Domain-based biophysical characterization of the structural and thermal stability of FliG, an essential rotor component of the Na⁺-driven flagellar motor

Abstract

Many bacteria move using their flagellar motor, which generates torque through the interaction between the stator and rotor. The most important component of the rotor for torque generation is FliG. FliG consists of three domains: FliG_N, FliG_M, and FliG_C. FliG_C contains a site(s) that interacts with the stator. In this study, we examined the physical properties of three FliG constructs, FliG_Full, FliG_MC, and FliG_C, derived from sodium-driven polar flagella of marine Vibrio. Size exclusion chromatography revealed that FliG changes conformational states under two different pH conditions. Circular dichroism spectroscopy also revealed that the contents of α-helices in FliG slightly changed under these pH conditions. Furthermore, we examined the thermal stability of the FliG constructs using differential scanning calorimetry. Based on the results, we speculate that each domain of FliG denatures independently. This study provides basic information on the biophysical characteristics of FliG, a component of the flagellar motor.

The flagellum is an organelle for locomotion and is composed of a filament, hook, and basal body. The basal body, which contains the rotor, comprises several ring structures, the L, P, MS, and C rings. The stator, which is an ion-conducting energy-converting complex, surrounds the rotor and generates torque through interaction with components of the rotor [1,2]. The flagellar motor uses mainly two kinds of ions, H⁺ or Na⁺, as the energy source and the stator is composed of two kinds of membrane proteins. Escherichia coli and Salmonella enterica have proton driven motors and the stator is composed of the MotA and MotB proteins. On the other hand, Vibrio has Na⁺ driven motors and the stator is composed of the proteins PomA and PomB [3,4]. MotA and PomA have four transmembrane regions, whereas MotB and PomB have only one. MotA and MotB, and PomA and PomB form a complex with a stoichiometry of A4:B2 [5,6]. It has been suggested that MotB and PomB undergo a dynamic conformational change in order to bind to the peptidoglycan layer [7,8].

To reach a favorable condition, bacteria have a chemotactic system to regulate the rotational direction of the flagellar motor. In the flagellar motor, the C ring plays a role in the directional change, counterclockwise (CCW) and clockwise (CW), thus it is also called the switch complex. In E. coli, which has peritrichous flagella, the CCW rotation of the flagella leads to the forward pushing of the bacterial cell body, leading to a smooth swimming motion. On the other hand, the CW rotation leads to the unwinding of the flagellar filament bundle and the bacteria tumble [9]. The C ring is composed of three proteins, FliG, FliM, and FliN, which are cytoplasmic, not transmembrane proteins [10]. FliM is divided into three domains, FliM_N, FliM_M, and FliM_C. FliG is also divided into three domains, FliG_N, FliG_M, and FliG_C. FliG_N interacts with FliF, which is the MS ring component. FliG_M and FliG_C interact with FliM [11,12]. FliM_N interacts with CheY-P, the phosphorylated form of CheY, which is a response regulator [13,14]. When CheY-P binds FliM_N, the flagella rotate CW. If the concentration of CheY-P decreases, the flagella rotate CCW. The crystal structures of the C ring component proteins, FliG, FliM, and FliN, have been revealed. The structural difference of FliG between the CW and the CCW state has been proposed from the crystal structures of FliG [15–19]. Recently, in the C ring structure it was shown for the FliM_M:FliG_M heterodimer and it was also detected that higher-order assemblies include a parallel back-to-front arrangement of the FliM_M units [20]. Furthermore, NMR analysis has been reported for Thermotoga maritima [12,21], and the HSQC spectra obtained revealed the interaction between FliF or FliM and FliG.

The most important rotor component for torque generation is FliG, which interacts with the stator protein, PomA or MotA [22]. It has been shown that the electrostatic interactions between some conserved charged residues of the cytoplasmic loop of MotA (MotA_loop) and the C-terminal domain of FliG (FliG_C) are important for torque generation of the H⁺-driven flagellar motor of E. coli [1]. On the other hand, the Na⁺-driven flagellar motor of marine Vibrio contains such conserved charged residues in PomA and FliG, but single mutations of these conserved residues do not strongly affect the motility [23,24]. Compared with E. coli, the number of conserved charged residues of the flagellar motor in Vibrio is greater, and the contribution of each residue for torque generation may be smaller in Vibrio motor [2,25]. Furthermore, we found that a specific interaction between the charged residues is critical for the correct assembly of the stators around the rotor and is important for torque generation [2].

It has been reported that there are other important residues or motifs of FliG for motility in addition to the charged residues. For example, in E. coli and Helicobacter pylori, the motif of FliG_C containing the MFXF sequence is thought to be important for the motor switch [19]. In Vibrio alginolyticus, the three-residue deletion mutant, ΔPSA, which corresponds to ΔPEV in T. maritima, showed loss of motility [26]. In addition to this mutant, three other Mot⁻ mutants, L259Q, L270R, and L271P of FliG, were reported in V. alginolyticus [27]. We have characterized the physical properties of the C-terminal domain (G214–L351) of wild-type FliG and its non-motile phenotype mutant derivatives [28]. The CD spectra and size exclusion chromatography did not show a significant difference between the wild-type and mutant FliG proteins, however, the DSC data were very different between the mutants. This possibly means that the secondary structure is not affected by the mutation but the tertiary structure is.

In this study, we made new constructs of plasmids containing the fliG gene and overexpressed FliG of marine Vibrio, FliG_MC, and full length FliG in addition to the C-terminal domain of FliG (FliG_C). We then characterized the physical properties of FliG using these truncated Vibrio FliG proteins.

Materials and Methods

Strains and plasmids

The strains and plasmids used in this study are shown in Table 1. Routine DNA manipulations were carried out according to standard procedures using the E. coli strain DH5α. E. coli strain BL21 was used for protein expression.

Table 1.

Strains and plasmids used in this study

Strain or plasmid	Description	Source or reference
Escherichia coli
BL21	Host for overexpression from the T7 promoter
DH5a	Recipient for cloning experiments
Plasmid
pColdI	Cold shock expression vector	Takara
pCold-FliGC	pColdI/FliG (G214–L351)	[28]
pCold-FliGMC	pColdI/FliG (G122–L351)	This study
pRAY201	pColdI/FliG (Full)	[31]

Protein expression in E. coli and purification

The expression plasmids were introduced into E. coli strain BL21. Cells were grown and induced as described previously [28]. Cells were harvested by centrifugation, suspended in TN buffer [50 mM Tris-HCl (pH 8.0), 0.5 M NaCl] containing 20 mM imidazole and protease inhibitors, and were disrupted by sonication (LV 5.0, duty cycle 50%, 1 min×5). After the suspension was centrifuged at 22,000×g for 15 min, the supernatant was ultracentrifuged at 100,000×g for 30 min. The supernatant of the soluble fraction was subjected to affinity column chromatography using the His Trap HP 5 ml (Ni-NTA) column (GE Healthcare). Proteins were eluted with an imidazole concentration gradient (from 20 mM to 500 mM) and collected by 1 ml fraction. The collected fractions were subjected to SDS-PAGE and Coomassie brilliant blue (CBB) staining to examine the purity of the proteins.

SDS-PAGE

Samples for SDS-PAGE were mixed with the SDS sample buffer and boiled at 95°C for 10 min. SDS-PAGE was performed using a 12% polyacrylamide gel.

Size exclusion chromatography

The sample was subjected to the gel filtration column using a Superdex 200 10/300 (GE healthcare). Size exclusion chromatography was performed using either 50 mM sodium phosphate (pH 6.5) or 50 mM Tris-HCl (pH 8.0) buffer at a flow rate of 0.5 ml/min, and samples were fractionated by 1 ml. The molecular weight was estimated using the marker proteins, ribonuclease A, carbonic anhydorase, ovalbumin, conalbumin, aldolase, ferritin, and thyroglobulin (GE healthcare).

Circular dichroism spectroscopy

Circular dichroism spectroscopy was measured using a J-720W instrument (Jasco) as described previously [28]. All sample concentrations were 0.1 mg/ml and buffer conditions were 50 mM sodium phosphate (pH 6.5 or pH 8.0). The spectra obtained were analyzed using Secondary structure prediction program (JASCO). The Yang equation was used as a reference to obtain the secondary structure ratio.

Differential scanning calorimetry (DSC)

The purified FliG constructs were inserted using a syringe into a Slide-A-Lyzer (PIERCE) and dialyzed against a 50 mM phosphate buffer (pH 6.5). The dialyzing buffer was replaced after 3 hours and further dialyzed overnight at 4°C. The dialyzed sample (650 μl of 1 mg/ml sample) was applied into a nano-DSC differential scanning calorimeter (TA Instruments) and the heat capacity and thermal stability were measured with a scan speed of 1°C/min from 15 to 90°C against the reference buffer of 50 mM phosphate (pH 6.5).

Results

Purification profiles of FliG

Recently we reported on the biophysical characteristics of the FliG C-terminal domain, FliG_C, from V. alginolyticus [28]. Using the same analytical techniques, we characterized two other constructs, FliG full length (FliG_Full), and FliG middle and C-terminal domain (FliG_MC), in addition to FliG_C (Fig. 1). We expressed these constructs in E. coli and purified them. As shown in the SDS-PAGE CBB staining results in Figure 2A, all constructs were expressed well and were purified as a single band. However, the expression levels of FliG_Full and FliG_MC were higher than that of FliG_C. Additional bands were detected in purified FliG_C, suggesting that FliG_C is digested during the purification step. FliG_C seems to be more unstable than FliG_Full and FliG_MC.

Figure 1.

(A): The crystal structure of full length FliG in Aquifex aeolicus. The N-terminal domain, middle domain, and C-terminal domain are colored blue, green, and red, respectively. G122 corresponds to the predicted N-terminal residue of FliG middle domain in Vibrio alginolyticus. G214 corresponds to the predicted N-terminal residue of FliG C-terminal domain in V. alginolyticus. (B): The three constructs of FliG used in this study. G122-FliG_MC corresponds to the middle and C-terminal domains and G214-FliG_C corresponds to the same C-terminal domain as (A).

Figure 2.

Purification profiles of FliG constructs. (A): CBB staining results. Pre indicates pre induction whole cell sample, post indicates post induction whole cell sample, and purify indicates the purification product sample. M: molecular weight markers. (B): The profiles of size exclusion chromatography. The arrows show the peaks derived from the purification product. The elution volume (ml) of the peak and the estimated molecular weight (kDa) are shown in each profile. Left, results under 50 mM sodium phosphate buffer at pH 6.5. Right, results under 50 mM Tris-HCl buffer at pH 8.0. Full: full length FliG_Full, MC: FliG_MC and C: FliG_C. ×2 and ×4 in brackets mean the multiplication of the vertical scales. (C): Calibration curves at pH 6.5 and at pH 8.0 were drawn using the molecular weight markers.

Size exclusion chromatography of FliG

We analyzed all the FliG constructs using size exclusion chromatography under two conditions: 50 mM Tris-HCl (pH 8.0) buffer and 50 mM sodium phosphate buffer (pH 6.5) (Fig. 2B, C). Under both conditions, all constructs were eluted as a single peak, but the peak position was dependent on the buffer condition. In the Tris-HCl buffer at pH 8.0, the estimated molecular weights of FliG_Full, FliG_MC, and FliG_C were ca. 155 kDa, ca. 87 kDa, and ca. 50 kDa, respectively. In the sodium phosphate buffer at pH 6.5, the estimated molecular weights of FliG_Full, FliG_MC, and FliG_C were ca. 58 kDa, ca. 53 kDa, and ca. 24 kDa, respectively. The estimated molecular weights under conditions of higher pH seem to be higher than those under lower pH. In particular, the estimated molecular weight of FliG_Full was dramatically different between the two conditions. These results suggest that the conformation of FliG, especially that of FliG_Full, was changed between the two buffer conditions.

Circular dichroism spectroscopy of FliG

The results of size exclusion chromatography suggest that the conformation of all FliG constructs changes depending on the buffer conditions. Next, we tried detecting the possibility of secondary structural changes of FliG using circular dichroism (CD) spectroscopy (Fig. 3). The CD spectra of all FliG constructs under both pH conditions show the negative peak around 207 nm and the negative shoulder around 222 nm (Fig. 3). These spectra are typical of α-helical conformation, which is supported by the fact that the crystal structures of FliG from various species are mainly composed of α-helices (Fig. 1). Comparing the spectra at pH 6.5 with those at pH 8.0, the negative peak and shoulder at pH 6.5 are deeper than those at pH 8.0 for all FliG constructs. Therefore, this shows that the structures of FliG at pH 6.5 have more α-helices than those at pH 8.0. Comparing the spectra among the FliG constructs, the spectra of FliG_MC show the most negative peak and shoulder under both pH conditions (Fig. 3D). Moreover, FliG_Full showed a more negative peak and shoulder than FliG_C. We estimated the content of α-helices in FliG_Full and FliG_MC. There was little difference under both pH conditions and the contents at pH 6.5 were higher than those at pH 8.5 (Fig. 3E).

Figure 3.

CD spectra of FliG constructs. (A), (B), and (C): The CD spectra data comparing pH 6.5 to pH 8.0. Solid lines show the results at pH 6.5 and dotted lines show the results at pH 8.0. (A): full length FliG; FliG_Full, (B): middle and C-terminal domain construct of FliG; FliG_MC and (C): C-terminal domain construct of FliG; FliG_C, respectively. (D): Data comparing CD spectra of all constructs at pH 6.5. Solid line: FliG_Full; dotted line: FliG_MC; and gray line: FliG_C. All sample concentrations were 0.1 mg/ml and buffer conditions were 50 mM sodium phosphate buffer pH 6.5 or pH 8.0. (E): The structural compositions are estimated using software from JASCO and the percentages of the structures for α-helix are indicated.

Differential scanning calorimetry of FliG

We used differential scanning calorimetry (DSC) to investigate the thermal stability of the FliG constructs (Fig. 4). We measured the FliG constructs at pH 6.5, in which they showed more negative molar ellipticities. FliG_C showed a clear single peak, suggesting that the thermal denaturation of FliG_C is a one-step process. Fitting this peak by a two state model, the T_m value is ca. 59°C and the corrected Van’t Hoff enthalpy, ΔH, is 208 kJ/mol, which are similar to our previous report [28]. In contrast, the DSC curves of FliG_MC and FliG_Full can be fitted with two and three peaks, respectively. The T_m values are ca. 42°C and ca. 60°C, and the estimated ΔH values are 197 kJ/mol and 95 kJ/mol for the first and second peaks of FliG_MC, respectively. The T_m values are ca. 53°C, ca. 59°C, and ca. 67°C, and the estimated ΔH values are 85 kJ/mol, 162 kJ/mol, and 53 kJ/mol for the first, second, and third peaks of FliG_Full, respectively. Importantly, the number of peaks is consistent with the number of domains in the FliG constructs. Therefore, our results imply that each domain of FliG thermally denatures independently and the thermal stability is affected by the interaction of the other domain or the linker between domains.

Figure 4.

DSC data of FliG constructs under 50 mM sodium phosphate buffer at pH 6.5. The scan range was 15 to 90°C and the scan rate was 1 degree/min. The vertical and horizontal axes show molar heat capacity and temperature, respectively. (A) The profiles of G214-FliG_C: C. The thin line and the solid line show the raw data and the results fitted by the two state scaled model, respectively. (B) The profiles of G122-FliG_MC: MC. The smooth lines show the results fitted by the two state scaled model. The solid line shows the result of the model sum. The line shows raw data. (C) The profiles of full length FliG_Full: Full. The thin smooth lines show the results fitted by the two state scaled model. The solid line shows the result of model sum. The line shows raw data. (D) The fitting data of FliG constructs. The solid lines show the result of model sum for full length FliG_Full, G122-FliG_MC, and G214-FliG_C.

Discussion

In this study, we investigated the biophysical characteristics of FliG by comparing FliG_Full with its fragments, FliG_MC and FliG_C. The results of size exclusion chromatography imply the conformational change of FliG between the two pH conditions (Fig. 2B, C). The estimated molecular weight of the constructs by size exclusion chromatography was larger than that calculated using their amino acid sequences. We cannot conclude whether this might be due to the formation of higher order complexes (dimer or trimer) or not. A recent study found that the conformation of FliG can be changed between an extended and compact form depending on the NaCl concentration [29]. Furthermore, it has been demonstrated that FliG exists as a monomer in solution but as a domain-swapped polymer in the flagellar motor [29].

FliG_Full showed the largest change in the apparent molecular weight between pH 6.5 and 8.0 among the constructs (Fig. 2). This implies that FliG_Full at pH 8.0 exists primarily in an extended form compared with that at pH 6.5, where it exists primarily in a compact form. FliG_MC and FliG_C showed similar profiles at both pH 6.5 and pH 8.0. The theoretical pIs of FliG_Full, FliG_MC, and FliG_C are 4.5, 4.4, and 4.3, respectively. Thus, we propose that the electrostatic interactions contribute to the pH dependent conformational changes in FliG. Simply speaking, there are more electrostatic charges under pH 8.0 since it is farther from the theoretical pIs of the constructs than pH 6.5, promoting FliG to form its extended conformation. This is consistent with the recent study that found that FliG exists in its extended form under lower NaCl concentrations but under higher NaCl concentrations in its compact form [29]. Usually, the high concentration of salt prevents the effects of electrostatic charge.

In the CD spectroscopic analysis, the constructs showed differences in their negative peaks between the two pH conditions (Fig. 3). All spectra at pH 6.5 showed higher negative peaks than those at pH 8.0. These results suggest that the structures under pH 6.5 seem to contain more secondary structures than those under pH 8.0. Furthermore, in Salmonella, it was reported that the CD spectra of FliG_C does not change under broad temperature conditions (15–45°C) [30], suggesting that temperature may not severely affect the structure of FliG compared to pH. As discussed above, at pH 6.5 FliG exists in its compact form, leading to the conclusion that the compact form of FliG has more secondary structures and the electrostatic interactions may have an important role for this conformation of FliG.

We measured the thermal stability of FliG constructs using DSC. In FliG_C, one peak appeared in the profile and the T_m value was ca. 60°C (Fig. 4A). This value is reasonable because CD spectroscopic analysis revealed that FliG_C does not change its secondary structure until 45°C [30]. In the case of FliG_Full and FliG_MC, there were several peaks in the profiles but the number of peaks corresponded with the number of domains of the constructs (Fig. 4B, C). Therefore, we propose that each domain of FliG independently denatures. Moreover, it is difficult to assign each domain a T_m value because there are various T_m values for the FliG constructs. This indicates that the interactions between domains also have some roles in thermal stability.

In summary, this study may support the recent report showing that in solution, FliG from E. coli has two conformations: an extended form and a compact form [29]. Here, using FliG from Vibrio we demonstrated that these conformations might be regulated by controlling pH conditions, suggesting an important role of electrostatic charges in the structural conformation of FliG. These findings implicate that the two conformations might be related to the state of the two rotational directions, CCW and CW, and these conformations are conserved among bacterial species. In the bacterial cell, it has been suggested that FliG forms a domain-swapped polymer when it assembles in the flagellar motor [20,29]. Therefore, it is essential to explore the mechanism of assembly for forming ring-like structures in the flagellar motor and the biophysical properties of FliG clarified in this study will assist to clarify the mechanism.

Significance.

Bacteria can swim by rotating their flagella. Interaction between the stator protein and the rotor protein makes the rotational driving force. FliG is the most important component of the rotor for the torque generation in the flagellar motor. Here, we purified FliG and its fragmants from Vibrio alginolyticus, which has a polar flagellum to characterize the biochemical/biophysical properties by gel-filtration chromatography, circular dichroism spectroscopy, and differential scanning calorimetry. The results suggest that FliG has at least two conformational states and that each domain of FliG is responsible for the denaturation. These provide information on the basic characteristics of FliG.

Abbreviations

PomA

Polar flagellar motility protein A

PomB

Polar flagellar motility protein B

SDS-PAGE

sodium dodecyl sulfate-polyacrylamide gel electrophoresis

CD

circular dichroism