Mechanosensitive Ion Channels of the Archaeon Haloferax volcanii *

Mechanosensitive (MS) ion channels have been documented in a variety of cells belonging to Eukarya andEubacteria. We report the novel finding of two types of MS ion channels in the cell membrane of the halophilic archaeonHaloferax volcanii, a member of the Archaeathat comprise the third phylogenetic domain. The two channels, MscA1 and MscA2, differed in their kinetic properties with MscA1 exhibiting more frequent open-closed transitions than MscA2. Both channels have large conductances that rectify between −40 mV and +40 mV where the conductance of MscA1 ranged from 380 to 680 picosiemens, whereas MscA2 ranged from 850 to 490 picosiemens. Both channels were blocked by submillimolar gadolinium. In addition, the channels of either membrane vesicles or detergent-solubilized membrane proteins remained functional upon reconstitution into artificial liposomes, a result that indicates that these channels are activated by mechanical force transmitted via the lipid bilayer alone. Subsequently a 37-kDa protein corresponding to the MscA1 channel activity was purified. With the possible functional similarity to bacterial MS channels, our finding of MS channels inArchaea emphasizes the ubiquity and importance of these channels in all domains of the evolutionary tree.

According to the recent revision the universal phylogenetic tree is composed of three domains: Eukarya, Eubacteria, and Archaea (formerly archaebacteria) (1)(2)(3)(4)(5)(6). From this scheme archaebacteria, which are prokaryotes like eubacteria, constitute an intermediary domain between eubacteria and eukaryotes, and although prokaryotes, archaebacteria are neither phylogenetically closer to eubacteria or to eukaryotes (7). As a distinct group of microorganisms Archaea comprise several different families of cells adapted to environments of certain habitats characterized by extreme temperatures such as in ocean hydrothermal vents, or high salt concentrations as occur in the Dead Sea (3,8).
The existence of ion channels in cell membranes of different organisms belonging to the eubacterial and eukaryotic phylogenetic domains has been well documented. In contrast, the existence of ion channels in cell membranes of Archaea was only recently documented with the discovery of porin-like channels in the archaebacterium Haloferax volcanii (formerly Halobacterium volcanii) (9). In the present study we report the finding of two types of MS ion channels in the plasma mem-brane of the same microorganism that seem to share many properties of the described bacterial MS 1 ion channels (10 -18). The finding of this class of channels in the cell membrane of an archaeon demonstrates that MS channels, as well as porins, are present in organisms belonging to all domains of the evolutionary tree and indicates the importance of these types of membrane pores in the phylogeny of ion channels.

EXPERIMENTAL PROCEDURES
Isolation of the Cell Envelope-Cells of H. volcanii were grown and membranes prepared as described previously (9). Cells were cultured in nutrient rich media (in mM: 3350 NaCl, 170 MgCl 2 , 200 MgSO 4 , 6 CaCl 2 , 26 KCl, 6.6 NaHCO 3 , 5.4 NaBr, plus 5 g of yeast extract per liter) until the absorbance A 600 was 1.0, washed three times by centrifugation with phosphate-buffered saline, and the pellet resuspended in 2 mM MgSO 4 , 5% sucrose, 100 mM NaCl, DNase (20 g/ml), 50 mM NaH 2 PO 4 , pH 7.6, and passed twice through a French press at 8,000 psi (Aminco, SLM Instruments, Inc.). The resulting suspension was centrifuged at 9,000 rpm for 20 min (Avanti J-25 I, Beckman) to eliminate cell debris. The supernatant was centrifuged for 25 min at 90,000 rpm (Beckmann TL-100). The pellet containing the membrane vesicle fraction was then resuspended in 10 mM HEPES-KOH, pH 7.0, and protein content was determined using the Dc protein assay (Bio-Rad). Aliquots of the suspension were stored at Ϫ20°C for further use. Alternatively, cell membrane extract following French press was layered on top of a discontinous sucrose gradient (25 and 60%) and spun down for 16 h at 25,000 rpm (SW28 Beckman rotor). Two major membrane fractions were collected as a "pink band" at the 25%/60% interface and a "black pellet." The protein content was measured as above. Aliquots (50 l) were stored at Ϫ20°C until required.
Membrane Solubilization and Protein Purification-Membrane fraction (pink band) was solubilized in 200 mM KCl, 10 mM HEPES-KOH, pH 7.0, buffer containing 100 mM octylglucoside and incubated for 30 min. Alternatively, membrane fractions after French press were washed by centrifugation in a high salt wash solution (700 mM NaCl, 10 mM Tris, pH 8.0) followed by a low salt wash (1 mM EDTA, 10 mM Tris, pH 8.0). For the subsequent steps either octylglucoside or Triton X-100 were used. The resulting pellet was resuspended in 1% Triton X-100 or 50 mM octylglucoside and cut with ammonium sulfate (at approximately 7% saturation in Triton X-100 or 17.5% in octylglucoside). The supernatant was collected after centrifugation at 90,000 rpm for 20 min (Beckman TL-100) and dialyzed overnight at 4°C against 10 mM Tris, pH 8.0, with either 0.5% Triton X-100 or 25 mM octylglucoside. The dialyzed fraction was concentrated on a Centricon 10 (Amicon), and the proteins were separated by preparative electrophoresis and electroelution in the presence of Triton X-100 or octylglucoside. The gel system consisted of a 4% stacking gel (pH 6.8) and a 9% resolving gel (pH 8.8) prepared as used in SDS-polyacrylamide gel electrophoresis, except that both gels contained 0.5% Triton X-100 or 25 mM octylglucoside. The protein fraction was mixed at a ratio of 2:1 with a buffer containing 10% glycerol, 65 mM Tris, pH 8.8, and 0.003% bromphenol blue before loading onto the gel. The electrophoretic separation was carried out at a constant current of 60 mA for 8 h. The resolving gel was cut in three parts identical in size, and the proteins were eluted from each part of the gel including the stacking gel using the Electro Eluter (Model 422, Bio-Rad) with a buffer of 192 mM glycine, 25 mM Tris containing 0.2% Triton X-100 or 25 mM octylglucoside. The electroeluted proteins were concentrated using Centricon 10 tubes (Amicon) and stored at 4°C for * This work was supported by Grants from the National Health and Medical Research Council of Australia (960591), the Australian Research Council (A09701150 and A337847), the Raine Medical Foundation, and the Center National de la Recherche Scientifique. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
§ Supported by a Healy Postdoctoral Fellowship. ʈ To whom correspondence should be addressed. Tel.: 618-9346-2986; Fax: 618-9346-3469; E-mail: bmartinac@receptor.pharm.uwa.edu.au. further use. In addition we introduced another step of purification before the electroelution. The membrane fractions solubilized in 25 mM octylglucoside were cut with ammonium sulfate (20%) and applied to a phenyl-Sepharose 6 fast flow column (1.5 ϫ 7 cm). The column was washed with 50 mM phosphate buffer (pH 6.8) containing 25 mM octylglucoside and 20% ammonium sulfate and then eluted by a linear ammonium sulfate gradient (20 to 0%). Active fractions were concentrated (Centricon 10, Amicon) and loaded onto a SDS-polyacrylamide gel to extract the proteins by electroelution as described above.
Liposome Preparation and Protein Reconstitution-Proteoliposomes were made essentially according to the method used for reconstitution of the large MS ion channel (MscL) of Escherichia coli (19). Alternatively, the solubilized protein fractions were incubated with azolectin liposomes (protein to lipid ratio of 1:100 or 1:30) for 30 min at room temperature and Bio-Beads SM-2 (Bio-Rad) were added (400 mg of Bio-Beads per ml of protein solution for 1% Triton X-100 or 50 mM octylglucoside) to remove the detergent. The suspension was rocked for 3-4 h at room temperature and then the Bio-Beads were allowed to settle. The supernatant was ultracentrifuged for 25 min at 90,000 rpm (TL-100, Beckman), and the resulting pellet was resuspended in 20 -50 l of 10 mM HEPES-KOH, pH 7.0. Aliquots of the liposomes were spotted onto glass slides and dehydrated/rehydrated as described to yield giant liposomes used for patch-clamp experiments.
Electrophysiology-For each experiment, a 2-l aliquot of proteoliposomes was placed in a 800-l chamber containing recording solution (in mM: 200 KCl, 40 MgCl 2 , 10 HEPES, pH 7.2, with KOH). Micropipettes of borosilicate glass (100 l microcapillaries, Sigma) were standardized by routinely testing a sample pipette by submersion in absolute ethanol and measuring the pipette bubble number (20), which was typically in the range 3.2-3.4. Pipettes were filled with recording solution. Membrane patches were obtained from unilamellar blisters from collapsing liposomes (21), and in some cases a brief application of suction (Ͻ20 mm Hg) was applied to form seals with resistances in the range 10 -50 G⍀.
Negative pressure (suction) was applied by syringe to the back of the pipette to activate MS ion channels (12). Pressure and current were recorded simultaneously using the pCLAMP6 program and Digidata 1200 A/D converter (Axon Instruments). Mechanosensitive ion channel currents were recorded by applying continuous pipette voltage either as voltage steps or voltage ramps in the range of Ϯ60 mV. The channel currents were observed once a threshold of applied pressure was achieved. Since the threshold activation pressure for bacterial MS channels reconstituted in liposomes is often reduced (but approaches a steady state) upon subsequent restretching of a particular patch, 2 only patches so "trained" were used in estimations of Boltzmann characteristics.
Data Analysis-Single channel currents were estimated either from amplitude histograms or using a cursor measurement following subtraction of the baseline. Threshold activation pressures were taken as the pressure at which the first channel opening to the full-level conductance was observed in a previously untested patch. Data are expressed as means and standard errors unless otherwise noted.

RESULTS
Reconstitution of native membrane vesicles yielded several ion channel types, predominantly exhibiting porin-like channel behavior (9). With slight modification of this method of Besnard, Martinac, and Ghazi (9), the presence of porin-like channel activity was minimized. Isolated liposome patches were examined for the presence of MS ion channels of similar type to those characteristic of E. coli; namely activated by mechanical tension in the lipid bilayer alone. Two separate MS channel conductances (MscA1 and MscA2) were observed in liposomes incorporated with native membrane vesicles (Fig. 1A). As well 2 A. C. Le Dain, unpublished observation. as observation in separate patches, on occasion both channel types were observed in the same patch. It was also possible to observe multiple channels of a given type in a particular patch. Both archaeal MS channel conductances were activated by application of negative pressure (suction) with a threshold activation pressure in previously untested patches for MscA1 of Ϫ121 Ϯ 11 mm Hg (n ϭ 4) and MscA2 of Ϫ160 Ϯ 13 mm Hg (n ϭ 7). Comparing the left panel with the right in Fig. 1A, the channel kinetics of the two conductances were different. MscA1 exhibited relatively brief openings, whereas MscA2 exhibited much longer openings with sustained activity while the applied pressure remained greater than the threshold activation pressure.
Detergent solubilization of the native membrane proteins by either octylglucoside or Triton X-100, followed by additional purification steps such as high salt wash of the membranes, ammonium sulfate cut, and protein electroelution from polyacrylamide gel, presented MS channel activity with similar electrophysiological characteristics (Fig. 1B). However, the threshold of activation for solubilized archaeal MS channels was reduced, where for MscA1 the threshold was Ϫ87 Ϯ 23 mm Hg (n ϭ 5) and for MscA2 was Ϫ77 Ϯ 7 mm Hg (n ϭ 5). Similar to MS channels from other preparations, as the applied pressure increased, the channel activity as measured by the open probability also increased ( Fig. 2A). The relationship between channel open probability and applied pressure could be well described by a Boltzmann distribution (Fig. 2B). The Boltzmann characteristics for both archaeal channels are summarized in Table I. For MscA2 the Boltzmann characteristics were also determined at ϩ40 mV (P 0.5 ϭ 29 Ϯ 6 mm Hg, 1/s ϭ 2.07 Ϯ 0.54; n ϭ 3) and Ϫ40 mV (P 0.5 ϭ 43 Ϯ 9 mm Hg, 1/s ϭ 2.82 Ϯ 0.95 mm Hg; n ϭ 4).
Since gadolinium is an inhibitor of MS channels (15,22), we tested sensitivity of the archaeal channels to this ion. Similar to the bacterial MS channels (23), both archaeal MS channels were blocked by submillimolar (0.2-0.5 mM) concentrations of gadolinium (data not shown).
Toward the molecular identification of these MS channels we introduced another step of purification in our isolation protocol. Octylglucoside membrane extracts cut with ammonium sulfate were applied onto a phenyl-Sepharose column. MS channel activities were found at the end of the eluate made by a linear ammonium sulfate gradient. Fig. 4 shows the protein pattern of one of the most active fractions (9 of 10 patches with activity) presenting both types of channels MscA1 and MscA2. This fraction contained two major protein bands that were electroeluted from the gel and reconstituted in liposomes. MS channel activity was only recovered from the electroeluted protein with an apparent molecular weight of approximately 37 kDa. The observed channels (5 of 10 patches active) corresponded to the MscA1 type (Fig. 4, right panel). DISCUSSION In the present study we report the discovery of two novel types of MS channels in the cell membrane of the archaeon H. volcanii. In analogy to the bacterial MS channels (18,24) we named the archaeal channels MscA1 and MscA2 and have found that both channels have many properties in common with bacterial MS channels but exhibit certain differences. When compared with E. coli MS channels, MscS (12) and MscL (19,25), which have a pressure sensitivity of Ϸ5 mm Hg per e-fold change in open probability, MscA1 and MscA2 exhibited an increased sensitivity of 1.4 -2.9 mm Hg per e-fold change in open probability. Second, unlike MscS, which is voltage-dependent (26), the archaeal MS channels do not show such significant voltage dependence (Table I). However, both of the archaeal MS channels possess large conductances in the range of several hundreds of picosiemens and seem to have a very low or minimal selectivity for cations over anions (data not presented), properties in common with bacterial MS channels (16,18,26,27). In addition, both archaeal channels were blocked by gadolinium at concentrations comparable to those reported to block bacterial MS channels (19,23). Significantly, MscA1 and MscA2 can be solubilized by detergents and reconstituted in artificial liposomes whereupon they remained fully functional when examined by the patch-clamp technique. This channel property is identical to that of the MscL, which allowed this ion channel to be to date the only cloned MS channel whose mechanosensitivity has been unambiguously documented (17,28). Furthermore, this functional property suggests that the archaeal channels belong to a family of MS channels that are activated by the mechanical force transmitted exclusively via the lipid bilayer (16,17). Interestingly, the archaeal MS channels possess a unique characteristic in that their conductive  properties alter in a nonlinear fashion with the pipette voltage, i.e. they rectify such that with the pipette voltage changing between Ϫ60 and ϩ60 mV, the conductance of MscA1 decreases whereas that of MscA2 increases. At present the importance of this channel property for the physiology of the H. volcanii remains unclear. In our view, the significance of discovering MS channel activity in the cell membrane of an archaeal cell is manifold. First, from an evolutionary perspective, this finding documents the ubiquity and importance of MS channels in the three domains of the phylogenetic tree. This channel ubiquity raises the question of the physiological function of these channels in prokaryotes. In bacteria, MS channels have been proposed to play a role in osmoregulation and more explicitly in the release of osmolytes (23). When cultivated in high osmolarity media, bacteria are capable of accumulating high concentrations of osmolytes (or osmoprotectants) such as trehalose or glycine betaine, the role of which is to counterbalance the external osmolarity. Upon a sudden shift to lower osmolarity media, these organic compounds are excreted (29 -31). Although the role of bacterial MS channels in osmolyte efflux is not proven, it is probable given the fast kinetics of this process (30,31). Moreover, a role of bacterial MS channels in the efflux of molecules of this size would be consistent with their apparently anomalous conductance; for example in the case of the MscL, molecules as large as polyamines readily pass through the pore (27). The study of osmoregulation is less advanced in archaea but it is noteworthy that accumulation of osmoprotectants has been documented in methanogens (32) and recently in haloalkaliphilic archaea (33). Although excretion of these species upon osmotic downshock has yet to be documented, it is highly plausible that archaeal MS channels play a similar role to that postulated for bacterial channels.
Second, the mechanism of gating the archaeal channels by the mechanical force transmitted by the lipid bilayer alone, supports the idea that mechanosensitivity may have evolved several times during evolution (28). Taking into account the two recognized mechanisms of mechanosensitivity (34), the present finding suggests that gating of MS channels according to the bilayer model (35)(36)(37) may represent a more general mechanism than that described by the tethered model (38 -41).
Finally, in view of a comparative analysis of Eukarya, Eubacteria and Archaea, our finding demonstrates that Archaea, or at least their halophilic branch, harbor eubacterial-like rather than eukaryote-like MS channel-type proteins in their cellular membranes. The fact that archaeal MS channels appear electrophysiologically similar to their bacterial counterparts is consistent with the idea that the prokaryotic Archaea proteins involved in transport are closer in function to those of Eubacteria rather than Eukarya (3,42). Although at present it is not known if the protein structure will prove to be as related, it is unlikely since one of the identified proteins has a molecular weight at least twice that of the MscL monomer. Elucidation of the physiological role of MS channels in Archaea as well as their molecular relatedness to bacterial MS channels will be the focus of further study.