If glutamate binds more tightly to the desensitized state, steady-state desensitization should occur at lower glutamate concentrations. We measured GSI-IX price the IC50 for desensitization by glutamate using concentration jumps, with pre-exposure to a

range of glutamate concentrations (see Experimental Procedures). For GluA2, as previously reported ( Plested and Mayer, 2009), the half-inhibitory concentration (IC50) was more than 100-fold lower than the EC50 for activation (9 ± 1 μM; n = 3–11 patches per point; Figure 4B). Glutamate is even more potent at inhibiting the slow recovering GluK2 receptor (IC50 = 700 ± 80 nM, n = 3–8 patches). Consistent with the much slower recovery of the TR mutant, glutamate also blocked activation potently, at about 1,000-fold lower concentration

than the EC50 for activation (IC50 = 240 ± 30 nM; n = 5–7 patches). The GluA2 Y768R single mutant was also more potently inhibited by glutamate than wild-type GluA2 (IC50 = 3 ± 0.3 μM, data not shown). These data demonstrate that the GluA2 TR mutant and GluK2 bind glutamate much more tightly in the desensitized state than wild-type GluA2 does. For our panel of GluA2 mutants, we also measured the rate of deactivation following a 1 ms pulse of saturating glutamate. This experiment approximates synaptic transmission, where the glutamate transient decays in about 1 ms (Clements et al., 1992). Slow-recovering receptors (for example, the GluA2 TR double mutant) had slower deactivation decays than wild-type GluA2 (Figure 4C and Table 1). We plotted the deactivation rate

and the desensitization rate of each mutant against the recovery rate (Figure 4D). Strikingly, ABT-199 order the deactivation rate and the recovery rate were strongly correlated (Pearson r = 0.82). The correlation also held for mutants where recovery was faster than wild-type channels, which tended to have faster deactivation. There was little correlation between the rate of recovery and rate of entry to desensitization (Pearson r = 0.01), which varied less than 2-fold across the Metformin entire panel (range 89–159 s−1, Table 1 and Table S1). The correlation between deactivation and recovery rates, accompanied by modest changes in glutamate apparent affinity, suggests that mutations in D2 might alter activation gating, in particular by lengthening apparent openings. We investigated this hypothesis by recording the activations of individual wild-type and A2 TR channels in 10 mM glutamate (Figure 5A). Long (8 s) applications of 10 mM glutamate to patches containing 5–100 channels produced an initial peak response, followed by well-spaced activations with two or three subconductance levels (Figure 5B), and a rare full conductance level, as previously reported (Zhang et al., 2008). The mean gap between activations of a single channel (corrected for the number of channels, as estimated from the peak response) was about 500 ms for WT (n = 4 patches) and about 2,500 ms for TR (n = 4).