Article Outline

Article Information

PubMed

Related Articles

  • …more

Copyright © 2007 The Biophysical Society. All rights reserved.
Biophysical Journal, Volume 92, Issue 8, 2771-2784, 15 April 2007

doi:10.1529/biophysj.106.098715

Channels, Receptors, and Electrical Signaling

2,2,2-Trifluoroethanol Changes the Transition Kinetics and Subunit Interactions in the Small Bacterial Mechanosensitive Channel MscS

Bradley Akitake*, 1, Robin E.J. Spelbrink†, 1, Andriy Anishkin*, J. Antoinette Killian†, Ben de Kruijff† and Sergei Sukharev*, Go To Corresponding Author,  

* Department of Biology, University of Maryland, College Park, Maryland
† Department of Biochemistry of Membranes, Institute of Biomembranes, Utrecht University, Utrecht, The Netherlands

Address reprint requests to Sergei Sukharev, Tel.: 301-405-6923.

1 B. Akitake and R.E.J. Spelbrink contributed equally to this work.

Abstract

2,2,2-Trifluoroethanol (TFE), a low-dielectric solvent, has recently been used as a promising tool to probe the strength of intersubunit interactions in membrane proteins. An analysis of inner membrane proteins of Escherichia coli has identified several SDS-resistant protein complexes that separate into subunits upon exposure to TFE. One of these was the homo-heptameric stretch-activated mechanosensitive channel of small conductance (MscS), a ubiquitous component of the bacterial turgor-regulation system. Here we show that a substantial fraction of MscS retains its oligomeric state in cold lithium-dodecyl-sulfate gel electrophoresis. Exposure of MscS complexes to 10–15vol.% TFE in native membranes or nonionic detergent micelles before lithium-dodecyl-sulfate electrophoresis results in a complete dissociation into monomers, suggesting that at these concentrations TFE by itself disrupts or critically compromises intersubunit interactions. Patch-clamp analysis of giant E. coli spheroplasts expressing MscS shows that exposure to TFE in lower concentrations (0.5–5.0vol.%) causes leftward shifts of the dose-response curves when applied extracellularly, and rightward shifts when added from the cytoplasmic side. In the latter case, TFE increases the rate of tension-dependent inactivation and lengthens the process of recovery to the resting state. MscS responses to pressure ramps of different speeds indicate that in the presence of TFE most channels reside in the resting state and only at tensions near the activation threshold does TFE dramatically speed up inactivation. The effect of TFE is reversible as normal channel activity returns 15–30min after a TFE washout. We interpret the observed midpoint shifts in terms of asymmetric partitioning of TFE into the membrane and distortion of the bilayer lateral pressure profile. We also relate the increased rate of inactivation and subunit separation with the capacity of TFE to perturb buried interhelical contacts in proteins and discuss these effects in the framework of the proposed gating mechanism of MscS.

Introduction

Since the mid-1960s, halogenated alcohols such as 2,2,2-trifluoroethanol (TFE) have been known to exert strong effects on protein secondary structure. More recently, these solvents have found new applications in the study of membrane proteins. Having a lower dielectric constant than water 1, TFE is often chosen as a nonpolar medium for spectroscopic determination of peptide conformations 2,3,4 and helical propensities 5,6. TFE also serves as a nonpolar cosolvent in studies of conformational equilibria and protein folding kinetics 7,8.

Although TFE is fully miscible with water at any ratio, the molecule forms microscopic clusters in aqueous solutions with the highest propensity for aggregation near 30vol.% 1,9. At these concentrations, TFE strongly stabilizes the α-helical and β-sheet structures of many soluble and amphiphilic peptides by reducing solvation of the backbone amide groups thus destabilizing extended coil conformations 10,11. TFE has also been proposed to associate with apolar side chains, providing a nonaqueous matrix for the hydrophobic collapse of polypeptides 12,13,14. TFE was shown to stabilize the secondary and tertiary structures of globular proteins subjected to denaturing agents or elevated temperatures 15. Finally, TFE has been shown to accelerate protein folding 7 and disfavor partially folded intermediates even at low concentrations 8.

In contrast to the stabilizing effects observed in soluble proteins, TFE predominantly destabilizes integral membrane proteins and their complexes. The bacterial potassium channel KcsA has been well studied in this regard. KcsA retains its tetrameric structure in nonionic detergents and even in SDS 16; however, it is completely disrupted into monomers by 20vol.% TFE present in a DDM detergent solution 17. Further increase of TFE to 35vol.% under such conditions leads to a reversible loss of secondary structure 18. Surrounding phospholipids, especially PE, stabilize the liposome-reconstituted KcsA complex against TFE, despite the fact that TFE concentrations above 20vol.% severely perturb membranes themselves 17.

TFE's ability to separate hydrophobic polypeptide chains has been utilized to improve the quality of samples for two-dimensional electrophoresis of membrane protein mixtures 19,20. More recently, a new proteomic approach to identify partners in stably associated detergent-resistant complexes has been designed. In this procedure, a change of protein mobility in gels upon exposure to TFE indicated that the components had altered their oligomeric state 21. Such analysis of the Escherichia coli inner membrane has identified ∼60 oligomeric proteins. One of these proteins is the mechanosensitive channel of small conductance (MscS), a ubiquitous component of the bacterial osmoregulation system and a highly convenient model system for mechanistic studies of mechanosensitive channel gating.

MscS, a product of E. coli mscS (formerly yggB) gene, is a stretch-activated (mechanosensitive) channel that acts as a release valve for small intracellular osmolytes in the event of acute osmotic downshock 22. Purification and reconstitution experiments proved that the channel opens in response to membrane tension transmitted directly trough the lipid bilayer 23,24. Functional patch-clamp analysis of MscS responses to pulses of hydrostatic pressure across the membrane indicate an adaptive multistate behavior, featuring tension-dependent transitions from the resting to open and then to inactivated states 22,25,26. The solved three-dimensional structure of MscS 27 revealed a heptameric assembly of identical subunits, each comprised of three transmembrane helices (TM1–TM3). The C-terminal ends of each subunit contribute to a large, hollow, cytoplasmic domain. The third transmembrane helix (TM3) lines the conducting pore and bears a characteristic kink at the cytoplasmic side 27. The MscS crystal structure laid the groundwork for several hypotheses about its gating mechanism, with proposed conformational transitions of either smaller 28 or larger scale 26,29,30. Thermodynamic analysis of dose-response curves, however, strongly suggested that the lateral protein expansion associated with the opening transition is large (∼8–18nm2) and must involve a substantial rearrangement of interhelical interactions 23,26.

In this work, we studied the oligomerization state and functional behavior of MscS in the presence of TFE. We report the conditions at which oligomeric MscS complexes remain stable in the presence of ionic detergents and the range of TFE concentrations at which breakdown into individual subunits occurs. We provide the first evidence that TFE, at concentrations much lower than those required for subunit separation, changes the equilibrium and transition kinetics between the functional states by reversibly driving the channel into the inactivated state. This new data suggests that TFE can be used for controlled perturbations of interhelical interactions in functional studies of membrane proteins.


Methods

Materials

Electrophoresis setups were purchased from Bio-Rad Laboratories (Emmen, The Netherlands). Lithium dodecyl sulfate (LDS) was purchased from USB (Cleveland, OH). Octylglucoside was obtained from LabScientific (Livingston, NJ). Ni2+ nitrilotriacetic acid agarose was obtained from Qiagen Benelux N.V. (The Netherlands). Anti-his6-C-term antibodies were purchased from Invitrogen (The Netherlands). Isopropyl-β-D-thiogalactopyranoside was obtained from Calbiochem (Los Angeles, CA). 2,2,2-Trifluoroethanol (TFE) was purchased from Merck (Darmstadt, Germany). 1,1,1,3,3,3-hexafluoroisopropanol was purchased from Acros Organics (Deventer, The Netherlands). Coomassie Brilliant Blue G-250 was purchased from ICN Biomedicals (Aurora, OH). Lithium dodecyl sulfate-polyacrylamide gel electrophoresis (LDS-PAGE) gradient gels were cast using a Hoefner SG30 gel maker while nongradient LDS gels were cast on BioRad Protean III casting systems. All other chemicals were of the highest quality commercially available.


Strains and expression constructs

PB111, a plasmid containing MscS with a C-terminal 6His tag, was a gift of Dr. Paul Blount (UT Southwestern, Dallas, TX). MJF465, a triple E. coli mutant (mscL−, mscS−, mscK−) 22, used in our work as a host strain was kindly provided by Dr. Ian Booth (University of Aberdeen, Scotland). The MscS S95C/I97C double mutant was generated with a single pair of complementary primers using a QuikChange mutagenesis kit (Stratagene, La Jolla, CA) and verified using automated sequencing.


Preparation of membrane vesicles

The PB111 construct containing MscS-his6 was transformed and expressed in MJF465 cells 22. Cells were grown from overnight culture in 800ml Luria-Bertani medium at 37°C to an OD600 of 0.6 and induced with 0.8mM isopropyl-β-D-thiogalactopyranoside for 1h. Cells were collected by centrifugation. The cell-pellet was washed with 50ml of 50mM potassium phosphate buffer pH 8 containing 5mM MgCl2 and resuspended in the same buffer. The suspension was passed twice through a French press at 1.1 kbar. Unbroken cells were removed by low-speed centrifugation and membrane vesicles were collected by ultracentrifugation in a Ti60 rotor (45k rpm, 45min, 4°C), resulting in ∼0.6g of cell membranes (wet weight). Membrane pellets were stored at −80°C until either being resuspended in 50mM phosphate buffer pH 8 or used for the purification of MscS-his6.


Purification of MscS-his6

His-tagged MscS was purified essentially as in Sukharev 23. An amount of 0.6g of membrane pellet was dissolved in 8ml of 50mM potassium phosphate buffer pH 8, 300mM NaCl, 20mM imidazole, and 3% (w/v) octylglucoside. This solution was cleared from insoluble particles by ultracentrifugation (45,000rpm, 45min, 4°C). The resulting solution was incubated with 0.5ml Ni2+ nitrilotriacetic acid slurry on ice for 1h. The slurry was poured into a column and eluted by gravity. The gel bed was washed with 10vol.mes of 300mM NaCl, 50mM potassium phosphate buffer pH 8, 20mM imidazole, and 1% (w/v) octylglucoside. Elution was performed stepwise with buffers containing 50, 75, and 200mM imidazole, using two gel-bed volumes for each step. Aliquots were run on an 11% SDS-PAGE gel and stained with Coomassie G-250. Fractions containing purified MscS were pooled and supplemented with 0.1% (w/v) Triton X-100. The protein solution was stored at 4°C.


TFE-induced dissociation of MscS detected by LDS-PAGE

Twenty-microliter samples of either MscS (0.3mg/ml) or a membrane preparation from MJF465 cells containing roughly 4mg/ml total protein were added to solutions of TFE in water for a total volume of 30μl. The samples were incubated at ambient temperature for 1h. Samples were cooled on ice before addition of 7.5μl ice-cold LDS-PAGE gel loading buffer. Samples were run on either 9.5% continuous or 8–18% gradient LDS-PAGE gels. In several experiments, TFE-exposed membrane vesicles were spun down and the TFE-containing buffer was carefully removed before dissolution in LDS.

To facilitate detection of oligomeric MscS, electrophoresis was performed at low temperature. Precipitation of dodecyl sulfate was prevented by replacing sodium dodecyl sulfate with lithium dodecyl sulfate in the gels and buffers. Otherwise, the gels and buffers were identical to those commonly used in SDS-PAGE. Electrophoresis setups, gels, and buffers were chilled before use and cooled continuously throughout each run. Gels were run at 120V until the blue dye-front reached the edge of the gel. Gels were stained with Coomassie Brilliant Blue G-250 in the case of purified protein or subjected to Western-blotting with anti-his6-COOH antibodies in the case of inner membrane vesicles. Precision Plus All-Blue protein standards were from BioRad Laboratories.


Electrophysiology

Patch-clamp recordings of MscS were performed using bacterial strains, equipment, and general techniques as previously described 26. Briefly, PB111, a plasmid construct containing MscS with a C-terminal his6 tag, was transformed and expressed in MJF465 strain 22. Voltage-clamp recordings were taken at +30mV (as measured in the pipette) from excised membrane patches of giant Escherichia coli spheroplasts. Patches and MscS activity were stimulated by reproducible ramps and pulses of negative pressure applied with a high-speed pressure-clamp apparatus HSPC-1 (ALA Scientific, Westbury, NY). Recording was conducted in symmetrical potassium buffer (200mM KCl, 90mM MgCl2, 10mM CaCl2, and 5mM HEPES titrated to pH 7.4 with KOH). TFE solutions were created by adding 99+% TFE (Sigma, St. Louis, MO) to the recording buffer for final concentrations of 0.5, 1.0, 2.0, 3.0, and 5.0vol.%. TFE solutions were made fresh before each experiment and solutions older than 3h were discarded.

Membrane patches were exposed to TFE from the cytoplasmic (bath) or periplasmic (pipette) faces. Exposure to TFE from the bath occurred after establishment of a gigaOhm seal and patch excision. Recording buffer in the bath chamber (∼4ml) was replaced with three chamber volumes of TFE solution through perfusion. The total time of perfusion was 3min, after which the system was allowed to rest for an additional 3min before stimulation. After cytoplasmic exposure, TFE could be “washed out” using the same perfusion technique with recording buffer replacing the TFE solution. Exposure to TFE from the pipette was accomplished by filling the electrode with TFE solution (1–5vol.%) behind a 3–5mm plug of pipette solution with 300mM sucrose to delay the onset of exposure. This diffusion-limited delay (2–10min) provided time to take control measurements.


Data collection and analysis

Axon pClamp 9.2 software (Axon Instruments, Foster City, CA) was employed to record integral or single-channel current with a bandwidth of 5–10kHz at a sampling rate of 30kHz. The pClamp software was also used to control the pressure application via output commands to the pressure clamp in episodic stimulation mode. Two-channel recordings of current and pressure versus time were then analyzed with Axon Clampfit 9.2. The maximal current (Gmax) achieved by the MscS population was calculated from traces as the average conductance after the pressure ramp reached its plateau. The midpoint pressure of activation (p1/2) was identified as the pressure at which the MscS population reached 1/2 Gmax. Fitting of the inactivation and recovery kinetics was also performed in Clampfit using built-in fit protocols. A standard exponential function with one or two terms was employed with a Levenberg-Marquardt search method.


Hydrophobicity analysis of MscS surfaces

The crystal structure of MscS (1MXM.pdb) 27 was used for mapping the hydrophobic and hydrophilic areas on the solvent-accessible surfaces of the entire protein. Estimations of the atomic solvent-exposed areas were performed using the web-based GETAREA program 31 with a probe radius of 1.4Å. The hydration energy was computed as the product of the exposed area for each individual atom and the corresponding atomic solvation energy parameter of Eisenberg 32. Hydration energies per amino-acid residue were introduced into the PDB structure file using the PDBAN program custom written in MatLab (The MathWorks, Natick, MA). The solvation energy density was mapped on the MscS solvent-accessible surface and visualized with color-code using VMD 32,33.



Results

TFE-induced dissociation of MscS oligomers

To assess the stability of MscS oligomers, the protein, either as a membrane preparation containing MscS-his6 or in purified, detergent-solubilized form, was incubated with varying concentrations of TFE before separation by LDS-PAGE. To assign the multimeric state of the gel-separated complexes, we attempted two sets of molecular weight markers. The first set was a commercial Precision Blue set (Bio-Rad) consisting of fully denatured soluble proteins (left side on all gels). As a second set we utilized disulfide-crosslinked subunits of the MscS S95C/I97C double cysteine mutant that formed ladders of products ranging from monomers to heptamers under nonreducing conditions (right side, Figure 1A).

Display large version of this figure
Figure 1
(A) Migration of MscS-6his and MscS-his6 S95C/I97C in E. coli membranes as visualized on anti-his6-C-term Western-blot of an 8–18% LDS-PAGE gradient gel. Lane 1 contains membrane vesicles of MJF465 overexpressing MscS-his6. The sample was kept on ice after exposure to LDS gel loading buffer. Lane 2 shows the same sample after boiling for 5min upon addition of LDS-PAGE gel loading buffer. Lane 3 was loaded with membranes of MJF465 overexpressing MscS-his6 S95C/I97C. The sample has undergone the same treatment as that in lane 1. Lanes 4 and 5 show MscS-his6 S95C/I97C loaded in membrane vesicles after boiling without and with 14 +mM of DTT, respectively. Positions of soluble molecular weight markers indicated on the left are used to provisionally designate the band positions. Marker positions based on disulfide-crosslinked MscS subunits providing better estimations for actual molecular weights are shown on the right with calculated molecular weights in parentheses (kDa). For reasons of legibility the pentamer and hexamer have been omitted. (B) TFE-induced dissociation of MscS-6his in E. coli membrane vesicles as visualized by anti-his6-C-term Western blotting on a continuous 9.5% LDS-PAGE gel. Lanes 1–8 were loaded with membrane vesicles of MJF465 overexpressing MscS-his6 that were incubated for 1h at room temperature with the indicated percentages of TFE. Densitometry indicated that the total protein intensity in lane 8 is reduced to a third compared to lane 1, presumably due to monomer aggregation. A molecular weight marker is shown on the left. (C) TFE-induced dissociation of purified oligomeric MscS-6his in 1% (w:v) octylglucoside and 0.1% w:v TX100 on a continuous 9.5% LDS-PAGE gel. Aliquots of purified MscS-his6 protein solution (0.3 mg/ml) were incubated with the indicated percentages of TFE for 1h at room temperature and analyzed by LDS-PAGE. The gel was stained with Coomassie Brilliant Blue G-250.

Electrophoresis on E. coli membranes overexpressing MscS-his6 was performed using a gradient-gel to allow for adequate resolution in the high-molecular-weight region. LDS-PAGE followed by Western-blotting with anti-his6-C-term antibodies revealed three bands (Figure 1A, lane 1). According to the soluble marker scale (left side), the upper band ran at 300 kDa, the second, most intensive band appeared to be close to 250 kDa, and one lightly stained band at 25 kDa. Boiling the sample before electrophoresis produced a single band of monomeric MscS at 25 kDa (Figure 1A, lane 2).

Since the mobility of membrane proteins in dodecyl-sulfate gels may deviate considerably from that of soluble proteins, electrophoresis standards made of soluble proteins may not provide accurate estimations of molecular weight. Therefore we utilized a double-cysteine mutant of MscS, which spontaneously cross-links under ambient atmospheric oxygen, to compare the migration patterns of known covalent homooligomers of MscS and assess the oligomeric state of the observed high-molecular-weight bands in unboiled MscS samples. Figure 1A, lane 3, shows that the covalent oligomers migrate mainly as two bands at the same location as the regular MscS oligomers. When the double-cysteine mutant was boiled before loading, a ladder of denatured, covalent oligomers was observed (Figure 1A, lane 4). The exact sequence-based molecular weights for these bands are presented in parentheses on the right side of the gel. The difference between the two scales shows that in an 8–18% polyacrylamide gel, denatured MscS monomers and dimers run slightly faster than soluble proteins of similar sizes, whereas larger cross-links (4×–7×) migrate slower. As expected, boiling the double-cysteine mutant in the presence of DTT caused most of the higher MW bands to disappear and the monomer band to increase in intensity (Figure 1A, lane 5).

We presume that the positions of covalently cross-linked oligomers of MscS itself (Figure 1A, lane 4) give more reliable estimations of MW than the soluble protein standards. Migration of the bands in this sample suggest that the upper band in lanes 1 and 3 represent intact heptamers, whereas the most intensively stained band near the 250 kDa soluble marker arises from tetramers of MscS subunits that partially retain tertiary structure. Therefore, to interpret these data, we propose assignment of molecular weights according to the disulfide-cross-linked multimers of MscS (Figure 1A, right). Using this interpretation, heptameric MscS is observed to run at a higher molecular weight than its covalently-linked, denatured heptamer. This result may seem surprising because compactly folded (nondenatured) proteins usually migrate in gels faster than their denatured counterparts. However, native MscS contains a bulky cagelike C-terminal domain, a feature that may cause the native form to migrate slower than the denatured protein.

To test whether MscS oligomers can be dissociated by exposure to TFE, membrane vesicles of a strain overexpressing MscS-his6 were incubated with TFE for 1h at ambient temperature, before being subjected to electrophoresis on continuous LDS-PAGE gels. Figure 1B shows that the upper bands disappear from the gel after exposure to TFE while a monomeric band appears. Both oligomeric forms of the protein disappear at concentrations of TFE >10vol.%, although some signal remains at high molecular weight. This residual signal may be the result of MscS aggregation. Aggregation may also explain the relatively low intensity of the monomeric band since such an effect was observed previously for KcsA upon exposure to high concentrations of TFE 17. To verify that the observed decomposition of MscS complexes to monomers is specifically due to the presence of TFE, but not a result of the combined action of TFE and LDS, in a separate experiment we pelleted the TFE-exposed membranes and carefully removed the TFE-containing buffer before adding the LDS sample buffer. This procedure led to a dilution of the residual TFE by at least 10-times. The resultant pattern of bands in the gel was similar to that in Figure 1B showing a breakdown between 10 and 15vol.% TFE (data not shown). This suggests that TFE present around and inside the membrane is, by itself, capable of disrupting intersubunit interactions in MscS.

To establish whether the effect of TFE on the MscS-his6 protein is dependent on the membrane context or it is an intrinsic property of the protein, preparations of purified protein in octylglucoside were also subjected to TFE-induced dissociation. The addition of minor amounts of Triton X-100 (0.1% w/v) was found to improve the stability of the purified protein in LDS-PAGE. Under these conditions, purified MscS migrates as a group of four bands with the most dense one, presumably tetrameric, migrating as the lower oligomer band seen in the membrane preparation gel (Figure 1C, lane 1). Exposure of MscS to 2–6vol.% TFE causes some bands to disappear, while simultaneously increasing the intensity of the heptameric and likely pentameric bands (Figure 1C, lanes 6–11). Apparently, even low amounts of TFE are sensed by the protein, causing it to migrate more slowly, likely due to the effect of “swelling” of hydrophobic cavities and voids 34.

Increasing the TFE concentration to 10vol.% causes complete dissociation of MscS into monomers (Figure 1C, lane 15). In this case no significant loss of protein was observed. The concentration of TFE resulting in a complete dissociation of MscS in detergent micelles was slightly lower than that required to achieve the same result in native membranes. Nevertheless, these concentrations are similar, which suggests that TFE-induced dissociation is an intrinsic property of the protein, which may be slightly stabilized by the lipid bilayer as compared to detergent micelles. The ability to dissociate MscS is not exclusive to TFE, as other alcohols such as 1,1,1,3,3,3-hexafluoroisopropanol produce the same effect on MscS albeit at lower concentrations (data not shown).


TFE effects on MscS activation by pressure ramps

As was shown previously 26, MscS steeply activates in response to 1s duration, linear ramps of negative pressure followed by a plateau (Fig. 2). After reaching saturating pressure, MscS stays open for the duration of pressure stimulus. In control experiments with a large number of channels per patch (50 or more), maximal current (Gmax) of the population reproduced itself within 10%. Using a typical size of patch pipettes, the midpoint pressure of activation (p1/2) varied in the range between 120 and 170mm Hg; however, within each patch, sequential sweeps grouped tightly around a single midpoint with <2% deviation around the mean 26.

Display large version of this figure
Figure 2
Dose-response curves of MscS measured with TFE in the pipette. The tip of the pipette was first filled ∼5mm with recording solution plus sucrose without TFE. The pipette was then backfilled with recording solution plus 5vol.% TFE. Shown are three traces taken immediately after patch formation before TFE could diffuse into the tip (right), after 20 (middle) and after 45min (left) of incubation. The bottom trace shows the linear ramp of the pressure gradient and the scale of the midpoint shift upon TFE application.

We tested the effects of TFE on MscS function in a range of concentrations between 0.5 and 5vol.%. Lower concentrations had no observable effect, whereas higher concentrations of TFE mechanically destabilized patches, thus precluding reliable measurements. Patches exposed to 1vol.% TFE from the pipette (periplasmic side of the membrane) displayed a slight (∼5mm Hg) leftward shift of the dose-response curves without any significant effect on Gmax. The time for development of the leftward shift at this concentration was long (>1h). When the concentration of TFE was increased to 3–5vol.%, larger decreases in p1/2 (leftward shifts) of ∼20mm Hg were observed. The ratio of midpoints for 5vol.% TFE in the pipette, as compared to control, was 0.93±0.04 (n=3). These concentration-dependent shifts occurred reproducibly in the course of 45-min incubations (Fig. 2). During most experiments Gmax, and the corresponding number of active channels in the population, remained essentially constant, falling well within previously established levels of control variability (8–10%).

Perfusion of TFE from the bath (cytoplasmic side of the membrane) even at low concentrations (0.5–2vol.%) invariably shifted p1/2 to the right by ∼10–40mm Hg (Figure 3AC). The peak ratio of midpoints for 2vol.% TFE in the bath, relative to control, was 1.13±0.08 (n=4). The presence of TFE in the bath appears to make the midpoint less stable from trace to trace when compared to controls. In all bath-perfusion experiments the initial and fastest midpoint movement was always to the right. However, in very long experiments (>2h), p1/2 and Gmax were observed to slowly return to the untreated level. We subsequently found that TFE is very volatile and evaporates from a 35mm Petri dish filled with 5vol.% TFE at a rate of ∼2μl/min. In the course of 100min its concentration is thus expected to drop by 80–90%. It was observed that the return of p1/2 and Gmax to control values occurs roughly within this time frame.

Display large version of this figure
Figure 3
Effects of TFE on dose-response curves of MscS measured in excised inside-out patches. TFE was applied to the cytoplasmic side (bath). After equilibration, 1s linear ramps of negative pressure reaching saturating level followed by a 1.5s plateau were applied in 3min intervals. Population responses to 0.5, 2.0, and 5.0vol.% TFE are presented in panels A–C, respectively. A right shift of the activation curve (midpoint change from 156 to 180mm Hg) in the presence of TFE is seen in panel A. Time course of current decline in the traces taken at 6–27min time points (B) indicate MscS inactivation at saturating pressures in the presence of TFE, not observed in controls. The midpoints for these traces are 161 (control), 161 (6min), 163 (9min), 166 (12min), and 179 (27min) mm Hg. Arrow in panel C points to a single channel transiently activated after exposure to 5vol.% TFE followed by complete inactivation. Plot of maximal current achieved by channel population as a percentage of the maximal current before TFE addition (D). The curves in all panels represent inward currents at +30mV pipette potential.

TFE presented to the cytoplasmic side reproducibly decreased Gmax of the MscS population as measured by standard 1s ramps of pressure. A measurable decline (>10%) was observed at 0.5vol.%, with nearly complete silencing of the entire population by a 5vol.% solution (Figure 3C). The concentration of TFE that causes 50% inactivation appears to fall between 0.7 and 1vol.%, due to natural variability in patches and spheroplast preparations. This concentration-dependent process of silencing was not instant but developed within the course of 7–20min (Figure 3D).

To verify that the decrease in Gmax was not due to a drastic change in single-channel conductance, we performed measurements of I-V curves in the presence and absence of TFE (Figure 4A). The single-channel conductance in the presence of 3vol.% TFE in the bath was essentially the same as in control except for a small deviation at strongly depolarizing voltages (−80mV pipette) where the open state current becomes noisy due to the increased presence of subconducting states. The pipette electrode potential has been tested independently in the presence of 5vol.% TFE, and we observed no systematic deviation >±1mV.

Display large version of this figure
Figure 4
Experiments illustrating the nature of Gmax reduction in the presence of TFE. (A) Current-to-voltage relationships for MscS in control (shaded circles) and in the presence of 3vol.% TFE (solid). (B) Effect of additional pressure on a TFE-silenced MscS population. The shaded trace represents the 4.09 nA MscS current in response to a 1s pressure ramp to 160mm Hg followed by a 1.5s plateau, recorded before TFE application. The solid trace shows the 0.44 nA response of the same patch 15min after application of 2vol.% TFE in the bath. The pressure protocol was extended by an additional ramp reaching higher pressure (200mm Hg), which did not produce any additional activity indicating that the active part of channel population is fully saturated by the first stimulus, whereas the rest is in an inactivated state. (C) The kinetics of recovery after TFE washout. The shaded trace represents the response of a freshly excised patch without TFE, whereas the lower trace (0min) shows residual current after a 20min exposure to 3vol.% TFE. A washout resulted in a gradual return of the population current to a pre-TFE level, within ∼20min. Shifting of the midpoint to the left occurs concomitantly with the recovery. The midpoints values are 131 (6min), 127 (9min), 122 (21min), and 119mm Hg in control.

To further demonstrate that the observed reduction of Gmax in the presence of TFE was not caused by the right shift of the activation curve, we stimulated the TFE-silenced population with a double-ramp protocol (Figure 4B). Before TFE application, the patch was tested with a saturating ramp of pressure followed by a plateau evoking a ∼4.09 nA current. After exposure to 2vol.% TFE for 15min, the current stimulated by the same ramp fell to 0.44nA. Additional pressure applied in the form of a second ramp to a higher plateau did not evoke any extra activity. The inset in Figure 4B shows expanded segments of these traces to illustrate again that the single-channel amplitudes before and after TFE addition are identical.

TFE-induced silencing was also found to be reversible. A washout of TFE with recording solution returned 80–100% of the inactivated population back to the active state even after complete silencing with the highest concentration of TFE tested (5vol.%). On washout, p1/2 typically shifted back to the left, returning to a pressure close to the control (before TFE exposure). A time course for the return of channel activity, after partial silencing with 3vol.% TFE and washout, is shown in Figure 4C. Only after 20min did Gmax return to the control level. This reproducible result suggests a slow process of TFE cleansing from some reservoir, possibly the lipid bilayer.


Stimulation by fast ramps and pulses: effects of TFE on inactivation and recovery

To address the nature of the TFE-silenced state of MscS, we investigated population responses to pressure ramps applied with different speeds as well as responses to steeply applied stimuli (pulses). Previously published data 26 demonstrated that the MscS population responds fully to fast (<3s) ramps of saturating pressure, but with slower ramps (10–90s), only a fraction of population reaches the conductive state. The part of the population that does not conduct appears to inactivate while the ramp passes slowly through a range of intermediate pressures. Figure 5A depicts MscS responses to short pressure ramps in the presence and absence of TFE. The set of control experiments without TFE (shaded) demonstrates that 0.1, 0.5, 1, and 2s ramps evoke essentially the same maximal current from the MscS population as our fastest (hardware-limited) test pressure pulses (10ms rise time, 250ms duration). Upon addition of 3vol.% TFE to the same patch (bath perfusion), a 2s ramp was observed to evoke <8% of the original Gmax. Progressively faster stimuli were found to activate larger fractions of channels population. A declining slope of Gmax during the pressure plateau at the end of each ramp reveals an increased propensity to inactivation. We know from the previous studies 22,25,26 that MscS displays the tendency to inactivate when subjected to intermediate pressure stimuli (above the threshold and below saturation). In the inactivated state, the channel does not conduct and is no longer responsive to even saturating stimuli. Traces recorded from the same patch with rectangular steps of subsaturating pressure (Figure 5B) show that indeed, 3vol.% TFE increases the rate of inactivation ∼10 times. These data presented in Fig. 5 reveal that MscS channels do not inactivate spontaneously from their resting state upon exposure to TFE as sharply applied stimuli can elicit activation of the channel population. At subsaturating pressures, TFE speeds up the process of inactivation, which appears to be the reason for the decreased fraction of active channels at slower rates of stimulus application.

Display large version of this figure
Figure 5
MscS responses to ramps and pulses. (A) Increasing the speed of linear ramps from zero to a saturating pressure (200mm Hg) evokes progressively larger current in patches treated with 3vol.% TFE on the cytoplasmic side (solid traces). Ramp speed was changed after stabilization of Gmax under the 1s stimulus, with each subsequent stimulus separated by 3min intervals. A control set of traces taken before TFE perfusion (shaded) shows no difference in the maximal current at this range of ramp speeds. (B) Responses to a step of subsaturating pressure (160mm Hg) recorded on the same patch before (shaded) and after TFE perfusion (solid). Arrowheads indicate the maximal level of conductance upon application of the pressure stimulus. The current decay time (τi) decreased from 2.8s in control to 0.25s, with TFE illustrating a higher rate of inactivation.

Recovery of the MscS population from the inactivated to the resting state was also found to be influenced by TFE. Previous experiments revealed that this process is kinetically complex, with full recovery taking ∼3min under zero applied pressure 26. A typical response of WT MscS to an intermediate stimulus, followed by a series of short saturating stimuli designed to test the kinetics of recovery, is shown in Figure 6AB. An applied 25s step of subsaturating pressure initially opens ∼95% of channel population. This spike of channel activity decays almost monoexponentially with a characteristic inactivation time (τi). The τi in MscS is not constant and becomes longer with increasing amplitude of the intermediate pressure stimulus 26. By the end of a 25s intermediate stimulus, the current approaches the baseline signifying that the entire population is now in a nonconductive state. A short (0.25s) test pulse of saturating pressure immediately after the 30s step (Figure 6B) reveals that most of the population is now unresponsive to the stimulus with the exception of a small variable fraction (∼0–15%) that still responds to the saturating pressure. A train of test pulses spaced at 1, 10, 30, and 60s after the intermediate pulse illustrates the kinetics of recovery. Recovery appears to be a multiexponential process with at least two components (τ1r and τ2r). We observed a relatively fast component in the beginning (τ1r=1.8s, ∼85–90% Gmax), followed by a much slower recovery to the initial Gmax (τ2r=18.9s). Although the control curve presented here is fit relatively well with two exponents, a third component with a longer characteristic time but smaller contribution may exist.

Display large version of this figure
Figure 6
Effect of TFE on the rate of MscS inactivation and recovery. Inside-out patches containing ∼125 active MscS channels were subjected to a pressure step of 25s duration, during which the channels transiently open and display inactivation kinetics. The pressure stimulus is followed by a train of 0.25s test pulses, which permit our monitoring the time course of recovery (A). A single test pulse of saturating pressure (preceding the stimulus by 3min) indicates the maximal current attainable in the patch (Gmax). (B) Representative current traces recorded before (shaded) and after (solid) perfusion of 0.5vol.% TFE in the bath. (C) The time course of recovery before (shaded circles) and after perfusion of 0.5vol.% TFE. The recovery kinetics was fit with one and two exponents, with and without TFE, respectively.

After perfusion of 0.5vol.% TFE on the cytoplasmic side, Gmax measured with a 1s ramp stabilized at 75–90% of its initial level. Experiments were carried out only after stabilization of Gmax. Even at this low concentration of TFE, inactivation after a stimulus near p1/2 was on average 2.6±0.8 times faster (mean±SD, n=6).

TFE markedly slows down the process of recovery from the inactivated state. Figure 5C shows the normalized conductance of the channel population as a function of time after the intermediate stimulus. The recovery curve from the TFE treated population was fit with a single exponent producing a characteristic τr of 10.6s. The recovery data for the TFE-treated population was fit better with one exponent than with two. This suggests a delay in the onset of the second, longer recovery component, observed in the control. For comparison, the initial part of the control recovery curve was fit with a single exponent producing characteristic time of τr of 2.4s. The fast stage of recovery of the TFE-treated population to 80% Gmax was therefore 4.2±0.4 times slower than untreated control (n=7) (Figure 6C).



Discussion

The results described above depict two types of events taking place at different concentrations of TFE in the aqueous solution. At lower TFE concentrations (1–5vol.%), we observe a dramatic effect on the kinetics of channel redistribution between the functional states, whereas at higher concentrations (10–15vol.%) MscS channels dissociate into monomeric form. It appears that the nature of these two effects is qualitatively the same and rests primarily on the capability of TFE to partition into membranes or detergent micelles and to perturb buried interhelical contacts.

Previous work 21 identified MscS as part of an oligomeric protein complex that survives solubilization in SDS at room temperature but becomes dissociated by TFE. In this study, we showed that MscS forms stable oligomers in cold, ionic-detergent (LDS) gel electrophoresis. Previously, oligomeric MscS could only be visualized by using Blue-Native PAGE 35.

Exposure of the protein in membrane vesicles to 15vol.% TFE was found to result in dissociation of oligomeric MscS into its monomeric subunits. A similar behavior was observed at 10vol.% for the purified, detergent-stabilized protein. This effect of TFE on MscS could potentially arise from two mechanisms. First, TFE could act via the lipid-phase by changing the packing properties of the bilayer as was observed for KcsA 17. Second, TFE may dissociate protein complexes by simply weakening the contacts between the subunits and/or associated lipids. Since we observe dissociation in MscS at approximately the same concentration, both in the context of the E. coli inner membrane and in detergent micelles, it seems likely that TFE works mainly by the latter mechanism, although the complexes are slightly more resistant to TFE when surrounded by the native lipid bilayer. Removal of free TFE from the system before membrane solubilization in LDS does not change the outcome, suggesting that TFE by itself critically compromises intersubunit interactions already in the membrane, and the dissociation of MscS does not appear to be a result of cooperative action between TFE and the detergent.

The existing data indicates a clear difference between TFE's effects on soluble and membrane-embedded proteins. The ability for TFE to stabilize helical conformations in peptides and accelerate protein folding has been explained by aggregation of TFE around the protein backbone, local exclusion of water from the competition for hydrogen bonds, and possibly by lowering the effective dielectric constant of the solvent 36. This mechanism is consistent with TFE's tendency to form microscopic clusters in aqueous solutions 1,9, partition into hydrophobic protein crevices 34, and promote desolvation of protein surfaces that normally form buried contacts 14,36. At the same concentrations (15–30vol.%) that stabilize soluble proteins, TFE completely disrupts KcsA and MscS as well as many other membrane complexes 17,21.

Soluble proteins are stabilized by the formation of a dehydrated core. They are held together by hydrophobic interactions as well as strong polar interactions in a largely nonaqueous environment. TFE does not interact strongly with hydrophobic side chains 15, and thus does not unfold the hydrophobic core of a soluble protein until the concentration in the surrounding aqueous solution exceeds 50%. Membrane proteins, on the other hand, have an inverted design when compared to typical soluble proteins 37,38. They have water-filled cavities with hydrocarbon-exposed hydrophobic rims, and are stabilized by interactions with the surrounding lipids. The lipid bilayer could be considered a two-dimensional anisotropic solvent for membrane proteins where the lipids exist in a liquid crystalline state. Lipid tails are relatively large and do not easily intercalate between the helices thus preserving interhelical contacts. In contrast, TFE is small and thus capable of wedging between helices and separating them. Helical separation may be initiated primarily at the membrane boundaries where the TFE concentration is expected to be the highest.

In the transmembrane part of the MscS crystal structure solved by Bass and co-workers 27 (Fig. 7) only the central helices (TM3) form intersubunit contacts. The peripheral helices TM1 and TM2 do not form a continuous lipid-facing wall, but protrude outward at an angle, forming deep hydrophobic crevices. Given that tilting of individual transmembrane helices in the bilayer is energetically unfavorable 39,40, the absence of tilt-stabilizing helical contacts between the TM1-TM2 pairs suggests that this unusual angle could be a result of delipidation. Several independent MD simulations showed that when embedded in lipids, without tension, this structure quickly collapses 30,41. This suggests that 1), the resting conformation should be more compact, consistent with the hypothesis proposed by Booth and co-workers, and supported by cross-linking studies 42,43; and 2), under certain conditions the peripheral helices can detach from the pore-lining TM3s, thus forming crevices. As shown by the color-coded map of the protein surface (Figure 7B), the crevices are largely hydrophobic and could be occupied by an apolar solvent such as TFE. Previous measurements of the adiabatic compressibility demonstrated an increase of protein (lactalbumin) volume in the presence of 10–20vol.% of TFE indicating induction of packing defects and preferential accumulation of the co-solvent in hydrophobic crevices 34. For membrane proteins, partitioning of TFE into the lipid would increase the chance of penetration into interhelical gaps and the crystal structure suggests where these gaps may form in MscS.

Display large version of this figure
Figure 7
The crystal structure of MscS (A) (1MXM.pdb) and a vertical slice through the channel shown in a space-filled representation (B). The solvent-accessible surfaces were created with the probe of 1.4Å in radius and colored according the normalized consensus hydrophobicity scale 32. The most hydrophobic areas (the pore constriction, TM2-TM3 crevices, and the distal parts of pore vestibules) are yellow and most hydrophilic regions (rings of charges in the vestibules and the equatorial part of the cytoplasmic cage) are blue. Computation and averaging of solvation parameters was performed using HISTAN (see Methods). The pore with the hydrophobic gate (denoted as G) is formed by the seven TM3 helices contributed from each subunit. Deep hydrophobic crevices separating the TM3 barrel from the peripheral TM1-TM2 helices might be a result of delipidation. In the crystal structure these crevices are likely filled by the detergent, whereas in our experiments, these spaces can potentially be occupied by TFE. It is thus possible that the peripheral helices in the TFE-induced inactivated state are less tilted than in the crystal structure. The putative positions of membrane boundaries are shown by horizontal lines. The position of the outer boundary implies that 26 N-terminal residues missing from the crystal structure may form a periplasmic extension of the transmembrane domain.

Based on the above considerations and previous work 26,42, our model of the MscS native resting state is schematically represented as a compact conformation with the TM1-TM2 pairs packed along the TM3s (Figure 8A). In the resting state, the TM1-TM2-TM3 interactions are strong enough to transmit mechanical forces from the lipid bilayer to the gate. Applied tension expands the entire barrel making it conductive (Figure 8B). A subsequent detachment of the pore-lining TM3 helices from the peripheral helices, accompanied by kink formation at Glycine-113, leads the channel into a tension-insensitive inactivated state (Figure 8C).

Display large version of this figure
Figure 8
A schematic illustration of the hypothetical conformations of the transmembrane domain of MscS in the closed (A), open (B), and inactivated (C) states and the mechanism of TFE-induced perturbations. The gate of MscS is located at the cytoplasmic side and thus the channel is predicted to be more sensitive to increased tension in the inner leaflet. Intercalation of TFE in the inner leaflet (A) would create an extra lateral pressure shifting the equilibrium to the closed state. Addition of TFE to the periplasmic side expands the outer leaflet and produces tension in the inner leaflet, promoting opening (B). When present on the cytoplasmic side in sufficiently high concentration, under membrane tension TFE penetrates the crevices between the TM1-TM2 pair of helices and TM3 facilitating the transition to the inactivated state (C). The black arrows show the locations of TFE-induced tension/pressure components, which superimpose with the externally applied far-field tension used as a stimulus.

Why are the dose-response curves susceptible to perturbation by low concentrations of TFE, and how does TFE promote inactivation of MscS? It is likely that the membrane acts as an apolar reservoir attracting TFE. Although the exact partitioning coefficient of TFE between membranes and aqueous solutions has not yet been measured, it is known that log P octanol/water is 0.41, indicating ∼2.5-times higher preference for the bulk organic phase. Having an OH group with hydrogen-bonding capacity, TFE, like ethanol 44,45, may preferentially accumulate at the polar-apolar interface of the membrane, where its concentration could be higher than that in the bulk. Intercalation of TFE into the interfacial layer may additionally change the dipole and the surface components of the membrane boundary potential, thus perturbing lipid-lipid interactions and local interactions with proteins. At concentrations of 10vol.% and above, TFE perturbs phosphatidylcholine liposomes based on permeability tests. Inclusion of phosphatidylethanolamine was found to make the bilayer more resistant to permeabilization by TFE 17.

The partitioning of TFE is clearly reflected by measurable shifts in the MscS activation dose-response curves. These shifts are dependent on the membrane face (cytoplasmic or periplasmic) to which TFE is applied (Figure 2 and Figure 3). One possible mechanism for TFE action is illustrated in Fig. 8 combined with a schematic representation of the functional cycle of MscS. When adding TFE to the periplasmic face of the patch (pipette), TFE intercalates into the outer leaflet and increases its area. Because the two leaflets of the membrane are area-coupled by the common midplane, the expansion of the outer leaflet of the membrane must create tension in the inner leaflet 46,47. Since the gate in MscS is located more toward the cytoplasm 27, channel activation is likely to be sensitive to tension in the inner leaflet (Figure 8B). Extra tension in the inner leaflet, created by TFE intercalation, should promote early activation of the MscS population. This was indeed the observed result, as addition of TFE to the periplasmic face caused a leftward-shift of the dose-response curves (Fig. 2). In contrast, when TFE is presented to the cytoplasmic face of the patch (bath perfusion), partitioning of TFE increases lateral pressure in the inner leaflet (Figure 8A), causing a right-shift of the activation curve (Figure 3AB). Increased pressure caused by TFE intercalation partially negates the applied tension. The fact that the magnitude of the right-shift is not always stable suggests that TFE can, given sufficient time, redistribute between the leaflets thus dissipating the asymmetric area perturbation. This interpretation, however, needs to be taken with caution, as it has not been demonstrated that excised patches of bacterial membrane lack lipid reservoirs at the edges, which may allow independent area expansion of each of the leaflets, thus uncoupling them. However, because the inner E. coli membrane is densely packed with integral proteins (50% by weight), it may be assumed that this greatly impedes slippage of the two leaflets, making this system similar to a closed liposome in terms of its response to amphipath incorporation.

Early data on the modulation of MscS-like channels by chlorpromazine, trinitrophenol, and lysophosphatidylcholine (LPC) showed that these substances invariably activate the channels when presented from the cytoplasmic side 48. In this respect, the action of these amphipaths is distinct from the observed inhibitory action of TFE, which lowers the activation threshold only when presented to the periplasmic side. This difference is the focus of further investigation. A strong activating effect of externally applied lysolipids has also been reported for the large mechanosensitive channel MscL. Spontaneous activation was observed in the presence of large concentrations of LPC, an effect that only occurred when LPC is applied asymmetrically 49. In this regard LPC, like TFE, may strongly perturb leaflet area. However, it is not known if TFE causes the same spontaneous positive curvature, a feature characteristic of LPC.

The increased propensity to inactivation in the presence of TFE can be explained by partial separation of TM1-TM2 pairs from the gate-forming TM3 helices and stabilization of this state by intercalating TFE. As illustrated by data in Fig. 5, TFE does not drive MscS inactivation at low tension, thus its partitioning into the interhelical crevices (at low concentrations) does not seem to occur spontaneously. Instead, TFE partitioning appears to be critically facilitated by membrane tension that, in the framework of our gating hypothesis (Fig. 8), normally drives TM2-TM3 separation. TFE occupying voids in the molecule would stabilize the inactivated state, preventing fast reassociation of the TM1-TM2 pairs with TM3 and thus recovery (Fig. 6). Such an effect would also result in a less compact conformation of the MscS channel consistent with a slight upshift of MscS bands observed in gel electrophoresis upon addition of TFE (Figure 1C). The sidedness of the inactivating effect of TFE, shown to be active only from the cytoplasmic side (Figure 2 and Figure 3 and Figure 4), supports the proposed location of the crevices as being accessible only from the cytoplasmic face. TFE added to the pipette does not cause inactivation, presumably because, after traversing the membrane core, it does not substantially accumulate in the cytoplasmic leaflet, as it would quickly partition out into the TFE-free aqueous compartment.

At the present stage we cannot firmly exclude that TFE in some way modifies the cytoplasmic “cage” domain leading to inactivation. It has been previously shown that the channel propensity to inactivation depends on the state of this cage domain, which can be altered either by truncating mutations 50 or by high-molecular-weight co-solvents 51. Additionally, it has been demonstrated that concentrations of TFE as low as 3–5vol.% can influence conformational distributions in soluble proteins 52,53. Besides the TM2-TM3 crevices, other apolar solvent-accessible areas of MscS, such as the pore vestibules (Fig. 7), could potentially act as sites of TFE accumulation. Although possible at higher concentrations, accumulation in the pore does not seem to occur in the tested range of 0.5–5vol.% as TFE was not observed to interfere with the single-channel conductance (Figure 4AB, inset). The slow onset of TFE action on wash-in (Figure 2 and Figure 3) and slow return of channel activity on washout (Figure 4C) are also more consistent with TFE partitioning into and out of a relatively large hydrophobic reservoir, a role more likely to be served by the membrane itself. A detailed comparison of the water-membrane partitioning coefficients with the concentration dependencies of their membranotropic actions for TFE and similar compounds is a current research focus and may clarify the above issues.


Conclusions and prospects

We have identified two ranges of concentrations for TFE, which cause separable effects on the MscS channel. The lower range (0.5–5vol.%) dramatically affects the kinetics of inactivation under tension with no effect on the oligomeric state of the channel complex. The higher range (10–15vol.%) causes larger perturbations, ultimately leading to subunit separation. The data above suggests that TFE can be used not only as a protein-denaturing proteomics tool, but also as a perturbing agent that biases membrane proteins toward specific conformational states or reduces transition barriers. We observed that the effects of TFE are consistent with the existing models of MscS activation and inactivation 26.

Future projects will certainly require a more quantitative analysis of TFE partitioning into cell membranes, liposomes, monolayers, and micelles and its effects on lateral pressure. A detailed kinetic analysis of MscS inactivation/recovery in the presence of different concentrations of TFE may suggest the characteristic times, distances, and pathways of TFE redistribution between the lipid bilayer and protein. Further understanding of the structural organization of MscS could be obtained by using TFE to probe the strength of intersubunit interactions in mutants with perturbed or stabilized helical contacts, thus the location of crevices filled with TFE in the inactivated state can be further specified. Also, perturbing the tight TM3-TM3 knob-into-hole packing in the resting state 27,28 with mutations may weaken the complex against TFE. If a decrease in stability is not observed in such mutants, we should search for alternative intersubunit contacts, not seen in the delipidated crystal structure.

As more atomic structures of membrane proteins become available and more realistic force fields for molecular simulations are developed, the utilization of nonaqueous co-solvents will become more useful and interpretable. Parameters for MD simulations of proteins in the presence of TFE are already available 36,54. The merging of computation with experimental research will be a powerful strategy in studies of function-defining conformational transitions in membrane proteins.


References

1. Hong, D.P., Hoshino, M., Kuboi, R., and Goto, Y. (1999). Clustering of fluorine-substituted alcohols as a factor responsible for their marked effects on proteins and peptides. J. Am. Chem. Soc. 121, 8427–8433. CrossRef | PubMed

2. Nelson, J.W., and Kallenbach, N.R. (1986). Stabilization of the ribonuclease S-peptide α-helix by trifluoroethanol. Proteins 1, 211–217. CrossRef | PubMed

3. Sonnichsen, F.D., Van Eyk, J.E., Hodges, R.S., and Sykes, B.D. (1992). Effect of trifluoroethanol on protein secondary structure: an NMR and CD study using a synthetic actin peptide. Biochemistry 31, 8790–8798. PubMed

4. Wienk, H.L., Czisch, M., and de Kruijff, B. (1999). The structural flexibility of the preferredoxin transit peptide. FEBS Lett. 453, 318–326. CrossRef | PubMed

5. Jasanoff, A., and Fersht, A.R. (1994). Quantitative determination of helical propensities from trifluoroethanol titration curves. Biochemistry 33, 2129–2135. PubMed

6. Luo, P., and Baldwin, R.L. (1999). Interaction between water and polar groups of the helix backbone: an important determinant of helix propensities. Proc. Natl. Acad. Sci. USA 96, 4930–4935. CrossRef | PubMed

7. Lu, H., Buck, M., Radford, S.E., and Dobson, C.M. (1997). Acceleration of the folding of hen lysozyme by trifluoroethanol. J. Mol. Biol. 265, 112–117. CrossRef | PubMed

8. Kumar, Y., Muzammil, S., and Tayyab, S. (2005). Influence of fluoro, chloro and alkyl alcohols on the folding pathway of human serum albumin. J. Biochem. (Tokyo) 138, 335–341. CrossRef | PubMed

9. Chitra, R., and Smith, P.E. (2001). Properties of 2,2,2-trifluoroethanol and water mixtures. J. Chem. Phys. 114, 426–435. CrossRef | PubMed

10. Cammers-Goodwin, A., Allen, T.J., Oslick, S.L., McClure, K.F., Lee, J.H., and Kemp, D.S. (1996). Mechanism of stabilization of helical conformations of polypeptides by water containing trifluoroethanol. J. Am. Chem. Soc. 118, 3082–3090. CrossRef | PubMed

11. Kentsis, A., and Sosnick, T.R. (1998). Trifluoroethanol promotes helix formation by destabilizing backbone exposure: desolvation rather than native hydrogen bonding defines the kinetic pathway of dimeric coiled coil folding. Biochemistry 37, 14613–14622. PubMed

12. Bodkin, M.J., and Goodfellow, J.M. (1996). Hydrophobic solvation in aqueous trifluoroethanol solution. Biopolymers 39, 43–50. CrossRef | PubMed

13. Rajan, R., and Balaram, P. (1996). A model for the interaction of trifluoroethanol with peptides and proteins. Int. J. Pept. Protein Res. 48, 328–336. PubMed

14. Reiersen, H., and Rees, A.R. (2000). Trifluoroethanol may form a solvent matrix for assisted hydrophobic interactions between peptide side chains. Protein Eng. 13, 739–743. PubMed

15. Banerjee, T., and Kishore, N. (2005). Does the anesthetic 2,2,2-trifluoroethanol interact with bovine serum albumin by direct binding or by solvent-mediated effects? A calorimetric and spectroscopic investigation. Biopolymers. 78, 78–86. CrossRef | PubMed

16. Cortes, D.M., and Perozo, E. (1997). Structural dynamics of the Streptomyces lividans K+ channel (SKC1): oligomeric stoichiometry and stability. Biochemistry 36, 10343–10352. PubMed

17. van den Brink-van der Laan, V., Chupin, J., Killian, A., and de Kruijff, B. (2004). Stability of KcsA tetramer depends on membrane lateral pressure. Biochemistry 43, 4240–4250. PubMed

18. Barrera, F.N., Renart, M.L., Molina, M.L., Poveda, J.A., Encinar, J.A., Fernandez, A.M., Neira, J.L., and Gonzalez-Ros, J.M. (2005). Unfolding and refolding in vitro of a tetrameric, α-helical membrane protein: the prokaryotic potassium channel KcsA. Biochemistry 44, 14344–14352. PubMed

19. Deshusses, J.M., Burgess, J.A., Scherl, A., Wenger, Y., Walter, N., Converset, V., Paesano, S., Corthals, G.L., Hochstrasser, D.F., and Sanchez, J.C. (2003). Exploitation of specific properties of trifluoroethanol for extraction and separation of membrane proteins. Proteomics 3, 1418–1424. CrossRef | PubMed

20. Zuobi-Hasona, K., Crowley, P.J., Hasona, A., Bleiweis, A.S., and Brady, L.J. (2005). Solubilization of cellular membrane proteins from Streptococcus mutans for two-dimensional gel electrophoresis. Electrophoresis 26, 1200–1205. CrossRef | PubMed

21. Spelbrink, R.E., Kolkman, A., Slijper, M., Killian, J.A., and de Kruijff, B. (2005). Detection and identification of stable oligomeric protein complexes in Escherichia coli inner membranes: a proteomics approach. J. Biol. Chem. 280, 28742–28748. CrossRef | PubMed

22. Levina, N., Totemeyer, S., Stokes, N.R., Louis, P., Jones, M.A., and Booth, I.R. (1999). Protection of Escherichia coli cells against extreme turgor by activation of MscS and MscL mechanosensitive channels: identification of genes required for MscS activity. EMBO J. 18, 1730–1737. CrossRef | PubMed

23. Sukharev, S. (2002). Purification of the small mechanosensitive channel of Escherichia coli (MscS): the subunit structure, conduction, and gating characteristics in liposomes. Biophys. J. 83, 290–298. Abstract | Full Text | PDF (234 kb) | PubMed

24. Okada, K., Moe, P.C., and Blount, P. (2002). Functional design of bacterial mechanosensitive channels. Comparisons and contrasts illuminated by random mutagenesis. J. Biol. Chem. 277, 27682–27688. CrossRef | PubMed

25. Koprowski, P., and Kubalski, A. (1998). Voltage-independent adaptation of mechanosensitive channels in Escherichia coli protoplasts. J. Membr. Biol. 164, 253–262. CrossRef | PubMed

26. Akitake, B., Anishkin, A., and Sukharev, S. (2005). The “dashpot” mechanism of stretch-dependent gating in MscS. J. Gen. Physiol. 125, 143–154. CrossRef | PubMed

27. Bass, R.B., Strop, P., Barclay, M., and Rees, D.C. (2002). Crystal structure of Escherichia coli MscS, a voltage-modulated and mechanosensitive channel. Science 298, 1582–1587. CrossRef | PubMed

28. Edwards, M.D., Li, Y., Kim, S., Miller, S., Bartlett, W., Black, S., Dennison, S., Iscla, I., Blount, P., Bowie, J.U., et al. (2005). Pivotal role of the glycine-rich TM3 helix in gating the MscS mechanosensitive channel. Nat. Struct. Mol. Biol. 12, 113–119. CrossRef | PubMed

29. Anishkin, A., and Sukharev, S. (2004). Water dynamics and dewetting transitions in the small mechanosensitive channel MscS. Biophys. J. 86, 2883–2895. Abstract | Full Text | PDF (1018 kb) | PubMed

30. Sotomayor, M., and Schulten, K. (2004). Molecular dynamics study of gating in the mechanosensitive channel of small conductance MscS. Biophys. J. 87, 3050–3065. Abstract | Full Text | PDF (1491 kb) | CrossRef | PubMed

31. Fraczkiewicz, R., and Braun, W. (1998). Exact and efficient analytical calculation of the accessible surface areas and their gradients for macromolecules. J. Comput. Chem. 19, 319–333. CrossRef | PubMed

32. Eisenberg, D., Schwarz, E., Komaromy, M., and Wall, R. (1984). Analysis of membrane and surface protein sequences with the hydrophobic moment plot. J. Mol. Biol. 179, 125–142. CrossRef | PubMed

33. Humphrey, W., Dalke, A., and Schulten, K. (1996). VMD: visual molecular dynamics. J. Mol. Graph. 14, 33–38. CrossRef | PubMed

34. Kanjilal, S., Taulier, N., Le Huerou, J.Y., Gindre, M., Urbach, W., and Waks, M. (2003). Ultrasonic studies of alcohol-induced transconformation in β-lactoglobulin: the intermediate state. Biophys. J. 85, 3928–3934. Abstract | Full Text | PDF (136 kb) | PubMed

35. Stenberg, F., Chovanec, P., Maslen, S.L., Robinson, C.V., Ilag, L.L., von Hippel, G., and Daley, D.O. (2005). Protein complexes of the Escherichia coli cell envelope. J. Biol. Chem. 280, 34409–34419. CrossRef | PubMed

36. Roccatano, D., Colombo, G., Fioroni, M., and Mark, A.E. (2002). Mechanism by which 2,2,2-trifluoroethanol/water mixtures stabilize secondary-structure formation in peptides: a molecular dynamics study. Proc. Natl. Acad. Sci. USA 99, 12179–12184. CrossRef | PubMed

37. White, S.H., and Wimley, W.C. (1999).