MotY interacts with both stator proteins MotB and MotD

In Vibrio species, the T ring is thought to achieve stator recruitment through the interaction between MotX and PomB. In contrast, the T ring of P. aeruginosa is composed solely of the MotY protein, and P. aeruginosa has two sets of stator units. To investigate the potential interactions between MotY and the stator units, we employed FRET, an assay using the energy transfer between two light-sensitive molecules where intermolecular distance is a key factor.

In a FRET assay, the interaction of a pair of donor and acceptor fluorescence proteins was examined. If an interaction exists, bleaching of the acceptor eliminates the energy transfer, causing an increase in donor fluorescence emission. We constructed an eCFP-MotY fusion as the donor, and eYFP-MotB and eYFP-MotD fusions as acceptors. FRET assays were performed for the eCFP-MotY/eYFP-MotB pair in the ∆fliC∆motAB∆motY strain, and for the eCFP-MotY/eYFP-MotD pair in the ∆fliC∆motCD∆motY strain. In both measurements, we found that CFP emission increased upon bleaching of YFP (Fig. 1), indicating energy transfer and protein interaction.

Fig. 1

FRET assays by acceptor bleaching with eCFP-MotY/eYFP-MotB and eCFP-MotY/eYFP-MotD pairs. (A) Representative FRET assay result for the eCFP-MotY/eYFP-MotB pair expressed in the ∆fliC∆motAB∆motY strain. Positive FRET is observed as an increase in CFP emission after YFP bleaching for 3 min. FRET efficiency E = 1.72 ± 0.96% (N = 16) indicates a strong interaction between MotY and MotB. (B) Representative FRET assay result for the eCFP-MotY/eYFP-MotD pair in the ∆fliC∆motCD∆motY strain. E = 1.54 ± 0.55% (N = 12) suggests a strong interaction between MotY and MotD

The FRET efficiency (E) was 1.72 ± 0.96% for the eCFP-MotY/eYFP-MotB pair (N = 16), and 1.54 ± 0.55% for the eCFP-MotY/eYFP-MotD pair (N = 12). Given that E ≥ 1% is typically regarded as the threshold for strong FRET [19], our results suggest that MotY interacts strongly with both MotB and MotD. As a control, ecfp-motY-pME6032 was transformed into ∆fliC∆motAB∆motY and ∆fliC∆motCD∆motY strains, and no enhancement in the CFP channel was observed after bleaching (Fig. S1). To confirm whether the FRET occurred at the motor, we constructed strain ∆fliC∆motAB∆motY∆fliG and conducted FRET assays by acceptor bleaching with eCFP-MotY/eYFP-MotB pairs (Fig. S2). FliG is the main component protein of the flagellar motor C ring [1], and the ΔfliG strain disrupted the assembly of the flagellar motor. We observed a notably reduced FRET signal (efficiency = 1.38 ± 0.33%, N = 10) compared to the ΔfliCΔmotABΔmotY strain. This suggests that while some interactions occur outside the fully assembled motor, a significant portion of the FRET signal is specific to motor-associated interactions.

We also test whether MotY, MotB, and MotD maintain their native function after fusion with fluorescent proteins. We conducted additional swimming plate assays on related strains (shown in Fig. S3). The introduction of exogenous eCFP-MotY and eYFP-MotD restored the motility of ∆motY and ∆motD strains, respectively, demonstrating that these fusion proteins retained their native functions. The strain expressing eYFP-MotB showed a reduced motility phenotype compared to the ∆motB strain. This suggests the existence of stator competition for occupancy. Specifically, eYFP-MotB likely occupied the original site of MotD due to its high expression level. As MotB does not support swimming in agar as well as MotD, this resulted in reduced motility. This observation is consistent with Baker et al. [20], who found that overexpression of MotAB protein in the wild-type strain caused a loss of swarming motility, further supporting the concept of dual stator competition for occupancy. In summary, the three fluorescent fusion proteins can reach the corresponding positions and function in the motor. Furthermore, we performed fluorescence imaging of the intracellular localizations of eCFP-MotY, eYFP-MotB, and eYFP-MotD (shown in Fig. S4). eYFP-MotB or eYFP-MotD bright spots were observed in some cells, although eCFP-MotY showed a more diffuse pattern.

Absence of MotY results in lower motor speed and higher motor switching rate

Structural analysis suggests that the T ring formed by MotY enhances motor torque generation, but direct evidence has been lacking. To address this, we attached micron-sized beads to the flagellar filament stubs of the motY knockout strain and used a high-speed camera to accurately record their trajectory, depicting single-motor level dynamic output.

We explored the effect of motY loss on the speed output of the motor. We measured the average rotation speeds for the ∆motY strain at 35.13 ± 10.56 Hz (counterclockwise, CCW) and 31.97 ± 10.32 Hz (clockwise, CW), nearly 30% lower than the wild-type (Fig. 2A), which rotated at 51.75 ± 9.06 Hz (CCW) and 48.51 ± 8.95 Hz (CCW) according to our previous study [18]. Motor torque decreased proportionally.

Fig. 2

MotY knockout affects multi-level motility-related phenotypes of P. aeruginosa. (A) Motor CCW and CW speeds for wild-type and ∆motY strains. The speed of the ∆motY strain (N = 142) is 30% lower than that of the wild-type strain. (B) Motor switching rate for wild-type and ∆motY strains. The switching rate of the ∆motY strain is 60% higher than that of the wild-type strain. ‘***’: significant difference (P-value<0.0001). (C) Typical time traces of motor rotation speed for the wild-type (top) and ∆motY mutant (bottom). (D) Swimming-plate assays for the wild-type/∆motY/∆motB∆motCD, ∆motAB/∆motAB∆motY/∆motB∆motCD and ∆motCD/∆motCD∆motY/∆motB∆motCD strains. Wild-type cells show a larger expansion radius than ∆motY mutants on swimming plates, suggesting that MotY deletion impairs the motility and environmental exploration ability of P. aeruginosa. Expansion of ∆motAB∆motY and ∆motCD∆motY was severely inhibited. All plates were photographed 16 h after inoculation

The switching rate of the wild-type strain was 0.42 ± 0.13 s− 1, while for the ∆motY mutant, it was 0.67 ± 0.43 s− 1 (Fig. 2B), approximately 60% higher and with a significantly more dispersed data distribution. One-way analysis of variance (ANOVA) and paired-sample t-tests showed significant statistical differences in both motor speed and switching rate between the two strains (P-value<10− 4). Typical time traces of the wild-type and ∆motY strains are shown in Fig. 2C.

Additionally, we used the swimming plate assay to observe the macroscopic expansion behavior of the wild-type and several mutants (Fig. 2D, Fig. S5). This revealed that deletion of motY affects the environmental exploration ability of the bacterial population, and the combined knockout of motY and individual stator units further inhibits expansion. In particular, the ∆motAB∆motY strain exhibits a phenotype similar to that of the ∆motB∆motCD strain, suggesting a near-loss of motility.

Here, we clarified the role of MotY in the output of the flagellar motor of P. aeruginosa through the bead assay. As a non-conserved protein, MotY does not exist in strains with biased motors such as Escherichia coli, and there is no similar T-ring structure in the corresponding strains. Therefore, we aimed to further explore the effect of MotY on the switching preference of unbiased motor and extend the understanding of the physiological significance of the T ring.

T ring is the structural basis for ensuring the unbiased motor characteristics

For biased flagellar motors, CW bias is controlled by the core chemotaxis protein CheY-P. The number of CheY-P molecules binding to corresponding sites on the motor C ring directly affects the CW bias, thus elegantly connecting the motility and chemotaxis systems [21,22,23]. P. aeruginosa features an unbiased motor (with a CW bias of 0.5), and it is speculated that intracellular CheY-P concentration influences the switching rate of the motor without affecting its CW bias [24]. This unbiased nature is surprising, especially considering that factors such as specific intracellular protein concentrations vary among individual cells. This suggests that there may be structural elements that ensure the unbiased nature of the P. aeruginosa motor, which we sought to identify.

Previously, we found that the knockout of certain stator-related genes (ΔmotAB and ΔmotCD) primarily affects motor output in terms of torque generation and switching rate, while maintaining CW bias highly consistent with the wild-type strain [18]. Similar conclusions were later drawn for the knockout of the motor modulator fliL [25]. Notably, the stator units (MotAB and MotCD) dynamically associate with or dissociate from the rotor, while FliL is a connecting protein interacting with the stator and rotor. Neither generates permanent motor structures.

The T ring formed by MotY is a non-conserved ring structure in flagellar motors. We further explored whether MotY affects motor CW bias. Figure 3 shows the CW bias distribution of several strains, with the ΔmotY mutant exhibiting an average value of 0.31 ± 0.17, significantly different from other strains. Additionally, exclusively CCW-rotating individuals appeared in the ΔmotY strain. Loss of MotY disrupts the unbiased nature of the P. aeruginosa flagellar motor and amplifies phenotypic differences between individuals within a population. Transformation of the ΔmotY mutant strain with a MotY expression plasmid restored the motor’s unbiasedness (Fig. 3), further supporting this conclusion.

Fig. 3

CW bias distribution of several P. aeruginosa strains. P. aeruginosa has an unbiased motor. The loss of stable structural proteins (FliL) or dynamic assembly proteins (MotAB or MotCD) does not alter this unbiased nature, while the loss of MotY disrupts this fundamental property. ‘***’: significant difference (P-value<0.0001), ‘ns.’: no significant difference (P-value>0.05). ∆motYRS indicates the ∆motY strain transformed with a plasmid expressing MotY

Stator assembly efficiency, but not overall flagellar assembly efficiency, is affected by MotY deletion

At the single motor level, the T ring increases torque generation and ensures CW bias characteristics. The known conserved motor ring structures are usually indispensable links in the flagellar assembly process. For example, flagella (including the motors) in MS ring or C ring mutants are unable to assemble completely [26,27,28]. We sought to investigate whether the absence of the T ring would affect the assembly of flagella to clarify its structural significance.

Previously, by introducing amino acid mutations into the flagellar filament protein FliC, in vivo visualization of flagellar filaments was achieved [29]. As the extension of flagellar filaments is the last step in flagellar assembly, flagellar assembly efficiency can be measured by quantitatively analyzing the growth of flagellar filaments. We performed flagellar fluorescence imaging for the wild-type and ∆motY strains (Fig. 4A). Statistical analysis showed that 84.40% (N = 1494) of ∆motY cells had clear flagellar filaments (Fig. S6A), similar to the wild-type strain (82.15%, N = 2095), suggesting that the loss of the T ring does not affect overall flagellar assembly efficiency.

Fig. 4

Effect of motY deletion on flagellar and stator unit assembly efficiencies in P. aeruginosa. (A) The motY mutant exhibited flagellar assembly efficiency similar to that of the wild-type strain. Endogenous flagellar filament protein FliC for the wild-type (top) and ∆motY (bottom) strains was labeled with maleimide-AlexaFluor568 and visualized by phase contrast and fluorescence microscopy. Scale bar: 10 μm. (B) MotY deletion affects the dynamic assembly efficiency of the stator units. Cellular localization of EGFP-fused MotB and MotD in the wild-type and ∆motY strains was observed using a TIRF microscope. Scale bar: 5 μm

We then explored whether the loss of MotY would affect the assembly of the stator units. Gene editing technology was used to fuse EGFP to the N-terminus of MotD. There is a strong asymmetry in the expression levels of MotAB and MotCD stator proteins, with MotAB expressing several times more than MotCD [30]. In light of this, we fused mEGFP, which disperses well and is less prone to aggregation, with MotB. The construction of these fluorescent fusion proteins enabled more accurate information gathering. As the dynamic assembly of stator units occurs near the cell membrane, we used TIRF microscopy to avoid background fluorescence interference (Fig. 4B). Statistical analysis of over 3,000 cells showed that the probability of unipolar fluorescent bright spots of the stator units (MotB and MotD) decreased by about 30% after motY knockout (Fig. S6B), indicating that motY deletion affects stator assembly efficiency.