Genetic Characterization of Conserved Charged Residues in the Bacterial Flagellar Type III Export Protein FlhA

For assembly of the bacterial flagellum, most of flagellar proteins are transported to the distal end of the flagellum by the flagellar type III protein export apparatus powered by proton motive force (PMF) across the cytoplasmic membrane. FlhA is an integral membrane protein of the export apparatus and is involved in an early stage of the export process along with three soluble proteins, FliH, FliI, and FliJ, but the energy coupling mechanism remains unknown. Here, we carried out site-directed mutagenesis of eight, highly conserved charged residues in putative juxta- and trans-membrane helices of FlhA. Only Asp-208 was an essential acidic residue. Most of the FlhA substitutions were tolerated, but resulted in loss-of-function in the ΔfliH-fliI mutant background, even with the second-site flhB(P28T) mutation that increases the probability of flagellar protein export in the absence of FliH and FliI. The addition of FliH and FliI allowed the D45A, R85A, R94K and R270A mutant proteins to work even in the presence of the flhB(P28T) mutation. Suppressor analysis of a flhA(K203W) mutation showed an interaction between FlhA and FliR. Taken all together, we suggest that Asp-208 is directly involved in PMF-driven protein export and that the cooperative interactions of FlhA with FlhB, FliH, FliI, and FliR drive the translocation of export substrate.


Introduction
The flagellum of Salmonella enterica is a supermolecular motor powered by an electrochemical potential difference of protons (PMF) across the cytoplasmic membrane. The flagellum consists of at least three parts: the basal body, the hook, and the filament. Flagellar assembly begins with the basal body, followed by the hook and finally the filament. Almost all the substructures of the flagellum lie beyond the cytoplasmic membrane. Most of flagellar proteins are transported to the distal end of the growing flagellum by the flagellar type III protein export apparatus [1][2][3][4]. The components of the export apparatus are highly homologous not only to those of the type III secretion system of pathogenic bacteria, which directly injects virulence effectors into eukaryotic host cell [5] but also to those of FOF1-ATP synthase, which consists of a water soluble F1 part, which is a ring complex having three catalytic sites for ATP synthesis/hydrolysis, and a membrane-integrated FO part, which mediates proton translocation [6][7][8].
The flagellar type III protein export apparatus consists of three soluble proteins (FliH, FliI, FliJ) and six integral membrane proteins (FlhA, FlhB, FliO, FliP, FliQ, FliR) ( Figure  S1) [9,10]. The export apparatus is believed to be located in the putative central pore of the basal body MS ring [11][12][13]. FliI is an ATPase [14] and forms a complex with FliH and FliJ [7,10,15,16]. FliI and FliJ bind to the FlgN-FlgK and FliT-FliD chaperone-substrate complexes [17][18][19]. The FliH-FliI-FliJ delivers export substrates to the export gate complex made up of the six integral membrane proteins [20,21]. A specific interaction of the FliH X -FliI 6 -FliJ ring complex with the docking platform formed by the cytoplasmic domains of FlhA and FlhB induces the initial entry of the substrates into the gate [7,22,23]. The export gate complex utilizes PMF across the cytoplasmic membrane as the energy source for the translocation of the substrates [22,24]. It has been shown that a fliO null mutant displays a weakly motile phenotype, suggesting that FliO is not directly involved in flagellar protein export [25]. Interestingly, a homologue of FliO is apparently absent in some other type III secretion systems [25].
FlhA is composed of an N-terminal integral membrane domain with eight predicted transmembrane (TM) helices (FlhATM, residues 1-327, 34.5 kDa) and a C-terminal cytoplasmic domain (FlhAC, residues 328-692, 40.5 kDa) ( Figure 1) [26]. A wellconserved hydrophilic cytoplasmic loop between TM-4 and TM-5 is indispensable for FlhA function, but little is known about its role in flagellar protein export [27]. FlhAC interacts with FliH, FliI, FliJ, the C-terminal cytoplasmic domain of FlhB (FlhBC), and the FliS-FliC and FliT-FliD chaperone-substrate complexes and initiates the translocation of the substrates [10,23,28,29]. FlhAC consists of four subdomains (D1, D2, D3, and D4) and a linker connecting FlhAC to FlhATM (Figure 1) [29][30][31]. The linker is involved in an interaction with FliJ [29]. The D2 subdomain is responsible for an interaction with the FliT-FliD and FliS-FliC complexes [29]. The D4 subdomain is dispensable for its function but is involved in the substrate specificity switching of the export apparatus [32]. Although FlhATM is required for the association of FlhA with the MS ring [12], it remains unknown whether it is directly involved in the PMFdriven protein export process.
The flagellar type III protein export apparatus shows many similarities with FOF1-ATP synthase [3,[6][7][8]. An inward-directed proton translocation through the FO part drives the rotation of the FO-c-ring as a rotor, and the c and e subunits act as a drive shaft to cause conformational changes in the F1 part that result in ATP synthesis. ATP hydrolysis by F1 drives the reverse rotation of the rotor, resulting in an outward-directed proton pumping. Two highly conserved charged residues, Arg-210 in the a subunit and Asp-61 in the c subunit of the E. coli FOF1-ATP synthase are critical for proton translocation [33]. Since FlhATM is the main component of the PMF-driven flagellar protein export gate, there is the possibility that the highly conserved charged residues of FlhATM may be involved in the energy transduction mechanism.
In order to clarify the role of FlhATM in PMF-driven flagellar protein export, we performed genetic analyses of eight highly conserved charged residues in the putative juxta-and transmembrane helices of FlhA. We show that only a negatively charged residue at position 208 of FlhA is critical for PMF-driven protein export and that most of substitutions are tolerated by the presence of FlhB, FliH and FliI. We also show an interaction between FlhA and FliR.

Alanine mutagenesis of conserved charged residues of FlhATM
To test the hypothesis that highly conserved charged residues of FlhATM is involved in the energy transduction mechanism, we identified eight highly conserved charged residues, Asp-45, Arg-85, Arg-94, Lys-203, Arg-206, Asp-208, Asp249 and Arg-270, in putative juxta-and trans-membrane helices of FlhA by multiple sequence alignment (Figures 1 and S2) and then replaced each with alanine. Immunoblotting with polyclonal anti-FlhAC antibody detected all the point mutant variants at the wild-type level (Figure 2A), indicating that these substitutions do not affect protein stability.
We next analyzed the motility of the flhA null mutant expressing each of the FlhA point mutant variants in soft agar plates ( Figure 2B). FlhA(R206A) and FlhA(R270A) fully complemented, FlhA(D45A) and FlhA(R85A) restored motility to a significant degree, and FlhA(K203A) to some degree. However, mutants with alanine substitution for Arg-94, Asp-208 and Asp-249 did not complement at all. To test whether their poor motility is due to their reduced export activity, we prepared the culture supernatants from the overnight culture of these flhA point mutants and analyzed the secretion levels of the hook-capping protein FlgD by immunoblotting with polyclonal anti-FlgD antibody ( Figure 2C). FlgD was detected at the wild-type level in the culture supernatants from the D45A, R85A, K203A, R206A, and R270A mutants but not from the R94A, D208A, and D249A mutants. FlhA(R94A), FlhA(D208A) and FlhA(D249A) also inhibited motility when expressed in the wild-type strain ( Figure 2D), indicating that these mutant proteins exert a negative dominance. This suggests that they can be incorporated into the export apparatus. Therefore, we conclude that Arg-94, Asp-208, and Asp-249 are critical for protein export.

Importance of the charge of critical residues
To probe the role of Arg-94, Asp-208 and Asp-249 in protein export, we mutated these three residues to the following two types: the same charge with a different length of side chain (Arg-to-Lys and Asp-to-Glu); and the oppositely charged residue (Arg-to-Asp and Asp-to-Lys). Immunoblotting with the polyclonal anti-FlhAC antibody revealed that all the mutant variants were as stable as the wild-type ( Figure 3A). At positions 94 and 249, neither types of mutations affected the secretion level of FlgD ( Figure 3B, lanes 3, 4, 8 and 9), indicating that these residues maintain the function of FlhA regardless of the charge type. While the D208E replacement still permitted the export of FlgD at the wild-type level ( Figure 3B, lane 5) the D208K and D208N mutations totally diminished the export (lanes 6 and 7). In agreement with these results, the D208K and D208N mutants as well as the flhA null mutant harboring the vector control accumulated much higher amounts of FlgD in the cytoplasm than the wild-type while the others did so more or less at the wild-type levels. FlhA(D208K) and FlhA(D208N) also exerted a dominant negative effect on wild-type motility (data not shown), indicating that they retain the ability to be incorporated into the export apparatus. These results suggest that a negatively charged residue at position 208 of FlhA is essential for the export function.

Protein secretion rate of slow motile mutants
We found that the motility of the flhA(D208E) (data not shown) and flhA(K203A) ( Figure 2B) mutants was worse than that of wildtype cells in soft agar plates whereas the levels of FlgD secretion by these mutants were at the wild-type level when grown overnight in ) and a C-terminal cytoplasmic domain (FlhAC) whose atomic structure was solved by X-ray crystallography [41]. FlhAC is involved in the early process of flagellar protein export along with FliH, FliI, FliJ, and the C-terminal cytoplasmic domain of FlhB. Closed circles in red and blue indicate invariant acidic and basic residues, respectively, which were identified by multiple sequence alignment of FlhA homologs ( Figure S1). doi:10.1371/journal.pone.0022417.g001 LB ( Figure 3B, lane 5, and Figure 2C, lane 6, respectively). These results raise the possibility that these mutants are slow secretors. Secretion rate measurement of the flagellar proteins requires the external onset control of flagellar gene expression. To do this, we inserted a Tn10d (T-POP) transposon upstream of the flagellar master flhDC operon, which is required for the expression of the entire flagellar regulon. As flhDC is transcribed from a tetracycline-inducible promoter PtetA only in the presence of tetracycline, flagellar gene expression can be externally controlled [34]. A flhA null mutant containing T-POP was transformed with pUC19-based plasmids encoding wild-type FlhA or each of the FlhA mutants. The transformants were grown at 30uC in LB until OD600 reached ca. 0.5,0.6. After induction with tetracycline, the culture supernatants were collected at regular time intervals and analyzed by immunoblotting with polyclonal anti-FlgD antibody ( Figure 4).
For wild-type FlhA, FlgD was detected at 60 min after induction and the amount of FlgD gradually increased as the incubation was continued. The D208E mutation caused a 15 min delay in the detection of FlgD. In agreement with this, when comparing the length of flagellar filaments labeled with a fluorescent dye, the filaments of these mutants were significantly shorter than those of the wild-type at 120 minutes after tetracycline induction (data not shown). When the incubation was further continued for 12 hours, the filaments grew nearly to the wild-type level. These results suggest that the D208E mutation reduces the rate of protein export. For the flhA(K203A) mutant, FlgD was not detected until 120 min after induction, indicating that this is a much slower secretor than the flhA(D208E) mutant.
Effect of flhA mutations on flagellar protein export in the DfliH-fliI double null mutant background It has been previously reported that a DfliH-fliI double null mutant is weakly motile [22]. We tested if the flhA mutations that did not affect protein export are also tolerated in the absence of FliH and FliI ( Figure 5). The weakly motile phenotype of the DfliH-fliI mutant was totally abolished by the flhA mutations ( Figure 5A), indicating that these mutant FlhA proteins have a more synergistic phenotype when combined with the flH-fliI double null mutation.
Minamino & Namba [22] isolated a DfliH-fliI bypass mutant whose second-site P28T mutation in FlhB significantly increases the probability of substrate entry into the gate in the absence of FliH and FliI. Therefore, we tested if the mutant FlhA proteins can work in this DfliH-fliI bypass mutant ( Figure 5B). Wild-type FlhA restored the motility of the DfliH-fliI flhB(P28T) DflhA mutant, in agreement with a previous report [22]. However, only FlhA(R270A) complemented this mutant to some degree although markedly less than the wild-type level, and the others did not show any complementation at all. We next investigated the effect of the flhB(P28T) mutation by itself on the function of these mutant FlhA proteins. In agreement with a previous report [22], wild-type FlhA fully restored the motility of a DflhA flhB(P28T) mutant ( Figure 6A). The presence of FliH and FliI allowed FlhA(D45A), FlhA(R85A), FlhA(R270A) and FlhA(R94K) to exert their function considerably even in the presence of the flhB(P28T) mutation but none of the remaining mutants ( Figure 6A and B). These results indicate a degree of defectiveness for the individual flhA alleles.
Tryptophan mutation can probe the structure and environment of transmembrane domains of membrane proteins [35][36][37]. We performed tryptophan mutagenesis to test if the DfliH-fliI bypass flhB(P28T) mutation and the FliH-FliI complex affect the structural architecture of FlhATM. The flhA(K203W) and flhA(R270W) mutants in the otherwise wild-type background remained functional mostly and to some degree, respectively, while the others were nonfunctional ( Figure S3). Since tryptophan substitution for residues facing the lipid preserves the function but is likely to disrupt the structure and function for residues within or adjacent to transmembrane segments, the results indicate that Lys-203 and Arg-270 of FlhA are likely to face the lipid or waterexposed channel of the export gate complex. We then analyzed the effect of the K203W and R270W mutations on motility of the DfliH-fliI flhB(P28T) mutant and the flhB(P28T) mutant. The motility of these mutants was significantly reduced compared with the wild-type ( Figure 6C). Since Lys-203 and Arg-270 are predicted to be located in the conserved cytoplasmic loop between TM-4 and TM-5 and within TM-6, respectively (Figure 1), these residues may face the interior of the gate complex in the presence of the flhB(P28T) mutation.

Isolation of pseudorevertants from the DfliH-fliI flhB(P28T) flhA(K203W) mutant
To investigate how the K203W mutation significantly reduces the export activity in the DfliH-fliI bypass mutant, gain-of-function mutants were isolated from the DfliH-fliI flhB(P28T) flhA(K203W) mutant. In total three pseudorevertants were obtained with significantly improved motility although not as good as that of the wild-type ( Figure 7A). DNA sequencing revealed that the second-site mutations are located in FliR, and they were G103A, G103C and G117D ( Figure 7B). These mutated residues are not conserved among FliR homologs ( Figure S4). They were all located in a cytoplasmic loop between TM-2 and TM-3, suggesting that the conserved loop between TM-4 and TM-5 in FlhA may interact with the cytoplasmic loop of FliR.
To test whether these extragenic fliR suppressors are allele specific for the parental flhA(K203W) mutation, we carried out P22-mediated transductional crosses. The DfliH-fliI DflhA flhB(P28T) fliR::Tn10 strain harboring the pUC19-based plasmid encoding each mutant FlhA protein was used as a recipient and the three pseudorevertants isolated were used as donors. A 50 ml solution of the overnight culture of the recipient and a 5 ml aliquot of phage lysates prepared from the donors were mixed together and then streaked on soft agar plates. After 40 hours incubation at 30uC, we looked for motility halos resulting from homologus recombination between the fliR::Tn10 and the suppressor fliR alleles. Motility halos appeared from the recipient expressing  To test the effect of each second-site fliR mutation by itself on motility, we isolated the second-site fliR mutants and analyzed their motility on soft agar plates. In all cases motility was nearly normal, just slightly less than the wild-type ( Figure S5B), indicating that they do not affect the function of FliR significantly.

Discussion
FlhA C not only acts as part of the docking platform of the export gate complex for the soluble export components, FliH, FliI, and FliJ, but also plays a critical role in PMF-driven flagellar protein export along with FliH, FliI, FliJ and FlhB C [23]. However, little is known about the role of FlhA TM in PMF-driven flagellar protein export. Here, we carried out genetic analysis of eight highly conserved charged residues of FlhA TM and provide evidence suggesting that Asp-208 is critical for the translocation of export substrate and that the cooperative interactions of FlhA with FlhB, FliH, FliI, and FliR facilitates PMF-driven flagellar protein export.
Asp-208 of FlhA is essential for flagellar protein export Two conserved charged residues of E. coli F O F 1 -ATP synthase, Arg-210 of the a subunit and Asp-61 of the c subunit, are critical for proton flow though F O , which is coupled with ATP synthesis/ hydrolysis by F 1 [38][39][40]. Asp-61 is directly involved in proton translocation [39]. Arg-210 is essential for preventing the shortcircuiting of proton flow without c-ring rotation in the F O proton channel [40]. Since there are many similarities between the flagellar type III protein export apparatus and F O F 1 -ATP synthase [3], the highly conserved acidic and basic residues of the transmembrane components of the flagellar export apparatus may be directly involved in the inwardly-directed proton flow. If so, the function of putative proton-binding sites in the export gate is presumably critical. In this study, we probed the function of highly conserved charged residues in trans-or juxta-transmembrane helices of FlhA by mutating these charged residues and found that only Asp-208 of FlhA is critical for the functioning of the export apparatus (Figure 2 and 3). The D208A, D208W, D208K, and D208N mutant alleles were dominant to the wildtype flhA gene and hence totally inhibited motility of wild-type cells. This suggests that these mutants retain the ability to be incorporated into the export apparatus. Because only the conservative D208E replacement permitted any function, the important feature of this residue appears to be either a negative side-chain charge or the ability to bind a proton. Asp-208 of FlhA is predicted to be located in the cytoplasmic juxtamembrane region (Figure 1), whereas both Arg-210 of a subunit and Asp-61 of c subunit are located at the center of a transmembrane helix. This suggests that Asp-208 in FlhA may not directly be involved in the proton translocation mechanism. But since Asp-208 is involved in PMF-driven flagellar protein export, we assume that a negative side-chain charge of Asp-208 may be required for an electrostatic interaction with other export component(s) and/or export substrates.

A degree of defectiveness for the individual flhA alleles is dependent on the presence of FlhB, FliH and FliI
In this study, we found that most of substitutions of evolutionally conserved charged residues except for Asp-208 were tolerated when FliH and FliI were present. In contrast, these substitutions totally abolished a weakly motile phenotype of a DfliH-fliI double null mutant. The P28T mutation in FlhB increases the probability of export substrate entry into the narrow pore of the export gate in the absence of FliH and FliI [22]. Here, we also found that those FlhA mutants did not function in the DfliH-fliI flhB(P28T) bypass mutant ( Figure 5B). FliH and FliI overcame the D45A, R85A, R94K and R270A defects even in the presence of the flhB(P28T) allele ( Figure 6A), indicating that the flhB(P28T) significantly affects the function of these mutant FlhA proteins and that FliH and FliI restore their function. Taken all together we suggest that these mutant FlhA proteins require the FliH-FliI complex and FlhB to exert their export function.
Tryptophan substitution at positions 203 and 270 in FlhATM significantly reduced the export activity when the FlhB(P28T) mutation was present ( Figure 6C). Tryptophan substitution is assumed to preserve the protein function when the mutated residues face lipid or water but to disrupt structure and function when the residues face adjacent protein segments [35][36][37]. Mutant FlhA(K203W) and FlhA(R270W) proteins retained the FlhA function when FlhB, FliH, and FliI were present ( Figure S3), indicating that the orientation of the side-chain of these tryptophan residues is significantly affected by FlhB, FliH and FliI. Therefore, we propose that cooperative interactions among FlhA, FlhB, FliH, FliI and export substrate may induce conformational changes in FlhATM to facilitate PMF-driven protein export.

Interaction of the well-conserved cytoplasmic loop between TM-4 and TM-5 of FlhA with a cytoplasmic loop of FliR
A highly conserved hydrophilic cytoplasmic loop is located between TM-4 and TM-5 of FlhA, but little is known about its role in flagellar protein export [27]. In this study, we found that the reduced motility of the DfliH-fliI flhB(P28T) flhA(K203W) mutant was significantly improved by extragenic mutations in the cytoplasmic loop between TM-2 and TM-3 of FliR ( Figure 7). Interestingly, these suppressors were allele specific for the flhA(K203W) mutation, suggesting an interaction between the conserved cytoplasmic loop of FlhA and the cytoplasmic loop of FliR. The FliR-FlhB fusion protein is partially functional, suggesting that FliR interacts with FlhB [13]. Therefore, it is likely that FlhA, FlhB and FliR are very close to each other in the export gate complex. This is supported by our finding that the P28T bypass mutation, which is close to the interface of TM-1 of FlhB, significantly affects the function of mutant FlhA.
The K203A mutation considerably reduced the secretion rate of FlgD in a way similar to the D208E mutation ( Figure 4). Since Lys-203 is likely to be close to Asp-208, we propose that the precise conformation of a negatively charged residue at position 208 of FlhA may be important for protein export and that the interaction of the conserved cytoplasmic loop of FlhA with the cytoplasmic loop of FliR is required for maintaining this conformation.

Bacterial strains, plasmids and media
Bacterial strains and plasmids used in this study are listed in Table 1. L-broth and soft tryptone agar plates were used as described previously [9,10]. Tetracycline-sensitive (Tc S ) plates, in which the tetracycline-resistant cells cannot grow, were prepared as described by Maloy and Nunn [41]. Ampicillin and tetracycline were added to LB at a final concentration of 100 mg/ml and 15 mg/ml, respectively.

Transductional crosses and DNA manipulations
P22-mediated transductional crosses were carried out using p22HTint as described previously [42]. DNA manipulations were carried out as described before [10]. Site-directed mutagenesis was carried out using QuickChange site-directed mutagenesis method as described in the manufacturer's instructions (Stratagene) with a plasmid, pMMHA004 (pUC19/His-FlhA), as a template DNA. The pairs of oligonucleotides shown in Table 2 were used to construct each FlhA mutant. All of the flhA mutations were confirmed by DNA sequencing. DNA sequencing reactions were carried out using BigDye v3.1 as described in the manufacturer's instructions (Applied Biosystems), and then the reaction mixtures were analyzed by a 3130 Genetic Analyzer (Applied Biosystems).

Construction of the flhA null strains
To delete the flhA gene form the chromosome, we used the l Red homologous recombination system developed by Datsenko and Wanner [43]. First, the entire flhA gene was replaced with the tetRA genes to construct a DflhA::tetRA strain as described [44]. Then, to remove the tetRA genes from the chromosome, the 112bp DNA was first synthesized and then was purified using a QIAquick PCR purification kit (QIAGEN). The DflhA::tetRA strain carrying pKD46, which has a temperature sensitive replicon [43], was grown in 5-ml L-broth containing ampicillin and 0.2% Larabinose at 30uC until OD 600 had reached 0.6. The cells were washed three times with ice-cold H 2 O and suspended in 50 ml of ice-cold H 2 O. 50 ml of cells were electroporated with 100 to 200 ng of purified 112-bp DNA using 0.1-cm cuvettes at 1.8 kV. Shocked cells were added to 1 ml SOC, and incubated for 1 h at 37uC. Then, one-half were spread onto Tc S plates and incubated at 42uC overnight. The constructs were conformed by DNA sequencing.

Motility assay
Fresh transformants were inoculated onto soft tryptone agar plates containing ampicillin and incubated at 30uC.

Preparation of whole cell and culture supernatant fractions and immunoblotting
Salmonella cells were grown at 30uC with shaking until the cell density had reached an OD 600 of ca. 1.4-1.6. Aliquots of culture proteins containing a constant number of cells were clarified by centrifugation. Cell pellets were resuspended in SDS-loading buffer normalized by cell density to give a constant amount of cells. The proteins in the culture supernatants were precipitated by 10% trichloroacetic acid (TCA), suspended in a Tris-SDS loading buffer and heated at 95uC for 5 min. After sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE), immunoblotting with polyclonal anti-FlgD or anti-FlhA C antibody was carried out as described previously [9]. Detection was done with an ECL immunoblotting detection kit (GE Healthcare).

Tetracycline-induced protein export rate assay
Salmonella cells with a T-POP insertion between the flhDC promoter and the transcription start site [34], were grown overnight in LB culture at 30uC and diluted 50-fold into 30 ml of fresh LB. When the cell density reached an OD 600 of 0.5, tetracycline was added to a final concentration of 15 mg/ml. Samples were taken at 0, 15, 30, 45, 60, 75, 90, 120 and 180 min after tetracycline addition. Proteins in the culture supernatant fraction were prepared by TCA precipitation as described above.    Figure S1 Schematic diagram of the bacterial flagellar type III protein export apparatus. The flagellar type III protein export apparatus consists of three soluble proteins, FliH, FliI, and FliJ and six integral membrane proteins, FlhA, FlhB, FliO, FliP, FliQ, and FliR. The integral membrane proteins are postulated to be located within the central pore of the MS ring and form the PMF-driven export gate complex. FlhA and FlhB have large cytoplasmic domains which project into the cavity within the C ring and form the docking platform for FliH, FliI, and FliJ as well as the substrates. FliI forms a heterotrimer with the FliH dimer in the cytoplasm. The FliH2FliI complex, along with FliJ and export substrates, is localized to the basal body C-ring through a specific interaction between FliH and FliN. FliI hexamerizes upon docking of the FliH-FliI-FliJ-substrate complex to the FlhA-FlhB platform and facilitates the entry of the N-terminal segment of a substrate into the gate. ATP hydrolysis by the FliI hexamer induces the dissociation of the FliHX-FliI6-FliJ complex from the gate. The export gate utilizes PMF across the cytoplasmic membrane as the energy source for the translocation of the export substrates into the central channel of the growing flagellar structure. (TIF)