α-Synuclein-112 Impairs Synaptic Vesicle Recycling Consistent With Its Enhanced Membrane Binding Properties

Synucleinopathies are neurological disorders associated with α-synuclein overexpression and aggregation. While it is well-established that overexpression of wild type α-synuclein (α-syn-140) leads to cellular toxicity and neurodegeneration, much less is known about other naturally occurring α-synuclein splice isoforms. In this study we provide the first detailed examination of the synaptic effects caused by one of these splice isoforms, α-synuclein-112 (α-syn-112). α-Syn-112 is produced by an in-frame excision of exon 5, resulting in deletion of amino acids 103–130 in the C-terminal region. α-Syn-112 is upregulated in the substantia nigra, frontal cortex, and cerebellum of parkinsonian brains and higher expression levels are correlated with susceptibility to Parkinson’s disease (PD), dementia with Lewy bodies (DLB), and multiple systems atrophy (MSA). We report here that α-syn-112 binds strongly to anionic phospholipids when presented in highly curved liposomes, similar to α-syn-140. However, α-syn-112 bound significantly stronger to all phospholipids tested, including the phosphoinositides. α-Syn-112 also dimerized and trimerized on isolated synaptic membranes, while α-syn-140 remained largely monomeric. When introduced acutely to lamprey synapses, α-syn-112 robustly inhibited synaptic vesicle recycling. Interestingly, α-syn-112 produced effects on the plasma membrane and clathrin-mediated synaptic vesicle endocytosis that were phenotypically intermediate between those caused by monomeric and dimeric α-syn-140. These findings indicate that α-syn-112 exhibits enhanced phospholipid binding and oligomerization in vitro and consequently interferes with synaptic vesicle recycling in vivo in ways that are consistent with its biochemical properties. This study provides additional evidence suggesting that impaired vesicle endocytosis is a cellular target of excess α-synuclein and advances our understanding of potential mechanisms underlying disease pathogenesis in the synucleinopathies.


INTRODUCTION
Synucleinopathies are a class of neurological disorders linked to overexpression and aggregation of α-synuclein, and they include Parkinson's disease (PD), Dementia with Lewy Bodies (DLB), and Multiple Systems Atrophy (MSA). In these diseases, α-synuclein aggregates throughout neurons, including axons and synapses, leading to cellular toxicity and neurodegeneration (Kramer and Schulz-Schaeffer, 2007;Schulz-Schaeffer, 2010;Scott et al., 2010;Burre et al., 2018;Sulzer and Edwards, 2019). Multiplication (duplication and triplication) of the α-synuclein gene (SNCA) and a number of point mutations in exons 2 and 3 lead to aberrant α-synuclein aggregation and are genetically linked to familial PD (Kruger et al., 1998;Singleton et al., 2003;Nussbaum, 2018). In addition, differential expression of several α-synuclein splice variants is observed in PD, DLB, and MSA (Beyer et al., 2004(Beyer et al., , 2008McLean et al., 2012;Cardo et al., 2014). Thus, it is increasingly important to understand how different α-synuclein variants impact neuronal function, as well as disease pathogenesis and progression.
In this study, we focus on the splice isoform α-syn-112, a 12-kDa protein comprising a deletion of 28 amino acids (a.a. 103-130) near the C-terminus (Ueda et al., 1994). α-Syn-112 is normally expressed in low levels in many human tissues, including skin, lung, kidney, and heart, with highest expression in the brain (Beyer et al., 2008). However, in parkinsonian, DLB and MSA brains, α-syn-112 is overexpressed in the substantia nigra, frontal cortex, and cerebellum (Beyer et al., 2004;Brudek et al., 2016). In addition, increased α-syn-112 levels are associated with PD risk (McCarthy et al., 2011). Compared to α-syn-140, α-syn-112 exhibits enhanced aggregation and fibrillation in vitro (Manda et al., 2014). While it is clear that excess α-syn-112 is associated with a number of neurodegenerative diseases, very little is known about its biochemical properties or neuronal functions.
We therefore set out to perform a more detailed characterization of α-syn-112, focusing on its possible roles at synapses. Under physiological conditions, α-syn-140 is expressed at the presynapse where it regulates synaptic vesicle clustering and trafficking (Bendor et al., 2013;Vargas et al., 2014;Logan et al., 2017;Atias et al., 2019). When overexpressed at mammalian synapses to levels comparable to those in familial PD, α-syn-140 impaired synaptic vesicle trafficking (Nemani et al., 2010;Scott et al., 2010), and altered the composition of presynaptic proteins (Scott et al., 2010). In line with these findings, we previously reported that acute introduction of α-syn-140 at a classical vertebrate synapse, the lamprey reticulospinal (RS) synapse, impaired synaptic vesicle recycling mediated by clathrin-mediated endocytosis and possibly bulk endocytosis (Busch et al., 2014;Medeiros et al., 2017;Banks et al., 2020). Similarly, acute introduction of α-syn-140 at mammalian synapses also impaired vesicle endocytosis with no observable effects on exocytosis (Xu et al., 2016;Eguchi et al., 2017). The synaptic deficits induced by α-syn-140 require proper membrane binding because point mutants with reduced lipid binding capacity exhibited greatly reduced effects on SV trafficking (Nemani et al., 2010;Busch et al., 2014). In comparison, there are no studies to date that have investigated how any of the related αsynuclein splice isoforms affect presynaptic functions, prompting this work.
α-Syn-140 or α-syn-112 (5 mg) was incubated for 2 h at RT with liposomes (∼34-36 µL), in HKE buffer (25 mM HEPES, pH 7.4, 150 mM KCl, 1 mM EDTA) in a 100 µL total volume. Samples were then added to the bottom of an Accudenz gradient (40%, 35%, 30%, 0%, in 800 µL total volume). Columns were centrifuged at 280,000 × g for 3 h at RT. After ultracentrifugation, columns were separated into eight fractions (100 µL each). The presence of liposomes was determined in each fraction by quantifying NBD fluorescence using a NanoDrop 3300 fluorospectrometer (Thermo Fisher Scientific). In parallel, the α-synuclein distribution in each fraction was determined by Western blotting. Quantification of α-synuclein band intensities was performed using ImageJ. Lipid bound protein (%) was calculated as the amount of protein in the first 3 fractions divided by the total protein in all 8 fractions. Data shown are representative of n = 3-5 independent experiments. GraphPad Prism 8 (GraphPad Software, Inc., La Jolla, CA, United States) was used to perform statistical analyses and generate graphs.

Membrane Recruitment Assays
Membrane recruitment assays were performed as described (Shetty et al., 2013). First, crude synaptosomes were prepared from two mouse brains, resuspended in 6 mL homogenization buffer (25 mM Tris-HCl; pH 8.0, 500 mM KCl, 250 mM sucrose, 2 mM EGTA), then added to the top of freshly prepared 0.65, 0.85, 1.00, 1.20M sucrose gradients and centrifuged at 100,000 × g for 2 h. Synaptosomes were collected from the 1/1.2M interface and resuspended in 20 mL buffer. Pure synaptosomes were centrifuged at 100,000 × g for 20 min. The pellet was resuspended in 4 mL ice-cold deionized water, and 250 mM HEPES-NaOH, pH 7.4 was added to a final concentration of 7.5 mM. The suspension was incubated on ice for 30 min and centrifuged at 100,000 × g for 20 min. The pellet was resuspended in 4 mL of 0.1M Na 2 CO 3 to strip peripheral proteins, incubated for 15 min at 37 • C, and centrifuged at 100,000 × g for 20 min. Pellet was resuspended in 2 mL cytosolic buffer (25 mM HEPES-NaOH, pH 7.4, 120 mM potassium glutamate, 2.5 mM magnesium acetate, 20 mM KCl, and 5 mM EGTA-NaOH, pH 8.0, filtered and stored at 4 • C), centrifuged again at 100,000 × g for 20 min, and resuspended in 2 mL of cytosolic buffer. Proteins were quantified using BCA. Mini cOmplete TM protease inhibitors (Roche) were added, and aliquots of purified membranes were flash frozen and stored at −80 • C until use.
Next, cytosol preparations were made from two mouse brains. To do that, brains were first washed and then homogenized with 2 mL of homogenization buffer (25 mM Tris-HCl, pH 8.0, 500 mM KCl, 250 mM sucrose, 2 mM EGTA, and 1 mM DTT) using 10 strokes at 5,000 rpm. The homogenate was transferred to a 3.5 mL ultracentrifuge tube and centrifuged at 160,000 × g for 2 h at 4 • C. The supernatant was exchanged into 3.5 mL cytosolic buffer. After measuring protein concentration and adding protease inhibitors, 100 µL aliquots were flash frozen and stored at −80 • C until use.
For the membrane recruitment assays, synaptic membranes (200 µg) were mixed with 250 µg brain cytosol proteins in 500 µl cytosolic buffer and supplemented with different concentrations of recombinant human α-syn-140 and α-syn-112 (rPeptide). A control experiment was prepared with only synaptic membranes and cytosolic buffer. Mixtures were incubated at 37 • C for 15 min. The samples were immediately centrifuged at 100,000 × g for 30 min at 4 • C. Pellets, now containing the synaptic membranes with bound proteins, were resuspended in 500 µL of cytosolic buffer at 4 • C. The resuspension was centrifuged at 100,000 × g for 30 min at 4 • C and resuspended in 90 µL of cytosolic buffer. For each sample, 20 µL aliquots were mixed with 5x loading buffer, run on 12% reducing SDS-PAGE gels, transferred to nitrocellulose membranes, and processed via Western blotting. Levels of α-synuclein recruited to synaptic membranes were detected for each condition by Western blot with a rabbit polyclonal pan-synuclein antibody (1:1000; Abcam ab53726; Cambridge, MA, United States) and quantified using ImageJ.

Microinjections and Stimulation
Recombinant human α-syn-140 and α-syn-112 were obtained from rPeptide, Inc. The recombinant α-syn-140 dimer (NC dimer) used in this study was a single polypeptide comprising two full-length copies of α-syn-140, as previously described (Pivato et al., 2012;Medeiros et al., 2017). All animal procedures were conducted in accordance with standards set by the National Institutes of Health and approved by the Institutional Animal Care and Use Committee at the Marine Biological Laboratory.

Electron Microscopy and Imaging
After fixation (overnight to 2 days), spinal cords were processed in 2% osmium, stained en bloc with 2% uranyl acetate, and embedded in Embed 812 resin, as previously described (Busch et al., 2014;Medeiros et al., 2017;Walsh et al., 2018). Ultrathin sections (70 nm) were counterstained with 2% uranyl acetate followed by 0.4% lead citrate. A JEOL JEM 200CX transmission electron microscope was used to acquire images of individual synapses at 37,000x magnification. For each experimental condition, images were acquired from at least n = 10-20 synapses collected from n = 2 axons/animals at distances of 25-150 µm from the injection site, which is where the protein had diffused based on the co-injected fluorescent dye. Images of control synapses were acquired from the same axons, but at greater distances from the injection site (>350 µm) in regions where the injected proteins had not diffused, providing an internal control for each experiment.
A morphometric analysis was performed on all synaptic membranes within 1 µm of the active zone, as previously described (Busch et al., 2014;Medeiros et al., 2017;Walsh et al., 2018;Banks et al., 2020). Image analysis was performed in FIJI 2.0.0. by a researcher blinded to experimental conditions. Measurements included the number of synaptic vesicles per synapse (per section), size of plasma membrane (PM) evaginations, number and size of large (>100 nm) irregularly shaped intracellular membranous structures ("cisternae"), and number and stage of clathrin coated pits (CCPs) and clathrin coated vesicles (CCVs). The sizes of PM evaginations were measured by first drawing a straight line (1 µm) laterally from the edge of the active zone to the nearest point on the axolemmal surface and then measuring the curved distance between these two points. Additionally, we also quantified the depth of the PM evaginations from the axolemmal surface to the deepest point within the evagination. CCP/V stages were defined as: stage 1initial clathrin coated bud; stage 2 -invaginated CCP without constricted neck; stage 3 -invaginated CCP with constricted neck; stage 4 -free CCV. GraphPad Prism 8 was used to generate graphs and for all statistical analyses.
Reconstruct software (Fiala, 2005) was used to generate a three-dimensional reconstruction of single synapses from four or five serial images. Fiduciary markers were used to align the serial images. Synaptic structures were added using trace slabs for PM and cisternae, spheres for synaptic vesicles (50 nm) and clathrincoated pits and vesicles (90 nm), and a Boissonnat surface for the active zone.
Giant RS synapses are en passant glutamatergic synapses that reside along the perimeter of the giant RS axons (Wickelgren et al., 1985;Brodin and Shupliakov, 2006). Stimulated control synapses exhibit a large and localized synaptic vesicle cluster, shallow plasma membrane (PM) evaginations, few clathrincoated pits (CCPs) and clathrin-coated vesicles (CCVs), and only a few cisternae, which are defined as large vesicular structures with a diameter > 100 nm ( Figure 6A). While we do not yet know the precise identities of cisternae, their morphologies are consistent with bulk and/or recycling endosomes (Morgan et al., 2013;Chanaday et al., 2019). By comparison, synapses treated with recombinant human αsyn-112 exhibited a drastic change in morphology, indicated by a loss of the synaptic vesicle cluster, large extended PM evaginations, and accumulation of cisternae and clathrin-coated pits and vesicles ( Figure 6B). Three-dimensional reconstructions show clearly the morphological alterations caused by α-syn-112, especially the loss of vesicles (blue) and buildup of PM (green) and cisternae (magenta) (Figures 6C,D). In addition, there were obvious changes in the number of CCPs and CCVs. Whereas stimulated control synapses have only a few CCPs, those treated with α-syn-112 have more CCPs and CCVs, suggesting deficits in vesicle fission and clathrin  (I) Quantification of the depth of PM evaginations reveals that α-syn-112 has an intermediary phenotype to monomeric and dimeric α-syn-140, consistent with its ability to dimerize on synaptic membranes. Bars indicate mean ± SEM from n = 21 to 26 synapses, 2 axons/animals. Asterisks indicate statistical significance by one-way ANOVA ** (p < 0.01); **** (p < 0.0001); n.s. = not significant.

DISCUSSION
This is the first study to investigate the lipid binding properties and synaptic effects of α-syn-112, which is both naturally occurring and overexpressed in multiple neurodegenerative diseases. We show here that α-syn-112 exhibits enhanced membrane binding in vitro compared to wild type α-syn-140 (Figures 2-4), including to synaptically relevant phosphoinositides such as PI(4)P and PI(4,5)P2 (Figure 3). α-Syn-112 also exhibits enhanced oligomerization (dimerization and trimerization) on synaptic membranes ( Figure 5) and impairs synaptic vesicle recycling when acutely introduced in excess (Figures 6, 7). In our previous studies, we showed that excess monomeric α-syn-140 impaired CCV uncoating at lamprey synapses (Medeiros et al., 2017;Banks et al., 2020), while dimeric α-syn-140 impaired an earlier stage of CCP fission (Medeiros et al., 2017(Medeiros et al., , 2018. Interestingly, although we injected recombinant α-syn-112 in the monomeric form (Figure 1C), the resulting synaptic phenotype was indicative of deficits in both CCP fission and CCV uncoating (Figure 6), which is consistent with its enhanced ability to dimerize on synaptic membranes (Figure 5). Further underscoring this result is that the depth of PM evaginations produced by α-syn-112 was also intermediate between monomeric and dimeric α-syn-140. We do not yet fully understand the oligomerization status of α-syn-112 once it enters the synaptic environment. However, this study nonetheless further emphasizes that different molecular species of α-synuclein can produce distinct effects at synapses (Medeiros et al., 2018), potentially compounding the cellular deficits if expressed in combination.
A key biochemical feature of α-syn-112 is its ability to bind phospholipid membranes with increased efficacy, as compared to wild type α-syn-140. In every example tested, α-syn-112 exhibited enhanced binding in vitro to anionic phospholipids, including many of the phosphoinositides and total brain lipids (Figures 2-4). The predicted structure for α-syn-112 involves a deletion of 28 amino acids (a.a. 103-130) in the C-terminal domain, which may result in an extended alpha helical region ( Figure 1B). Given that the membrane binding capacity of α-syn-140 is fairly evenly distributed throughout the alpha helical N-terminal domain (Davidson et al., 1998;Chandra et al., 2003;Burre et al., 2012), extending the alpha helix could result in the enhanced lipid binding that was observed. Additionally, we show that α-syn-112 also binds more strongly to a number of phosphoinositides, including PI, PI(3)P, PI(4)P, PI(4,5)P 2 , and PI(3,4,5)P 3 , though we did not detect any preferential selectivity amongst them (Figure 3). While interactions between α-syn-140 and PI(4,5)P 2 have been reported using giant unilamellar vesicles (Narayanan et al., 2005;Stockl et al., 2008), to our knowledge this is the first study that provides a more comprehensive and comparative assessment of αsynuclein binding to phosphoinositides. It is notable that such strong binding was observed when the phosphoinositide concentrations were only 5% of the total lipid composition (Figure 2), which is much less than the 30-50% anionic lipids normally used in these in vitro assays (Burre et al., 2012;Busch et al., 2014;Medeiros et al., 2017). Phosphoinositides are present in limiting amounts and tightly-controlled on cellular membranes (Di Paolo and De Camilli, 2006;Takamori et al., 2006;Balla, 2013;Schink et al., 2016), including on synaptic vesicles where they likely comprise < 10% of the total phospholipids (Takamori et al., 2006). Thus our results may be more reflective of what happens intracellularly and suggest that α-syn-112 binds to physiological synaptic membranes better than α-syn-140, which has implications for its potential toxicity. PI(4,5)P 2 is enriched on the PM and helps to recruit clathrin adaptor proteins to the membrane during initiation of clathrinmediated synaptic vesicle endocytosis (Ford et al., 2001;Di Paolo and De Camilli, 2006;Saheki and De Camilli, 2012). Thus, the strong binding of α-syn-140 and α-syn-112 to PI(4,5)P 2 may mask sites for clathrin coat initiation and inhibit early stages of vesicle endocytosis, which is consistent with the expanded PM evaginations observed after introducing either isoform to synapses (Figures 6, 7) (Busch and Morgan, 2012;Medeiros et al., 2017;Banks et al., 2020). Stronger binding to PI(4,5)P 2 may also explain in part why α-syn-112 has greater effects than α-syn-140 on the depth of PM evaginations (Figure 7). In addition, it is thought that PI(4,5)P 2 remains on the endocytic vesicle throughout CCP and CCV formation until it is dephosphorylated to PI(4)P by the uncoating protein, synaptojanin (Cremona et al., 1999;Saheki and De Camilli, 2012). Thus, strong binding of α-syn-140 and α-syn-112 to PI(4,5)P 2 and PI(4)P may also mask these lipids and alter the dynamics of the late stages of clathrin-mediated endocytosis and contribute to the fission and uncoating defects observed, along with mislocalization of the CCV uncoating protein (Hsc70), which we recently reported (Banks et al., 2020). Going forward, it will be important to advance our understanding of α-synuclein interactions with phosphoinositides, since misregulation of phosphoinositide levels and phosphoinositide-mediated membrane trafficking may contribute to neurodegenerative diseases (Fabelo et al., 2011;Nadiminti et al., 2018).
Another interesting finding is that α-syn112 has increased propensity for oligomerization on synaptic membranes (Figure 5). Like monomeric α-syn-140, dimeric α-syn-140 undergoes alpha helical folding in the presence of SDS micelles, binds strongly to PA-containing liposomes, and exhibits timedependent aggregation and fibrillation in vitro in biochemical assays (Pivato et al., 2012;Medeiros et al., 2017;Dong et al., 2018). α-Synuclein rapidly dimerizes and aggregates on membranes containing PS (Lv et al., 2019), which is one of the major anionic lipids comprising synaptic vesicles (Takamori et al., 2006). Under physiologic conditions, α-syn-140 multimers exist at synapses and participate in synaptic vesicle clustering, restricting vesicle motility during trafficking (Wang et al., 2014). When introduced in excess to synapses, dimeric α-syn-140 inhibited synaptic vesicle recycling and impaired CCP fission (Medeiros et al., 2017(Medeiros et al., , 2018. Because excess α-syn-112 also interfered with CCP fission (Figure 6), this suggests that the injected monomeric α-syn-112 protein dimerized upon interaction with synaptic membranes in vivo, consistent with its in vitro effects ( Figure 5). In future experiments, it will be interesting to determine the impacts of dimeric α-syn-112 on synaptic vesicle trafficking. Since oligomerization on membranes is associated with membrane penetration and toxicity (Tsigelny et al., 2012(Tsigelny et al., , 2015, formation of α-syn-112 or α-syn-140 dimers may be an important rate-limiting step in the early pathogenesis of the synucleinopathies.
In summary, like α-syn-140, α-syn-112 avidly binds phospholipid membranes and, when in excess, impairs synaptic vesicle recycling producing distinct effects on clathrinmediated endocytosis. Despite these similarities, α-syn-112's enhanced membrane binding properties and propensity for oligomerization may underlie the greater effects on synaptic membranes. In addition to providing the first insight into the synaptic toxicity caused by α-syn-112, this study further emphasizes the need for investigating the impacts of different α-synuclein isoforms and conformations on neuronal function, since doing so may help us better understand the cellular pathways leading to neurodegeneration.

DATA AVAILABILITY STATEMENT
The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

ETHICS STATEMENT
The animal study was reviewed and approved by the Institutional Animal Care and Use Committee at the Marine Biological Laboratory, Woods Hole, MA, United States in accordance with guidelines set by the National Institutes of Health.

AUTHOR CONTRIBUTIONS
LS, JE, KV, AM, and JM contributed to the conception and design of the study, as well as data acquisition, analysis and interpretation. KH contributed to data acquisition. All authors (LS, JE, KV, AM, KH, and JM) were involved in drafting the manuscript, have provided final approval of this manuscript for submission, and agreed to be accountable for all aspects of the work.

FUNDING
This study was supported by a research grant from the National Institutes of Health (NIH NINDS/NIA R01 NS078165 to JM), as well as research funds from the Marine Biological Laboratory (to JM).