Control of surface organelle number and placement is a crucial aspect of the cell biology of many Gram-positive and Gram-negative bacteria, yet mechanistic insights into how bacteria spatially and numerically organize organelles are lacking. Many surface structures and internal complexes are spatially restricted in the bacterial cell (e.g. type IV pili, holdfasts, chemoreceptors), but perhaps none show so many distinct patterns in terms of number and localization as the flagellum. In this review, we discuss two proteins, FlhF and FlhG (also annotated FleN/YlxH), which control aspects of flagellar assembly, placement and number in polar flagellates, and may influence flagellation in some bacteria that produce peritrichous flagella. Experimental data obtained in a number of bacterial species suggest that these proteins may have acquired distinct attributes influencing flagellar assembly that reflect the diversity of flagellation patterns seen in different polar flagellates. Recent findings also suggest FlhF and FlhG are involved in other processes, such as influencing the rotation of flagella and proper cell division. Continued examination of these proteins in polar flagellates is expected to reveal how different bacteria have adapted FlhF or FlhG with specific activities to tailor flagellar biosynthesis and motility to fit the needs of each species.
Flagella are complex molecular machines that facilitate swimming motility in many bacteria. Bacteria evolved elegant transcriptional control mechanisms to ensure temporal production of subsets of flagellar proteins conducive for proper flagellar biosynthesis (for extensive reviews, see Chilcott and Hughes, 2000; Chevance and Hughes, 2008). Although core flagellar components and basic principles for regulating flagellar gene expression are fairly conserved across motile bacteria, bacteria produce different flagellation patterns characteristic for each species. These patterns range from peritrichous flagellation (production of many flagella along the cell body) to polar flagellation (production of flagella exclusively at bacterial poles). Among polar flagellates, flagellation patterns include monotrichous (a single flagellum at one pole), amphitrichous (a flagellum at each pole), or lophotrichous flagellation (a few flagella at a single pole). Determining how bacteria – particularly polar flagellates – exquisitely control spatial and numerical parameters of flagellar biosynthesis is one of the more recent challenges in the field of bacterial motility. Two proteins implicated in spatial and numerical control of flagellar biosynthesis in polar flagellates are the FlhF GTPase and FlhG (a putative ATPase also annotated as FleN or YlxH). Both FlhF and FlhG are members of the SIMIBI class of nucleotide-binding proteins, which often are involved in partitioning or localizing other factors within prokaryotic or eukaryotic cells (Leipe et al., 2002). Recent investigation has revealed that FlhF and FlhG are produced by Bacillus subtilis and influence spatial aspects of peritrichous flagellation in this bacterium (Guttenplan et al., 2013). However, these proteins are not produced by other peritrichous organisms, such as Escherichia coli or Salmonella species, which raises questions regarding how broadly these proteins influence peritrichous flagellation.
Initial studies in Vibrio cholerae and Pseudomonas aeruginosa suggested that FlhF is involved in spatial regulation of flagellar biosynthesis whereas FlhG numerically controls flagellation (Dasgupta et al., 2000; Dasgupta and Ramphal, 2001; Correa et al., 2005; Murray and Kazmierczak, 2006; Green et al., 2009). However, more recent studies have shown that FlhF and FlhG interact in some species, namely Campylobacter jejuni, Vibrio alginolyticus and B. subtilis (Parrish et al., 2007; Kusumoto et al., 2008; Bange et al., 2011). Thus, these interactions may influence FlhF and FlhG activities in vivo so that each protein can function properly to regulate spatial and numerical parameters of flagellar biosynthesis. Furthermore, data are accumulating that FlhF–FlhG cognate pairs likely possess differing biochemical properties, which may contribute to the different flagellation phenotypes across bacterial species. In this review, we attempt to summarize previous findings regarding the activities of FlhF and FlhG proteins and present new questions for future investigations of how FlhF, FlhG, or pairs of these proteins may mediate spatial and numerical parameters of flagellar biosynthesis and other cellular activities in different polar flagellates.
Biochemical properties of FlhF and FlhG
FlhF, the signal sequence-binding protein Ffh (SRP54 in Archaea and Eukarya) and the signal-recognition particle (SRP) receptor FtsY (SRα in Eukarya) comprise the SRP-GTPase subfamily of the SIMIBI class of nucleotide-binding proteins (Leipe et al., 2002). In biology, GTPases often function as molecular switches, in which transitions between inactive GDP-bound and active GTP-bound conformations are integral to the regulation of many fundamental cell processes. Ffh (in complex with SRP RNA) and FtsY together constitute the universally conserved machinery that targets ribosomes in complex with nascent proteins to membrane channels for translocation (reviewed in Doudna and Batey, 2004). When bound to GTP, the Ffh and FtsY proteins form a heterodimer with a composite active site (Egea et al., 2004). GTP hydrolysis is reciprocally activated within this heterodimer, and SRP RNA and membrane lipids play crucial roles in this activation process (Zhang et al., 2009). Recent elegant work has revealed how dimerization and conformational changes of Ffh and FtsY are integrated with binding of ‘cargo’ (complexes of ribosome and nascent chain protein) and its delivery to the membrane translocation machinery (Saraogi et al., 2011). Much less is known, however, about the regulation and function of the FlhF GTPase.
FlhF proteins share significant homology with Ffh and FtsY within their N and G domains, but have N-terminal extensions (or B domains) that are unique (Fig. 1A). The first and only crystal structure to date of a FlhF family member was solved in 2007 (Bange et al., 2007). In this study, the crystal structure of B. subtilis FlhF lacking the B domain (NG–FlhF) formed a composite active site via a dimer interface composed entirely by interactions between G domains. In contrast, the FtsY–Ffh heterodimer interface is mediated by N domain interactions from each protein and regulated by three motifs that are absent from FlhF.
FtsY and Ffh both possess conserved arginines in the G2 subdomain of the GTPase domain that form an ‘arginine finger’ to stimulate GTP hydrolysis (Egea et al., 2004). Asymmetry of the side-chain conformation of these two arginines within the heterodimer may allow for sequential hydrolysis of nucleotide. No such asymmetry exists within the NG–FlhF homodimer, and the arginine finger is disordered within the GTP bound structure (Bange et al., 2007; Fig. 1A). Subsequently, Bange et al. demonstrated that FlhG (originally annotated as YlxH) associates with the GTP-bound NG–FlhF dimer and stimulates its GTPase activity approximately fivefold (Bange et al., 2011). Such regulation of GTPase activity by external factors, though not observed for Ffh–FtsY, is commonly described for eukaryotic signalling GTPases (Bourne et al., 1991). A co-crystal structure consisting of NG–FlhF and the N-terminal 23 residues of FlhG, which are necessary and sufficient for stimulation of GTPase activity, revealed that the Gln8 side-chain of FlhG inserts into the composite active site of the NG–FlhF homodimer and repositions the G2 arginine finger residue to stabilize the transition state geometry of the nucleotide substrate (Bange et al., 2011; Fig. 1B). A similar mechanism of activation by regulator of G protein signalling (RGS) proteins occurs for eukaryotic monomeric Gα proteins (Ross and Wilkie, 2000). RGS binding locks the flexible switch regions containing catalytic residues of Gα into their active conformation with a resulting GTP-hydrolysis rate enhancement of 100- to 1000-fold (Ross and Wilkie, 2000).
Even though FlhG stimulates the GTPase activity of B. subtilis FlhF in vitro, the consequences of this activation for peritrichous flagellation are not known. In polar flagellates, the ability of FlhG to influence the GTPase activity of FlhF has yet to be explored. Furthermore, the ATPase activity for any FlhG protein of a polar flagellate has yet to be verified. For many polar flagellates, it is not known whether FlhF–FlhG cognate pairs interact in vivo and if these interactions influence respective GTPase or ATPase activities to regulate the function of each protein in flagellar assembly or organization. It is possible that differences in specific domains of FlhF and/or FlhG in polar flagellates give rise to different in vivo activities such as nucleotide binding or hydrolysis that ultimately result in different flagellation patterns. Overall, the FlhF proteins that have been analysed in polar flagellates are most similar within their GTPase domains (27–39% identity; 49–62% similarity) and are more diverse in other regions of the proteins. Additionally, FlhG orthologues in polar flagellates display homology (30–59% identity; 55–79% similarity) across approximately 80% of the length of the proteins, with the putative ATPase domains being the most highly conserved (Fig. 1B). The FlhG N-terminal extension with the DQAxxLR motif encompassing the Gln8 residue is conserved among many FlhG orthologues (Bange et al., 2011; Fig. 1B). However, this N-terminal domain is lacking from the FlhG proteins of the Pseudomonads, Xanthomonads and Stenotrophomonads. Thus, it is unclear whether P. aeruginosa FlhF and FleN (the P. aeruginosa FlhG orthologue) do interact and whether FleN or some other protein stimulates the GTPase activity and biological function of FlhF. It is conceivable that FlhF–FlhG cognate pairs may interact in some polar flagellates but not in others, which consequently results in the different flagellation patterns produced by these bacteria.
Contributions of FlhF to polar flagellar biogenesis
FlhF appears to be required for proper spatial arrangement of flagella at the poles in many polar flagellates, but mutation of flhF results in additional effects on flagellation in some bacteria (Pandza et al., 2000; Murray and Kazmierczak, 2006; Balaban et al., 2009; Green et al., 2009). P. aeruginosa and Pseudomonas putida ΔflhF mutants produce a flagellum that is no longer restricted to the pole (Pandza et al., 2000; Murray and Kazmierczak, 2006). This misplaced flagellum rotates and powers swimming motility, but swimming paths are no longer straight and motility is impaired as viscosity increases (Murray and Kazmierczak, 2006). Whether this is simply due to the increased drag associated with movement off the longitudinal axis of the cell, or to alterations between the assembled flagellum and proteins that dynamically associate with it (e.g. flagellar motor components) is not known. In other organisms such as C. jejuni, V. alginolyticus and V. cholerae, deletion of flhF causes aflagellation in most cells, which is likely attributed to the requirement for FlhF to stimulate wild-type levels of flagellar gene expression (Hendrixson and DiRita, 2003; Correa et al., 2005; Kusumoto et al., 2008; Balaban et al., 2009; Green et al., 2009). However, it is not known why FlhF is required for transcription of certain flagellar genes. Of note, in a small number of V. cholerae flhF mutant cells that produced a flagellum, the flagellum was not localized to the pole, which was similar to the misplaced flagellum phenotype of a P. aeruginosa ΔflhF mutant (Murray and Kazmierczak, 2006; Green et al., 2009).
For monotrichous flagellation in V. cholerae, it is thought that a new flagellum is synthesized at the old pole of a daughter cell after division. Like in other polar flagellates, FlhF is targeted to the bacterial pole in V. cholerae (Murray and Kazmierczak, 2006; Kusumoto et al., 2008; 2009; Ewing et al., 2009; Green et al., 2009). Polar localization of FlhF in V. cholerae is likely a prerequisite for FlhF mediating the polar localization of FliF, which composes the inner membrane MS ring of the flagellum (Green et al., 2009). The N domain of V. cholerae FlhF was required, but not sufficient to localize FlhF to the pole. FlhF truncations lacking the N-terminal B domain or C-terminal G domain were polarly localized in V. cholerae, but these proteins did not support motility or recruit FliF to the pole, indicating that all domains are required for FlhF-dependent activities in flagellation (Green et al., 2009). Furthermore, bacteria expressing a FlhF mutant deficient in GTP binding accumulated a significant level of FliF at a pole, suggesting that the GTPase activity of FlhF is not essential for the polar targeting of FliF. However, it is unclear what FlhF may be recognizing at the pole for its localization or how FlhF mediates the polar localization of the MS ring.
It is not clear why deletion of flhF has such varied consequences for flagellar assembly in different polarly flagellated organisms. Many of these observations are consistent with a model in which FlhF positively influences flagellar assembly, presumably at the site where FlhF itself resides. The work in V. cholerae would argue that FlhF might exert its role in many polar flagellates by recruiting components such as FliF to form essential flagellar substructures necessary for initial stages in flagellar biosynthesis (Green et al., 2009). However, a single flagellum – albeit mislocalized – is consistently produced in flhF mutants in Pseudomonas species, which could reflect a lower barrier to flagellar assembly initiation in these organisms relative to other polar flagellates (Pandza et al., 2000; Murray and Kazmierczak, 2006).
Correlating FlhF GTPase activity and function
Bacillus subtilis, C. jejuni and P. aeruginosa FlhF proteins bind and hydrolyse GTP in vitro, indicating at least a low level of FlhF GTPase activity without any exogenous activating factors (Balaban et al., 2009; Bange et al., 2011; Schniederberend et al., 2013). Additionally, V. cholerae FlhF binds GTP, but GTPase activity has yet to be verified (Green et al., 2009). Many works have examined how amino acid mutations in FlhF predicted or documented to alter GTPase activity and/or GTP binding affect flagellation phenotypes or motility. In general, the observations are consistent with the idea that FlhF function, like that of eukaryotic GTPases, is determined by its nucleotide-bound state. However, mutations that alter GTP binding or hydrolysis by FlhF have vastly different effects on flagellar motility, assembly, number or placement in different polar flagellates, suggesting that different FlhF proteins may promote specific flagellation patterns or alter flagellar function that are suitable to the needs of each species.
In C. jejuni, mutation of D321 and R324 (the G2 arginine finger residue) in FlhF altered flagellar assembly and reduced GTPase activity, although it is unclear if these FlhF mutants were specifically defective in GTP binding or hydrolysis (Balaban et al., 2009; Fig. 1A). Unlike wild-type C. jejuni, which mainly produces bacteria with a single flagellum at one or both poles, these mutants demonstrated diverse flagellation phenotypes ranging from a significant number of wild-type flagellated bacteria, aflagellate bacteria and bacteria with non-polar flagella (Balaban et al., 2009). Additionally, a small number of bacteria produced extra polar flagella. One interpretation of these data is that the normal GTPase activity of FlhF is required for C. jejuni to exclusively construct polar flagella. In contrast, V. cholerae FlhF with D321A or R324A mutations, which are also predicted to reduce its GTPase activity, supported motility, while mutations in other residues (K295, D367 and D429) that reduced binding to GTP agarose abolished flagellar assembly and motility (Green et al., 2009). Of note, GFP fusions to FlhFK295A and FlhFD367A localized to the pole, suggesting that polar localization of FlhF was independent of GTP binding or hydrolysis, but GTP binding appears to be required to assemble polar flagella.
In P. aeruginosa, FlhF, but not its GTPase activity, is required for polar flagellar assembly. Instead, FlhF GTPase activity affects flagellar function for the swimming motility of P. aeruginosa. Point mutations that rendered purified P. aeruginosa FlhF enzymatically inactive in vitro (K222A and D294A in the G1 and G3 subdomains respectively), or altered its hydrolytic activity (R251G, the arginine finger residue) all supported polar flagellar assembly in the ΔflhF mutant background (Schniederberend et al., 2013; Fig. 1A). FlhFD294A did not bind GTP or dimerize, while FlhFK222A bound GTP, but dimerized less well than wild-type FlhF. Single-cell analysis of swimming velocity revealed that FlhFD294A supported faster swimming in a greater proportion of bacteria than FlhFK222A, while FlhFR251G-expressing bacteria were predominantly non-motile. FlhFR251G bound GTP and dimerized, but this mutant protein demonstrated a defect in GTP/GDP exchange, suggesting that the mutation causes a delay in homodimer dissociation and product release (Schniederberend et al., 2013). These findings are consistent with a model in which FlhF in a GTPase-independent manner functions as a ‘place setter’ for flagellar assembly, but the GTPase activity is required for flagellar function as a motility organelle, perhaps via regulating interactions with components of the flagellar motor. Any possible P. aeruginosa FlhF interaction partners associated with the flagellar motor, however, remain unknown. Taken together, these varied phenotypes of flhF mutants in polar flagellates highlight the diverse means by which FlhF in GTPase-dependent and -independent activities shapes flagellation patterns and/or flagellar function in individual polarly flagellated species.
Numerical regulation of flagellar biosynthesis
Initial studies of FlhG homologues in various polar flagellates suggested that the protein controls numerical parameters of flagellar biosynthesis. A P. aeruginosa mutant lacking fleN (an flhG orthologue) produced three to six flagella at a pole instead of only one (Dasgupta et al., 2000; Dasgupta and Ramphal, 2001). These flagella conferred movement, but not correct directional swimming. In V. cholerae, mutation of flhG also caused a switch from monotrichous to lophotrichous flagellation with eight to 10 flagella at a single pole and, in some cells, amphitrichous flagellation with flagella at both poles (Correa et al., 2005).
These P. aeruginosa and V. cholerae fleN/flhG mutants overexpressed many classes of flagellar genes, suggesting that hyperflagellation may be a result of increased flagellar protein production (Dasgupta et al., 2000; Dasgupta and Ramphal, 2001; Correa et al., 2005). In P. aeruginosa, FleN interacts with the FleQ master transcriptional regulator that governs flagellar gene expression (Dasgupta and Ramphal, 2001), which likely allows FleN to finely control FleQ as a transcriptional activator. An elevated expression of the master transcriptional regulator FlrA (a FleQ orthologue) was observed in V. cholerae ΔflhG, which is believed to contribute to overexpression of flagellar genes and hyperflagellation (Correa et al., 2005). Although increased flagellar gene expression via a hyperactive or overexpressed master transcriptional regulator may dysregulate numerical control of flagellation in some flhG mutants, this explanation is not valid for all polar flagellates. For example, C. jejuni lacks such a master transcriptional regulator yet deletion of flhG caused C. jejuni cells to shift from producing a single flagellum at one or both poles to assembling multiple polar flagella with only a mild increase in the expression of a subset of flagellar genes (Balaban and Hendrixson, 2011).
Interactions between FlhF and FlhG may occur in vivo in some polar flagellates which suggests an alternative hypothesis for numerical control of flagellar biosynthesis. Regulation of flagellar number may be achieved by FlhG modulating some activity of FlhF important for flagellation, such as its GTPase activity, polar localization, or interaction with flagellar components. In most polar flagellates, FlhF positively influences flagellar biosynthesis. FlhF is required for flagellar gene expression in V. cholerae, C. jejuni and Helicobacter pylori (Hendrixson and DiRita, 2003; Niehus et al., 2004; Correa et al., 2005; Balaban et al., 2009), while overexpression of FlhF in Pseudomonas and Vibrio species causes hyperflagellation that may or may not be restricted to the poles, depending on the species (Pandza et al., 2000; Kusumoto et al., 2006; Green et al., 2009; Schniederberend et al., 2013). In contrast, overproduction of FlhG in P. aeruginosa or V. alginolyticus reduces flagellation (Dasgupta and Ramphal, 2001; Kusumoto et al., 2006). If FlhG influences FlhF GTPase activity in vivo, FlhG modulation of cycles of GTP binding and hydrolysis by FlhF might influence whether FlhF is competent for some step in flagellar assembly, such as polar localization, interaction with flagellar subunits, or nucleating flagellar assembly. Indeed, V. alginolyticus FlhG affects the amount of FlhF that can accumulate at a pole (Kusumoto et al., 2008). The finding that FlhF GTPase mutants alter flagellation differently among polar flagellates provides some support to the idea that differences in FlhF–FlhG interactions and possible resultant effects on their biochemical properties may underlie the variety of flagellation patterns and numbers observed in polar flagellates.
What does FlhF do?
Currently, specific mechanistic roles of FlhF in flagellar biosynthesis or motility are unknown. Because flagella either mislocalize or are not produced in flhF mutants, FlhF may affect the assembly, activity, and/or polar localization of flagellar substructures such as the MS ring or the flagellar type III secretion system (T3SS). Like the MS ring, the flagellar T3SS is one of the earliest formed flagellar substructures and is responsible for secretion of almost all remaining flagellar proteins. Thus, its placement is crucial in deciding where a flagellum will form. However, most peritrichous bacteria do not have FlhF or FlhG homologues and assemble flagella nonetheless. Thus, some unique aspect of the assembly or activity of the flagellar T3SSs in polar flagellates may exist and require FlhF for these T3SSs to be formed or secretion-competent at poles. If so, FlhF may directly interact with T3SS components or assist in nucleation of these components into a mature system with other flagellar structures such as the MS ring. Furthermore, in the case of monotrichous flagellation, the bacterial poles may possess discernable features after division that either allow for or prevent flagellation so that the production of a flagellum occurs only at one pole. An intriguing possibility is that flagellar biosynthesis may be linked to cell division cues that generate structurally different old and new poles, with FlhF perhaps differentiating between the poles to begin flagellar biosynthesis specifically at one pole.
In addition to its flagellar polar-positioning function, FlhF may have other regulatory roles that have diverged in some species. In P. aeruginosa, certain FlhF mutant proteins facilitate polar flagellation, but alter flagellar function (Schniederberend et al., 2013). An effect on flagellar rotation might be mediated by the ability of FlhF to associate with the flagellar motor and other components that ‘tune’ rotation, such as stator units, or molecular brakes and clutches (Delalez and Armitage, 2009).
Even though most peritrichous organisms, including E. coli, Salmonella species, and many others, do not produce FlhF and FlhG, these proteins are found in B. subtilis. By analysing the distribution of basal bodies, peritrichous flagellation along the cell body of B. subtilis was found to be fairly ordered, rather than random (Guttenplan et al., 2013). Basal bodies in wild-type cells were dispersed in a grid-like pattern, symmetrically distributed on either side of the midcell, and discouraged at poles. Furthermore, flagellar basal bodies were separated from each other by a minimum distance greater than that predicted by chance. Thus, peritrichous flagellation in at least B. subtilis is specifically programmed by the cell. It is currently unknown if peritrichous flagellation in other bacterial species is similarly ordered, rather than a random distribution of the organelles. In an ΔflhF mutant, basal bodies accumulated towards a pole, suggesting that FlhF discourages polar flagellation in B. subtilis, which is opposite to its function in many polar flagellates (Guttenplan et al., 2013). Additionally, mutation of flhG resulted in aggregation of flagellar basal bodies and a loss of the normal spacing distance between flagella (Guttenplan et al., 2013). From these data, Guttenplan et al. proposed that FlhG may determine the distance between individual FlhF proteins in B. subtilis, which allows formation of flagellar basal bodies at some distance from each other. Although FlhF and FlhG spatially influenced flagellation in B. subtilis, they did not influence numerical parameters of flagellation. These results further suggest that FlhF–FlhG pairs may have co-evolved in individual motile bacterial species to facilitate flagellation patterns and numbers specific for each species.
A role for FlhG in influencing cell division in C. jejuni
In addition to numerically regulating flagellar biosynthesis, C. jejuni FlhG influences spatial parameters of division (Balaban and Hendrixson, 2011). A striking phenotype of the C. jejuni flhG mutant was the abundant presence of minicells during normal growth. Minicells are products of cell division when the FtsZ polymerizes into the Z-ring at a pole to divide the cell, rather than at the cellular midpoint to promote symmetrical division (Adler et al., 1967; Davie et al., 1984). In many bacteria, spatial regulation of Z-ring formation is controlled by the Min system (for detailed reviews, see Barak and Wilkinson, 2007; Lutkenhaus, 2007). The MinD ATPase is a SIMIBI family member with homology to FlhG proteins (Fig. 1B). However, MinD itself does not inhibit Z-ring formation. Instead, MinD complexes with MinC, the actual inhibitor of FtsZ polymerization. MinCD complexes form at a pole to inhibit Z-ring formation and dissociate due to stimulation of MinD-mediated ATP hydrolysis by MinE. After dissociating from a pole, MinCD complexes form at the other pole and are eventually dissociated at this pole. The net effect of this process is that MinCD complexes oscillate between poles and inhibit Z-ring formation at polar regions. Because MinCD concentrations are minimal at the cellular midpoint, FtsZ forms the Z-ring at the midpoint to promote symmetrical division.
Whereas many bacteria encode a complete Min system, Campylobacter species lack these genes. As such, it is possible that Campylobacter species have adapted FlhG with some properties common to MinD that allows it to influence inhibition of division at poles. Expression of H. pylori or V. cholerae FlhG proteins, but not the respective MinD proteins moderately restored normal division to C. jejuni ΔflhG (Balaban and Hendrixson, 2011). These findings raise the question of whether FlhG, the Min system, or both contribute to spatial regulation of division in other polar flagellates. It is currently unknown if FlhG or another as yet unidentified protein directly inhibits Z-ring formation.
Further examination of C. jejuni revealed that flagellar components function with FlhG to inhibit division at poles. In mutants lacking FlhF, FliF (the inner membrane MS ring component), FliM or FliN (the latter two proteins are cytoplasmic C ring components that form part of the flagellar switch complex), minicells formed at similar levels as an flhG mutant (Balaban and Hendrixson, 2011). Furthermore, minicell formation in these mutants was suppressed by overexpression of flhG, which suggested that FlhG works in a mechanism with and possibly downstream of these flagellar proteins to prevent division at poles. Currently, it is not known how flagellar proteins may influence an FlhG-dependent mechanism to inhibit division at poles. FliF, FliM and FliN compose the base of the flagellum with domains accessible to the cytoplasm, which is where FlhG is predicted to reside. It is possible that FlhG may require MS and C ring structures to localize to poles or be spatially arranged to function in a mechanism to inhibit division.
Proposed models for how FlhF and FlhG function in different polar flagellates
The most pronounced effect of flhF or flhG mutation in polar flagellates is a gross alteration in flagellar placement or number, but the existence of a single model for how these proteins function to mediate specific flagellation patterns across polarly flagellated species is unlikely. FlhF and FlhG are present in many polar flagellates, yet as a group, these bacteria produce monotrichous, lophotrichous and amphitrichous flagellation. It is possible that intrinsic biochemical properties of each FlhF–FlhG cognate pair may create the specific flagellation pattern for each polar flagellate. For instance, whether FlhG regulates the GTPase activity of its FlhF partner may dictate how many flagella form or whether flagella form at a specific pole. In the case of the lophotrichous flagellation pattern of H. pylori, FlhG may only weakly regulate the GTPase activity of its cognate FlhF, thereby resulting in a relatively hyperactive FlhF with a prolonged GTP-bound ‘ON’ state relative to monotrichous organisms, which may facilitate construction of multiple flagella at one pole. Presented below are two proposed models for how FlhF and FlhG may function in C. jejuni and P. aeruginosa to mediate flagellation patterns or contribute to flagellar function. Future research will be necessary to test and further refine such models to accurately detail how these organisms achieve specific flagellation patterns and flagellar activity required by each species.
Data indicate that FlhF, FlhG, the MS ring and C ring components of C. jejuni are linked together in mechanisms that impact not only spatial and numerical parameters of flagellation, but also spatial regulation of cell division (Balaban et al., 2009; Balaban and Hendrixson, 2011). One possible model for how these components may work together to mediate these biological processes is presented in Fig. 2.
Immediately after division, a C. jejuni daughter cell possesses a flagellum only at the old pole. Thus, flagellar biosynthesis must occur at the new pole to complete amphitrichous flagellation. FlhF, but not its GTPase activity, is required for expression of σ54-dependent flagellar genes (Hendrixson and DiRita, 2003). However, C. jejuni FlhF is polarly localized and alteration of its GTPase activity affected flagellar placement (Balaban et al., 2009; Ewing et al., 2009; Balaban and Hendrixson, 2011), indicating that the GTPase activity is important for normal amphitrichous flagellation. One possible scenario in wild-type bacteria is that FlhF, in a GTPase-dependent manner, polarly localizes and promotes one round of flagellar biosynthesis thereby completing amphitrichous flagellation. FlhF may help decide where flagellation occurs by influencing an early step in flagellar biosynthesis, such as promoting the flagellar T3SS, the MS ring and/or the C ring to form at the new pole. FlhG also polarly localizes, but it is unknown if its localization is dependent on FlhF or flagellar structures (Balaban and Hendrixson, 2011). Once at the pole, FlhG may subsequently alter GTP binding or hydrolysis by FlhF so that FlhF is incompetent for another round of flagellar assembly, ensuring that only one flagellum is produced at the new pole. Positioning of FlhG at the new pole is likely a prerequisite for the protein to function with the MS and C rings to prevent division at this pole and encourage Z-ring formation at the cellular midpoint for symmetrical division.
This unusual mechanism to spatially control division may have contributed a biological advantage for C. jejuni to produce amphitrichous flagellation, which is a fairly rare flagellation pattern in motile bacteria. By having a flagellum-influenced cell division system, C. jejuni must produce a flagellum at each pole to inhibit division at both poles via FlhG. If C. jejuni possessed monotrichous flagellation, division could occur at an aflagellated pole to result in minicells, which decreases the number of viable daughter cells generated per a round of division. If the bacterium produced peritrichous flagella, division may be inhibited throughout the cell. Thus, amphitrichous flagellation allows C. jejuni to not only promote motility, but also prevent polar division and encourage symmetrical division to guarantee production of viable daughter cells.
The biology of P. aeruginosa FlhF and FleN differs from that of C. jejuni in several significant regards. FlhF appears to play a minor role, if any, in influencing flagellar gene expression. The primary phenotype of P. aeruginosa ΔflhF bacteria is the assembly of a single, mislocalized flagellum that can power swimming motility (Murray and Kazmierczak, 2006). FleN, unlike C. jejuni FlhG, likely controls flagellar number by modulating expression of flagellar genes via a master transcriptional regulator (Dasgupta et al., 2000; Dasgupta and Ramphal, 2001). Cell division defects are also not reported for ΔflhF or ΔfleN mutants, suggesting that a MinD-like function is not a feature of FleN. Lastly, FlhF mutant proteins with altered nucleotide binding and hydrolysis functions do not alter the monotrichous flagellation pattern, although they do influence the motility of the assembled organelle. Since overexpression of wild-type or mutant FlhF alleles is associated with increased numbers of bacteria expressing more than one polar flagellum, the protein may interact with, guide or localize an early component of the flagellar T3SS or MS ring to the pole. In this regard, it is interesting that flhF is located immediately downstream from and co-transcribed with the flagellar T3SS component flhA (Wurtzel et al., 2012). However, it is unknown if these two proteins interact.
Ultimately, data regarding FlhF in the biology of P. aeruginosa largely suggest a polar-positioning function independent of GTPase activity, but a GTPase-dependent effect on flagellar motility. The notion that FlhF behaviour is altered by its nucleotide-bound state is consistent with the observations that FlhF dimerization depends on nucleotide binding, similar to what is reported for the FtsY–Ffh heterodimer. Whether FleN can influence the GTP hydrolysis rate of P. aeruginosa FlhF is not clear because the activating helix of FlhG is not present in FleN (Fig. 1B). The observation that FlhF mutants with slowed GDP/GTP exchange do not support flagellar rotation does suggest, however, that FlhF maintains ongoing interactions with the assembled flagellum (and perhaps more specifically, motor and stator components) that are sensitive to the nucleotide-bound state of FlhF.
Understanding spatial and numerical regulation of flagella and other organelles are new areas for exploring how bacteria genetically programme the assembly and distribution of surface-associated structures. Undoubtedly, FlhF and FlhG influence spatial and numerical parameters of flagellar biosynthesis, yet these proteins appear to function in a species-specific manner among the polar flagellates to result in different flagellation patters. Thus, future analysis will likely uncover different mechanisms by which these proteins function in the various polar flagellates. It is expected that answering questions regarding how flagella are organized in motile bacteria will add to our growing knowledge of how bacteria organize and assign regions of the cell to enable the formation of structures with specific activities critical for bacterial fitness, such as locomotion, protein secretion, adherence, environmental sensing and proper cell division.
This work was supported by NIH Grant R01AI075051 and the Burroughs Wellcome Fund Investigator of Pathogenesis of Infectious Disease Award to B.I.K., and NIH Grants R01AI065539 and 1R21AI103643 to D.R.H.