Loss of Doc2-Dependent Spontaneous Neurotransmission Augments Glutamatergic Synaptic Strength

Abstract

Action potential-evoked vesicle fusion comprises the majority of neurotransmission within chemical synapses, but action potential-independent spontaneous neurotransmission also contributes to the collection of signals sent to the postsynaptic cell. Previous work has implicated spontaneous neurotransmission in homeostatic synaptic scaling, but few studies have selectively manipulated spontaneous neurotransmission without substantial changes in evoked neurotransmission to study this function in detail. Here we used a quadruple knockdown strategy to reduce levels of proteins within the soluble calcium-binding double C2 domain (Doc2)-like protein family to selectively reduce spontaneous neurotransmission in cultured mouse and rat neurons. Activity-evoked responses appear normal while both excitatory and inhibitory spontaneous events exhibit reduced frequency. Excitatory miniature postsynaptic currents (mEPSCs), but not miniature inhibitory postsynaptic currents (mIPSCs), increase in amplitude after quadruple knockdown. This increase in synaptic efficacy correlates with reduced phosphorylation levels of eukaryotic elongation factor 2 and also requires the presence of elongation factor 2 kinase. Together, these data suggest that spontaneous neurotransmission independently contributes to the regulation of synaptic efficacy, and action potential-evoked and spontaneous neurotransmission can be segregated at least partially on a molecular level.SIGNIFICANCE STATEMENT Action potential-evoked and spontaneous neurotransmission have been observed in nervous system circuits as long as methods have existed to measure them. Despite being well studied, controversy still remains about whether these forms of neurotransmission are regulated independently on a molecular level or whether they are simply a continuum of neurotransmission modes. In this study, members of the Doc2 family of presynaptic proteins were eliminated, which caused a reduction in spontaneous neurotransmission, whereas action potential-evoked neurotransmission remained relatively normal. This protein loss also caused an increase in synaptic strength, suggesting that spontaneous neurotransmission is able to communicate independently with the postsynaptic neuron and trigger downstream signaling cascades that regulate the synaptic state.

Keywords: Doc2; spontaneous neurotransmitter release; synaptic scaling.

Figure 1.

Loss of Doc2-like proteins decreases the frequency but increases the amplitude of spontaneous excitatory events. A, Representative traces of AMPA-mEPSC recordings in GFP-infected (Control) or Doc2A, Doc2B, Doc2G, and Rabphilin quadruple knockdown (Doc/Rph KD) neurons in rat hippocampal cultures. B, Cumulative probability histograms of AMPA-mEPSC interevent intervals from Control and Doc/Rph KD neurons. AMPA-mEPSC interevent interval is significantly increased, suggesting a decreased frequency, in Doc/Rph knockdown neurons compared with control neurons (p = 0.0001; Kolmogorov–Smirnov test; data collected from 9–30 neurons per condition from two to six independent cultures). C, Average AMPA-mEPSC amplitudes from Control and Doc/Rph KD neurons. Doc/Rph KD neurons exhibit significantly increased AMPA-mEPSC amplitudes (p = 0.036; Student's t test; n = 9–30 neurons from two or six independent cultures). D, Rank-order plot of AMPA-mEPSC amplitudes from Control and Doc/Rph KD neurons analyzed in C. The slopes of the linear fits of the two curves indicate a multiplicative increase (289% scaling) of AMPA-mEPSC amplitudes in Doc/Rph KD neurons compared with Control. *, p < 0.05.

Figure 2.

shRNA-resistant Doc2B rescues AMPA-mEPSC parameters altered by loss of Doc2-like proteins. A, Representative traces of AMPA-mEPSC recordings in GFP-infected (Control), Doc/Rph KD, or Doc/Rph KD with coexpression of shRNA-resistant Doc2B (Doc2B rescue) neurons in rat hippocampal cultures. B, Cumulative probability histograms of AMPA-mEPSC interevent intervals from Control, Doc/Rph KD, and Doc/Rph KD + Doc2B rescue neurons. AMPA-mEPSC interevent interval is significantly increased in Doc/Rph knockdown neurons but not in Doc/Rph KD + Doc2B rescue neurons compared with Control (Control vs Doc/Rph KD: Bonferroni-corrected, p < 0.05; Control vs Doc/Rph KD + Doc2B rescue: uncorrected, p > 0.05; Kolmogorov–Smirnov test). Data were collected from 8–19 neurons per condition from two to five independent cultures. C, Average AMPA-mEPSC amplitudes from Control, Doc/Rph KD, and Doc/Rph KD + Doc2B rescue neurons analyzed in B. A significant increase in AMPA-mEPSC amplitude was observed in Doc/Rph KD neurons but not in Doc/Rph KD + Doc2B neurons compared with Control (Control vs Doc/Rph KD: Bonferroni-corrected, p < 0.001; Control vs Doc/Rph KD + Doc2B rescue: uncorrected, p = 0.21; ANOVA followed by Student's t test). *, Bonferroni corrected p < 0.05.

Figure 3.

Loss of Doc2-like proteins decreases the frequency of spontaneous inhibitory events but does not alter their amplitudes. A, Representative traces of GABA-mIPSC recordings in Control or Doc/Rph KD neurons in rat hippocampal cultures. B, Cumulative probability histograms of GABA-mIPSC interevent intervals from Control and Doc/Rph KD neurons. GABA-mIPSC interevent interval is significantly increased, suggesting a decreased frequency, in Doc/Rph knockdown neurons compared with control neurons (p = 0.0001; Kolmogorov–Smirnov test; data collected from 8 neurons per condition from three independent cultures). C, Average GABA-mIPSC amplitudes from Control and Doc/Rph KD neurons analyzed in B. Doc/Rph KD neurons do not exhibit significantly different GABA-mIPSC amplitudes (p = 0.94; Student's t test).

Figure 4.

Network-driven excitatory activity is not altered by loss of Doc2-like proteins. A, Representative traces of network-driven EPSCs from Control and Doc/Rph KD neurons in rat hippocampal cultures. B, No significant difference was observed in average total charge transfer (Qc) between Control and Doc/Rph KD neurons from 3 min of recording (p = 0.91; Student's t test; n = 12 neurons from five independent cultures). C, No significant difference was observed in average EPSC burst number between Control and Doc/Rph KD neurons analyzed in B (p = 0.94; Student's t test).

Figure 5.

Loss of Doc2-like proteins reduces levels of phosphorylated eEF2 at synapses. A, Representative images from Control and Doc/Rph KD neurons after 5 min incubation in TTX and then 15 min incubation in the presence of TTX and antibodies against the luminal domain of synaptotagmin1 to measure spontaneous synaptic vesicle trafficking (green). After fixation, neurons were processed for phospho-eEF2 immunocytochemistry (red). Areas with closely apposed red and green puncta (white arrows), which are representative of presynaptic synaptotagmin1 aligned with postsynaptic phospho-eEF2 detection, were analyzed for antibody intensity. B, A significant decrease in phospho-eEF2 levels was observed at synapses after Doc/Rph KD (p = 0.00005; Student's t test; n = 362–588 synapses from five independent cultures). C, A significant decrease in synaptotagmin1 luminal antibody levels was observed after Doc/Rph KD at the same synapses analyzed in B (p = 0.001; Student's t test), suggesting a decrease in spontaneous vesicle recycling. D, Representative immunoblots showing total eEF2, phospho-eEF2, and GDI loading control levels in neuronal protein samples collected from Control and Doc/Rph KD neurons. *p < 0.05

Figure 6.

Synaptic scaling of AMPA-mEPSC amplitudes after loss of Doc2-like proteins requires eEF2 kinase. A, Representative traces of AMPA-mEPSC recordings in control-treated or Doc/Rph KD neurons in hippocampal cultures from eEF2K KO mice or their littermate controls (eEF2K WT). B, Cumulative probability histograms of AMPA-mEPSC interevent intervals from control-treated or Doc/Rph KD neurons from eEF2K KO mice or their littermate controls (eEF2K WT). Doc/Rph KD neurons from both eEF2K KO and eEF2K WT exhibited significantly increased interevent intervals compared with control-treated neurons, but control-treated eEF2K KO neurons did not differ significantly from eEF2K WT (eEF2K WT vs eEF2K WT + Doc/Rph KD: Bonferroni-corrected, p < 0.005; eEF2K WT vs eEF2K KO + Doc/Rph KD: Bonferroni-corrected, p < 0.05; eEF2K WT vs eEF2K KO: uncorrected, p > 0.05; Kolmogorov–Smirnov test; n = 17 neurons from five independent cultures). B, Symbol legend refers to both B and C. C, Cumulative probability histograms of AMPA-mEPSC amplitudes from data analyzed in B. Doc/Rph KD neurons from eEF2K WT mice exhibited significantly increased AMPA-mEPSC amplitudes compared with control-treated neurons, but control-treated and Doc/Rph KD neurons from eEF2K KO mice did not differ significantly from eEF2K WT (eEF2K WT vs eEF2K WT + Doc/Rph KD: Bonferroni-corrected, p < 0.001; eEF2K WT vs eEF2K KO + Doc/Rph KD: uncorrected, p > 0.05; eEF2K WT vs eEF2K KO: uncorrected, p > 0.05; Kolmogorov–Smirnov test). Inset, Rank-order plot of mEPSC amplitudes from control treated (black squares) and Doc/Rph KD (green squares) eEF2K WT neurons. The slopes of the linear fits of the two curves indicate a multiplicative increase (51% scaling) of AMPA-mEPSC amplitudes in Doc/Rph KD neurons compared with control. To improve the linear fits of the slope, additional analyses of the initial portion (mEPSCs 0–200) and later portion (mEPSCs 200–350) of the plot were conducted and showed a 44% and 53% increase in mEPSC amplitudes, respectively, following Doc/Rph KD.