Synaptic Zn2+ inhibits neurotransmitter release by promoting endocannabinoid synthesis

Abstract

Although it is well established that many glutamatergic neurons sequester Zn(2+) within their synaptic vesicles, the physiological significance of synaptic Zn(2+) remains poorly understood. In experiments performed in a Zn(2+)-enriched auditory brainstem nucleus--the dorsal cochlear nucleus--we discovered that synaptic Zn(2+) and GPR39, a putative metabotropic Zn(2+)-sensing receptor (mZnR), are necessary for triggering the synthesis of the endocannabinoid 2-arachidonoylglycerol (2-AG). The postsynaptic production of 2-AG, in turn, inhibits presynaptic probability of neurotransmitter release, thus shaping synaptic strength and short-term synaptic plasticity. Zn(2+)-induced inhibition of transmitter release is absent in mutant mice that lack either vesicular Zn(2+) or the mZnR. Moreover, mass spectrometry measurements of 2-AG levels reveal that Zn(2+)-mediated initiation of 2-AG synthesis is absent in mice lacking the mZnR. We reveal a previously unknown action of synaptic Zn(2+): synaptic Zn(2+) inhibits glutamate release by promoting 2-AG synthesis.

Figure 1.

Zn²⁺ reduces synaptic strength by lowering Pr. A, Time course of EPSC peak amplitude reversible depression by 100 μm, 5 min Zn²⁺ application. (During Zn²⁺ application: 67.9 ± 9.0% of baseline, n = 7, p < 0.01; after Zn²⁺ washout: 94.5 ± 9.5% of baseline, n = 7, p = 0.7). Top, Representative averaged traces of EPSCs before (pre), during Zn²⁺, and after removal of Zn²⁺ (washout). B, PPR (20 ms interstimulus interval) is increased after Zn²⁺ application (PPR: baseline: 1.96 ± 0.11; after Zn²⁺: 2.42 ± 0.13; n = 7, p < 0.01). C, Zn²⁺-induced EPSC depression is associated with a decrease in Pr. 1/CV² is decreased after Zn²⁺ (1/CV² in Zn²⁺: 0.38 ± 0.1% of baseline, n = 7, p < 0.01). Effect of baclofen (2–5 μm) on PPR and 1/CV² (PPR: baseline: 2.00 ± 0.15; after baclofen: 2.42 ± 0.10; n = 6, p < 0.01; 1/CV² in baclofen: 0.27 ± 0.07% of baseline, n = 6, p < 0.01). Effect of NBQX (0.5 μm) on PPR and 1/CV² (PPR: baseline: 1.84 ± 0.25; after NBQX: 1.89 ± 0.08; n = 6, p = 0.47; 1/CV² in NBQX: 0.90 ± 0.08% of baseline, n = 8, p = 0.55). D, Cumulative probability plot of mEPSC frequencies shows that Zn²⁺ application decreased frequencies (left). Cumulative probability plot of mEPSC amplitudes shows Zn²⁺ application did not affect amplitudes (right). Top, Sample traces of mEPSCs before and after Zn²⁺ application (mean mEPSC frequency: baseline: 1.89 ± 0.3 Hz; after Zn²⁺: 1.22 ± 0.5 Hz; n = 4, p < 0.05 paired t test; mean mEPSC amplitude: baseline: 13.79 ± 0.81 pA; after Zn²⁺: 13.35 ± 0.94 pA; n = 4, p > 0.12 paired t test). Summary data represent mean ± SEM. Asterisks indicate statistically significant differences, p < 0.05.

Figure 2.

Zn²⁺-induced decrease in Pr is mediated by endocannabinoid signaling. A, Zn²⁺-induced EPSC depression is blocked by 1 μm AM-251 (EPSC amplitude in control: 66.4 ± 4.8% of baseline, n = 8, p < 0.01; in AM-251: 91.4 ± 7% of baseline, n = 6, p = 0.3). B, PPR (20 ms interstimulus interval) is increased after Zn²⁺ application in control conditions, but this increase is blocked in the presence of AM-251 (PPR in control: baseline: 1.88 ± 0.12; after Zn²⁺: 2.36 ± 0.11, n = 8, p < 0.01; PPR baseline in AM-251: 1.86 ± 0.2; after Zn²⁺: 1.91 ± 0.12; n = 6; p = 0.19. C, 1/CV² is decreased after Zn²⁺ application in control conditions, but this decrease is blocked in the presence of AM-251 (1/CV² in Zn²⁺: 0.45 ± 0.12%, n = 8, p < 0.01; 1/CV² in Zn²⁺ + AM-251: 0.97 ± 0.11% of baseline, n = 6, p = 0.78). D, Time course of baseline EPSC amplitude before and after application of AM-251 (EPSC amplitude after AM-251 application: 107.6 ± 10.2% of baseline, n = 5, p = 0.34). Summary data represent mean ± SEM. Asterisks indicate statistically significant differences, p < 0.05.

Figure 3.

Zn²⁺-induced decrease in Pr requires rises in postsynaptic Ca²⁺ and postsynaptic activation of a GPCR pathway. A, Zn²⁺-induced EPSC depression is blocked by 10 mm intracellular BAPTA or 0.5 mm intracellular GDP-β-S (BAPTA: EPSC amplitude: 99.9 ± 8.5% of baseline, n = 5, p = 0.3; GDP-β-S: EPSC amplitude: 105.87 ± 4.39% of baseline, n = 7, p = 0.38). Our previous studies have shown that GDP-β-S does not affect parallel fiber EPSCs in fusiform cells; yet, application of GDP-β-S decreases the input resistance of fusiform cells (Zhao and Tzounopoulos, 2011). Thus, to avoid possible complications in the interpretation of our findings due to changes in input resistance and to allow diffusion of GDP-β-S in fusiform cells, we investigated the effect of GDP-β-S on Zn²⁺-mediated modulation of Pr after the GDP-β-S-mediated effect on input resistance reached a steady state: Zn²⁺ application took place ∼30 min after break-in of the cell. B, PPR (20 ms interstimulus interval) was not increased after Zn²⁺ application in the presence of BAPTA and GDP-β-S (BAPTA PPR: baseline: 1.95 ± 0.1; PPR after Zn²⁺: 1.92 ± 0.1; n = 5; p = 0.49; GDP-β-S: PPR baseline: 1.83 ± 0.1; PPR after Zn²⁺: 1.82 ± 0.1; n = 7; p = 0.8). C, 1/CV² was not decreased after Zn²⁺ application in the presence of BAPTA and GDP-β-S (BAPTA: 1/CV²: 0.8 ± 0.12% of baseline, n = 5, p = 0.61; GDP-β-S: 1/CV²: 0.87 ± 0.13% of baseline, n = 7, p = 0.78). Summary data represent mean ± SEM.

Figure 4.

mZnR, a Zn²⁺-sensing G_q-protein-coupled receptor, mediates Zn²⁺-induced increase in intracellular Ca²⁺ in DCN neurons. A, Intracellular Ca²⁺ rise was monitored following a 100 Hz, 1 s stimulation of the molecular layer of the DCN in the presence or absence of tricine (10 mm, top traces). Slices from WT and KO mice were loaded with fura-2. The conditions for the different traces are indicated at the bottom of the bar graph. Bar graph showing averaged changes of Ca²⁺ rises. The residual Ca²⁺ response in the slices from mZnR KO and ZnT3 KO mice is not further attenuated by tricine and is similar to the response observed in the presence of tricine in slices from WT mice (ΔF/F: mZnR WT Control: 0.062 ± 0.005, n = 8; mZnR KO Control: 0.027 ± 0.003, n = 9, p < 0.01; mZnR WT in tricine: 0.023 ± 0.01, n = 9, p < 0.01 when compared with mZnR WT control; mZnR KO in tricine: 0.027 ± 0.003, n = 7, p = 0.48 when compared with mZnR WT in tricine; ZnT3 WT control: 0.064 ± 0.006, n = 4; ZnT3 KO control: 0.028 ± 0.004, n = 8, p < 0.01 when compared with ZnT3 WT control; ZnT3 WT in tricine: 0.034 ± 0.01, n = 8, p < 0.01 when compared with WT control; ZnT3 KO in tricine: 0.026 ± 0.004, n = 8, p = 0.52 when compared with ZnT3 KO control). B, Representative response illustrating extracellular Zn²⁺-dependent changes in 2 μm Newport Green (a nonpermeable, extracellular Zn²⁺ sensor) fluorescence in the DCN molecular layer after 1 s, 100 Hz stimulation of the parallel fibers in a slice from WT, ZnT3 KO, and mZnR KO mice. C, Average responses showing that Newport Green fluorescence in response to 1 s, 100 Hz stimulation of the parallel fibers was abolished in ZnT3 KO mice, but was normal in mZnR KO mice (ΔF: ZnT3 WT: 0.018 ± 0.004, n = 7 slices; ZnT3 KO: 0.001 ± 0.0004, n = 11 slices, p < 0.01 compared with ZnT3 WT; mZnR KO: 0.015 ± 0.004, n = 8 slices). Summary data represent mean ± SEM. Asterisks indicate statistically significant differences, p < 0.05.

Figure 5.

mZnR is necessary for Zn²⁺-induced decrease in Pr. A, Zn²⁺-induced reversible depression of EPSCs was observed in WT mice but not in mZnR KO littermates (EPSC amplitude mZnR WT: 65.7 ± 8% of baseline, n = 6, p < 0.01; mZnR KO: 107.4 ± 5.6% of baseline, n = 5, p = 0.27). B, PPR (20 ms interstimulus interval) is increased after Zn²⁺ application in mZnR WT mice but is unaffected in mZnR KO littermates (mZnR WT PPR baseline: 1.97 ± 0.05; PPR after Zn²⁺: 2.32 ± 0.06; n = 6, p < 0.01; mZnR KO: PPR baseline: 1.88 ± 0.07; PPR after Zn²⁺: 1.90 ± 0.03; n = 5, p = 0.75). C, 1/CV² is decreased after Zn²⁺ application in mZnR WT mice but is unaffected in mZnR KO littermates (mZnR WT 1/CV²: 0.36 ± 0.11% of baseline, n = 6, p < 0.01; mZnR KO: 1/CV²: 0.79 ± 0.11% of baseline, n = 5, p = 0.15). Summary data represent mean ± SEM. Asterisks indicate statistically significant differences, p < 0.05.

Figure 6.

Quantal properties and depolarization-induced endocannabinoid signaling are not altered in mZnR KO mice. A, Cumulative probability plot of mEPSC frequencies shows that frequencies are not different between mZnR WT and KO mice (mean mEPSC frequency mZnR WT: 3.03 ± 0.45 Hz, n = 5; mZnR KO: 3.52 ± 0.59 Hz; n = 5, p = 0.51). Top, Sample traces of mEPSCs from mZnR WT and KO mice. B, Cumulative probability plot of mEPSC amplitudes shows that amplitudes are not different between mZnR WT and KO mice (mean mEPSC amplitude mZnR WT: 16.62 ± 2.09 pA, n = 5; mZnR KO: 15.56 ± 0.65 pA; n = 5, p = 0.64). C, Cumulative probability plot of mEPSC rise times shows that rise times are not different between mZnR WT and KO mice (mean mEPSC rise time mZnR WT: 0.63 ± 0.05 ms, n = 5; mZnR KO: 0.66 ± 0.05 ms; n = 5, p = 0.61). D, Cumulative probability plot of mEPSC decay times shows that decay times are not different between mZnR WT and KO mice (mean mEPSC decay time: mZnR WT: 1.3 ± 0.09 ms, n = 5; mZnR KO: 1.43 ± 0.07 ms; n = 5, p = 0.28). E, Time course of DSE, induced by 5 s depolarization to 10 mV in mZnR WT and KO mice (mZnR WT DSE: 37 ± 3.1% of control, n = 5; mZnR KO DSE: 41 ± 5.3% of control, n = 6, p = 0.35). F, Time course of 2 μm WIN on baseline synaptic transmission of mZnR WT and KO mice (mZnR WT: 31.8 ± 0.77% of baseline, n = 5, mZnR KO: 30.84 ± 1.9%, n = 5, p = 0.32). Summary data represent mean ± SEM.

Figure 7.

Zn²⁺-induced decrease in Pr is mediated by a PLC- and DGL-dependent metabolic pathway. A, Zn²⁺-induced reversible depression of EPSCs was blocked by bath application of 5 μm U73122 or 10 μm THL (EPSC amplitude control: 67.60 ± 6.10% of baseline, n = 5, p < 0.01; U73122: 98.30 ± 3.66% of baseline, n = 6, p = 0.33; THL: 105.40 ± 7.70% of baseline, n = 6, p = 0.42). B, PPR (20 ms interstimulus interval) is increased after Zn²⁺ application in control conditions, but this increase is blocked in the presence of either U73122 or THL (PPR Control baseline: 1.91 ± 0.08; after Zn²⁺: 2.44 ± 0.16; n = 5; p < 0.01; U73122 baseline: 2.09 ± 0.08; after Zn²⁺: 2.06 ± 0.09; n = 6; p = 0.59; THL baseline: 1.78 ± 0.11; after Zn²⁺: 1.72 ± 0.10; n = 6; p = 0.3). C, 1/CV² is decreased after Zn²⁺ application in control but this decrease is blocked in the presence of either U73122 or THL (1/CV²: after Zn²⁺ 0.46 ± 0.11% of baseline, n = 6, p < 0.01; + U73122 1.27 ± 0.24% of baseline, n = 6, p = 0.41; + THL: 1.13 ± 0.26% of baseline, n = 6, p = 0.56). D, Bath application of U73122 does not affect baseline synaptic transmission (EPSC amplitude: 104.9 ± 2.9%, of baseline, n = 5, p = 0.25). E, Bath application of THL does not affect baseline synaptic transmission (EPSC amplitude: 101.5 ± 2.1%, n = 5, p = 0.35). F, Bath application of either U73122 or THL does not affect 1/CV² (1/CV² U73122: 1.03 ± 0.08% of baseline, n = 5, p = 0.52; THL: 0.99 ± 0.12% of baseline, n = 5, p = 0.59). Summary data represent mean ± SEM. Asterisks indicate statistically significant differences, p < 0.05.

Figure 8.

LC/MS/MS analysis reveals that Zn²⁺ induces a mZnR-dependent increase of 2-AG synthesis in the DCN. A, The two different AG isomers (2-AG and 1-AG) were chromatographically resolved and base peak separations were achieved. 2-AG and 2-AG-d8 were detected following the 379.3 to 287.2 m/z and 387.3 to 295.2 transitions, respectively, using a triple quadrupole. In parallel accurate mass determination using a Velos Orbitrap resulted in mass confirmations at the 2 ppm level for standards and for tissue-obtained 2-AG (inserts). B, 2-AG identity was confirmed by co-elution with internal standard (A) and by high-resolution MS and MS/MS data obtained on the 2-AG peak, where a similar fragmentation pattern was observed between the 2-AG standard and the DCN tissue samples. C, Molecular structures of product ions obtained upon ion trap collision-induced dissociations of 2-AG shown in B. D, Zn²⁺ application (100 μm for 5 min) increased 2-AG levels in a PLC and mZnR-dependent manner. HPLC-MS measurements of 2-AG levels in DCN slices from WT and mZnR KO littermates in the presence of Zn²⁺ and/or U73122 (5 μm, PLC inhibitor applied for 30 min before Zn²⁺ application; mZnR WT Control: 1 ± 0.06 pmol/μg protein, n = 8; mZnR WT in Zn²⁺: 2.2 ± 0.1 pmol/μg protein, n = 8, p < 0.01; mZnR WT in U73122: 0.97 ± 0.51 pmol/μg protein, n = 6 when compared with mZnR WT control; p = 0.6; mZnR WT in U73122 and Zn²⁺: 0.89 ± 0.37 pmol/μg protein, n = 7; p = 0.42 when compared with mZnR WT control; mZnR KO: 1.3 ± 0.08 pmol/μg protein in control, n = 6; mZnR KO in Zn²⁺: 1.2 ± 0.02 pmol/μg protein, n = 4, p = 0.5 when compared with mZnR KO control). Summary data represent mean ± SEM. Asterisks indicate statistically significant differences, p < 0.05.

Figure 9.

Synaptically released Zn²⁺ modulates STF via endocannabinoid signaling. A1, STF is enhanced after Zn²⁺ chelation (STF control: 58.24 ± 3.4%, n = 9; tricine: 101.5 ± 11.16%, n = 9, p < 0.01). Summary graph showing the time course of the change in EPSC amplitude after 100 Hz, 1 s parallel fiber stimulation, in control and after endogenous Zn²⁺ chelation with tricine (100 Hz, 1 s stimulation was applied before and after tricine for the same and for each cell; 10 mm tricine was bath applied for at least 10 min before 100 Hz, 1 s stimulation). The high-frequency stimulus was delivered under current clamp to allow postsynaptic spiking. Top, Representative traces of EPSCs before the train (pre-burst), and after the 100 Hz train (post-burst). A2, AM-251-enhanced baseline STF and blocked the enhancement of STF after Zn²⁺ chelation (STF control in AM-251: 90.32 ± 3.40%, n = 10, p < 0.05 when compared with control from A1; tricine in AM-251: 90.55 ± 11.16%, n = 10, p = 0.28 when compared with control in AM-251). Summary graph showing the time of the change in the EPSC amplitude after 100 Hz, 1 s parallel fiber stimulation as in A1, but now in the presence of AM-251 (pre-incubation of slices with 1 μm AM-251 for at 1 h and bath application during the experiment). Top, Representative traces of EPSCs before the train (pre-burst), and after the 100 Hz train (post-burst). A3, Average values of STF from A1 and A2. Percentage STF magnitude is the average of first three points after the train. B1, STF is revealed after Zn²⁺ chelation; synaptic Zn²⁺ modulates the threshold for STF (STF control: 6.66 ± 6.10%, n = 6, p = 0.25; tricine: 33.95 ± 3.31%, n = 6, p < 0.01). Summary graph showing the time course of the change in the EPSC amplitude after 30 Hz, 1 s parallel fiber stimulation, in control and after endogenous Zn²⁺ chelation with tricine (30 Hz, 1 s stimulation was applied before and after tricine for the same and for each cell; 10 mm tricine was bath applied for at least 10 min before 30 Hz, 1 s stimulation). The high-frequency stimulus was delivered under current clamp to allow postsynaptic spiking. B2, AM-251 enhanced baseline STF and blocked the enhancement of STF after Zn²⁺ chelation. Summary graph showing the time of the change in the EPSC amplitude after 100 Hz, 1 s parallel fiber stimulation as in B1, but now in the presence of AM-251 (pre-incubation of slices with 1 μm AM-251 for at 1 h and bath application during the experiment; STF control in AM-251: 34.27 ± 5.97%, n = 6, p < 0.01 when compared with control from B1; tricine in AM-251: 38.16 ± 6.23%, n = 6, p = 0.38 when compared with control in AM-251). B3, Average values of STF from B1 and B2. Percentage STF magnitude is the average of first three points after the train. Summary data represent mean ± SEM. Asterisks indicate statistically significant differences, p < 0.05.

Figure 10.

Tricine does not affect either baseline synaptic transmission or Pr. A, Time course of the effect of tricine (10 mm) on EPSC amplitude (EPSC amplitude: 106.4 ± 6.7% of control, n = 6, p = 0.57). B, C, PPR (PPR control: 1.71 ± 0.11; tricine: 1.72 ± 0.10, n = 6, p = 0.51) (B) and 1/CV² (1/CV² in tricine: 1.28 ± 0.18% of control, n = 6, p = 0.3) (C) did not change after tricine application. Summary data represent mean ± SEM. Asterisks indicate statistically significant differences, p < 0.05.

Figure 11.

Synaptically released Zn²⁺-induced modulation of STF is absent in mZnR KO and in ZnT3 KO mice. In A1–A3 baseline STF, elicited by 100 Hz, 1 s stimulation is increased in mZnR KO mice; tricine-mediated enhancement of STF is absent in mZnR KO littermates. A1, Summary graph showing the time course of the change in EPSC amplitude after 100 Hz, 1 s parallel fiber stimulation, in control and after endogenous Zn²⁺ chelation with tricine in mZnR WT mice (100 Hz, 1 s stimulation was applied before and after tricine for the same and for each cell; 10 mm tricine was bath applied for at least 10 min before 100 Hz, 1 s stimulation; STF mZnR WT: 45.41 ± 6.65%, n = 10; mZnR KO: 87.28 ± 13.54%, n = 10, p < 0.01). A2, Summary graph showing the time course of the change in EPSC amplitude after 100 Hz, 1 s parallel fiber stimulation, in control and after endogenous Zn²⁺ chelation with tricine as in A1, but now in mZnR KO mice (STF: mZnR KO control: 87.28 ± 13.54%, n = 10; mZnR KO in tricine: 76.83 ± 17.10%, n = 10, p = 0.30). A3, Average values of STF from A1 and A2. Percentage STF magnitude is the average of the first three points after the train. In B1–B3 baseline STF elicited by 100 Hz, 1 s stimulation is increased in ZnT3 KO mice; tricine-mediated enhancement of STF is absent in ZnT3 KO littermates. B1, Summary graph showing the time course of the change in EPSC amplitude after 100 Hz, 1 s parallel fiber stimulation, in control and after endogenous Zn²⁺ chelation with tricine in ZnT3 WT mice (conditions are the same as in A1; STF ZnT3 WT: 39.28 ± 3.97%, n = 9; ZnT3 KO: 87.36 ± 10.92%, n = 9, p < 0.01). B2, Summary graph showing the time course of the change in EPSC amplitude after 100 Hz, 1 s parallel fiber stimulation, in control and after endogenous Zn²⁺ chelation with tricine in ZnT3 KO mice (conditions are the same as in A1; STF ZnT3 KO control: 87.36 ± 10.92%, n = 9; ZnT3 KO in tricine: 71.6 ± 12.29%, n = 9, p = 0.42). B3, Average values of STF from B1 and B2. Percentage STF magnitude is the average of the first three points after the train. In C1–C5 baseline STF elicited by 30 Hz, 1 s stimulation is revealed in mZnR KO and in ZnT3 KO mice. C1, C2, Summary graphs showing the time course of the change in EPSC amplitude after 30 Hz, 1 s parallel fiber stimulation, in mZnR WT and in mZnR KO littermates (STF mZnR WT: 14.35 ± 4.41%, n = 6; mZnR KO: 42.7 ± 14.10%, n = 6 p < 0.01 when compared with mZnR WT). C3, C4, Summary graphs showing the time course of the change in EPSC amplitude after 30 Hz, 1 s parallel fiber stimulation, in ZnT3 WT and in ZnT3 KO littermates (STF ZnT3 WT: 14.25 ± 6.91%, n = 6; ZnT3 KO: 49.34 ± 6.01%, n = 6, p < 0.01 when compared with ZnT3 WT). C5, Average values of STF from C1–C4. Percentage STF magnitude is the average of the first three points after the train. Summary data represent mean ± SEM. Asterisks indicate statistically significant differences, p < 0.05.

Figure 12.

Quantal properties, baseline Pr and depolarization-induced endocannabinoid signaling are not altered in ZnT3 KO mice. A, Top, Sample traces of mEPSCs from ZnT3 WT and KO mice. Cumulative probability plot of mEPSC frequencies shows that frequencies are not different between ZnT3 WT and KO mice (mean mEPSC frequency WT: 3.25 ± 0.60 Hz, n = 5; KO: 3.57 ± 0.73 Hz; n = 5, p = 0.74). B, Cumulative probability plot of mEPSC amplitudes shows that amplitudes are not different between ZnT3 WT and KO mice (mean mEPSC amplitude WT: 17.50 ± 1.2 pA, n = 5; KO: 16.7 ± 0.33 pA; n = 5, p = 0.54). C, Cumulative probability plot of mEPSC rise times showing that rise times are not different between ZnT3 WT and KO mice (mean mEPSC rise time: WT: 0.67 ± 0.06 ms, n = 5; KO: 0.59 ± 0.03 ms; n = 5, p = 0.3). D, Cumulative probability plot of mEPSC decay times showing that decay times are not different between ZnT3 WT and KO mice (mean mEPSC decay time WT: 1.44 ± 0.08 ms, n = 5; KO: 1.42 ± 0.07 ms; n = 5, p = 0.84). E, PPR is similar in ZnT3 WT and KO littermates (PPR WT: 2.13 ± 0.15, n = 7; KO: 2.16 ± 0.11, n = 7, p = 0.85). F, DSE is similar in ZnT3 WT and KO mice. Time course of DSE, induced by a 5 s depolarization to 10 mV (DSE WT: 33.5 ± 4.9% of control, n = 5; KO DSE: 35 ± 3.2% of control, p = 0.54). Summary data represent mean ± SEM.

Figure 13.

Schematic representation illustrating the mechanism of Zn²⁺-induced decrease in Pr. Synaptically released Zn²⁺, mZnR activation, rises in postsynaptic Ca²⁺ and PLC stimulation are necessary for 2-AG synthesis. 2-AG is produced by cleavage of diacylglycerol by DGL. 2-AG diffuses to the presynaptic terminal where it decreases Pr.