| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
* Department of Biology, University of Maryland, College Park, Maryland; and Department of Biochemistry of Membranes, Institute of Biomembranes, Utrecht University, Utrecht, The Netherlands
Correspondence: Address reprint requests to Sergei Sukharev, Tel.: 301-405-6923; E-mail: sukharev{at}umd.edu.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONCONCLUSIONS AND PROSPECTSREFERENCES |
|---|
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONCONCLUSIONS AND PROSPECTSREFERENCES |
|---|
), TFE is often chosen as a nonpolar medium for spectroscopic determination of peptide conformations (2–4) 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 30 vol % (1,9). At these concentrations, TFE strongly stabilizes the 
-helical and ß-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 (12–14). 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 20 vol % TFE present in a DDM detergent solution (17). Further increase of TFE to 35 vol % 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 20 vol % 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 (8–18 nm2) 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 |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONCONCLUSIONS AND PROSPECTSREFERENCES |
|---|
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 800 ml Luria-Bertani medium at 37°C to an OD600 of 0.6 and induced with 0.8 mM isopropyl-ß-D-thiogalactopyranoside for 1 h. Cells were collected by centrifugation. The cell-pellet was washed with 50 ml of 50 mM potassium phosphate buffer pH 8 containing 5 mM 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, 45 min, 4°C), resulting in 0.6 g of cell membranes (wet weight). Membrane pellets were stored at –80°C until either being resuspended in 50 mM 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 0.6 g of membrane pellet was dissolved in 8 ml of 50 mM potassium phosphate buffer pH 8, 300 mM NaCl, 20 mM imidazole, and 3% (w/v) octylglucoside. This solution was cleared from insoluble particles by ultracentrifugation (45,000 rpm, 45 min, 4°C). The resulting solution was incubated with 0.5 ml Ni2+ nitrilotriacetic acid slurry on ice for 1 h. The slurry was poured into a column and eluted by gravity. The gel bed was washed with 10 volumes of 300 mM NaCl, 50 mM potassium phosphate buffer pH 8, 20 mM imidazole, and 1% (w/v) octylglucoside. Elution was performed stepwise with buffers containing 50, 75, and 200 mM 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 (0.3 mg/ml) or a membrane preparation from MJF465 cells containing roughly 4 mg/ml total protein were added to solutions of TFE in water for a total volume of 30 µl. The samples were incubated at ambient temperature for 1 h. Samples were cooled on ice before addition of 7.5 µl 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 120 V 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 +30 mV (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 (200 mM KCl, 90 mM MgCl2, 10 mM CaCl2, and 5 mM 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 5.0 vol %. TFE solutions were made fresh before each experiment and solutions older than 3 h 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 (4 ml) was replaced with three chamber volumes of TFE solution through perfusion. The total time of perfusion was 3 min, after which the system was allowed to rest for an additional 3 min 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 (1–5 vol %) behind a 3–5 mm plug of pipette solution with 300 mM sucrose to delay the onset of exposure. This diffusion-limited delay (2–10 min) 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 5–10 kHz at a sampling rate of 30 kHz. 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 1.4 Å. 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 |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONCONCLUSIONS AND PROSPECTSREFERENCES |
|---|
|
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. Fig. 1 A, 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 (Fig. 1 A, 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 (4x–7x) 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 (Fig. 1 A, lane 5).
We presume that the positions of covalently cross-linked oligomers of MscS itself (Fig. 1 A, 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 (Fig. 1 A, 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 1 h at ambient temperature, before being subjected to electrophoresis on continuous LDS-PAGE gels. Fig. 1 B 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 >10 vol %, 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 Fig. 1 B showing a breakdown between 10 and 15 vol % 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 (Fig. 1 C, lane 1). Exposure of MscS to 2–6 vol % TFE causes some bands to disappear, while simultaneously increasing the intensity of the heptameric and likely pentameric bands (Fig. 1 C, 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 10 vol % causes complete dissociation of MscS into monomers (Fig. 1 C, 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 1 s 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 170 mm Hg; however, within each patch, sequential sweeps grouped tightly around a single midpoint with <2% deviation around the mean (26).
|
5 mm 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 (>1 h). When the concentration of TFE was increased to 3–5 vol %, larger decreases in p1/2 (leftward shifts) of 20 mm Hg were observed. The ratio of midpoints for 5 vol % 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 (0.5–2 vol %) invariably shifted p1/2 to the right by 10–40 mm Hg (Fig. 3, A–C). The peak ratio of midpoints for 2 vol % 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 (>2 h), 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 35 mm Petri dish filled with 5 vol % TFE at a rate of 2 µl/min. In the course of 100 min 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.
|
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 (Fig. 4 A). The single-channel conductance in the presence of 3 vol % TFE in the bath was essentially the same as in control except for a small deviation at strongly depolarizing voltages (–80 mV 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 5 vol % TFE, and we observed no systematic deviation >±1 mV.
|
4.09 nA current. After exposure to 2 vol % TFE for 15 min, the current stimulated by the same ramp fell to 0.44 nA. Additional pressure applied in the form of a second ramp to a higher plateau did not evoke any extra activity. The inset in Fig. 4 B 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 (5 vol %). 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 3 vol % TFE and washout, is shown in Fig. 4 C. Only after 20 min 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 (<3 s) ramps of saturating pressure, but with slower ramps (10–90 s), 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. Fig. 5 A 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 2 s ramps evoke essentially the same maximal current from the MscS population as our fastest (hardware-limited) test pressure pulses (10 ms rise time, 250 ms duration). Upon addition of 3 vol % TFE to the same patch (bath perfusion), a 2 s 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 (Fig. 5 B) show that indeed, 3 vol % 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.
|
3 min 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 Fig. 6, A and B. An applied 25 s step of subsaturating pressure initially opens 95% of channel population. This spike of channel activity decays almost monoexponentially with a characteristic inactivation time (
i). The i in MscS is not constant and becomes longer with increasing amplitude of the intermediate pressure stimulus (26). By the end of a 25 s intermediate stimulus, the current approaches the baseline signifying that the entire population is now in a nonconductive state. A short (0.25 s) test pulse of saturating pressure immediately after the 30 s step (Fig. 6 B) 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 60 s after the intermediate pulse illustrates the kinetics of recovery. Recovery appears to be a multiexponential process with at least two components (1r and 2r). We observed a relatively fast component in the beginning (1r = 1.8 s, 85–90% Gmax), followed by a much slower recovery to the initial Gmax (2r = 18.9 s). 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.
|
TFE markedly slows down the process of recovery from the inactivated state. Fig. 5 C 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 r of 10.6 s. 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 r of 2.4 s. 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) (Fig. 6 C).
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONCONCLUSIONS AND PROSPECTSREFERENCES |
|---|
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 15 vol % TFE was found to result in dissociation of oligomeric MscS into its monomeric subunits. A similar behavior was observed at 10 vol % 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 (15–30 vol %) 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 (Fig. 7 B), 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 10–20 vol % 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.
|
,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 (Fig. 8 A). 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 (Fig. 8 B). 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 (Fig. 8 C).
|
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 10 vol % 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 (Figs. 2 and 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 (Fig. 8 B). 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 (Fig. 8 A), causing a right-shift of the activation curve (Fig. 3, A and B). 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 (Fig. 1 C). The sidedness of the inactivating effect of TFE, shown to be active only from the cytoplasmic side (Figs. 2–4
), 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 3–5 vol % 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 0.5–5 vol % as TFE was not observed to interfere with the single-channel conductance (Fig. 4, A and B, inset). The slow onset of TFE action on wash-in (Figs. 2 and 3) and slow return of channel activity on washout (Fig. 4 C) 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 |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONCONCLUSIONS AND PROSPECTSREFERENCES |
|---|
).
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.
|
FOOTNOTES |
|---|
Submitted on October 3, 2006; accepted for publication January 9, 2007.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONCONCLUSIONS AND PROSPECTSREFERENCES |
|---|
2. Nelson, J. W., and N. R. Kallenbach. 1986. Stabilization of the ribonuclease S-peptide -helix by trifluoroethanol. Proteins. 1:211–217.[CrossRef][Medline]
3. Sonnichsen, F. D., J. E. Van Eyk, R. S. Hodges, and B. D. Sykes. 1992. Effect of trifluoroethanol on protein secondary structure: an NMR and CD study using a synthetic actin peptide. Biochemistry. 31:8790–8798.[CrossRef][Medline]
4. Wienk, H. L., M. Czisch, and B. de Kruijff. 1999. The structural flexibility of the preferredoxin transit peptide. FEBS Lett. 453:318–326.[CrossRef][Medline]
5. Jasanoff, A., and A. R. Fersht. 1994. Quantitative determination of helical propensities from trifluoroethanol titration curves. Biochemistry. 33:2129–2135.[CrossRef][Medline]
6. Luo, P., and R. L. Baldwin. 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.
7. Lu, H., M. Buck, S. E. Radford, and C. M. Dobson. 1997. Acceleration of the folding of hen lysozyme by trifluoroethanol. J. Mol. Biol. 265:112–117.[CrossRef][Medline]
8. Kumar, Y., S. Muzammil, and S. Tayyab. 2005. Influence of fluoro, chloro and alkyl alcohols on the folding pathway of human serum albumin. J. Biochem. (Tokyo). 138:335–341.
9. Chitra, R., and P. E. Smith. 2001. Properties of 2,2,2-trifluoroethanol and water mixtures. J. Chem. Phys. 114:426–435.[CrossRef]
10. Cammers-Goodwin, A., T. J. Allen, S. L. Oslick, K. F. McClure, J. H. Lee, and D. S. Kemp. 1996. Mechanism of stabilization of helical conformations of polypeptides by water containing trifluoroethanol. J. Am. Chem. Soc. 118:3082–3090.[CrossRef]
11. Kentsis, A., and T. R. Sosnick. 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.[CrossRef][Medline]
12. Bodkin, M. J., and J. M. Goodfellow. 1996. Hydrophobic solvation in aqueous trifluoroethanol solution. Biopolymers. 39:43–50.[CrossRef][Medline]
13. Rajan, R., and P. Balaram. 1996. A model for the interaction of trifluoroethanol with peptides and proteins. Int. J. Pept. Protein Res. 48:328–336.[Medline]
14. Reiersen, H., and A. R. Rees. 2000. Trifluoroethanol may form a solvent matrix for assisted hydrophobic interactions between peptide side chains. Protein Eng. 13:739–743.
15. Banerjee, T., and N. Kishore. 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][Medline]
16. Cortes, D. M., and E. Perozo. 1997. Structural dynamics of the Streptomyces lividans K+ channel (SKC1): oligomeric stoichiometry and stability. Biochemistry. 36:10343–10352.[CrossRef][Medline]
17. van den Brink-van der Laan, V., J. Chupin, A. Killian, and B. de Kruijff. 2004. Stability of KcsA tetramer depends on membrane lateral pressure. Biochemistry. 43:4240–4250.[CrossRef][Medline]
18. Barrera, F. N., M. L. Renart, M. L. Molina, J. A. Poveda, J. A. Encinar, A. M. Fernandez, J. L. Neira, and J. M. Gonzalez-Ros. 2005. Unfolding and refolding in vitro of a tetrameric, -helical membrane protein: the prokaryotic potassium channel KcsA. Biochemistry. 44:14344–14352.[CrossRef][Medline]
19. Deshusses, J. M., J. A. Burgess, A. Scherl, Y. Wenger, N. Walter, V. Converset, S. Paesano, G. L. Corthals, D. F. Hochstrasser, and J. C. Sanchez. 2003. Exploitation of specific properties of trifluoroethanol for extraction and separation of membrane proteins. Proteomics. 3:1418–1424.[CrossRef][Medline]
20. Zuobi-Hasona, K., P. J. Crowley, A. Hasona, A. S. Bleiweis, and L. J. Brady. 2005. Solubilization of cellular membrane proteins from Streptococcus mutans for two-dimensional gel electrophoresis. Electrophoresis. 26:1200–1205.[CrossRef][Medline]
21. Spelbrink, R. E., A. Kolkman, M. Slijper, J. A. Killian, and B. de Kruijff. 2005. Detection and identification of stable oligomeric protein complexes in Escherichia coli inner membranes: a proteomics approach. J. Biol. Chem. 280:28742–28748.
22. Levina, N., S. Totemeyer, N. R. Stokes, P. Louis, M. A. Jones, and I. R. Booth. 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][Medline]
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.
24. Okada, K., P. C. Moe, and P. Blount. 2002. Functional design of bacterial mechanosensitive channels. Comparisons and contrasts illuminated by random mutagenesis. J. Biol. Chem. 277:27682–27688.
25. Koprowski, P., and A. Kubalski. 1998. Voltage-independent adaptation of mechanosensitive channels in Escherichia coli protoplasts. J. Membr. Biol. 164:253–262.[CrossRef][Medline]
26. Akitake, B., A. Anishkin, and S. Sukharev. 2005. The "dashpot" mechanism of stretch-dependent gating in MscS. J. Gen. Physiol. 125:143–154.
27. Bass, R. B., P. Strop, M. Barclay, and D. C. Rees. 2002. Crystal structure of Escherichia coli MscS, a voltage-modulated and mechanosensitive channel. Science. 298:1582–1587.
28. Edwards, M. D., Y. Li, S. Kim, S. Miller, W. Bartlett, S. Black, S. Dennison, I. Iscla, P. Blount, J. U. Bowie, and I. R. Booth. 2005. Pivotal role of the glycine-rich TM3 helix in gating the MscS mechanosensitive channel. Nat. Struct. Mol. Biol. 12:113–119.[CrossRef][Medline]
29. Anishkin, A., and S. Sukharev. 2004. Water dynamics and dewetting transitions in the small mechanosensitive channel MscS. Biophys. J. 86:2883–2895.
30. Sotomayor, M., and K. Schulten. 2004. Molecular dynamics study of gating in the mechanosensitive channel of small conductance MscS. Biophys. J. 87:3050–3065.
31. Fraczkiewicz, R., and W. Braun. 1998. Exact and efficient analytical calculation of the accessible surface areas and their gradients for macromolecules. J. Comput. Chem. 19:319–333.[CrossRef]
