Neuromodulator Signaling Bidirectionally Controls Vesicle Numbers in Human Synapses

SUMMARY

Neuromodulators bind to pre- and postsynaptic GPCRs, are able to quickly change intracellular cAMP and Ca²⁺ levels, and are thought to play important roles in neuropsychiatric and neurodegenerative diseases. Here, we discovered in human neurons an unanticipated presynaptic mechanism that acutely changes synaptic ultrastructure and regulates synaptic communication. Activation of neuromodulator receptors bidirectionally controlled synaptic vesicle number within nerve terminals. This control correlated with changes in the levels of cAMP-dependent protein kinase A-mediated phosphorylation of synapsin-1. Using a conditional deletion approach, we reveal that the neuromodulator-induced control of synaptic vesicle numbers was largely dependent on synapsin-1. We propose a mechanism whereby non-phosphorylated synapsin-1 ‘latches’ synaptic vesicles to presynaptic clusters at the active zone. cAMP-dependent phosphorylation of synapsin-1 then removes vesicles. cAMP-independent dephosphorylation of synapsin-1 in turn recruits vesicles. Synapsin-1 thereby bidirectionally regulates synaptic vesicle numbers and modifies presynaptic neurotransmitter release as an effector of neuromodulator signaling in human neurons.

In Brief:

Detailed characterization of human and mouse synapses reveals a mechanism whereby neuromodulators impact synaptic transmission by altering synaptic vesicle numbers in presynaptic nerve terminals.

Graphical Abstract

INTRODUCTION

Unlike the fast neurotransmitters glutamate and γ-aminobutyric acid (GABA), neuromodulators do not primarily interact with ligand-gated ion channels but with G protein-coupled receptors (GPCRs). Via these receptors, neuromodulators control neural circuits, regulate high-level cognitive functions, and contribute to multiple brain disorders (Arnsten, 2006; Schultz, 2015; Thiele and Bellgrove, 2018). GPCRs constitute the largest family of membrane proteins in animals. Hundreds of non-sensory GPCRs have been identified, of which more than 90% are expressed in brain (Pierce et al., 2002; Vassilatis et al., 2003). GPCRs encompass drug targets in numerous therapeutic areas, including various neurological diseases (Huang et al., 2017). Neuromodulator activation of GPCRs induces second messenger responses that can acutely lead to biochemical and physiological modulation of a target cell and may also change its transcriptional and metabolic activity (Civelli, 2012).

Serotonergic input into neural networks, for instance, activates at least 14 structurally and pharmacologically different receptors, and triggers responses like a hyperpolarization-induced decrease in firing rate, a stimulation of phospholipase C (PLC) via G_αq proteins to modulate intracellular Ca²⁺, and an activation or inactivation (G_αs/i) of adenylate cyclase (AC) to regulate cAMP levels (with the exception of 5-HT₃ receptors, which gate a cation-permeable ion channel) (Bockaert et al., 2006; Lesch and Waider, 2012; Millan et al., 2008). Dysfunction of serotonin signaling is not only associated with psychiatric disorders, but serotonergic activation of 5-HT_2A/2C and 5-HT₇ receptors also regulates neuronal excitability in the auditory system (Tang and Trussell, 2015). Noradrenergic synaptic transmission, conversely, is important for arousal, attention and cognition, and activates GPCRs that modulate PLC or AC as well (Schwarz and Luo, 2015).

Despite the prominent role of neuromodulators in physiology and pathology, a systematic understanding of their functions in neurons is lacking. Their acute functions at human synapses in particular are not well understood. Here, we used human induced neurons as a system to screen for acute actions of neuromodulators on synapses and synaptic vesicles (SVs). Unexpectedly, we found opposing, highly significant effects on modulation of SV number that were mediated by cAMP signaling and culminated in the differential protein kinase A (PKA)-mediated phosphorylation of the presynaptic phosphoprotein synapsin-1 (Syn1). Moreover, we applied a conditional knockout (cKO) approach in human neurons and discovered that changes in SV number were largely dependent on Syn1. This previously undescribed presynaptic modulatory mechanism with Syn1 as the central bidirectional effector triggers an acute increase of SV number after activation of the adrenergic receptor α2A, and an acute decrease in SV number after activation of the receptor 5-HT₇ by serotonin or by rapid elevations of cAMP levels via AC activation.

RESULTS

Neuromodulators acutely change the number of SV-associated synaptic puncta in human nerve terminals.

Because of their impact on the human synapse, we sought to investigate the acute effects of neuromodulators on human neurons in an unbiased screen using light-microscopy-based analyses of synaptic puncta. We generated induced excitatory neurons by forced expression of the transcription factor neurogenin 2 (Ngn2) in human pluripotent stem cells (Figures S1A–D) (Zhang et al., 2013). We incubated numan neurons for 30 mins with norepinephrine, adenosine, dopamine, histamine, or serotonin (each 100 µM, to maximally activate receptors), and quantified presynaptic puncta densities on dendrites after immunolabeling the neurons for the SV proteins synapsin (pan-synapsin antibody), synaptophysin (Syph), and synaptotagmin-1 (Syt1), and for the active zone protein piccolo. Strikingly, we detected a robust increase in SV-associated puncta densities, but not in piccolo puncta densities, in norepinephrine- or adenosine-treated cells (Figures 1A–E). Serotonin treatment had the opposite effect, with a highly significant reduction of Syph and Syt1, but not of synapsin, puncta densities and again with no effect for piccolo puncta densities (Figures 1A–E, S1D–H).

Figure 1. Neuromodulation acutely alters synaptic vesicle numbers in human synapses.

A Human neurons (1 month old) before and after acute treatment with norepinephrine stained for presynaptic markers Syph, synapsin (‘Pan-Syn’) and MAP2.

B-E Synaptic puncta density of pan-synapsin (B), Syph (C), Syt1 (D), and piccolo (E) of acutely treated human neurons. Human neurons were acutely (30 mins) treated with the following reagents and concentrations before fixation: PKA inhibitor H89 (5 µM), norepinephrine (‘Norepi’, 100 µM), guanfacine (‘Guanf’, 5 µM), adenosine (100 µM), dopamine (100 µM), histamine (100 µM), serotonin (‘Sero’, 100 µM), 5-HT₇ receptor antagonist DR4485 (10 µM), and serotonin (100 µM) in combination with DR4485 (10 µM). Control samples were treated with the matching volume of solvent. Synaptic puncta areas are shown in Figure S1 E–H.

F Phospho-Syn1 protein levels of acutely treated human neurons with different concentrations (5-100 µM) of neuromodulators and forskolin as a control. Norepinephrine induces decrease of phosphorylation, and serotonin induces increase of phosphorylation of Syn1.

G Phospho-Syn1 protein levels of acutely treated human neurons with different concentrations of the α2A adrenergic receptor agonists guanfacine, tizanidine, and the PKA inhibitor H89.

H-M Representative EM micrographs of chemically fixed human neurons after acute H89 or guanfacine treatments (5 µM, 30 mins). Summary graphs show values for number (I), distribution (J), clustering (K), and diameter (L) of SVs. M shows PSD length.

Data are means ± SEM; statistical significance for B-G, I, K and M was assessed by one-way ANOVA and for J by two-way ANOVA (*, P < 0.05; **, P < 0.01; ***, P < 0.001).

After screening the expression levels of neuromodulator receptors in these neurons by RNA sequencing, we sought to pharmacologically identify the receptors responsible for the observed ‘puncta effects’. Notably, the effect of norepinephrine was recapitulated with guanfacine (‘Guanf’), an agonist for the α2A adrenergic receptor (a G_αi-coupled GPCR that localizes to axons (Figure S1K)). The serotonin effect, conversely, was blocked by the 5-HT7 receptor (G_αs-coupled) antagonist DR4485 (Figures 1A–E). Since both GPCRs mediate changes of cAMP levels (α2A decreases while 5-HT7 increases cAMP levels) and PKA is a major downstream effector of cAMP (Walsh et al., 1968), we next tested the PKA inhibitor H89. Recapitulating the norepinephrine and guanfacine results, acute application of H89 also led to an increase in the density of SV-associated puncta (Figures 1A–E, S1D–H).

Because all the treatments affected SV-associated markers but not the active zone marker piccolo, these results suggest that our assay reflects dynamics in SVs rather than changes in the number of synapses. Together, these data indicate that activation of the neuromodulator receptors α2A and 5-HT7 differentially change the dynamics of SVs in human neurons by regulating PKA activity.

Syn1 phosphorylation is downstream of neuromodulator receptor-mediated cAMP signaling.

How could a neuromodulator-mediated change in cAMP levels and PKA activity impact SV dynamics? Since synapsins are abundant SV proteins that are downstream targets of PKA (Dolphin and Greengard, 1981; De Camilli et al., 1983; Huttner et al., 1983), we next analyzed phosphorylation of synapsins by immunoblotting. We used a phospho-synapsin antibody that specifically recognizes the phosphorylated PKA-site (serine-9) of synapsins (Hosaka et al., 1999) to screen human neurons treated with neuromodulators (Figure 1F). RNA sequencing and immunoblotting revealed that induced human excitatory neurons mainly express Syn1 but not Syn2 or Syn3 (Figure 2F). Matching our results from the SV-associated puncta analysis, we detected an increase of phospho-Syn1 levels after acute serotonin treatment and a decrease after acute norepinephrine or adenosine treatment, but no change after acute incubation with histamine, acetylcholine or dopamine (Figure 1F). The AC activator forskolin (to raise intracellular cAMP levels), as expected, increased levels of phospho-Syn1 as a positive control (Figures 1F, 5G). Activation of the adrenergic receptor α2A by the agonists guanfacine and tizanidine or inhibition of PKA by H89 similarly decreased phosphorylation of Syn1 (Figure 1G), reflecting the effect of norepinephrine. However, overall protein levels of Syn1, Syph, Syt1, and PSD95 did not change with the various treatments, suggesting that the treatments selectively changed Syn1 phosphorylation without alterating global levels of synaptic proteins (Figure S1I). Additionally, we found that co-treatment of human neurons with DR4485 and serotonin blocked phosphorylation of Syn1 (Figure S1J). Together with previous reports, these results thus suggest that the serotonin-mediated increase in Syn1 phosphorylation leads to dissociation of Syn1 from SVs, explaining why Syn puncta densities do not decrease after serotonin treatment since synapsin remains in the terminal after it dissociates from the vesicles (Benfenati et al., 1989; Hosaka et al., 1999). In summary, our data show that the phosphorylation of Syn1 inversely correlates with changes in SV-associated puncta densities, indicating that Syn1 phosphorylation at serine-9 acts downstream of neuromodulators in the cAMP signaling pathway.

Figure 2. Neuromodulator effects on synaptic vesicle numbers are mediated by Syn1.

A-D Puncta densities of acutely treated mouse hippocampal cultures (DIV14) for 30 mins with the following reagents: H89 (5 µM), norepinephrine (100 µM), and serotonin (100 µM). Synaptic puncta areas are shown in Figure S2 A–D.

E Summary of experimental design and phospho-Syn1 protein levels of acutely treated mouse hippocampi (60 mins) after dissection of 2-months-old C57BL/6 mice. For each condition 3 hippocampi were tested.

F In contrast to mouse neurons, human neurons express very little Syn2. Immunoblot showing expression of Syn1 and Syn2 from mouse brain (6 months old) and human neurons (1 month after induction). Total protein stain (Ponceau S) was used as loading control.

G and H Three (Figure S2C) independent Syn1 conditional KO (cKO) stem cell clones were generated and two of them (termed clone #1 and clone #2) were used for production of human neurons. Immunoblot (plus ponceau staining for Syn1 cKO clones #1 and #2) and confocal micrograph of cKO Syn1 (clone #1) human neurons at 1 month after neuronal induction. Cre mediated deletion of Syn1 abolishes expression of both Syn1 splice variants: Syn1A and Syn1B. Inactive Cre (ΔCre) was used as control and represents the Syn1 WT. Cre and ΔCre were delivered as lentiviruses at day −1 (1 day before neuronal induction; Figure S1A).

I-M Puncta densities of acutely guanfacine (‘Guanf’) and PKA inhibitor H89 treated (5 µM, 30 mins) and control neurons stained for Syph (E), Syt1 (F), Pan-Synapsin (‘Pan-Syn’, G), Bassoon (‘BSN’,H), and MAP2 for clone #2. Treatments increase puncta densities of SV-associated proteins in WT but not in matching isogenic Syn1-deficient human neurons.

N Expression of Syn1 10 days after transduction of 4 different Syn1B rescue constructs (Figures S2N and S2O) in 2-months-old cultures using 3 different antibodies: mouse monoclonal specific for Syn1 (top), phospho-Syn1 (middle) and pan-synapsin (bottom) binding to the N-terminus of Syn1.

O-Q Representative confocal micrographs (O) and summary graphs (P, Q) showing presynaptic puncta densities of Syn1 cKO neurons with and without expression of Syn1B rescue constructs for phospho-mimetics.

Data are means ± SEM; statistical significance for B-E, J-M and P, Q was assessed by one-way ANOVA (*, P < 0.05; **, P < 0.01; ***, P < 0.001). Broken lines in H and M indicate separation between two different blots.

Figure 5. Deletion of Syn1 from human synapses lowers SV numbers but does not impact synapse numbers.

A, B and C Puncta densities of acutely forskolin treated (10 µM, 30 mins) and control and Syn1-deficient human neurons stained for pan-Synapsin (A), Syt1 (A), Syph (B), Homer (C) and MAP2. Puncta areas are depicted in Figure S4.

D Representative EM micrographs (fast-freeze) and summary graphs showing synapse densities in WT and isogenic Syn1-deficient human neurons. No significant change in synapse densities were detected after deletion of Syn1. SVs: Synaptic vesicles, PSD: Postsynaptic density, M: Mitochondrion, CV: Clathrin coated vesicle, LVs: Large vesicles.

E and F Protein levels in induced neurons produced from independent SYN1 cKO clones #1 and #2. Values in F represent mean protein levels observed in SYN1 KO neurons normalized to those of matching isogenic WT controls (dashed line). TuJ1 is used as loading control.

G Protein levels in WT human neurons with and without acute forskolin treatment. No significant changes in Syt1, Syph, and Syn1 levels were detected. However, levels of N-terminally (serine 9) phosphorylated Syn1 (P-Syn1) increased significantly. Values were normalized to controls (without forskolin) and corrected for blotting and loading variations using TuJ1 as an internal standard.

Data are means ± SEM; statistical significance was assessed using Student’s t test (*, P < 0.05; **, P < 0.01; ***, P < 0.001). Numbers of images/independent experiments analyzed are shown in the bars or indicated as n.

α2A adrenergic receptor activation or PKA inhibition acutely increase the number of SVs in human nerve terminals.

To determine whether our light-microscopy data reflect changes in the organization or number of SVs, we performed EM analyses of human neurons acutely treated with H89 or guanfacine. Both treatments rapidly increased the number of SVs in the nerve terminal close to the active zone (~200 nm distance) compared to control neurons (~45 versus ~60 SVs per synapse) (Figures 1H–J). Moreover, both treatments decreased the inter-vesicle distance, reflecting increased clustering of SVs (Figure 1K). However, the diameter of SVs or the length of the postsynaptic density (PSD) were unchanged (Figures 1L and 1M). Thus, inhibition of Syn1-phosphorylation by α2A adrenergic receptor activation or by PKA inhibition causes rapid accumulation and denser packing of SVs in human nerve terminals, suggesting among others that increases in SV-associated puncta densities observed by light microscopy can reflect increases in SV number.

Neuromodulators acutely change the number of SV-associated synaptic puncta in mouse hippocampal nerve terminals.

Next, we assessed how generalizable our findings are. We acutely treated cultured mouse hippocampal neurons (DIV14) with norepinephrine, serotonin, or H89, and stained them for presynaptic markers. Essentially, we obtained the same results as in human neurons: H89 and norepinephrine increased, and serotonin decreased (again, except for synapsin itself) the puncta densities of SV-associated proteins but not of the active zone marker protein bassoon (Figures 2A–D, S2A–D). We also performed acute treatments ex vivo on hippocampi from 2-months-old mice. Again similar to cultured human neurons, norepinephrine decreased, and serotonin increased, Syn1 phosphorylation (Figure 2E), suggesting that neuromodulators act in mouse hippocampal neurons similar to human neurons.

Syn1 deletion blocks neuromodulator-mediated control of SV dynamics.

So far, our data show that rapid changes in SV numbers by neuromodulators or by PKA inhibition correlate with alterations in Syn1 phosphorylation. To determine if these findings are mechanistically linked, we generated a conditional KO (cKO) mutation of the X-chromosomal Syn1 (SYN1) gene in male H1 embryonic stem (ES) cells, and analyzed two independent cKO clones (referred to as clone #1 and #2; Figures S2F–H). Neurons derived from ES cells carrying the SYN1 cKO allele fully expressed Syn1 when transduced with lentiviruses producing mutant inactive Cre-recombinase (ΔCre), but lacked Syn1 when transduced with lentiviruses expressing active Cre recombinase (Cre) (Figure 2H). Human neurons deficient in Syn1 appeared to develop normally, and exhibited no immediately obvious morphological changes (Figure 2G, S5N). Because Syn1 is the predominant synapsin variant in human induced neurons (and to some degree in post-mortem human brain samples) (Figures 2F, S2E), the cKO of Syn1 effectively generates human neurons deficient for all synapsins (Figure 2H).

We next analyzed by immunocytochemistry the effects of the SYN1 deletion on acute treatments with guanfacine or H89 (Figures 2I–M). Guanfacine and H89 increased the SV-associated puncta density, but not the bassoon-associated puncta density, in ΔCre-transduced wild-type (WT) as expected. However, guanfacine and H89 had no effect in Syn1-deficient neurons (Figures 2I–M, S2I–M).

The lack of an effect of guanfacine or H89 on the SV-associated puncta density in synapsin-deficient neurons suggests that PKA-dependent phosphorylation of the Syn1 A-domain mediates the neuromodulator effect. To further test this hypothesis, we performed rescue experiments with WT and mutant Syn1B, using A-domain mutations that mimic (S9D) or prevent phosphorylation of serine-9 (S9A and S9K; Figures 2N, S2N, S2O). Expression of S9D-mutant Syn1B decreased the Syt1-associated puncta density, whereas expression of Synt1B S9A or S9K increased the Syt1-associated puncta density both in synapsin-deficient and in WT human neurons (Figure 2O, 2P, S2P, S2R). None of the Synt1B mutants altered the bassoon-associated puncta density (Figure 2Q). WT Syn1B had also no effect (Figure 2N–2Q). These findings support the notion that the PKA-site at serine-9 in the A-domain of Syn1 regulates SV dynamics in presynaptic terminals.

Rapid high-pressure freeze EM uncovers spatial properties of the Syn1-dependent control of SV dynamics.

Next, we used rapid high-pressure freeze EM, yielding high-definition images, to characterize the fine structure of human WT and Syn1-deficient synapses in combination with optogenetic stimulation (Figure 3A) (Watanabe et al., 2013; Watanabe et al., 2014). We generated WT and synapsin-deficient human neurons that co-expressed a fast variant of channelrhodopsin-2 (‘ChetaTC’; (Berndt et al., 2011)) from the two SYN1 cKO lines, and performed optogenetic stimulation combined with fast-freeze (‘flash-and-freeze’) EM. We stimulated the neurons for 3 sec with a 10 Hz stimulus train immediately before rapid freezing for EM (Figure 3B). This brief stimulus train was chosen to excite the neurons without causing vesicle depletion. To visualize the stimulation efficiency by EM, we stimulated the neurons in the presence of extracellular ferritin that cannot penetrate the plasma-membrane (Figure 3C). We observed a large increase in SVs containing ferritin (around 3% of all presynaptic vesicles; Figure 3C), confirming that the optogenetic stimulation efficiently induced SV endocytosis.

Figure 3. Synapsin deletion and acute forskolin treatments similarly decrease the number of SVs in human presynaptic terminals.

A Flow diagram of fast-freeze EM experiments. Cells were frozen down after 6-8 weeks of incubation.

B Sample traces of Syn1 cKO human neurons showing 30 APs triggered by optogenetic stimulation (10Hz for 3 s).

C Summary graph of average number of ferritin-containing presynaptic SVs per synapse. SVs were classified according to their shape: circular or non-circular.

D and E WT and Syn1 deficient human neurons. Arrows in middle panels point to mitochondria found on both sides of the synapse.

F Number of SVs per synaptic profile of Syn1 cKO clones #1 and #2. The conditional deletion of Syn1 (Cre +) reduces the number of SVs. Acute (5 mins) forskolin treatment (forskolin +) reduces the number of SVs mainly in the WT (Cre −).

G-I PSD-lengths, SV diameter and synaptic cleft width for Syn1 cKO clones #1 and #2.

Data are means ± SEM per synaptic profile; numbers of images/independent experiments analyzed are shown in the bars. Statistical significance for F, G and I was assessed using one-way ANOVA (*, P < 0.05; ***, P < 0.001) or Student’s t-test for C (***, P < 0.001).

Synapses from WT control (ΔCre-expressing) and Syn1-deficient (Cre-expressing) neurons exhibited a similar overall morphology, with abundant presynaptic vesicles and less abundant, larger endomembrane structures, including endosome-like vesicles on both the pre- and postsynaptic sides (Figures 3D, 3E, S3A–S3D, S4A–D). The numerical parameters of human synapses were similar to those of rodent neurons (Rosahl et al., 1995), with an SV diameter of ~34 nm, a synaptic cleft width of ~27 nm, and a PSD length of ~300 nm (Figures 3G–3I). The Syn1 deletion caused no significant change in these parameters. However, the Syn1 deletion induced a large decrease in SV numbers (~40%; Figures 3D–F and 4A). Optogenetic stimulation had no significant effect on any structural parameter of presynaptic terminals, including SV number, in either control or Syn1-deficient neurons (Figures 3D–3I, 4C, 4D). In contrast to presynaptic vesicles, the Syn1 deletion did not decrease postsynaptic vesicle numbers (Figures S4A–S4D), or cause significant changes in the number of presynaptic large vesicles (Figures S3A–D), of pre- or postsynaptic dense-core vesicles (Figures S3E–F, S4E–F), or of pre- and postsynaptic coated vesicles (Figures S3G–H, S4G–H). Thus, the loss of synapsins has similar effects on the ultrastructure of synaptic terminals in human neurons as in mouse neurons (Rosahl et al., 1995).

Figure 4. Reduced number of synaptic vesicles in the reserve pool and decreased SV clustering in Syn1-deficient and forskolin treated human neurons.

A-D Syn1 cKO with (B) and without (A) acute forskolin (forskolin) treatment (10 µM, 5 mins) C and D Comparison of WT and SYN1-deleted human neurons after optogenetic stimulation. Right panels represent zoom-ins.

E and F Distribution plots of presynaptic SVs per synapse that are not docked or tethered to the active zone.

G Summary graph of SV clustering determined by the average distance of each SV to its nearest neighbor.

H and I Summary graphs of number of presynaptic docked and tethered SVs.

Data are means ± SEM per synaptic profile. Numbers of images/independent experiments analyzed are shown in the bars or graphs. Statistical significance for E and F was assessed using two-way ANOVA repeated measurements (* P < 0.05; *** P < 0.001) and for G-I using one-way ANOVA comparing data to the control ‘-Cre -forskolin 0 APs’ (*, P < 0.05; **, P < 0.01; ***, P < 0.001).

Next, we asked whether treatment of neurons with forskolin alters the structure of WT or synapsin-deficient human synapses. Unexpectedly, a 5 min application of forskolin caused a large loss of SVs from control synapses similar to that induced by the Syn1 deletion (Figures 3F, 4A, 4B). In contrast, forskolin had little effect on SV numbers in Syn1-deficient synapses with already decreased SV numbers. Again, optogenetic stimulation of neurotransmitter release had no significant effect on forskolin-treated synapses in either control or Syn1-deficient neurons (Figures 3F, 4A, 4B).

Plots of the distance of SVs from the active zone revealed that after the Syn1 deletion or after forskolin treatment, SVs distant from the active zone – the presumptive ‘reserve pool’ (Sudhof, 2004) – were selectively lost (Figures 4E, 4F). However, SVs continued to be tightly clustered at the active zone, and there was no change in docked and tethered SVs (Figures 4H, 4I). We also detected a significant increase in the average distance between SVs in the cluster in both the Syn1-deficient and forskolin-treated synapses, suggesting that SVs are clustered less tightly (Figure 4G). For both the change in SV distribution and in the distance between SVs in the cluster, the Syn1 deletion occluded the effect of forskolin, indicating that they are in a common pathway. Viewed together, these data suggest that the Syn1 deletion and forskolin-induced increases in cAMP both cause a selective loss of SVs in the peripheral reserve pool of the SV cluster and a decrease in the inter-SV distance in the cluster.

Syn1 deletion or forskolin treatment decrease SV-associated puncta densities but not the number of synapses.

Based on the data so far, we propose that cAMP-dependent phosphorylation of Syn1 physiologically regulates SVs by stimulating Syn1 dissociation from SVs, thereby freeing vesicles from physical synapsin-based constraints and allowing vesicles to diffuse away. This hypothesis is consistent with the differential distribution of synapsins in the vesicle cluster (Kempf et al., 2013) and with the mobility of SVs in neurons (Darcy et al., 2006).

To test this hypothesis, we analyzed the localization of SV-associated proteins (Syt1 and Syph) in human neurons with and without the Syn1 deletion or forskolin treatment (30 mins) by immunocytochemistry (Figures 5A–C, S5). We found that stimulation of cAMP synthesis with forskolin had no effect on the density and size of synapsin-, homer- or bassoon-containing puncta, but caused a modest decrease in the apparent density of Syt1- and Syph-containing synaptic puncta (Figures 5A–4C and S5). Matching the results of the serotonin experiment, forskolin treatment did not decrease synapsin-positive puncta densities because synapsin dissociates from the SVs, and presumably does not acutely leave the nerve terminals (Figures S5J–M). Examining the effect of the Syn1 deletion, we observed a qualitatively similar result. The Syn1 deletion had no effect on the density of PSD95-, homer- and bassoon-positive synaptic puncta, but decreased the density of Syt1- and Syph-positive synaptic puncta, consistent with the similar consequence of forskolin and Syn1 deletion on nerve terminals as revealed by EM (Figures 5A–C, S5A–I, S5M). In support of this conclusion, synapse density analyses using EM also demonstrated that the Syn1 deletion did not cause a significant change in synapse density (Figure 5D), again demonstrating that immunocytochemistry signals for Syt1- and Syph-containing synaptic puncta reflect a decrease in SV numbers rather than synapse numbers.

Syn1 deletion but not forskolin treatment decreases SV protein levels.

The question now arises whether the decrease in SV numbers per presynaptic cluster in Syn1-deficient neurons corresponds to a true loss of SVs or could be due to their diffusion out of the cluster. To address this question, we quantified the levels of synaptic proteins as a proxy for overall SV number. We detected a modest but significant decrease (~20%) in the levels of three SV proteins tested (Syph, SV2A, and Syt1) and of syntaxin-1 (Figures 5E and 5F). The levels of other synaptic protein, including the SV protein synaptobrevin-2, were unchanged. Acute treatment of neurons with forskolin, conversely, did not alter SV protein levels, although immunoprobing with the synapsin antibody specific for the PKA phosphorylation site in the A-domain (serine-9) of synapsins confirmed significant stimulation of Syn1-phosphorylation (Figure 5G) (Hosaka et al., 1999). We thus conclude that on the one hand, activation of the norepinephrine-α2A-cAMP-PKA-Syn1 pathway (or likely any inhibition of AC-mediated cAMP production) acutely increases the number of SVs, and that on the other hand, activation of the serotonin-5-HT7-cAMP-PKA-Syn1 pathway (or any increase in cAMP) rapidly reduces SV number in terminals.

Syn1 deletion and cAMP-mediated phosphorylation of Syn1 impair short-term synaptic plasticity.

Our hypothesis that the Syn1 deletion and forskolin-induced cAMP increases similarly unlatch vesicles from a Syn1-based constraint suggests that the Syn1 deletion and forskolin-induced cAMP increases should have similar physiological consequences. To test this prediction, we analyzed synaptic transmission in human neurons as a function of the Syn1 deletion and/or forskolin treatment.

The Syn1 deletion and forskolin, either alone or in combination, caused no change in basal synaptic transmission, measured as the amplitude of excitatory postsynaptic currents (EPSCs) elicited by isolated action potentials (Figures 6A–C, S6J). Also, the intrinsic electrical properties of neurons were unchanged by the Syn1 deletion or by forskolin treatment (Figures S6A–S6C). On the other hand, when we measured miniature mEPSCs, we observed a decrease in the mEPSC frequency but not amplitude in synapsin-deficient neurons (Figures S6P–T), indicating an impaired availability of SVs for fusion. However, the Syn1 deletion and forskolin separately increased the paired-pulse depression of EPSCs induced by tandem stimuli (Figures 6D–F, S6K). The Syn1 deletion occluded the effect of forskolin, and forskolin treatment of Syn1-deficient neurons produced no effect. Increases in paired-pulse depression (or decreases in paired-pulse facilitation) are generally considered proxies for an increase in release probability (Thomson, 2000). Thus, similar to the EM results, deleting Syn1 and increasing cAMP levels by forskolin treatment (which induces Syn1 dissociation from SVs) produce the same increase in release probability and occlude each other’s effects.

Figure 6. Syn1 deletion and brief forskolin applications similarly alter neurotransmitter release and occlude each other mechanistically.

A-C Representative traces (A) and amplitude summary graphs (B and C) of EPSCs evoked by isolated extracellular stimulus in control and Syn1-deficient neurons derived from ES cell clones #1 and #2.

D Representative traces of evoked EPSCs amplitudes after paired-pulse stimulations (interval 50 ms) with and without acute forskolin (forskolin) treatment.

E and F Paired-pulse ratios (PPRs) of both clones with (middle) and without (left) acute forskolin treatment and ratio of PPRs (right) with and without forskolin.

G Representative traces of evoked EPSC amplitudes during stimulus train (10Hz for 10 s).

H and J Summary graphs of normalized amplitudes (to first response) during stimulus train.

I and K Summary graphs of cumulative amplitudes during stimulus train. Slope was estimated for stimulus numbers 51-100 indicated by solid line. RRP (readily releasable pool) is indicated by broken line crossing y-axis.

L and M Measurements of release induced by hypertonic sucrose (0.5M) for different durations (0.5-20s). M Total charge transfer and peak amplitude of fitted traces (indicated in L).

Data are means ± SEM; statistical significance for B and C was assessed using one-way ANOVA and for E, F and H-K by two-way ANOVA (*, P < 0.05; **, P < 0.01; ***, P < 0.001). Numbers of images/independent experiments analyzed are shown in the graphs.

Next, we expanded our analysis of short-term synaptic plasticity by examining the effect of extended high-frequency stimulus trains (10 Hz for 10 sec). As expected from the observed paired-pulse depression, the high-frequency stimulus train caused massive synaptic depression in both control and Syn1-deficient neurons (Figures 6G, 6H, 6J, S6D–G, S6L–O). Syn1-deficient neurons, however, exhibited two significant differences from control neurons: depression set in faster, and the steady-state level of the EPSCs at the end of the stimulus train was lower (Figures 6H, 6J). Interestingly, forskolin treatment again replicated the Syn1 deletion phenotype in control neurons, and the Syn1 deletion also occluded the effect of forskolin on high-frequency stimulus trains. Thus, acutely increasing the cAMP levels in neurons results in the same functional change as deletion of Syn1.

The decrease in steady-state responses during high-frequency stimulus trains could either be due to a decrease in the capacity of the readily-releasable pool (RRP) of vesicles, or to a decrease in the replenishment rate of SVs from the reserve pool into the RRP. To differentiate between these two possibilities, we analyzed the cumulative EPSC amplitudes during the stimulus trains and the response of synapses to hypertonic sucrose, which measures the size of the RRP (Figures 6I, 6K, S6H, S6I) (Rosenmund and Stevens, 1996). The Syn1 deletion or forskolin treatment both similarly decreased the replenishment rate of vesicles during the stimulus train, suggesting that increasing cAMP directly decreased RRP replenishment during stimulus trains (Figures 6I, 6K) but did not significantly alter sucrose responses (Figures S6H, S6I). Finally, to address vesicle fusogenicity we applied a 0.5 M sucrose solution for different time periods (0.5-20 sec), but could not detect any significant changes in the sucrose-sensitivity in synapsin-deficient neurons (Figure 6L, 6M). In summary, these results argue that rapid cAMP induction or Syn1 deletion impairs the replenishment rate of vesicles but not the initial RRP size under baseline conditions. Phosphorylation or deletion of Syn1 decrease SV replenishment and steady-state responses during stimulus trains.

Synapsin function is maintained in long-term cultures of human neurons.

Human neurons derived from ES or induced pluripotent stem cells are generally immature, raising the possibility that some of the changes observed may reflect a transient developmental impairment. Human induced neurons, however, can be cultured more than 2 years, allowing long-term studies (Figures 7A–7B, S7D–S7M). ‘Old’ human neurons display extensive axonal and dendritic arborizations and synaptic networks, and occasionally exhibit spine-like structures on dendrites (Figures S7D–S7L). Analysis of Syn1 and Syn2 expression in old neurons, monitored via immunocytochemistry and immunoblotting, revealed that Syn1 expression by human neurons increased dramatically (>50-fold) during 18 months of culture (Figures 7A–7C). Syn2 expression also increased (~22-fold). Nevertheless, old Syn1-deficient human neurons continue to show an impairement in spontaneous synaptic transmission (Figures S7J–L), and a large decrease in SV-associated, but not homer-associated, puncta density (Figure S7N). These results replicate those of ‘young’ Syn1-deficient neurons (Figures 5A–5C), suggesting that the Syn1 deletion phenotype persists during long-term in vitro development of human neurons.

Figure 7. Correcting Syn1-dependent abnormalities in human neuron by interfering with neuromodulator receptors - potential therapeutic effects.

A Syn2-expression of WT neurons after 2 and 18 months imaged with identical settings.

B Summary graphs depicting Syn1 (left) and Syn2 (right) intensities normalized to MAP2 intensities over different time points of culture period.

C Representative immunoblot of 1- and 4-months-old Syn1 cKO human neurons, showing Syn1 and Syn2 protein levels. Equal loading is shown by total protein stain Ponceau S (Ponc).

D-H Representative confocal micrographs (D) and summary graphs showing puncta densities of acutely guanfacine (Guanf) (5 µM, 30 mins) treated and control 4 months old human neurons stained for L1CAM, Syn2 (E), pan-synapsin (F), Syph (G), bassoon (H), and MAP2.

I-M Representative immunoblots and summary graphs of 4-months-old mutant and WT neurons showing protein levels for Syt1 (J), Syph (K), Syn2 (L), and PSD95 (M) after 1 week of treatment with indicated reagents. Treatment with guanfacine partly rescues SV-associated protein levels in Syn1 deficient human neurons.

Data are means ± SEM; statistical significance for C was assessed by Student’s t test (* p<0.05, ** p<0.01, *** p<0.001) and for F-I and K-N by one-way ANOVA (*, P < 0.05; **, P < 0.01; ***, P < 0.001). Numbers of images/independent experiments analyzed are shown in the bars or indicated as n.

Guanfacine reverses impairments caused by SYN1 mutations.

Multiple point mutations in the human SYN1 gene were identified in patients with intellectual disability, epilepsy, and/or autism (Fassio et al., 2011; Giannandrea et al., 2013; Guarnieri et al., 2017; Lignani et al., 2013) (Figure S2H), raising the question whether pharmacological engagement of Syn2 by guanfacine might have a potential therapeutic effect on more mature SYN1-deficient neurons. To address this question, we acutely treated (30 mins) WT and Syn1-deficient human neurons with guanfacine and compared these by immunocytochemistry to neurons acutely treated with serotonin, H89, norepinephrine or control. Similar to our results in 1-month-old neurons (Figures 2I–2M, S2I–S2M), guanfacine, H89, and norepinephrine treatment of the WT 4-months-old neurons resulted in an increase in puncta densities of synapsin and Syph, and also (as anticipated) for Syn2 (Figure 7D, 7E). There was no change for Syn2 puncta after serotonin treatment, mirroring our previous results (Figures 1B, S1F and S5K), and no significant changes for bassoon, again pointing to an increase in SV number rather than in synapse number (Figures 7D–7H). Interestingly, however, in 4-months-old neurons as opposed to 1-month-old neurons we detected a less pronounced but (in relation to the control) a similar puncta density increase in the Syn1-deficient neurons (Figures 7D–7H, S7A), suggesting that Syn2, which has the same PKA-phosphorylation site in its A-domain as Syn1, might be subjected to similar phosphorylation dynamics as Syn1 and bypass the loss of Syn1. We further tested this hypothesis by immunoblotting (again using the antibody against the phosphorylated A-domain), and found that Syn2 indeed becomes phosphorylated after acute forskolin treatment and dephosphorylated after acute H89 or guanfacine treatments (Figure S7B). Together, our data suggest that activation of the α2A-adrenergic receptor in human neurons 4 months of age or older can, at least partly, regulate SV number in Syn1-deficient neurons, and that this regulation may occur via Syn2. Similar results were also obtained with chronic treatments of 4-months-old neurons with guanfacine (5 µM for 1 week with medium exchange every other day; Figures 7I–7M, S7C). Importantly, we now detected a significant increase of Syt1, Syph, and Syn2 levels after guanfacine treatment in Syn1-deficient but not in WT neurons, suggesting that the reduction in SV number caused by the Syn1 deletion can be partially rescued by continued activation of the α2A-adrenergic receptor (Figures 7I–7M). Together, these results indicate that guanfacine can increase the total number of SVs in Syn1-deficient neurons with initially decreased SV number. Moreover, guanfacine, which is already in use for treatment of attention deficit-hyperactivity disorder (ADHD), likely represents a potential therapeutic drug for patients with pathogenic SYN1 mutations (Posey and McDougle, 2007).

DISCUSSION

Syn1 the first SV protein identified, represents the most abundant phosphoprotein of synapses and one of the most abundant SV components that in rat brain accounts for 6% of the total vesicle protein (De Camilli et al., 1983; Dolphin and Greengard, 1981; Huttner et al., 1983; Takamori et al., 2006). SYN1 is the founding member of a family of three genes that are primarily expressed in neurons (Hosaka and Sudhof, 1998; Kao et al., 1998; Sudhof et al., 1989). All synapsins share a short N-terminal membrane-attachment sequence (the A-domain), a short linker sequence (the B-domain), and a large, highly conserved central sequence (the C-domain, Figure S2H) (Sudhof et al., 1989). These highly homologous N-terminal and central domains are followed in synapsins by variable C-terminal domains. As peripheral membrane proteins, synapsins are attached to SVs via their A-domains. The A-domains are phosphorylated by cAMP-dependent PKA and CaM Kinase I (Huttner et al., 1981; Kennedy and Greengard, 1981); this phosphorylation causes dissociation of synapsins from SVs (Hosaka et al., 1999).

In excitatory synapses of mice, deletion of Syn1 induced a selective impairment in short-term plasticity; this impairment was aggravated by double deletion of both Syn1 and Syn2, and was associated with epilepsy. EM revealed that the number of SVs was decreased approximately 50% in Syn1/2 double-deficient nerve terminals, but the number of synapses was not changed (Li et al., 1995; Rosahl et al., 1993; Rosahl et al., 1995). The initial amount of release is maintained in synapsin-deficient excitatory neurons, but the reserve pool of SVs is decreased, and synaptic depression ensues during longer stimulus trains (Fornasiero et al., 2012; Orenbuch et al., 2012; Rosahl et al., 1993; Rosahl et al., 1995; Vasileva et al., 2012).

As a result of our study, we hypothesize that short-term plasticity and the sustained release capability of synapses are physiologically regulated by neuromodulators that alter presynaptic levels of cAMP, which in turn controls the organization of SVs by inducing or preventing the PKA-dependent dissociation of synapsins from SVs (see graphical abstract). This hypothesis constitutes an unanticipated regulatory mechanism of short-term presynaptic plasticity that contributes to the modification of synaptic transmission by neuromodulators. The proposed mechanism is consistent with previous studies on the function of Syn1 (Gitler et al., 2004; Rosahl et al., 1995; Ryan et al., 1996), but now shows that Syn1 also controls the number of SVs in nerve terminals in a bidirectional manner downstream of neuromodulator-mediated changes in cAMP levels.

The decrease in SV number observed here agrees well with previous mouse studies (Rosahl et al., 1995; Gitler et al., 2004; Orenbuch et al., 2012; Vasileva et al., 2012; Orlando et al., 2014), but the rapid decrease in SV number induced by cAMP stimulation was unexpected. Even more surprising was the finding that inhibition of Syn1 phosphorylation rapidly increased SV numbers in nerve terminals (Figures 1, 2). Clearly, reserve pool SVs are highly dynamic, consistent with studies on the exchange of vesicles between presynaptic terminals (Darcy et al., 2006). The cAMP- and synapsin-dependent mobility of reserve vesicles provides an unappreciated local mechanism of regulation for vesicle pools.

Our results thus propose a molecular mechanism by which presynaptic modulators – and possibly presynaptic activity under certain circumstances – regulate neurotransmitter release by controlling Syn1 phosphorylation. Moreover, our results provide a potential mechanistic explanation for the deleterious effect of SYN1 mutations in human neurons that may impair brain function by blocking the phosphorylation-dependent regulation of SV clusters. Our results also raise the possibility that guanfacine may be a therapeutic agent in patients with SYN1 mutations by enhancing the action of Syn2 after the loss of Syn1.

Although we provide multiple lines of evidence for our hypotheses, our data have several limitations and raise multiple questions. For example, GPCR cAMP signaling and thereby synapsin function appear to be quite different between excitatory and inhibitory mouse neurons (Chiappalone et al., 2009; Gitler et al., 2004), and the question arises whether the Syn1 deletion phenotype might differ between excitatory and inhibitory human neurons as well. Furthermore, we here analyzed only Syn1 because the type of human neurons we examined primarily express this isoform, but it is likely that Syn2 and Syn3 perform similar functions. Moreover, results from Betz and colleagues showing that acutely blocking phosphatases by okadaic acid disperses SV clusters in frog motor nerve terminals very much resembles our findings with acute forskolin treatment (Betz and Henkel, 1994). However, it is puzzling that SV mobility from similar preparations using synapsin‐null mutant mice was unchanged (Gaffield and Betz, 2007). Another important question regards the nature of the action of Syn1 on the vesicle cluster: Does Syn1 act as an ATPase with a substrate that latches SVs together, or as an ATPase in which ATP is a structural component of a cytomatrix-like support (Esser et al., 1998)? Since actin is absent from the vesicle cluster (Fernandez-Busnadiego et al., 2010), actin binding by Syn1 cannot be responsible in the cluster itself, but it could shape the periphery of the cluster. A third important question concerns the nature of the electrophysiological phenotype. Whereas the decrease in steady-state responses during high-frequency stimulus trains observed in Syn1-deficient or forskolin-treated synapses can be plausibly accounted for by the loss of peripheral SVs in the cluster, no ready explanation exists for the striking change in paired-pulse ratios induced under the same conditions. If the increased paired-pulse depression is due to an increased release probability, why is there no change in the single pulse triggered EPSC amplitude during isolated synaptic responses? Syn1 likely does more than organize the reserve pool of vesicles, and may in fact perform different additional functions that produce a general shift in the SV supply chain. Addressing these questions will be the next major challenge in solving the synapsin enigma.

STAR★METHODS TEXT

LEAD CONTACT AND MATERIALS AVAILABILITY

Further information and requests for resources and reagents should be directed to and will be fulfilled by the Lead Contact, Christopher Patzke (patzke@stanford.edu).

All renewable reagents generated in this study are available from the Lead Contact without restriction.

EXPERIMENTAL MODEL AND SUBJECT DETAILS

Human Embryonic Stem Cells

Male human embryonic stem cells (ESCs) line H1 (WA01 (WiCell Research Resources WiCell, WI) were purchased from WiCell (normal karyotype, passage 40) and used for all stem-cell-derived human neurons experiments.

The present study was approved by Stem Cell Research Oversight (SCRO) at Stanford University Research Compliance Office, Stanford University (SCRO 518: Studying brain diseases affecting synaptic transmission by using human induced neurons).

Mice

Experiments involving animals were approved by the Stanford IACUC, Administrative Panel on Laboratory Animal Care (APLAC) Research Compliance Office, Stanford University and all procedures conformed to NIH Guidelines for the Care and Use of Laboratory Animals. All mice were housed in the Stanford animal facility under supervision of the Stanford animal care unit; all mice were healthy and not kept in a sterile facility. Mice were not involved in previous procedures and were drug or test naïve. Mouse lines used: 2-months-old male and female C57BL/6 wild type for ex vivo neuromodulator treatment of hippocampi, and CD1 wild type (postnatal day 1, male and female) for primary hippocampal or astrocyte culture.

HEK293T cells

HEK293T cells purchased from ATCC, expanded and stocked in aliquots in liquid N₂. They were used for production of all lentiviral constructs, as well as Adeno-Associated-Viral constructs.

Samples from human subjects

No primary human samples have been used in this study.

METHOD DETAILS

Gene targeting in human embryonic stem cells

Experiments were performed as described (Patzke et al., 2016). H1 (male) human ES cells (WiCell Research Resources WiCell, WI) were maintained as feeder-free cells in mTeSR1 medium (Stem Cell Technologies). H1 cells were transduced by recombinant adeno-associated viruses (rAAVs) carrying the targeting construct (Figure S2F) to generate conditional hemizygous mutant ES cells, by flanking synapsin-1 (X chromosome) exon 2 (ENSE00000867032, encoding for aa in the C-domain) with loxP sites, which results in a frameshift and a premature stop-codon upon cre-recombination. Additionally the targeting construct includes a puromycin-resistance gene flanked by frt sites and homology arms (around 1.3kb length) for homologous recombination. Puromycin (1 µg/ml) was added and kept in mTeSR1 medium starting 2 days after transduction. The surviving ES cells were allowed to grow into colonies, and individually picked using dispase to lift up colonies. Subsequently accutase was used to create single cells for better cell survival during clonal expansion. 5 correctly targeted colonies out of 50 colonies in total were confirmed by PCR screening using oligo sequences GGGAAAAAGATCCATGGAGAAATTGACATTAAA and GCCTTGCTCATAACACCAAGGTACCCTTC, resulting in a 333bp band for a correctly targeted clone or 299bp for the untargeted wild type. Correctly mutated alleles were confirmed by immunoblot after flp-and cre-recombination and iN differentiation. Flp-recombination restored the allele and resulted in protein expression of synapsin-1, and cre-recombination resulted in a complete absence of synapsin-1 protein.

Mouse hippocampal culture and tissue dissection

For cell cultures, mouse hippocampi of newborn P0 CD1 WT mice were dissected and plated onto matrigel-coated coverslips. After 14 days of incubation cells were incubated with neuromodulators (norepinephrine or serotonin) or H89 containing growth media for 30 mins. For the ex-vivo-experiment intact hippocampi from 2-months-old C57BL/6 WT mice were dissected and incubated for 1h in neuromodulator (norepinephrine, histamine or serotonin) growth media. Subsequently, tissue was homogenized (using pipette and very thin insulin syringe) and boiled in SDS-PAGE loading buffer and assayed by immunoblots. Loading controls were whole protein stain Ponceau S solution and antibody TuJ1.

Mouse glial cells were cultured from the forebrain of newborn wild type CD1 mice. Briefly, newborn mouse forebrain homogenates were digested with papain and EDTA for 15 min, cells were dissociated by harsh trituration to avoid growing of neurons, and plated onto T75 flasks in DMEM supplemented with 10% FBS. Upon reaching confluence, glial cells were trypsinized and replated at lower density a total of two times to remove potential trace amounts of mouse neurons before the glia cell cultures were used for co-culture experiment with iN cells.

Viral constructs

The following lentiviral constructs were used: 1. FUW-TetO-Ng2-T2A-puromycin expressing TetO-Ng2-T2A-puromycin cassette (TetO promoter drives expression of full-length mouse Ngn2 and of puromycin via the cleavage-peptide sequence T2A; Figure 3B). 2. F-Ubiquitin-W (FUW)-rtTA containing rtTA. 3. FUW-TetO-EGFP expressing EGFP. 4. FUW-Flp-IRES-Blast to express Flp-recombinase in combination with the blasticidin resistance gene. 5. FUW-GFP::Cre to express Cre-recombinase in order to delete SYN1 exon 2 and create via frameshift a null-allele or GFP::ΔCre for the wild type control. 6. Channelrhodopsin-2 (E123T/T159C; ChetaTC) under control of a neuron-specific synapsin promoter. Infection rates were monitored by YFP. 7. FSW-EGFP to express EGFP under the control of the human Syn1 promotor. One Adeno-associated-virus (AAV-DJ, (Grimm et al., 2008)) construct was used for gene targeting of X-chromosomal Syn1 allele of the male hESCs H1 (also shown in Figure S2F): the construct contains sequences from the region encoding exon 2 flanked by loxP sites and a puromycin resistance cassette flanked by frt sites (Frt-PGK-Puro-SV40polyA-Frt) adjacent to the 5’ loxP site. For homologous recombination the 5’ arm of the construct includes 1.6 kb of sequences located upstream of exon 2. The 3’ arm contains 1.3 kb of sequences located downstream of exon 2.

Virus generation

Lentiviruses were produced as described (Zhang et al., 2013) in HEK293T cells (ATCC, VA) by co-transfection with three helper plasmids (pRSV-REV, pMDLg/pRRE and vesicular stomatitis virus G protein (vsv-g) expression vector) with 12 µg of lentiviral vector DNA and 6 µg of each of the helper plasmid DNA per 75 cm2 culture area) using calcium phosphate. Lentiviruses were harvested in the medium 48 hr after transfection, pelleted by centrifugation (49,000×g for 90 min), resuspended in MEM, aliquoted, and frozen at −80ºC. Adeno-associated-virus DJ (AAV-DJ) was used to deliver the targeting construct for generation of cKO cells. AAV-DJ was produced in HEK293T cells by co-transfection of pHelper, pDJ, and AAV vector (8.5 µg of DNA per 75 cm2 culture area) using calcium phosphate. Cells were harvested 72 hr after transfection in PBS/1mM EDTA and following one freezing thawing cycle. AAVs were collected from cytoplasm using Benzonase nuclease at a final concentration of 50 units/ml at 37°C for 30min Cell debris was cleared by slow centrifugation (3,000×g for 30 min) then AAVs were isolated after fast centrifugation (400,000×g for 120 min) in iodixanol (gradient from 15-60%) from the 40% layer and further concentrated using centricon concentrating tube (100,000 MWCO, Millipore UFC0910024) according to the manufacturer’s suggested protocol.

Generation of human induced neurons (iN Cells)

iN cell generation has been described previously (Zhang et al., 2013). Briefly, targeted and flp-recombined (after blasticidin selection) human ES cells were treated with Accutase (Innovative Cell Technologies) and plated as dissociated cells in 24-well plates (1 × 10⁴ cells/well) on day −1 (Figure S1A). Cells were plated on matrigel (BD Biosciences)-coated coverslips in mTeSR1 containing 2 mM thiazovivin (Bio Vision). At the same time point, lentiviruses prepared as described above (0.3 µl/well of 24-well plate) were added. Two main types of lentiviruses were used for co-infection: the lentiviruses used for iN cell induction as described, and lentiviruses expressing either Cre-recombinase (to create a null allele) under control of the ubiquitin promoter or an inactive mutated ΔCre-recombinase for the wild type control. Additionally, lentiviruses expressing ChetaTC (for the flash-and-freeze experiments) and EGFP (for visualization in Fig. 1C) were used. On day 0, the culture medium was replaced with N2-supplement in DMEM/F12 including non-essential amino acids (NEAA, Invitrogen) containing human BDNF (10 ng/ml, PeproTech), human NT-3 (10 ng/ml, PeproTech), and mouse Laminin-1 (0.2 µg/ml, Invitrogen). Doxycycline (2 µg/ml, Sigma) was added on day 0 to induce TetO gene expression and retained in the medium until the end of the experiment. On day 1, a 24 hr puromycin selection (1 µg/ml) period was started. On day 2, mouse glia cells were added in Neurobasal-A medium supplemented with B27/Glutamax (Invitrogen) containing BDNF, NT3 and Laminin-1; Ara-C (2 µM, Sigma) was added to the medium to inhibit astrocyte proliferation. After day 2, 50% of the medium in each well was exchanged every 2 days. Fetal bovine serum (5%) was added to the culture medium from day 10 onward to support astrocyte viability, and iN cells were assayed after at least 35 days or as indicated.

Rescue-experiments using PKA-site mutants

Using PCR-based site-directed mutagenesis, we employed mutations that mimic the phosphorylated state of serine-9 (S9D) or prevent phosphorylation (S9A and S9K) of synapsin-1B in the A-domain. Wild type human synapsin-1B cDNA (from Harvard PlasmID) served as template for the PCR reactions. The following primer pairs were used to create mutant cDNA: for S9D, cggcgccgcctgGATgacagcaactttatg and cataaagttgctgtcATCcaggcggcgccg: for S9A, cgccgcctgGCGgacagcaactttatg and cataaagttgctgtcCGCcaggcggcg; for S9K, cggcgccgcctgAAGgacagcaactttatg and cataaagttgctgtcCTTcaggcggcgccg. Wild type and mutant cDNA was PCR-amplified using the primers ggTCTAGAgcagccatgaactacctgcggcgc and ccGGATCCctaaggctgggcctgggc and inserted into a lentiviral expression plasmid (FSW) containing the human Syn1-promotor and used for production of recombinant lentiviruses. To best correlate to data obtained from acute neuromodulator treatments, we super-infected 2-month-old cultures with lentiviruses to express wild type and mutant human synapsin-1B in wild type and KO human neurons. After 10 days of incubation we analyzed the cells by immunoblots and immunocytochemistry. For immunoblots we used 3 different antibodies to detect Syn1: Mouse monoclonal anti Synapsin-1 (clone 10.22) that binds to the D-domain present only in Syn1; Rabbit polyclonal anti Synapsin (Pan-Synapsin, E028) that binds to the A-domain present in all synapsins and Rabbit polyclonal anti Phospho-Synapsin (U342) that binds only to the wild type A-domain containing a phosphorylated serine-9 (Figure 2N). For immunocytochemistry we used the mouse monoclonal anti Synapsin-1 (clone 10.22) antibody to detect the subcellular localization of wild type and mutant Syn1. Further immunocytochemistry analyses of the neurons were performed using antibodies for Pan-Synapsin (E028), Synaptophysin (c7.2), Synaptotagmin-1 (mAb 48 (asv 48)) and for Bassoon (Sigma).

Immunofluorescence and immunoblotting

Immunofluorescence staining was performed essentially as described (Zhang et al., 2013). Briefly, cultured iN cells were fixed in 4% paraformaldehyde and 4% sucrose in PBS for 20 min at room temperature, washed three times with PBS, and incubated in 0.2% Triton X-100 in PBS for 10 min at room temperature. Cells were blocked in PBS containing 5% goat serum for 1 hr at room temperature. Primary antibodies were applied overnight at 4°C, cells were washed in PBS three times, and fluorescent-labeled secondary antibodies (Alexa 405, Alexa 488, Alexa 546, and Alexa 647, 1:1000) were applied for 2 hr at room temperature. The following antibodies were used in immunocytochemisty experiments: Human Nuclei (clone 235-1, Millipore, 1:1000) MAP2 (CPCA-MAP2, EnCor. 1:2000), Bassoon (Sigma, 1:500), Pan-Synapsin (E028; 1:1,000. Antigene: NYLRRRLSDSNFMANLPNGYMTDLQRPQP), Phospho-Synapsin (U342, 1:1,000. Antigene: CYLRRRLSDSNF, bold = phosphorylated), PSD95 (7E5-1B8; Pierce/Thermo; 1:500), Syn1 (10.22; Synaptic Systems; 1:1000), Synapsin2 (19.31, Synaptic Systems; 1:1000), Synaptophysin (c7.2; Synaptic Systems; 1:1000), Synaptotagmin-1 (41.1; Synaptic Systems; 1:1000; mAb 48 (asv 48 from DSHB)), and GFP (Invitrogen; A11122; 1:2000). Images were taken using a Nikon A1RSi confocal microscope system, with a 10x, 20x or 60x objective at RT. All quantitative immunoblotting experiments were performed with fluorescently labeled secondary antibodies (LiCor, 1:5000). Samples were separated by SDS-PAGE under reducing conditions and transferred onto nitrocellulose membranes, before quick stain with Ponceau solution (0.1% Ponceau S in 5% acetic acid) and destain with PBS. Blots were blocked in Tris-buffered saline containing 0.1% Tween 20 (Sigma) and 5% fat-free milk for 2 hr at room temperature. The blocked membrane was incubated in blocking buffer containing the primary antibody overnight at 4°C, followed by three to five washes. The washed membrane was incubated in washing buffer containing secondary antibody for 2 hr at room temperature. Blots were scanned with the Odyssey-system (LiCor), followed by quantification with ImageStudio software (LiCor). For immunodetection, the following antibodies were used (1:500): NeuN (ABN78, Millipore), TuJ1 (MMS-435P, Covance), Complexin 1/2 (L668), SNAP25 (P913), Synaptobrevin-2 (P939), Synapsin (E028), Phospho-Synapsin (U342, 1:1,000), Syn1 (10.22; Synaptic Systems; 1:1000), Syntaxin-1 (438B), Synaptophysin (Synaptic Systems, 7.2), calmodulin-associated serine/threonine kinase (BD, clone 7/CASK), L1CAM (UJ127.11, Sigma), SynCAM (T2412), PSD95 (AbCAM, ab76115), Munc18 (610336; BD Transduction Lab), N-Cadherin (13A9; Millipore), Rab3A (T957), Dynamin1 (E027), SV2A (P915), Synapsin2 (19.41, Synaptic Systems; 1:1000).

Quantification of synaptic puncta density

For synaptic puncta analyses, images were acquired using a Nikon A1RSi confocal microscope system (60x objective at RT) and the puncta density was determined using the software Nikon NIS-Elements.

Electrophysiology

All electrophysiological recordings were performed using whole-cell patch clamp. Briefly, five weeks after iN induction, on the day of recording, a coverslip containing relatively low-density induced human neurons was placed in a recording chamber mounted onto an Axioskop 2F upright microscope (Zeiss) equipped with DIC and fluorescence capabilities. Cells were approached under DIC with ~2 MΩ pipettes pulled from borosilicate glass (Warner instruments, Inc) using a vertical Narishige PC-10 puller (Japan), and impaled until forming high resistance seal (GΩ) between the recording pipette and the cell membrane. Then, whole-cell configuration was established by gentle application of negative pressure trough the recording pipette. The recording pipette contained (in mM): 125 K-gluconate, 20 KCl, 10 HEPES, 0.5 EGTA, 4 ATP-Magnesium, 0.3 GTP-Sodium, 10 Na-Phosphocreatine, osmolarity: 312 mOsm; pH 7.2 adjusted with KOH. Electrical signals were recorded at 25 kHz with a two-channel Axoclamp 700B amplifier (Axon Instruments), digitalized with a Digidata 1440 digitizer (Molecular devices) that was in turn controlled by Clampex 10.1 (Molecular Devices). All recordings were performed at ~24°C.

For whole-cell voltage-clamp recordings, neurons were maintained at −80 mV holding potentials. Series resistance varied between 4-10 MΩ. All induced-neurons in which series resistance was higher than that were not included in the analysis. In all current-clamp experiments, the membrane potential was maintained at ~ −80 mV by constant injection of negative current through the recording pipette. Samples in the recording chamber were continuously perfused with oxygenated (95% O2/5%CO2) bath solution containing (in mM): 125 NaCl, 2.5 KCl, 1 MgCl2, 2 CaCl2, 25 glucose, 1.25 NaH2PO4, 0.4 ascorbic acid, 3 myo-inositol, 2 Na-pyruvate, and 25 NaHCO3 pH 7.4., and 315 mOsm. Detection and analysis of Na/K currents, and action potentials properties was performed offline with Clampfit10.1 software (Molecular Devices).

To record spontaneous (sEPSCs) or miniature excitatory postsynaptic currents (mEPSCs), cell were held at −80 mV holding potentials. sEPSCs and mEPSCs were recorded in the absence and presence of TTX (0.5 uM) in the bath solution, respectively.

To measure evoked synaptic transmission (EPSCs) and paired-pulse plasticity, we used extracellular stimulation. Briefly, a bipolar electrode (WPI) was placed 200 um away from recorded cells and brief (100 µs) square pulses of voltage were applied via a stimulus isolation unit (Model 2100 Isolated Pulse Stimulator (A-M Systems)). Evoked currents were recorded at −80 mV holding potentials in voltage-clamp and repeated 10 times at 0.1 Hz, and Coefficient of Variation was calculated from standard deviation of EPSC amplitudes, divided by the mean EPSC amplitude of each cell.

To measure paired-pulse ratios, a second identical stimulus was applied at different intervals (see results), and the ratio between the peak amplitude of the second and the first response was computed.

Similarly, trains of presynaptic action potential were triggered via a bipolar electrode while holding postsynaptic cells at −80 mV in voltage clamp. For each cell, the responses of at least 3-5 trains were recorded (at intervals of 2 minutes), and averaged offline.

To trigger evoked postsynaptic action potentials (18-months-old cells), cells were maintained at near −70 mV resting potential in current clamp, and step current injections of increasing intensities (overall from 20 to 480 pA, with increments ranging from 20 to 100 pA) were applied trough the recording pipette.

Measurements of RRP by hypertonic sucrose solution

For measurement of the readily-releasable pool (RRP) of fusion-competent SVs the patch pipette solution contained (in mM) 136 KCl, 17.8 HEPES, 1 EGTA, 0.6 MgCl2, 4 ATP-Mg, 0.3 GTP-Na, 12 phosphocreatine and 50 units/ml phosphocreatine kinase (300mOsm, pH 7.4). The recording chamber was constantly perfused with extracellular solution containing (in mM) 140 NaCl, 2.4 KCl, 10 HEPES, 2 CaCl2, 4 MgCl2, 10 glucose (pH adjusted to 7.3 with NaOH, 300 mOsm). A hypertonic 500 mM sucrose solution was applied for 5 seconds using a fast-flow system (Rosenmund and Stevens, 1996). The RRP was quantified by integrating the transient component of the evoked current.

Optogenetic stimulation and fast freeze EM

‘Flash-and-freeze’ experiments were essentially performed as previously described (Watanabe et al., 2013; Watanabe et al., 2014). Briefly, sapphire disks with cultured cells (6-8 weeks) were frozen using a HPM 100 high-pressure freezer (Leica) combined with a custom-built light stimulation device. All experiments were performed at room temperature to match electrophysiological recording settings. We stimulated the cells with 30 10-ms light pulses at 10 Hz before freezing. Controls for each experiment were always frozen on the same day from the same batch of cultured cells. Following high-pressure freezing, samples were transferred into vials containing 1% osmium tetroxide (EMS), 1% glutaraldehyde (EMS), and 1% milliQ water, in anhydrous acetone (EMS). The freeze-substitution was performed in an automated freeze-substitution device (AFS2, Leica) with the following program: −90°C for 5-7 h, 5°C per hour to −20°C, 12 h at −20°C, and 10°C per hour to 20°C. Following en bloc staining with 0.1% uranyl acetate, the samples were embedded into epon and cured for 48 h in a 60°C oven. Serial 40-nm sections were cut using a microtome (Leica UCT) and collected onto formvar-coated single-slot grids. Sections were stained with 2.5% uranyl acetate and lead citrate before imaging. For ferritin application, cells were immersed in solution containing ferritin (0.25 mg ml−1) with a reduced concentration of calcium (1 mM) for 5 min to reduce ferritin uptake by spontaneous activity during incubation. Cells were then transferred into solution containing 4 mM calcium and 1 mM magnesium and frozen 100 ms after light stimulation. To improve the contrast of ferritin, specimens were not stained with uranyl acetate.

The images were scored blind using an analysis program developed for ImageJ and Matlab. The active zones were defined as the portion of presynaptic membrane directly juxtaposed to the post-synaptic density. We defined docked vesicles as those contacting the plasma membrane and tethered vesicles as those with a visible tether and within 30 nm of the plasma membrane. The diameter of vesicles and the width of the synaptic cleft were measured from the outer leaflet of membrane. Vesicles were scored as clathrin-coated only if distinctive coats were visible, which can lead to underscoring.

The nearest neighbor distances were calculated using the ImageJ plugin developed by Yuxiong Mao provided online at https://icme.hpc.msstate.edu/mediawiki/index.php/Nearest_Neighbor_Distances_CalculCalcu_with_ImageJ.

Synapse density in cultures was measured by taking random clusters of pictures (EM) (usually 5×5 at 12,000x magnification) followed by manual counting of synapses per area. For this purpose we defined synapses as structures that clearly present a cluster of synaptic vesicles in the presynaptic compartment that opposes a contrast-rich PSD.

Chemical fixation for EM (in Figure 1)

4% Paraformaldehyde + 0.25% Glutaraldehyde in 0.1 M Sodium Cacodylate buffer pH 7.4 at room temperature for 45 mins. Washing in PBS, after-fixation with osmium (0.5% OsO4) in PBS and again washing followed by staining with 2% uranyl acetate. Samples were then embedded into epon and cured for 48 h in a 60°C oven. Sectioning and imaging were performed as in the previous paragraph.

Acute treatments

For acute treatments of human induced neurons or mouse hippocampal neurons the following reagents have been used: Solubilized (10 µM) forskolin (Tocris, Cat. No. 1099) was added to the cell culture medium 5 mins before beginning of electrophysiological or high-pressure freezing experiments. For immunocytochemistry or immunoblot experiments of cultured neurons 30 mins of incubation were used for the following solubilized reagents (concentrations if not otherwise indicated in the corresponding Figure): Forskolin (10 µM), the PKA-antagonist H89 (5 µM, Tocris, Cat. No. 2910), α2A adrenergic receptor agonist guanfacine (5 µM, Tocris, Cat. No. 1030), α2A adrenergic receptor agonist tizanidine (10 nM – 10 µM, Tocris, Cat. No. 3609), 5-HT7 receptor antagonist DR4485 (10 µM, Tocris, Cat. No. 5005), serotonin (100 µM, Tocris, Cat. No. 3547), norepinephrine (100 µM, Tocris, Cat. No. 5169), adenosine (100 µM, Tocris, Cat. No. 3624), dopamine (100 µM, Tocris, Cat. No. 3548), histamine (100 µM, Tocris, Cat. No. 3545), acetylcholine (100 µM, Tocris, Cat. No.2809). For treatment of ex vivo dissected mouse hippocampi 60 mins incubation with norepinephrine (100 µM), histamine (100 µM) or serotonin (100 µM) was performed. All incubations were performed in neuronal growth media (Neurobasal-A medium supplemented with B27/Glutamax, AraC (2 µM), Doxycycline (2 µg/ml) and 5% fetal bovine serum) at 37°C in the tissue culture incubator.

Long-term treatments

Solubilized reagents as indicated (and listed in the acute treatments paragraph) were added to the cell culture media for 1 week. Media were changed 50% every other day.

RNA-Sequencing data

Data were retrieved from GTExPortal (The Broad Institute of MIT and Harvard) (www.getexportal.org) using ‘gene expression in tissues’ for available brain tissues from 88 to 173 donors (post mortem) (Data source: Release V7 (dbGaP Accession phs000424.v7.p2). Box plots are shown as median and 25th and 75th percentiles.

QUANTIFICATION AND STATISTICAL ANALYSIS

All data shown are means ± SEMs; number of measured cells and/or independent experiments are indicated inside each bar, or mentioned in the figure.

All statistical analyses were performed using two-tailed Student’s t-test, one-way ANOVA, two-way ANOVA, or two-way ANOVA repeated measurements comparing the test sample to the control sample examined in the same experiments.

DATA AND CODE AVAILABLITY

All data reported here are published with this paper as a supplementary excel file (Table S1).

Supplementary Material

1

Figure S1. Related to Figure 1. Generation of induced human neurons, and measurements of synaptic puncta and protein levels after acute neuromodulator treatments

A Single transcription factor overexpression converts pluripotent stem cells into human neurons. Flow diagram of all experiments, except for 2-25-months-old neurons. Stem cells (day −1) were infected with lentiviruses expressing Ngn2, rtTA, and neuronal induction was activated by doxycyclin on the following day.

B Summary graphs of number of cells per area, positive for MAP2, human nuclei (HuNu), and DAPI. Human neuronal cultures (after 1 month of induction) used in all experiments (indicated in A). Cells from same micrographs were manually scored. Importantly, not a single MAP2-positive but HuNu-negative cell was found, indicating the complete absence of mouse neurons. Overlapping somata were leading to slight underestimation of MAP2-positive cells.

C Micrographs showing cells on day 1 and on day 14; upper panels: bright field; lower panels: ES cells were coinfected at day −1 with an EGFP-expressing lentivirus for visualizing human neurons.

D Representative confocal micrograph: Robust synapse-formation as visible in this example of a single human neuron in higher magnification showing MAP2, postsynaptic PSD95, and presynaptic Synapsin.

E Representative confocal micrographs depicting human neurons before and after acute treatment (30 mins) with H89 (5 µM), norepinephrine (‘Norepi’, 100 µM), or serotonin (‘Sero’, 100 µM), stained for MAP2 and the presynaptic vesicle marker Synaptotagmin-1 (‘Syt1’).

F-I Summary graphs of synaptic puncta area of Pan-Synapsin (E), Synaptophysin (F), Synaptotagmin-1 (G), and Piccolo (H) of acutely treated human neurons.

J Representative immunoblots and summary graph of synaptic protein levels after acute norepinephrine (100 µM) or guanfacine (5 µM) treatment of human neurons. Treatments do not significantly change synaptic protein levels. Values were normalized to controls (without reagent) and corrected for blotting and loading variations using TuJ1 as an internal standard.

K Representative immunoblot and summary graph of Phospho-Synapsin1 (serine 9) protein levels of acutely treated human neurons (30 mins) with serotonin (‘Sero’, 100 µM), 5-HT₇ receptor antagonist DR4485 (10 µM), and serotonin (100 µM) in combination with DR4485 (10 µM).

L Representative confocal micrographs showing expression of α2A adrenergic receptor on a subset of axons but not on dendrites of human neurons.

Data are means ± SEM; numbers of images/independent experiments analyzed are shown in the bars or graphs. Statistical significance for E-J was assessed using one-way ANOVA (*, P < 0.05; **, P < 0.01; ***, P < 0.001).

2

Figure S2. Related to Figure 2. Neuromodulator treatment of mouse neurons, expression of Synapsins in post mortem human brain tissues, and rescue experiments with synapsin-1 serine 9 point mutations

A-D Summary graphs of presynaptic puncta areas of acutely treated mouse hippocampal neurons (DIV14). BSN: bassoon. Puncta densities are shown in Figure 2 A–D.

E Summary plots of bulk RNA-sequencing results from different human brain tissues collected post mortem by GTEx consortium, Broad Institute, MIT and Harvard (https://gtexportal.org/home). Data are shown for neuronal markers: Synapsin-1, Synapsin-2, Synapsin-3, PSD95 (DLG4), L1CAM, STXBP1 (Munc18.1), SHANK3, and Neurexin1. In all analyzed human brain tissues the RNA levels for synapsins follow the order Syn1>Syn2>Syn3.

F and G Targeting strategy and genotyping PCR. The SYN1 gene was mutated by homologous recombination in male H1 ES cells using a recombinant AAV. Primers P1 and P2 were used for genotyping. 3 clones out of 30 tested clones were hemizygously targeted. Numbers 10 and 17 (subsequently referred to as clone #1 and clone #2) were used for the rest of the study.

H Structure of the human SYN1 gene and protein, depicting pathogenic mutations. ASD: autism spectrum disorder; E: epilepsy ID: intellectual disability.

I Flow diagram of all experiments using Syn1 cKO human neurons, except for 2-25-months-old neurons.

J-M Summary graphs of presynaptic puncta areas of acutely treated Syn1 cKO neurons (derived from clone #2). Guanfacine is a α2A adrenergic receptor agonist. H89 is an inhibitor of protein kinase A. Example pictures and puncta densities are shown in Figure 2 I–M.

N Flow diagram of rescue experiments, using human Syn1B recombinant lentiviruses on Syn1 cKO neurons (2 months after induction). Cells were allowed to express human (hs) Syn1-promotor driven Syn1B for 10 days, before functional analyses.

O Schematic representation of Syn1B rescue constructs for (de)phospho-mimetics. The serine 9 residue was mutated into alanine (S9A) (mimicking constitutive dephosphorylation), aspartate (S9D) (mimicking constitutive phosphorylation) or lysine (S9K) to create an additional positive charge. Additionally, WT Syn1B and untransduced neurons (Ctrl) were used.

P Confocal micrographs depicting human neurons expressing rescue constructs, using a Syn1 specific monoclonal antibody (clone 10.22, epitope different from Syn1 N-terminus). Syn1 adopts an almost homogenous axonal localization in the case of Syn1B S9D. The S9A and S9K mutations accumulate in nerve terminals.

Q and R Summary graphs of presynaptic puncta areas of rescued Syn1 cKO neurons (derived from clone #2). Example pictures and puncta densities are shown in Figure 2 O–Q.

Data are means ± SEM; numbers of images/independent experiments analyzed are shown in the bars or graphs. Statistical significance for B-D, J-M, and Q-R was assessed using one-way ANOVA (*, P < 0.05; **, P < 0.01; ***, P < 0.001).

3

Figure S3. Related to Figures 3–4. EM quantifications of presynaptic vesicles in human neurons

A Representative electron micrographs of synapses of Syn1 cKO neurons (derived from clone #1) showing large presynaptic circular and endosome-like vesicles (black arrows).

B Summary graphs of relative number of presynaptic large vesicles in human neurons after 35 days of culture. Large vesicles were defined as vesicles with a diameter greater than 60 nm. All endosome-like vesicular structures were scored as non-circular vesicles.

C Summary graphs of average diameter of large circular presynaptic vesicles.

D Summary plots for relative diameter distribution of large circular presynaptic vesicles. The majority of diameters are around 100 nm. No major change is observed following acute forskolin treatment and/or optogenetic stimulation.

E and F Representative electron micrograph of human WT (clone #1) neurons after 35 days of culture showing one dense core vesicle (arrow head) and summary graphs of relative number of presynaptic dense core vesicles.

G and H Representative electron micrographs of human neurons after 35 days of culture with presynaptic clathrin coated vesicles (arrows) and summary graphs of relative number of presynaptic clathrin coated vesicles which have been classified according to their sub-cellular localization. Scoring was done as follows: Plasma membrane vesicles are vesicles fused with the plasma membrane. Endosomatic vesicles are vesicles fused to endosome-like vesicles. Cytoplasmic vesicles are all other non-fused vesicles in the presynaptic cytoplasm.

Data are means ± SEM; Statistical significance was assessed using one-way ANOVA; numbers of images/independent experiments analyzed are shown in the bars. Forskolin: ‘FSK’

4

Figure S4. Related to Figures 3–4. EM quantifications of postsynaptic vesicles in human neurons

A Representative electron micrographs of human iN cells after 35 days of culture (control: left; Syn1-deficient: right). Note that circular and non-circular (endosome-like) postsynaptic vesicles are visible.

B Summary graphs for the relative number of postsynaptic vesicles of control and Syn1-deficient human neurons, with and without optogenetic stimulation and with or without acute forskolin treatment. All endosome-like vesicles were scored as non-circular vesicles.

C Summary graphs of average diameter of postsynaptic vesicles.

D Summary plots for relative diameter distribution of circular postsynaptic vesicles. The majority of diameters are around 50 nm and second most frequent values are around 70-80 nm. Few values are above 100 nm.

E and F Representative electron micrograph of human WT iN cells after 35 days of culture showing one dense core vesicle (arrow head) postsynaptically and summary graphs of relative number of postsynaptic dense core vesicles.

G and H Representative electron micrographs of human iN cells after 35 days of culture with postsynaptic clathrin coated vesicles (arrows) and summary graphs of relative number of postsynaptic clathrin coated vesicles which have been classified according to their sub-cellular localization. Scoring was done as follows: Plasma membrane vesicles are vesicles fused with the plasma membrane. Endosomatic vesicles are vesicles fused to endosome-like vesicles. Cytoplasmic vesicles are all other non-fused vesicles in the postsynaptic cytoplasm.

Data are means ± SEM; Statistical significance was assessed using one-way ANOVA; numbers of images/independent experiments analyzed are shown in the bars. Forskolin: ‘FSK’

5

Figure S5. Related to Figure 5. Pre- and postysynaptic ‘puncta’ after acute forskolin treatment or deletion of Syn1

A Representative confocal micrographs depicting Synaptotagmin-1 or Synaptophysin puncta and MAP2 positive dendrites of WT and isogenic Syn1-deficient human neurons.

B and C Summary graphs for puncta areas of acutely forskolin (‘FSK’) treated human neurons (10 µM for 30 min) and conditionally SYN1 deleted neurons. Puncta densities are shown in Fig. 4 A and B.

D-I Representative confocal micrographs (D and G) and summary graphs for presynaptic Bassoon and postsynaptic PSD95 puncta densities (E and H) and areas (F and I).

J-L Representative confocal micrographs and summary graphs for presynaptic Phospho (P)-Pan-Syn and Syn1, as well as Pan-Syn (see Fig. 4A) puncta densities and areas with and without acute FSK treatment.

M Summary graphs for postsynaptic Homer puncta area. Puncta densities are shown in Fig. 4C.

N MAP2-intensities per cell assessed by analyses of immnuo-labeled human neurons. No overall change is detected in SYN1-deficient cells from both clones.

Data are means ± SEM; statistical significance was assessed using Student’s t test (*, P < 0.05). Numbers of images/independent experiments analyzed are shown in the bars. Forskolin: ‘FSK’

6

Figure S6. Related to Figure 6. Electrophysiology of human neurons

Intrinsic neuronal properties, miniature EPSCs, evoked EPSCs during single stimulus, paired stimuli, stimulus train and sucrose response of human neurons conditionally mutated for Syn1 and/or acutely treated with forskolin.

A and B Capacitance (A) and Input resistance (B) for both clones of hemizygous Syn1-mutant iN cells (day 35).

C Representative traces of analyses of the action potential firing properties of control and hemizygous Syn1-mutant iN cells (day 35). Neurons held in current-clamp mode were injected with increasing current pulses.

D-G Summary graphs of absolute EPSC amplitudes during stimulus train of both conditional mutants (after 35 days of culture) without (D and F) and with (E and G) acute treatment of 10 µM forskolin. Representative traces and more summary graphs are shown in Fig. 6G–K.

H and I Measurements of release induced by hypertonic sucrose to assess the size of the RRP of vesicles. Representative traces from both clones (H) and summary graphs as absolute values (I). These experiments were performed by applying 0.5M sucrose for 5 s.

J Detailed analysis of synaptic responses evoked by single action potentials showing summary graphs for amplitude, coefficient of variation (C.V.), charge transfer, rise time (20%-80%) and half-width of the responses. No changes in the C.V. were observed.

K-O Repetition of paired-pulse ratios (PPRs) using different stimulus intervals (20-1000ms) and responses during stimulus trains (10Hz) for clone #2. These data are generated using the same sets of human neurons used for the sucrose-puffing experiments shown in Figure 6L and M and represent repetitions of the measurements shown in Figure 6D–K.

P-T Representative traces (P) of miniature excitatory postsynaptic currents (mEPSCs), cells were held at −80 mV holding potentials. mEPSCs were recorded in the presence of TTX (0.5 uM) in the bath solution. Q (clone #1) and S (clone #2), mEPSC frequency: cumulative plots of inter-event intervals (inset: mean frequency). R (clone #1) and T (clone #2), mEPSC amplitude: cumulative plots of amplitudes (inset: mean amplitude). Deletion of Syn1 leads to decreased spontaneous neurotransmitter release.

Data are means ± SEM; numbers of images/independent experiments analyzed are shown in the bars or graphs; statistical significance for A, B, I, J, K, Q, R, S, T was assessed by Student’s t-test, for D-G and M-O by two-way ANOVA (*, P < 0.05; **, P < 0.01; ***, P < 0.001).

7

Figure S7. Related to Figure 7. Analyses of human neurons after 4, 18 and 25 months of culture

A Summary graphs (Syn1 clone #2) showing puncta areas of acutely guanfacine (Guanf) (5 µM, 30 mins), serotonin (Sero) (100 µM), H89 (100 µM) and norepinephrine (100 µM) treated and control 4-months-old human neurons stained for Synapsin-2 (Syn2), pan-synapsin (pan-Syn), Synaptophysin (Syph) and Bassoon (BSN). Example pictures and puncta densities are shown in Figure 7 D–H.

B Representative immunoblot and summary graph showing protein levels of phosphorylated Syn2 (highlighted by *) of acutely treated 4-months-old human neurons (Cre Syn1 cKO clone #2, ΔCre in left lane shown for reference). Forskolin (‘FSK’) treatment increases and H89, and guanfacine treaments decrease levels of Phsopho-Syn2 (serine 9) similar to Phospho-Syn1 levels in Figures 5F and 1G.

C Summary graph of 4-months-old mutant and wild type neurons showing protein levels for Synapsin-1 after 1 week of treatment with indicated reagents. A representative immunoblot and protein levels of other synaptic markers are shown in Figure 7I–M.

D DIC picture of a wild type (clone #2) 18-months-old human neuron with patch-pipette.

E Representative confocal micrographs of 18-months-old Syn1 cKO human neurons (4 left upper panels), transfected with GFP and stained for MAP2. Representative confocal micrographs showing Syn1-expression of WT and conditional KO human neurons (top row 2, right panels). Representative confocal micrographs of an 18-months-old wild type (ΔCre, Syn1 cKO clone #2) human neuron, showing spine-like structures on the dendrites (middle row 2 right panels and zoom-in bottom panel). Two weeks before fixation and staining cells were infected with lentivirus expressing EGFP under the control of the human Synapsin-1-promotor.

F Density of dendritic spine-like structures, and MAP2-intensities per cell (18 months, mutant and wild type) as indicator for dendritic morphology.

G Representative traces of AP firing properties of control and SYN1 mutant neurons (clone #2). Neurons were maintained at near −80 mV in current-clamp mode and were depolarized with increasing current pulses for 2 s. Note that the AP-firing pattern looks very different from 1-month old cells shown in Figure S6C.

H and I Summary graphs of the intrinsic properties including capacitance, excitability (rheobase), input resistance and resting membrane potential.

J-L Decreased frequency of spontaneous EPSCs recorded in the absence of TTX. Representative traces (J), cumulative plot of the spontaneous EPSC inter-event interval (K) (inset: mean frequency) and cumulative plot of the spontaneous EPSC amplitude (L) (inset: mean amplitude).

M Representative confocal micrographs of 25-months-old human neurons (ΔCre and Cre, Syn1 KO clone #2) transduced for expression of EGFP (like E) and stained for dendritic MAP2 and axonal L1CAM, or (Cre, clone #2) stained for MAP2 and Pan-Synapsin.

N Representative confocal micrographs of dendrites of human neurons after 25 months of culture period showing Synaptophysin- and Synaptotagmin-1 -containing puncta, and summary graphs (Syn1 clone #2) for synaptic densities and area of Synaptophysin-, Synaptotagmin-1- and Homer-containing puncta.

Data are means ± SEM; Numbers of images/independent experiments analyzed are shown in the bars or graphs or indicated as n. Statistical significance for F, H, I, K, L and N was assessed using Student’s t test and using one-way ANOVA for A-C (*, P < 0.05; **, P < 0.01; ***, P < 0.001).

Supplementary table

Table S1. Related to Figures 1–7 and S1–S7. Combined data of the paper.

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
Mouse monoclonal anti-Synaptotagmin-1	Synaptic Systems	Clone 41.1
Mouse monoclonal anti-Synaptotagmin-1	DSHB	Clone asv 48
Mouse monoclonal anti Human Nuclei	Millipore	Clone 235-1
Rabbit polyclonal anti phospho-Synapsin	Hosaka et al. 1999	U342
Chicken polyclonal ant MAP2	Encor	CPCA-MAP2
Rabbit polyclonal anti Synapsin	This paper	E028
Mouse monoclonal anti PSD95	Pierce / Thermo	Clone 7E5-1B8
Mouse monoclonal anti Synapsin-1	Synaptic Systems	Clone 10.22
Mouse monoclonal anti Synapsin-2	Synaptic Systems	Clone 19.31
Mouse monoclonal anti Synaptophysin	Synaptic Systems	Clone 7.2
Rabbit polyclonal anti GFP	Invitrogen	Cat# A11122
Rabbit polyclonal anti Bassoon	Sigma	Cat# SAB5200101
Rabbit polyclonal anti Bassoon	Synaptic Systems	Cat# 141 003
Rabbit polyclonal anti Piccolo	Synaptic Systems	Cat# 142 003
Rabbit polyclonal anti Homer-1b/c	This paper	YZ6085
Rabbit polyclonal anti PSD95	This paper	L667
Rabbit polyclonal anti-Adrenergic Receptor α-2A	Sigma	Cat# SAB4500548
Mouse monoclonal TuJ1	Covance	Cat# MMS-435P
Mouse monoclonal anti L1CAM	Sigma	Clone UJ 127.11
Rabbit polyclonal TuJ1	Sigma	Cat# T2200
Rabbit polyclonal anti Complexin 1/2	This paper	L668
Rabbit polyclonal anti SNAP25	This paper	P913
Rabbit polyclonal anti Synaptobrevin-2	This paper	P939
Rabbit polyclonal anti Syntaxin-1	This paper	438B
Mouse monoclonal anti calmodulin-associated serine/threonine kinase	BD transduction laboratories	Clone 7/CASK
Rabbit polyclonal anti SynCAM	This paper	T2412
Rabbit monoclonal anti PSD95	AbCAM	Cat# ab76115
Rabbit polyclonal anti L1CAM	Wolff et al. 1988	N/A
Mouse monoclonal anti Munc18	BD transduction laboratories	Cat# 610336
Mouse monoclonal anti N-Cadherin	Millipore	Cat# 13A9
Rabbit polyclonal anti Rab3A	This paper	T957
Rabbit polyclonal anti Dynamin1	This paper	E027
Rabbit polyclonal anti SV2A	This paper	P915
Mouse monoclonal anti Synapsin-2	Synaptic Systems	Clone 19.41
Rabbit polyclonal anti NeuN	Millipore	Cat# ABN78
Chemicals, Peptides, and Recombinant Proteins
Accutase	Innovative Cell	Cat# AT104
Ponceau S	Sigma-Aldrich	Cat# P3504-10G
Benzonase	Sigma-Aldrich	Cat# E1014-5KU
B-27 supplement	Thermo Fisher	Cat# 12587010
BDNF	PeproTech	Cat# 450-02
Blasticidin	Sigma-Aldrich	Cat# 203350
Doxycycline	Sigma-Aldrich	Cat# D9891
DMEM media	Thermo Fisher	Cat# 11965092
DMEM/F-12 media	Thermo Fisher	Cat# 11320082
HyClone fetal bovine serum (FBS)	GE Healthcare	Cat# SH30071.03
MEM media	Thermo Fisher	Cat# 51200038
Matrigel	BD Biosciences	Cat# 356230
mTeSR1 media	STEMCELL Technologies	Cat# 85850
N-2 supplement	Thermo Fisher	Cat# 17502048
Neurobasal-A Medium	Thermo Fisher	Cat# 10888022
NT-3	PeproTech	Cat# 450-03
Puromycin	Sigma-Aldrich	Cat# P8833
Thiazovivin	Bio Vision	Cat# 1681
Laminin-1	Invitrogen	Cat# 23017-015
Trypsin / EDTA	Thermo Fisher	Cat# 25300120
Papain, Suspension	Worthington Biochemical Corporation	Cat# LS003127
Cytosine β-D-arabinofuranoside hydrochloride (AraC)	Sigma-Aldrich	Cat# C6645
Glutamax	Thermo Fisher	Cat# 35050079
Forskolin	Tocris	Cat# 1099
H89	Tocris	Cat# 2910
Guanfacine	Tocris	Cat# 1030
Tizanidine	Tocris	Cat# 3609
DR4485	Tocris	Cat# 5005
Serotonin	Tocris	Cat# 3547
Norepinephrine	Tocris	Cat# 5169
Adenosine	Tocris	Cat# 3624
Dopamine	Tocris	Cat# 3548
Histamine	Tocris	Cat# 3545
Acetylcholine	Tocris	Cat# 2809
Dispase	Stem Cell Technologies	Cat# 07923
TTX	Fisher Scientific	Cat# 50-753-2807
EDTA (Ethylenediaminetetraacetic acid)	Sigma-Aldrich	Cat# E6758-100G
Non-Essential Amino Acids Solution (NEAA)	Thermo Fisher	Cat# 11140076
cationized ferritin	Sigma-Aldrich	Cat# F7879-2ML
Experimental Models: Cell Lines
Human H1-ESCs	WiCell Research Institute	Cat# WA01
HEK293T cells	ATCC	Cat# CRL-11268
Experimental Models: Organisms/Strains
C57BL/6J mouse	The Jackson Laboratory	Cat# 000664
CD-1 IGS mouse	Charles River	Cat# Crl:CD1(ICR)
Oligonucleotides
GGGAAAAAGATCCATGGAGAAATTGACATTAAA	IDT	N/A
GCCTTGCTCATAACACCAAGGTACCCTTC	IDT	N/A
CGGCGCCGCCTGGATGACAGCAACTTTATG	IDT	N/A
CATAAAGTTGCTGTCATCCAGGCGGCGCCG	IDT	N/A
CGCCGCCTGGCGGACAGCAACTTTATG	IDT	N/A
CATAAAGTTGCTGTCCGCCAGGCGGCG	IDT	N/A
CGGCGCCGCCTGAAGGACAGCAACTTTATG	IDT	N/A
CATAAAGTTGCTGTCCTTCAGGCGGCGCCG	IDT	N/A
GGTCTAGAGCAGCCATGAACTACCTGCGGCGC	IDT	N/A
CCGGATCCCTAAGGCTGGGCCTGGGC	IDT	N/A
Recombinant DNA
FUW-TetO-Ng2-T2A-puromycin (lentiviral vector)	Zhang et al. 2013	52047 Addgene
FUW-rtTA (lentiviral vector)	Zhang et al. 2013	20342 Addgene
FUW-TetO-EGFP (lentiviral vector)	Vierbuchen et 2010	30130 Addgene
FSW-Syn-EGFP (lentiviral vector)	This paper	N/A
FUW-Flp-IRES-Blast (lentiviral vector)	This paper	N/A
FUW-GFP::Cre (lentiviral vector)	This paper	N/A
FUW-GFP::ΔCre (lentiviral vector)	This paper	N/A
FSW-ChetaTC (lentiviral vector)	Berndt et al. 2011	N/A
AAV vector for targeting hSYN1	This paper	N/A
pHelper for AAV	Grimm et al. 2008	N/A
pDJ for AAV	Grimm et al. 2008	N/A
Human Syn1B	The ORFeome Collaboration	HsCD00297161 Harvard PlasmID
FSW-Syn-Human Syn1B S9A (lentiviral vector)	This study	N/A
FSW-Syn-Human Syn1B S9D (lentiviral vector)	This study	N/A
FSW-Syn-Human Syn1B S9K (lentiviral vector)	This study	N/A
pRSV-REV (lentiviral helper plasmid)	Dull et al. 1998	12253 Addgene
pMDLg/pRRE (lentiviral helper plasmid)	Dull et al. 1998	12251 Addgene
vsv-g expression vector (pMD2.G) (lentiviral helper plasmid)	Gift from Didier Trono	12259 Addgene
Software and Algorithms
ImageJ	NIH	RRID: SCR_003070
ImageStudio software	LiCor	RRID:SCR_013715
pClamp	Molecular Devices	RRID:SCR_011323
NND plug-in for ImageJ	Yuxiong Mao	https://icme.hpc.msstate.edu/mediawiki/index.php/Nearest_Neighbor_Distances_CalculCalcu_with_ImageJ
Nikon NIS-Elements	Nikon	RRID:SCR_014329
GraphPad Prism	GraphPad Prism	RRID:SCR_002798
Other
GTExPortal	The Broad Institute of MIT and Harvard	www.getexportal.org

Highlights:

• Neuromodulator-GPCR signaling rapidly changes synaptic vesicle numbers

• cAMP-dependent decrease of synapsin-1 phosphorylation causes vesicle recruitment

• cAMP-dependent synapsin-1 phosphorylation triggers removal of synaptic vesicles

• Neuromodulators-cAMP-synapsin-1 pathway bidirectionally controls vesicle numbers