Minimal molecular determinants of isoform-specific differences in efficacy in the HCN channel family

Figure 1.. Voltage dependence and sequence comparison of HCN1, HCN2 and chimeric HCN2 with C-linker and CNBD of HCN1. (A) Representative current traces in the inside-out configuration and conductance–voltage curves obtained in the presence and absence of 10 µM cAMP. The same patch was recorded in both conditions. Tail currents from HCN2 (n = 19 patches), HCN2/1 (n = 19), and HCN1 (n = 21) were normalized to the maximum fitted tail current recorded in the presence of cAMP for each patch. Data presented are mean ± SEM. (B) Sequence alignment and secondary structures of mouse HCN1–4. The 18 residues that differ between HCN1 and HCN2 are highlighted in red.

Figure 2.. C-linker residues critical for isoform-specific modulation of HCN2 by cAMP. Top: Alignment of the C-linker of HCN1 and HCN2 illustrating in red the residues that were found to have effects on the response of HCN2 to cAMP. The remaining residues that differ between isoforms are shown in blue. (A–G) Normalized conductance–voltage curves in the presence and absence of cAMP obtained by measuring tail currents of WT HCN2 and various C-linker substitutions with equivalent HCN1 residues. Mutations that significantly decreased the cAMP-dependent ΔV1/2 are labeled red. (H) Summary of the shifts in the midpoints of activation (ΔV1/2) induced by cAMP for the mutants shown in A–G. *, P = 0.02; **, P < 0.00003 versus WT HCN2. The numbers of patches recorded per mutant were HCN2 (n = 19), D489E G493S (n = 21), D489E G497D G493S (n = 21), K534R (n = 18), M485I K534R (n = 14), S514T K534R (n = 19), and M485I G497D S514T (n = 18). Data presented are mean ± SEM.

Figure 3.. CNBD residues critical for isoform-specific modulation of HCN2 by cAMP. Top: Alignment of the CNBD of HCN1 and HCN2 illustrating in red the residues that were found to have effects on the response of HCN2 to cAMP. The remaining residues that differ between isoforms are shown in blue. (A–E) Normalized conductance-voltage curves in presence and absence of cAMP obtained by measuring tail currents of HCN2 carrying various CNBD substitutions. Labeled red are CNBD mutations in the background of the previously identified nonneutral C-linker mutations. (F) Summary of the shifts in the midpoints of activation (ΔV1/2) induced by cAMP for the mutants shown in A–E. The background of the CNBD mutations tested (the triple mutant M485I G497D S514T) is abbreviated by a plus sign (+) in the mutant labels. *, P < 0.04; **, P < 0.00001 versus the triple mutant M485I, G497D, and S514T. The numbers of patches recorded per mutant were M485I G497D S514T (n = 18), M485I G497D S514T V562A S563G (n = 21), M485I G497D S514T L565I (n = 22), M485I G497D S514T L565I (n = 21), M485I G497D S514T S575T (n = 21), and HCN2/1 (n = 19). Data presented are mean ± SEM.

Figure 4.. The voltage-dependent modulation of the HCN1minimal mutant (with substitutions M485I, G497D, S514T, V562A, and S563G) is similar to the HCN2/1 chimera. (A) Representative traces of inside-out recordings for the HCN1minimal mutant in the absence and presence of 10 µM cAMP. For comparison, because of the slow kinetics in this mutant, recordings obtained with 10-s voltage pulses to ensure that the currents are saturated even in absence of cAMP are shown. (B) Conductance–voltage curves for the HCN1minimal mutant normalized to the maximum tail current in the presence of cAMP. (C) Tail currents at saturating voltages in absence of cAMP for WT HCN2 (n = 14 patches), WT HCN1 (n = 21), HCN2/1 (n = 19), and the HCN1minimal mutant (n = 21), calculated relative to the maximum tail current in saturating voltages and cAMP for each patch. *, P = 0.01; **, P < 0.00003 compared with WT HCN2. (D) Additional mutations L565I and S575T do not affect cAMP-dependent ΔV1/2 in the HCN1minimal (Min) background. n = 19 for WT HCN2, 21 for HCN1minimal, 19 for Min + L565I, 18 for Min + S575T, 17 for M485I G497D S514T L565I S575T, 20 for Min + L565I S575T, and 19 for HCN2/1. *, P = 0.03; **, P = 5 × 10−17…

Figure 5.. The extent of cAMP-induced shifts correlates with intrinsic midpoints of activation of various mutants. Plot of the cAMP-induced shifts in V1/2 (ΔV1/2) against the V1/2 in the absence of cAMP for all mutants reported in this study. The correlation coefficient is 0.92.

Figure 6.. Allosteric models recapitulate the various voltage- and ligand-dependent properties of WT HCN2, the HCN2/1 chimera, and the HCN1minimal mutant. (A) Schematic representation of the various allosteric elements and interactions implemented in Scheme 4. Binary elements for the pore (P), voltage sensors (VS), linker (L), and binding domain (CNBD) are characterized by equilibrium constants (Keq) between two states. Coupling factors (Θ) characterize the interactions between elements. All the parameters, except those marked in red, were kept constant to fit the kinetic and steady-state data for WT HCN2, HCN2/1, and the HCN1minimal mutant. (B–D) Model fits (red) and experimental data (black) of current traces in response to voltage steps in the presence and absence of cAMP. Reference data for the apparent Po-voltage curves (right) were obtained from the steady-state conductances at the end of each pulse and are compared with the maximum probability of the open state calculated for the model. Refer to Tables S2, S3, and S4 for a complete list of all parameters.