A multiple sequence alignment of human α-, β-, γ- and lamprey γ-synucleins was created using the msa package in R software. The GenBank accession numbers for the sequences used in the alignment were: human α-synuclein (NM_000345.4); human β-synuclein (NM_001001502.3); human γ-synuclein (NM_003087.3); and lamprey γ-synuclein (JN544525.1). Disorder probability and charge distribution analyses were performed using the PrDOS bioinformatics tool and EMBOSS package, respectively.

Standard Western blotting procedures were performed as previously described (Busch et al., 2014). For all experiments, 10 μg of rat brain lysates and 20 μg of lamprey CNS (brain and spinal cord) lysates were separated on 10% or 12% SDS-PAGE gels. Primary antibodies were as follows: anti-pan synuclein rabbit polyclonal antibody raised against amino acids 11-26 within the N-terminal domain of human α-synuclein (1:1000; ab6176; Abcam, Cambridge, MA); an anti-synaptobrevin (VAMP) mouse monoclonal antibody (1:1000; 1933-SYB, clone SP10, PhosphoSolutions, Aurora, CO); and an anti-SV2 mouse monoclonal antibody (1:1000; Developmental Studies Hybridoma Bank; University of Iowa, Iowa City, IA, United States). The SV2 antibody was deposited to the DSHB by Buckley, K.M., and it labels SV clusters in all vertebrates tested, including lampreys (DSHB Hybridoma Product SV2) (Buckley and Kelly, 1985; Lau et al., 2011; Busch and Morgan, 2012; Busch et al., 2014). Secondary antibodies were goat anti-rabbit or goat anti-mouse HRP conjugated IgG (H + L), as appropriate (Thermo Scientific, Waltham, MA, United States). Protein bands were detected using Pierce™ ECL Western blotting substrate (Thermo Scientific, Waltham, MA, United States) and imaged using an Azure Imaging System 300 (Azure Biosystems; Dublin, CA, United States).

Animal procedures were approved by the Institutional Animal Care and Use Committee at the MBL in accordance with standards set by the National Institutes of Health. Late larval lampreys (Petromyzon marinus; 11–13 cm; 5–7 years old; M/F) were anesthetized in 0.1–0.2 g/ L Tricaine-S (Syndel; Ferndale, WA, United States). Segments of spinal cord (2–3 cm) were dissected, pinned ventral side up in a Sylgard petri dish, and stripped of meninx. Axonal microinjections were performed as previously described (Busch et al., 2014; Walsh et al., 2018; Banks et al., 2020; Soll et al., 2020; Román-Vendrell et al., 2021). The antibodies injected included an anti-pan-synuclein antibody (ab6176; Abcam), anti-SV2 antibody (DSHB); and anti-synaptobrevin (VAMP) antibody (1933-SYB/SP10, PhosphoSolutions). Rabbit polyclonal IgGs (ab37415; Abcam) were also injected in order to provide a negative, isotype control for the synuclein antibody. Prior to injection, all antibodies were diluted in lamprey internal solution (180 mM KCl and 10 mM HEPES K⁺; pH 7.4) to a final pipet concentration of 0.5 mg/ml and mixed with fluorescein dextran (0.1 mM; 70 kDa; Thermo Fisher) in order to monitor the injections in real time. Antibodies were then loaded into glass microelectrodes and microinjected into giant RS axons using small, repeated pulses of N₂ (5–30 ms, 40 psi, 0.2–0.3 Hz) delivered through a Toohey spritzer. Injections lasted 15–20 min and typically resulted in a 20-100x dilution of the antibody within the axon, based on the fluorescence of the co-injected dye. After injection, the spinal cords were immediately fixed in 3% glutaraldehyde, 2% paraformaldehyde in 0.1 M Na cacodylate, pH 7.4 overnight.

After fixation, spinal cords were processed for electron microscopy, thin sectioned at ∼70 nm, and counterstained with 2% uranyl acetate and 0.4% lead citrate, as detailed previously (Busch et al., 2014; Walsh et al., 2018; Banks et al., 2020; Soll et al., 2020; Román-Vendrell et al., 2021). Ultrastructural images were obtained at ×37,000 magnification using a JEOL JEM 200CX electron microscope. Serial images of giant reticulospinal synapses were acquired from at least n = 20 synapses, from 2-3 axons/lampreys for each experimental condition. Synapses were analyzed from three different regions of the injected axon (representing different concentration ranges), based on axonal diffusion patterns of the fluorescein dextran: >400 μm from the injection site, beyond where the injected protein had diffused (Untreated Control); 150–390 μm from the injection site (Low concentration); and 30–140 μm from the injection site (High concentration). Each experiment was therefore internally controlled, reducing variance due to natural differences in SV clusters between axons and animals.

Morphometric analysis was performed on a single image per synapse, taken at or near the center of the active zone, as previously described (Busch et al., 2014; Walsh et al., 2018; Banks et al., 2020; Soll et al., 2020; Román-Vendrell et al., 2021). These included synaptic vesicles (SVs), plasma membrane, cisternae (putative endosomes), as well as clathrin-coated pits (CCPs) and clathrin-coated vesicles (CCVs). SVs were defined as small, clear round vesicles <100 nm in diameter, while “cisternae” were defined as larger vesicles that were >100 nm in diameter. Plasma membrane evaginations were determined by drawing a straight 1 μm line from the edge of the active zone to the nearest position on the axolemma on both sides of the synapse, then measuring the curved distance between these points, and averaged. CCPs and CCVs were staged as detailed previously (Morgan et al., 2004). After obtaining measurements for each organelle, a total membrane analysis was performed on each synapse to determine if and how synaptic membranes were redistributed under each experimental condition. SV and CCP/V membrane areas were calculated by multiplying the surface area of a sphere (4πr²) by the number of each type of vesicle at each synapse. Membrane areas associated with plasma membrane and cisternae were obtained by multiplying the length of plasma membrane evaginations and summed cisternae perimeters, respectively, by the section thickness (70 nm). In addition, the SV distribution and nearest neighbor analysis were determined using a Python script (based on script in Banks et al., 2020; https://github.com/kfouke/Morgan-Lab), which measured the distance from the center of each SV to the nearest point on the active zone and to the center of the nearest SV. Graphing and statistical analyses, including ANOVA and linear regressions, were performed in GraphPad Prism 9.0.

The goal of this study was to determine whether synuclein modulates SV clustering at resting synapses, given its known interactions with synapsin. Mammals, including humans, express three isoforms of synuclein: α-synuclein, β-synuclein, and γ-synuclein. As in mammals, lampreys, which are jawless vertebrates, also express three isoforms of synuclein: two γ-synucleins (as observed in other fishes) and a third synuclein isoform that remains unassigned because it did not reliably group in a phylogenetic analysis with other α-, β-, or γ-synuclein orthologs (Busch and Morgan, 2012; Smith et al., 2013). Our prior study demonstrated that the most abundant isoform expressed within lamprey giant RS neurons is a γ-synuclein (GenBank: JN544525), whereas the other two synuclein isoforms were expressed at low or undetectable levels (Busch and Morgan, 2012). We therefore focus our analysis on this abundant γ-synuclein isoform and refer the reader to our prior study for details on the other lamprey synuclein isoforms (Busch and Morgan, 2012). The primary amino acid sequence of lamprey γ-synuclein is 56, 54, and 55% identical to full length human α-, β-, and γ-synuclein, respectively (Figure 1A). The highly conserved N-terminal domain of synuclein (a.a. 1-95), which folds into an amphipathic α-helix and contains the non-amyloid component (NAC; a. a. 60-95), is 66% identical and 78% similar between human α-synuclein and lamprey γ-synuclein orthologs (Figure 1A). In contrast, the C-termini are more variable between all synuclein orthologs. We previously reported the predicted structure of lamprey γ-synuclein, which is an N-terminal α-helix followed by a less structured random coil at the C-terminus, similar to the structure of human α-synuclein bound to lipid micelles (Ulmer et al., 2005; Busch and Morgan, 2012).

To further compare the structures of human α-synuclein and lamprey γ-synuclein, we ran disordered region and charge distribution predictions on their sequences using PrDOS (https://prdos.hgc.jp/cgi-bin/top.cgi) and EMBOSS (https://www.bioinformatics.nl/cgi-bin/emboss/charge), respectively. Although the lamprey synuclein protein is shorter in length, the protein disorder prediction analysis revealed nearly identical disorder probability between lamprey γ-synuclein and human α-synuclein (Figure 1B). Both orthologs possess intrinsically-disordered regions (IDRs) at their C-termini (Figure 1B). Similarly, the charge distributions across human α-synuclein and lamprey γ-synuclein proteins are nearly identical (Figure 1C). As another indicator of conservation, a pan-synuclein antibody (ab6176; Abcam) raised against a peptide in the N-terminal domain of human α-synuclein (a.a. 11-26) recognized monomeric synuclein in both lamprey CNS and rat brain lysates (Figure 1D). The high degree of conservation is further supported by our prior study, which demonstrated that the N-terminal domain of lamprey γ-synuclein can also bind avidly to small lipid vesicles in vitro, like human α-synuclein (Busch et al., 2014). Moreover, when introduced in excess at stimulated lamprey synapses, recombinant lamprey γ-synuclein, its N-terminal domain, and human α-synuclein all phenocopied each other and inhibited SV endocytosis (Busch et al., 2014). Thus, the current data indicate that lamprey γ-synuclein and human α-synuclein are highly conserved both structurally and functionally, though we acknowledge that the variation in amino acid sequence within the C-terminus may lead to some functional differences that are as yet to be determined.

To determine whether synuclein regulates SV clustering in vivo, we acutely introduced the pan-synuclein antibody (ab6176; Abcam) to lamprey giant RS synapses via axonal microinjection (Figure 2A). We previously reported that this pan-synuclein antibody recognizes all three lamprey synucleins (Busch and Morgan, 2012). However, we predict that the antibody injections predominantly disrupt one of the γ-synuclein isoforms (GenBank: JN544525), since that is the only synuclein isoform expressed at appreciable levels within the lamprey giant RS neurons (Busch and Morgan, 2012).

RS synapses are en passant, glutamatergic synapses that reside along the perimeter of the giant RS axons within the ventral spinal cord. Axonal microinjection therefore delivers the antibody directly to the presynaptic SV clusters (Figure 2A). Within the dissected spinal cord, the giant RS axons and synapses are quiescent unless exogenously stimulated with action potentials. Thus, in the absence of stimulation, we can determine how different perturbations affect resting SV clusters at unstimulated synapses. Untreated, resting RS synapses possessed a large, tight cluster of SVs adjacent to the active zone (Figure 2B). Similarly, unstimulated synapses treated with Control IgG antibodies (Control Ab; Rabbit IgG isotype control; Thermo Fisher) also exhibited large SVs clusters without any noticeable changes in morphology (Figure 2C). In contrast, the pan-synuclein antibody (Synuclein Ab; ab6176; Abcam) severely disrupted the SV clusters, leaving only a few vesicles around the active zone (Figure 2D). Quantitative analysis confirmed that treatment with the synuclein antibody caused a >75% reduction in the number of SVs at resting synapses (Figure 2E; Untreated: 121 ± 7 SVs, n = 62 synapses, 5 axons; Control Ab: 130 ± 10 SVs, n = 29 synapses, 3 axons; Syn Ab: 29 ± 2 SVs, n = 23 synapses, 2 axons; ANOVA, p < 0.0001, Tukey’s post hoc). Although the remaining SVs sometimes appeared misshapen, their mean diameters were not significantly altered (Untreated: 51.5 ± 0.6 nm; Syn Ab-Low 50.6 ± 0.7 nm; Syn Ab-High 49.2 ± 0.8 nm; n = 206–209 SVs, 23–31 synapses; ANOVA p = 0.90; Tukey’s post hoc).

To determine whether the SV membrane was redistributed to other synaptic compartments, such as the plasma membrane, cisternae (putative endosomes), and/or clathrin-coated pits or vesicles (CCP/Vs), we performed a total membrane analysis for each synapse. The dramatic reduction of SV membrane was only partially compensated by modest increases in plasma membrane evaginations and cisternae, resulting in a 40% net loss of total membrane at synapses (Figure 2F; Untreated: 1.44 ± 0.06 μm², n = 62 synapses, 5 axons; Control Ab: 1.50 ± 0.09 μm², n = 29 synapses, 3 axons; Syn Ab: 0.87 ± 0.04 μm², n = 23 synapses, 2 axons; ANOVA, p < 0.0001, Tukey’s post hoc). Thus, acute disruption of synuclein function at resting synapses induced a severe depletion of SVs at the active zone. The loss of total membrane suggests that the SVs escaped away from the immediate synaptic area.

To determine the specificity of the phenotype produced by the synuclein antibody, we also injected RS axons with antibodies raised against two other abundant SV proteins, synaptic vesicle glycoprotein 2 (SV2) and vesicle-associated membrane protein (VAMP/synaptobrevin) (Figure 3A). Like synapsin, VAMP2 is another known binding partner of α-synuclein (Burre et al., 2010; Diao et al., 2013; Sun et al., 2019). The SV2 antibody (DSHB) is a mouse monoclonal raised against synaptic vesicles purified from the electric ray (Discopyge ommata), and it recognizes synaptic vesicle clusters in all vertebrates tested, including lampreys (Buckley and Kelly, 1985; Lau et al., 2011; Busch and Morgan, 2012; Busch et al., 2014). Previous studies using immunofluorescence and immunogold EM techniques showed that this SV2 antibody strongly labels the SV clusters at resting lamprey giant synapses, demonstrating that it penetrates the SV cluster and reaches accessible epitopes (Bloom et al., 2003; Busch et al., 2014; Banks et al., 2020). The VAMP antibody (1933-SYB/SP-10; PhosphoSolutions) is a mouse monoclonal raised against crude synaptic immunoprecipitate from human brain. As shown by Western blotting, these antibodies recognized protein bands of the predicted molecular weights for SV2 (95 kDa) and VAMP (∼16 kDa) in both lamprey CNS and rat brain lysates (Figure 3B).

Compared to untreated control synapses, resting synapses treated with either the anti-SV2 antibody or anti-VAMP antibody had a normal appearance with a large pool of tightly clustered SVs, and no disruptions were detected (Figures 3C–E). There was no significant difference in the number of SVs after treatment with either the SV2 or VAMP antibody (Figure 3F; Untreated: 140 ± 10 SVs, n = 38 synapses, 4 axons; SV2 Ab: 158 ± 11 SVs, n = 24 synapses, 2 axons; VAMP Ab: 121 ± 10 SVs, n = 25 synapses, 2 axons; ANOVA, p = 0.085). Likewise, there was no change in the total membrane or any of the individual synaptic compartments (i.e., SVs, PM, cisternae, CCP/Vs) (Figure 3G; Untreated: 1.51 ± 0.09 μm², n = 38 synapses, 4 axons; SV2: 1.69 ± 0.14 μm², n = 24 synapses, 2 axons; VAMP: 1.42 ± 0.10 μm², n = 25 synapses, 2 axons; ANOVA, p = 0.147). Therefore, the loss of synaptic vesicles observed with the pan-synuclein antibody was specific to disrupting synuclein and not simply a non-specific phenotype induced by interfering with any SV-associated protein.

We next examined whether there was a dose-dependent effect of the synuclein antibody. Axonal microinjection results in a concentration gradient of synuclein antibody down the axon, due to lateral diffusion, with the highest concentration around the injection site (see Figure 2A). Based on the diffusion pattern of the co-injected fluorescent dye, we categorized synapses as receiving a “high” concentration (20–140 μm from the injection site) or “low” concentration of antibody (150–390 μm from the injection site). At distances farther from the injection site (>400 μm), the synapses received no antibody, thus providing untreated synapses as an internal control. Indeed, the pan-synuclein antibody induced a dose-dependent loss of SVs with increasing antibody concentration (Figure 4A). At many synapses, smaller subclusters of SVs were observed detaching from the main vesicle cluster, suggesting that synaptic vesicles were de-clustering into smaller units (Figure 4A, red arrowheads). De-clustering of individual SVs was also observed (Figure 4A, green arrowheads). In contrast, after injection of the Control IgG antibodies, the SV clusters remained large and tightly clustered at all distances evaluated (Figure 4B). At synapses treated with the synuclein antibody, the SV cluster sizes were positively correlated with distance from the injection site, indicating a dose-dependent response, whereas this was not observed with the Control antibodies (Figure 4C) (Synuclein Ab: slope = 0.1158, R² = 0.394; Control Ab: slope = -0.0011, R² < 0.0001; linear regression, p < 0.0001). This dose-dependence was also apparent in the binned distribution of SVs (Figure 4D) (two-way ANOVA, p < 0.01, Tukey’s post hoc). Synuclein perturbation with the pan-synuclein antibody (high concentration) induced a ∼50–80% loss of SVs at all distances evaluated (Figure 4E). This included a loss of the docked SVs closest to the active zone (Figure 4E). Synuclein is therefore necessary for proper SV clustering at synapses, including the maintenance of docked SVs and the distal reserve pool.

To further explore the phenotype, we generated 3D reconstructions of synuclein antibody-treated synapses. At untreated synapses, SVs are typically tightly clustered into a single, large SV cluster at the active zone (Figure 5A). Further supporting a role for synuclein in SV clustering, the pan-synuclein antibody-treated synapses often exhibited smaller, discrete SV subclusters, which were dispersed away from the active zone (Figures 5B,C, red arrowheads). These miniature SV subclusters were clearly separated from the main SV cluster and varied in size and shape (Figures 5B–G, red arrowheads). Single SVs were also observed as part of the de-clustering phenotype (Figures 5D–G, green arrowheads). To quantify the SV de-clustering, we performed a nearest-neighbor analysis, which measured the shortest distance between the center of each SV and its nearest neighbor. Synaptic vesicles that are tightly clustered together have a nearest neighbor within ∼50–60 nm, which is approximately the diameter of a synaptic vesicle at lamprey synapses (Medeiros et al., 2017; Román-Vendrell et al., 2021). As expected, the untreated synapses were tightly clustered (Figure 5H, gray). In contrast, synapses treated with the pan-synuclein antibody were de-clustered, most especially with the higher concentration of antibody (Figure 5H, blue and red). With the synuclein antibody, the nearest neighbor distances significantly increased, ranging from ∼50 nm close to the active zone to ∼200 nm farther from the active zone (Figure 5H, blue and red) (Untreated: slope = 0.0447, R² = 0.0730, n = 2986 SVs, 30 synapses, 2 axons; Synuclein Ab Low: slope = 0.0530, R² = 0.2440, n = 2418 SVs, 31 synapses, 2 axons; Synuclein Ab High: slope = 0.1645, R² = 0.0805, n = 676 SVs, 23 synapses, 2 axons; linear regression; multiple comparisons p = 0.002). Thus, synuclein perturbation resulted in a de-clustering and dispersion of synaptic vesicles away from the active zone. In comparison, synapses treated with the higher concentrations of the Control, SV2, and VAMP antibodies remained on average ∼50–70 nm from their nearest neighbor, reflective of tight SV clustering (Figure 5I; Control IgG Ab: slope = 0.0222, n = 2928 SVs, 29 synapses, 3 axons; SV2 Ab: slope = 0.0289, n = 4079 SVs, 24 synapses, 3 axons; VAMP Ab: slope = 0.0122, n = 3032 SVs, 25 synapses, 2 axons; linear regression, multiple comparisons p = 0.028). Thus, synuclein regulates SV clustering at resting synapses, and disruption of synuclein function disperses SVs into both smaller discrete clusters and individual SVs.