Gabapentin efficacy was highest at −60 to −40 mV, leading to a −9 mV shift in the voltage dependence of KCNQ2/3 activation (1 µM gabapentin), but gabapentin also augmented currents at positive membrane potentials (Fig. 3, B and E). Dose-response studies showed that at −60 mV, gabapentin exhibited an EC50 value for KCNQ2/3 activation of 4.2 ± 0.13 nM (n = 5–7); at 10 nM, gabapentin increased KCNQ2/3 current 3.5-fold at −60 mV (Fig. 3F; Supplemental Fig. 1; Supplemental Table 1). The ability to activate KCNQ2/3 at subthreshold potentials enabled gabapentin to shift the membrane potential (EM) of KCNQ2/3-expressing oocytes by > −10 mV (EC50, 4.2 nM) (Fig. 3G). Parallel studies showed that pregabalin failed to activate KCNQ2/3 even at 1 µM, and began to inhibit KCNQ2/3 at 10 µM and above (Fig. 3, E and F; Supplemental Fig. 2; Supplemental Table 2). Pregabalin likewise failed to shift the oocyte EM (Fig. 3G). Compared with the established KCNQ2/3 opener and anticonvulsant retigabine, gabapentin acted as a potent partial agonist. Thus, retigabine (30 µM) shifted the voltage dependence of KCNQ2/3 activation by −30 mV (Fig. 3H) and increased current at −60 mV by 6-fold (Fig. 6I), however, the EC50 value for retigabine was in the micromolar (not nanomolar) range (Fig. 3J), which is ∼1000-fold less potent than gabapentin (see Supplemental Fig. 3; Supplemental Table 3). In comparison, we recently found that GABA, which also acts at KCNQ3-W265, activates KCNQ2/3 with an EC50 value of 0.85 µM at −60 mV, increasing the current by 4-fold (Manville et al., 2018). Thus, gabapentin and GABA exhibit similar efficacy but gabapentin is 200-fold more potent.