Mechanical stimuli activate gene expression via a cell envelope stress sensing pathway

Mechanosensitive mechanisms are often used to sense damage to tissue structure, stimulating matrix synthesis and repair. While this kind of mechanoregulatory process is well recognized in eukaryotic systems, it is not known whether such a process occurs in bacteria. In Vibrio cholerae, antibiotic-induced damage to the load-bearing cell wall promotes increased signaling by the two-component system VxrAB, which stimulates cell wall synthesis. Here we show that changes in mechanical stress within the cell envelope are sufficient to stimulate VxrAB signaling in the absence of antibiotics. We applied mechanical forces to individual bacteria using three distinct loading modalities: extrusion loading within a microfluidic device, direct compression and hydrostatic pressure. In all cases, VxrAB signaling, as indicated by a fluorescent protein reporter, was increased in cells submitted to greater magnitudes of mechanical loading, hence diverse forms of mechanical stimuli activate VxrAB signaling. Reduction in cell envelope stiffness following removal of the endopeptidase ShyA led to large increases in cell envelope deformation and substantially increased VxrAB response, further supporting the responsiveness of VxrAB. Our findings demonstrate a mechanosensitive gene regulatory system in bacteria and suggest that mechanical signals may contribute to the regulation of cell wall homeostasis.


Results and discussion
VxrAB signaling is activated by extrusion loading.To investigate the role of mechanical stress on VxrAB signaling, we used a custom microfluidic device to apply controlled, reproducible mechanical loads in a process we call "extrusion loading".Extrusion loading uses fluid pressure to push bacteria into narrow tapered channels with sub-micron dimensions (Fig. 1B) 18,32 .Cells become lodged in the tapered channels and experience mechanical forces as they are deformed by the channel walls.The pressure difference (∆P) across the tapered channel regulates the magnitude of mechanical stress experienced by a trapped cell.Cells submitted to greater pressure difference inside the tapered channels travel further into the tapered channels and experience greater deformation by the channel walls and greater magnitudes of mechanical stress.Analytical and finite element models indicate that extrusion loading results in increases in axial tensile stress (along the length of a rod-like cell), reductions in hoop tensile stress (circumferentially around the cell envelope), and increases in octahedral shear (shape-changing) stress in the cell envelope 18 .
We used a transcriptional P murJ :msfGFP fusion as a well-established reporter for VxrAB signaling (Fig. 1A).MurJ encodes for lipid II flippase for peptidoglycan precursor, and flipping the peptidoglycan precursor into the periplasmic space is critical for cell wall assembly 33 .The response regulator VxrB has a high affinity for direct binding of the murJ promoter 28 , and murJ expression is consequently strongly controlled by VxrAB 27 , rendering this construct a robust readout of VxrAB activation.We applied extrusion loading to P murJ :msfGFP cells for two hours to allow sufficient time for transcription and protein folding, then measured the fluorescence of individual cells to quantify the msfGFP expression under MurJ promoter control.The fluorescence of P murJ :msfGFP cells increased with increasing magnitude of extrusion loading (Fig. 1E), supporting the idea that VxrAB signaling is mechanosensitive (although it is possible that the fluorescence saturates after a differential pressure of 3 kPa).
Vibrio cholerae cells naturally exhibit a crescent shape (Fig. 1C).The straightening of a crescent-shaped cell inside the microfluidic device during extrusion loading results in additional mechanical stresses including greater tensile stresses on the concave side of the cell and compressive stresses on the convex side.To determine whether the mechanosensitive fluorescent response was solely due to the stresses caused by cell straightening, we created rod-shaped V. cholerae by deleting crvA (Fig. 1D) 34 and submitted ΔcrvA P murJ :msfGFP cells to extrusion loading.CrvA is a periplasmic polymer protein that is responsible for curvature in V. cholerae; removal of crvA results in rod-shaped cells 34 .The fluorescence of ΔcrvA P murJ :msfGFP cells also increased with increasing pressure difference (Fig. 1F, magenta), suggesting that cell curvature is not necessary for the mechanosensitive fluorescent response of P murJ :msfGFP cells during extrusion loading.The difference in slope of the fluorescence vs pressure difference between the crescent-shaped and the rod-shaped cells may be explained by the additional stresses that crescentshaped cells experience from straightening.All subsequent experiments were performed with rod-shaped, ΔcrvA cells.The autofluorescence signal of ΔcrvA non-GFP producing cells was not increased at greater magnitudes of extrusion loading (Supplementary Fig. S1) nor was the reporter fluorescence related to magnitude of mechanical stress in the first minutes after the initiation of extrusion loading (Supplementary Fig. S2).
To confirm that the mechanosensitive response was due to VxrB-activated expression of P murJ :msfGFP, we used a mutant with a partial deletion of the VxrB binding site on the MurJ promoter (ΔcrvA P murJ ΔvxrB box :msfGFP); we previously established that this mutant indeed lacks VxrAB-responsiveness 28 .Under extrusion loading, the fluorescence of such mutant cells showed only small, negative correlations with magnitude of pressure (Fig. 1G, www.nature.com/scientificreports/blue).Cells deficient in vxrAB do not show a mechanosensitive response (Supplementary Fig. S3).Similarly, a mutation in the conserved phosphorylated histidine residue in VxrA in its native chromosomal locus (H301A strain) that results in a phenocopy of a ∆vxrAB strain (Supplementary Fig. S4) leads to a poor relationship between mechanical loading and P murJ signaling (Fig. 1H, p = 0.05).Taken together, these data demonstrate that VxrB activation through the canonical histidine kinase function is required for the mechanosensitive increase in P murJ :msfGFP expression.
In addition to its role in MurJ induction, VxrAB signaling is also autoregulated (Fig. 1A) 28 .To confirm that the mechanosensitive response of VxrAB was not limited to the MurJ promoter, we also used a P vxrAB :msfGFP reporter strain as an alternative readout of VxrAB activation.Cell fluorescence of P vxrAB :msfGFP cells increased at greater pressure difference (Fig. 1I), confirming that mechanical stress in the cell envelope regulates VxrAB signaling (note that the basal activity of the VxrAB promoter differs from that of MurJ promoter resulting in different total amounts of cell fluorescence).We conclude that VxrAB activation is sensitive to mechanical stress from extrusion loading.
VxrAB signaling is activated by compression.To confirm that the mechanosensitive response of VxrAB signaling is not specific to the microfluidic device environment or the forms of mechanical stresses generated by extrusion loading, we investigated the response of VxrAB signaling to other methods of mechanical loading.
To apply compression, cells were sandwiched between agarose gel and a weighted glass slide (Fig. 2A).At greater magnitudes of applied force, cells experienced greater mechanical loading and deformed to a greater cell width in the plane of imaging (Supplementary Fig. S5).Compression causes increases in tensile stress in the axial and hoop directions and overall increase in octahedral shear (shape-changing) stress within the cell envelope.P murJ :msfGFP cells were submitted to compression loading for two hours (the same duration that cells experienced under extrusion loading), and the fluorescence of individual cells was then measured to quantify expression of P murJ :msfGFP.Although there was greater variability in cell fluorescence, on average cell fluorescence increased 30% with increasing magnitude of applied compression force (Fig. 2B, left), indicating that VxrAB signaling is sensitive to compression loading.In contrast, upon deleting the vxrB box in the promoter (ΔvxrB box), the relationship between cell fluorescence and applied compression force was not detectable (Fig. 2B, right), suggesting that the mechanosensitive increase in cell fluorescence for P murJ :msfGFP cells under compression also required VxrB activation.We again used P vxrAB :msfGFP cells as a secondary reporter for VxrAB signaling.Cell fluorescence of P vxrAB :msfGFP cells increased with increasing applied compression force (Fig. 2C), further supporting the idea that the mechanosensitive response to compression is mediated by VxrAB.We note that msfGFP fluorescence shows larger variance among individual cells during compression loading than during extrusion loading, which is likely because this loading modality does not rigorously control the cell orientation relative to loading, the amount of load applied to each individual, or cell-cell contact, all of which may alter the mechanical stress state within the cell envelope.
To supplement these findings, we assessed the role of hydrostatic pressure in VxrAB induction.Hydrostatic pressure exerts force perpendicular to all surfaces of the cell envelope (Fig. 2D-F).Extreme hydrostatic pressure (> 50 MPa) causes changes to RNA synthesis and DNA replication and can cause cell death 35 ; here we apply mild hydrostatic pressure of 0-100 kPa (a bacterium 10 m underwater experiences approximately 100 kPa hydrostatic pressure), in contrast a cell leaving a host gastrointestinal system and entering brackish water is expected to experience an increase in turgor pressure of 500 kPa.Hydrostatic pressure increases the compressive stresses perpendicular to the cell envelope.The magnitude of mechanical stresses caused by hydrostatic pressure are typically much smaller than the hoop and longitudinal stresses (oriented parallel to the cell envelope surface) generated by extrusion loading or compression, but under conditions of applied hydrostatic pressure can become noticeable.Hydrostatic pressure was applied to cells suspended in liquid media in a custom chamber connected to a microfluidic pump.After 2 h of incubation at room temperature, cells were removed and visualized.Average cell fluorescence increased 28% with increasing magnitude of applied compression force (Fig. 2E, left), supporting the idea that VxrAB signaling is also sensitive to hydrostatic pressure.Upon deleting the vxrB box in the promoter, cell fluorescence did not vary with applied hydrostatic pressure (Fig. 2E, right), confirming that the mechanosensitive increase in cell fluorescence for P murJ :msfGFP cells required VxrB activation.As in the compression experiments, the P vxrAB :msfGFP reporter strain were consistent with the observations with P murJ :msfGFP (Fig. 2F).The variance in fluorescence was greater than that seen in compression loading, a finding we attribute to the fact that mechanical stresses in the cell envelope are limited to radial compression and are not as well controlled (cell contact with the walls of the device and other cells is completely uncontrolled).As a result, the mechanosensitive response to hydrostatic pressure was dominated by a subset of cells (upper portion of cell population in Fig. 2E,F).We conclude that VxrAB signaling is responsive to diverse methods of applying mechanical load to the cell envelope.

Cell wall turnover affects mechanosensitivity of VxrAB signaling.
VxrAB is known to respond to cell wall damage caused by cell wall targeting antibiotics and to the activity of the endopeptidase ShyA through an unknown mechanism (Fig. 1A) 27 .The endopeptidase ShyA plays a key role in cell wall homeostasis and cell elongation through controlled cell wall degradation in V. cholerae 36 .Removal of ShyA may therefore modulate cell envelope stiffness, which may affect VxrAB and the associated mechanosensory response.To explore the possibility that VxrAB mechanosensitive signaling is dependent on upstream signals from ShyA, we exposed mutants in which shyA was deleted to extrusion loading and measured the resulting changes in VxrAB signaling.
Indeed, ΔshyA P murJ :msfGFP cells exhibited noticeably perturbed physiology with blebs, curves, and abnormal shape, suggesting the possibility of impaired cell envelope mechanical properties that might reduce cell stiffness (Fig. 3A).Cells with shyA deleted were wider than cells with normal shyA expression outside of the microfluidic device (Fig. 3B, top vs. bottom), an observation expected if the stiffness of the cell envelope were reduced.We speculate that deletion of ShyA may lead to reductions in the stiffness of the cell envelope, for example by causing overexpression of other cell wall lytic enzymes.ShyA deficient cells did not travel as far into the tapered channels during extrusion loading as cells with normal ShyA expression (Fig. 3C), potentially suggesting that cells defective in major endopeptidase activity experienced insufficient cell deformation compared with cells with normal ShyA expression.However, because they are initially wider, ΔshyA P murJ :msfGFP cells actually experience greater mechanical strain/deformation than the P murJ :msfGFP cells despite not traveling as far in the tapered channel.Indeed, when comparing cell width inside and outside the tapered channels, ΔshyA P murJ :msfGFP cells experienced a greater percentage decrease (average of 40% decrease, 0.91 ± 0.11 μm undeformed v. 0.54 ± 0.09 μm loaded, mean ± SD) in cell width as compared to P murJ :msfGFP cells (average of 25% decrease in cell width, 0.67 ± 0.06 μm undeformed v. 0.51 ± 0.03 μm loaded).Larger deformations within the tapered channels would be expected to enhance the response of mechanisms sensitive to mechanical strain within the cell envelope.Consistent with this possibility, the baseline fluorescence of cells with shyA deleted (ΔshyA P murJ :msfGFP, 8.84e + 04 on average, Fig. 3D) under extrusion loading was greater than that in cells with intact shyA (4.04e + 04 on average, Fig. 1F).However, the response was poorly correlated with the magnitude of extrusion loading (a slight negative relationship rather than positive, Fig. 3D).When ShyA was reintroduced in trans (ΔshyA ShyA + + P murJ :msfGFP), the mechanosensitive response was partially restored: cell fluorescence increased with increasing pressure difference with a slope closer to that seen in the ΔcrvA P murJ :msfGFP (Fig. 3E, compared to pink line in Fig. 1F).Since cells in tapered channels for both groups experience the same pressure difference, the poor correlation between fluorescence and pressure difference in the ΔshyA P murJ :msfGFP cells is not due to differences in applied mechanical force.Instead we interpret this finding to be very high, perhaps saturating, induction of the VxrAB signaling under the greater magnitudes of deformation experienced by ΔshyA cells.An additional possibility is that the absence of ShyA results in changes in PG structure in a way that makes it nonconducive to VxrAB-mediated mechanosensing.Although the mechanotransduction mechanisms that make VxrAB mechanosensitive remain unclear, these findings demonstrate how cell envelope stiffness can regulate the response of VxrAB signaling to mechanical loading.

Conclusions
We have demonstrated that the two-component signaling system VxrAB is activated by mechanical stress in the cell envelope caused by distinct mechanical loading modalities.These findings indicate that mechanosensitive mechanisms within the cell envelope can regulate gene expression involved in cell wall remodeling.
Our findings regarding mechanosensitivity of VxrAB are consistent with the idea that mechanical stress and strain may contribute to cell wall homeostasis.The ability of mechanical stress and strain to contribute to the remodeling of a load-bearing component such as the cell wall is a powerful means of maintaining the function of the cell envelope.While we are unaware of other studies directly investigating the effects of mechanical stress on cell wall maintenance in other bacteria, there are other two-component systems that regulate cell wall remodeling.The two component system WalKR in Bacillus subtilis responds to degradation of cell wall constituents to regulate cell wall remodeling 29 .Additionally, recent findings indicate that the outer membrane is also capable of carrying mechanical loads 23 , opening the possibility that the synthesis and transport of outer membrane components may be sensitive to mechanical stress.
One limitation of this study is the variance in fluorescence response among cells at each applied load magnitude.The magnitude of variance follows the same pattern as the degree of control of mechanical loading of each cell (most control in the microfluidic system, less in compression loading, even less in hydrostatic pressure), suggesting that variation in mechanical stresses experienced by individual cells is the primary cause.Even with the most controlled loading system it is likely that there remains variance from cell to cell that may be caused by cell physiology or natural variance in cell envelope stiffness within the population.Further study would be required to understand the sources of this variance in more detail.Despite this limitation, the large number of cells examined makes it possible to detect meaningful trends relating applied mechanical load to VxrAB signaling.
The mechanosensitive nature of VxrAB suggests potential applications in the field of synthetic biology.Gene regulatory mechanisms that respond to the external environment are key tools in the field of synthetic biology.Systems that respond to target chemicals, light, temperature, and pH have been used to control synthetic gene circuits 37 .Our work demonstrates that two-component systems can respond to mechanical stress in the cell envelope, providing an additional mechanical mechanism for stimulating gene circuits for synthetic biology applications.

Methods
Microfluidic device manufacturing.The methods for microfluidic device manufacturing used here are similar to what is described in our previously published works 18,32 except the etch depth was reduced to accommodate the smaller cell dimensions of Vibrio cholerae as compared to E. coli examined in prior work.Fused silica wafers (100 mm diameter and 500 µm thick, WF3937X02031190, Mark Optics, Santa Ana, CA, USA) were patterned using Deep UV photolithography in the cleanroom facility at Cornell NanoScale Facility Science and Technology Facility (Ithaca, NY, USA).Clean fused silica wafers were first coated with ~ 55 nm of chrome using the AJA Sputter Deposition Tool (AJA International, Scituate MA, USA).The Gamma Automatic Coat-Develop Tool (Suss MicroTec Gamma Cluster Tool, Garching Germany) was then used to apply a ~ 60 nm coat of antireflective coating (ARC, DUV 42P, Brewer Science, Rolla, MO, USA) and ~ 510 nm coat of photoresist (UV210, MicroChem, Westborough, MA, USA).The photoresist was exposed to our custom microfluidic device pattern using the ASML Deep UV stepper (Veldhoven Netherlands), then the photoresist was immediately developed using the Gamma Automatic Coat-Develop Tool.The pattern was transferred from the photoresist to the antireflective coating using plasma etching in the Oxford 82 Tool (Oxford, Abingdon, UK), then transferred to the chrome layer with plasma etching using the Plasma-Therm 770 ICP tool (Plasma-Therm St. Petersburg FL, USA).Any remaining anti-reflective coating was removed with a plasma oxygen clean in the Oxford 82 Tool.Finally, the pattern was transferred to the fused silica with plasma etching using the Oxford 100 Tool (Oxford, Abingdon, UK).The remaining chrome was removed using a wet chemical bath.Through-holes were laser-etched at the microfluidic device inlets and outlets using a Versalaser (VLS3.50,Universal Laser Systems, Scottsdale, AZ, USA).
The wafer feature dimensions were characterized using a profilometer, an atomic force microscope, and a scanning electron microscope.The target etch depth for the channels was at least two standard deviations greater than the cell width (> 0.80 µm) so the cells could flow freely through all of the feeder channels and would only get stuck in the narrow constriction of the tapered channels.The feeder channel etch depth was characterized using a profilometer (P-7, KLA Inc, Milpitas CA, USA).The profilometer tip was too wide to measure the tapered channels, so an atomic force microscopy high aspect ratio tip was used to measure the etch depth of tapered channels (Veeco Icon Bruker, Billerica MA, USA).Channel etch depth was 0.88 ± 0.03 µm.A scanning electron microscope (Zeiss Ultra 55 SEM microscope, Oberkocken Germany) was used to measure the tapered channel inlet width (1.48 ± 0.07 µm) and outlet width (0.35 ± 0.02 µm).The tapered channel inlet is wide enough that cells can enter, and the outlet is narrow enough to prevent cells from flowing out.
The patterned wafers were bonded to 100 mm diameter and 170 µm thick fused silica cover wafers (WF3937X0073119B Mark Optics, Santa Ana CA, USA).Cover wafer thickness was chosen to be the same thickness as standard coverslips.Both the patterned wafers and the cover wafers were MOS/RCA cleaned before bonding.Wafers were gently hand bonded, then underwent nitrogen annealing for 5 h at 1100 °C.Wafers were allowed to sit at least one week before use to let the bond mature.

Microfluidic device design.
The microfluidic device design has been described previously in our published works 18,32 .Fluid pressure pushes bacterial cells into narrow tapered channels, a process we refer to as extrusion loading (Supplementary Fig. S6).The cell experiences deformation and mechanical stress and strain in the cell envelope as it is constricted by the channel walls.The fluidic pressure is higher at the wide inlet of the channel and lower at the narrow outlet of the channel, pushing the cells towards the outlet.We refer to the difference in pressure between the wide inlet and the narrow outlet of the tapered channel as the pressure differential (ΔP).Cells in tapers with a greater pressure differential experience a greater magnitude of mechanical loading, travel further into the tapered channel, and deform more than cells in tapers with a lesser pressure differential.The distance the cell travels into the tapered channel is dependent on the magnitude of the pressure differential, initial cell width, cell stiffness and cell width prior to entrance into the tapered channel.www.nature.com/scientificreports/Hundreds of cells under multiple loading magnitudes are observed during each experiment.Twelve sets of five tapered channels are put in parallel and connected by a bypass channel (Supplementary Fig. S6).Pressure is highest at the inlet of the bypass channel and lowest at the outlet of the bypass channel due to pressure loss from hydraulic resistance; therefore, the tapered channels near the bypass inlet and outlet experience the greatest pressure differential (ΔP) (Supplementary Fig. S6).Ten bypass channels are put in parallel and connected by feeder channels to a single entry port for the microfluidic device (Supplementary Fig. S6).While trapped in the tapered channels cells remain alive for hours and are observed elongating and dividing.
Loading cells into the microfluidic device.Fluid pressure was used to load the cells into the microfluidic device.Pressure was generated using a PneuWave Pump (CorSolutions, Ithaca NY, USA).PEEK tubing (Idex 360 µm OD × 150 µm ID, Lake Forest IL, USA) was attached to the PneuWave pump.Isopropanol alcohol was flushed through the PEEK tubing for 5 min for sterilization, then M9 minimal media was flushed through the tubing for 5 min to clear the isopropanol alcohol and prepare the tubing for cells.A magnetic connector lever arm (Fluidic Indexing Probe, CorSolutions, Ithaca NY, USA) and gasket (N-123-03 IDEX, Lake Forest IL, USA) were used to form a compression seal connecting the PEEK tubing to the microfluidic device.M9 minimal media was run through the microfluidic device to pre-wet all of the channels and push out any bubbles.
The tubing and connector were disconnected from the microfluidic device, and cell culture was flushed through the tubing at 60 kPa applied pressure for 5 min.The tubing was then reattached to the microfluidic device, and cells were flowed into the device.Applied pressure was maintained at 60 kPa for the remainder of the experiment.Imaging began two hours after all pressure levels in the microfluidic device were loaded with cells.The two-hour time point was determined from preliminary experiments; the fluorescent signal increased from the start of the experiment until 2 h, then remained constant from 2 to 6 h.

Microfluidic device hydraulic circuit pressure calculations.
The fluidic pressure at the tapered channels within the microfluidic device could not be directly measured, so it was calculated using the pressure at the entry of the device (measured with the PneuWave Pump) and a hydraulic circuit model with the Hagen-Poiseuille law (Eq.1).ΔP is the difference in pressure between upstream end and downstream end of a channel, Q is the flow rate, and R h is the hydraulic resistance of the channel.
The hydraulic resistance of each linear channel in the microfluidic device was determined individually using either Poiseuille flow (Eqs. 2 and 3) or Plane Poiseuille flow (Eq.4) where µ is fluid viscosity (assumed to be the viscosity of water = 8.9e−4 Pa s), L is the length of the channel, A is the area of the channel cross-section, r is the hydraulic radius, P is the perimeter of the channel cross-section, and H is the height of the channel.Poiseuille flow (Eqs. 2 and 3) was used for channels where the ratio of the channel cross-section width to channel crosssection height was less than 20, and Plane Poiseuille flow (Eq.4) was used for channels where the ratio was greater than 20.
The hydraulic resistance of the entire microfluidic device was determined by combining the hydraulic resistance of all of the individual linear channels.Channels in parallel were combined using Eq. ( 5) where R Total is the combined hydraulic resistance of channels 1 to n.
Segments in series were combined using Eq. ( 6) where R Total is the combined hydraulic resistance of channels 1 to n.
Due to the complex geometry of the devices, these hydraulic circuit calculations were performed using a custom script in MATLAB (v.2019a, Mathworks, Natick, MA, USA).
Hydraulic resistance of tapered channels with cells.When a cell occupies a tapered channel, the cell partially blocks fluid flow and thereby causes an increase in the hydraulic resistance of that channel.Since the fluid flow profile around the cell is irregular (the tapered channel has a trapezoidal cross-section and a cell has a circular cross-section), Eqs. ( 2) and (3) were not appropriate.To determine the hydraulic resistance of a tapered channel occupied by a cell, we created a model using COMSOL Multiphysics (v 4.3, Stockholm, Sweden).The hydraulic resistance of a tapered channel occupied by a cell is an order of magnitude greater than a tapered (1) (3) r = 2A P ; (4) R h = 12µL AH 2 . (

Figure 1 .
Figure 1.VxrAB signaling responds to mechanical stress from extrusion loading.(A) When activated, inner membrane histidine kinase VxrA phosphorylates response regulator VxrB, which regulates gene expression of regulons that include murJ and vxrAB.Transcriptional msfGFP fusions for murJ and vxrAB were used as reporters for VxrAB signaling.ShyA function promotes VxrA signaling through an unknown mechanism.(B) Extrusion loading involves forcing bacteria under fluid pressure into tapered channels.(top) Cells deform more and experience greater mechanical loading at a greater pressure difference in tapered channels.(bottom) ΔcrvA P murJ :msfGFP Vibrio cholerae expressing GFP while trapped within tapered channels are shown (a "set" of tapered channels with the same differential pressure is shown).(C) Vibrio cholerae are crescent-shaped.(D) Vibrio cholerae with crvA deletion are rod-shaped.(E) Single cell fluorescence of P murJ :msfGFP cells vs pressure difference.Solid line is a linear regression.(F) Single cell fluorescence of ΔcrvA P murJ :msfGFP cells (pink) and (G) ΔcrvA ΔvxrB box P murJ :msfGFP cells (blue) vs pressure difference.Solid lines are linear regressions.Slope of ΔcrvA P murJ :msfGFP cell (pink) is greater than the slope of the ΔcrvA ΔvxrB box P murJ :msfGFP cells (blue) (p < 0.001).(H) The vxrA H301A mutant with modified histidine kinase.(I) Single cell fluorescence of ΔcrvA P vxrAB :msfGFP cells vs pressure difference.Solid line is a linear regression. https://doi.org/10.1038/s41598-023-40897-w