Keywords

Introduction

The bacterial flagellum, which is a large filamentous assembly, is one of the organelles involved in motility in various environments such as liquids and solid surfaces. Salmonella enterica serovar Typhimurium (Salmonella) has several flagella on the cell body. Each flagellum is a supramolecular protein complex made of about 30 different proteins, and the copy number of each component protein varies from a few to tens of thousands. More than 60 proteins are involved in flagellar construction, force generation and chemotactic behavior in Salmonella (Table 8.1). The Salmonella flagellum is made of five distinct structural and functional parts: a base body, a hook, a junction, a filament and a filament cap (Fig. 8.1). The base body is embedded in the cellular membranes and works as a bi-directional rotary motor energized by the electrochemical potential difference of protons (H+) across the cytoplasmic membrane (proton motive force, PMF). The hook, junction, filament and filament cap structures are located outside the bacterial cell body. The filament acts as a helical propeller to generate propulsion. The hook functions as a universal joint that smoothly transmits rotational force produced by the bi-directional rotary motor to the helical propeller. The junction connects the hook and filament. The filament cap supports the self-assembly of the filament protein, flagellin, into the long helical filament structure (Minamino and Imada 2015; Nakamura and Minamino 2019).

Table 8.1 Bacterial flagellar proteins in Salmonella
Full size table
Fig. 8.1
figure 1

Schematic diagram of the bacterial flagellar motor complex structure. The flagellum is composed of a filament, a hook and a basal body. Each building block synthesized in the cytoplasm is translocated via a type III protein export apparatus into the central channel of the growing axial structure. OM—outer membrane, PG—peptidoglycan layer, CM—cytoplasmic membrane

Full size image

The Salmonella flagellar motor rotates in both counterclockwise (CCW) and clockwise (CW) directions. When all the flagellar motors rotate in CCW direction, long helical flagellar filaments form a bundle structure behind a cell body, due to bending flexibility of each hook structure. As a result, the flagellar bundle produces thrust enough to allow Salmonella cells to go straight. When one or more motors switch the rotational direction from CCW to CW, the bundle structure is partially disrupted, and so the cells stop swimming and change the swimming direction (Fig. 8.2a) (Morimoto and Minamino 2014). Each flagellar motor is placed under a control of the sensory signaling network, which modulates the switching frequency of flagellar motor rotation, and hence bacterial cells carry out chemotaxis by a biased random walk toward various chemicals, pH and temperature to move toward more suitable stimuli (Fig. 8.2b, c) (Berg 2003).

Fig. 8.2
figure 2

a Swimming pattern of Salmonella cells. Salmonella cells undergoes continuous run-and-stop pattern, thereby showing random walk motions in liquids. When the flagellar motor rotates in counterclockwise (CCW) direction, Salmonella cells can go straight. In contrast, when the motor switches its rotational direction from CCW to counterclockwise (CW), the cells stop and change the swimming direction. b Biased random walk motion to attractants. Salmonella cells sense gradients of chemical attractants in the environments to move towards more favorable stimuli. c Schematic diagram of a sensory signaling transduction pathway responsible for chemotaxis. The Salmonella phosphorelay signal transduction network is formed by methyl-accepting chemotaxis protein (MCP), the CheW adopter protein, the CheA kinase, the CheY chemotaxis signaling protein, the CheZ phosphatase, the CheB methylesterase and the CheR methyltransferase. CheY-P binds to the switch complex of the flagellar motor to chnage the direction of flagellar motor rotation from CCW to  CW

Full size image

Methyl-accepting chemotaxis protein (MCP) is a transmembrane protein with a large cytoplasmic domain and senses temporal changes in environmental stimuli. The adapter protein CheW binds to the cytoplasmic domain of the MCP, allowing the CheA kinase to efficiently associate with the MCP. The cytoplasmic domain of MCP controls the autophosphorylation activity of the CheA kinase to generate the chemotaxis signal. A phosphorylated form of the CheA kinase transfers its phosphate group to CheY, which acts as the chemotaxis signaling protein. Phosphorylated CheY (CheY-P) binds to two cytoplasmic rotor component proteins, FliM and FliN, in the flagellar motor. As a result, the motor switches the direction of rotation from CCW to CW. The CheZ phosphatase accelerates the dephosphorylation of CheY-P to induce the dissociation of CheY from the motor to allow the motor to rotate CCW again. This chemotaxis signaling network adapts to changes in the concentration of chemical stimuli over a wide dynamic range. CheB and CheR are involved in the adaption mechanism. CheR is the methyltransferase to induce the methylation of the cytoplasmic domain of the MCP to modulate the autophosphorylation activity of the CheA kinase. The CheA kinase also transfers the phosphate group to the CheB methylesterase to promote the demethylation of the MCP to reduce the probability of autophosphorylation of the CheA kinase (Fig. 8.2c) (Bi and Sourjik 2018).

The Salmonella flagellar motor consists of a rotor and about ten stator units placed around the rotor. The rotor is a macromolecular ring complex made of four flagellar proteins, FliF, FliG, FliM and FliN. The transmembrane MS ring is formed by FliF. FliG, FliM and FliN assemble into the C ring on the cytoplasmic surface of the MS ring in this order. The C ring also works as a directional switching device to switch between CCW and CW states of the flagellar motor. Each stator unit is composed of two integral membrane proteins, MotA and MotB, which form a proton channel complex to convert the energy of the proton flow through the channel into the rotational force (Minamino et al. 2018; Nakamura and Minamino 2019).

To construct the axial structure of the flagellum such as the hook and filament outside the cell body, the basal body has a type III protein export apparatus at its proximal end (Macnab 2003). The type III protein export apparatus is also a macromolecular protein complex consisting of a transmembrane export gate complex powered by PMF across the cytoplasmic membrane and an associated cytoplasmic ATPase ring complex and transports building blocks of the axial structure from the cytoplasm to the distal end of the nascent structure where each building block self-assembles into the axial structure. The export gate complex is located inside the central pore of the MS ring whereas the cytoplasmic ATPase ring complex associates with the C ring (Minamino 2014, 2018). In this chapter, we describe the structure and function of the bacterial flagellar motor complex in Salmonella.

Axial Structure

The Salmonella flagellum is composed of several basal body rings and an axial structure (Fig. 8.1). The axial structure is a helical, tubular structure consisting of six distinct structural parts: the proximal rod (FlgB, FlgC, FlgF), the distal rod (FlgG), the hook (FlgE), the junction (FlgK, FlgL), the filament (FliC, FljB), and the filament cap (FliD). The N- and C-terminal regions of each building block are disordered in solution but adopt a coiled coil in the inner core domain (D0) of the building block when they assemble into the axial structure (Fig. 8.3a) (Furukawa et al. 2002; Saijo-Hamano et al. 2004; Samatey et al. 2001). Hydrophobic interactions between the D0 domains of the building blocks are essential not only for the self-assembly of the axial structure but also for structural and mechanical stability of the entire tubular structure (Yonekura et al. 2003; Fujii et al. 2009; Maki-Yonekura et al. 2010; Saijo-Hamano et al. 2019). Except for the cap structure, which has five-fold rotational symmetry (Yonekura et al. 2000), the axial structure is a helical assembly composed of 11 protofilaments (Fig. 8.3b) (Yonekura et al. 2003; Fujii et al. 2009, 2017; Maki-Yonekura et al. 2010; Kato et al. 2019). The basic helical line that passes through all the subunits is called a 1-start helix, and there are approximately 11 subunits per two turns of the 1-start helix (Fig. 8.3b).

Fig. 8.3
figure 3

Axial structure in the bacterial flagellar motor complex. a Schematic diagram of cross section of the flagellar filament. Flagellar filament protein is composed of domains D0, D1, D2 and D3. b Arrangement of building blocks in the flagellar axial structure composed of 11 protofilaments. Arrows show the helical lines. The 11-start helical line constitutes each protofilament. The numbers of the subunit starting from the subunit 0 along the 1-start helical line are shown. c Representative atomic structures of flagellar axial component proteins, FlgG (PDB ID, 6JZR), FlgE (PDB ID, 6KFK), FlgK (PDB ID, 2D4Y), FlgL (PDB ID, 2D4X), FliC (PDB ID, 1UCU) and FliD (PDB ID, 5H5T) derived from Salmonella. CM—cytoplasmic membrane, PG—peptidoglycan layer, OM—outer membrane

Full size image

The rod is straight and rigid and acts as a drive shaft of the flagellar motor (Fujii et al. 2017). The hook is supercoiled and highly flexible in bending and functions as a universal joint (Samatey et al. 2004; Fujii et al. 2009; Kato et al. 2019). The filament is also supercoiled but rigid against bending and so functions as a helical propeller (Yonekura et al. 2003; Maki-Yonekura et al. 2010). The Salmonella filament normally adopts a left-handed supercoil to form a flagellar bundle structure behind a cell body for straight swimming and switches from the left-handed to right-handed supercoils when the flagellar motor switches the direction of rotation from CCW to CW (Calladine 1975, 1976; Maki-Yonekura et al. 2010). Since the rod and hook structurally look similar to each other, the hook is directly connected with the rod (Chevance et al. 2007; Fujii et al. 2017). In contrast, because there are structural and mechanical differences between the hook and filament, the junction is a cushioning structure to connect them (Homma and Iino 1985). The filament cap promotes the self-assembly of flagellin subunits into the filament at the distal end (Ikeda et al. 1985, 1996; Yonekura et al. 2000).

Electron cryo-microscopy (cryo-EM) image analyses of the unusually elongated rod structure, namely polyrod, which is caused by the G65V substitution in the distal rod protein FlgG, have revealed that FlgG is composed of three domains, D0, Dc, and D1 from the inside to outside of the polyrod structure (Fig. 8.3c) (Fujii et al. 2017; Saijo-Hamano et al. 2019). Intermolecular interactions between FlgG subunits in the polyrod structure are very tight, and hence the rod is straight and rigid against bending to act as a drive shaft. Because FliE is an axial component protein of the basal body and interacts with the proximal rod protein FlgB, it is thought to form the most proximal part of the rod structure that is directly connected with the MS ring and/or the transmembrane export gate complex of the type III protein export apparatus (Minamino and Macnab 2000b; Minamino et al. 2000).

The hook is a short, curved structure composed of about 120 copies of the hook protein FlgE. FlgE consist of four domains, D0, Dc, D1 and D2, arranged from the inside to the outside of the hook structure (Fig. 8.3c) (Samatey et al. 2004; Fujii et al. 2009). Although the D0, Dc and D1 domains of FlgE are highly homologous to those of FlgG, there are structural and functional differences in the Dc domain between FlgE and FlgG (Chevance et al. 2007; Fujii et al. 2017; Saijo-Hamano et al. 2019). The Dc domain of FlgG contains a FlgG-specific YQTIRQPGAQSSEQTTLP sequence (GSS), whereas FlgE does not. Insertion of the GSS into a corresponding region of the Dc domain of FlgE makes the hook straight and rigid in a way similar to the FlgG rod structure, suggesting that the GSS is important for the rigidity of the rod structure (Hiraoka et al. 2017). Axial packing interactions between a triangular loop of the D1 domain and the D2 domain in each protofilament of the hook structure are responsible for hook supercoiling (Samatey et al. 2004; Kato et al. 2019). However, a deletion of either a triangular loop of the D1 domain or domain D2 makes the hook straight but retains its bending flexibility, allowing the filaments to form a flagellar bundle behind a cell body (Sakai et al. 2018). Because there are gaps between FlgE subunits in the entire hook structure, these gaps are likely to contribute to bending flexibility of the hook (Fujii et al. 2009; Horvath et al. 2019; Kato et al. 2019; Shibata et al. 2019).

The filament is composed of more than 30,000 flagellin subunits and is a major target for antigens. Salmonella contains two flagellin genes fliC and fljB on the chromosomal DNA and expresses either FliC or FljB to form the filament on the cell surface. The alternate expression frequency of these two flagellin genes is about 10–3–10–5 per cell per generation (Stocker 1949; Gillen and Hughes 1991), and such an autonomous switching changes the motility function as well as antigen response of Salmonella cells (Yamaguchi et al. 2020). The flagellin subunit is composed of four domains D0, D1, D2, and D3, arranged from the inner core to the outer surface of the filament (Fig. 8.3a, c) (Namba and Vonderviszt 1997). The D0 and D1 domains are responsible for the formation of the supercoiled form of the filament structure, and a β-hairpin in the D1 domain is thought to act as a structural switch that changes between left-handed and right-handed helical forms of the filament structure (Samatey et al. 2001). Because the D3 domain, which occupies the outermost part of the filament, is recognized as H antigens by host immune systems, the primary sequence of the D3 domain differs significantly between FliC and FljB. Interestingly, there are structural differences in the D3 domain between the FliC and FljB filaments, and such differences cause a significant difference in the helical propeller function of the filament under viscous conditions (Yamaguchi et al. 2020). The D3 domain of the flagellin molecule is dispensable for filament formation but contributes to thermal stability of the entire filament structure (Muskotal et al. 2010; Furukawa et al. 2016).

The filament growth occurs at the distal end of the growing filament structure (Minamino 2014). Purified flagellin subunits can self-assemble into a long helical filament in vitro (Asakura 1970). However, flagellin subunits require a filament cap formed by FliD to polymerize into the filament structure in vivo (Homma et al. 1984). If the FliD cap is missing, flagellin monomers cannot assemble into the filament and hence are secreted into the culture media. This suggests that the FliD cap prevents newly exported flagellin molecules from leaking into the culture media to ensure that each flagellin molecule has sufficient time to be incorporated into the filament. Interestingly, the stoichiometry of the filament cap structure is different among bacteria species. The Salmonella FliD cap has five-fold rotational symmetry (Fig. 8.3c), the Escherichia coli and Pseudomonas aeruginosa FliD caps have six-fold rotational symmetry, and the Serratia marcescens and Bdellovibrio bacteriovorus FliD caps form a four-fold rotational symmetry structure (Yonekura et al. 2000; Maki-Yonekura et al. 2003; Postel et al. 2016; Ko et al. 2017; Song et al. 2017; Cho et al. 2019). These differences might be related to the structural diversity of the flagellin molecule.

Rotor

The Salmonella basal body has the C ring (FliG, FlM, FliN), the MS ring (FliF), the P ring (FlgI) and the L ring (FlgH), which are located in the cytoplasm, the cytoplasmic membrane, the peptidoglycan (PG) layer and the outer membrane, respectively (Fig. 8.4a). The MS ring and the C ring together form a rotor ring complex of the flagellar motor. The most proximal part of the rod is thought to be directly connected with the MS ring. The L and P rings together function as a bushing for the distal rod structure, and the inner surface of the LP ring complex is likely to be very smooth for high-speed rotation of the rod as a drive shaft (Minamino et al. 2008).

Fig. 8.4
figure 4

Flagellar motor complex structure. a Cartoon of basal body rings. The basal body contains four ring structures named the C ring (FliG, FliM, FliN), the MS ring (FliF), the P ring (FlgI) and the L ring (FlgH). The MS ring and the C ring form the rotor ring complex of the flagellar motor. Crystal structures of Aquifex aeolicus FliG (PDB ID, 3HJL), Thermotoga maritima FliM middle domain (FliMM) (PDB ID, 2HP7) and the Salmonella FliMC-FliN fusion protein (PDB ID, 4YXB). FliG consists of three compactly folded domains, FliGN, FliGM and FliGC that is able to be divided into two FliGCN and FliGCC subdomains and two linker helices, HelixNM connecting FliGN and FliGM and HelixMC connecting FliGM and FliGCN. Two highly conserved charged residues, Arg284 (R) and Asp292 (D) in the torque helix of FliGCC are responsible for the interaction with the stator protein MotA. b Structure of the stator unit of the flagellar motor. The stator unit consists of four MotA and two MotB proteins and acts as a transmembrane proton channel to couple the proton flow to torque generation. Highly conserved Pro173 of MotA (P173) and Asp33 of MotB (D33) are directly involved in the proton translocation. The periplasmic domain of MotB (MotBC) forms a homodimer and binds to the PG layer, allowing the MotA/MotB complex to act as a stator unit in the motor. A crystal structure of the MotBC dimer (PDB ID, 2ZVY) is shown. A linker region connecting MotB-TM and MotBC is dispensable for flagellar motor rotation but coordinates the proton channel activity of the MotA/MotB complex to stator assembly into the motor. The plug segment suppresses premature proton flow through the proton channel until the MotA/MotB complex become an active stator unit in the motor. CM—cytoplasmic membrane, PG—peptidoglycan layer, OM—outer membrane

Full size image

FliF molecules with two transmembrane (TM) helices assemble into the MS ring in the cytoplasmic membrane. Over-expression of FliF alone from a pET-based plasmid vector results in the self-assembly of FliF subunits into the MS ring (Ueno et al. 1992). However, when FliF is expressed from the chromosomal DNA, FliF requires FliG for efficient formation of the MS ring in the cytoplasmic membrane (Morimoto et al. 2014). The N-terminal domain of FliG (FliGN) binds to the C-terminal cytoplasmic domain of FliF (Kihara et al. 2000; Levenson et al. 2012). Intermolecular interactions between FliGN domains and between the middle (FliGM) and C-terminal (FliGC) domains of FliG are responsible for FliG ring formation on the cytoplasmic surface of the MS ring (Baker et al. 2016; Kim et al. 2017). The FliM/FliN complexes with a 1 FliM: 3 FliN stoichiometry bind to the FliG ring through an interaction between FliGM and FliM to form the C ring wall with 34-fold rotational symmetry (Paul et al. 2011; Vartanian et al. 2012; Lam et al. 2013). Because the torque helix in the C-terminal portion of FliGC (FliGCC) is directly involved in the interaction with the stator protein MotA to generate the rotational force, FliGCC is believed to be located at the top of the C ring wall (Fig. 8.4a) (Zhou et al. 1998a). It has been reported that there is a symmetry mismatch between the MS and C rings (Suzuki et al. 2004), but recent high-resolution cryo-EM image analysis of the Salmonella MS ring have shown that the MS ring also has about 34-fold rotational symmetry (Johnson et al. 2020).

Many Salmonella mutants with altered motility have been isolated. Among motility mutants, mutations in FliG, FliM, and FliN confer a CCW switch bias phenotype or a CW switch bias phenotype even in the presence of the chemosensory signal transduction pathway. This suggests that these three C ring proteins are involved in directional switching of the flagellar motor (Yamaguchi et al. 1986). Genetic and biochemical analyses have revealed that the chemotaxis signaling protein CheY-P directly binds to the intrinsically disordered N-terminal region of FliM and the C-terminal domain of FliN in the C ring and promotes a highly cooperative conformational change in the FliG ring structure, allowing the motor to rotate in CW direction (Dyer et al. 2009; Sarkar et al. 2010). A conserved MFXF motif in a flexible hinge connecting the N-terminal portion of FliGC (FliGCN) and FliGCC is thought to induce a 180º rotation of the torque helix of FliGCC relative to FliGCN when the motor switches the direction of rotation from CCW to CW (Lam et al. 2012; Miyanoiri et al. 2017). An in-frame deletion of three residues (Pro169–Ala170–Ala171) near the N-terminal portion of the linker helix connecting FliGM and FliGCN (HelixMC) causes an extremely strong CW switch bias phenotype. The PAA deletion induces not only a distinct orientation of HelixMC relative to FliGM but also the rotation of FliGCC relative to FliGCN, suggesting that HelixMC is postulated to act as a gear to coordinately switch the FliG ring structure between CCW and CW states (Minamino et al. 2011a; Kinoshita et al. 2018a, b). In addition to CheY-P, cyclic di-GMP (c-di-GMP), which is a second messenger molecule to induce biofilm formation, binds to YcgR to regulate chemotaxis through interactions of the c-di-GMP bound form of YcgR with FliG and FliM in the C ring as well as MotA (Boehm et al. 2010; Fang and Gomelsky 2010; Paul et al. 2010).

The FliM1/FliN3 complex labeled with a fluorescent protein is visualized to exchange rapidly between the basal body and its cytoplasmic pool by high-resolution single molecule imaging techniques whereas neither FliF nor FliG does not, indicating that the C ring wall has a highly dynamic nature whereas both MS and FliG rings do not (Delalez et al. 2010; Fukuoka et al. 2010). The copy number of FliM1/FliN3 complex is estimated to be about 1.3 times higher in the CCW motor than in the CW motor (Delalez et al. 2014). Because the rapid turnover of the C ring wall depends on CheY-P, the directional switching of the flagellar motor seems to induce the dissociation of weakly bound FliM1/FliN3 complexes from the C ring wall (Lele et al. 2012; Branch et al. 2014; Delalez et al. 2014). The number of the FliM1/FliN3 complexes in the C ring increases with a decrease in the cytoplasmic concentration of CheY-P, suggesting that the motor adapts to changes in the steady-state level of CheY-P by tuning the copy number of the FliM1/FliN3 complex associated with the C ring wall (Lele et al. 2012; Branch et al. 2014; Delalez et al. 2014). However, because both purified CCW and CW motors have the C ring with 34-fold rotational symmetry (Sakai et al. 2019), it remains unknown how the FliG ring can accommodate additional FliM1/FliN3 complexes to the C ring.

Stator

MotA and MotB form a transmembrane proton channel that acts as a stator unit of the flagellar motor to couple the proton flow through the proton channel to torque generation (Fig. 8.4a) (Larsen et al. 1974; Manson et al. 1977; Ravid and Eisenbach 1984). MotA and MotB form a heterohexamer with a 4 MotA: 2 MotB stoichiometry and has two distinct proton channels (Fig. 8.4b) (Braun et al. 2004; Kojima and Blair 2004). MotA assembles into a homo-tetramer (Braun et al. 2004; Kim et al. 2008; Takekawa et al. 2016), and the MotB homo-dimer binds to the central pore of the MotA tetramer (Braun and Blair 2001). MotA has four TM helices (TM1–TM4), two short periplasmic loops between TM1 and TM2 and between TM3 and TM4, a relatively large cytoplasmic loop between TM2 and TM3 and a C-terminal cytoplasmic tail (Zhou et al. 1995). MotB has an N-terminal cytoplasmic tail, followed by a single TM helix and finally a large C-terminal periplasmic domain containing a peptidoglycan-binding (PGB) motif (MotBC) (Kojima et al. 2009). The MotB-TM helix forms a proton channel with the TM3 and TM4 helices of MotA (Braun and Blair 2001; Braun et al. 2004). Highly conserved Asp33 of MotB and Pro173 of MotA, which are located in the proton channel, are involved in the proton flow through the channel (Zhou et al. 1998b; Braun et al. 1999; Kojima and Blair 2001; Che et al. 2008). The cytoplasmic loop of MotA contains highly conserved Arg90 and Glu98 residues that are responsible for the interaction with the torque helix of FliGCC (Zhou et al. 1998a). MotBC forms a dimer, and its dimerization is critical for efficient targeting and stable anchoring of the stator units to the PG layer. As a result, the MotA4/MotB2 complex can become an active stator unit around a rotor (Kojima et al. 2009, 2018). Interestingly, site-specific disulfide bridge experiments have shown that MotBC is in relatively close proximity to the P ring of the basal body (Fig. 8.4a) (Hizukuri et al. 2010).

A flexible linker region of MotB (residues 51–100) connecting MotB-TM and the MotBC domain are dispensable for flagellar motor rotation (Fig. 8.4b) (Muramoto and Macnab 1998). Because an in-frame deletion of residues 53–66 of MotB considerably reduces cytoplasmic pH due to undesirable proton leakage through unassembled MotA4/MotB2 proton channel complex, residues 53–66 of MotB are postulated to form a plug segment to suppress the proton leakage through the proton channel until the MotA4/MotB2 complex is incorporated into a motor (Hosking et al. 2006; Morimoto et al. 2010a). Although the MotA4/MotB2 complex with a deletion of residues 51–100 of MotB is still functional, such a large deletion does not induce a considerable proton leakage through unassembled MotA4/MotB2 proton channel complexes (Muramoto and Macnab 1998). The MotB (L119P) substitution in the PGB domain of MotBC not only causes massive proton leakage though unassembled MotA4/MotB (Δ51–100)2 complexes (Kojima et al. 2009; Morimoto et al. 2010a) but also increase the binding affinity of the PGB domain for the PG layer (Kojima et al. 2018). The MotB (L119P) substitution induces an order-to-disorder transition of the N-terminal portion of the PGB domain (Kojima et al. 2018). Since a 5 nm extension of the PGB domain of MotB from the MotB-TM helix is required for the binding of the PGB domain to the PG layer (Kojima et al. 2008), such an order-to-disordered transition of the N-terminal portion of the PGB domain should occur when the MotA4/MotB2 complex encounters a rotor to become an active stator unit in the motor (Kojima et al. 2018). In agreement with this, the 5 nm extension process of the PGB domain of MotS derived from Bacillus subtilis, which is a MotB homologue, has been directly visualized using high-speed atomic force microscopy (Terahara et al. 2017a). Therefore, these observations suggest that the flexible linker region of MotB coordinates not only proper proton channel formation of the MotA4/MotB2 complex but also efficient and proper anchoring of the PGB domain to the PG layer.

Each stator complex binds to and dissociates from a rotor ring complex during torque generation by the flagellar motor, suggesting that the stator complex has a highly dynamic nature (Leake et al. 2006). The number of functionally active stators in the motor varies in response to changes in the environment, such as external load and external proton concentration (Fukuoka et al. 2009; Lele et al. 2013; Tipping et al. 2013; Che et al. 2014; Terahara et al. 2017b; Suzuki et al. 2019). For example, the maximum number of active stator units incorporated into the flagellar motor is about ten when the motor operates at high loads (Reid et al. 2006; Pourjaberi et al. 2017). In contrast, the stator number decreases from ten to a few with a decrease in external load (Yuan and Berg 2008; Nakamura et al.  2020). Thus, the flagellar motor autonomously controls the number of active stator units around the rotor in response to changes in external loads. It has been shown that an in-frame deletion of residues 72–100 of MotB results in much steeper decrease in the stator number with a decrease in the external load as compared to the wild-type motor (Castillo et al. 2013). This suggests that the linker region of MotB modulates the binding affinity of the PGB domain for the PG layer in a load-dependent manner. Interestingly, certain point mutations in the cytoplasmic loop of MotA also change the load-sensitivity of the MotA4/MotB2 complex (Pourjaberi et al. 2017). Because the cytoplasmic loop of MotA directly binds to FliG to generate torque (Zhou et al. 1998a; Morimoto et al. 2010b; Morimoto et al. 2013), this cytoplasmic loop seems to act as a load sensor to coordinate the number of active stators incorporated into the motor in response to changes in external load. Recently, it has been shown that the localization efficiency of the MotA4/MotB2 complex is also increased by lowering extracellular pH, that is, increasing proton concentration, and that such an external pH-dependent localization pattern of the MotA4/MotB2 complex is not affected by the motB(D33N) mutation, which inhibits the proton flow through the MotA4/MotB2 proton channel complex (Suzuki et al. 2019). Since it has been reported that the PGB domain of MotS induces a disordered-to-order transition of the PGB domain in a Na+-dependent manner to allow the MotP4/MotS2 complex to become an active stator unit in the B. subtilis flagellar motor (Terahara et al. 2017a), it is possible that the folding efficiency of the PGB domain of MotB depends on external pH values.

Type III Protein Export Apparatus

The assembly of the axial structure begins with the rod, followed by the hook, the junction, the filament cap in this order and finally the filament (Macnab 2003). For the self-assembly of each building block at the distal end of the growing structure, a type III protein export apparatus unfolds the building blocks synthesized in the cytoplasm and then transports them into the central channel of the growing structure (Minamino 2014; Minamino et al. 2020a). PMF across the cytoplasmic membrane and ATP are utilized as the energy sources to drive flagellar protein export by the type III protein export apparatus (Minamino and Namba 2008; Paul et al. 2008). The type III protein export apparatus is composed of a PMF-driven export gate complex consisting of five membrane proteins FlhA, FlhB, FliP, FliQ and FliR, and a cytoplasmic ATPase ring complex made of three cytoplasmic proteins, FliH, FliI and FliJ (Fig. 8.5) (Minamino 2014). These component proteins show sequence and functional similarities to those of the injectisome of pathogenic bacteria such as Salmonella spp, Shigella spp and Yersinia spp, which is involved in direct injection of effector proteins into eukaryotic host cells for bacterial infection (Galán et al. 2014). Furthermore, the overall structure of the cytoplasmic ATPase ring complex is similar to F-type and V-type rotary ATPase families, suggesting that the type III protein export apparatus of the flagellum and the injectisome share a common evolutionary origin with these rotary ATPase families (Imada et al. 2007, 2016; Ibuki et al. 2011). In addition, four cytoplasmic proteins, FlgN, FliA, FliS and FliT, function as flagellum-specific export chaperones to support the export of their cognate building blocks by the type III protein export apparatus (Table 8.1).

Fig. 8.5
figure 5

Structures of the flagellar type III export apparatus component proteins. The type III protein export apparatus is composed of five transmembrane proteins, FlhA, FlhB, FliP, FliQ and FliR and three cytoplasmic protein FliH, FliI and FliJ. In addition, flagellar export chaperones such as FlgN and FliT interact with the type III protein export apparatus proteins to promote the export of their cognate building blocks. Crystal structures of the C-terminal cytoplasmic domain of Salmonella FlhA (FlhAC) (PDB ID, 3A5I), the C-terminal cytoplasmic domain of Salmonella FlhB (FlhBC) (PDB ID, 3B0Z), the Salmonella FliP/FliQ/FliR complex with a 5 FliP (green): 4 FliQ (blue): 1 FliR (red) stoichiometry (PDB ID, 6R69), Salmonella FlhE (PDB ID, 4QXL), Salmonella FliH2FliI complex with a 2 FliH: 1 FliI stoichiometry (PDB ID, 5B0O), Salmonella FliJ (PDB ID, 3AJW) and the Salmonella FlgN (PDB ID, 5B3D) and FliT (PDB ID, 3A7M) export chaperones are shown. FlhAC consists of four compactly folded domains, D1, D2, D3 and D4 and a flexible linker connecting FlhAC and the N-terminal transmembrane region of FlhA. Hinge loops of the FlgN and FliT chaperones are responsible for helical rearrangements of these chaperone structure to modulate the binding affinity for their binding partners. CM—cytoplasmic membrane

Full size image

FliP, FliQ, FliR form a helical assembly with a 5 FliP: 4 FliQ: 1 FliR stoichiometry inside the central pore of the MS ring (Fig. 8.5) (Fabiani et al. 2017; Fukumura et al. 2017; Kuhlen et al. 2018; Butan et al. 2019). The FliP5/FliQ4/FliR1 complex is thought to be connected with the most proximal part of the rod to form a continuous protein channel inside the growing axial structure (Kuhlen et al. 2018). FlhA forms a homo-nonamer through intermolecular interactions between its C-terminal cytoplasmic domains (FlhAC) (Abrusci et al. 2013; Kawamoto et al. 2013; Morimoto et al. 2014; Terahara et al. 2018) and acts as an energy transducer along with the cytoplasmic ATPase complex (Minamino et al. 2011b, 2016b; Erhardt et al. 2017). FlhA associates not only with the FliP5/FliQ4/FliR1 helical structure but also with the MS ring (Kihara et al. 2001; Fukumura et al. 2017). A single copy of FlhB associates with FliP and FliR (Kuhlen et al. 2019). The assembly of the export gate complex begins with the formation of the FliP5/FliR1 complex with the help of the transmembrane protein FliO, followed by the assembly of FliQ and FlhB presumably in this order and finally the assembly of FlhA (Morimoto et al. 2014; Fabiani et al. 2017; Fukumura et al. 2017; Kuhlen et al. 2018). FlhAC and the C-terminal cytoplasmic domain of FlhB (FlhBC) project into the central cavity of the C ring (Fig. 8.5) and form a docking platform for the cytoplasmic ATPase complex, flagellar export chaperones and building blocks (Bange et al. 2010; Minamino et al. 2010, 2012). These two FlhAC and FlhBC domains mediate well-organized protein targeting and transport for efficient assembly of the axial structure with the help of the cytoplasmic ATPase complex and the molecular ruler protein FliK (Table 8.1) (Minamino and Macnab 2000a; Fraser et al. 2003; Hirano et al. 2009; Kinoshita et al. 2013; Minamino et al. 2016a, 2020b; Inoue et al. 2018, 2019).

FliI is a member of the Walker-type ATPase family (Fan and Macnab 1996) and forms a homo-hexamer to hydrolyze ATP at an interface between FliI subunits in a way similar to the α3β3 hetero-hexameric ring complex of F1-ATPase (Claret et al. 2003; Imada et al. 2007; Minamino et al. 2006). FliJ binds to the center of the FliI hexamer in a way similar to the γ subunit of the F1-ATPase (Ibuki et al. 2011). The FliH dimer, which structurally looks similar to the peripheral stalk of the F1-ATPase (Imada et al. 2016), binds to the N-terminal domain of each FliI subunit (Okabe et al. 2009) and anchors the FliI6FliJ ring complex to the flagellar base through interactions of the extreme N-terminal region of FliH with FliN and FlhA (Bai et al. 2014; Gonzalez-Pedrajo et al. 2006; Minamino et al. 2009; Hara et al. 2012). ATP hydrolysis by the FliI ATPase activates the PFM-driven export gate complex through an interaction between the FliJ stalk and FlhA, allowing the export gate complex to transport building blocks from the cytoplasm to the distal end of the growing structure in a PMF-dependent manner (Minamino et al. 2011b; Ibuki et al. 2013). The FliI ATPase shows rapid exchanges between the basal body and the cytoplasmic pool, suggesting that the cytoplasmic ATPase complex seems to be a dynamic structure (Bai et al. 2014).

When the cytoplasmic ATPase complex is missing, the transmembrane export gate complex utilizes sodium motive force across the cell membrane in addition to PFM (Minamino et al. 2016b). FlhA acts as the transmembrane ion channel of the export gate complex to conduct not only protons but also sodium ions (Minamino et al. 2016b). The transmembrane export gate complex functions as a protein/cation antiporter that couples the inward-directed cation flow through the FlhA channel to the outward-directed protein transport (Minamino et al. 2011b). The flhE gene forms an operon together with the flhB and flhA genes, and its gene product is secreted via the Sec translocon into the periplasm (Minamino et al. 1994). FlhE is postulated to act as a plug to prevent the FlhA ion channel from translocating large numbers of protons along an electrochemical difference of protons across the cytoplasmic membrane (Lee et al. 2015), but it remains unknown how it works. High-resolution local pH measurements around the type III protein export apparatus have shown that a loss of the ATPase ring complex increases the local pH inside the C ring, indicating ATP hydrolysis by the ATPase ring complex facilitates the proton channel activity of the FlhA channel coupled with flagellar protein export (Morimoto et al. 2016).

The FlgN, FliS and FliT chaperones not only suppresses premature aggregation and/or proteolysis of their cognate building blocks in the cytoplasm (Auvray et al. 2001; Bennett et al. 2001; Aldridge et al. 2003) but also facilitate their docking to the type III protein export apparatus (Thomas et al. 2004; Evans et al. 2006; Kinoshita et al. 2013). The FlgN, FliS and FliT chaperones adopt an α-helical structure (Fig. 8.5) (Evdokimov et al. 2003; Imada et al. 2010; Kinoshita et al. 2016), and the binding of their cognate building block to these chaperones induces helical rearrangements of the chaperones through flexible hinge loops, allowing the building blocks to be efficiently transferred to the type III protein export apparatus, followed by efficient unfolding and the subsequent translocation of the building blocks by the PMF-driven export gate complex (Imada et al. 2010; Kinoshita et al. 2013, 2016; Furukawa et al. 2016).

Conclusions

Detailed domain structures of bacterial flagellar component proteins have been revealed at an atomic resolution by X-ray crystallography. In addition, cryo-EM image analyses of the bacterial flagellum have provided deep insights into the mechanical functions of the rod, the hook and the filament. Cryo-electron tomographic techniques have revealed the in situ structures of the transmembrane part of the basal body structure. However, while static structural knowledge is increasing, there are still many mysterious points that do not connect to dynamic functions of the bacterial flagellum. Biophysical and molecular genetic approaches combined with structural studies are required to more accurately investigate the dynamic mechanism of the flagellar motor complex.