through media of increasing viscosity than does a single polar flagellum (1), this observation is not likely to apply to movement powered by a single lateral flagellum. Indeed, the increased resistance to moving a rod-shaped particle “sideways” is likely to have a profound effect on the motility of the ΔflhF mutant under both swimming and swarming conditions.