Patch-clamp electrophysiology of giant plasma membrane vesicles (GPMVs) measures MscL pressure sensitivity.

A) Schematic representation of GPMV formation. Human bone osteosarcoma epithelial cells (U2OS) expressing MscLGFP were treated with N-ethylmaleimide (NEM) vesiculation agent to form GPMVs that contain MscLGFP. In certain cases, exogenous amphiphiles, such as poloxamers, were added to U2OS cells during the vesiculation process to generate GPMVs with modified membrane compositions. B) Microscopy images of GPMVs from non-transfected cells (bottom) or GPMVs from cells transfected with MscLG22S tagged with GFP (top). DIC images (left) and GFP epifluorescence (right) are shown. The scale bar is 10 μM. C) Representative electrophysiology traces of MscL pressure sensitivity over time. MscLG22S (top) recordings show MscL activation in response to pressure ramp (bottom). MscL activation sensitivity is defined as the pressure at which the first MscL channel activates (ex.∼65 mmHg for the shown GPMV). GPMVs from non-transfected cells (no MscL) exhibit no change in current in response to pressure (middle). Recordings were performed at -10 mV in 10 mM HEPES pH 7.3, 150 mM NaCl, 2 mM CaCl2. Pipette internal solution consists of 10 mM HEPES pH 7.3, 100 mM NaMES, 10 mM Na4BAPTA, 10 mM NaCl, 10 mM EGTA. D) Example MscLG22S single channel events recorded from GPMVs derived from U2OS cells at different voltages. E) Resulting single channel current-voltage relationship used in determination of the MscLG22S conductance. The current amplitude at each voltage (shown in D) is plotted and the slope of the linear fit was used to determine channel conductance. MscLG22S conductance is 1.113 nS (95% confidence 1.020 to 1.207 nS), n = 17 individual recordings at various holding potentials.

Poloxamer P124 decreases MscL sensitivity and modulates KA and kc.

A) Schematic and chemical structure of poloxamer P124. Poloxamer may be found in a hairpin or transverse conformation within the membrane. B) MscLG22S activation is desensitized in the presence of P124. MscLG22S activation is defined here as the pressure where the first MscLG22S channel is activated. The pressure at first channel opening was measured for GPMVs prepared in solutions containing 0 – 0.1% wt/vol P124, n ≥ 13 independent GPMVs were recorded. C) Images of micropipette aspiration of GPMVs measuring membrane mechanical properties. Micropipette aspiration is used to measure membrane areal changes in response to applied tension. The area expansion modulus (KA) and bending rigidity (kc) of GPMV membranes treated with poloxamer were measured. D) Sample aspiration curve from a micropipette aspiration experiment. A stress-strain curve was plotted to calculate KA and kc of GPMV membranes using micropipette aspiration techniques at high and low tension, respectively. E, F) KA and kc decrease in the presence of P124. GPMV KA and kc were measured using micropipette aspiration techniques after inclusion of P124 in the GPMV formation buffer. kc and KA were calculated at low- and high-pressure regimes, respectively, where KA is corrected for relaxation of thermal undulations using mean kc. P124 % wt/vol is as listed, pH 7.3. Bath and pipette solution consist of 10 mM HEPES pH 7.3, 150 mM NaCl, 2 mM CaCl2 in micropipette aspiration studies. Mean activation pressure, KA, kc are plotted with error bars that represent standard error of the mean. p-values in B, E, and F were generated by ANOVA using Dunnett’s multiple comparisons test compared to no poloxamer. **** p ≤ 0.0001, *** p ≤ 0.001, * p ≤ 0.05, non-significant (ns) p >0.05, n>10.

Membrane area expansion modulus and bending rigidity are good predictors for MscL pressure sensitivity.

A, B, C) Membrane KA and kc, but not fluidity, predict MscLG22S activation pressure. GPMVs were treated with various poloxamers or a detergent (C12E8) during formation and membrane properties (KA, kc, and fluidity) and MscLG22S pressure sensitivity were measured. Mean KA (A), kc (B), and fluidity (via anisotropy) (C) are reported as a function of MscLG22S activation pressure when treated with the same amphiphile at a given concentration. Decreased anisotropy indicates a relative increase in membrane fluidity. For all plots, Pearson’s r was used to determine correlation strength, n > 10. D) Summary of key correlations (KA and kc) between GPMV membrane properties and MscLG22S activation sensitivity across a variety of amphiphile additives.

Poloxamer does not impact MscL through pore occlusion or membrane thickness changes but likely through a reduction in interfacial tension.

A) Poloxamers could prevent MscLG22S activation through pore occlusion. B) MscLG22S pore size, as inferred by conductance calculations, does not change in the presence of P124. MscLG22S conductance in the presence of P124 is plotted; the slope of the linear fit was used to determine channel conductance. C12E8 serves as a non-polymeric control. C) MscLG22S conductance quantification in the presence of P124 and C12E8. Mean conductance and 95% confidence interval are plotted from the linear fit shown in (B). D) Chemical activation of MscLG22C. Fluo4 Ca2+ imaging of the internal space of the vesicle reveals that Ca2+ influx can be detected when MscLG22C is activated with MTSET but not in the presence of MscLG22S. DIC images show GPMV outline. E) P124 does not prevent MscL activation. GPMVs containing MscLG22C were formed in the presence of various P124 concentrations as indicated. Fluo4 fluorescence was quantified after GPMV incubation with MTSET and 5 mM Ca2+ and is plotted as integrated intensity / vesicle area. Mean is plotted with standard error off the mean, and the dashed horizontal line is the signal for MscLG22S in 0% P124 (negative control, (-)CT). *** p ≤ 0.001, non-significant (ns) p > 0.5, n > 30 vesicles per condition. p-values were generated by ANOVA using Dunnett’s multiple comparisons test compared to no poloxamer. F) Poloxamers may prevent MscLG22S activation by increasing membrane thickness. G) Schematic of coarse-grain molecular dynamics simulations of DOPC bilayers (translucent) containing P124 (opaque). H) Membrane surface area increases with increasing poloxamer concentration. I) P124 does not increase membrane thickness in molecular simulations. Error bars represent standard error of the mean, n = 3. J) Poloxamers may affect MscL channel opening by reducing interfacial tension in the membrane when under tension. K) Distribution of the PEO blocks normal to the lipid bilayer in P124-containing membranes as determined by coarse-grained molecular dynamics simulations. The normal coordinate at 0 is defined as the middle of the lipid bilayer, and the approximate location of the hydrophobic region is shaded dark gray and the hydrophilic region is shaded light gray. Under tension, the PEO groups adsorb to the bilayer, suggesting adsorption is a thermodynamically favorable process that lowers the interfacial tension.

TREK-1 activation also correlates with membrane properties.

A) Mouse TREK-1 single-channel potassium currents are observed in GPMVs in response to pressure. TREK-1 currents were measured using patch-clamp electrophysiology with an integrated pressure controller in asymmetric potassium conditions, pH 7.3. B) TREK-1 is outward rectifying with a conductance of 140 pS. TREK-1 current was quantified under various holding potentials. TREK-1 currents were larger under negative voltages and the slope of the I/V relationship under negative voltages was used to measure outward rectifying conductance. Individual electrophysiology measurements are plotted, n > 15. C) TREK-1 activation pressure is increased in the presence of a poloxamer (0.1% P124) and a detergent (1 μm C12E8). Mean activation pressure is plotted and dots represent individual measurements, error bars represent standard error of the mean. p-values were generated by ANOVA using Dunnett’s multiple comparisons test compared to no poloxamer. **** p ≤ 0.0001, *** p ≤ 0.001, n>10. D, E) Membrane KA and kc predict TREK-1 activation pressure. GPMVs were treated with P124 at two concentrations and C12E8 at the highest concentration and TREK-1 activation or membrane mechanical properties were measured. Mean KA or kc and TREK-1 activation pressure were plotted. F) Membrane fluidity does not predict TREK-1 pressure sensitivity. Mean membrane anisotropy and TREK-1 activation pressure were plotted. Decreased anisotropy indicates a relative increase in membrane fluidity. Pearson’s r in (D-F) was determined to measure correlation strength, error bars represent standard error of the mean, n > 10 for each measurement.

MscLG22S reduces the pressure required for activation compared to MscLWT but does not change channel conductance.

A) MscL activation pressure is reduced in the G22S mutant. The pressure at first channel opening was measured to quantify MscL activation sensitivity. Individual points represent independent GPMV measurements. Mean is plotted with error bars which represent standard error of the mean. B) Conductance calculations for MScLG22S and MscLWT demonstrate that MscL channel size and selectivity is not affected by the G22S mutation. MscLG22S conductance = 1.113 nS (95% confidence interval 1.020 – 1.207). MscLWT conductance = 0.9033 nS (95% confidence interval 0.5761 – 1.231)

U2OS cell lines stably expressing MscLG22S tagged with GFP.

MscLG22S expressed as a fusion protein with GFP was transfected into U2OS cells and selected using geneticin for 4 weeks. MscL is localized to various membranes within the cell including the plasma membrane.

GPMVs budding off U2OS cells.

DIC micrograph demonstrates Giant Plasma Membrane Vesicles (GPMVs) formation by budding off from adherent mammalian cells. This process creates large diameter, bilayer vesicles comprised of the cellular plasma membrane without cytoskeletal attachments or intracellular membranes.

Differing GPMV formation methods yield similar MscL activation pressures, but crosslinking reagents may reduce the probability of channel opening.

GPMV formation is commonly performed using two different methods, NEM or PFA/DTT. While PFA/DTT is the most commonly used reagent mixture, this method induces crosslinking of lipids and proteins that is not amenable to mechanosensitive channel activation. In the presence of PFA/DTT, after >30 attempts over three independent GPMVs preparations, only three instances of functional MscL channels were observed. In contrast, NEM preparation allowed reliable observation of MscL activation. The two methods for GPMV formation did not demonstrate different activation pressure thresholds of MscL. p-values to determine significance between means was measured using Student’s t-test. non-significant (ns) p > 0.05, n > 30.

Schematic and properties of amphiphiles used in this study.

We used various synthetic amphiphiles to exogenously alter membrane mechanical properties. These amphiphiles belong to two classes, poloxamer (Pluronic) and small nonionic detergent. Poloxamer is a class of triblock copolymers which consist of a polypropylene oxide hydrophobic block sandwiched by two polyethylene oxide hydrophilic blocks. Poloxamers have similar properties to surfactants in their ability to insert into bilayers but have much larger molecular weights. The detergent C12E8 we used has a much smaller molecular size in comparison to poloxamer.

Sample electrophysiology data analysis workflow.

We calculated MscLG22S pressure sensitivity using an integrated pressure controller to quantify the pressure at which the first MscLG22S channel activated. Higher pressures open more than one channel as demonstrated by multiples of original channel current. Inset demonstrates stochastic channel openings due to the applied pressure. We applied enough pressure to pop the GPMV membrane to ensure the pressure was high enough to activate mechanosensitive channels in the GPMVs and to enable the quantification of membrane stability.

Example results of full traces of MscLG22S and no MscL electrophysiology recordings.

A) MscLG22S exhibits stochastic activation behavior in response to pressure in GPMVs. When applied pressure is returned to 0 mmHg, MscLG22S closes and current returns to baseline. This observation suggests that the change in current is due to channel activation and not membrane rupture. B) No MscL control demonstrates no channel currents under MscLG22S activation conditions. Channel currents are not observed up to pressures high enough to rupture the membrane. The open/close pressure ramp in (A) was only used to demonstrate channel closure after pressure reduction. All other recordings were performed using the rate of pressure change in (B) for all MscLG22S activation calculations.

Lysophosphatidylcholine (LPC) activation confirms MscLG22S behavior in GPMVs formed from U2OS cells.

A) LPC inserts into membranes and induces local changes in curvature on short timescales. These changes in curvature are known to reduce the pressure required to activate MscL (Nomura et al., 2012). On longer timescales, LPC equilibrates across bilayer leaflets, and this equilibration neutralizes curvature such that there is no expected effect on MscLG22S activation. B) MscLG22S activation pressure is reduced upon LPC introduction. After incubation for >1hr, LPC no longer alters MscLG22S activation pressure as expected. p-values were generated by ANOVA using Dunnett’s multiple comparisons test compared to no poloxamer. ** p ≤ 0.01, non-significant (ns) p > 0.05, n> 10.

MALDI-TOF mass spectra of pure poloxamer A) P124 and B) P188. Samples in panes C–E are poloxamer-integrated GPMVs. In separate experiments, data was acquired in the higher mass regions, m/z 1000–4000 and m/z 4000–12500 to detect C) P124 and D) P188 in GPMVs, respectively. Also, data was acquired in a lower mass region m/z 500–1000 to detect E) GPMV lipids of the P124-GPMVs and F) GPMV lipids of the P188-GPMVs. The inset in pane E is the matrix, CHCA, collected in the mass range m/z 500–1000.

GPMV diameter increases in the presence of various poloxamers and a detergent.

A) DIC phase contrast micrographs demonstrate the effect poloxamer and detergent (C12E8) on GPMV size. B) GPMV diameter was quantified for all GPMVs large enough to quantify. Individual points represent the diameter of a single GPMV, mean and standard error of the mean are plotted. p-values were generated using ANOVA Dunnett’s test of multiple comparisons compared to no poloxamer. **** p ≤ 0.0001, * p ≤ 0.05, non-significant (ns) p >0.05, n > 10.

Poloxamer effect on membrane stability.

We quantified membrane rupture pressure during our electrophysiology recordings of MscLG22S activation to measure if various poloxamers exhibited a strengthening effect on the membrane as has been previously discussed (Zaki and Carbone, 2017). The effect of P124 (A), P188 (B), P407 (C), and P184 (D) were measured, and we observed a membrane strengthening effect in the presence of some poloxamers. However, P407 exhibited a decrease in membrane stability which we hypothesize is due to the large molecular weight of this poloxamer. Each point represents the rupture pressure of a single GPMV, mean is plotted, and error bars represent standard error of the mean. p-values were generated by ANOVA using Dunnett’s test for multiple comparisons compared to no poloxamer (0%). * p ≤ 0.05, non-significant (ns) p > 0.05, n >10 for most samples except P407 (n > 3) which was prone to popping prior to the initiation of the electrophysiology recording.

P124 prevents MscLG22S activation at high concentrations.

Some electrophysiology recordings contained membranes that remained stable in the presence of high pressure (> -80 mmHg), yet MscLG22S did not activate. Recordings in which we observed activation of MscLG22S are presented as a percent of all recordings for a given condition (% active). The dashed line represents the percent of recordings in which MscLG22S openings were observed in the absence of poloxamer. n>12 for all conditions. p-values were generated using Fisher’s exact test comparing the distribution of values to 0% P124, * p ≤ 0.05, non-significant (ns) p >0.05, n>10.

Poloxamer incubation time alters poloxamer effect on membrane properties.

Incubation time noticeably, but does not significantly, alter membrane KA in this experiment. GPMVs were incubated with 0.1% wt/vol P188 at 4 °C for either 1, 3, or 4 hours and membrane KA was measured using micropipette aspiration techniques. p-values were generated by ANOVA using Dunnett’s test for multiple comparisons compared to 1-hour, non-significant (ns) p > 0.05, n > 6 independent GPMVs.

P124 effect on membrane fluidity.

Fluorescence anisotropy was used to measure relative differences in membrane fluidity between GPMV membranes containing various amounts of P124. Decreased anisotropy indicates a relative increase in membrane fluidity. Mean and standard error of the mean are plotted. p-values were generated by ANOVA using Dunnett’s multiple comparisons test compared to no poloxamer, non-significant (ns) p >0.05, n>10.

Poloxamer P124 alters GPMV mechanical properties.

A) P124 decreases KA and increases MscLG22S activation pressure. KA is plotted as a function of MscL activation pressure. B) P124 decreases kc and increases MscLG22S activation pressure. Mean kc is plotted as a function of mean MscLG22S activation pressure. C) Membrane fluidity does not correlate with MscLG22S activation pressure. Membrane anisotropy was used to measure relative changes in membrane fluidity in the presence of P124. Decreased anisotropy indicates a relative increase in membrane fluidity. GPMVs were mixed with diphenylhexatriene (DPH) to a final concentration of 50 μM and fluorescence anisotropy was measured using a fluorimeter with fluorescence polarization capabilities. For all plots, Pearson’s r was calculated to measure correlation strength, n>10.

C12E8 effect on MscL activation pressure, KA, kc, and fluidity.

A) C12E8 increases MscLG22S activation pressure with increasing concentration. MscLG22S activation pressure in the presence of C12E8 in the buffer during GPMV formation was measured using patch-clamp electrophysiology and quantified as the pressure at the first channel activation. B) C12E8 did not prevent MscLG22S activation to the same extent as P124 (Figure 2 – figure supplement 1). The percent of recordings which activated MscLG22S were quantified. p-values were generated using Fisher’s exact test compared to 0% C12E8. non-significant (ns) p > 0.05, n > 10. C12E8 decreased KA (C), kc (D) as measured by micropipette aspiration. Points represent individual GPMV measurements which were not repeated. E) C12E8 noticeably but non-significantly increased membrane fluidity as measured by fluorescence anisotropy, where a decreased anisotropy indicates a relative increase in membrane fluidity. Mean and standard error of the mean for all plots is indicated by the bar and error bars p-values in (A), (C-E) were generated by ANOVA using Dunnett’s test of multiple comparisons compared to no C12E8. **** p ≤ 0.0001, *** p ≤ 0.001, ** p ≤ 0.01, non-significant (ns) p >0.05, n > 10.

MscL activation sensitivity in the presence of various alternate poloxamers.

A-C) MscLG22S activation pressure is altered by poloxamer identity and concentration. The pressure of MscLG22S first channel opening was calculated in the presence of various poloxamers at increasing concentration of poloxamer in the buffer P188 (A), P407 (B), and P184 (C). Individual points represent independent GPMV recordings. Higher concentrations of P407 (B) destabilized the membrane and led to rupture at pressures lower than the expected activation pressure of MscLG22S. For such conditions, no active MscLG22S currents were observed in the small number of P407 GPMVs which were successfully recorded. Mean and standard error of the mean are indicated by the black bar and error bars. The amount of poloxamer in the GPMV formation buffer is indicated in the plot. p-values were generated using ANOVA Dunnett’s test for multiple comparisons compared to no poloxamer (0%). ** p ≤ 0.01, non-significant (ns) p > 0.05, n > 10 for all samples except for P407, the inclusion of which led to membrane instability which prevented MscLG22S activation (B). D-E) MscLG22S activation percentage is modulated by poloxamers. P188 (D) had no depression effect on MscLG22S activation percentage while P407 (E) demonstrated a significant reduction in activation probability. P184 (F) showed a noticeable, but non-significant reduction in MscLG22S activation probability. p-values were generated using Fisher’s exact test compared to 0% poloxamer. * p ≤ 0.05, non-significant (ns) p > 0.05, n > 3.

Other poloxamers effect on KA, kc, and fluidity.

Membrane mechanical properties were measured in the presence of poloxamer molecules P188 and P184. A, B) With increasing P188 concentration, KA is reduced and kc is slightly but non-significantly reduced. C) P184 generally decreases membrane KA at and above 0.0001% wt/vol in the GPMV formation buffer. D) P184 exhibits a non-monotonic effect on membrane kc where low concentrations decrease kc, but higher concentrations increase kc. E) membrane anisotropy was measured in the presence of P188, P407 and P184. No significant differences in membrane fluidity were observed in the presence of these poloxamer molecules. Decreased anisotropy indicates a relative increase in membrane fluidity. Individual points represent single GPMV measurements, mean and standard error of the mean are plotted as a black bar with error bars, pH 7.4. p-values were generated using ANOVA Dunnett’s test for multiple comparisons compared to no poloxamer (0%). *** p ≤ 0.001, ** p ≤ 0.01, * p ≤ 0.05, non-significant (ns) p >0.05, n > 10.

A non-ionic detergent decreases MscL sensitivity and alters membrane properties similarly to poloxamer.

A) C12E8 decreases area expansion modulus and increases MscLG22S activation pressure (Figure 3-figure supplement 1A,C). KA was measured using micropipette aspiration. Mean KA is plotted as a function of mean MscLG22S activation pressure for each concentration of C12E8. B) C12E8 decreases membrane kc and increases MscLG22S activation pressure (Figure 3-figure supplement 1A,D). Mean kc is plotted as a function of mean MscLG22S activation pressure. C) Membrane fluidity and MscLG22S activation pressure positively correlate in the presence of C12E8. Mean fluorescence anisotropy of DPH is plotted as a function of MscLG22S activation pressure. Decreased anisotropy indicates a relative increase in membrane fluidity. For all plots, Pearson’s r was calculated to measure correlation strength, n>10.

Validation of GPMV calcium influx assay.

MscLG22C is a chemically-activatable mutant of MscL which opens in the presence of MTSET through binding to cysteine groups which induces forced hydration of the hydrophobic gate. Calcium and MTSET were added to GPMVs formed from cells incubated with Fluo4 Ca2+ indicator and Ca2+ influx was detected as an increase in fluorescence and was quantified as integrated intensity/area. GPMVs formed from non-transfected cells (U2OS), MscLG22S-containing GPMVs treated with MTSET, and MscLG22C-containing GPMVs not treated with MTSET (G22CGFP(-MTSET)) did not exhibit increases in fluorescence in response to Ca2+. p-values were generated by ANOVA compared to no MscL (U2OS). ** p ≤ 0.01, non-significant (ns), p > 0.05, n > 10 independent GPMVs.

MscLG22C pressure sensitivity is restored in the presence of DTT.

The chemically activatable mutant, MscLG22C, used in Ca2+ imaging assays was further characterized using patch-clamp electrophysiology to confirm channel activity. Without DTT (top) the channel is not mechanically activatable as described previously (Levin and Blount, 2004) due to disulfide bonds which prevent channel opening. In the presence of DTT, channel mechanosensitivity is restored and stochastic channel activation is observed with increasing pressure.

Coarse-grain molecular dynamics simulations confirm the effect of P124 on membrane elasticity.

Membrane KA was measured at increasing concentrations of P124 and was found to decrease similar to experimental results. Error bars represent standard error of the mean, n=3.

U2OS cells stably expressing mTREK-1.

mTREK-1 tagged with GFP localizes to the membrane as shown in GFP images for transfected cells (bottom). Non-transfected U2OS lines (top) do not show GFP signal. DIC images are shown to demonstrate cell boundary and confluency.

TREK-1 is retained in GPMV membranes.

TREK-1 tagged with GFP is localized to the GPMV membrane in transfected cells (bottom). No GFP signal is detected in GPMVs from non-transfected cells (U2OS, top).

TREK-1 mechanosensitive currents are only observed in GPMVs formed from stable cells expressing TREK-1.

A) We observed TREK-1 activation where single channel currents were ∼4.5 pA at -50 mV. TREK-1 activation was observed in most patches above -30 mmHg. B) To further confirm that we observed TREK-1 channels in our electrophysiology recordings, we performed patch-clamp electrophysiology on GPMVs formed from U2OS cells that were untransfected, referred to as untransfected GPMVs, under identical conditions as our TREK-1 recordings which required higher holding potentials to observe single channels due to the smaller conductance of TREK-1 compared to MscLG22S. We applied enough pressure to pop the GPMV membrane to ensure the pressure was high enough to activate any native mechanosensitive channels in the GPMVs. We did not observe any channels in untransfected GPMVs before membrane popping. Recordings are performed at -50 mV in asymmetric potassium buffer conditions. Bath solution was composed of 10 mM HEPES, 150 mM KCl, 2 mM MgCl2 for excised-patch recordings on TREK-1 with outward rectifying potassium currents. n > 6 recordings on independent patches.