Revealing molecular determinants governing mambalgin-3 pharmacology at acid-sensing ion channel 1 variants

Abstract

Acid-sensing ion channels (ASICs) are trimeric proton-gated cation channels that play a role in neurotransmission and pain sensation. The snake venom-derived peptides, mambalgins, exhibit potent analgesic effects in rodents by inhibiting central ASIC1a and peripheral ASIC1b. Despite their distinct species- and subtype-dependent pharmacology, previous structure-function studies have focussed on the mambalgin interaction with ASIC1a. Currently, the specific channel residues responsible for this pharmacological profile, and the mambalgin pharmacophore at ASIC1b remain unknown. Here we identify non-conserved residues at the ASIC1 subunit interface that drive differences in the mambalgin pharmacology from rat ASIC1a to ASIC1b, some of which likely do not make peptide binding interactions. Additionally, an amino acid variation below the core binding site explains potency differences between rat and human ASIC1. Two regions within the palm domain, which contribute to subtype-dependent effects for mambalgins, play key roles in ASIC gating, consistent with subtype-specific differences in the peptides mechanism. Lastly, there is a shared primary mambalgin pharmacophore for ASIC1a and ASIC1b activity, with certain peripheral peptide residues showing variant-specific significance for potency. Through our broad mutagenesis studies across various species and subtype variants, we gain a more comprehensive understanding of the pharmacophore and the intricate molecular interactions that underlie ligand specificity. These insights pave the way for the development of more potent and targeted peptide analogues required to advance our understating of human ASIC1 function and its role in disease.

Keywords: ASIC; Allosteric modulation; Electrophysiology; Gating modifier; Ligand selectivity; Protein-protein interaction; Specificity; Venom peptide.

Fig. 1

Multiple substitutions combined in rASIC1a alter Ma-3 pharmacology to resemble that at rASIC1b (a) Cryo-EM complex of Ma-1 bound to hASIC1a (PDB: 7CFT) showing a single ASIC1a subunit with domains coloured and labelled. (b) The complex structure showing: (i) the core pharmacophore of four residues on the thumb domain (red sticks). Although these residues are unlabelled, they are Tyr316, Asn320, Phe350, and Tyr358 using rASIC1a numbering, and (ii) residues at the interface of subunits that were mutated in this study are shown as brown sticks and labelled. Labels show rASIC1a residues which are all the same as hASIC1a except for A178 which is a valine in hASIC1a. (c) Concentration-response curves for Ma-3 at individual mutants of rASIC1a, and (d) combination mutants of rASIC1a. See Supplementary Table 2 for full Hill equation fits and statistical comparisons. (e) Maximal inhibition by 1 µM Ma-3 at each channel tested. All data use a conditioning pH of 7.45 and stimulating pH of 6. Welch’s one-way ANOVA with Dunnett’s multiple comparisons test where P < 0.05 is considered significant compared to rASIC1a (purple hash, #) and rASIC1b (green asterisk, *)

Fig. 2

Ma-3 induces distinct changes in the gating properties of rASIC1a, rASIC1b, and the rASIC1a SQRE mutant (a) Example traces showing current kinetics after Ma-3 application; 10 nM at rASIC1a, 300 nM at rASIC1b and rASIC1a SQRE mutant. Scale bar: abscissa 3 s, ordinate 1000 nA. (b–d) Average raw rise and decay times analysed from concentration-response data for (b) rASIC1a, (c) rASIC1b, and (d) rASIC1a SQRE mutant (conditioning pH 7.45 and stimulating pH 6). (e − g) The pH dependence of activation (up-pointing triangle, conditioning pH 8) and steady-state desensitisation (SSD; down-pointing triangle, stimulating pH 5) in the absence and presence of 100 nM Ma-3 for (e) rASIC1a, (f) rASIC1b, and (g) rASIC1a SQRE mutant. Details of the Hill equation fits are reported in Table 1. Solid lines represent pH dependence curves normalised to the maximal control current (left y-axis of I/I_control), and dashed lines are for data scaled to the maximal current observed in the presence of Ma-3. All data are mean ± SEM and n = 5–8

Fig. 3

Inter-domain coordination of conformational rearrangements during gating regulates the activity of Ma-3. (a) Schematic and structural representation of rat ASIC1a: ASIC1b chimeras. ASIC1a and ASIC1b are identical in the C-terminal two-thirds of the protein (grey box and line in the schematic) and only differ in the first third of the protein (residues 1–186 of rASIC1a and interchanged with rASIC1b). The schematic shows the N-terminal sequences from ASIC1b in orange, and the ASIC1a sequence in green. ASIC1a has a shorter N-terminus than ASIC1b, and the chimeras use the first amino acid of ASIC1a as the starting point. The extended N-terminal sequence of rASIC1b is intracellular. Bottom panels have the chimeric regions mapped onto the Ma-1:hASIC1a complex structure (PDB: 7CFT) of one monomer in the trimer, with the identical region shown in grey (the additional two monomers are coloured pink and cyan). (b) Activity of 300 nM Ma-3 at each wild-type and chimeric channel using a conditioning pH of 7.45 and stimulating pH of 5. (c) Comparison of the apo hASIC1a structure (PDB: 7CFS) and Ma-1 bound hASIC1a (PDB: 7CFT) highlighting the different position of Asp355 (green stick), and its proximity to Arg175 (brown stick), under these conditions. The structures were aligned, and distance measurements are made in PyMol 2.6. (d) Concentration-response curves for Ma-3 at rASIC1a D355A (IC₅₀ = 15.38 nM; pIC₅₀ 95% CI = 7.76–7.86; slope 95% CI = 1.69–1.26), and comparison to the R175C mutant and wild-type channels. (e) Concentration-response curves for Ma-3 at rASIC1b mutants. See Supplementary Table 3 for full Hill equation fits and statistical comparisons. For panels d and e, the conditioning pH was 7.45 and stimulating pH 6. All data in graphs are mean ± SEM and n = 5–7

Fig. 4

Comparison of the Ma-3 pharmacophore between rASIC1a and rASIC1b. (a) Amino acid sequence of Ma-3, with circles below residues that were mutated to alanine in this study. The background colouring highlights the residues that form the three finger loops of mambalgins and are colour coordinated with panel b. (b) Structure of Ma-1 from the cryo-EM complex model (PDB: 7CFT) highlighting the three finger loops. (c) pIC₅₀ values from Hill equation fits for concentration-response data of Ma-3 and mutants at rASIC1a (upper panel) and rASIC1b (lower panel). Dashed bars represent data for which complete fits are not possible due to lack of higher concentration data (see Supplementary Fig. 4). Statistics with Welch’s one-way ANOVA with Dunnett’s multiple comparisons test. (d) Ma-1 structure as in panel b, with side chains of residues that show a > 10-fold loss in IC₅₀ values compared to wild-type for rASIC1a and rASIC1b are shown in blue, Lys8 in orange, and Met16 in purple. (e) Concentration-response curves for Ma-3 wild type, K8A, and M16A at rASIC1a, hASIC1a and rASIC1b. (f) Example current traces at rASIC1a (30 nM peptide) and rASIC1b (300 nM peptide). Scale bar: abscissa 4 s, ordinate 500 nA. (g) Complex structure highlighting labelled channel residues within 5 Å of K8 (orange, left box) and M16 (purple, right box). (h) Activity of 300 nM Ma-3 WT (wild-type), K8A, and M16A mutants at rASIC1a, rASIC1a SQRE (S83T/Q84E/R175C/E177G), and rASIC1b. (i) Activity of peptides as in panel h, towards rASIC1b mutants (note: TECG is T128S/E129Q/C208R/G210E and the reverse mutant to rASIC1a SQRE). P values above bars in panels h and i, are determined by a two-way ANOVA with Dunnett’s correction for multiple comparisons. All data use a conditioning pH of 7.45 and stimulating pH of 6, are n = 5–8, and with full Hill equation fits and statistical comparisons reported in Supplementary Tables 4–6

Fig. 5

Mutation of residue 291 in the lower thumb domain can explain the Ma-3 potency shift between rat and human ASIC1a. (a–b) The pH-dependence of activation for (a) rASIC1a and (b) hASIC1a in control (grey), in the presence of 100 nM Ma-3 in the pH 8 conditioning solution (black and pink), and 100 nM Ma-3 scaled to its own maximum (dotted line without symbols). (c) Cryo-EM complex of Ma-1 bound to hASIC1a (PDB: 7CFT) with mutated residues labelled using rASIC1a numbering. Asn291 of rASIC1a is a Lys in hASIC1a, and the Asp298 and Leu299 residues shown are insertions in hASIC1a relative to rASIC1a (side chains not built in the cryo-EM model). (e–f) Concentration-response curves for Ma-3 at (d) rASIC1a individual substitutions that differ between species variants, (e) local mutations in rASIC1a spatially close to position 291, and (f) the reverse substitution at position 291 in the hASIC1a background (all concentration-response data use a conditioning pH of 7.45 and stimulating pH of 6). All data are mean ± SEM and n = 5–6. See Supplementary Tables 7–9 for full Hill equation fits and statistical comparisons

Fig. 6

Two amino acid substitutions define the unique pharmacology of Ma-3 at hASIC1b. (a) Cryo-EM complex of Ma-1 bound to hASIC1a (PDB: 7CFT). Residues that determine the rat to human ASIC1b selectivity are shown as sticks and labelled. Labels use rASIC1b residues and numbering. Cys208 is a Gln in hASIC1b, and the equivalent of Arg175 in hASIC1a. Asn324 is a Lys in hASIC1b, and the equivalent of Lys291 in hASIC1a. (b–f) Concentration-response curves (top) and pH-dependence of activation (bottom) for (b) rASIC1b, (c) hASIC1b, (d) rASIC1b C208Q, (e) rASIC1b N324K, and (f) rASIC1b C208Q/N324K. Concentration-response curves are performed using pH 6 stimulus (open symbols and dashed lines) and pH 5 stimulus (solid symbols and lines). pH-dependence of activation data is shown for control (black; each respective channel tested for that figure panel in the absence of Ma-3 – i.e. for panel b, control is rASIC1b without Ma-3), in the presence of 300 nM Ma-3 (coloured), and 300 nM Ma-3 scaled to its own maximum (coloured dashed line without symbols). All data use a conditioning pH of 7.45, are mean ± SEM, and n = 5–6. See Supplementary Tables 10 and 11 for full Hill equation fits and statistical comparisons

Fig. 7

The Ma-3 pharmacophore at hASIC1b. (a) Concentration-response curve of Ma-3 at hASIC1b activated by pH 6 and pH 5 (n = 5–6). (b) Ma-3 and mutants (1 µM) tested for activity at hASIC1b activated by pH 6 (top) and pH 5 (bottom). Violet indicates statistically significant differences from Ma-3 (Welch’s one-way ANOVA with Dunnett’s multiple comparisons test). Data are n ≥ 5 with all points shown to give exact replicates. All data use a conditioning pH of 7.45, and are mean ± SEM. See Supplementary Tables 12 and 13 for details of inhibition and statistical comparisons of panel b data

Fig. 8

Residues surrounding the core pharmacophore of mambalgins determine selectivity. (a) Cryo-EM complex of Ma-1 (grey) bound to hASIC1a trimer (pink, brown, and yellow monomers; PDB: 7CFT). (b) Zoomed in inset of black box from panel a, with mambalgin molecules removed. Side chains of the conserved ASIC1 pharmacophore that drives the binding affinity is shown in red. The three peripheral regions identified here that modulate mambalgin activity are shown in green. All side chains are labelled using rASIC1a residues and numbering. (c) The same image as in panel b, but also showing Ma-1 (backbone in grey). Mambalgin residues in blue are pharmacophore residues that are important for activity at ASIC1a and ASIC1b. Distant from this cluster are Lys8 (orange) and Met16 (magenta) that are important for the inhibition of rASIC1b but not rASIC1a. See Supplementary Fig. 6 for sequence alignment highlighting positions mutated in this study