Notice: Wiley Online Library will be unavailable on Saturday 27th February from 09:00-14:00 GMT / 04:00-09:00 EST / 17:00-22:00 SGT for essential maintenance. Apologies for the inconvenience.
The single polar flagellum of Pseudomonas aeruginosa is an important virulence and colonization factor of this opportunistic pathogen. In this study, the annotation of the genes belonging to the fla regulon was updated and their organization was analysed in strains PAK and PAO1, representative type-a and type-b strains of P. aeruginosa respectively. The flagellar genes are clustered in three non-contiguous regions of the chromosome. A polymorphic locus flanked by flgJ and fleQ in Region I contains a glycosylation island in PAK. The expression and ordered assembly of the complex multicomponent flagellum is intricately regulated. Dedicated flagellar genes fleQ, fleS, fleR, fliA, flgM and fleN encode proteins that participate in the regulation of the flagellar transcriptional circuit. In addition, expression of the flagellum is coordinately regulated with other P. aeruginosa virulence factors by the alternative sigma factor σ54, encoded by rpoN. In order to gain insight into the hierarchical regulation of flagellar genes, deletion mutations were constructed in fleQ, fleR, fliA and rpoN. The transcriptional impact of these mutations was examined by transcriptional profiling using a P. aeruginosa whole genome microarray. Analysis of the transcriptomes generated for each of these mutants indicates a four-tiered (Classes I-IV) hierarchy of transcriptional regulation. Class I genes are constitutively expressed and include the transcriptional regulator fleQ and the alternative sigma factor fliA (σ28). Class II genes including fleSR, encoding a two-component regulatory system require FleQ and RpoN (σ54) for their transcriptional activation. Class III genes are positively regulated by the activated response regulator FleR in concert with RpoN. The transcription of Class IV genes is dependent on the availability of free FliA following the export of the FliA specific antisigma factor FlgM through the basal body rod-hook structure (assembled from Class II and III gene products). Two previously uncharacterized genes, which are coordinately regulated with known flagellar genes have been identified by genome-wide analysis and their role in flagellar biogenesis was analysed.
If you can't find a tool you're looking for, please click the link at the top of the page to "Go to old article view". Alternatively, view our Knowledge Base articles for additional help. Your feedback is important to us, so please let us know if you have comments or ideas for improvement.
Flagella expressed on the surface of a number of bacterial species serve as a primary means of locomotion or motility. In many pathogenic bacterial species, flagellar motility is required in order to reach the site of infection as the first step in the establishment of a bacterial infection. A role for flagella in virulence has been demonstrated for numerous pathogenic bacteria including Pseudomonas aeruginosa, Campylobacter jejuni, Proteus mirabilis, Vibrio cholerae and Helicobacter pylori (Moens and Vanderleyden, 1996; Ottemann and Miller, 1997).
In P. aeruginosa, an opportunistic human pathogen that possesses a single polar flagellum, non-flagellated mutants have been shown to be defective in virulence in a burned mouse model of infection (Montie et al., 1982) and demonstrate reduced invasion of cultured corneal epithelial cells (Fleiszig et al., 2001). Moreover, disruption of fliD, the gene encoding the flagellar cap, reduces adhesion of P. aeruginosa to human respiratory mucin (Arora et al., 1998). Despite the apparent importance of flagella in colonization and virulence, clinical isolates of P. aeruginosa from chronically colonized cystic fibrosis patients have a propensity to be non-motile (Mahenthiralingam et al., 1994). Explanations for this phenomenon are still lacking but the absence of a flagellum during chronic colonization may be a means by which the organism evades a potent inflammatory response mediated by flagellin (Hayashi et al., 2001) the most abundant flagellar protein. Thus, a complete understanding of the regulation of flagellum assembly in the organism may provide clues about the behaviour of this organism in vivo when more is known about genomic conservation in states of chronic colonization.
Expression of flagellar motility is an energetically expensive process, involving the expression of more than 40 different genes. The products of these genes include regulatory proteins, structural components of the flagellar secretion and assembly apparatus, proteins involved in generating the rotational motor force and the chemosensory machinery that controls flagellar-based chemotaxis. To ensure maximal efficiency and accuracy during flagellar biogenesis, bacteria utilize hierarchical regulatory networks involving transcriptional and post-translational mechanisms to control the ordered expression of the individual components of the flagellar organelle. Detailed studies of the regulation of flagellar assembly have been carried out in Escherichia coli, Salmonella typhimurium (Macnab, 1996), Caulobacter crescentus (Wu and Newton, 1997), V. cholerae (Prouty et al., 2001) and V. parahaemolyticus (Kim and McCarter, 2000; McCarter, 2001). A recent review has compared the different aspects of regulation in these organisms (Aldridge and Hughes, 2002). Whereas significant differences exist between the regulatory programs used by these organisms, a salient feature in all cases is that the flagellar genes can be classified based on their temporal expression and on their dependence on various nested transcriptional regulators.
In peritrichously flagellated E. coli and S. typhimurium, three classes of promoters have been described for the flagellar operons (Macnab, 1996). Class 1 consisting of the master regulator flhCD, is transcribed by σ70 (Kutsukake, 1997) and regulated by cAMP receptor protein CRP (Soutourina et al., 1999). The promoters directly activated by FlhCD comprise Class 2, which includes genes for various structural components and the alternative sigma factor FliA (σ28). The Class 3 promoters require free FliA for their transcription and are dependent on the secretion of the FliA-specific antisigma factor FlgM through the assembled hook-basal body rod. The FliA–FlgM interaction thus regulates flagellar biogenesis via a post-translational mechanism.
In C. crescentus, a four-level gene hierarchy has been described in the biogenesis of its single polar flagellum. CtrA, the cell-cycle response regulator, occupies the first level and in turn activates Class II genes in response to cell cycle cues. The transcriptional factor FlbD and sigma factor RpoN, both Class II gene products, activate the transcription of the subsequent Class III and IV genes (Gober and England, 2000).
In the monoflagellate V. cholerae, a transcriptional hierarchy with four classes of genes has been described (Prouty et al., 2001). Class I consists of the transcriptional activator flrA. FlrA in association with σ54 (RpoN) transcribes the Class II genes, which code for structural and regulatory components including a two-component system (flrBC) and the alternative sigma factor σ28 encoded by fliA. Class III genes are dependent on FlrC and σ54 for their activation. They code for the structural components needed for the completion of the hook-basal body rod structure. Subsequent secretion of the antisigma factor FlgM through the assembled hook is postulated to allow transcription from the Class IV FliA-dependent genes.
In P. aeruginosa, transcription of flagellar genes has been shown to require a number of regulatory proteins including the transcription factor FleQ (Arora et al., 1997) and the two component system FleSR (Ritchings et al., 1995). Successful flagellar biosynthesis also requires three different sigma factors (σ70, σ54 and σ28) (Totten et al., 1990; Starnbach and Lory, 1992; Dasgupta et al., 2002). In addition, the FliA antisigma factor FlgM has been shown to modulate the activity of FliA and expression of the FliA-dependent flagellin gene fliC (Frisk et al., 2002). Although previous work has implicated these regulators in P. aeruginosa flagellar biosynthesis, interconnection of these regulatory components in a hierarchical expression circuit has not been established.
We took advantage of recently available resources and tools to gain an understanding of the interplay of the various flagellar regulators of P. aeruginosa and to construct a putative flagellar assembly pathway. These resources include the completed P. aeruginosa strain PAO1 genome sequence (Stover et al., 2000), the availability of an oligonucleotide microarray (GeneChip®P. aeruginosa Genome Array, Affymetrix) based on this sequence and a limited number of genes from other strains, and the increased collection of publicly available non-PAO1 sequence information. In this report, we present the organization of the various flagellar genes, their hierarchical regulation by RpoN, FleQ, FleR and FliA, and propose a model for the assembly of the flagellum in P. aeruginosa. In addition, we describe two novel P. aeruginosa genes involved in the biogenesis and function of the flagellum. Their role in flagellar biogenesis was hypothesized based on their coordinate regulation with known flagellar genes and confirmed by mutational analysis.
Results and discussion
Annotation and organization of the flagellar regulon in representative type-a and type-b P. aeruginosa strains
As a first step in studying the regulation of flagellar biogenesis, we carried out a thorough analysis to identify and annotate the genes present in the P. aeruginosa genome with a known or predicted role in flagellar biogenesis or function. This analysis was based on the complete genome sequence of P. aeruginosa strain PAO1 (Stover et al., 2000), information from the community-wide annotation project (available at http://www.pseudomonas.com), data deposited in GenBank by various investigators for other P. aeruginosa strains and numerous studies carried out over the past decade. Previous studies have shown that P. aeruginosa strains can express one of two structurally distinct flagellin proteins encoded by the fliC gene, which are designated as type-a and type-b (Spangenberg et al., 1996). The genome of the sequenced strain PAO1 encodes type-b flagellin. However strains expressing type-a flagellin represent the majority of clinical isolates surveyed (Arora et al., 2001). As differences in flagellar components and genetic organization exist between type-b and type-a expressing strains, representative type-b strain PAO1 and type-a strain PAK were both annotated.
Based on the above resources, we were able to identify 41 genes encoding structural/assembly or regulatory components of the flagellar organelle in the sequenced PAO1 genome (Table 1). An additional nine genes known or predicted to be involved in flagellar chemotaxis control were also included in the table of annotations based on their linkage to flagellar genes. Each of the 50 flagellar or linked chemotaxis genes identified in the PAO1 genome was assigned to one of 17 putative operons based on criteria described (Fig. 1; Table 1).
Table 1. . Annotation of P. aeruginosa flagellar biogenesis genes.
Basal body MS ring, mounting plate for motor/switch (E)
Motor/switch, mounted onto MS ring, rotor component (H)
Negative regulator of FliI (H)
Chaperone, export of hook proteins (H)
Hook length control (H)
Flagellar export pathway (H)
Flagellar export pathway (H)
Flagellar export pathway (H)
Flagellar export pathway (H)
Flagellar export pathway (H)
Flagellar export pathway (E)
Polar flagellar site determinant (H)
Flagellar number regulator (E)
σ28sigma factor (E)
Switch regulator (H)
CheY phosphatase (H)
CheY and CheB kinase (H)
Chemoreceptor methylesterase (H)
Motor rotation (H)
Motor rotation (H)
Positive regulator of CheA (H)
Chemotaxis regulator (H)
Chemoreceptor methylesterase (H)
P ring assembly component (H)
σ28-specific antisigma factor (E)
Initiation of filament assembly (H)
a. Unified nomenclature for genes encoding P. aeruginosa type-a and -b flagellar components and linked chemotaxis proteins. Genes are grouped according to their putative organization in operons.
b. Flagellin type specifies whether an allele of the gene is present in the type-a (a) and/or type-b (b) reference strain.
c. The source of the sequence for each gene is provided. Genes not present in a given reference strain are denoted NP. PA numbers refer to the unique identification number assigned to each gene as part of the PAO1 genome sequence annotation project ( http://www.pseudomonas.com ). GenBank accession numbers are given for strain specific sequences found in PAK. Parentheses indicate an alternative gene name was provided in the accession. NA indicates that the DNA sequence is not available for the PAK allele but that its existence is predicted based on the cross-hybridization of PAK genomic DNA to oligonucleotide probes derived from the corresponding PAO1 gene sequence.
d. Given is the per cent identity between type-a and -b strain alleles. NA signifies that only one sequence is available.
e. Function of the gene product was assigned based on studies conducted in P. aeruginosa (E) or predicted based on homology (H) to proteins of known function found in other flagellar systems.
f. The role of the fleP gene product in P. aeruginosa flagellar biogenesis is described in this work.
These operons are located in three regions of the chromosome (Fig. 1). Most of the structural genes coding for the basal body rod, rings, hook, filament, cap and basal body (flgBCDE, flgFGHIJKL, fliCfleL, fliDS and fliEFGHIJ) are clustered in Region I of the PAO1 genome. The transcriptional regulator fleQ and two-component system fleSR are also located within this region. Region II contains genes encoding the hook length regulator, switch, export apparatus, flagellar placement determinant, flagellar number regulator and alternative sigma factor FliA (fliK, fliLMNOPQRflhB, flhA, flhFfleN, fliA). This region also includes genes encoding motor and chemotaxis proteins (cheYZ, cheAB, motAB, cheW). Region III consists of genes coding for the flagellar export apparatus, antisigma factor (flgA, flgMN) and additional chemotaxis regulatory proteins (cheVR).
The complete genome sequence of a type-a strain is not currently available; however, from the numerous PAK-specific flagellar gene sequences that have been deposited in GenBank, sequence heterogeneity in the flagellar genes flgK, flgL, fliC, fleL, fliD and fliS found in Region I flanked by flgJ and fleQ have been reported between PAK and PAO1 (Table 1; Arora et al., 2001). This region also contains a polymorphic gene fleP, whose product plays a novel role in P. aeruginosa flagellar biogenesis (see below). Strain PAK as well as other type-a strains possess an additional gene in the fliDS operon, designated fliS′ (because of its sequence similarity to fliS), which is located between fliS and fleP (Table 1; Fig. 1; Arora et al., 1998). The role of fliS′ in flagellar biogenesis is unknown and it was included in this study because of its linkage and homology to fliS. Another defining difference between the studied type-a strain and the type-b strains is that all type-a strains carry a glycosylation island, a cluster of genes demonstrated to be involved in the glycosylation of the type-a flagellin (Arora et al., 2001). In strain PAK, the island is composed of 14 genes arranged in several putative operons and located between flgL and fliC (Table 1; Fig. 1). In the type-b strain PAO1, the island is replaced by three genes of unknown function and a putative glycosyltransferase (Arora et al., 2001).
Sequence comparison of 17 other PAK flagellar genes shows at least 99% identity to the equivalent PAO1 genes (Table 1). However, it remains to be determined whether additional evolutionary divergence has occurred between the type-a and -b flagellar systems as many PAK flagellar genes have yet to be sequenced. In order to examine potential sequence differences we have adopted a comparative hybridization approach utilizing GeneChip®P. aeruginosa Genome Arrays (Affymetrix). This commercial oligonucleotide array carries probe sets for all annotated PAO1 genes as well as probe sets specific for the PAK alleles flgK, flgL, fliC, fliD, fliS, fliS′, fleP and the 14 glycosylation island genes vioA-orfN. Genomic DNA, prepared from both PAO1 and PAK, was fragmented, labelled and hybridized to separate GeneChips® as previously described (Wolfgang et al., 2003). Normalized hybridization intensities for each probe set have been included as supplemental data (Table S1). The hybridization profile is presented in Fig. 1. PAK genomic DNA showed levels of hybridization equivalent to that of PAO1 for all PAO1 flagellar gene probes with the exception of flgK, flgL, fliC, fleL, fliD, fliS and fleP mapping in Region I (Fig. 1A). The absence of hybridization to probes for these genes reflects the known sequence polymorphisms discussed above. The absence of, or marginal (orfN ) hybridization of PAK DNA to the PAO1 sequence based probes indicates that significant sequence heterogeneity can be readily detected (Fig. 1B). Whereas minor sequence variation is likely to exist within the unsequenced PAK flagellar genes, relative to PAO1, the level of hybridization seen in Fig. 1A indicates that the set of genes sufficient for flagellar biogenesis in PAO1 are present in the PAK genome and they share a high degree of homology with PAO1 genes. However, this does not rule out the possibility that additional flagellar genes not identified in this study are present in type-a strains.
Transcriptional profiling in P. aeruginosa and elucidation of the regulon controlling flagellar gene expression
Previous reports have shown that the expression of some flagellar gene(s) necessary for the assembly of the flagellar organelle is under the control of several key regulatory factors including RpoN, FleQ, FleR and FliA (Totten and Lory, 1990; Starnbach and Lory, 1992; Ritchings et al., 1995; Arora et al., 1997; 1998; Jyot et al., 2002). The coordination of these regulators and their involvement in the expression of other flagellar genes is unclear. In order to gain a more complete understanding of the regulatory pathways that control the expression of genes encoding flagellar components, we again utilized GeneChip®P. aeruginosa Genome Arrays (Affymetrix) to compare the expression of genes in wild type and isogenic fleQ, fleR, fliA and rpoN non-polar deletion mutants. Because most clinical isolates express type-a flagellin, analysis of the transcriptional regulatory network controlling flagellar expression was carried out in PAK, a model type-a P. aeruginosa strain. We have previously shown that of the 5900 probe sets included on the array 5678 hybridize to PAK genomic DNA (Wolfgang et al., 2003). For transcriptional profiling experiments only these 5678 probe sets were used. A complete list of the 5678 PAK specific probe sets used for transcriptional profiling is included as supplemental data (Table S2).
We first analysed the global effect of each of the deletion mutations on transcript levels (Table S2). Figure 2 summarizes the transcriptional profiling results, indicating the number of genes showing significantly altered transcript levels in each of the regulatory mutants relative to the wild-type parent. The identity of these genes and magnitude of transcriptional expression change in each mutant is provided as supplemental data (Tables S3-6). Although known flagellar genes exclusively showed reduced transcription in the mutants (consistent with previous studies), the genome-wide analysis showed additional genes significantly activated and repressed in the mutant backgrounds. Many of these genes showed altered expression in multiple mutants, indicating an overlapping regulatory network (Fig. 2).
Genes showing altered regulation in multiple mutants in the microarray studies were of particular interest. Forty-two genes showed significantly reduced expression in both fleQ and rpoN mutants (Fig. 2A). Of these, 26 showed reduced expression in the fleR mutant, indicating that their transcription was FleR dependent. The other 16 FleR-independent genes were not differentially regulated in the FliA mutant, therefore they could be categorized as genes directly regulated by FleQ. Because FleQ controls the expression of the fleSR operon, it is conceivable that FleQ indirectly controls the expression of the 26 FleQ/FleR/RpoN-dependent genes through the two-component system FleSR. 20 of these genes are known or predicted flagellar genes (Fig. 2A). They included 12 genes previously implicated in flagellar biogenesis, three genes that likely encoded motility/chemotaxis functions (PA1458, PA1459, PA2654), and PA1462 (which belongs to the ParA family of chromosome segregation proteins and may be part of the PA1458/59 operon) (Table 1). Of these fliC, flgMN and fliS′ additionally displayed reduced transcription in the fliA mutant. Among the six genes not previously implicated in flagellar expression, five showed FliA dependence: PA1096 (fleP), PA3713, PA4310, PA4326, and PA4633 (Fig. 2B). Figure 3A shows the relative expression levels for 56 PAK genes that have been shown or implicated in flagellar biogenesis. Figure 3B indicates the location of previously mapped or deduced binding sites for the various transcription factors implicated in the regulation of expression of genes in the flagellar regulon. The chemotaxis genes mapping in Regions I, II and III were not included for further analyses.
The microarray results were validated using lacZ-reporter constructs in the fleQ, fleR and fliA mutant backgrounds. Because the rpoN deletion has a pleiotropic effect it was not included in the lacZ reporter assays. Promoters of genes (vioA-orfN) predicted to be involved in the glycosylation of flagellin was not included in the reporter assay since they would be analysed in a separate study on regulation of glycosylation of flagellin. In general, most of the FleQ, FleR and FliA regulatory effects were verifiable by the reporter assay (Table 2). However, microarray analysis revealed two discrepancies with previously published reports and with the promoter-lacZ reporter analysis. The operon transcribed from the promoter preceding the fliD gene was FleR-dependent by microarray analysis (Fig. 3A), but not by reporter assay (Table 2). The pattern of RNA abundance corresponding to the individual genes in the operon (Table S2) suggests that the fleR mutation may have resulted in enhanced degradation of the entire RNA transcript from the 3′ end, which would not be apparent from the analysis of promoter activity in a fusion construct. Similarly, in the microarray studies fliS′ and fleP belonging to the same operon as fliD show reduced transcription in the fliA mutant background probably because of the enhanced degradation of these gene transcripts at the 3′ end of the operon.
Table 2. . Analysis of the fold upregulation or downregulation of various P. aerug inosa promoters in a β-galactosidase (LacZ)-reporter assay.
Folda down (–) or up (+) regulation of promoter activityb in:
Promoter positively regulated by:
. The promoter activities were determined in the wild-type PAK strain and various mutants and the fold change calculated with respect to the wild type.
The promoter activity was determined using the β-galactosidase assay. Each fold change depicted is an average of two assays along with the standard deviation.
−17 ± 1.2
−28.0 ± 1.3
1.0 ± 0.2
FleQ and FleR
−2.5 ± 0.3
−2.7 ± 0.2
+1.9 ± 0.2
FleQ and FleR
−14.8 ± 1.2
−8.4 ± 1.0
−21.5 ± 1.4
FleQ,FleR and FliA
−3.1 ± 0.2
+1.2 ± 0.1
+1.1 ± 0.1
−1.2 ± 0.1
−1.7 ± 0.1
−1.4 ± 0.1
−6.8 ± 0.8
1.0 ± 0.1
1.0 ± 0.1
−3.0 ± 0.4
+1.4 ± 0.1
+1.1 ± 0.0
pfliK-lacZ in attB site
−1.8 ± 0.2
−1.7 ± 0.2
1.0 ± 0.0
FleQ and FleR
−3.1 ± 0.6
−1.4 ± 0.1
+1.2 ± 0.1
−23.6 ± 1.0
1.0 ± 0.1
1.0 ± 0.0
−9.1 ± 0.5
1.0 ± 0.1
1.0 ± 0.1
+1.3 ± 0.1
+2.6 ± 0.1
1.0 ± 0.0
−8.2 ± 0.9
+1.6 ± 0.2
+1.6 ± 0.2
−2.1 ± 0.2
−2.0 ± 0.2
−1.9 ± 0.2
FleQ,FleR and FliA
The other discrepancy pertained to the transcripts for the fliLMNOPQRflhB genes, which did not display significant changes by microarray in any of the mutants, suggesting that the transcription of these genes was independent of FleQ, FleR, RpoN and FliA. However, previous promoter activity studies conducted in our laboratory on the fliLMNOPQRflhB promoter indicated that it is dependent on FleQ (Dasgupta et al., 2000; Jyot et al., 2002). We verified the previous work by examining the activity of a fliL-lacZ promoter fusion in both wild-type PAK and a fleQ mutant where a threefold difference in β-galactosidase activity was detected (Table 2). The disparity in microarray and reporter results for this particular operon cannot be accounted for by the overall low levels of expression (Table S2) and thus represents an exception to the trend of agreement between these independent analysis methods.
Transcriptional profiling and reporter analysis in the various mutant backgrounds allowed us to identify and distinguish direct and indirect effects of the regulatory factors on the expression of the flagellar regulon. The fliCfleL operon and the flgMN operon show dependence on RpoN, FleQ, FleR and FliA. The regulatory regions of these operons contain a consensus FliA (σ28) binding site and putative RpoN binding sites. The initiation of transcription at sites lying in close proximity to the predicted FliA binding sites in fliCfleL and flgMN has been demonstrated previously by primer-extension analysis (Starnbach and Lory, 1992; Frisk et al., 2002). We therefore conclude that their expression involves a regulatory cascade involving RpoN, FleQ, FleR and FliA. Very likely, FleQ controls the RpoN-dependent expression of fleSR, and FleR indirectly regulates the FliA dependent expression of the fliCfleL and flgMN promoters by a mechanism that we discuss below.
In P. aeruginosa, the activity of FliA is modulated by its antisigma factor FlgM (Frisk et al., 2002). In other flagellated bacterial species including S. typhimurium, FlgM directly binds to FliA and prevents it from activating transcription of FliA-dependent promoters (Aldridge and Hughes, 2002). The inhibition is postulated to be reversed following secretion of FlgM through the hook-basal body rod structure as demonstrated in S. typhimurium (Karlinsey et al., 2000). As the synthesis of components of the hook-basal body rod structure in P. aeruginosa is dependent on FleQ, FleR and RpoN (Fig. 3A; Table 2) and the secretion of FlgM is dependent on the completion of the hook basal body rod, we hypothesize that the secretion of FlgM would be compromised in a fleQ or fleR mutant background.
Completion of the hook substructure serves as a checkpoint for the transcription of FliA dependent genes such as fliC
In order to test the hypothesis that in P. aeruginosa FliA activity is dependent on completion of the basal body rod-hook structure and subsequent secretion of the FlgM antisigma factor, synthesis of flagellin from the FliA dependent fliC gene was assessed in various mutants by SDS-PAGE and immunoblotting with FliC-specific antisera (Fig. 4). Lysates from wild-type PAK displayed immunoreactivity to a cluster of protein bands suggestive of glycosylated flagellin (lane 1). These bands were not detected in the negative control lysates of the fliC mutant as expected (lane 8). Reduced amounts of flagellin were detected relative to wild type in equivalent lysates from the fleQ (lane 2), fleR (lane 3), and fliA (lane 4) regulatory mutants. This reduction in flagellin synthesis is consistent with the reduction in fliC expression seen in these mutants (Fig. 3A and Table 2). To further test the hypothesis that the flagellin synthesis defect results from reduced secretion of FlgM through the basal body rod-hook structure, insertion mutants were constructed in fliF and flgE. The structural proteins FliF (MS ring) and FlgE (flagellar hook) are implicated in the successful assembly of the basal body rod-hook structure (Chilcott and Hughes, 2000). We were unable to detect flagellin in lysates of the fliF mutant (Fig. 4, lane 5) and significantly reduced amounts were present in the flgE mutant (lane 7) compared to the wild type. As the expression of fliF and flgE are dependent on FleQ and FleR, respectively (Fig. 3A), we conclude that the reduction in fliC expression seen in fleQ and fleR mutants can be accounted for by the reduced expression of these and other basal body rod-hook structure components, resulting in the intracellular accumulation of FlgM and the inhibition of FliA activity. In the lysate of a PAO flhA mutant constructed for a previous study (Fleiszig and Arora, 2001) FliC was not detectable (Fig. 4, lane 6), thus indicating that in the absence of FlhA, predicted to be a structural component of the export apparatus, FliC expression was impaired. This can be best explained by the inability to export FlgM in the flhA mutant.
Coordinate regulation reveals new flagellar genes
Coordinate expression of genes can indicate functional relatedness between known and unknown genes. We used genome-wide transcriptional profiling to search for genes co-regulated with known flagellar biosynthesis operons. As shown in Fig. 2, we identified six genes not previously associated with flagellar expression or function whose expression was dependent on the flagellar transcriptional regulators FleQ, FleR, RpoN and in five cases FliA. Two of these genes (PA4310 and PA4633) have previously been annotated as chemotactic transducers. The remaining four genes (PA1967, PA1096, PA3713, PA4326) specify proteins of unknown function not previously implicated in flagellar biogenesis. In order to assess the potential role of the products of these genes, we have undertaken an analysis of strains carrying mutations in these hypothetical ORFs. We were able to successfully disrupt three out of the four genes; repeated attempts to engineer a mutation in PA4326 failed, suggesting that it may encode a protein whose basal level of expression is essential in P. aeruginosa. Among the engineered strains, mutations in PA1096 (PAK::fleP) and PA3713 (PAK::3713) resulted in reduced motility in the soft-agar motility assay (Fig. 5). Because of the effect of PA1096 on the P. aeruginosa pilus (below), we named this gene fleP. Motility of P. aeruginosa with a mutation in PA1967 was not reduced compared to the wild type.
To explore the basis for the observed motility phenotypes of these strains, we examined the fleP and PA3713 mutants by electron microscopy. Negative staining revealed that the average length of polar type IV pili in the fleP mutant (3.08 µm) were significantly longer than the pili in wildtype PAK (1.66 µm). The twitching motility phenotype of the fleP mutant was comparable to that of the wild-type PAK. The length of pili in a PA3713 mutant was comparable to the wild type. Complementing the fleP mutant with fleP in trans on a plasmid (pET-fleP) rescued the motility defect with an intermediate motility zone to that of the wild type and fleP mutant (Fig. 5). On electron microscopic examination in most of the fleP mutant bacteria, we observed that flagella had detached from the cells (data not shown). In the few cells that displayed an intact attached flagellum, the filament lacked the characteristic wavy contour. The polar type IV pili length and flagellar morphology in the pET-fleP complemented strain was comparable to the wild-type strain. We therefore infer that fleP is involved in a novel function responsible for maintaining the length of type IV pili and stable flagellar attachment to the bacterial pole. The flagellum of the PA3713 mutant appeared normal.
Flagellar structural genes are usually conserved across the genomes of flagellated bacteria, but differences in regulatory genes are not uncommon (Aldridge and Hughes, 2002). Peritrichous bacteria such as E. coli and S. typhimurium lack homologues of fleQ, fleS, fleR, flhF, fleN, and fleL found in the monoflagellates V. cholerae, V. parahaemolyticus and P. aeruginosa. More subtle differences in flagellar regulation may separate and distinguish the monoflagellate species from each other. fleP appears to represent a novel flagellar gene specific for genus Pseudomonas, although there appears to be sequence polymorphisms among the FleP homologues from different strains. The 96 amino acid FleP polypeptide from PAK is identical to a protein referred to as orf4 in strain DG1, a type-a strain (GenBank accession L43064) however, it shares only 43% identity with the corresponding protein encoded in the genome of PAO1 (Stover et al., 2000). Using the protein blast algorithm (Altschul et al., 1997), we identified two homologues of FleP in unfinished genome projects of Pseudomonas putida KT2440 and Pseudomonas syringae pv. syringae B728a, sharing 45% and 44% identity, respectively, with the P. aeruginosa PAK FleP. The genome of P. fluorescens Pf0–1 also contains a gene encoding a homologue of PAK FleP (36% sequence identity) (Casaz et al., 2001). In all of these Pseudomonas species, the fleP homologue is located in a large cluster of flagellar biogenesis genes, between homologues of fliS and fleQ.
The deduced sequence of PA3713 specifies a secreted protein of 620 amino acids. It showed significant level of relatedness (44% identity) to a hypothetical unknown protein identified as a translation product of a Ralstonia metallidurans gene (GenBank accession ZP00022954). Although its expression is dependent on RpoN, FleQ, FleR, and FliA, we were unable to identify any recognizable binding sites for RpoN and FliA upstream of this apparently monocistronic gene. The regulatory effects observed in the microarray analysis are therefore very likely indirect. Because no obvious morphological alterations were seen in the flagellar structure of the mutant lacking PA3713, we hypothesize that this gene product participates in flagellar function, perhaps in chemotaxis.
Transcriptional hierarchy of the P. aeruginosa flagellar regulon consists of four classes of genes that are differentially regulated
Peritrichously flagellated and monoflagellated bacteria show sequence similarities in flagellar proteins which suggest that their functions are conserved. The hierarchical organization of the regulatory networks that control flagellar biogenesis however, varies with flagellar number and placement. Bacteria with peritrichous flagella display a three-tiered hierarchy of transcriptional regulation of their fla regulon (Macnab, 1996), whereas the monoflagellates V. cholerae (Prouty et al., 2001), V. parahaemolyticus (Kim and McCarter, 2000) and C. crescentus (Gober and England, 2000) use a four-tiered hierarchy. The results presented here indicate that P. aeruginosa, a monoflagellate, utilizes a four-tier hierarchy (Fig. 6) involving the RpoN, FleQ, FleS, FleR, FliA and FlgM regulatory proteins to control and coordinate the transcription of the flagellar regulon.
Class I genes are externally regulated
Based on our present understanding of flagellar biogenesis, FleQ is the master regulator of the flagellar regulon in P. aeruginosa. It directly or indirectly regulates the expression of the majority of flagellar gene promoters (Fig. 3A; Table 2) with the exception of fliA. Consequently, we have grouped the promoters of fleQ and fliA as Class I. Although these promoters are primarily regulated by factors outside of the flagellar regulon (Fig. 6), they are independently controlled. In a recent report (Dasgupta et al., 2002) we provide evidence indicating the involvement of σ70 in the transcription of the fleQ gene and its repression by Vfr, the P. aeruginosa homologue of E. coli CRP. Additionally, fleQ expression appears to be modestly influenced by any disruption of the flagellar biogenesis pathway (Fig. 3A; Table 2).
The molecular mechanism that controls the expression of fliA remains unknown. The transcription of fliA appears to be constitutive and not dependent on σ54 or other flagellar regulators tested here (Fig. 3A; Table 2). The putative fliA promoter region was analysed for the presence of a σ70 consensus, although a complete consensus binding site was not identified, a putative −10 σ70 binding site with a four out of six match was found. The potential binding site was subjected to site directed mutagenesis and the activity of the altered promoter (placΩmutfliA) was compared to that of the wild type using a lacZ fusion. The mutation did not affect fliA promoter activity (data not shown). It therefore appears that fliA transcription is initiated by a novel mechanism, perhaps involving an as yet uncharacterized sigma factor.
Class II genes require FleQ and RpoN (σ54) for transcriptional activation
Promoter fusions and microarray studies indicate that transcriptional activation of fleSR, fliEFGHIJ, flhA, flhFfleN, flgA and fliLMNOPQRflhB genes are directly regulated by FleQ and σ54, and consequently are classified as Class II (Fig. 3; Table 2; Fig. 6). The fliDSS′fleP operon shows dependence on FleQ in the reporter assays (Table 2, Arora et al., 1998) but the microarray data suggests that the operon is both FleR and FleQ dependent. However, Western blots show that the amount of the FliD protein synthesized in a fleR mutant is comparable to the amounts synthesized in the wild-type PAK strain, and it is drastically reduced in a fleQ mutant (unpubl. obs.). Therefore we have classified the fliDSS′fleP operon as FleQ dependent belonging to Class II (Fig. 6).
The Class II genes encode structural components of the basal body, MS ring, P ring, motor, switch, flagellar export apparatus and the filament cap. Regulatory proteins expressed from Class II genes include FlhF, FleN, FleS and FleR (Table 1). FlhF is the polar flagellar determinant in P. putida (Pandza et al., 2000) and its close homologues are present in the genomes of a variety of mono-flagellated Gram-negative microorganisms. It is very likely that FlhF determines flagellar localization in other polar flagellates including P. aeruginosa. FleN, the antiactivator of FleQ, downregulates FleQ activity through direct interactions and thereby plays a crucial role in maintaining a single flagellum (Dasgupta et al., 2000; Dasgupta and Ramphal, 2001). FleS and FleR comprise a two-component system in which FleS is a sensor kinase for the response regulator FleR (Ritchings et al., 1995). In V. cholerae, the FleS homologue FlrB activates the FleR homologue FlrC by phosphorylation (Correa et al., 2000). We therefore hypothesize that in P. aeruginosa, the phosphorylation-dependent activation of FleR by its cognate sensor kinase FleS is necessary for the transcriptional progression from Class II to Class III promoters and serves as an additional check point in the flagellar biogenesis of this organism.
The majority of sensor kinases are integral membrane proteins, which activate their cognate response regulators in response to external stimuli. Analysis of the predicted secondary structure of FleS indicates that it is a cytoplasmic protein. The signal sensed by P. aeruginosa FleS is unknown; given its probable localization to the cytoplasm, the signal is unlikely to originate from the extracellular environment.
Class III genes require activated FleR and σ54
The promoters of operons flgBCDE, flgFGHIJKL and fliK are grouped under Class III as their expression required RpoN, FleQ and FleR (Fig. 1A; Table 2; Fig. 6). The apparent involvement of FleQ is an indirect consequence of the fleSR promoter being FleQ dependent (Class II). The expression of the fleSR promoter also requires RpoN as do all other Class II promoters. In addition to its indirect effect RpoN is also required for the expression of Class III promoters as evidenced by the identification of σ54 binding sites in each promoter element (Fig. 3B). Class III promoters control the expression of genes coding for the basal body-rod, L ring, hook, hook-cap scaffold and hook-filament junctional proteins (Table 1).
Class IV genes require the gene products of the preceeding classes and FliA
Among the flagellar operons the transcription of two operons namely, fliCfleL and flgMN were significantly reduced in the fliA mutant (Fig. 3A; Table 2). This is consistent with earlier reports showing fliC and flgMN to be FliA (σ28) dependent (Starnbach and Lory, 1992; Frisk et al., 2002). The cheABmotABcheW, cheVR operons showed a similar trend (Table S1, data not shown) and all genes showing a FliA dependence were grouped under Class IV.
Unlike the other Class IV gene, fliC, which lost most of its promoter activity in the rpoN, fleQ, fleR and fliA mutants, the flgMN promoter retained partial transcriptional activity in all the mutant backgrounds (data not shown). A similar observation with the flgMN promoter was made in an earlier report (Frisk et al., 2002). This indicated that the observed basal level of its transcription was independent of RpoN, FleQ, FleR and FliA. It probably depends on another unknown transcriptional or sigma factor not examined in this study. The regulation of flgM is multifactorial in the enteric sysytem. In S. typhimurium, the flgM promoter is dually regulated by FlhCD and FliA and is classified under Class 2 and 3a respectively (Macnab, 1996). We hypothesize, that in P. aeruginosa, FlgM synthesized in a FliA independent manner binds to free FliA and inhibits its activity initially. Following the simultaneous expression of other Class II and III genes that contribute to the completion of the hook-basal body rod structure, FlgM is then secreted by a similar mechanism as described for S. typhimurium (Karlinsey et al., 2001) allowing for Class IV gene activation by free FliA (Fig. 6).
By adopting a detailed order of transcription of the flagellar genes as seen in other flagellated bacteria, P. aeruginosa functions to make flagellar synthesis more efficient by temporally and spatially regulating transcription to express proteins when needed in an ordered manner.
Model proposed for flagellar assembly in P. aeruginosa
In free-swimming P. aeruginosa, FleQ is the master regulator of flagellar biogenesis. However, there is no distinctive consensus sequence in the FleQ binding sites of the FleQ regulated Class II genes examined (Jyot et al., 2002). The lack of a FleQ binding consensus in these gene promoters suggests that FleQ probably has variable affinity for the different promoters it controls. This may represent a mechanism by which this single transcription factor could control the timing and level of expression of various components required at different stages in the flagellar assembly process. A similar mechanism controlling the expression of flagellar gene promoters has been described in E. coli (Kalir et al., 2001).
The proposed model for the assembly of the single polar flagellum in P. aeruginosa presented in Fig. 7 is based on the present understanding of transcriptional and post-translational regulation of the various flagellar genes. Initiation of the assemblage is most likely to begin with FlhF determining the polar placement site for the new flagellum. FlhF is thus an example of a Class II gene product that is required early on in the biogenesis of this organelle. The other Class II gene products comprising the MS ring, switch, basal-body, and export apparatus are subsequently assembled in the maturing flagellum. FliD, the filament cap protein, though expressed from a Class II gene is probably not assembled into the growing flagellar organelle until the hook and the hook-filament junctional proteins expressed from Class III genes are incorporated into the structure. Therefore though fliD is a class II gene, its product is required later in the assembly process. Meanwhile, simultaneous synthesis of FleN, another Class II gene product, helps maintain the monoflagellate status by inhibiting FleQ activity and downregulating further synthesis of most of the structural components. The FleSR two-component system subsequently activates the Class III genes coding for proteins needed for the completion of the hook-basal body structure. This allows secretion of the antisigma factor FlgM and consequent transcription of FliA dependent genes coding for flagellin (FliC) and some chemotaxis proteins. fleL, a gene co-transcribed with fliC, implicated in the filament length control, maintains the flagellar filament length through a mechanism that yet remains to be understood. The Pseudomonas specific Class II gene product FleP influences stability of the mature flagellar filament by an unknown mechanism.
Promoter analysis and operon assignment
The terminator search program in the GCG Wisconsin package was used to search for factor independent transcriptional terminators. Intergenic regions greater than 100 base pairs were analysed for promoter activity, using the β-galactosidase reporter assay.
Genome content analysis
Genomic DNA hybridizations were preformed as previously described (Wolfgang et al., 2003). Genome arrays (Affymetrix) were washed and scanned according to the manufacturers cDNA protocol. Hybridization intensity data was extracted from the scanned array images and intrachip normalizations were performed using Affymetrix Microarray Suite 5.0 software. The mean average difference of the probe sets was scaled to 500 for interchip comparisons. A presence/absence determination was made by comparing hybridization intensity between strains. For PAO1 based probes values derived from the PAO1 genomic DNA hybridization were used as baseline. For the PAK based probes hybridization values for PAK DNA were used as baseline. Probe sets with a ratio of greater than 0.25 were considered present. Ratios between 0.1 and 0.25 were considered marginal and below 0.1 was designated absent.
Construction of P. aeruginosa mutant strains
The fleQ (PAKΔFleQ), fleR (PAKΔFleQ), fliA (PAKΔFliA) and fliC (PAKΔFliC) deletion mutants were derived from the wild-type strain PAK by introducing mutant alleles into the chromosome as described (Hoang et al., 1998). The mutant alleles for fleQ and fleR deleted for residues 2–475 for FleQ and 2–471 for FleR, respectively, were constructed in a plasmid by ligating two 1 kb fragments flanking the desired region to be deleted. The flanking regions were amplified by polymerase chain reaction (PCR) using the chromosomal DNA of PAK as template, specific primers enlisted in Table 3 and Taq DNA polymerase (Invitrogen). The fleQ flanking regions were amplified using primer pairs QKD1-QD2 and QD3-QD4. Primer pairs RD1-RD2 and RD3-RD4 were used to amplify the fleR flanking regions. The resulting PCR products were either directly cloned into the pCR2.1 TOPO vector (Invitrogen) or digested with restriction enzymes to facilitate cloning into pUC19. These recombinants were further manipulated to ligate the respective fragments flanking either fleQ or fleR. The insert from the resulting plasmid was excised and cloned into pEXGwGm resulting in pEX-Q1234 and pEX-R1234 harbouring the fleQ and fleR mutant alleles respectively.
The mutant allele for fliA and rpoN was assembled by essentially deleting most of the coding region of the gene using a modified version of a PCR based technique termed splicing by overlap extension (SOE) (Warrens et al., 1997). Chromosomal DNA from strain PAK was used as template and the PCR carried out using the GC-Rich PCR system (Roche). Flanking primers were designed to anneal 800–1200 base pairs upstream and downstream of each deletion/splice junction respectively. These primers were tailed with attB1 and attB2 sequences, respectively, as described in the Gateway Cloning Technology Instruction Manual (Invitrogen). The resulting PCR products were cloned using gateway technology into pEXGmGW, which was adapted from pEX18Gm (Hoang et al., 1998) using the gateway Vector Conversion Kit (Invitrogen).
The fliF (Arora et al., 1996), mutant constructed by insertion of an antibiotic resistance marker into the coding region has been previously described. For the construction of the flgE and fleP mutants, a gentamicin cassette was inserted into a unique restriction site in the coding regions of the genes thereby disrupting the reading frame. This was achieved by allelic exchange of the wild-type chromosomal copy of the gene for a disrupted plasmid copy based on homologous recombination. pBS-flgEGm and pKΔ4.5PGm containing the disrupted plasmid copy for flgE and fleP respectively were electroporated into the PAK strain and selected for double crossovers displaying gentamicin resistance, thus generating the flgE and fleP mutants respectively.
All the mutants were confirmed by PCR analysis for their respective deletion or insertion and tested in a motility assay for non-motility.
Rapid gene disruption
Candidate flagellar genes PA1967 and PA3713 were mutagenized by a previously described technique (Wolfgang et al., 2003). Briefly, an internal fragment of each target gene was amplified by PCR an tailed with attB1 and attB2 sequences by inclusion in the PCR primers. The products were cloned using Gateway Technology (Invitrogen) in pEXGmGW. Chromosomal integration was selected as described (Hoang et al., 1998) and maintained by the propagation in the presence of gentamicin (75 µg ml−1).
Construction of promoter-lacZ fusions in pDN19lacΩ
Putative promoter containing regions (500–650 bp) located upstream of various genes including, flgB, flgF, fleL, flgA, cheY, cheA and cheV were amplified from the PAK genome by PCR using Pfu polymerase (Stratagene) and the products cloned in the appropriate orientation upstream of the promoterless lacZ in pDN19lacΩ yielding placΩflgB, placΩflgF, placΩfleL, placΩflgA, placΩcheY, placΩcheA, and placΩcheV, respectively. A 5′ and a 3′ primer designed to amplify the respective promoter regions of each gene is listed in Table 3. The results of the promoter studies with the chemotaxis genes have not been addressed in this study. The putative fliK promoter region was cloned into the mini-CTXlacZ vector upstream of the promoterless lacZ, resulting in pfliK-lacZ. This plasmid was integrated into the attB site of the various strains tested. The rest of the promoter constructs used in this study were available in the laboratory.
RNA extraction, labelling and hybridization
Overnight bacterial cultures were diluted to a starting OD600 of 0.01 in 100 ml of LB medium and grown with shaking (300 r.p.m.) at 37°C until an OD600 of 0.5. Cells harvested by centrifugation, RNA was isolated and cDNA was synthesized according to the Affymetrix Expression Analysis Protocol Guide. Briefly, 12 µg of RNA from each sample was converted to cDNA using Superscript II Reverse transcriptase (Invitrogen) and a semirandom decamer NS5. The reaction also included RNA spike controls provided by the Cystic Fibrosis Foundation. The resulting cDNA was purified, partially digested with DNase I and end-labelled with ddUTP-biotin according to the Expression Analysis Protocol (Affymetrix). The resulting targets were hybridized to GeneChip P. aeruginosa Genome Arrays. The arrays were washed, stained and scanned according to the manufacturer's recommendation.
Hybridization intensity data was extracted from the scanned array images and intrachip normalizations were performed using Affymetrix Microarray Suite 5.0 software.
Static cultures were grown overnight at 37°C with the respective antibiotics when needed. A drop of the culture was allowed to adhere to a carbon-coated grid for 10 s, drained off, the grid rinsed in a drop of saline and adherent cells negatively stained with a 2% aqueous solution of phosphotungistic acid for 10 s. Samples were examined with a Hitachi H-7000 transmission electron microscope. The average length of the type IV pilus was determined by measuring the lengths of 10–12 pili on an electron microscope.
Site directed mutagenesis
Primer pairs fliAmut1and fliAmutcomp2 were used in the Quick-change site directed mutagenesis kit (Stratagene, La Jolla, CA) to generate pBS-mutfliApr according to the protocol provided in the kit. Briefly, 20 ng of column purified (plasmid mini kit, Qiagen, Valencia, CA) plasmid template (pBS-fliApr) was used in a 50 µl amplification reaction containing 1 µl of dNTPs, 1 µl of Pfu polymerase, 1.25 µl of each primer and 5 µl of reaction buffer. It was subjected to a cycling profile of initial denaturation for 30 s at 95°C, followed by 13 cycles of denaturation (95°C for 30 s), annealing (60°C for 1 min) and extension (68°C for 8 min). The contents were then treated with DpnI to digest the original plasmid template. One microlitre of the post-digestion amplification reaction was used to transform E. coli XL1 blue cells and transformants were selected on LB ampicillin (100 µg ml−1) plates. A clone with the desired site-specific mutagenesis confirmed by sequencing using the T7 primer was subsequently used for further characterization. The insert from pBS-mutfliApr was cloned upstream of the promoterless lacZ in pDN19lacΩ in the right orientation to yield placΩmutfliA.
The promoter-lacZ fusions and the vector pDN19lacΩ were electroporated into wild-type PAK and mutants PAK-ΔFleQ, PAK-ΔFleR and PAK-ΔFliA. The strains were grown to late log phase (A600 of 0.7–1.0) in LB medium containing either streptomycin or tetracycline and their β-galactosidase activity assayed (Miller, 1972).
SDS-polyacrylamide gel electrophoresis (PAGE) and Western analysis
Three microlitres of the bacterial lysate (cell pellet from 1.0 ml of bacterial culture resuspended and boiled in 100 µl of SSB) was resolved on a 10% polyacrylamide gel in duplicate. Proteins in one set were stained with Coomassie brilliant blue and from the other transferred to a PVDF membrane (Sambrook et al., 1989). The blot was developed using rabbit anti-FliC antibodies as the primary antibody and alkaline phosphatase conjugated anti-rabbit IgG as the secondary antibody. NBT and BCIP served as the substrate for the colour reaction.
This work was supported by NIH grant AI45014 to R.R. GeneChipâPseudomonas aeruginosa Genome Arrays were subsidised by Cystic Fibrosis Foundation Therapeutics. M.C.W. was supported by a Cystic Fibrosis Foundation Post-doctoral Research Fellowship. A.L.G. is a Howard Hughes Medical Institute Predoctoral Fellow.