Fig. 3. Agonist unbinding rates contribute to deactivation and affinity. A, A previously established kinetic model was modified and used to estimate the microscopic unbinding rate constants for different agonists. The development and performance of the model are described in detail in Jones and Westbrook (1995,1997). B, Agonist-specific patch current deactivation (noisy lines) can be simulated (smooth lines) solely by changes in the unbinding rate. The model was optimized to fit 5 and 505 msec saturating GABA pulses and was then allowed to fit the deactivation phases for different agonists with only the unbinding rate (koff) as a free parameter. Records were averaged, normalized to the same maximum open probability (Po
) (Jones and Westbrook, 1995, 1997), and aligned at the peak. The rates were (in sec−1)kon = 5 × 106m−1, α1 = 1100, β1= 200, α2 = 142, β2 = 2500,d1 = 13, r1 = 0.2, d2 = 1250,r2 = 25, p = 2, andq = 10−2m−1. The asterisk denotes a net counterclockwise motion at steady state (see Materials and Methods). The best-fitting unbinding rates were (in sec−1) 40 for muscimol, 131 for GABA, 1125 for THIP, and 4500 for β-alanine.C, D, The unbinding rate was closely correlated with both the patch-current deactivation rate (C) and the apparent affinity (D). Unbinding rates from fitting 5 msec (closed circles and solid line) and 505 msec (open circles and dashed line) pulse responses are shown with SEM bars for the y-axis and 95% confidence limits of the fit for the x-axis. InC, the lines are regression fits to the power function: 1/τ =akoffb, wherea was 0.56 and b was 0.61 for 5 msec pulses and a was 0.90 and b was 0.53 for 505 msec pulses. In D, EC50 =akoffb, wherea was 1.5 × 10−7 andb was 1.18 for 5 msec pulses and a was 3.2 × 10−7 and b was 1.05 for 505 msec pulses.