Mechanism of influenza A M2 transmembrane domain assembly in lipid membranes

Abstract

M2 from influenza A virus functions as an oligomeric proton channel essential for the viral cycle, hence it is a high-priority pharmacological target whose structure and functions require better understanding. We studied the mechanism of M2 transmembrane domain (M2TMD) assembly in lipid membranes by the powerful biophysical technique of double electron-electron resonance (DEER) spectroscopy. By varying the M2TMD-to-lipid molar ratio over a wide range from 1:18,800 to 1:160, we found that M2TMD exists as monomers, dimers, and tetramers whose relative populations shift to tetramers with the increase of peptide-to-lipid (P/L) molar ratio. Our results strongly support the tandem mechanism of M2 assembly that is monomers-to-dimer then dimers-to-tetramer, since tight dimers are abundant at small P/L's, and thereafter they assemble as dimers of dimers in weaker tetramers. The stepwise mechanism found for a single-pass membrane protein oligomeric assembly should contribute to the knowledge of the association steps in membrane protein folding.

Figure 1. A glimpse at Influenza A M2.

a In the structure of M2 tetramer one recognizes the transmembrane domain and C-terminal amphipathic helix as determined by Sharma et al. by solid state NMR (PDB ID: 2L0J). The surrounding lipid bilayer is added for clarity. The C_β-atom of residue L46, mutated to cysteine for spin-labeling, is shown as red sphere in each of four protomers of M2. b Diagram representing the full length M2 shows the position of the transmembrane domain (TMD) formed by the residues 25–46 colored in dark red. The amino acid sequence of M2TMD_21–49 used in this study is emphasized with L46C residue colored in red.

Figure 2

DEER data for M2TMD_21–49 spin-labeled at position L46C in DOPC:POPS (85:15% molar ratio) membranes recorded at pH 5.5 (in red) and pH 8 (in blue). a Background corrected and normalized (Methods, Supplementary Figs. 1 and 7) time-domain DEER data for P/L molar ratios of 1:235 and 1:2,300. The DEER modulation depth (Δ) is indicated. b DEER modulation depths for P/L 1:160, 1:253 and 1:2,300 are visualized as bar plots. The margins of error due to experimental uncertainties and limitations of data analysis were found to be within ± 5%. The expected values for DEER modulation depth of 0.23 and 0.54 for dimer (D) and tetramer (T), respectively, are indicated by dashed lines. However, it is recognized that for a tetramer the theoretical maximum of 0.54 would be unlikely due to slightly less than unity spin-labeling efficiency and effects caused by uniformity of the magnetic component, B₁, of microwave field over the sample in the resonator (Methods).

Figure 3. Experimental time-domain DEER data for M2TMD_21–49 spin-labeled at position L46C.

The protein was reconstituted into DOPC:POPS (85:15% molar ratio) at pH 5.5 a, and pH 8 b. Normalized raw data are compiled in the left a and b panels with P/L ratios given for each signal. Background-corrected data are plotted in the right a and b panels such as to have asymptotic value of zero and are normalized to have the amplitude at zero time equal to the modulation depth. (cf. Supplementary Fig. 1). The data for P/L’s 1:18,800, 1:9,400, 1:4,100, 1:2,300 1:1,650, 1:1,300, 1:820 1:500, 1:235 (only for pH 5.5) and 1:160 with progressively increasing modulation depths (Δ) values are shown. The colors of raw signals (left panels) match the colors of background corrected signals (right panels). c The modulation depth values plotted for samples spanning the whole range of P/L’s, namely 1:18,800, 1:9,400, 1:4,100, 1:2,300, 1:1,650, 1:1,300, 1:820, 1:500, 1:235, and 1:160, for both pH’s 5.5 and 8, plus an extra point 1:3,170 available for pH 8. The error bars correspond to the estimated error margins of ± 5% (cf. note in Fig. 1 caption).

Figure 4

Experimental time-domain DEER signals for M2TMD_21–49 spin-labeled at position L46C and reconstituted in β-DDM at pH 5.5. Raw data are shown on the left with P/D ratios provided for each signal. Background corrected and normalized data are on the right.

Figure 5. Concentration profiles of M2TMD_21–49 monomers, dimers and tetramers as a function of P/L and P/D molar ratios.

DEER signal modulation depth concentration profiles (filled squares) are plotted in the top row vs. M2 molar fraction for DOPC/POPS lipid mixture in a and b and β-DDM in c. The fits of these experimental data to the equilibrium model based on tandem momomer↔dimer↔tetramer vs. that for monomer↔tetramer model (for lipid and detergent at pH 5.5) are plotted in solid red and dashed green lines, respectively in a and c upper panels. Clearly, the tandem model is necessary to describe the M2TMD assembly pathway. Respective equilibrium constants, k_2d, k_4d for the tandem model are 15·10⁻⁶ MF and 448·10⁻⁶ MF, pointing to strong binding for dimers but relatively weaker bound tetramers. In β-DDM these constants both are weak and close to each other, being 264·10⁻⁶ and 644·10⁻⁶ MF. Concentration profiles of populations for M2TMD_21–49 monomers (M), dimers (D), and tetramers (T) are plotted in a and b lower panels for DOPC/POPS lipid membranes at pH’s 5.5 and 8, respectively, and for β-DDM in c lower panel. The populations of each fraction are expressed as M2 percentages of total M2TMD_21–49 monomer concentration, C_M2TMD = C_M + C_D + C_T. Here C_M2TMD is expressed as M2 molar fraction, 1/(1+A/P), with A/P is amphiphile-to-protein molar ratio.