Substitution of N-terminal region of BD abolishes con action: refutation of sufficiency hypothesis. (A) Bars show substitutions in the N-terminal portion of the BD, with mean values of g_{(30 mM cGMP)}/g_{(3 mM cGMP)}, including previously reported data for Construct 6 (Young et al., 2001). (B) Dose–response relation of Construct 8 activated with cGMP and cAMP (down-triangles and up-triangles, respectively; n ≥ 4 for all points). Conductances were converted to P_o using the methods of Fig. 3, fixing P_o = 0.337 at 3 mM cGMP. Black curves are Hill equation fits (see Materials and methods). Fitted parameters (±SE) are: cGMP, K_1/2 = 368 ± 26 µM, h = 0.85 ± 0.04, and P_max = 0.399 ± 0.006; cAMP, K_1/2 = 520 ± 220 µM, h = 0.81 ± 0.23, and P_max = 0.58 ± 0.05. For comparison, gray curves show X-fA2 dose–response data (Young et al., 2001) for cGMP (solid) and cAMP (dashed). (C) Model of the C-linker and BD from X-fA2 (see Results). (Left) Ribbon diagram highlights helix αA and strand β1 (orange) with the three unconserved residues (magenta) whose substitution in Construct 8 abolished con action; numbering follows fCNGA2 sequence (Fig. 1 alignment) with the substituted amino acid from rCNGA4 in parentheses. Hash marks (#) indicate three loops (αB′-αC′, PB cassette, and β4-β5), which were built without template residues. The canonical-binding site (dashed ellipse) uses key ligand contacts with R529 and T530 (green), which are universally conserved in CNG channels. (Right) Electrostatic potential mapped to molecular surface.