The role of the C2A domain of synaptotagmin 1 in asynchronous neurotransmitter release.

Following nerve stimulation, there are two distinct phases of Ca2+-dependent neurotransmitter release: a fast, synchronous release phase, and a prolonged, asynchronous release phase. Each of these phases is tightly regulated and mediated by distinct mechanisms. Synaptotagmin 1 is the major Ca2+ sensor that triggers fast, synchronous neurotransmitter release upon Ca2+ binding by its C2A and C2B domains. It has also been implicated in the inhibition of asynchronous neurotransmitter release, as blocking Ca2+ binding by the C2A domain of synaptotagmin 1 results in increased asynchronous release. However, the mutation used to block Ca2+ binding in the previous experiments (aspartate to asparagine mutations, sytD-N) had the unintended side effect of mimicking Ca2+ binding, raising the possibility that the increase in asynchronous release was directly caused by ostensibly constitutive Ca2+ binding. Thus, rather than modulating an asynchronous sensor, sytD-N may be mimicking one. To directly test the C2A inhibition hypothesis, we utilized an alternate C2A mutation that we designed to block Ca2+ binding without mimicking it (an aspartate to glutamate mutation, sytD-E). Analysis of both the original sytD-N mutation and our alternate sytD-E mutation at the Drosophila neuromuscular junction showed differential effects on asynchronous release, as well as on synchronous release and the frequency of spontaneous release. Importantly, we found that asynchronous release is not increased in the sytD-E mutant. Thus, our work provides new mechanistic insight into synaptotagmin 1 function during Ca2+-evoked synaptic transmission and demonstrates that Ca2+ binding by the C2A domain of synaptotagmin 1 does not inhibit asynchronous neurotransmitter release in vivo.


Introduction 4 8
Fast, synchronous release is the large burst of neurotransmitter release that peaks within 4 9 milliseconds (ms) of the arrival of the action potential. At most healthy synapses, the majority of 5 0 release occurs during the synchronous phase [1]. Synaptotagmin 1, which contains two Ca 2+ -5 1 binding domains, C 2 A and C 2 B [2], is essential for coupling Ca 2+ binding to efficient, 5 2 synchronous release [3][4][5][6].

3
Asynchronous release can last from 10's of ms to 10's of seconds [7] Indeed, at some specialized synapses, such as certain hippocampal interneurons, asynchronous 5 7 release is predominant [14].

8
In addition to being the Ca 2+ sensor for synchronous release, synaptotagmin 1 is 5 9 proposed to regulate asynchronous release. Increases in asynchronous release are reported in 6 0 1 1 noise. The average resting membrane potential was calculated from the mean potential during the 2 2 0 280 ms prior to stimulation and was used to determine the beginning of the response post 2 2 1 stimulation. Total charge transfer represents the total evoked response and was measured from 2 2 2 the beginning of the response to 300 ms post stimulation. To assess asynchronous release, we 2 2 3 divided the total response into synchronous, beginning of response to 20 ms, and asynchronous, [17]. Statistical significance between genotypes was determined using a one-way ANOVA with Dunnett's correction. Statistical analyses were performed using Prism 7 (GraphPad software, La Jolla, CA).

3 0
Each data set was collected from at least 9 samples for all experiments. Unpaired student t-tests with t values were used to compare datasets with Gaussian distributions between two genotypes. Paired t-tests were employed when comparing values within a given electrophysiological trace. One-way ANOVAs with F values and Dunnett's correction were used to compare datasets with 2 3 4 Gaussian distributions across all 3 genotypes. Kruskal-Wallis tests with Dunn's correction were 2 3 5 used to compare datasets with non-Gaussian distributions across all 3 genotypes. An alpha p-2 3 6 value of 0.05 was considered significant. Raw data will be posted at 2 3 7 https://lib.colostate.edu/find/csu-digital-repository/. To test the function of Ca 2+ binding by the C 2 A domain of synaptotagmin during vesicle 2 4 1 fusion events, we completed the first direct comparison of two disparate mutations that both 2 4 2 block Ca 2+ binding ( Fig 1A). Mutating the third and fourth of the Ca 2+ -binding aspartates to 2 4 3 asparagines (syt D-N , Fig 1A) blocks Ca 2+ binding by removing key negative charges required to 2 4 4 coordinate Ca 2+ [20, 30, 31]. Mutating the second aspartate to a glutamate (syt D-E , Fig 1A)   negatively-charged presynaptic membrane. A. Top, The C 2 A Ca 2+ -binding pocket of wild type 2 5 1 synaptotagmin (syt WT ). Prior to Ca 2+ entry (left), 5 negatively charged (magenta) aspartate 2 5 2 residues repel the negatively-charged presynaptic membrane (large arrows). Ca 2+ binding 2 5 3 neutralizes the negative charge of the pocket (right), resulting in the penetration of the 2 5 4 presynaptic membrane by hydrophobic residues (grey). A. Below, By replacing two aspartate 2 5 5 residues with neutral (green) asparagines, the syt D-N mutation in C 2 A blocks Ca 2+ binding and 2 5 6 partially neutralizes the negative charge of the pocket. Importantly, this partial neutralization also 2 5 7 decreases the electrostatic repulsion of the presynaptic membrane (small arrows), which may 2 5 8 mimic Ca 2+ binding. By replacing one aspartate residue deep in the Ca 2+ -binding pocket with a 2 5 9 larger, negatively-charged glutamate residue (magenta, larger), the syt D-E mutation in C 2 A 2 6 0 blocks Ca 2+ binding by steric hindrance while maintaining electrostatic repulsion of the  [4, 17], there were no significant differences in expression of the synaptotagmin transgenes. synaptotagmin is appropriately concentrated at the neuromuscular junction in all three genotypes 2 8 0 (Fig 1C and [4,17]). binding by the C 2 A domain is critical for efficient synchronous neurotransmitter release and 2 9 7 support the hypothesis that the syt D-N mutation participates in triggering membrane fusion by 2 9 8 mimicking Ca 2+ binding ( Fig 1A).  We also compared the effect of these mutations on spontaneous neurotransmitter release, We found no significant differences in mEJP amplitudes in any genotype The effect of these mutations on the frequency of mEJP events is consistent with Ca 2+ -binding pocket is the key characteristic of synaptotagmin required to clamp spontaneous 3 2 7 fusion events. The differential effect of these mutations on both spontaneous and synchronous between its C 2 domains and the presynaptic membrane is required to prevent vesicle fusion (see 3 3 0 Fig 1A). solution to the neuromuscular junction (Fig 3A-C). To control for the increase in spontaneous Fig 2D), event frequencies were normalized to the mean 3 3 9 mEJP frequency prior to sucrose application ( Fig 3C) Dunn's correction). Therefore, the decrease in synchronous release in the P[syt D-E ] mutant is not to control (*p < 0.0001). There was no significant change in paired pulse ratios between control 3 5 7 and P[syt D-N ]. Error bars are SEM, and n's within bars represent number of fibers tested. ns = 3 5 8 not significant.  Opposite effects of Ca 2+ -binding mutants on asynchronous release 3 7 5 To quantitatively assess the timing of neurotransmitter release, we counted individual release events that occurred between 280 ms prior to stimulation and 580 ms after stimulation 3 7 7 (Fig 4, Table 1, and see [16,17]). Latency histograms of the mean number of events/stimulation before and after the stimulus are shown (Fig 4 A-C, right). A single, multi-quantal, synchronous 3 7 9 response occurred during the first 20 ms following stimulation (Fig 4 A-C). The time course of Asynch, Table 1 Mean Async), as event frequency was not elevated in any genotype during the 3 8 2 300 -580 ms period post stimulation (Fig 4A,  increase in asynchronous release ( Fig 4D, Table 1, p = 0.001, t(59) = 3.46, paired t-test). events between 280 ms before to 300 ms after stimulation (large dotted arrow). Stimulation histograms. Data were parsed into 20 ms bins from 280 ms before to 580 ms after single  Values are shown for the 280 ms pre stimulation (Mean Prestim), the 20 -300 ms asynchronous Consistent with the increased mEJP frequency reported in voltage traces (Fig 2E),  increase in the number of asynchronous release events compared to control (*, Fig 4E,  comparison). Similar to the event frequency analysis (Fig 4E), analysis of the charge transfer analysis (Fig 4E), the asynchronous phase of release was increased in the P[syt D-N ] mutant (41.7 4 5 5 ± 6.9 pC, mean ± SEM) relative to control (7.8 ± 10.4 pC, p = 0.01, Dunnett's correction).

5 6
Importantly, there was no difference in the asynchronous phase of release in the P[syt D-E ]  the electrostatic repulsion of the presynaptic membrane was maintained. . Therefore, the release probability is significantly reduced (Fig 3D,E), fewer vesicles are triggered to fuse, and the EJP amplitude is dramatically reduced (Fig 2A, only 60% (Fig 2E). The differential impact of the C 2 A and C 2 B syt D-N mutations in larvae 5 1 5 suggests that the ability of the C 2 B domain to clamp spontaneous release is dominant by this 5 1 6 stage in development. Importantly, the rate of spontaneous release is not increased by the syt D-E 5 1 7 mutation in C 2 A (Fig 2E and [4]). Thus, the decreased electrostatic repulsion between the C 2 A increase in spontaneous fusion events, while the maintenance of electrostatic repulsion in the 5 2 0 syt D-E mutation prevents them.

2 1
Synaptotagmin null mutants show an increase in asynchronous release, indicating that the 5 2 2 presence of synaptotagmin 1 inhibits aberrant asynchronous fusion events [15,16]. Since this 5 2 3 ability was maintained in a mutant that lacked the C 2 B domain, the C 2 A domain was postulated 5 2 4 to provide the inhibition of asynchronous fusion [16]. In addition, various syt D-N mutations that 5 2 5 blocked Ca 2+ binding by the C 2 A domain in either synaptotagmin 1 or Doc2 (another C 2 domain 5 2 6 based Ca 2+ sensor) resulted in increased asynchronous release [17,49]. Together, these results

2 7
suggested that Ca 2+ binding by the C 2 A domain of synaptotagmin (or Doc2) directly inhibited 5 2 8 asynchronous fusion events.

2 9
However, the differential effects on asynchronous release between the two distinct Ca 2+ 5 3 0 binding mutants presented in this study are not consistent with this inhibition hypothesis. Rather, our findings are consistent with synaptotagmin's role as an electrostatic switch. In the syt D-N 5 3 2 mutation, the removal of electrostatic repulsion by C 2 A could constitutively mimic bound Ca 2+ 5 3 3 [18]. This would be expected to trigger fusion events for a longer period of time following Ca 2+ 5 3 4 influx, resulting in increased asynchronous release. Since blocking Ca 2+ binding with our syt D-E 5 3 5 mutation (which maintains electrostatic repulsion) did not result in a similar increase, our data 5 3 6 demonstrate that the increase in asynchronous release caused by the syt D-N mutation is an 5 3 7 artifact.

3 8
Our comparison of the syt D-N to syt D-E mutations refutes the asynchronous inhibition 5 3 9 hypothesis. However, all of the disparate changes in synchronous, spontaneous, and 5 4 0 asynchronous release seen in synaptotagmin null and C 2 A domain mutants could be explained by asynchronous Ca 2+ sensor has unimpeded access to SNARE complexes to trigger increased 5 4 6 asynchronous release. The competition hypothesis is also consistent with results from the 5 4 7 different Ca 2+ binding mutations. In the C 2 A syt D-N mutation, the presence of synaptotagmin 1 5 4 8