Identification of the functional core of the influenza A virus A/M2 proton-selective ion channel

Abstract

The influenza A virus M2 protein (A/M2) is a homotetrameric pH-activated proton transporter/channel that mediates acidification of the interior of endosomally encapsulated virus. This 97-residue protein has a single transmembrane (TM) helix, which associates to form homotetramers that bind the anti-influenza drug amantadine. However, the minimal fragment required for assembly and proton transport in cellular membranes has not been defined. Therefore, the conductance properties of truncation mutants expressed in Xenopus oocytes were examined. A short fragment spanning residues 21-61, M2(21-61), was inserted into the cytoplasmic membrane and had specific, amantadine-sensitive proton transport activity indistinguishable from that of full-length A/M2; an epitope-tagged version of an even shorter fragment, M2(21-51)-FLAG, had specific activity within a factor of 2 of the full-length protein. Furthermore, synthetic fragments including a peptide spanning residues 22-46 were found to transport protons into liposomes in an amantadine-sensitive manner. In addition, the functionally important His-37 residue pK(a) values are highly perturbed in the tetrameric form of the protein, a property conserved in the TM peptide and full-length A/M2 in both micelles and bilayers. These data demonstrate that the determinants for folding, drug binding, and proton translocation are packaged in a remarkably small peptide that can now be studied with confidence.

Fig. 1.

Modular structure of A/M2 and the sequences of A/M2 variants. The dark gray box indicates the FLAG tag domain, red letters indicate Ala substitutions, and dashes indicate deleted residues.

Fig. 2.

pH activation and amantadine sensitivity of A/M2 constructs in Xenopus oocytes. Representative recordings for these A/M2 constructs are shown. Amantadine sensitivity was evaluated by bathing oocytes in pH 5.5 solution containing 100 μM amantadine once the inward current reached maximum amplitude.

Fig. 3.

Relative specific activity of M2(21-51)-FLAG, M2(21-61)-FLAG, and full- length A/M2-FLAG protein. (A) Representative recording traces of M2(21-51)-FLAG, M2(21-61)-FLAG, and full-length A/M2-FLAG protein expressed in oocytes. (B) Relative specific activity of A/M2-FLAG; the current was plotted against the immunofluorescence signal for each cell. The slope of the regression line provides the relative specific activity. (C and D) Relative specific activity of M2(21-61)-FLAG (C) and M2(21-51)-FLAG (D).

Fig. 4.

synM2(22-46) (A), synM2(19-62) (B), and A/M2 (C) proton fluxes were determined by proteoliposome assay with an electrical potential (≈−120 mV) at both pH_out and pH_in of 7.40. Cumulative protons transported per tetramer are plotted vs. time after liposome dilution into assay buffer after a 3-s delay for mixing (see SI Text). Protein-free control liposomes exhibit minimal background flux. All constructs are inhibited by 99 μM amantadine. A titration of synM2(22-46) with amantadine (A Inset) shows the fraction of the uninhibited initial rate (F) vs. negative log of amantadine concentration (M).

Fig. 5.

pH dependence of the thermodynamic stability (pK_app) of synM2(22-46) (A) (60–100 μM peptide in DPC, 1:150–1:250 monomer/detergent) and A/M2 (B) (50 μM protein in C14-sulfobetaine, 1:150 monomer/detergent) in the absence (●, ■) and presence (▲) of amantadine (3- to 10-fold excess over protein tetramer depending on detergent conditions). The curve in A was generated by using literature values for pK_as for the monomer (21) and tetramer (3), and the value of pK_app at limiting high pH as the sole fitting variable. The pK_a for protonation of His-37 in the monomeric form of full-length A/M2 has not been determined, so this variable was treated as an adjustable parameter in B. The computed value of the pK_a for His-37 in the monomer was 6.2, close to the value of 6.8 found for synM2(22-46) in micelles (21).

Fig. 6.

Spectral and isothermal calorimetric titrations of amantadine to synM2(22-46). (A) CD titration of synM2(22-46) at pH 7.4 and synM2(22-46) S31N at pH 8 in DPC micelles with amantadine and rimantadine. Peptide monomer concentrations were ≈40 μM; peptide/DPC ratios were 1:40–1:50. The ratio of the mean residue ellipticity (θ₂₂₃/θ₂₀₉) is shown vs. the concentration of amantadine (circles) or rimantadine (squares) for the WT syn(22-46) peptide (closed symbols) and the drug-resistant synM2(22-46) S31N (open symbols). The curve through the rimantadine data for synM2(22-46) shows the titration expected for a very tight-binding inhibitor of M2 (i.e., under conditions in which the total tetramer concentration is much greater than the dissociation constant for drug binding). The arrows show the position at which a break would be expected for a stoichiometry of 1 drug per tetramer versus 4 drugs per tetramer. (B) ITC titration of synM2(22-46) in POPC/POPG bilayers (1:100 peptide/lipid, pH 7.4) with amantadine, and the titration of the same drug into lipid bilayers that did not contain peptide. Buffer conditions for CD and ITC titrations are given in Materials and Methods.