NMDARu currents were always accompanied by an increase in [Ca2+]i (Figure 5Di). The kinetics of the NMDARu currents were rapid, mean time

to peak of 1.36 ± 0.29 ms (n = 6; Figure 5Dii). Importantly, we only observe NMDARu currents and their associated increase in [Ca2+]i when photolysis is directed at boutons. Directing the photolytic spot at points along the collateral failed to generate either. This is illustrated in Figures 5Ei and 5Eii, where an NMDARu current and increase in [Ca2+]i are seen at the bouton, whereas there find more is no response when the spot is moved 2 μm away from the bouton. Because both voltage-dependent relief of the Mg2+ block and glutamate binding are requisite steps for the activation of the NMDAR (Mayer et al., 1984 and Nowak et al., 1984), we used these features to explore the mechanism by which presynaptic NMDARs generate large Ca2+ transients. Initially, we increased the level of extracellular Mg2+ to 10 mM. Superfusion of 10 mM Mg2+ (Figures 6Aii and 6Aiii) significantly reduced the probability of observing a large event (ACSF

θ = 0.185 ± 0.075; 10 mM Mg2+ θ = 0.009 ± 0.018; n = 5; Figures 6Aii–6Aiv), whereas the amplitude of these events remains unchanged (Figure 6Av). In contrast, the absence of Mg2+ from the extracellular solution did not change the probability of observing a large Ca2+ event (ACSF θ = 0.19 ± 0.079; Mg2+-free θ = 0.197 ± 0.078; n =

5; Figure 6Biv) but did increase the amplitude of both large and small events (Figure 6Bv). We manipulated the release of glutamate INCB018424 from the boutons by modifying the duration of the AP. This was achieved by lowering the extracellular concentration of K+ ions to 0.1 mM, thereby reducing the duration of the AP, or by applying Urease 4-aminopyridine (4-AP, 40 μM) to increase AP duration (Qian and Saggau, 1999). As expected, low K+ conditions significantly decreased the width of the AP (ACSF: τ [ms] = 2.35 ± 0.01; 0.1 mM K+: τ = 1.65 ± 0.01; n = 4; p < 0.0001). With the duration of the AP reduced, the probability of observing large Ca2+ events was significantly decreased compared to control (ACSF θ = 0.178 ± 0.075; 0.1 mM K+ θ = 0.134 ± 0.043; n = 4; Figure 7Aiv). In contrast, 4-AP enhanced spike duration (ACSF: τ [ms] = 2.14 ± 0.07; 40 μM 4-AP: τ = 14.36 ± 2.7; n = 5; p < 0.0001) and significantly increased the probability of observing a large event (ACSF θ = 0.196 ± 0.063; 40 μM 4-AP θ = 0.006 ± 0.013; n = 5; Figure 7Biv). These results indicate that in normal K+ conditions, the depolarization arising from a single AP invading the bouton is adequate to relieve the Mg2+ block of the NMDAR, but this is not the case when the AP duration is curtailed. Enhancing the duration of the AP increases Ca2+ influx and consequently transmitter release (Mintz et al.