Substantial evidence links the small presynaptic protein α-syn to neurodegenerative diseases, collectively called synucleinopathies. The normal function of α-syn has been investigated for over a decade, and a prevailing view is that α-syn is a physiologic attenuator of neurotransmitter release. Modest overexpression of α-syn dampens synaptic responses (Nemani et al., 2010; Scott et al., 2010; Wang et al., 2014; Sun et al., 2019; Atias et al., 2019), and analogously, eliminating α-syn leads to phenotypes consistent with augmented synaptic release (Abeliovich et al., 2000; Yavich et al., 2004; Yavich et al., 2006; Senior et al., 2008; Greten-Harrison et al., 2010; Anwar et al., 2011), although the latter has not been seen in all studies (Burré et al., 2010). At a cellular level, synaptic attenuation is likely mediated by effects of α-syn on vesicle organization and trafficking, which are even seen in minimal in vitro systems, where recombinant α-syn clusters small synaptic-like vesicles (Diao et al., 2013; Sun et al., 2019). An emerging model is that α-syn plays a role in the organization and mobilization of SVs, that in turn regulates SV-recycling and neurotransmitter release; however, underlying mechanisms are unknown.

Work over several decades has shown that temporal and spatial regulation of the SV cycle is achieved by the cooperative effort of diverse groups of proteins, such as Muncs/SNAREs – orchestrating SV docking, priming, fusion – and sequential assembly of a variety of endocytosis-related proteins that build a platform for efficient membrane retrieval. Reasoning that an understanding of functional α-syn partners would offer meaningful insight into α-syn function, we have been combining SV-recycling assays with structure-function approaches to identify the protein-network in which α-syn operates at the synapse. Using this approach, we recently found that the physiologic effects of α-syn at the synapse requires synapsins (Atias et al., 2019). While modest over-expression of α-syn in wild-type (WT) cultured hippocampal neurons attenuated SV recycling, there was no effect in neurons lacking all synapsins, indicating that synapsins were necessary to enable α-syn functionality. Reintroduction of the canonical synapsin isoform (synapsin Ia) reinstated α-syn mediated attenuation, confirming functional cooperation between α-syn and synapsins (Atias et al., 2019).

Synapsins are a family of cytosolic proteins with known roles in maintaining physiologic SV clusters (Cesca et al., 2010; Orenbuch et al., 2012; Zhang and Augustine, 2021), and recent work supports a model where SVs are confined within synapsin-based protein condensates (Milovanovic et al., 2018). Alternative splicing of three synapsin genes gives five major isoforms. Both synapsins and synucleins are peripherally associated with SVs via the N-terminus, while C-terminal regions are more variable and structurally disordered (Song and Augustine, 2023). Depending on the isoform, the C-terminus of synapsin has two to three structurally distinct domains, and substantial evidence indicates that this domain-variability leads to isoform-specific functions (Song and Augustine, 2015; Song and Augustine, 2023). Reasoning that identifying the specific synapsin domain/isoform that bound to α-syn and facilitated α-syn function would offer mechanistic insight into α-syn biology, we systematically evaluated effects of each synapsin isoform in enabling the physiologic effects of α-syn. For these experiments, we used pHluorin assays that report exo/endocytic SV recycling. pHluorin is a pH-sensitive GFP that acts as a sensor for pH changes, and in our experiments, the probe is tagged to the transmembrane presynaptic protein synaptophysin, and targeted to the interior of SVs [called ‘sypHy’, see Royle et al., 2008]. In resting SVs, sypHy is quenched, as the pH is acidic (~5.5). However, upon stimulation, SVs fuse with the presynaptic plasma membrane, resulting in pH-neutralization and a concomitant rise in fluorescence, which is subsequently quenched as the vesicles are endocytosed and reacidified (Figure 1A). Fluorescence fluctuations in this assay are a measure of SV exo/endocytosis, and at the end of the experiment, all vesicles can be visualized by adding NH₄Cl to the bath (alkalinization). As reported previously, overexpression of h-α-syn attenuated SV recycling in WT hippocampal cultured neurons, but there was no effect in neurons from mice lacking all synapsins – synapsin triple knockout or TKO mice (Figure 1B–C).

The synapsin family has five main isoforms, Ia Ib, IIa, IIb, and IIIa (Figure 1D). To determine synapsin isoforms that enable α-syn functionality, we overexpressed h-α-syn in cultured neurons from synapsin TKO mice and systematically reintroduced each synapsin isoform, with the goal of identifying synapsin isoforms that reinstated α-syn-induced synaptic attenuation (see plan in Figure 1E). Reintroduction of Synapsin Ia (containing domains A-E) in this setting reinstated α-syn functionality (Figure 1F, left-panel; these changes are due to altered exocytosis, see Figure 1—figure supplement 1A–B). However, interestingly, only synapsins Ia, IIa, and IIIa enabled h-α-syn-mediated synaptic attenuation, whereas synapsins Ib and IIb had no effect (Figure 1F, quantified in Figure 1G). These effects were likely due to exocytosis as noted above, as quantification of the fluorescence decay-kinetics – which is a measure of endocytosis – did not reveal any changes (Figure 1—figure supplement 1C–E). One prediction of the pHluorin experiments is that the synapsin isoforms that allow for α-syn functionality would also be the ones that bind to α-syn. To test this, we performed co-immunoprecipitation experiments in neuronal cell lines, where we co-transfected neuro-2a cells with myc-tagged h-α-syn and each synapsin isoform (fluorescent-tagged), immunoprecipitated the synapsin isoform, and determined amounts of co-immunoprecipitated α-syn by western blotting (schematic in Figure 2A). While synapsins Ia, IIa, and IIIa bound robustly to h-α-syn, binding of synapsins Ib and IIb was much lower (Figure 2B, quantified in Figure 2C). Taken together, these experiments indicate that only three synapsin isoforms (Ia, IIa, and IIIa) can robustly bind to α-syn and reinstate functional effects of α-syn in this setting. Since the E-domain, within the variable C-terminus, is common to these three synapsin isoforms – and absent in the others (see domain-structure in Figure 2D) – we reasoned that the E-domain was the bona fide α-syn binding-site, and also responsible for facilitating α-syn functionality.

In parallel experiments, we also narrowed down the reciprocal region in α-syn bound to synapsin. Toward this, we designed GST-pulldown assays to test the interaction of various h-α-syn sequences with mouse brain synapsins. In these experiments, beads with GST-tagged h-α-syn (WT, deletions, and scrambled variants) were incubated with mouse brain lysates, and brain synapsins binding to α-syn were evaluated by western blotting (Figure 2E). Figure 2F shows how the scrambled variants were designed. While synapsins bound to GST-tagged WT-h-α-syn, deletion of the C-terminus (α-syn 96–140) eliminated this interaction (Figure 2G, lanes 1–3). Regions within amino acids 96–110 of α-syn were critical in binding synapsin, as this minimal region bound to synapsin (Figure 2G, lanes 4–5), and scrambling the amino acids within this region – while keeping the other sequences intact – eliminated this interaction (Figure 2G, lanes 6–7). Data from all western blots is quantified in Figure 2G – bottom. Together, these experiments identify amino-acids 96–110 of α-syn as the region binding to synapsin.

To test if the E-domain was necessary for enabling α-syn functionality, we generated a synapsin-Ia construct where the amino acid sequences of the E-domain were scrambled (Figure 3A, synapsin-Ia^ScrE). As shown previously, expression of WT synapsin-Ia enables α-syn-mediated synaptic attenuation in neurons lacking all synapsins (Figure 1F, leftmost panel). We reasoned that if the E-domain enabled α-syn functions and mediated synapsin/α-syn interactions in these experiments, scrambling this region should abolish such synapsin-dependent functions. Towards this, we used pHluorin assays in synapsin TKO neurons, asking if synapsin-Ia^ScrE would fail to reinstate α-syn functionality (schematic in Figure 3B). Indeed, while overexpressed h-α-syn was able to attenuate synaptic responses in the presence of WT-Synapsin-Ia in synapsin TKO neurons, Synapsin-Ia^ScrE failed to have any effect (Figure 3C), despite the detection of similar quantities of both at synapses (Figure 3—figure supplement 1). Analogously, in neuro2a co-immunoprecipitation experiments to test binding of WT-Synapsin-Ia or Synapsin-Ia^ScrE to α-syn, WT h-α-syn bound to Synapsin-Ia, but not to Synapsin-Ia^ScrE (Figure 3D), indicating that the E-domain is critical in mediating this interaction.

Next, we tested if the synapsin-E domain was sufficient for enabling α-syn functionality. Toward this, we first over-expressed the E-domain in synapsin TKO neurons, along with h-α-syn and sypHy (Figure 3—figure supplement 2A), with the overall intention of evaluating SV-recycling in this setting. However, we found that the E-domain by itself was not targeted to synapses (Figure 3—figure supplement 2B) – consistent with the known biology of synapsins (Gitler et al., 2004) – and expectedly, the E-domain had no effect on SV-recycling in pHluorin assays (Figure 3—figure supplement 2C). To allow the E-domain to operate in a context where it would be ‘functionally available’, we fused the synapsin E-domain to the C-terminus of sypHy. Since in this scenario, the small synapsin fragment would be localized to the cytosolic surface of SVs and target to synapses (Figure 3E and Figure 3—figure supplement 2D), we reasoned that such placing of the E-domain in the right cellular context may be sufficient to enable α-syn functionality. Indeed, forced targeting of the synapsin E-domain to the surface of SVs enhanced α-syn enrichment in synapses (Figure 3—figure supplement 2E), and restored α-syn mediated synaptic attenuation in synapsin null neurons (Figure 3F), suggesting that the E-domain was sufficient to reinstate the functional interplay between α-syn and synapsins. Collectively, the evidence makes a strong case that the synapsin E-domain is both necessary and sufficient to allow α-syn functionality at synapses.

Previous studies have shown that loss of all synapsins disrupt the tight clustering of SVs that is normally seen in cultured hippocampal neurons, leading to a reduced number of SVs within the bouton-boundary and an increase in vesicles spilling out into the adjacent axon [(Orenbuch et al., 2012), and see Figure 4A]. One possibility in our α-syn over-expression experiments is that excessive α-syn can bind to endogenous synapsin molecules (presumably via the E-domain) and prevent the normal functionality of synapsins (i.e. ability to cluster SVs). Dispersion of SVs can be quantified using ‘full-width half-max’ (FWHM) analysis, which is a quantitative measure of the extent of protein-dispersion at synapses (Orenbuch et al., 2012; Wang et al., 2014). Briefly, combined attenuation and dispersion of synaptic proteins would cause an increase in FWHM (see Figure 4B). As shown in Figure 4C, loss of synapsins lead to an overall reduction in the intensity of SV-staining at boutons (Figure 4C, left), as well as increased FWHM (Figure 4C, right). To examine SV dispersion in a α-syn over-expression setting, we cultured neurons from WT or synapsin TKO mice, and transduced either h-α-syn alone (in neurons from WT mice), or h-α-syn, along with various synapsin isoforms (in neurons from synapsin TKO mice – see strategy in Figure 4D). As shown in Figure 4E, over-expression of h-α-syn led to an attenuation/dispersion of SV-intensities (increased FWHM) in WT neurons, but had no effect in synapsin TKO neurons. Over-expression of Ia/IIa synapsin isoforms (but not Ib/IIb isoforms) also led to SV dispersion (IIIa was not tested). At first glance these data seem to contradict studies from many groups showing that α-syn clusters SVs (Diao et al., 2013; Wang et al., 2014; Sun et al., 2019), but we surmise that the AAV-mediated over-expression of α-syn in this setting creates a scenario where excessive α-syn binds to and displaces native synapsin molecules from SVs, or may disrupt synapsin-based protein condensates (Hoffmann et al., 2021; Song and Augustine, 2023; Hoffmann et al., 2023).

Precise organization of vesicles at synapses is critical for synaptic function (Denker and Rizzoli, 2010). Typically, each synapse has clusters of dozens to hundreds of SVs, and these vesicles are classified into different pools based on their ability to participate in exocytosis, and their physical proximity to the site of exocytosis (active zone). SVs within the readily releasable pool are docked at the active zone and can rapidly fuse with the plasma membrane in response to an action potential. On the other hand, SVs within the much larger reserve pool are distal to the active zone and are thought to help replenish SVs following exocytosis. Actively recycling vesicles comprise the recycling pool. Studies over several decades have shown that functional perturbation of synapsins selectively reduces the number of SVs in the reserve pool, establishing a role for synapsin in maintaining SV clusters within this pool [reviewed in Zhang and Augustine, 2021]. Previous studies have also explored the role of the E-domain in various model systems. Microinjecting domain-E antibodies into lamprey giant axons dispersed the distal cluster of SVs (Pieribone et al., 1995), suggesting that this domain has a role in organizing the reserve pool. Injection of a peptide from the E-domain into squid giant synapses also dispersed the distal SV cluster, while docked SVs remained intact (Hilfiker et al., 1998), indicating that interfering with this domain in different ways resulted in the same phenotype – disruption of the reserve pool of SVs.

Thus, the current view is that the synapsin E-domain has an important role in maintaining the distal reserve pool SV clusters, although this domain has other independent roles in SV exocytosis that are not well defined (Song and Augustine, 2015). In this context, α-syn has also been long thought to play roles in SV organization and trafficking. First, the N-terminus of α-syn adopts a helical structure in the presence of small synaptic-like vesicles (Burré et al., 2018), and can also directly modulate vesicle shape (Varkey et al., 2010). In cell-free systems, recombinant α-syn can cluster synaptic-like vesicles (Diao et al., 2013; Sun et al., 2019), and experiments with cultured neurons also support the idea that α-syn can cluster SVs (Wang et al., 2014). For example, induced multimerization of α-syn at synapses clusters synaptic vesicles (Wang et al., 2014), and α-syn overexpression also diminished vesicle trafficking between synaptic boutons (Scott and Roy, 2012), which may reflect clustering of SVs by α-syn. Adjacent vesicles may also be directly tethered by α-syn (Fusco et al., 2016; Lautenschläger et al., 2018), thus α-syn-dependent organization and corralling of SVs are important clues to its function. Interestingly, a recent study showed that injection of an antibody to the N-terminus of α-syn into lamprey giant axons also led to a loss of SVs (Fouke et al., 2021) – resembling the SV disruption caused by synapsin E-domain injections (Pieribone et al., 1995; Hilfiker et al., 1998) – although both reserve and readily-releasable pools were depleted with α-syn injections. Our results support co-regulation of SV organization by both α-syn and synapsin, involving the synapsin E-domain. Additionally, α-syn has also been implicated in promoting SNARE complex formation (Burré et al., 2010), facilitating endocytosis (Vargas et al., 2014), and may participate in fusion-pore opening (Logan et al., 2017). Further work is needed to clarify whether these different effects of α-syn are linked, or whether they reflect functionality in distinct neuronal states (for instance in resting versus active neurons). In summary, our studies open the door to further mechanistic investigations into the functional interacting partners of α-syn, which will be important to uncover the myriad functions of this enigmatic protein. More broadly, our structure-function experiments place α-syn in a functional context with its interacting partners at the synapse, offering new insight into α-syn biology.