Recovery from paired pulse depression of fast HC EPSCs was slowed by inhibiting endocytosis with dynasore, dyngo-4A, or DIP. (A) Recovery from paired pulse depression of the fast EPSC component plotted as a function of interpulse interval (control, n = 11 pairs; dynasore [80 µM], n = 15 pairs; dyngo-4A [30 µM], n = 10; DIP [1 mM], n = 7). The amplitude of the second EPSC was normalized to the first EPSC in each pair. Recovery of the fast EPSC component was fit with a double exponential function. (B) The recovery observed with shorter interpulse intervals is replotted with the data points for dyngo-4A offset by 0.1 s from the others for ease of comparison. In control conditions, the fast time constant for recovery from paired pulse depression averaged 618 ms (solid lines and filled circles), slowing to 2.52, 3.24, or 3.06 s, respectively, in recordings where dynasore (dashed lines and open circles), dyngo-4A (dotted lines and downward triangles), or DIP (long dashes and upward triangles) were included in the rod patch pipette solution. The fast time constants obtained with dyngo-4A (P < 0.001, F test) and DIP (P = 0.0003, F test) were significantly slower than control. (C) Recovery from paired pulse depression of slow EPSC components was fit with a single exponential function. Time constants did not differ significantly among the four conditions (control, 732 ms; dynasore, 635 ms; dyngo-4A, 1,056 ms; DIP, 584 ms; P = 0.25, F test). (D) Paired pulse responses of I_Ca charge transfer (Q_Ca) showed a slight enhancement with short intervals and no paired pulse depression. Paired pulse responses of Q_Ca also did not differ among the four conditions. Error bars represent ±SEM.