SEARCH

SEARCH BY CITATION

Summary

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

MotA and MotB form a transmembrane proton channel that acts as the stator of the bacterial flagellar motor to couple proton flow with torque generation. The C-terminal periplasmic domain of MotB plays a role in anchoring the stators to the motor. However, it remains unclear where their initial binding sites are. Here, we constructed Salmonella strains expressing GFP-MotB and MotA-mCherry and investigated their subcellular localization by fluorescence microscopy. Neither the D33N and D33A mutations in MotB, which abolish the proton flow, nor depletion of proton motive force affected the assembly of GFP-MotB into the motor, indicating that the proton translocation activity is not required for stator assembly. Overexpression of MotA markedly inhibited wild-type motility, and it was due to the reduction in the number of functional stators. Consistently, MotA-mCherry was observed to colocalize with GFP-FliG even in the absence of MotB. These results suggest that MotA alone can be installed into the motor. The R90E and E98K mutations in the cytoplasmic loop of MotA (MotAC), which has been shown to abolish the interaction with FliG, significantly affected stator assembly, suggesting that the electrostatic interaction of MotAC with FliG is required for the efficient assembly of the stators around the rotor.


Introduction

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

Many bacteria can swim towards favourable conditions by rotating their flagella. The bacterial flagellum acts as a rotary motor powered by the flow of proton or sodium ion by the electrochemical potential gradient across the cytoplasmic membrane. The flagellum consists of at least three parts: the basal body, the hook and the filament. The basal body is embedded within the cell membranes and acts as a rotary motor. The hook and filament extend outwards in the cell exterior. The filament acts as a helical propeller. The hook connects the filament with the basal body and functions as a universal joint to smoothly transmit torque produced by the motor to the helical filament (Minamino and Namba, 2004; Minamino et al., 2008).

In Escherichia coli and Salmonella enterica, MotA, MotB, FliG, FliM and FliN, are responsible for torque generation (Berg, 2003; Kojima and Blair, 2004; Sowa and Berry, 2008). MotA and MotB are transmembrane proteins that form the stator as well as the proton channel (Block and Berg, 1984; Blair and Berg, 1988; 1990; Wilson and Macnab, 1990; Stolz and Berg, 1991; Tang et al., 1996). The stator unit consists of four copies of MotA and two copies of MotB (Braun and Blair, 2001; Braun et al., 2004; Kojima and Blair, 2004). FliG, FliM and FliN form the C ring on the cytoplasmic face of the MS ring, which is formed by 26 copies of FliF in the cytoplasmic membrane (Francis et al., 1994; Suzuki et al., 2004). The MS ring and C ring together act as the rotor. FliG, FliM and FliN are responsible not only for torque generation but also for switching the direction of motor rotation (Yamaguchi et al., 1986). The electrostatic interactions of the cytoplasmic loop of MotA with FliG are important for torque generation (Zhou et al., 1998a). A high-resolution observation of flagellar motor rotation has revealed a fine stepping motion with 26 steps per revolution (Sowa et al., 2005; Nakamura et al., 2010), and the number corresponds to that of FliG subunits in the C ring (Suzuki et al., 2004). Although it remains unknown how the stator complex couples the proton flow to torque generation, two highly conserved residues, Pro-173 in MotA and Asp-33 in MotB, are involved in the energy coupling mechanism (Sharp et al., 1995; Togashi et al., 1997; Zhou et al., 1998b; Braun et al., 1999; Kojima and Blair, 2001; Che et al., 2008; Nakamura et al., 2009a). Two evolutionally conserved charged residues, Arg-90 and Glu-98, in the cytoplasmic loop of MotA, are responsible for the interactions with charged residues of FliG for torque generation (Zhou and Blair, 1997; Zhou et al., 1998a).

Up to about 11 copies of the MotA/B complex are installed into a motor to act as stators (Reid et al., 2006). Stator resurrection experiments have shown that abrupt drops in the rotation rate occur frequently (Block and Berg, 1984; Blair and Berg, 1988; Sowa et al., 2005). Consistently, total internal reflection fluorescence microscopy has revealed the turnover of GFP-fused MotB between the membrane pool and the basal body (Leake et al., 2006). These results suggest that the association of the MotA/B complex to its target sites on the flagellar basal body is highly dynamic. Since MotB has a highly conserved peptidoglycan (PG)-binding motif in its C-terminal periplasmic domain (MotBC) (De Mot and Vanderleyden, 1994), presumably to anchor the MotA/B complex to be the stator around the rotor, the PG binding of MotBC must occur only upon stator assembly. It has been shown that the dimerization of MotBC is responsible for the proper targeting and stable anchoring of the MotA/B complex around the rotor (Kojima et al., 2008; 2009). Recently, site-directed disulphide cross-linking experiments have revealed an interaction between MotBC and the P ring of the motor (Hizukuri et al., 2010).

It has been shown that polar localization of the PomA/B complex of Vibrio alginolyticus, which are homologues of the MotA/B complex of a Na+-driven flagellar motor, is greatly affected by changes in the external Na+ concentration (Fukuoka et al., 2009). This suggests that Na+ is required not only for torque generation but also for the efficient assembly process of PomA/B (Fukuoka et al., 2009). Furthermore, Shewanella oneidensis MR-1 has two distinct stators consisting of either MotA/B or PomA/B within a single flagellar motor and selects the stator to use at the level of protein localization in response to the Na+ concentration (Paulick et al., 2009).

The proton channel of the Salmonella MotA/B complex is also regulated to open only upon stator assembly by the plug segment consisting of residues 53–66 of MotB, which is just after its single transmembrane segment. The deletion of the plug segment results in proton leakage through free MotA/B complexes not assembled into the motor and thereby impairs cell growth (Hosking et al., 2006; Morimoto et al., 2010). Interestingly, however, the deletion of residues 51–100 of Salmonella MotB still allows a functional stator to be formed with MotA despite that the plug is missing (Muramoto and Macnab, 1998; Kojima et al., 2009). But the crystal structure of a MotBC fragment corresponding to residues 99–276 of MotB indicates that it is too small to reach the PG layer if connected directly to the transmembrane helix by the deletion of residues 51–100, suggesting that a relatively drastic conformational change in the N-terminal portion of MotBC should occur upon stator assembly to allow the stator to be anchored and open the proton channel (Kojima et al., 2009). Therefore, it is quite possible that an interaction between MotA and FliG initiates the binding of the MotA/B complex to the motor and trigger conformational changes in MotBC.

Although the structures of the entire basal body containing the stators have been visualized by electron cryotomography (Murphy et al., 2006; Liu et al., 2009; Kudryashev et al., 2010), the resolution is still limited. Since the stator is missing in highly purified flagellar basal bodies (Minamino et al., 2008), relatively high-resolution structures of the basal body by electron cryomicroscopy and single particle image analysis do not show the binding mode of the stators (Thomas et al., 2006). Therefore, it is still unclear how the MotA/B complex is localized to the basal body to be the stator.

To understand the localization mechanism of the MotA/B complex, we investigated the subcellular localization of GFP-MotB and MotA-mCherry by fluorescence microscopy. We show that overexpression of MotA inhibits wild-type motility due to the reduced number of functional stators in the flagellar motor. This suggests that MotA can be installed into the motor even in the absence of MotB. We also show that the R90E mutation in the cytoplasmic loop of MotA significantly affects the subcellular localization of GFP-MotB, suggesting that the electrostatic interaction between the cytoplasmic loop of MotA and FliG is required for the efficient assembly of the stators around the rotor.

Results

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

Effect of the motB(D33N) mutation on the subcellular localization of the stator complex

A highly conserved aspartic acid residue of Vibrio PomB, Asp-24, is thought to be a Na+ binding site. The D24N mutation in PomB abolishes polar localization of the PomA/B complex, suggesting that the binding of Na+ to Asp-24 is critical for stator assembly into the motor (Fukuoka et al., 2009). In contrast, the E. coli motB(D32N) and motB(D32A) alleles, which are non-functional, exert a strong negative-dominant effect on motility of wild-type cells, suggesting that the MotA/B(D32N) and MotA/B(D32A) complexes can be installed into the motor (Zhou et al., 1998b; Kojima and Blair, 2001). Therefore, to investigate whether analogous D33N and D33A mutations of Salmonella MotB affect stator assembly, we constructed Salmonella strains that encode either GFP-MotB, GFP-MotB(D33N) or GFP-MotB(D33A) on the chromosome under the control of the promoter of the motAB operon and analysed their subcellular localization by fluorescence microscopy (Fig. 1). In agreement with a previous report (Leake et al., 2006), GFP-MotB was partially functional (Fig. 1A). Most of the GFP-MotB molecules were intact in the cells as judged by immunoblotting with polyclonal anti-MotB antibody although a very faint band of MotB was seen as a cleaved product (Fig. 1B). To examine the subcellular localization of GFP-MotB, cells expressing GFP-MotB were observed by epi-illumination fluorescence microscopy (Fig. 1C). We used a Salmonella strain encoding GFP-FliG as a positive control because FliG forms part of the C ring (Francis et al., 1994). This strain was also partially functional (Fig. 1A) and GFP-FliG was mostly intact in the cell (Fig. 1B). Fluorescent spots of GFP-FliG and GFP-MotB were observed within the cell bodies of each strain (Fig. 1C). Some spots of GFP-FliG and GFP-MotB were observed at the centre of rotation of cells tethered to a coverslip by a single flagellar filament (Movies S1–S4). In agreement with this, the fluorescent spots of GFP-FliG and GFP-MotB were observed at the base of the filament labelled with a fluorescent dye (Fig. 1D). In contrast, when GFP-MotB was expressed in the fliF flgM double mutant, which allows GFP-MotB to be expressed from the motAB promoter even in the presence of the fliF mutation (Kutsukake and Iino, 1994), no spots were detected, and only diffuse fluorescent signals were observed throughout the cell body (Fig. S1). These results demonstrate that the GFP-labelled MotA/B complex can be installed into the motor. The D33N and D33A mutations caused a loss-of-function of GFP-MotB (Fig. 1A), in agreement with a previous report (Zhou et al., 1998b). However, these mutations did not affect the subcellular localization of GFP-MotB (Fig. 1C and D). These results suggest that proton binding to Asp-33 of MotB is not critical for stator assembly around the rotor.

image

Figure 1. Effect of the motB(D33N) mutation on the subcellular localization of GFP-MotB. A. Swarming motility of SJW1103 (wild type, indicated as WT), YVM003 (GFP-MotB), YVM004 (GFP-FliG), YVM007 [GFP-MotB(D33N)] and YVM030 [GFP-MotB(D33A)] in semi-solid agar plates. Plates were incubated at 30°C for 7 h. B. Immunoblotting, using polyclonal anti-MotB (upper panel) and anti-FliG (lower panel) antibodies, of whole-cell proteins prepared from the same cells. The positions of molecular mass markers (kDa) are shown on the left. Arrowheads indicate the positions of MotB, GFP-MotB, FliG and GFP-FliG. C. Fluorescence (epi) and bright-field images (BF) of YVM003 (GFP-MotB), YVM004 (GFP-FliG), YVM007 [GFP-MotB(D33N)] and YVM030 [GFP-MotB(D33A)]. The cells were grown overnight in LB at 30°C and observed by fluorescence microscopy. D. Colocalization of GFP-MotB, GFP-MotB(D33N) and GFP-FliG with the flagellar filaments labelled with the Alexa fluorescent dye. The cells were treated with polyclonal anti-FliC antibody and Alexa Fluor 594-conjugated anti-rabbit secondary antibody. The fluorescence images of GFP-MotB, GFP-MotB(D33N) and GFP-FliG (green) and the filament labelled with Alexa (red) were merged in the right panel.

Download figure to PowerPoint

Na+ motive force across the cytoplasmic membrane is critical for polar localization of the PomA/B complex (Fukuoka et al., 2009). Therefore, we investigated whether PMF is required for the subcellular localization of GFP-labelled stators. Since PMF can be collapsed by a protonophore, carbonyl cyanide m-chlorophenylhydrazone (CCCP), we analysed the effect of CCCP on stator assembly (Fig. 2). In the presence of 50 µM CCCP, the rotation of tethered cells expressing GFP-MotB was totally inhibited (Fig. 2B, Movies S5–S7), indicating that PMF is totally collapsed. However, the number and the intensity of the fluorescent spots of GFP-MotB were both mostly unchanged by a CCCP treatment for 30 min (Fig. 2A, Table 1). Interestingly, new fluorescent spots occasionally appeared over the treatment period. When CCCP was removed by replacing the motility buffer, the tethered cells rotated again (Fig. 2B, Movies S5–S7). These results suggest that PMF is not required for stator assembly around the rotor in the proton-driven flagellar motor.

image

Figure 2. Effect of CCCP on the subcellular localization of GFP-MotB. A. Fluorescence images of YVM003 (GFP-MotB) before and after the treatment with 50 µM CCCP for 30 min at room temperature. B. Bright-field (BF) and fluorescence (epi) images of tethered rotational GFP-MotB cell. When the tethered cell was treated with 50 µM CCCP, and the rotation immediately stopped. After washing by the motility medium without CCCP, the rotation restarted. White straight arrows indicate the fluorescent spot at the centre of rotation. Arc-shaped arrows indicate the rotation of the cell by which the image of the cell body is blurred.

Download figure to PowerPoint

Table 1.  Effect of CCCP on the localization of GFP-MotB.
 GFP-MotB
50 µM CCCPNumber of spots per cellFluorescence intensity (A.U.)
1.9 ± 0.9 (n = 176)1486 ± 257 (n = 40)
+2.3 ± 0.8 (n = 149)1439 ± 301 (n = 40)

Multicopy effect of MotA on motility of wild-type cells

To test if MotA alone can be installed into a motor, we first analysed the multicopy effect of MotA on motility of wild-type cells in semi-solid agar plates. When MotA was over-produced (Fig. 3A, lane 2), the swarming size decreased significantly compared with the vector control (Fig. 3B). We also measured free-swimming motility in liquid media by phase-contrast microscopy and found that the swimming speed of the cells overexpressing MotA was significantly reduced (Fig. 3C). In agreement with a previous report (Wilson and Macnab, 1988), cell growth was not impaired by the overexpression of MotA (Fig. 4A). Consistently, neither membrane potential nor intracellular pH was changed by the overexpression (Fig. 4B), indicating that PMF across the cytoplasmic membrane was not affected either. Therefore, we conclude that the poor motility of the cells overexpressing MotA is not a consequence of either growth impairment or reduced PMF.

image

Figure 3. Multicopy effect of MotA on motility of wild-type cells. A. Immunoblotting, using polyclonal anti-MotA and anti-MotB antibody, of whole-cell proteins prepared from the same transformants. The positions of molecular mass markers (kDa) are shown on the left. Arrowheads indicate the positions of MotA and MotB. Since our polyclonal anti-MotA antibody cannot detect the chromosomal expression level of MotA, no band of MotA was seen in lane 1. B. Swarming motility assay of SJW1103 (wild type) transformed with pKK223-3 (Vector, V) and pKSS12 (MotA) on soft agar plates. The plates were incubated at 30°C for 6 h. C. Free-swimming speed measured by phase-contrast microscopy. Measurements were performed at c. 23°C. Vertical bars are standard errors.

Download figure to PowerPoint

image

Figure 4. Effect of MotA over-production on growth rate and the proton motive force across the cytoplasmic membrane. A. Growth curve of the wild-type strain SJW1103 transformed with pKK223-3 (V) (open circles) or pKSS12 (MotA) (closed circles). The cells were grown in T-broth at 30°C with shaking. The optical density at 600 nm (OD600) was measured. B. Relative membrane potential (Δψ) and transmembrane proton gradient (ΔpH). The membrane potential was measured using TMRM. Intracellular pH was measured with pHluorin at an external pH of 7.0. Both Δψ and ΔpH are normalized to those of the vector control. The Δψ of the cells harbouring pKK223-3 or pKSS12 is the average of 135 cells and 111 cells respectively. The data of ΔpH are the average of six independent experiments. Vertical bars indicate standard errors.

Download figure to PowerPoint

MotA interacts with FliG, which is not only responsible for torque generation and the switching of rotational direction of the motor but is also needed for export of flagellar proteins (Yamaguchi et al., 1986; Minamino and Macnab, 1999). To test whether excess MotA might titrate FliG away from the motor, resulting in the inhibition of flagellar protein export, we observed wild-type cells overexpressing MotA by high-intensity dark-field light microscopy and found that the number of flagellar filaments was not affected (data not shown). Chemotactic behaviour was also not changed (data not shown). These results indicate that the reduced motility is neither due to reduction in the flagellar protein export activity nor due to increase in the probability of CW-biased rotation.

To investigate whether the poor motility is due to a reduced torque, we next carried out bead assays of motor rotation with polystyrene and fluorescent beads (Fig. 5 and Table 2). The torque–speed relationships of Salmonella MM3076iC cells overexpressing MotA are shown in Fig. 5A in comparison with a vector control. Rotation rates of the motors of the cells overexpressing MotA were reduced over the entire range of observation under various amounts of load. The zero-speed torque, as obtained by zero-speed extrapolation of the torque–speed curve, was c. 60% of the vector control. The zero-torque speeds as deduced by extrapolation of the regression lines also decreased from 225 Hz of the vector control to 145 Hz when MotA was over-produced. The torque at high load is dependent on the number of stators while the motor speed near zero load is independent of the number of functional stators (Ryu et al., 2000; Reid et al., 2006; Yuan and Berg, 2008). These results suggest that free MotA can be installed into the motor to reduce torque even without forming the MotA/B complex and also inhibit high-speed rotation at near zero load. It is also possible that excess MotA may induce the formation of anomalous stator complexes, thereby disturbing motor function.

image

Figure 5. Effect of MotA over-production on the torque–speed relationship of the flagellar motor. A. Torque–speed curve of MM3076iC [Δ(cheA-cheZ), fliC(Δ204–292)] carrying pKK223-3 (left panel) or pKSS12 (MotA) (right panel). Rotation measurements of individual flagellar motors were carried out by tracking the position of 1.0 µm (closed circle), 0.8 µm (open circle), 0.5 µm (closed triangle) and 0.1 µm beads (open triangle) attached to the sticky flagellar filament. All the measurements were performed at c. 25°C. B. Speed histogram of MM3076iC carrying pKK223-3 (left) or pKSS12 (right). Rotation rates of single flagellar motors labelled with 1.0 µm beads were determined from the power spectrum using 1 s data windows (1024 points) at an interval of 0.1 s (Table 2). The number labels above the data represent the units corresponding to multiples of 7 Hz. The peaks of rotation rates correspond to these units, indicating the presence of different numbers of functional stators in the motor.

Download figure to PowerPoint

Table 2.  Effect of MotA over-production on flagellar motor rotation.
BeadVectorMotA
Rotation rate (Hz)Torque (pN nm)Rotation rate (Hz)Torque (pN nm)
1.0 µm65 ± 71705 ± 19040 ± 181122 ± 440
n = 17n = 27
0.8 µm101 ± 151519 ± 17261 ± 24960 ± 355
n = 32n = 29
0.5 µm169 ± 32854 ± 15281 ± 38424 ± 198
n = 38n = 48
0.1 µm218 ± 57449 ± 115130 ± 62270 ± 130
n = 13n = 20

Multicopy effect of MotA on the subcellular localization of GFP-labelled stators

To test whether overexpression of MotA actually reduces the number of functional stators in a rotating flagellar motor, we analysed the multicopy effect of MotA on the subcellular localization of GFP-MotB (Fig. 6A). When both MotA and MotB were expressed in the GFP-MotB encoding strain (Fig. 6B, lane 3), the fluorescent spots of GFP-MotB drastically decreased (Fig. 6A, top right) compared with the vector control (Fig. 6A, top left), indicating that non-labelled MotB molecules assembles into the motor along with MotA. Overexpression of MotA also reduced not only the number of fluorescent spots of GFP-MotB but also their fluorescence intensity although not to the levels seen in the cells overexpressing MotA/B (Fig. 6A, top middle). In contrast, when MotA alone or MotA/B was overexpressed in the GFP-FliG encoding strain, the number of GFP-FliG spots was not changed (Fig. 6A, bottom row). These results suggest that a certain number of the MotA/GFP-MotB complexes are replaced with MotA in the motor.

image

Figure 6. Multicopy effect of MotA on the subcellular localization of GFP-MotB and GFP-FliG. A. Fluorescence (epi) and bright-field images (BF) of the Salmonella YVM003 (gfp-motB) and YVM004 (gfp-fliG) strains transformed with pKK223-3 (V), pKSS12 (MotA) and pKSS13 (MotA/B). The cells were grown overnight at 30°C and observed by fluorescence microscopy. B. Immunoblotting, using polyclonal anti-MotA (upper panel) and anti-MotB (lower panel) antibodies, of whole-cell proteins prepared from the same transformants. Lane 1, YVM003 carrying pKK223-3; lane 2, YVM003 harbouring pKSS12; lane 3, YVM003 carrying pKSS13; lane 4, YVM004 carrying pKK223-3; lane 5, YVM004 harbouring pKSS12; lane 6, YVM004 carrying pKSS13. Arrowheads indicate the positions of MotA, MotB and GFP-MotB.

Download figure to PowerPoint

To estimate the number of functional stators in the motor of Salmonella cells overexpressing MotA, the position of the bead was determined at a sampling rate of 1 kHz for 10 s, and the rotation rates were determined from the power spectra calculated from 1-s-long data windows that were shifted every 0.1 s over the entire 10 s data (Fig. 5B). The vector control showed a relatively narrow distribution in the speed histogram with the speed ranging from 40 to 80 Hz (Fig. 5B, left). Since a single stator unit can drive the proton-driven flagellar motor labelled with 1.0 µm beads at c. 7 Hz (Reid et al., 2006; Nakamura et al., 2010), the maximum number of stators in the fully functional motor is estimated to be 11, in agreement with a previous report (Reid et al., 2006). In contrast, the motor speeds of MotA-expressing cells were distributed over a wide range from 15 to 70 Hz and also showed several distinct peaks separated by a roughly equal interval of around 7 Hz in the histogram (Fig. 5B, right). These results let us assume that the number of functional stators decreased to 5 ± 3 by the overexpression of MotA.

Fluorescence microscopy observation of MotA-mCherry

To obtain more direct evidence that MotA can assemble into the motor even in the absence of MotB, we constructed Salmonella YVM018 strain (gfp-fliG, motA-mCherry, ΔmotB) with or without MotB expressed from a pTrc99A-based plasmid, and analysed the subcellular colocalization of MotA-mCherry and GFP-FliG by epifluorescence microscopy (Fig. 7). MotA-mCherry was not functional (data not shown). However, fluorescent spots of MotA-mCherry were occasionally observed at the same position as those of GFP-FliG whether MotB is present or not (Fig. 7A and B), although the probability of fluorescent-spot observation of MotA-mCherry was much less than that of GFP-MotB, presumably due to its loss-of-function phenotype. When MotA-mCherry was expressed in the fliF flgM double mutant background, only diffuse fluorescent signals were observed throughout the cell body (Fig. S2). These results suggest that MotA can be incorporated into the flagellar motor to some extent even without MotB and that the localization of MotA is dependent on its interaction with other components of the flagellar motor.

image

Figure 7. Subcellular localization of MotA-mCherry. A. Fluorescence images of YVM018 (gfp-fliG, motA-mCherry, ΔmotB) transformed with pTrc99A (V) or pHMK1602 (MotB). The cells were grown in LB at 30°C for 5 h and observed by fluorescence microscopy. The fluorescence images of GFP-FliG (green) and MotA-mCherry (red) were merged in the right panel. B. Immunoblotting, using polyclonal anti-RFP (upper panel) and anti-MotB (lower panel) antibodies, of whole-cell proteins prepared from the same transformants. Arrowheads indicate the positions of MotA-mCherry and MotB.

Download figure to PowerPoint

Effect of the MotA(R90E) mutation on the motility

Two charged residues, Arg-90 and Glu-98, which are in the cytoplasmic loop of MotA, are thought to interact with charged residues of FliG for torque generation (Zhou and Blair, 1997; Zhou et al., 1998a). It has been reported that the motA(R90E) and motA(E98K) alleles are recessive, although the MotA(R90E) and MotA(E98K) proteins are as stable as wild-type MotA (Zhou and Blair, 1997). These results suggest that these mutant proteins are not efficiently incorporated into the motor. Since it has been shown that there is fundamental difference between E. coli and Salmonella flagellar motors (Nakamura et al., 2009a), we analysed the effects of the Salmonella MotA(R90E) and MotA(E98K) mutations on the motility in liquid media (Fig. 8A and B, and Movies S8 and S9). When the expression levels of MotA(R90E) and MotB were almost the same as those expressed from the chromosome of wild-type cells, almost all of the cells expressing the MotA(R90E)/B complex were non-motile. However, an increment of the expression level of MotA(R90E)/B allowed about 70% of the cells to become motile, and their swimming speed reached about 60% of the wild-type levels. To test whether no motility of the motA(R90E) mutant under low-induction condition is due to the poor assembly of the MotA(R90E)/B complex around the rotor, we constructed the Salmonella motA(R90E), gfp-motB strain and analysed the subcellular localization of GFP-MotB. The MotA(R90E) substitution drastically decreased both the number and intensity of the fluorescent spots of GFP-MotB (Fig. 8C), although the expression level of GFP-MotB was not changed (Fig. 8D). These suggest that the R90E mutation just reduces the binding affinity of the stator to the rotor. In contrast, the motA(E98K) allele was non-functional even under high-induction condition and did not exert any dominant negative effect on wild-type motility (data not shown), in agreement with a previous report (Zhou and Blair, 1997). This indicates that Glu-98 is critical for the assembly of the MotA/B complex into the motor.

image

Figure 8. Motility of cells producing MotA(R90E) and MotB in liquid media. A. Immunoblotting, using polyclonal anti-MotB antibody, of whole-cell proteins prepared from SJW1103 (wild type) harbouring pBAD24 (V), and SJW2241(ΔmotAB) carrying pBAD24 (V), pYC20 (WT) or pYC20(R90E) (R90E) were incubated at 30°C for 5 h in LB with 0%, 0.002%, 0.02%, 0.2% arabinose. Arrowhead indicates the position of MotB. B. Free-swimming fraction and speed of SJW2241(ΔmotAB) carrying pYC20 (WT) (open bar) and pYC20(R90E) (R90E) (shaded bar) measured by phase-contrast microscopy. Swimming fraction is the number fraction of swimming cells. Swimming speed is the average of more than 20 cells, and vertical lines are standard errors. If the fraction of motile cells was less than 5% of total cells, the swimming speed is presented as zero. Measurements were performed at c. 23°C. C. Bright-field (BF) and fluorescence (epi) images of YVM003 (MotA/GFP-MotB) and YVM031 (MotA(R90E)/GFP-MotB). The cells were grown overnight in LB at 30°C and observed by fluorescence microscopy. D. Immunoblotting, using polyclonal anti-MotB antibodies, of whole-cell proteins prepared from the same strains. Arrowhead indicates the position of GFP-MotB.

Download figure to PowerPoint

Discussion

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

MotB consists of the N-terminal cytoplasmic region, one transmembrane helix, and the C-terminal periplasmic region named MotBC, which contains the PG-binding motif (Berg, 2003; Kojima and Blair, 2004; Sowa and Berry, 2008). Since the MotBC fragment inhibits motility of wild-type cells when exported into the periplasm, MotBC is proposed to be required for proper targeting and stable anchoring of the MotA/B complex to the motor (Kojima et al., 2008). The stators can be replaced even during motor rotation, indicating that the stator does not permanently bind to the motor and the peptidoglycan layer (Leake et al., 2006). The proton-conductivity of free MotA/B complexes is suppressed by the plug segment in the MotBC region when they are not associated with the motor (Hosking et al., 2006; Morimoto et al., 2010). However, since the MotA/B complex is not co-purified with the hook-basal body (Minamino et al., 2008), it remains unclear how the MotA/B complex finds and assembles into the motor and activates its proton-conductivity to become a functional stator. In the present study, we analysed the multicopy effect of MotA on the torque–speed relationships of the flagellar motor and the subcellular localization of GFP-labelled stators and obtained evidence suggesting that MotA can be installed into the flagellar motor even in the absence of MotB.

MotA/B(D33N) and MotA/B(D33A) can assemble into the motor

Stator assembly in a Na+-driven flagellar motor responds to Na+ concentration (Fukuoka et al., 2009; Paulick et al., 2009). Mutations in the amino acid residues of PomB involved in the Na+-relay mechanism significantly interfere with the polar localization of the PomA/B complex (Fukuoka et al., 2009). In this study, we found that the motB(D33N) and motB(D33A) mutations, which abolish proton translocation through the proton channel (Zhou et al., 1998b; Morimoto et al., 2010), does not affect the localization of GFP-MotB at the base of the flagellar filament (Fig. 1). This is in agreement with previous data showing that E. coli MotB(D32N) and MotB(D32A) exert a negative dominant effect on wild-type motility (Zhou et al., 1998b; Kojima and Blair, 2001). Therefore, stator assembly in the proton-driven motor is not dependent on the proton conductive activity of the MotA/B complex.

It has been reported that the stators of the Na+-driven motor detach from the rotor when motor is de-energized and stopped (Sowa et al., 2005; Fukuoka et al., 2009). In this study, the treatment of the proton-driven motor by 50 µM CCCP, which totally inhibited the rotation of tethered cells, did not affect the subcellular localization of GFP-MotB (Fig. 2), suggesting that the stators remain to exist around the rotor even in the absence of PMF. In agreement with this, it has been shown that the stator actually switches its functional state between the active and inactive ones without detaching from the rotor completely when PMF is largely reduced (Nakamura et al., 2010). Therefore, we conclude that the assembly of the MotA/B complex into the motor is not obligatorily linked to the process of the proton translocation through the proton channel of the MotA/B complex.

Overexpression of MotA inhibits motility of wild-type cells

It has been shown that the in-frame deletion of residues 51–100 in Salmonella MotB does not markedly impair motility (Muramoto and Macnab, 1998; Kojima et al., 2009). In agreement with this, the flagellar motors produced torque at wild-type level upon full induction of the MotA/B(Δ51–100) complex from an IPTG-inducible promoter (S. Nakamura, unpubl. data), suggesting that MotA/B(Δ51–100) can be properly incorporated into the motor. However, the crystal structure of a MotBC fragment corresponding to residue 99–276 of MotB is so small that MotB(Δ51–100) cannot reach the PG layer if the fragment is directly connected to the transmembrane helix by the deletion of residues 51–100 (Kojima et al., 2009), raising the question of how the MotA/B complex finds the stator binding site on the motor. Here, we showed that overexpression of MotA greatly reduces wild-type motility both in semi-solid agar plates and in liquid media (Fig. 3). Neither the growth rate (Fig. 4A), flagellation (data not shown), nor chemotactic behaviour (data not shown) was changed by overexpression of MotA. However, the torque–speed relationship of the flagellar motor was significantly affected (Fig. 5). The rotation rate of the proton-driven motor at a given load is proportional to PMF (Gabel and Berg, 2003). Since neither the membrane potential nor the transmembrane proton gradient was changed by overexpression of MotA (Fig. 4B), the approximately 40% reduction in the zero-speed torque by over-production of MotA is not the consequence of reduced PMF. Since the zero-speed torque is dependent on the number of functional stators in the motor (Ryu et al., 2000; Reid et al., 2006), this suggests that MotA occupies the stator binding sites of the motor and reduces the number of functional stators. This is further supported by our data showing that overexpression of MotA reduces not only the number of the fluorescent spots of GFP-MotB but also their fluorescence intensity compared with the vector control (Fig. 6) and that MotA-mCherry is localized at the base of the flagellum to some extent even in the absence of MotB (Fig. 7). We therefore conclude that MotA alone can be installed into the motor. Interestingly, high-speed rotation at near zero load was also decreased by overexpression of MotA (Fig. 5). Since one stator unit can spin the motor at the same speed as many stators do at near zero load (Ryu et al., 2000; Yuan and Berg, 2008), we suggest that MotA incorporated into the motor may attach to the rotor and somehow interfere with high-speed rotation.

Two charged residues of E. coli MotA, Arg-90 and Glu-98, are involved in electrostatic interactions with charged residues of FliG for torque generation (Zhou and Blair, 1997; Zhou et al., 1998a). It has been reported that the motA(R90E) and motA(E98K) alleles are recessive (Zhou and Blair, 1997), suggesting that these mutant proteins are not efficiently incorporated into the motor. Consistently, we also found that Salmonella motA(R90E) and motA(E98K) mutants displayed essentially the same phenotype as those of E. coli when their expression levels were the same as those of wild-type cells (Fig. 8). Interestingly, however, an increase in the expression level of the MotA(R90E)/B complex by more than 10-fold allowed 70% of the cells to swim in liquid media. It was confirmed by markedly decreased number and intensity of fluorescent spots of GFP-MotB (Fig. 8C and D) that no motility of the motA(R90E) mutant results from poor stator assembly around the rotor when the expression level of MotA(R90E)/MotB was the same as that of MotA/B expressed from the chromosome. Since the loss-of-function phenotype of the motA(R90E) and motA(E98K) alleles are considerably suppressed by the fliG(D289K) and fliG(R281V) mutations, respectively (Zhou et al., 1998a), our present results suggest that the interactions between MotA Arg-90 and FliG Asp-289 and between MotA Glu-98 and FliG Arg-281 are critical not only for torque generation but also for the assembly of the stators into the motor.

Experimental procedures

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

Bacterial strains, plasmids and media

Bacterial strains and plasmids used in this study are listed in Table 3. P22-mediated transduction was carried out as described (Yamaguchi et al., 1986). DNA manipulations, site-directed mutagenesis and DNA sequencing were carried out as described previously (Saijo-Hamano et al., 2004). L-broth (LB), semi-solid agar plates and motility medium were used as described previously (Minamino and Macnab, 1999; Minamino et al., 2003). Ampicillin was added to LB at a final concentration of 100 µg ml−1.

Table 3.  Strains and plasmids used in this study.
Strains and plasmidsRelevant characteristicsSource or reference
  • a.

    The YVM018 strain carrying pNSK9, which encodes MotA and MotB on pTrc99A, displayed the Che- phenotype presumably due to the reduced expression levels of the chemotaxis genes caused by a deletion of MotB.

Salmonella
SJW1103Wild type for motility and chemotaxisYamaguchi et al. (1984)
SJW2241ΔmotA-motBKomoriya et al. (1999)
MM3076iCΔ(cheA-cheZ), fliC(Δ204–292)Nakamura et al. (2009b)
MMPH001ΔaraBAD::pHluorinMorimoto et al. (2010)
YVM003gfp-motBThis study
YVM004gfp-fliGThis study
YVM007gfp-motB(D33N)This study
YVM018agfp-fliG, motA-mCherry, ΔmotBThis study
YVM030gfp-motB(D33A)This study
YVM031motA(R90E), gfp-motBThis study
YMM002gfp-motB, ΔflgM::kanThis study
YMM010gfp-motB, ΔflgM::kan, fliF::Tn10This study
YMM019gfp-fliG, motA-mCherry, ΔmotB, ΔflgM::kanThis study
YMM020gfp-fliG, motA-mCherry, ΔmotB, ΔflgM::kan, fliF::Tn10This study
Plasmids
pKK223-3Expression vectorGE Healthcare
pTrc99AExpression vectorGE Healthcare
pBAD24Expression vectorGuzman et al. (1995)
pKSS12pKK223-3/MotAS. Sugiyama
pKSS13pKK223-3/MotA + MotBMorimoto et al. (2010)
pNSK9pTrc99A/MotA + MotBChe et al. (2008)
pHMK1602pTrc99A/MotBH. Matsunami
pYC20pBAD24/MotA + MotBMorimoto et al. (2010)
pYC20(R90E)pBAD24/MotA(R90E) + MotBThis study
pYC20(E98K)pBAD24/MotA(E98K) + MotBThis study

Construction of gfp-fliG, motA-mCherry or gfp-motB strains

To construct Salmonella gfp-fliG, motA-mCherry or gfp-motB strains, the fliG, motA or motB gene on the chromosome was replaced by the gfp-fliG, motA-mCherry or gfp-motB allele, respectively, by using the λ Red homologous recombination system developed by Datsenko and Wanner (2000). The gfp-fliG, motA-mCherry and gfp-motB alleles are placed under the control of their native promoters.

Preparation of whole-cell proteins and immunoblotting

Salmonella cells were grown overnight at 30°C in LB with shaking. Cell pellets were suspended in a SDS-loading buffer and normalized by cell density to give a constant amount of cells. After sodium dodecyl sulphate-polyacrylamide gel electrophoresis (SDS-PAGE), immunoblotting with polyclonal anti-MotA, anti-MotB, anti-GFP (MBL) and anti-RFP (MBL) antibodies was carried out as described previously (Minamino and Macnab, 1999).

Motility assay

For swarming motility, fresh colonies were inoculated in semi-solid agar plates and incubated at 30°C. For measurements of swimming speeds, cells were observed under a phase-contrast microscope at room temperature, and their motile behaviour was recorded on DVD. The swimming speed of individual cells was analysed as described before (Minamino et al., 2003).

Fluorescence microscopy

To observe bacterial cell bodies and epifluorescence of GFP, mCherry and tetramethyl rhodamine methyl ester (TMRM), we used an inverted fluorescence microscope (IX-71, Olympus) with a 150× oil immersion objective lens (UApo150XOTIRFM, NA 1.45, Olympus) and an Electron-Multiplying Charge-Coupled Device (EMCCD) camera (C9100-02, Hamamatsu Photonics). Epifluorescence was excited by a 100 W ultra-high-pressure mercury lamp (USHIO) with the following fluorescence mirror units: U-MGFPHQ (Olympus) for GFP; U-DM-CY3-2 (Olympus) for TMRM; and mCherry-A (Semrock) for mCherry. Fluorescence images of GFP and mCherry were captured by the EMCCD camera with every 1 s exposure. Sequential fluorescence images were also acquired every 100 ms. Fluorescence image processing was performed with the ImageJ version 1.42 software (National Institutes of Health).

Labelling of flagellar filaments with a fluorescent dye

Salmonella cells were attached to a coverslip (Matsunami glass, Japan), and unattached cells were washed away with motility medium. A 1 µl aliquot of polyclonal anti-FliC antibody and 1 µl of anti-rabbit IgG conjugated with Alexa Fluor® 594 (Invitrogen) were suspended in 100 µl of motility medium. A 50 µl aliquot of the buffer containing the fluorescent dye was applied to the cells attached to the coverslip. After washing with motility medium, cells were observed by a fluorescence microscope (Axio Observer, Carl Zeiss Microimaging) with a 100× oil immersion objective lens (αPlan-Apo 100×, NA 1.46, Carl Zeiss Microimaging) and a CCD camera (AxioCam MRm, Carl Zeiss Microimaging).

Measurement of intensities of fluorescent spots of GFP-MotB

The fluorescence images of GFP-MotB before and after a CCCP treatment for 30 min were subjected to 10 iterations of 2D blind deconvolution using the AutoDeblur software (Media Cybernetics). Then, the intensity of each fluorescent spot was determined by fitting a two-dimensional Gaussian function with an image processing program we developed based on the Igor Pro 6.06j software (WaveMetrics).

Measurement of the membrane potential across the cytoplasmic membrane

The membrane potential of SJW1103 (wild-type) carrying pKK223-3 or pKSS12 was measured as described by Lo et al. (2007). Cells were suspended in motility medium plus 10 mM EDTA for 10 min. Cells were washed with motility medium and were incubated in motility medium containing 0.1 µM TMRM (Invitrogen) for 10 min at room temperature. Calculation of the membrane potential was carried out as described (Lo et al., 2007) with minor modifications.

Measurement of intracellular pH

Intracellular pH measurement with a ratiometric fluorescent pH indicator protein, pHluorin (Miesenböck et al., 1998), was carried out at an external pH of 7.0 as described before (Nakamura et al., 2009b; Morimoto et al., 2010). The fluorescence-excitation spectra of MMPH001 (ΔaraBAD::pHluorin) expressing MotA were recorded by a fluorescence spectrophotometer (RF-5300PC, Shimadzu).

Bead assays

Salmonella MM3076iC cells harbouring pKK223-3 or pKSS12 were grown in LB containing ampicillin for 4 h at 30°C with shaking. Bead assays with polystyrene beads with diameters of 1.0 and 0.8 µm (Invitrogen) or with fluorescent beads with diameters of 0.5 µm (FluoSpheres®‘yellow/green’, excitation 470 nm, emission 518 nm; Invitrogen) and 0.1 µm (FluoSpheres®‘orange’, excitation 540 nm, emission 560 nm; Invitrogen) were carried out as described before (Che et al., 2008). In the measurements using 1.0 µm and 0.8 µm beads, phase-contrast images were projected onto a quadrant photodiode and the signals were recorded at 1 ms intervals. In the measurements using 0.5 µm and 0.1 µm fluorescent beads, fluorescence images were captured by an EMCCD camera (iXon, Andor) every 0.4 ms. Torque calculation was carried out as described previously (Che et al., 2008; Nakamura et al., 2009b). To produce speed histograms, the rotation rate of each cell was sampled at 1 kHz for 10 s and the average speed was determined from a power spectrum using 1 s long data windows (1024 points) at 0.1 s intervals (speed resolution was about 1 Hz) as described by Reid et al. (2006).

Acknowledgements

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

We thank May K. Macnab, M. Homma and S. Kojima for critical reading of the manuscript and helpful comments, S. Sugiyama for his kind gift of pKSS12, H. Matsunami for his kind gift of pHMK1602 and C.J. Lo for technical advice. S.N. and Y.V.M. are research fellows of the Japan Society for the Promotion of Science. This work has been supported in part by Grants-in-Aid for Scientific Research from the Ministry of Education, Culture, Sports, Science and Technology of Japan to K.N. and T.M.

References

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information

Supporting Information

  1. Top of page
  2. Summary
  3. Introduction
  4. Results
  5. Discussion
  6. Experimental procedures
  7. Acknowledgements
  8. References
  9. Supporting Information
FilenameFormatSizeDescription
MMI_7391_sm_FigS1-2.pdf457KSupporting info item
MMI_7391_sm_MovieS1.mov272KSupporting info item
MMI_7391_sm_MovieS2.mov263KSupporting info item
MMI_7391_sm_MovieS3.mov236KSupporting info item
MMI_7391_sm_MovieS4.mov263KSupporting info item
MMI_7391_sm_MovieS5.mov110KSupporting info item
MMI_7391_sm_MovieS6.mov81KSupporting info item
MMI_7391_sm_MovieS7.mov96KSupporting info item
MMI_7391_sm_MovieS8.mov964KSupporting info item
MMI_7391_sm_MovieS9.mov919KSupporting info item

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.