A hydrophobic gate. (a) Representative single-channel current traces were recorded in the absence of agonists from WT and L9′S mutants (different subunits). The mutations increase constitutive openings. (b) Representative, low-resolution unliganded single-channel current traces from AChRs having only background (bkg: αA96H + αY93H) or Ala mutation in different subunits at the gate on this bkg. Boxed insets: Higher-resolution traces. (c) Scatter plot showing a linear correlation between diliganded (ΔΔG₂) and unliganded (ΔΔG₀) gating free energy changes (27-point mutations at 9′, 12′, and 13′; Table S1). The straight-line slope is 1.0 ± 0.09 (R² = 0.9) indicating that the mutations do not affect the coupling constant. (d) Range-energies (ΔΔG_n values for different mutations obtained in the presence or absence of agonists) versus position in the pore. Solid lines are splines; dotted lines mark 9′, 13′, and 16′. The largest ranges are associated with 9′ and 13′. (e) Representative single-channel current traces and dwell-time distribution histograms from AChRs having a point mutation at 9′ position in the absence of agonists. The splines are fitted exponentials (dotted, component; solid, sum). (f) Scatter plots showing the correlation of ΔΔG₀ versus ΔΔG^transfer in kcal.mol⁻¹ for L9′/V13′ (each symbol is a different mutation). ΔG^transfer from Fauchere and Pliska (1983) or Radzicka et al. (1988). Slopes of the linear regression lines are 1.88 ± 0.2 (R² = 0.81) and 0.39 ± 0.04; dotted lines, 95% confidence limit.