1. Microbiology and Infectious Disease
Download icon

The flagellar motor of Vibrio alginolyticus undergoes major structural remodeling during rotational switching

  1. Brittany L Carroll
  2. Tatsuro Nishikino
  3. Wangbiao Guo
  4. Shiwei Zhu
  5. Seiji Kojima
  6. Michio Homma  Is a corresponding author
  7. Jun Liu  Is a corresponding author
  1. Department of Microbial Pathogenesis, Yale School of Medicine, United States
  2. Microbial Sciences Institute, Yale University, United States
  3. Division of Biological Science, Graduate School of Science, Nagoya University, Furo-cho, Chikusa-ku, Japan
Research Article
  • Cited 0
  • Views 578
  • Annotations
Cite this article as: eLife 2020;9:e61446 doi: 10.7554/eLife.61446

Abstract

The bacterial flagellar motor switches rotational direction between counterclockwise (CCW) and clockwise (CW) to direct the migration of the cell. The cytoplasmic ring (C-ring) of the motor, which is composed of FliG, FliM, and FliN, is known for controlling the rotational sense of the flagellum. However, the mechanism underlying rotational switching remains elusive. Here, we deployed cryo-electron tomography to visualize the C-ring in two rotational biased mutants in Vibrio alginolyticus. We determined the C-ring molecular architectures, providing novel insights into the mechanism of rotational switching. We report that the C-ring maintained 34-fold symmetry in both rotational senses, and the protein composition remained constant. The two structures show FliG conformational changes elicit a large conformational rearrangement of the rotor complex that coincides with rotational switching of the flagellum. FliM and FliN form a stable spiral-shaped base of the C-ring, likely stabilizing the C-ring during the conformational remodeling.

Introduction

Many bacteria navigate complex environments by controlling the flagellar rotational switch between counterclockwise (CCW) and clockwise (CW). Escherichia coli and Salmonella enterica (henceforth, Salmonella) use a ‘run-and-tumble’ approach for controlling movement, by which flagella rotating in a CCW sense drive the cell body forward, and when the rotation sense switches to CW the bacterium tumbles through the medium to change direction (Berg, 2003; Chevance and Hughes, 2008; Terashima et al., 2008). Vibrio alginolyticus has a unique three-step swimming pattern with forward, reverse, and flick motions; where CCW rotation propels the cell body forward, CW rotation drives the bacterium in reverse, and a flicking motion occurs upon CW-CCW rotation to change swimming direction, analogous to the tumble (Xie et al., 2011).

The motor is the most intricate part of the flagellum, not only responsible for flagellar assembly and rotation but also essential for rotational switching. Spanning from the cytosol through the outer membrane, the motor consists of a series of rings, with the L-ring at the outer membrane, the P-ring located within the periplasmic space, the MS-ring embedded in the cytoplasmic membrane, and the C-ring inside the cytoplasm (Homma et al., 1987; Francis et al., 1992; Ueno et al., 1992; Francis et al., 1994). The stator subunits, embedded in the cytoplasmic membrane, generate torque via ion flow to rotate the C-ring (Blair, 2003; Sato and Homma, 2000a). Different from E. coli and Salmonella, V. alginolyticus possesses several Vibrio-specific features: H-, T-, and Orings. The O-ring is located on the outside of the outer membrane (Zhu et al., 2017), the H-ring facilitates the outer membrane penetration of the flagellum (Terashima et al., 2010; Zhu et al., 2018), and the T-ring contacts the H-ring and stators, presumably acting as a scaffold to hold the stators (Terashima et al., 2006; Zhu et al., 2019).

Flagellar rotation is powered by an electrochemical gradient across the cell membrane,driving ion flow through the stator complex (Berg, 2003; Terashima et al., 2008; Kojima and Blair, 2004; Li et al., 2011). In Salmonella and E. coli, H+ ions are conducted through the stator units. In Vibrio species, Na+ ions are conducted through the stator units. The prevailing idea is that MotA and MotB (in the H+ motor) or PomA and PomB (in the Na+ motor) form a membrane-bound stator subunit (Sato and Homma, 2000a; Kojima and Blair, 2004; Braun et al., 2004; Sato and Homma, 2000b). Recent cryo-EM studies have shown that MotA and MotB assemble in 5:2 stoichiometry (Santiveri et al., 2020; Deme et al., 2020). The A and B subunits have four and one transmembrane helices, respectively, and two helices from the A subunit and one from the B subunit form an ion channel that contains an essential ion-binding aspartyl residue (Hosking et al., 2006). Inactive stator complexes diffuse through the cytoplasmic membrane and interact with the rotor to generate a series of conformational changes within PomA and PomB that open the ion channel and facilitate binding to the peptidoglycan layer (Hosking et al., 2006; Mino et al., 2019; Zhu et al., 2014; Fukuoka et al., 2009; Sudo et al., 2009; Kojima et al., 2018). However, exactly how the stator assembles around the motor and interacts with the C-ring is still not well understood.

The C-ring is essential for flagellar rotation and rotational switching. The C-ring structure is conserved among diverse species, with repeating subunits consisting of FliG, FliM, and FliN, with FliN sometimes being supplemented or replaced by FliY (Kojima and Blair, 2004; Zhao et al., 1995). FliG, located closest to the cell membrane (henceforth referred to upper portion) of the C-ring, interacts with the MS-ring, the lower C-ring, and the stator via three domains (Lee et al., 2010). The interaction of the N-terminal domain of FliG (FliGN) with FliF tethers the C-ring to the MS-ring and is necessary both for assembly and rotation of the flagellum (Ogawa et al., 2015; Lynch et al., 2017; Xue et al., 2018). The middle domain of FliG (FliGM) interacts with FliM, holding the upper, membrane-proximal portion of the C-ring in contact with its lower, membrane-distal portion (Brown et al., 2002; Brown et al., 2007; Minamino et al., 2011). Lastly, the C-terminal domain of FliG (FliGC) interacts with a cytoplasmic loop of PomA (or MotA in proton-driven motors) via interactions of oppositely charged residues, connecting the stator and rotor (Lloyd and Blair, 1997; Yakushi et al., 2006; Takekawa et al., 2014). FliM also has three domains (Park et al., 2006). The N-terminal domain of FliM (FliMN) binds to the chemotaxis signaling protein phosphoryl CheY (CheY-P) to trigger switching from CCW to CW rotation (Paul et al., 2011; Vartanian et al., 2012). The FliMM serves as a connection between the base of the C-ring and FliG (Brown et al., 2002; Minamino et al., 2011). FliMC forms a heterodimer with FliN when fused via a flexible linker (Notti et al., 2015; Dos Santos et al., 2018). The third protein, FliN, a small single-domain protein, dimerizes with FliM or itself to form the base of the C-ring (Brown et al., 2005). Proposed models for FliG, FliM, and FliN assembly include 1:1:4 (Sarkar et al., 2010a; Sarkar et al., 2010b) or a 1:1:3 (McDowell et al., 2016) stoichiometry. Two FliN homodimers form a ring at the base of each C-ring subunit in the first model (Sarkar et al., 2010a), while a FliM:FliN heterodimer and a FliN homodimer create a spiral base of the C-ring in the second model (McDowell et al., 2016).

It has been proposed that FliG undergoes a dramatic conformational change from an open, extended form during CCW rotation to a closed, compact form during CW rotation (Lee et al., 2010). Two α-helices play an important role in determining the FliG conformation: helixMN connects FliGN and FliGM, and helixMC connects FliGM and FliGC. It was suggested that these helices are rigid and extended in the open CCW conformation, while they become disordered to switch into the compact CW conformation. Each domain contains armadillo (ARM) repeat motifs; ARMN interacts with the adjacent FliG monomer, and ARMM and ARMC interact either inter- or intramolecularly, depending upon the CCW or CW rotational sense, respectively (Lee et al., 2010; Brown et al., 2002; Minamino et al., 2011). The domain-swapping mechanism was proposed to coordinate with the conformational change in helixMC (Lee et al., 2010). Another biochemical study suggested that the domain swapping occurs during C-ring assembly, with FliG in solution existing as a monomer, and an equilibrium favoring FliG oligomers in the C-ring (Baker et al., 2016).

The FliG conformational change occurs about a hinge region first predicted by in silico modeling and characterized via mutational analysis (Van Way et al., 2004). Two recently characterized fliG variants in V. alginolyticus (G214S and G215A), previously characterized in E. coli (G194S and G195A [Van Way et al., 2004]), appear to hinder the conformational change sterically (Nishikino et al., 2018; Nishikino et al., 2016), resulting in motors that rotate primarily in a single direction. The wildtype flagella rotate in a CW:CCW ratio of 1:3, while the fliG-G214S mutant produces a CCW-biased phenotype (CW:CCW 1:9), and fliG-G215A mutant produces a CW-locked phenotype (Nishikino et al., 2016). These adjacent residue substitutions create opposite motility phenotypes and are located in the Gly-Gly flexible linker, a hinge region whose conformation is believed to be dependent upon helixMC (Nishikino et al., 2016).

To characterize C-ring dynamics during switching, we used cryo-electron tomography (cryo-ET) to determine in situ motor structures of the V. alginolyticus CCW-biased mutant G214S and the CW-locked mutant G215A. We found that a large conformational change of the C-ring occurs between the CCW- and CW-motors. Docking the previously solved homologous structures into our cryo-ET maps, we built two models of the C-ring in CCW and CW rotation.

Results

Visualization of intact flagellar motors in CCW and CW rotation states

To address the mechanism of rotational switching, we utilized cryo-ET to visualize the C-rings of two FliG mutants (G214S and G215A) of V. alginolyticus. These mutants were previously shown to have CCW-biased and CW-locked flagellar motors, respectively (Nishikino et al., 2018). Each V. alginolyticus cell usually possesses a single, polar, sheathed flagellum, which limits the amount of information per cell for in situ motor structure determination. We therefore introduced the fliG-G214S and fliG-G215A mutations (Table 1) into the flhG KK148 strain, which produces multiple flagella at the cell pole (Kusumoto et al., 2006). Tomograms were collected using a 300kV Titan Krios electron microscope (Thermo Fisher) with Volta phase plate, a GIF energy filter and a post-GIF K2 detector electron detector (Gatan). Multiple flagella can be readily seen in a typical cryo-ET reconstruction (Figure 1A,D). In total, 2221 CCW-biased motors from G214S cells and 1618 CW-locked motors from G215A cells were used to determine in situ flagellar motor structures (Figure 1B,E and Table 1). Due to increased contrast from the Volta phase plate and focused classification, we observed 34-fold symmetry of the C-ring in both variants (Figure 1C,F and Figure 1—figure supplement 1). To begin to address potential differences in the switch complex, we measured the diameter of the C-ring in Fiji ImageJ (Schindelin et al., 2012). To describe the observed changes, we will refer to the top of the C-ring as the outermost region facing the cytoplasmic membrane. Starting from the top portion in the fliG G214S (CCW) the diameter is 46.2 nm, the middle 46.6 nm, and the bottom 49 nm (Figure 1C,I). In fliG G215A (CW) motor the diameters are 49.0 nm, 46.6 nm, and 48.3 nm, respectively (Figure 1F,I). Thus, the CW-motor is about 20.8 Å wider than the CCW-motor at the top of the C-ring, even though the number of C-ring subunits remains unchanged (Figure 1). To understand the interaction between the C-ring and the stator, we also analyzed the densities on the top of the C-ring. We were able to resolve the stator ring in the CW-locked motor but not in the CCW-biased motor (Figure 1G,H). Interestingly, the diameter of the stator ring is 48.3 nm, falling in between the CCW-biased motor and CW-locked motor diameters (Figure 1H).

Figure 1 with 1 supplement see all
In situ structures of V. alginolyticus flagellum motor in CCW and CW rotation.

(A) A representative tomogram slice showing two motors in V. alginolyticus fliG G214S mutant (white arrow) Scale bar 100 nm. (B) A medial cross-section of the in situ flagellar motor structure. Scale bar 25 nm. (C) A perpendicular cross-section of the motor showing the top portion of the C-ring (blue arrow in B). (D) A representative tomogram slice of the CW-motor in V. alginolyticus fliG G215A CW-locked variant. (white arrow) Scale bar 100 nm. (E) A medial cross-section of the in situ flagellar motor structure. Scale bar 25 nm. (F) A perpendicular cross-section of the motor showing the top portion of the C-ring (purple arrow in E). (G) A medial cross-section of the flagellum motor with density for the stator (orange arrow). (H) A perpendicular cross-section of the motor showing the stator ring (orange arrow in G). (I) Diameters of the C-ring measured at the top, middle, and bottom from perpendicular cross-sections for both the CCW- and CW-motors. Abbreviations: outer membrane (OM), cytoplasmic membrane (CM).

Table 1
Cryo-ET data used in this study.
Strain genotypeTomogramsMotors
fliG-G214S2592221
fliG-G215A2201618

To increase the resolution of the C-ring, we refined the C-ring structures with 34-fold symmetry applied to approximately 18 Å for the CCW and 19 Å for CW rotations (Figure 2A,D). The increased resolution allows us to reveal two unique conformations of the C-ring (Figure 2). We observed a significant lateral shift upon switching from CCW to CW, while the overall diameter of the C-ring varies slightly between the two mutants. These data demonstrate that the C-ring undergoes a large lateral conformational change during flagellar switching that leads to a difference in FliG presentation evident from the diameter change at the top of the C-ring. The protein composition of the C-ring likely remains the same as both CCW- and CW-motors have 34-fold symmetry, and the middle and bottom regions of the C-ring are similar in diameter.

Focused refinement of the C-ring reveals differences in the CCW- and CW-motors.

(A) A medial cross-section of the focused refinement of the C-ring in V. alginolyticus fliG G214S CCW-biased variant. Scale bar 25 nm. (B) A side view of the CCW C-ring. Scale bar 10 nm. (C) A side view of the segmented map shows the top (blue), middle (light blue), and bottom (green) regions. (D) A medial cross-section of the C-ring in fliG G215A CW-locked variant. Scale bar 25 nm. (E) A side view of the focused refined CW-locked C-ring. Scale bar 10 nm. (F) A side view of the focused refined C-ring further segmented into the top (purple), middle (gray blue) and bottom (green) regions of the C-ring. The asterisk highlights the additional density (white) observed in only the CW-motor that we speculate to be CheY-P. Abbreviations: cytoplasmic membrane (CM), export apparatus and ATPase (EXP).

Molecular architectures of the CCW and CW C-rings

To understand the molecular detail, we built pseudo-atomic models of the C-ring by docking available crystal structures from the C-ring proteins into the in situ maps derived from sub-tomogram averaging. First, we used homologous structures deposited in the PDB and SCWRL4 (Krivov et al., 2009) to map the V. alginolyticus amino acid sequence onto the open (PDB-3HJL [Lee et al., 2010]) and closed (PDB-4FHR [Vartanian et al., 2012]) conformations of the FliG structure. Specifically, we used the full-length extended FliG with crystal packing from PDB-3HJL to build the CCW model. The monomer generated from crystal packing fits into the cryo-ET map surprisingly well. The compact FliG from the PDB-4FHR model was used for the CW model, with the FliGN domain from PDB-3HJL, as there are no other structures of FliGN. Second, we used I-TASSER (Roy et al., 2010; Yang et al., 2015) to generate the structures of FliM and FliN. Third, the models were placed into the cryo-ET map using UCSF ChimeraX (Goddard et al., 2018). The known protein-protein interfaces were preserved during docking and while the unknown protein interface between the FliMN heterodimer and FliN homodimer was refined with the Rosetta protein-protein docking function (Lyskov and Gray, 2008) before being placed into the cryo-ET map. Phenix (Afonine et al., 2013) was used to optimize the placement and to assess the overall fit using rigid body refinement.

We built the C-ring subunit with FliG at the top, FliM in the middle, and three FliN molecules at the base (see Figure 2C,F for clarification), supporting the 1:1:3 (FliG, FliM, and FliN) model in lieu of the 1:1:4 model, as there is no additional density for a fourth FliN molecule. The 1:1:3 model suggests a FliM-FliN heterodimer interacts with the FliN homodimer characterized by mass spectroscopy, as had been modeled into a previously solved cryo-EM map (McDowell et al., 2016; Thomas et al., 2006). To understand the differences in the CCW and CW structures, we first built the CCW model into four adjacent subunit patch. From this patch model single subunit was real space rigid body refined with the CCW subunit having CCmask of 0.69, and CCbox of 0.83 and the CW subunit had a CCmask of 0.66 and CCbox of 0.82 (Figure 3A,C). We also generated maps from the models using ChimeraX molmap at 18 Å for the CCW model and 19 Å for the CW model. The in silico generated maps exhibit features similar to the experimental map (Figure 3—figure supplement 1) and (Figure 3—figure supplement 2).

Figure 3 with 3 supplements see all
Molecular architectures of the C-ring in CCW and CW rotation.

(A) Two views varying by 90o of a single C-ring subunit model fitted in the in situ map of fliG G214S (white). (B) A side view of the CCW-motor with the model shown as surface expanded for symmetry. The density map for the MS-ring, CM, and EXP are shown in gray for reference. FliG is alternating as dark blue, and aqua, FliM is colored light blue, and FliN is shown in green. The charged residues of FliG that interact with the stator are red spheres, and glycine 214 is shown as a yellow sphere. (C) Two views varying by 90o of a single C-ring subunit model in cartoon fitted into the cryo-ET density of fliG G215A (white). (D) A side view of the CW model rendered as surfaces expanded for symmetry. FliG is alternating as purple, and mauve, FliM is gray blue, and FliN is green. The charged residues of FliG that interact with the stator are red spheres, and glycine 215 is shown as a yellow sphere. The alternating coloring of FliG reflects the different crystal structures used to build the model. To build the CCW model, we used the extended FliG, PDB-3HJL, with crystal packing. To build the CW model, we used the compact FliG, in PDB-4FHR. These models support the domain swapping hypothesis. Abbreviations: cytoplasmic membrane (CM), export apparatus and ATPase (EXP).

We expanded our subunit model to the whole C-ring by applying 34-fold symmetry (Figure 3B,D). Importantly, the expanded model fits into the experimental map without clashes, and preserves the architecture of the base observed in the maps. In silico CCW and CW maps from the entire C-ring model were also generated using ChimeraX molmap, showing that the two models yield distinct maps similar to their respective rotational sense (Figure 3—figure supplement 1).

Our pseudo-atomic models can be further substantiated from the biochemical data. The ARMM (Figure 3—figure supplement 2, orange [Park et al., 2006]) and ARMC (Figure 3—figure supplement 2, green) motif orientations have been preserved from the crystal structures. The EHPQR motif (Glu144, His145, Pro146, Gln147, and Arg179, Figure 3—figure supplement 2, lime spheres) of FliG is located in ARMM and oriented to interact with FliMM (Figure 3—figure supplement 2, green sticks). Residues at the FliM-FliM interface had previously been identified in Salmonella and E. coli correspond to Asn56, His63, Asp184, Pro185, and Met187 in V. alginolyticus and they could interact with the neighboring FliM (Figure 3—figure supplement 2, magenta) (Park et al., 2006; Sakai et al., 2019). The N-terminus of FliMM is oriented facing outward toward the additional density of the CW-locked mutant in a position in which it can bind to CheY-P (Figure 3—figure supplement 2, asterisk and D) (Lee et al., 2001a). It is important to note that these residues were used to identify unique location in our model, but due to limited resolution their precise location is not known. Our models of the C-ring are thus consistent with both the previous biochemical data and our cryo-ET maps.

FliM/FliN interactions hold the C-ring subunits together

The tilting observed between the CCW- and CW-motors is centered on FliM, which serves as a structural protein that holds the C-ring subunits together (Figure 3). However, without a full-length crystal structure, the relative orientations of FliMN, FliMM, and FliMC are unknown. The previous crystal structures and solution NMR studies showed FliMM interacting with FliG (Vartanian et al., 2012; Dyer et al., 2009; Lam et al., 2013). FliMC shows sequence homology to FliN, and structural homology to the FliN homodimer, as shown from the structure of a crystallized FliMC-FliN fusion protein (Notti et al., 2015; McDowell et al., 2016). To place FliM into our map, two I-TASSER models, FliMM and FliMC, and place the domains separately into our cryo-ET map, and then connected the linker regions (Figure 3). We lacked sufficient information to confidently place FliMN, and we therefore left it out of our model. The position of FliMN, which binds CheY-P, is flexible (discussed below) (Lee et al., 2001b). Our model shows that FliMM forms the middle portion of the C-ring and allows for the tilting of the C-ring, as FliMC tethers FliM to FliN via the heterodimer. We observed the same continuous helical density at the base of the C-ring as shown in Salmonella (Thomas et al., 2006) and proposed to be formed via alternating FliMC-FliN heterodimers and FliN homodimers (McDowell et al., 2016) connecting adjacent subunits (Figure 4A,B). This region, known as a spiral (McDowell et al., 2016), is uniform within both the CCW and CW cryo-ET maps, and remains largely unaltered upon rotational switching (Figure 4C). We will refer to this as the FliMCFliN3 (MCN3) spiral from here on. The expansion of the C-ring pseudo-atomic model preserves the continuous uniform shape (Figure 4A,B). By building the patch model initially we were able to preserve the similarity of the MCN3 spiral by removing potential bias from the segmentation of a monomer. The hydrophobic patch residues, originally identified in Salmonella (McMurry et al., 2006), that interact with FliH are facing toward the ATPase such that the interaction could occur (Figure 4D). It is important to note that there may be slight movements of the MCN3 proteins during switching, which we would need higher resolution to observe. Taken together, these data suggest that perhaps the MCN3 spiral is important for keeping the C-ring subunits connected during the conformational remodeling and rotation.

FliMC and FliN form a continuous spiral structure at the base of the C-ring.

(A) A bottom view of the fliG G214S (CCW) motor map alone (light blue), and the model (cartoon) expanded for symmetry in the in situ map (gray transparent). (B) A bottom view of the fliG G215A (CW) motor map alone (gray blue), and the model (cartoon) expanded for symmetry in the in situ map (gray transparent). (C) A view from the outside periphery, and (D) a view from the center or inside looking out, are surface representations of six subunits that form a portion of MCN3 spiral from the CCW model, with one FliMC (light blue), and three FliN molecules (green, red, and orange). The hydrophobic patch of FliN (McMurry et al., 2006), L68, A93, V111, V113, Y118, that had been shown to interact of the FliH (cyan) points toward the center of the C-ring.

Presentation of FliGC to the stator in the CCW- and CW-motors

To characterize the observed conformational change, we created a superposition of the CCW and CW C-ring subunits in the CCW- and CW-motors. Using UCSF Chimera (Pettersen et al., 2004) an axis was drawn through the center of the model, FliGMC shown in red, FliGMC and FliMM in green, and FliMM alone in orange. This revealed a 23 Å shift of FliG with a 37o tilt about FliM (Figure 3—figure supplement 3). FliG and FliM tilt 20o outward upon switching from CCW to CW rotation (Figure 3—figure supplement 3). Furthermore, comparison of just the axes shows that the movement characterizing the large conformational change can be attributed to FliMM (Figure 3—figure supplement 3 insets). This suggests that, upon switching, FliG is presumably presented to the stator units very differently as a result of the rearrangement of FliMM.

Extra density around the CW C-ring

We observed an extra ring around the CW-locked C-ring (Figure 2). We speculate that it is formed by CheY-P, as there are several pieces of evidences in line with this model. First, the density is above the noise level in the CW-motors but not in the CCW-motors, which biochemically makes sense as CheY-P binds to CW rotating motors (Paul et al., 2011; Vartanian et al., 2012; Dyer et al., 2009; Lee et al., 2001b). The cells should have endogenous CheY, as there has been no alteration to the CheY gene. Second, density similar in location and shape has been shown two recent studies to be CheY-P. Chang et al. have used GFP-tagged CheY-P to show that, in Borrelia burgdorferi, CheY-P occupies the same position (Chang et al., 2020). Rossmann et al. also identified similar density for the CheY homolog, CleD in Caulobacter crescentus (Rossmann et al., 2020). Third, when we place the crystal structure of CheY into the model, there is about 60 Å gap between the C-ring density and this additional density (Figure 3—figure supplement 2). This distance may be bridged by the 42 residues of FliMN to connect CheY-P to FliMM.

Discussion

The flagellar motor structures we determined provide direct evidence that the C-ring possesses two distinct conformations in CCW and CW rotation. The dynamics of the C-ring appear to be confined to the upper two-thirds of the structure, with the spiral base that connects the C-ring subunits remaining relatively static. We observed a 37o lateral and 20o medial tilting of the C-ring (Figure 3—figure supplement 3), which almost certainly alters its interactions with the stator. In our pseudo-atomic model, the charged residues (Lys284, Arg301, Asp308, and Asp309 in FliGC) that are known to interact with the stator are located at the top of the C-ring. FliGC moves about 40 Å laterally (Video 1). This large remodeling of the C-ring presumably enables distinct interactions between the stator and the C-ring (Video 2). FliGCN contains the ARMC motif that interacts with FliGM and FliMM, and FliGCC contains the charged residues that interact with the stator. A flexible linker attaching the two domains would allow for a range of movement in FliGCC relative to FliGCN and FliGM, thus suggesting that the presentation of the charged residues varies greatly depending upon the conformation of FliG.

Video 1
Focused refinement of the C-ring reveals conformational change during rotational switching.

The video starts with a fliG G214S motor, a mutation leads to a strong preference for CCW rotation, without symmetry applied. The C-ring is highlighted in blue. The video then zooms in to a focused refined C-ring, with much higher resolution, and turns in the CCW (looking down from the flagellum) direction. The pseudo-atomic model appears as cartoon in the map, as the map fades the model turns into surface rendering. Transparent density, that is likely CheY-P, binds and triggers the conformational change of rotor into the CW architecture. The video toggles between CCW and CW pseudo-atomic models in surface rendering to visualize the global changes of the C-ring architecture. It ends with the model spinning in the CW direction. The color scheme is the same as the paper. The video was rendered using ChimeraX.

Video 2
C-ring rearrangement leads to altered FliGC presentation to the stator.

The video starts with a side view of the CCW pseudo-atomic model in surface rendering, and the view rotates so that we are looking down on the C-ring from the flagellum. We can see the charged residues of FliGCC that interact with the stator. The video toggles between CCW and CW models, and asterisks appear to mark the charged residues of every other FliG. This highlights the large change,~40 Å, in FliG upon switching. A morphing of four adjacent subunits shows the dramatic movement of FliGC relative to the stator from the top view, and side view, the CCW model is shown as transparent. The video ends with the CW pseudo-atomic model. The video was rendered using ChimeraX.

Notably, our cryo-ET model shows 34-fold symmetry for both the CCW and CW rotating C-rings, with only a modest diameter change in FliG. It has previously been hypothesized, based on experiments utilizing fluorescently tagged FliM and FliN, that the FliG, FliM, and FliN composition of the C-ring changes during rotational switching (Delalez et al., 2014; Delalez et al., 2010; Hosu and Berg, 2018; Lele et al., 2012). A recent single-particle cryo-EM study suggested that inter-subunit spacing and interactions of the C-ring account for motor switching in Salmonella (Sakai et al., 2019). That study reported a 9 Å decrease in C-ring diameter upon switching to CW. We report identical diameters at the middle of the C-ring in both rotational directions. However, we observed ~21 Å increase in the diameter and ~40 Å lateral change at FliGC of the CW-motor. In a parallel study, Chang et al. used a B. burgdorferi with a CheX deletion or CheY3 deletion to investigate the switching mechanism of spirochetes (Chang et al., 2020). Similar conformational changes of the C-ring were observed in both species; however, the stator complexes were resolved in B. burgdorferi, as they appear to be less dynamic than those in V. alginolyticus. Visualization of the stator complex provided direct evidence that the conformational change of the C-ring alters the rotor-stator interaction. The model of stator-rotor interactions controlling the rotational direction of the motor is further bolstered by the recent high-resolution cryo-EM models of MotAB (Santiveri et al., 2020; Deme et al., 2020). Santiveri et al., 2020 further postulated that MotA rotates CCW in Campylobacter jejunie by showing that the conformations of wildtype MotAB and MotAB trapped in the protonated state are nearly identical, thus ruling out a dramatic conformational change in the stator unit. Therefore, we conclude that the change in rotational direction occurs from a large movement of the charged residues of FliGC relative to the stator.

Our in situ structures support that FliG, FliM, and FliN interact in a 1:1:3 ratio, as suggested for the non-flagellar homolog Spa33 in Shigella (McDowell et al., 2016). This stoichiometry favors the model in which FliM holds the individual C-ring subunits together by interacting with FliGM and forming a heterodimer with FliN. The FliM-FliN spiral connects adjacent C-ring subunits, allowing for dramatic movement without dissociation of the subunits. The FliGM-FliMM interface, perhaps dynamic in its own right, has been suggested to be involved in flagellar switching (Sakai et al., 2019; Dyer et al., 2009; Pandini et al., 2016). It was shown, using NMR, that the CheY-P binding to FliM displaces FliGC, and that the dramatic rearrangement of FliGMC is possible because of the flexibility of FliG (Dyer et al., 2009). In particular, the GGPG loop in FliMM is suggested to be critical for rotation of FliGM relative to FliMM (Pandini et al., 2016). Most recently, point mutations targeting FliMM were shown to restore CCW rotation in a CW-biased mutant (Sakai et al., 2019). These results, combined with our findings, lead us to suggest that the conformational change in FliG results in a rotation about FliMM that leads to the tilting of the C-ring subunit to alter the presentation of the charged residues of FliGC to the stator.

Given that the putative CheY-P-ring is likely associated with the CW C-ring (Figure 1), we propose a CCW-to-CW switching model (Figure 5, Videos 1 and 3). CheY-P binding to the FliM results in a conformational change in FliG and alters the interactions between the C-ring and stator. FliMC and FliN create a spiral base that holds the C-ring together during these dynamic rearrangements. FliMC and FliN may also be involved in relaying the conformational change to adjacent C-ring subunits. By placing the FliG crystal structures of the open and closed variations within our cryo-ET image, we provide additional evidence to support the proposed model in which the ARMM and ARMC domains toggle between inter- and intramolecular interactions (Lee et al., 2010). NMR data suggest that FliGC is the domain that moves and that FliGM remains in stable contact with FliM (Dyer et al., 2009). These changes in the C-ring structure produce the two directions of flagellar rotation. Understanding the interactions of the C-ring with the stator and the MS-ring is essential to elucidate the mechanism of rotational switching and transmission of the rotation to the flagellar filament. We could not resolve the detailed interactions between the cytoplasmic portion of the stator and the C-ring in both rotational directions, and we could not confirm directly that the orientation of FliGCC alters the presentation of its charged residues to the charged residues of PomA. Restricting the movement of FliGCC via truncations of the flexible linker may address the importance of FliGCC mobility. With more data and further classification and refinement of the CCW-biased motor, we expect to explain the FliGCC and stator interactions in great detail. With the rapid development of cryo-ET, it is increasingly possible to reveal motor structure in situ at higher resolution, which will further our understanding of the flagellar assembly and function.

A model for the rotational switching.

(A) A cartoon of the intact CCW-motor with the C-ring highlighted in blue, and the other conserved regions that have been resolved in teal, the Vibrio-specific proteins in light green, PomAB is colored gray as it has yet to be resolved in detail in Vibrio. (B) A cartoon of the intact motor that is rotating CW, C-ring is colored purple, the other conserved structures in coral, PomAB in gray, and CheY-P in pink. (C) A cartoon representation of the C-ring rearrangement upon CheY-P binding and rotational switching to the CW sense. In CCW rotation, the intermolecular interactions of FliG ARMM and ARMC are possible because helixNM and helixMC are ordered. The Gly-Gly flexible region is depicted by red circles. The stator (gray) is shown to interact with FliGC. Upon CheY-P binding, the C-ring undergoes a conformational change that produces CW rotation. During this transition, the center of mass tilts laterally and slightly outward. In CW rotation there are intramolecular interactions of FliG ARMM and ARMC. CheY-P (dark purple) is shown bound to FliM, and the stator (gray) interacts with FliGC. We hypothesize that the conformational change in FliG is initiated in vivo by CheY-P binding, and this switches the rotational direction of the C-ring by changing how FliGC interacts with the stator. FliG undergoes a conformational change and FliM tilts about the base of the spiral. The spiral ensures that the C-ring subunits are connected while FliG and FliM undergoes large conformational changes.

Video 3
A model for the rotational switching.

The psudeo-atomic CCW model in surface rendering side view is rotated such we are looking up the motor toward the flagellum. The MCN3 spiral is in view, and the hydrophobic patch of FliN that interact with FliH is highlighted. The view then rotates back through the side view to top view, where the charged residues of FliGC are highlighted. The view returns to side view and zooms into a patch of four adjacent subunits. The CCW subunits morph into the CW architecture, highlight the large lateral change about FliM, and the domain swapping occurs as the FliG ARMMC interactions change from intramolecular to intermolecular. The MCN3 spiral exhibits little change. The video ends with the CW psudeo-atomic model. The video was rendered using ChimeraX.

In summary, we have used cryo-ET to visualize a major conformational change of the C-ring during rotational switching due to a single point mutant in FliG. Molecular modeling attributes the tilt component of the conformational change to FliM. Gyration of FliG about FliM presents the charged residues of FliG to the stator in a manner that controls the rotational sense. These movements within the C-ring subunits may be relayed throughout the switch complex by interactions between FliM and FliN within the spiral.

Materials and methods

Bacterial strains, plasmids, and growth condition

Request a detailed protocol

Bacterial strains used in this study are listed in Table 2. To introduce the fliG deletion, NMB328 was constructed from KK148 using pHIDA3 by allelic exchange as previous reported (Kusumoto et al., 2006; Le Roux et al., 2007). V. alginolyticus strains were cultured at 30°C on VC medium (0.5% [wt/vol] polypeptone, 0.5% [wt/vol] yeast extract, 3% [wt/vol] NaCl, 0.4% [wt/vol] K2HPO4, 0.2% [wt/vol] glucose) or VPG medium (1%[wt/vol] polypeptone, 3% [wt/vol] NaCl, 0.4% [wt/vol] K2HPO4, 0.5% [wt/vol] glycerol). If needed, chloramphenicol and Isopropyl β-D-1-thiogalactopyranoside (IPTG) were added at final concentrations of 2.5 µg/ml and 1 mM, respectively. E. coli was cultured at 37°C in LB medium (1% [wt/vol] Bacto tryptone, 0.5% [wt/vol] yeast extract, 0.5% [wt/vol] NaCl). When culturing E. coli β3914 strain, 2,6-diaminopimelic acid was added to the LB medium to a final concentration of 300 µM. If needed, chloramphenicol was added at final concentrations of 25 µg/ml.

Table 2
Strains and plasmids.
Strains or plasmidsGenotype or descriptionReference or source
V. alginolyticus
VIO5Wild-type strain of a polar flagellum
(Rif+ Pof+ Laf-)
Okunishi et al., 1996
KK148VIO5 flhG (Multi-Pof+)Kusumoto et al., 2008
NMB328KK148 ΔfliG (Pof-)This study
E. coli
DH5αHost for cloning experimentsGrant et al., 1990
S17-1recA hsdR thi pro ara RP-4 2-tc::Mu-Km::Tn7 (Tpr Smr)Simon et al., 1983
β3914β2163 gyrA462 zei-298::Tn10
(Kmr Emr Tcr)
Le Roux et al., 2007
Plasmids
pMMB206Cmr, PtacPlacUV5Morales et al., 1991
pNT1fliG in pMMB206Takekawa et al., 2014
pSW7848Suicide plasmid for allele exchangeVal et al., 2012
pHIDA3SacI fragment of fliF-fliG (55–175 internal deletion of FliG) in pSW7848Mino et al., 2019

Mutagenesis

To introduce mutations (G214S or G215A) in the fliG gene on plasmid pNT1 site-directed mutagenesis was performed using the QuikChange method, as described by the manufacturer (Stratagene). All constructs were confirmed by DNA sequencing. Transformation of V. alginolyticus with plasmid pNT1 was performed by conjugational transfer from E. coli S17-1, as described previously (Okunishi et al., 1996).

Sample preparation

Request a detailed protocol

The methods of sample preparation, data collection, data processing, and sub-tomogram analysis were followed as described previously (Zhu et al., 2018). V. alginolyticus cells were cultured overnight at 30°C on VC medium, diluted 100 × with fresh VPG medium, and cultured at 30°C. After 4 or 5 hr, cells were collected and washed twice and finally diluted with TMN500 medium (50 mM Tris-HCl at pH 7.5, 5 mM glucose, 5 mM MgCl, and 500 mM NaCl). Colloidal gold solution (10 nm diameter) was added to the diluted Vibrio sp. samples to yield a 10 × dilution and then deposited on a freshly glow-discharged, holey carbon grid for 1 min. The grid was blotted with filter paper and rapidly plunge-frozen in liquid ethane.

Data collection and processing

Request a detailed protocol

The frozen-hydrated specimens of NMB328 was transferred to a Titan Krios electron microscope (Thermo Fisher Scientific). The microscope is equipped with a 300-kV field emission gun (FEG), a GIF energy filter, and a post-GIF K2 Summit direct electron detector (Gatan). The images were collected at a defocus near to 0 µm using a Volta phase plate and the energy filter with a 20 eV slit. The data were acquired automatically with SerialEM software (Mastronarde, 2005). For better data collection, the phase shift is normally distributed in the range of 0.33π to 0.67π. A total dose of 50 e-/Å (Chevance and Hughes, 2008) was distributed among 35 tilt images covering angles from −51° to +51° at tilt steps of 3°. For every single tilt series collection, the dose-fractionated mode was used to generate 8 to 10 frames per projection image. Collected dose-fractionated data were first subjected to the motion correction program to generate drift-corrected stack files (Li et al., 2013). The stack files were aligned using gold fiducial markers and volumes reconstructed using IMOD and Tomo3d, respectively (Kremer et al., 1996; Agulleiro and Fernandez, 2015). In total, 259 tomograms of CCW-state motor (G214S mutation) and 220 tomograms of CW-state motor (G215A mutation) were generated.

Sub-tomogram analysis with I3 package

Request a detailed protocol

Bacterial flagellar motors were detected manually, using the I3 program (Winkler, 2007; Winkler et al., 2009). We selected two points on each motor, one point at the C-ring region and another near the flagellar hook. The orientation and geographic coordinates of selected particles were estimated. In total, 1618 and 2221 sub-tomograms of V. alginolyticus motors from CW-motor and CCW-motor, respectively, were used for sub-tomogram analysis. The I3 tomographic package was used on the basis of the ‘alignment by classification’ method with missing wedge compensation for generating the averaged structure of the motor, as described previously (Zhu et al., 2017). To resolve the symmetry of the C-ring, we used focused refinement to the C-ring density. We then used multivariate statistical analysis for 3D classification, which allowed us to determine the 34-fold symmetry of the C-ring. It’s important to point out that we did not resolve the 34-fold symmetry by imposing different symmetries. Furthermore, the C-ring structures in the two states share the same symmetry, however they have different conformations. We did similar analysis on the stator region to resolve 13 stator units in CW-motors. However, we were unable to resolve the stator units in CCW-motors. The reason behind the difference is currently unknown.

Model generation

Request a detailed protocol

The V. alginolyticus C-ring proteins were generated using I-TASSER version 5.1 (Yang et al., 2015). The proteins were trimmed to the homologous structures deposited in the PDB to avoid large clashes. Using I-TASSER we have generated two models of Vibrio FliM, FliM full length, and FliMC with the linker region before. The full-length FliM generated by I-TASSER had the correct topology for FliMM, but the orientation FliMC relative to FliMM was incorrect. Furthermore, the folding of FliMC is similar to the previously solved crystal structure, but different enough that the FliM-FliN heterodimer was unable to be modeled. To circumvent this problem, we ran I-TASSER with just the last 223 residues of FliM, including the FliMC and the flexible region immediately before. Due to the unknown relative location of the flexible regions it was necessary to trim. FliG, FliM, and FliN were hand guided by the literature into segmented patches of the CCW-biased and CW-locked cryo-ET maps corresponding to four subunits, and fit using the ChimeraX (Goddard et al., 2018) fit to map function. Mainly, four PDB models were used the full-length FliG (3HJL [Lee et al., 2010]), FliG-FliM (4FHR [Vartanian et al., 2012]), FliM-FliN fusion (4YXB [Notti et al., 2015]), and the FliN-dimer (1YAB [Brown et al., 2005]).

Model refinement

Request a detailed protocol

The patch models were refined using PHENIX-1.17.1 Real Space Refinement (Afonine et al., 2013) to move the protein domains relative to one another while preserving the known architecture of the C-ring subunits. The unknown protein-protein interfaces were refined in Rosetta_2019.35 using the protein-protein docking scripts (Lyskov and Gray, 2008). The optimized single subunit model was then rigid body refined into a single subunit segmentation of the corresponding C-ring tomography map using PHENIX Real Space Refinement.

References

  1. 1
  2. 2
  3. 3
  4. 4
  5. 5
  6. 6
  7. 7
  8. 8
  9. 9
  10. 10
    Molecular mechanism for rotational switching of the bacterial flagellar motor
    1. Y Chang
    2. K Zhang
    3. BL Carroll
    4. X Zhao
    5. NW Charon
    6. SJ Norris
    7. MA Motaleb
    8. C Li
    9. J Liu
    (2020)
    Nature Structural & Molecular Biology 72:71–84.
  11. 11
  12. 12
  13. 13
  14. 14
  15. 15
  16. 16
  17. 17
  18. 18
  19. 19
  20. 20
  21. 21
  22. 22
  23. 23
  24. 24
  25. 25
  26. 26
  27. 27
  28. 28
  29. 29
  30. 30
  31. 31
  32. 32
  33. 33
  34. 34
  35. 35
  36. 36
  37. 37
  38. 38
  39. 39
  40. 40
  41. 41
  42. 42
  43. 43
  44. 44
  45. 45
  46. 46
  47. 47
  48. 48
  49. 49
  50. 50
  51. 51
  52. 52
  53. 53
  54. 54
  55. 55
  56. 56
  57. 57
  58. 58
  59. 59
  60. 60
  61. 61
  62. 62
  63. 63
  64. 64
  65. 65
  66. 66
  67. 67
  68. 68
  69. 69
  70. 70
  71. 71
  72. 72
  73. 73
  74. 74
  75. 75
  76. 76
  77. 77
  78. 78
  79. 79
  80. 80
  81. 81
  82. 82
  83. 83
  84. 84
  85. 85
  86. 86
  87. 87

Decision letter

  1. Edward H Egelman
    Reviewing Editor; University of Virginia, United States
  2. John Kuriyan
    Senior Editor; University of California, Berkeley, United States
  3. Edward H Egelman
    Reviewer; University of Virginia, United States
  4. Chi Aizawa
    Reviewer; Hiroshima, Japan
  5. Susan M Lea
    Reviewer; University of Oxford, United Kingdom

In the interests of transparency, eLife publishes the most substantive revision requests and the accompanying author responses.

[Editors’ note: the authors submitted for reconsideration following the decision after peer review. What follows is the decision letter after the first round of review.]

Thank you for submitting your work entitled "The flagellar motor of Vibrio alginolyticus undergoes major structural remodeling during rotational switching" for consideration by eLife. Your article has been reviewed by four peer reviewers, one of whom is a member of our Board of Reviewing Editors, and the evaluation has been overseen by a Senior Editor. The following individuals involved in review of your submission have agreed to reveal their identity: Chi Aizawa (Reviewer #2); Susan M Lea (Reviewer #3); Richard M Berry (Reviewer #4).

Our decision has been reached after consultation between the reviewers. Based on these discussions and the individual reviews below, we regret to inform you that your work will not be considered further for publication in eLife at this stage.

All reviewers agreed that an atomic model of the C-ring that allowed a molecular explanation for switching would be of great interest. The key issue is whether there is enough information in the tomograms to uniquely position existing atomic structures to enable this question to be addressed, and whether one can exclude significant conformational changes within these structures. At present the data shown do not allow the uniqueness and quality of the fit to be addressed. This paper is also unusual in that no estimates of resolution are provided. A further concern was the preprint of Chang et al. recently posted on bioRxiv. Is it possible that the mechanism proposed in Chang et al. only applies to Borrelia burgdorferi, while the mechanism in the present paper only applies to Vibrio alginolyticus? If this is true, that should have been stated in both papers, but it is stated in neither. In summary, all reviewers felt that a substantially revised paper that addressed all of these concerns should be considered again for publication in eLife.

Reviewer #1:

I had a difficult time judging the significance of the work. In some ways, the paper comes across as an incremental advance over previous studies on the bacterial flagellar motor. My main concerns had to do with the cryo-ET image analysis. This may be the first cryo-EM paper I have ever read where there is neither any mention of, nor discussion of, the resolution the reconstructed volumes. It is hard to understand how this can be absent, and it is difficult to understand the reliability of the models without some discussion of this point. Given the recent paper (Johnson et al., 2020) showing symmetry variation in the MS-ring complex, one wonders how reliable the overall imposition of a C34 symmetry is in the present reconstruction. There is no analysis of possible variance within these volumes, nor is there any information on what happens when other symmetries are imposed, both locally and globally. This is important, as the authors take one of the conclusions of their work to be that the composition of the C-ring remains constant between the two states. It is not clear how reliable such a conclusion can be at the limited resolution with the same imposed symmetry for both volumes.

Reviewer #2:

This is a beautiful work on conformational changes of the C ring of the flagellar motor.

The authors showed two conformations of the C ring in which rotation is biased to either CCW or CW.

Although I did not understand the details of the image processing using many sophisticated software, the results obtained are quite reasonable and convincing. I highly recommend this work to be published.

Reviewer #3:

The molecular details underlying how the bacterial flagellar motor switches direction of rotation is a long-studied question. This paper aims to address this question by using sub-tomogram averaging to generate volumes from directionally locked forms of the V. alginolyticus C-ring. This is an important question, being at the heart of how many bacteria move in response to their environment, and new data are therefore welcome. However, there are some issues to do with how the experiments are reported in light of past work and critical details of interpretation of the data that need to be made more explicit to allow a fuller understanding of how the new data enhance our current models. We would like to thank the authors for making the two volumes available as part of the review package to help with assessing the work.

Key prior work that provides the background to the current manuscript:

1) Prior structural work on intact C-rings by De Rosier and colleagues (Thomas et al., 2006) led to a moderate resolution, C34 reconstruction, of the C-ring with sufficient detail that permitted fitting of a spiral of a FliMN homologue at the base of the ring (Mc Dowell et al., 2016). No formal resolution estimate was provided for the earlier reconstruction, but fitting of domains of known structure in a robust fashion was possible, at least at the bases of the C-ring.

2) Sakai et al., 2019, reported low resolution analyses of locked C-ring motors from Salmonella and demonstrate that there is only a minor change in diameter at the mid-point of the C-ring (~10A).

3) We (Johnson et al., 2020) have recently reported molecular detail for the MS ring on which the C-ring is assembled showing that, at least in the context of the isolated MS-ring, assemblies with different symmetries ranging from C32-C35 exist. This tallied with earlier analyses of C-ring stoichiometry, including Sakai (Sakai et al., 2019).

Novelty of these data:

These volumes are clearly the highest resolution views of the C-ring in locked CW and CCW states to date, combined with atomic structures of component proteins this gives a great opportunity to try and build a more detailed model of the complete C-ring in both states.

Major issues to be addressed:

1) Hand of the reconstructions? There are no details on how this key step was performed.

2) Symmetry of the assembly – there is no detail on how the tomograms were treated to assess the correct symmetry to apply, with only a simple statement on p10 that “our cryo-ET model shows 34-fold symmetry for both the CCW and CW rotating C-rings”. Full details, ideally including images, on how symmetry was assigned must be provided and evidenced in the manuscript. This is especially important in light of the previous data showing variation in C-ring symmetry as, in the extreme, the relatively subtle differences between the CW and CCW volumes could reflect different proportions of 34mer in the samples being averaged rather than true differences between the states.

3) Fitting of coordinates – it is not possible to judge the quality and uniqueness of the fit of coordinates to the volumes in the current version of the manuscript. A single correlation coefficient statistic is given in the text for each volume and it is unclear (1) if this is for the full model or simply the FliMN spiral and (2) how each domain of the model contributes to this statistic. The figures that do show the fit show substantial extrusion of secondary structural elements outside of the volume and unfilled regions within the volume. This is perhaps unsurprising for a volume at modest resolution, but given that the authors go on to propose a radical structural rearrangement and then discuss the location of specific amino acids, the evidence for the uniqueness of the fit needs to be provided. Inspection of the volumes did not trivially allow the viewer to see a correlation in shape between the densities ascribed to the crucial FliG domains in the figures. There is also some confusion surrounding the FliMc/FliN spiral, as the text states that this portion of the structure is largely unaltered between the two states, but Figure 2 and Figure 2—figure supplement 1 show significant differences in the side views, with what looks like a 90 degree rotation of the spiral. How is this justified given the relatively featureless resolution of the maps?

4) Discussion of novelty throughout should acknowledge the incremental nature of the current findings c.f. the earlier work. In particular the extensive discussion c.f. the work of Sakai should acknowledge that, at the resolution of both studies, both could be summarized as “no major difference in diameter on switching”.

5) CheY binding – this is the most speculative part of the manuscript as a ring of density around the CCW ring is assigned as CheY-P largely on the basis of the groups work that will be presented in another manuscript (cited as Chang et al. in preparation). Particularly in light of another manuscript recently reported in bioRxiv that demonstrates the N-terminus of FliM is not required for CheY-P binding induced switching (Eisenback et al.), the authors should either provide further supporting experiments or more clearly mark this as speculative.

Reviewer #4:

The paper presents insitu structures of the bacterial flagellar motor in CW and CCW states, and a corresponding detailed ("pseudoatomic") structural model of each state. Models are consistent with previous biochemical data on the proximity of various amino acid residues of the constituent proteins.

The work is novel and important, and of sufficient general interest for publication in eLife: it allows analysis of the structural nature of the flagellar switch, which has been a canonical example of a macromolecular switch since the discovery of flagellar switching in the 1970s.

1) There is a lack of clarity, perhaps even an apparent internal contradiction, as to whether the structure at the filament-distal end of the motor (the FliMC/ FliN3 super-helix, hereafter "MCN3 helix") is different between CW and CCW structures. This is an important point, worthy even of inclusion in the Abstract, as any changes in this part of the C-ring caused by the 2 point mutations in FliG would indicate long-range effects of very small mutational changes. Specifically: the MCN3 helix seems to be very different between CW and CCW in Figures 2C/D and Figure 2—figure supplement 1 B/D – with fairly substantial rearrangement of protein chains. But the density maps for CW and CCW in Figure 2A/B look very similar, and the lack of changes in the MCN3 helix is confirmed in Figure 4 and in the text. Is it possible that the repeating structure of 34 units has been cut differently between Figures 2C and D, and also Figures 2—figure supplement 1B and D? Or alternatively, that the outwardly similar shape of the MCN3 helix conceals a very different internal arrangement of protein chains? Or something else? This apparent contradiction needs to be addressed and removed. A pseudoatomic model allows and demands this level of inspection, with appropriate caveats about the confidence in the model itself.

2) Closely related to comment 1, it is hard to match corresponding parts of the CW and CCW models – particularly in Figures 2C/D and Figure 2—figure supplement 1B/D. In reality there must be a pathway for switching between the CW and CCW structures. The authors should explore the transition between the two, and supply one or more animations to illustrate the conclusions that they can draw from their model regarding this important point – which is the main result of the paper. Such animations would be a powerful illustration, clarification and explanation of Figures 2-4 (and perhaps also Figure 5). They would serve 3 important purposes:

a) Identify corresponding parts, including perhaps the possibility that the repeating structure has been cut differently to isolate the units displayed in Figure 2 (see comment 1).

b) Illustrate either a smooth pathway between CW and CCW states (i.e. requiring no significant unfolding of secondary structure to avoid steric clashes, as proposed in the FliG model of Lee et al., 2010); or the lack of one, indicating rather the need for substantial disruption of interactions during a switch. The current best model for the switch is that of conformational spread (e.g. Bai et al., Science 2010), in which an energetic penalty for boundaries between domains of CW and CCW structure is a key parameter. This paper might be able to add structural data to that model.

c) Highlight and illustrate changes in the orientation and location of FliGc, which are the "output" of the switch and most relevant in terms of understanding how it can cause a change in rotation direction upon interaction with the stator. These are discussed in the text, but it is hard to see them in the existing figures.

3) Regarding the extra density peripheral to the C-ring in the CW state, visible in Figure 1 and suggested to be CheY later in the paper (Figure 2—figure supplement 1, subsection “Extra diffuse density around the CW C-ring”). The authors should cite references that proposed that FliMN is an unstructured "fishing rod" that serves to bind and concentrate CheY, which their hypothesis and data support. Also the alternative explanation (see e.g. Figure 7 in J Mol Biol. 430(22):4557-4579.) that this density could be extra copies of FliM, outside the ordered C-ring. This would reconcile previous observations of substantial recruitment of FliM to the C-ring (although in E. coli) during CW rotation with the lack of expansion of the 34-fold ring observed here.

4) Figure 5 legend is wrong – D and C are described as if they were parts of B and A respectively. More importantly, the figure presents a model where the domain-swap (visible in B as swapping of colour-labelled domains of each protein chain into the neighbouring structural unit) is present in one directional state (CCW, B) but not the other (CW, D). This may be a misunderstanding of the domain-swap model of the flagellar switch. Alteration of the domain-swap during switching is not necessary, nor is it proposed by Baker et al., 2016 (as incorrectly stated in the Introduction). In fact, quite the opposite: the model proposes that the ARM-ARM interface is very strong, the main interaction holding the FliG ring together, and therefore very unlikely to open except under the special conditions that prevail during assembly of the FliG ring. Instead, the cartoon of the CCW state, Figure 5B, appears entirely consistent with the model of Lee et al., 2010, and the switch to a domain-swapped alternative of Figure 5D would then be a data-supported refinement and confirmation of the model of Lee et al., 2010. Alternative models propose non-domain swapped structures (like the cartoon of Figure 5D) for both CW and CCW. Perhaps the authors could comment on whether their data support one or other model.

[Editors’ note: further revisions were suggested prior to acceptance, as described below.]

Thank you for resubmitting your work entitled "The flagellar motor of Vibrio alginolyticus undergoes major structural remodeling during rotational switching" for further consideration by eLife. Your revised article has been evaluated by John Kuriyan (Senior Editor), a Reviewing Editor, and three reviewers.

The manuscript has been improved but there are is a remaining issue that needs to be addressed before acceptance, as outlined below:

In general, the authors have addressed the concerns raised in the previous review. But the decision letter for the previous manuscript stated: "A further concern was the preprint of Chang et al. recently posted on bioRxiv. Is it possible that the mechanism proposed in Chang et al. only applies to Borrelia burgdorferi, while the mechanism in the present paper only applies to Vibrio alginolyticus? If this is true, that should have been stated in both papers, but it is stated in neither." This is not answered either in the response nor in the revised paper.

https://doi.org/10.7554/eLife.61446.sa1

Author response

[Editors’ note: the authors resubmitted a revised version of the paper for consideration. What follows is the authors’ response to the first round of review.]

Reviewer #1:

I had a difficult time judging the significance of the work. In some ways, the paper comes across as an incremental advance over previous studies on the bacterial flagellar motor. My main concerns had to do with the cryo-ET image analysis. This may be the first cryo-EM paper I have ever read where there is neither any mention of, nor discussion of, the resolution the reconstructed volumes. It is hard to understand how this can be absent, and it is difficult to understand the reliability of the models without some discussion of this point.

We thank the reviewer for providing critical comments on our insufficient descriptions on cryo-ET methods and the resulting structures. We have addressed the concerns. We hope that the importance of this work is better clarified.

Given the recent paper (Johnson et al., 2020) showing symmetry variation in the MS-ring complex, one wonders how reliable the overall imposition of a C34 symmetry is in the present reconstruction. There is no analysis of possible variance within these volumes, nor is there any information on what happens when other symmetries are imposed, both locally and globally. This is important, as the authors take one of the conclusions of their work to be that the composition of the C-ring remains constant between the two states. It is not clear how reliable such a conclusion can be at the limited resolution with the same imposed symmetry for both volumes.

We agree with the reviewer that the symmetry is important for this study. We have been studying the Vibrio motor for many years and published several papers including Zhu et al., 2017. We could not resolve the symmetry of the C-ring until this study. To overcome the limitation, we first used a relatively new technique (Volta phase plate) to collect cryo-ET data, resulting in reconstructions with better contrast. We then used multivariate statistical analysis for 3D classification, which allowed us to determine the 34-fold symmetry of the C-ring. It’s important to point out that we did not resolve the 34-fold symmetry by imposing different symmetries. Furthermore, the C-ring structures in the two states share the same symmetry, however they have different conformations. Again the 34-fold symmetric features were observed first before we impose the symmetry to increase the resolution.

Reviewer #2:

This is a beautiful work on conformational changes of the C ring of the flagellar motor.

The authors showed two conformations of the C ring in which rotation is biased to either CCW or CW.

Although I did not understand the details of the image processing using many sophisticated software, the results obtained are quite reasonable and convincing. I highly recommend this work to be published.

We thank the reviewer for the positive comments. We have provided more details in the Materials and methods. Hopefully the revised manuscript is easier to understand.

Reviewer #3:

The molecular details underlying how the bacterial flagellar motor switches direction of rotation is a long-studied question. This paper aims to address this question by using sub-tomogram averaging to generate volumes from directionally locked forms of the V. alginolyticus C-ring. This is an important question, being at the heart of how many bacteria move in response to their environment, and new data are therefore welcome. However, there are some issues to do with how the experiments are reported in light of past work and critical details of interpretation of the data that need to be made more explicit to allow a fuller understanding of how the new data enhance our current models. We would like to thank the authors for making the two volumes available as part of the review package to help with assessing the work.

Key prior work that provides the background to the current manuscript:

1) Prior structural work on intact C-rings by De Rosier and colleagues (Thomas et al., 2006) led to a moderate resolution, C34 reconstruction, of the C-ring with sufficient detail that permitted fitting of a spiral of a FliMN homologue at the base of the ring (McDowell et al., 2016). No formal resolution estimate was provided for the earlier reconstruction, but fitting of domains of known structure in a robust fashion was possible, at least at the bases of the C-ring.

2) Sakai et al., 2019, reported low resolution analyses of locked C-ring motors from Salmonella and demonstrate that there is only a minor change in diameter at the mid-point of the C-ring (~10A).

3) We (Johnson et al., 2020) have recently reported molecular detail for the MS ring on which the C-ring is assembled showing that, at least in the context of the isolated MS-ring, assemblies with different symmetries ranging from C32-C35 exist. This tallied with earlier analyses of C-ring stoichiometry, including Sakai (Sakai et al., 2019).

Novelty of these data:

These volumes are clearly the highest resolution views of the C-ring in locked CW and CCW states to date, combined with atomic structures of component proteins this gives a great opportunity to try and build a more detailed model of the complete C-ring in both states.

We thank the reviewers for reading and providing insightful feedback. We appreciated their understanding of the importance of the work. We have revised the manuscript to incorporate the key prior work and highlight our major findings in this study.

Major issues to be addressed:

1) Hand of the reconstructions? There are no details on how this key step was performed.

This is a critical concern. It is challenging to directly determine the handedness of cryo-ET reconstructions. We have three pieces of data to support the handedness of the reconstructions in this study. First, we have independently determined two structures of the C-rings. The base of the C-ring shares similar right-handed spiral structure in both states. Importantly, the right-handed spiral structure matches well with the crystal structure of a FliN homolog (Mc Dowell et al., 2016). Second, the exact same cryo-ET protocol was used to generate 3D reconstructions of periplasmic flagella in spirochetes. Consistent with previous findings, periplasmic flagella form a right-handed ribbon in our reconstructions. Third, our recent in situ structures of HIV-1 Env spikes match well with near-atomic structures of purified Env spikes by cryo-EM single particle analysis. Together, we believe that our reconstructions in this study have correct handedness.

2) Symmetry of the assembly – there is no detail on how the tomograms were treated to assess the correct symmetry to apply, with only a simple statement on p10 that “our cryo-ET model shows 34-fold symmetry for both the CCW and CW rotating C-rings”. Full details, ideally including images, on how symmetry was assigned must be provided and evidenced in the manuscript. This is especially important in light of the previous data showing variation in C-ring symmetry as, in the extreme, the relatively subtle differences between the CW and CCW volumes could reflect different proportions of 34mer in the samples being averaged rather than true differences between the states.

We agree with the reviewers that the symmetry of the C-ring is important for this study. Therefore, we have provided the detailed steps as requested by the reviewer 1. We first used Volta phase plate to increase the contrast of each reconstruction. We then used multivariate statistical analysis for 3D classification, which allowed us to resolve 34-fold symmetric features of the C-ring. Two independent approaches were used to determine the structures of the C-ring in both states. The resulting structures of the C-ring are considerably different.

3) Fitting of coordinates – it is not possible to judge the quality and uniqueness of the fit of coordinates to the volumes in the current version of the manuscript. A single correlation coefficient statistic is given in the text for each volume and it is unclear (1) if this is for the full model or simply the FliMN spiral and (2) how each domain of the model contributes to this statistic. The figures that do show the fit show substantial extrusion of secondary structural elements outside of the volume and unfilled regions within the volume. This is perhaps unsurprising for a volume at modest resolution, but given that the authors go on to propose a radical structural rearrangement and then discuss the location of specific amino acids, the evidence for the uniqueness of the fit needs to be provided. Inspection of the volumes did not trivially allow the viewer to see a correlation in shape between the densities ascribed to the crucial FliG domains in the figures. There is also some confusion surrounding the FliMc/FliN spiral, as the text states that this portion of the structure is largely unaltered between the two states, but Figure 2 and Figure 2—figure supplement 1 show significant differences in the side views, with what looks like a 90 degree rotation of the spiral. How is this justified given the relatively featureless resolution of the maps?

We understand how it is very challenging to judge the placement of fitted, pseudoatomic models. However, we believe the model was necessary to help explain the cryo-ET maps. To address this issue, we did three things:

1) We made a supplemental figure of in silico maps reconstructed from the model at the resolution of our experimental maps. This was to show the level of detail that we should expect to see. The maps were similar to our experimental maps, but allow the reader to visualize what is different. We think that this will help with the idea of what can be interpreted from our pseudoatomic model at our resolution.

2) We clarified in the text that the amino acids are only talked about in relative location to help confirm the location of the proteins.

“It is important to note that these residues were used to identify unique locations in our model, but due to limited resolution their precise location is not known”

3) We made it clear that the spiral is not altered specifically in shape.

“This region, known as a spiral, like that of a spiral bound notebook, is uniform within both the CCW and CW cryo-ET maps, and remains largely unaltered upon rotational switching (Figure 4C). […] By building the patch model initially we were able to preserve the similarity of the MCN3 spiral by removing potential bias from the segmentation of a monomer. […] It is important to note that there may be slight movements of the MCN3 proteins during switching, that we would need higher resolution to observe.”

4) Discussion of novelty throughout should acknowledge the incremental nature of the current findings c.f. the earlier work. In particular the extensive discussion c.f. the work of Sakai should acknowledge that, at the resolution of both studies, both could be summarized as “no major difference in diameter on switching”.

We agree with the reviewers that our finding is an important extension of earlier work from many different labs. We have revised our manuscript with extensive discussions of other related studies.

5) CheY binding – this is the most speculative part of the manuscript as a ring of density around the CCW ring is assigned as CheY-P largely on the basis of the groups work that will be presented in another manuscript (cited as Chang et al. in preparation). Particularly in light of another manuscript recently reported in bioRxiv that demonstrates the N-terminus of FliM is not required for CheY-P binding induced switching (Eisenback et al.), the authors should either provide further supporting experiments or more clearly mark this as speculative.

We agree with the reviewers that this is speculative. Further experiments will be needed to support the model.

Reviewer #4:

The paper presents in situ structures of the bacterial flagellar motor in CW and CCW states, and a corresponding detailed ("pseudoatomic") structural model of each state. Models are consistent with previous biochemical data on the proximity of various amino acid residues of the constituent proteins.

The work is novel and important, and of sufficient general interest for publication in eLife: it allows analysis of the structural nature of the flagellar switch, which has been a canonical example of a macromolecular switch since the discovery of flagellar switching in the 1970s.

1) There is a lack of clarity, perhaps even an apparent internal contradiction, as to whether the structure at the filament-distal end of the motor (the FliMC/ FliN3 super-helix, hereafter "MCN3 helix") is different between CW and CCW structures. This is an important point, worthy even of inclusion in the Abstract, as any changes in this part of the C-ring caused by the 2 point mutations in FliG would indicate long-range effects of very small mutational changes. Specifically: the MCN3 helix seems to be very different between CW and CCW in Figures 2C/D and Figure 2—figure supplement 1B/D – with fairly substantial rearrangement of protein chains. But the density maps for CW and CCW in Figure 2 A/B look very similar, and the lack of changes in the MCN3 helix is confirmed in Figure 4 and in the text. Is it possible that the repeating structure of 34 units has been cut differently between Figures 2C and D, and also Figures 2—figure supplement 1B and D? Or alternatively, that the outwardly similar shape of the MCN3 helix conceals a very different internal arrangement of protein chains? Or something else? This apparent contradiction needs to be addressed and removed. A pseudoatomic model allows and demands this level of inspection, with appropriate caveats about the confidence in the model itself.

We thank the reviewer for the critical concern. Our two structures of the C-ring suggested that the base of the C-ring remains in similar conformation. We therefore built the model based on the crystal structure of the FliN homolog (in McDowell et al., 2016). Our models were previously refined based on the segmented maps, which were unfortunately not reliable. Given the concerns and suggestions from the reviewer, we have used a different strategy, where we built a patch of 4 subunits, to refine the models. We believe that current models are consistent with the maps with 34-fold symmetry.

2) Closely related to comment 1, it is hard to match corresponding parts of the CW and CCW models – particularly in Figures 2C/D and Figure 2—figure supplement 1B/D. In reality there must be a pathway for switching between the CW and CCW structures. The authors should explore the transition between the two, and supply one or more animations to illustrate the conclusions that they can draw from their model regarding this important point – which is the main result of the paper. Such animations would be a powerful illustration, clarification and explanation of Figures 2-4 (and perhaps also Figure 5). They would serve 3 important purposes:

a) Identify corresponding parts, including perhaps the possibility that the repeating structure has been cut differently to isolate the units displayed in Figure 2 (see comment 1).

b) Illustrate either a smooth pathway between CW and CCW states (i.e. requiring no significant unfolding of secondary structure to avoid steric clashes, as proposed in the FliG model of Lee et al., 2010); or the lack of one, indicating rather the need for substantial disruption of interactions during a switch. The current best model for the switch is that of conformational spread (e.g. Bai et al., Science 2010), in which an energetic penalty for boundaries between domains of CW and CCW structure is a key parameter. This paper might be able to add structural data to that model.

c) Highlight and illustrate changes in the orientation and location of FliGc, which are the "output" of the switch and most relevant in terms of understanding how it can cause a change in rotation direction upon interaction with the stator. These are discussed in the text, but it is hard to see them in the existing figures.

We agree and have made appropriate animations.

Video 1 is a summary of the project. Starting with the low resolution CCW map and moving to the locally refined C-ring map. Then building our pseudo-atomic model. We should how the symmetrized model switches between the CCW- and CW conformations by a simple toggle. And then move back out to the CW rotation.

Video 2 is a simple look at switching from above the C-ring. This shows how the charged residues of FliG change position relative to the stator if the stator were resolved in our maps.

Video 3 provides detail at how switching may occur in the domain swapping hypothesis. Here with zoom into a patch of the motor and watch how the CCW subunits morph into the CW subunits.

3) Regarding the extra density peripheral to the C-ring in the CW state, visible in Figure 1 and suggested to be CheY later in the paper (Figure 2—figure supplement 1, subsection “Extra diffuse density around the CW C-ring”). The authors should cite references that proposed that FliMN is an unstructured "fishing rod" that serves to bind and concentrate CheY, which their hypothesis and data support. Also the alternative explanation (see e.g. Figure 7 in J Mol Biol. 430(22):4557-4579.) that this density could be extra copies of FliM, outside the ordered C-ring. This would reconcile previous observations of substantial recruitment of FliM to the C-ring (although in E. coli) during CW rotation with the lack of expansion of the 34-fold ring observed here.

We agree that our data alone does not identify the density as CheY-P, although it was recently suggested in two different studies. Alternatively, it could be also formed by extra FliM or FliN as suggested by the reviewer.

4) Figure 5 legend is wrong – D and C are described as if they were parts of B and A respectively. More importantly, the figure presents a model where the domain-swap (visible in B as swapping of colour-labelled domains of each protein chain into the neighbouring structural unit) is present in one directional state (CCW, B) but not the other (CW, D). This may be a misunderstanding of the domain-swap model of the flagellar switch. Alteration of the domain-swap during switching is not necessary, nor is it proposed by Baker et al., 2016 (as incorrectly stated in the Introduction). In fact, quite the opposite: the model proposes that the ARM-ARM interface is very strong, the main interaction holding the FliG ring together, and therefore very unlikely to open except under the special conditions that prevail during assembly of the FliG ring. Instead, the cartoon of the CCW state, Figure 5B, appears entirely consistent with the model of Lee et al., 2010, and the switch to a domain-swapped alternative of Figure 5D would then be a data-supported refinement and confirmation of the model of Lee et al., 2010. Alternative models propose non-domain swapped structures (like the cartoon of Figure 5D) for both CW and CCW. Perhaps the authors could comment on whether their data support one or other model.

Our resolution doesn’t not allow for the conformation of either model. As we push to increase resolution of the C-ring in tomography, hopefully we will be able to confirm via a true atomic resolution model. We built our models following the literature as closely as possible. Since many papers have suggested FliG is extended in the CCW rotation we used the full length FliG structure, 3HJL. To build this into the map we used the crystal packing that led to the domain swapping hypothesis. This FliG monomer assembled due to crystal packing fits into the map with little movement necessary. To build the CW model we used the “compact” FliG that has been proposed during CW rotation from the crystal structure 4FHR. The figures have been updated to clarify this by alternating the coloring of FliG. We have also addressed this in the text. We think that this is a piece of evidence to support this model, but it is not direct evidence.

We have updated our previous miscitation. We greatly thank the reviewer for noticing this.

“The domain swapping mechainism was proposed to coordinate with the conformational change in HelixMC (Lee et al., 2010). Another biochemical study suggested that the domain swapping occurs during C-ring assembly, with FliG in solution existing as a monomer, and an equilibrium favoring FliG oligomers in the C-ring (Baker et al., 2016).”

[Editors’ note: what follows is the authors’ response to the second round of review.]

The manuscript has been improved but there are is a remaining issue that needs to be addressed before acceptance, as outlined below:

In general, the authors have addressed the concerns raised in the previous review. But the decision letter for the previous manuscript stated: "A further concern was the preprint of Chang et al. recently posted on bioRxiv. Is it possible that the mechanism proposed in Chang et al. only applies to Borrelia burgdorferi, while the mechanism in the present paper only applies to Vibrio alginolyticus? If this is true, that should have been stated in both papers, but it is stated in neither." This is not answered either in the response nor in the revised paper.

First, we would like to thank the editors and the reviewers for your time and consideration of our revised manuscript. To some extent, the results in the preprint of Chang et al. matches with what we observed in Vibrio alginolyticus. We have updated our manuscript to make this clearer, although we had a sentence in the Discussion that may be overlooked. We replaced, “This finding is consistent with another structural characterization of the CW/CCW motors in B. burgdorferi (Chang et al., 2020).” with:

“In a parallel study, Chang et al. used a B. burgdorferi with a CheX deletion or CheY3 deletion to investigate the switching mechanism in spirochetes (Chang et al., 2020). We observed similar conformational changes of the C-ring in both species; however, the stator complexes were well resolved in B. burgdorferi, as they appear to be less dynamic than those in V. alginolyticus. Visualization of the stator complex provided direct evidence that the conformational change of the C-ring alters the rotor-stator interaction. The model of stator-rotor interactions controlling the rotational direction of the motor is further bolstered by the recent high-resolution cryo-EM structure of MotAB. Santiveri et al. further postulate that MotA rotates counterclockwise in Campylobacter jejunie by showing the conformations of wildtype MotAB and MotAB trapped in the protonated state, are nearly identical, thus, ruling out a dramatic conformational change in the stator unit.”

To further elaborate, the two manuscripts fit nicely together. We used different but analogous approaches to end up at the similar model in which the C-ring underwent a comparable conformational change in both species in different rotational directions. In V. alginolyticus we were able to generate single point mutation variants that locked the motors in CW or CCW rotation. This enabled us to be truly confident in the rotational direction of the flagellum motor. In B. burgdorferi we created motors that were half CCW by deleting CheY3, or all CW by deleting CheX, the kinase responsible for phosphorylation of CheY3. Spirochetes have a unique mode of movement as the flagella are located at both cell poles, therefore during the “run” one cell pole is rotating CCW and the other rotating CW (potentially due to an unknown protein). In B. burgdorferi we were able to resolve density for the stator complex, this has been shown before but not to such resolution. The ability to resolve the stator helps identify FliG-MotA interactions. We are working on trying to visualize the stator in the V. alginolyticus motors, but the stator occupancy seems to be more dynamic making this a very challenging task. The B. burgdorferi visualization of the stator, reinforced by the new stator cryo-EM structures recently put on bioRxiv, provides additional evidence for our model. We were unable to discuss much about the stator involvement in this manuscript. But comparison of our CCW and CW C-rings revealed an unambiguously similar pattern with the movement of the C-ring repositioning FliGC thus altering the FliG-MotA interaction.

https://doi.org/10.7554/eLife.61446.sa2

Article and author information

Author details

  1. Brittany L Carroll

    1. Department of Microbial Pathogenesis, Yale School of Medicine, New Haven, United States
    2. Microbial Sciences Institute, Yale University, West Haven, United States
    Contribution
    Formal analysis, Investigation, Visualization, Methodology, Writing - original draft
    Contributed equally with
    Tatsuro Nishikino
    Competing interests
    No competing interests declared
  2. Tatsuro Nishikino

    Division of Biological Science, Graduate School of Science, Nagoya University, Furo-cho, Chikusa-ku, Nagoya, Japan
    Contribution
    Conceptualization, Data curation, Formal analysis, Investigation, Writing - original draft, Writing - review and editing
    Contributed equally with
    Brittany L Carroll
    Competing interests
    No competing interests declared
  3. Wangbiao Guo

    1. Department of Microbial Pathogenesis, Yale School of Medicine, New Haven, United States
    2. Microbial Sciences Institute, Yale University, West Haven, United States
    Contribution
    Formal analysis
    Competing interests
    No competing interests declared
  4. Shiwei Zhu

    1. Department of Microbial Pathogenesis, Yale School of Medicine, New Haven, United States
    2. Microbial Sciences Institute, Yale University, West Haven, United States
    Present address
    Howard Hughes Medical Institute, Yale School of Medicine, New Haven, United States
    Contribution
    Data curation, Formal analysis
    Competing interests
    No competing interests declared
  5. Seiji Kojima

    Division of Biological Science, Graduate School of Science, Nagoya University, Furo-cho, Chikusa-ku, Nagoya, Japan
    Contribution
    Conceptualization, Supervision, Funding acquisition, Investigation, Writing - review and editing
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0002-5582-8935
  6. Michio Homma

    Division of Biological Science, Graduate School of Science, Nagoya University, Furo-cho, Chikusa-ku, Nagoya, Japan
    Contribution
    Conceptualization, Formal analysis, Supervision, Writing - original draft, Writing - review and editing
    For correspondence
    g44416a@cc.nagoya-u.ac.jp
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0002-5371-001X
  7. Jun Liu

    1. Department of Microbial Pathogenesis, Yale School of Medicine, New Haven, United States
    2. Microbial Sciences Institute, Yale University, West Haven, United States
    Contribution
    Conceptualization, Supervision, Funding acquisition, Validation, Investigation, Writing - original draft, Writing - review and editing
    For correspondence
    jliu@yale.edu
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0003-3108-6735

Funding

Japan Society for the Promotion of Science (JP16H04774)

  • Seiji Kojima

Japan Society of Ultrasonics in Medicine (JP18K19293)

  • Seiji Kojima

National Institute of Allergy and Infectious Diseases (AI087946)

  • Jun Liu

National Institute of General Medical Sciences (GM107629)

  • Brittany L Carroll
  • Shiwei Zhu
  • Jun Liu

Japan Society for the Promotion of Science (JP18K06155)

  • Tatsuro Nishikino

National Institutes of Health (R01AI087946)

  • Seiji Kojima

Japan Society for the Promotion of Science (JP17J11237)

  • Brittany L Carroll
  • Shiwei Zhu
  • Jun Liu

Japan Society for the Promotion of Science (JP17J11237)

  • Tatsuro Nishikino

The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication.

Acknowledgements

We thank Michael Manson for critical reading and suggestions. We thank A Abe in our laboratory for technical support in this research. We thank Shenping Wu for her support on Krios. This work was supported in part by JSPS KAKENHI Grant Numbers JP16H04774 and JP18K19293 (to SK), JP18K06155 (to TK), and Program for leading Graduate Schools of Japan, Science for the Promotion of Science (JP17J11237 to TN). TN was supported in part by the Integrative Graduate Education and Research program of Nagoya University. BLC, SZ and JL were supported by grants GM107629 and R01AI087946 from National Institutes of Health. Molecular graphics and analyses were performed with UCSF ChimeraX, developed by the Resource for Biocomputing, Visualization, and Informatics at the University of California, San Francisco, with support from National Institutes of Health R01-GM129325 and the Office of Cyber Infrastructure and Computational Biology, National Institute of Allergy and Infectious Diseases.

Senior Editor

  1. John Kuriyan, University of California, Berkeley, United States

Reviewing Editor

  1. Edward H Egelman, University of Virginia, United States

Reviewers

  1. Edward H Egelman, University of Virginia, United States
  2. Chi Aizawa, Hiroshima, Japan
  3. Susan M Lea, University of Oxford, United Kingdom

Publication history

  1. Received: July 26, 2020
  2. Accepted: September 4, 2020
  3. Accepted Manuscript published: September 7, 2020 (version 1)
  4. Version of Record published: September 21, 2020 (version 2)

Copyright

© 2020, Carroll et al.

This article is distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use and redistribution provided that the original author and source are credited.

Metrics

  • 578
    Page views
  • 126
    Downloads
  • 0
    Citations

Article citation count generated by polling the highest count across the following sources: Crossref, PubMed Central, Scopus.

Download links

A two-part list of links to download the article, or parts of the article, in various formats.

Downloads (link to download the article as PDF)

Download citations (links to download the citations from this article in formats compatible with various reference manager tools)

Open citations (links to open the citations from this article in various online reference manager services)

Further reading

    1. Microbiology and Infectious Disease
    Jason Neidleman et al.
    Research Article

    The latent reservoir is a major barrier to HIV cure. As latently infected cells cannot be phenotyped directly, the features of the in vivo reservoir have remained elusive. Here, we describe a method that leverages high-dimensional phenotyping using CyTOF to trace latently infected cells reactivated ex vivo to their original pre-activation states. Our results suggest that, contrary to common assumptions, the reservoir is not randomly distributed among cell subsets, and is remarkably conserved between individuals. However, reservoir composition differs between tissues and blood, as do cells successfully reactivated by different latency reversing agents. By selecting 8–10 of our 39 original CyTOF markers, we were able to isolate highly purified populations of unstimulated in vivo latent cells. These purified populations were highly enriched for replication-competent and intact provirus, transcribed HIV, and displayed clonal expansion. The ability to isolate unstimulated latent cells from infected individuals enables previously impossible studies on HIV persistence.

    1. Microbiology and Infectious Disease
    Kilian Dekoninck et al.
    Research Article

    OmpA, a protein commonly found in the outer membrane of Gram-negative bacteria, has served as a paradigm for the study of b-barrel proteins for several decades. In Escherichia coli, OmpA was previously reported to form complexes with RcsF, a surface-exposed lipoprotein that triggers the Rcs stress response when damage occurs in the outer membrane and the peptidoglycan. How OmpA interacts with RcsF and whether this interaction allows RcsF to reach the surface has remained unclear. Here, we integrated in vivo and in vitro approaches to establish that RcsF interacts with the C-terminal, periplasmic domain of OmpA, not with the N-terminal b-barrel, thus implying that RcsF does not reach the bacterial surface via OmpA. Our results suggest a novel function for OmpA in the cell envelope: OmpA competes with the inner membrane protein IgaA, the downstream Rcs component, for RcsF binding across the periplasm, thereby regulating the Rcs response.