An arginine residue in the outer segment of hASIC1a TM1 affects both proton affinity and channel desensitization

Abstract

Acid-sensing ion channels (ASICs) respond to changes in pH in the central and peripheral nervous systems and participate in synaptic plasticity and pain perception. Understanding the proton-mediated gating mechanism remains elusive despite the of their structures in various conformational states. We report here that R64, an arginine located in the outer segment of the first transmembrane domain of all three isoforms of mammalian ASICs, markedly impacts the apparent proton affinity of activation and the degree of desensitization from the open and preopen states. Rosetta calculations of free energy changes predict that substitutions of R64 in hASIC1a by aromatic residues destabilize the closed conformation while stabilizing the open conformation. Accordingly, F64 enhances the efficacy of proton-mediated gating of hASIC1a, which increases the apparent pH50 and facilitates channel opening when only one or two subunits are activated. F64 also lengthens the duration of opening events, thus keeping channels open for extended periods of time and diminishing low pH-induced desensitization. Our results indicate that activation of a proton sensor(s) with pH50 equal to or greater than pH 7.2-7.1 opens F64hASIC1a, whereas it induces steady-state desensitization in wildtype channels due to the high energy of activation imposed by R64, which prevents opening of the pore. Together, these findings suggest that activation of a high-affinity proton-sensor(s) and a common gating mechanism may mediate the processes of activation and steady-state desensitization of hASIC1a.

Figure S1.

Cartoon of putative β11-β12 desensitization mechanism. A single subunit of the channel trimer is shown in closed, open, and desensitized conformations based on the corresponding structures of cASIC. The ECD is divided in subdomains marked by different colors. The acidic pocket is delineated by a dashed oval and encompasses negatively charged residues located in the thumb, finger, and palm. Residues L415 and N416 in the β11-β12 linker together with Q276 in the adjacent β9 strand are shown inside the orange square. (A) Closed or resting conformation (PDB accession no. 5WKV) with a relaxed thumb and expanded acidic pocket, and the side chain of N416 facing the interior of the lower palm. (B) Open conformation (PDB accession no. 4NTW). Here, the acidic pocket is collapsed owing to displacement of the α5-helix of the thumb, the lower palm is expanded, and the upper part of pore is also expanded, removing the constriction of the gate. (C) The desensitized state. The acidic pocket remains collapsed, the lower palm is contracted, residues L415 and N416 in the β11-β12 linker have undergone isomerization (180° flipping), and the transmembrane domain narrows, shutting the pore. Structures of cASIC in SSD (PDB accession no. 6VTK) and desensitized by low pH are almost identical (PDB accession no. 3IJ4).

Figure 1.

Residue Trp-66 located close to the external side of TM1 in lASIC produces channels without desensitization. (A) lASIC and functional chimeras (CH1 and CH2) made with sequences of hASIC1a are represented above the corresponding current traces activated by pH 6.5. The continuous line indicates duration of the stimulus and dashed line the zero-level current. The experiment was repeated three times with different batches of oocytes. 10 oocytes were examined for each chimera. (B) Alignment of TM1 and TM2 amino acid sequences of chicken, human, and lamprey ASIC channels. In red are residues present only in lamprey. (C) Replacement of Arg-64 by Trp or Phe in hASIC1 induces nondesensitizing currents. The experiment was repeated three times with different batches of oocytes. 10 oocytes were measured for each mutant. (D) Replacement of Trp-66 in lASIC (W64R) for the corresponding residue in hASIC1a produces complete desensitization, whereas replacement by another aromatic side chain (W66F) maintains desensitization and increases the level of constitutive current at pH 7.4. The experiment was repeated three times with different batches of oocytes. 10 oocytes were examined for each construct.

Figure 2.

Protein structure modeling and stability calculations suggest that large aliphatic and aromatic amino acids at positions 64/66 in hASIC1a/lASIC favor the open channel conformation. (A) Model of hASIC1a with each domain depicted with a different color. For clarity, only one subunit is shown and the other two subunits are transparent. The position of R64 (W66 in lASIC) on the outer segment of TM1 is shown as yellow sphere. The membrane is depicted schematically and divided into membrane core (MC), which corresponds to the region occupied by the fatty acid chains, and transition region (TR), which is the area occupied by the glycerol-phosphate and lipid head groups. TMD, transmembrane domain. (B) Homology models of hASIC1a in closed, open, and desensitized states and their interdependencies according to the proton-mediated gating cycle of ASICs. (C) Residues surrounding R64 in hASIC1a and W66 in lASIC in structural models of the closed, open, and desensitized states. For clarity, only two subunits, which make interactions with R64 or W66 in the depicted models, are shown and colored blue and green, respectively. Possible side chain hydrogen bonds are indicated as dashed lines. (D) Computationally predicted stability changes (ΔΔG) for mutations of R64 in hASIC1a (left) or W66 in lASIC (right) to every other amino acid. A negative ΔΔG value indicates that protein stability relative to WT is increased by the mutation, whereas a positive value for ΔΔG indicates a destabilization. ΔΔG values for F/W64^hASIC1a and R/F66^lASIC, for which currents are shown in Fig. 1, B and C, are indicated with a black frame. Values for proline were off-scale due to incompatible backbone geometries in the starting model and are not shown in the plots (hASIC1a closed: 60.2 ± 2.2 REU, open: 45.9 ± 2.2 REU, desensitized: 54.5 ± 1.9 REU; lASIC closed: 46.5 ± 2.4 REU, open: 56.2 ± 3.6 REU, desensitized: 47.7 ± 2.6 REU).

Figure S2.

Comparison of Rosetta homology models of ASIC channels with experimental structures of cASIC1. (A and B) Homology models of hASIC1 (A) and lASIC (B) are compared with crystal structures of cASIC1 in closed (PDB accession no. 5WKV), open (PDB accession no. 4NTW), and desensitized (PDB accession no. 4NYK) conformations. The root-mean-squared distance (RMSD) deviations between all backbone (bb) or side chain (sc) heavy atoms resulting from the structural alignments in A and B are shown next to the structural models.

Figure S3.

Cartoon representations of xray–determined structures of cASIC1a- and Rosetta-predicted models of hASIC1a R64F. (A) Residue neighborhood around R65 (yellow) in the closed (PDB accession no. 5WKV), open (PDB accession no. 4NTW), and desensitized (PDB accession no. 4NYK) state structures of cASIC1a. The TM segments of two cASIC1a subunits are depicted as ribbons and colored blue and green, respectively. The third subunit is not shown for clarity. Residue side chains are depicted as sticks, and polar contacts are indicated as dashed lines. (B) Residue neighborhood around the substituted residue F64 (yellow) in Rosetta models of the closed, open, and desensitized state structures of hASIC1a-R64F.

Figure S4.

Breakdown of computationally predicted protein stability changes for hASIC1a-R64 and lASIC-W66 mutants by Rosetta score term. (A and B) For every amino acid substitution at residue 64 in hASIC1a (A) or residue 66 in lASIC (B), the contributions of the individual score terms in the RosettaMembrane energy function to the stability change of the closed, open, and desensitized state structure, respectively, are shown. LJ, Lennard-Jones; Elec., electrostatic interaction energy.

Figure 3.

R64^hASIC1a and R66^lASIC enable channel desensitization by the β11-β12 linker mechanism. (A) Representative lASIC current traces exhibiting sustained currents in the presence of pH 6.0. The substitution R66^lASIC in the outer segment of TM1 produces rapid and complete desensitization upon exposure to pH 6.0, which is reverted back to nondesensitizing when the β11-β12 linker mechanism is disabled by the mutation Q278G. (B) Increasing proton concentration response curves of activation (red curve) and SSD (blue curve) currents. Lines are the fit to the Hill function. Each data point represents four to seven independent measurements. (C) Typical proton-induced current of hASIC1a WT desensitizes completely. Channels with the substitution R64F exhibit slow and incomplete desensitization. Impairing function of the β11-β12 linker by the mutation Q276G eliminates desensitization in hAQSIC1a-R64F-Q276G. (D) Concentration response curves of channel activation (orange curve) and SSD (green curve) shown in C fit to the Hill function. Each data point represents four to five independent measurements.

Figure 4.

F64^hASIC1a exhibits long opening events and a subconductance state. (A) Representative example of an excised outside-out membrane patch containing a large number of channels activated by external pH 7.1 for 1 min. (B) Representative example of excised outside-out membrane patches containing a single channel activated by pH 7.2. Dashed lines indicate three current levels: closed (C), subconductance (S), and fully open (O). The subconductance state appears when channels transition to the open and to the closed states. Holding potential, −60 mV. External solution 100 mM Na⁺; pipette solution 100 mM K⁺. The shown trace represents 1 out of 15 patches, all showing the subconductance state. (C) All-points histogram of a continuous recording of activity of a single channel exhibits three conductance states of magnitude: 0, 0.7, and 1.5 pA. The line is the Gaussian fit to the data. (D) Histogram of duration of opening events in seconds. (E) Representative example of multichannel hASIC1a patch activated with pH 6.5. (F) Representative examples of unitary currents. (G) All-points histogram of a continuous recording of activity of a single channel exhibits two conductance states of magnitude: 0 and 1.5 pA. (H) Histogram of duration of opening events of WT hASIC1a.

Figure 5.

Large hydrophobic residues at positions 64/66 in hASIC1a/lASIC increase the apparent pH_50a of ASICs. (A) Proton concentration-response curves of hASIC1a with various amino acid substitutions in position R64. Each data point represents the mean of four or five independent measurements ± SD. Mutants were measured in different experiments, but each one always had as control a group of WT oocytes to ensure the pH of solutions for activation were correctly calibrated; thus, the number of cells of the WT curve is 38. Curves are fit to the Hill equation. Values of pH_50a and n (Hill coefficient) are presented in Table 1. (B) Proton concentration-response curves of lASIC1 with Arg or Cys substituting Trp-66. Lines in A and B are the fit of the data to the Hill equation. Each data point is the mean of four to six independent measurements. Error bars represent the SD. (C) Representative examples of currents activated by increasing concentration of protons of mASIC2a wt and the mutant mASIC2a-R63F and the corresponding concentration-response curves of activation. Each data point represents the average of at least five independent cells; error bars represent the SD. Lines are the fit to the Hill equation. (D) Representative examples of currents activated by increasing concentration of protons of mASIC3 wt and mutant mASIC3-R64F and the corresponding concentration-response curves of activation. Each data point represents the average of at least three independent cells; error bars represent the SD. Lines are fit to the Hill equation. (E) Simple kinetic schemes comparing WT and F64^hASIC1a response to high and low proton concentrations indicated by large- and small-type size [H⁺]. Only three states are shown: C, closed; O, open; D, desensitized. The thick red arrow indicates the favored pathway, i.e., opening rather than SSD. The thin blue arrows indicate transitions that are slowed or markedly unfavored (broken line) by F64. wt, wild type.

Figure 6.

Opening of lASIC, CH1, and F64^hASIC1a with low proton concentrations is associated with movement of the α5-helix. (A) Representative currents of lASIC P85C-D358C double mutant stimulated by pH 7.0 before and after treatment with 5 mM of the disulfide-reducing agent TCEP for 1 h. Dashed line is zero current level. (B) Maximal currents elicited by pH 7.0 in 10 independent cells expressing lASIC P85C-D358C before and after treatment with TCEP. Horizontal dashed lines indicate the mean current value. Asterisks represent significant statistical difference, **, P = 0.0003. (C) Representative currents of lASIC before and after treatment with TCEP for 1 h. (D) Values of maximal current elicited by pH 7.0 before and after TCEP treatment of 10 independent cells expressing lASIC.

Figure S5.

Effect of a reducing agent on the activity of S83C-Q358C CH1. (A) Representative example of whole cell currents from cells expressing CH1:P85C-Q358C channels activated with pH 7.0 before and after treatment with the reducing agent TCEP for 1 h. (B) Summary of current values of before (n = 12) and after treatment (n = 10) with TCEP reducing agent. Two oocytes were damaged and not included in the TCEP results. Asterisks indicate statistically significant difference between the two groups by t test; **, P < 0.001.