INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

The M₂ ion channel protein of influenza A virus has a high selectivity for protons (3, 8, 17, 21, 24). Even though the M₂ protein is only a minor component of the viral envelope (30), the ion channel activity is nonetheless essential in the life cycle of the virus. Influenza virus enters cells via endocytosis, and in the low-pH environment found in the endosomal compartment, the M₂ ion channel is activated (reviewed in reference 14). Protons flow into the virion, acidifying the virion interior, a pH change which is thought to weaken protein-protein interactions between the viral matrix protein and the ribonucleoprotein (RNP) complex (reviewed in references 10 and 14), a prerequisite in the uncoating process. A considerable body of evidence indicates the M₂ ion channel is inhibited by the antiviral drug amantadine (3, 21, 29). In virus-infected cells treated with amantadine, the viral matrix protein fails to dissociate from the RNP, and import of the RNA into the nucleus does not occur. Hence, the RNPs cannot undergo replication in the nucleus and the viral life cycle is blocked (reviewed in references 10, 14, and 15).

The mature M₂ protein is 96 amino acid residues (the N-terminal methionine is cleaved [27]), and it spans the membrane once: it has 23 extracellular residues, a transmembrane (TM) domain of 19 residues, and a 54-residue cytoplasmic tail (16). The M₂ protein is disulfide linked (11, 25), and the ion channel active form is a homotetramer (22). Models of the structure of the M₂ ion channel based on experimentally determined data (13, 20) and by molecular dynamics principles (1, 23, 31) indicate that the M₂ TM domain forms a left-handed -helical coiled-coil containing a central ion-conducting pore through the axis of symmetry. The majority of models predict occlusion of the ion-conducting pore by the inward facing histidine residue 37. The importance of histidine residue 37 in proton conduction is highlighted by the finding that (i) the curve of pH activation of the channel possesses a midpoint very near the pK_a of histidine (28), (ii) substitution of His-37 for other residues causes altered ion channel properties (20, 28), (iii) coordination of Cu²⁺ ions with His-37 results in inhibition of channel activity (5), and (iv) the extent of the diminution of ion conduction in deuterium oxide solvent is consistent with protons binding to the channel as they traverse the pore (17).

To identify experimentally the residues which line the ion-conducting pore of the M₂ channel, each residue in the presumed TM domain was changed individually to cysteine. The ion currents of the resulting mutant ion channels were tested for their susceptibility to modification by aqueous sulfhydryl-specific reagents. A change in current of a given mutant on addition of the sulfhydryl-specific reagent is interpreted as indicating that in the wild-type protein, the residue is exposed to the aqueous surface of the molecule and forms part of the ion-conducting pore. By application of reagents that do not readily cross the membrane to both the extracellular and intracellular sides of the channel, it was possible to investigate the point at which the channel is most narrow.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Construction of M₂ TM domain cysteine mutants. The cysteine substitution mutants in the M₂ TM domain (influenza virus A/Udorn/72) were those described previously (20). In vitro synthesis of mRNA was performed using a mMessage mMachine T7 transcription kit (Ambion, Austin, Tex.).

Microinjection and culture of oocytes. Oocytes from Xenopus laevis were prepared for mRNA injection and injected with M₂ mRNA as described previously (5).

Electrophysiological recordings. At 24 to 48 h after mRNA injection, whole cell currents were recorded with a two-electrode voltage-clamp apparatus consisting of a differential preamplifier (model MEZ-7101; Nihon Kohden, Tokyo, Japan) that recorded the voltage difference between a pipette (filled with 3 M KCl) located in the cell and another in the surrounding bath. A voltage-clamp amplifier (Nihon Kohden CEZ-1100) provided feedback current to the oocyte through a second intracellular pipette. Oocyte currents were recorded in standard Barth's solution (0.3 mM NaNO₃, 0.71 mM CaCl₂, 0.82 mM MgSO₄, 1.0 mM KCl, 2.4 mM NaHCO₃, 88 mM NaCl, 15.0 mM morpholineethanesulfonic acid [pH 6.2] or 15.0 mM HEPES [pH 7.5]). To check for nonspecific leaks, amantadine hydrochloride (100 mM stock in Barth's solution; Sigma Chemical Co., St. Louis, Mo.) was diluted and applied at a working concentration of 100 mM for 2 to 5 min at the end of measurements from an oocyte.

Reagents. 2-Aminoethyl methanethiosulfonate hydrobromide (MTSEA), [2-(trimethylammonium)ethyl] methanethiosulfonate bromide (MTSET), and 2-[(5-fluoresceinyl)aminocarbonyl]ethyl methanethiosulfonate (MTS-fluorescein) were obtained from Toronto Research Chemicals (Toronto, Ontario, Canada). Reagents were stored at 20°C in a desiccator. Individual aliquots of reagent powder were weighed out. For each experiment, an aliquot was freshly diluted in Barth's solution and immediately applied to the oocyte.

Polypeptide analysis. Metabolic labeling of oocytes with [³⁵S]cysteine was performed as described previously (12, 21). Oocytes were homogenized in radioimmunoprecipitation buffer and extracted with trifluoromethane as described previously (12, 30). Polypeptides were analyzed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) under nonreducing conditions as described previously (11). Radioactivity was detected using a Fuji BAS 1000 Bioimager (Fuji Medical Instruments, Stanford, Conn.), and fluorescence was detected using a STORM imager (Molecular Dynamics, Perkin-Elmer, Norwalk, Conn.).

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Effect of extracellular application of MTS reagents on the activity of M₂ protein cysteine substitution mutants. To determine which residues line the pore of the M₂ ion channel, a set of mutants was generated in which each residue of the M₂ TM domain (residues 25 to 44) was replaced by cysteine (20). The accessibility of the cysteine residues to aqueous sulfhydryl-specific reagents applied both extracellularly and intracellularly was tested. Oocytes were injected with mRNA for a given M₂ mutant, and membrane currents were recorded by maintaining the oocyte at a holding voltage of 20 mV, with test pulses taken to 80 mV at 10-s intervals. The pH of the bath medium was lowered from 7.5 to 6.2 to activate the M₂ ion channel, and the membrane current was allowed to reach a steady-state amplitude. A sulfhydryl-specific reagent was then added to the bath, and its effect on the current was recorded. After the membrane current reached a new steady-state amplitude, the reagent was washed out. Finally, amantadine (100 µM) was added to determine M₂-specific currents. At the end of each step, the current-voltage relationship was determined by both a series of voltage clamp steps and a voltage ramp. As a control, the currents were recorded from an M₂ ion channel that did not contain a cysteine residue. As shown in Fig. 1A, addition of MTSEA or its washout did not cause a perturbation to the recorded current for the nonsubstituted (cysteineless) protein. An example of the effect of MTSEA on a cysteine-containing M₂ ion channel, M₂-G34C, is shown in Fig. 1B. Immediately upon addition of MTSEA (2.5 mM) the current decreased, with maximum inhibition observed within 30 s. This decrease was not reversible upon washout, as expected for a covalent modification. To determine that this effect was specific to the MTS reagent and was not caused by a mechanical disturbance of the oocyte, the bath medium was changed between two solutions of identical composition. No change in current was observed when the buffer was exchanged.

View larger version (21K): [in this window] [in a new window]   FIG. 1.   Effect of extracellular MTSEA on the conductances of M2 cysteine substitution mutants. (A) Application of MTSEA had no effect on the currents of a cysteineless M2 ion channel. (B) Time course of inhibition of the M2-G34C ion channel by MTSEA. Current was immediately decreased upon addition of MTSEA (2.5 mM) to the bathing medium and did not recover upon washout. (C) The conductances of only two mutants (M2-A30C and M2-G34C) were decreased by extracellular application of 2.5 mM MTSEA (*, P < 0.05). Two of the mutants (labeled "ia") had activity that was too low to measure inhibition accurately in this assay.

Activity of the M₂-L26C ion channel. Extracellular application of MTS reagents did not identify any residues that were exposed to the aqueous pore that were situated closer to the extracellular mouth of the channel than Ala-30. However, if the channel forms a helical coiled-coil bundle, then at least one additional N-terminal residue should line the pore between the extracellular opening and Ala-30. One residue in this region that could not be tested was Leu-26, because when substituted by cysteine the activity of the resulting mutant ion channel was too low to measure inhibition accurately. One possible explanation for the low activity of M₂-L26C is that introduced cysteine residues form an intersubunit disulfide bond (2), leading to decreased channel activity. Therefore, we tested whether treatment with the reducing agent dithiothreitol (DTT) would reduce such bonds, thereby increasing the currents observed.

View larger version (81K): [in this window] [in a new window]   FIG. 2.   The M2-L26C mutant protein forms reversible intersubunit disulfide bonds. Radiolabeled oocyte lysates were immunoprecipitated with the anti-M2 antibody 14C2 (30). Oocytes were either untreated controls (lane 1), DTT treated (lane 2), or DTT treated and then H2O2 treated (lane 3). Note that the ratio of M2 dimers (D) to monomers (M) decreases upon DTT treatment and is restored upon H2O2 treatment. , nonglycosylated species; g and 2g, mono- and diglycosylated species, respectively. Positions of molecular weight (MW) markers are shown at the right in kilodaltons.

Fluorescent labeling of M₂ cysteine substitution mutants. One possibility for the lack of effect of MTSEA on cysteine residues substituted at positions N-terminal to Ala-30 is that the channel pore widens near its extracellular mouth, such that even if a MTS reagent reacts with a cysteine in this region, the modifying moiety is too small to occlude the pore. To test whether such biochemical modification was occurring, fluorescently conjugated MTS derivatives were used. Oocytes expressing the M₂-L26C, M₂-V27C, and M₂-A30C ion channels were metabolically labeled with [³⁵S]cysteine and then incubated in Barth's solution (pH 7.5) to which MTS-fluorescein (ca. 500 µM) was added. Excess unreacted reagent was thoroughly washed out before the oocytes were lysed. The M₂ proteins were immunoprecipitated, the peptides were analyzed by SDS-PAGE, and the gels were analyzed for both radioactivity, to visualize total protein expression, and fluorescence, to detect the covalently linked fluorescent MTS reagent. By superimposing the two gel images, it could be determined which M₂ proteins were labeled with the fluorescent reagent.

View larger version (57K): [in this window] [in a new window]   FIG. 3.   MTS-fluorescein labeling of M2 cysteine substitution mutants. [35S]cysteine-labeled oocytes were incubated in a solution containing MTS-fluorescein reagent (two 5-min applications of ca. 500 µM), and immunoprecipitations were then performed. The resulting polyacrylamide gel was analyzed both for radioactivity (A) and for fluorescence (B). By superimposing the two gel images, it was determined that only the M2-A30C nonglycosylated and monoglycosylated monomeric species were fluorescently labeled. Other details are as for Fig. 2.

Effect of intracellular application of MTS reagents. The currents of the M₂ ion channel are thought to be in the order of 1 to 10 fA (17), too small to examine the effect of MTS reagents in an outside-out patch. Instead, we employed the technique of direct microinjection of a concentrated reagent solution while the oocyte was under two-electrode voltage clamp (26).

View larger version (17K): [in this window] [in a new window]   FIG. 4.   Effect of intracellular MTS reagent on the conductances of M2 cysteine substitution mutants. (A) The conductance of only one mutant, M2-W41C, was decreased by intracellular application of 2 mM MTSET. The M2-H37C mutant (labeled "ia") had activity that was too low to measure inhibition accurately in this assay. (B) M2-W41C currents were immediately decreased upon microinjection of a concentrated solution of MTSET (40 mM) into the oocyte. The trend line shown in gray is the extrapolated current in the absence of MTSET. Note that impaling the oocyte with the injection pipette while under two-electrode voltage clamp had no significant effect on the stability of the recording. (C) Mock injection with water had no effect on the observed current. (D) M2-G34C currents were not affected by microinjection of MTSET, in contrast to the inhibition observed when sulfhydryl-specific reagent was applied extracellularly (Fig. 1).

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

Extracellular application of MTS reagents inhibited currents observed from the M₂-A30C and M₂-G34C ion channels, whereas intracellular application inhibited currents from the M₂-W41C ion channel. These data indicate that residues Ala-30, Gly-34, and Trp-41 line the aqueous pore of the M₂ ion channel, results which are consistent with the M₂ TM domain forming a coiled coil (Fig. 5). These results agree with conclusions drawn from circular dichroism spectroscopy studies of the M₂ TM domain (4) and molecular dynamics simulations (23, 31), and in particular they confirm the identity of the residues that we predicted in a previous study to be pore lining (20).

View larger version (25K): [in this window] [in a new window]   FIG. 5.   Model of the proposed TM domain of the M2 protein, showing top view as seen from the extracellular side (A) and a cross-section in the plane of the lipid bilayer (B). Residues which were identified as facing the ion-conducting aqueous pore are indicated. The model of the M2 channel is taken from that calculated in reference 20. The figure was generated using the program GRASP (18).

The finding that the M₂-G34C ion channel could be inhibited only by extracellular addition of sulfhydryl reagent while the M₂-W41C could be inhibited only by intracellular addition of sulfhydryl reagent indicates that the narrowest part of the pore lies between positions 34 and 41. This is consistent with our hypothesis that His-37 serves as the selectivity filter for the channel via a proton-relay mechanism (17, 20). We were not able to test the M₂-H37C ion channel directly in this assay, however, as mutating this functionally critical residue to cysteine abolished channel activity. Nonetheless, a body of evidence suggests that the His-37 residue forms part of the ion-conducting pathway. This evidence includes inhibition of M₂ ion channel activity by transition metals (5), mutagenesis studies, and pH titration data (28).

Influenza viruses grown in the presence of the M₂-specific inhibitor amantadine acquire mutations in the TM domain of the M₂ protein. These mutations occur predominantly at Leu-26, Val-27, Ala-30, Ser-31, and Gly-34 (6, 8, 9, 25). As we show here that two of these residues, Ala-30 and Gly-34, form part of the interior surface of the channel pore, the data are consistent with amantadine binding to a site within the pore of the M₂ channel, in the fashion of an open pore blocker.

In our experiments, inhibition of the M₂ cysteine substitution proteins by MTS reagents was measured at pH 6.2, conditions at which the M₂ ion channel is activated. Therefore, the results obtained here most likely reflect the conformation of the channel in the ion-conducting state. From studies of disulfide bond formation in M₂ cysteine substitution proteins, it was concluded that conformational changes may occur near the cytoplasmic end of the channel pore when the medium is changed from pH 7.5 to pH 6.2 (2). However, we could not test this region of the protein at neutral pH due to technical limitations of our assay in applying MTS reagents intracellularly.

Labeling experiments using a fluorescently conjugated MTS derivative applied extracellularly confirmed that Ala-30 is a pore-lining residue. The fact that a reagent as large as MTS-fluorescein (molecular weight of 513.54) could penetrate so deeply into the pore suggests that the pore diameter may be surprisingly large near the extracellular opening. Results concerning pore-lining residues in the external region of the protein were inconclusive, however.

One mutant ion channel with a cysteine substituted near the extracellular mouth of the pore, M₂-L26C, had very low activity which could be increased upon incubation in DTT. Subsequent treatment with H₂O₂ partially reversed this effect. Immunoprecipitation experiments confirmed that this was due to the reduction and subsequent re-formation of intersubunit disulfide bonds between the introduced cysteines on neighboring monomers. The fact that cysteines introduced at this position can form reversible disulfide bonds very readily might suggest that Leu-26 is also a pore-lining residue, as discussed previously (2). However, alternative explanations are possible. The monomer -helices may rotate with respect to one another, briefly exposing minor conformations in which introduced cysteines are located favorably for disulfide bond formation. When such bonds are formed, they may lock the protein in a twisted conformation that does not conduct current. It is only upon the reduction of these bonds that the protein may return to its native structure, in which case it can conduct at least a small amount of current.

Attempts to test directly the accessibility of the reduced cysteine side chain of M₂-L26C to modification by sulfhydryl-specific reagents did not yield measurable changes in current, nor was this residue labeled by a fluorescently conjugated reagent. The cysteine side chain in this mutant protein may therefore be inaccessible to reagents applied in the aqueous medium. However, it is always a concern that a mutated residue may not occupy the analogous position in the structure of the mutant protein as the corresponding wild-type residue. The low activity observed from the M₂-L26C ion channel could also be due to the amino acid change causing an alteration to the protein structure. Similarly, the activity observed from the M₂-V27C ion channel is markedly different from that of the wild-type channel, in that although increased current is observed upon lowered pH, the channel conductance does not change, nor is the current sensitive to amantadine. Given these limitations, it may not be possible to identify either Leu-26 or Val-27 as pore lining by the substituted-cysteine accessibility method.

The results presented here confirm the current model of the M₂ ion channel, in which the TM domains of each subunit adopt an -helical conformation. Our results also agree with the details of other studies (2, 20), which locate residues Ala-30, Gly-34, His-37, and Trp-41 on the surface of the ion-conducting pore. Furthermore, our observation that an obstruction in the vicinity of His-37 prevents passage of MTS reagents introduced from either side of the membrane is consistent with the hypothesis that His-37 serves as the selectivity filter for the channel (20). These results therefore indicate that detailed biophysical measurements of the structure and activity of the M₂ ion channel with methods such as site-directed spin labeling (7), solid-state nuclear magnetic resonance spectroscopy (13), and unnatural amino acid substitution (19) will yield a more detailed understanding of the function of this model ion channel.