pH-dependent conformation, dynamics, and aromatic interaction of the gating tryptophan residue of the influenza M2 proton channel from solid-state NMR

Abstract

The M2 protein of the influenza virus conducts protons into the virion under external acidic pH. The proton selectivity of the tetrameric channel is controlled by a single histidine (His(37)), whereas channel gating is accomplished by a single tryptophan (Trp(41)) in the transmembrane domain of the protein. Aromatic interaction between these two functional residues has been previously observed in Raman spectra, but atomic-resolution evidence for this interaction remains scarce. Here we use high-resolution solid-state NMR spectroscopy to determine the side-chain conformation and dynamics of Trp(41) in the M2 transmembrane peptide by measuring the Trp chemical shifts, His(37)-Trp(41) distances, and indole dynamics at high and low pH. The interatomic distances constrain the Trp41 side-chain conformation to trans for χ1 and 120-135° for χ2. This t90 rotamer points the Nε1-Cε2-Cζ2 side of the indole toward the aqueous pore. The precise χ1 and χ2 angles differ by ∼20° between high and low pH. These differences, together with the known changes in the helix tilt angle between high and low pH, push the imidazole and indole rings closer together at low pH. Moreover, the measured order parameters indicate that the indole rings undergo simultaneous χ1 and χ2 torsional fluctuations at acidic pH, but only restricted χ1 fluctuations at high pH. As a result, the Trp(41) side chain periodically experiences strong cation-π interactions with His(37) at low pH as the indole sweeps through its trajectory, whereas at high pH the indole ring is further away from the imidazole. These results provide the structural basis for understanding how the His(37)-water proton exchange rate measured by NMR is reduced to the small proton flux measured in biochemical experiments. The indole dynamics, together with the known motion of the imidazolium, indicate that this compact ion channel uses economical side-chain dynamics to regulate proton conduction and gating.

Figure 1

Several Trp⁴¹ rotamers that have been proposed in the literature using various experimental techniques (Table 1). Population percentages for α-helices from the Penultimate Rotamer Library are shown (42). (Left column) Side view of two of the four helices; both His³⁷ and Trp⁴¹ side chains are depicted. Note the shortest distances between these two residues are between two adjacent helices rather than from the same helix. (Right column) Top view of the four helices and the Trp41 sidechains. (a) t90 rotamer (χ₁ = 180°, χ₂ = +90°). (b) t-105 rotamer (χ₁ = 180°, χ₂ = −105°). (c) m95 rotamer (χ₁ = −60°, χ₂ = +95°).

Figure 2

Trp⁴¹ chemical shifts at pH 4.5 and 8.5 in LW-M2TM bound to the VM+ membrane. (a) Aromatic region of the 1D ¹³C cross-polarization spectra at pH 4.5 and 8.5. (b) Aliphatic region of the 1D ¹³C DQF spectra at the two pH conditions. The lipid peaks were suppressed by the double-quantum filter. Note the chemical shift changes of Trp⁴¹ and Leu⁴⁰ Cα and Cβ peaks, indicating pH-induced small conformation changes of the helix backbone. (c) 2D ¹³C-¹³C DQF spectrum of LW-M2TM at pH 8.5. (d) 2D ¹⁵N-¹³C correlation spectrum of the peptide at high pH. (e) Summary of Trp⁴¹¹³C and ¹⁵N chemical shifts at high pH (red) and low pH (green). (Black) High-pH chemical shifts of M2(22–62) in DOPC/DOPE bilayers (38) for sites that have >0.5 ppm chemical shift difference from the M2TM values. (Bold) Chemical shifts that differ by >0.5 ppm between high and low pH. All ¹³C chemical shifts are reported on the tetramethylsilane scale.

Figure 3

¹³C-¹⁹F REDOR distance measurements of ¹³C-labeled His³⁷ and 5-¹⁹F-labeled Trp⁴¹ in VM bound M2TM. (a) Approximate spatial arrangement of His³⁷ of one helix and Trp⁴¹ of the neighboring helix. (b) Representative ¹³C-detected ¹⁹F-dephased REDOR control (S₀, black) and dephased (S, red) spectra of the peptide at pH 8.5. The spectra were measured at 243 K under 3.3 kHz MAS. (c) Representative REDOR spectra of the peptide at pH 4.5. The spectra were measured under 3.0, 3.3, and 6.7 kHz MAS in order to detect all His³⁷ aromatic ¹³C signals without overlap from the carbonyl sidebands (asterisks). (d) ¹³C-¹⁹F REDOR S/S₀ intensity as a function of mixing time for the high-pH (red) and low-pH (green) peptides. De novo two-spin simulations are shown. Best-fit distances are highlighted for the high-pH data (red) and the low-pH data (green).

Figure 4

Total RMSD between the measured and modeled Cα-F, Cγ-F, Cε1-F, Cδ2-F, and ¹⁹F-¹⁹F distances as a function of Trp⁴¹ (χ₁, χ₂) torsion angles. Different backbone structure models for M2TM at high (a and b) and low (c and d) pH are used for structural modeling. Only the shortest of the four possible distances in the tetramer for each two-spin combination is used to compare with the experimental data. (a) PDB:3LBW backbone structure. (b) PDB:2KQT backbone structure. (c) PDB:3C9J backbone structure. (d) PDB:2KAD backbone structure. Between the two high-pH backbone structures, consensus (χ₁, χ₂) results were found at (−175°, 120°). For the low-pH state, the best-fit Trp⁴¹ rotamer depends on the backbone structure. For the PDB:2KAD structure (d), the best-fit solution is (χ₁, χ₂) = (−155°, 135°), whereas the PDB:3C9J structure (c) shows two rotamer minima without steric conflict. (Crosses) Rotamers with steric clashes, which are not considered further (see Fig. S2).

Figure 5

His³⁷-Trp⁴¹ contacts and Trp⁴¹ side-chain conformation at high and low pH. (a) Best fits of the ¹³C-¹⁹F REDOR and ¹⁹F-¹⁹F CODEX data at high pH, using PDB:2KQT for the backbone structure and a Trp⁴¹ rotamer of χ₁ = −175° and χ₂ = +120°. (Red and black curves) Model-dependent five-spin and two-spin best-fit simulations. (b) Best fits of the ¹³C-¹⁹F REDOR and ¹⁹F-¹⁹F CODEX data at low pH, using PDB:2KAD as the backbone structure and a Trp⁴¹ rotamer of χ₁ = −155° and χ₂ = +135°. (c) His-Trp region of the high-pH structure (PDB:2KQT), with Trp⁴¹ (χ₁, χ₂) angles of (−175°, 120°). (d) His-Trp region of the low-pH structure (PDB:2KAD), where the Trp⁴¹ rotamer is (−155°, 135°). (Dashed lines) Relevant His-Trp ¹³C-¹⁹F distances. At low pH, the imidazole and indole rings approach each other more closely, due to the combined effect of the helix backbone orientation change and Trp⁴¹ (χ₁, χ₂) changes.

Figure 6

Trp⁴¹ side-chain dynamics at high pH (red) and low pH (green) in VM+ membranes at 303 K. (a and b) Representative ¹³C-¹H DIPSHIFT curves for (a) Cδ1-H and (b) Cζ3-H couplings. (c) Cε2-Nε1 REDOR dephasing curve. (d) Summary of all dipolar order parameters of the indole ring at high and low pH. Cδ1-H and Cζ3-H bonds show different order parameters between high (red) and low (green) pH. Most order parameters have an experimental uncertainty of ±0.05. (e) Calculated order parameters as a function of the standard deviations (σ) of the Gaussian fluctuations around the Cα-Cβ and Cβ-Cγ bonds. Only values that agree with the measured low-pH S_CH are shown. Small-amplitude Gaussian fluctuations with σ₁ ≈ 30° and σ₂ ≈ 15° (shaded area) agree with all measured order parameters. (f) Calculated order parameters that agree with the high-pH S_CH values. Only torsional motion around the χ₁ axis is compatible with the experimental data, with σ₁ ≈ 25° (indicated by a star).

Figure 7

Trp⁴¹ conformational dynamics and aromatic interaction with His³⁷. (a) Low-pH scenario. The measured equilibrium (χ₁, χ₂) rotamer of (−155°, 135°) is shown as position 2, together with two limiting rotamers based on the measured order parameters. The (−125°, 150°) rotamer (position 3) is closer to His³⁷ whereas the (175°,120°) rotamer (position 1) is further away from His³⁷. These angles were chosen to be one standard deviation (30° for σ₁ and 15° for σ₂) from the average angle. (b) High-pH scenario. The equilibrium rotamer is (−175°, 120°) (position 2), and the limiting rotamers are (−150°,120°) (position 1) and (160°,120°) (position 3), based on a σ₁ of 25° and no χ₂ fluctuation.