α-Synuclein Mutation Inhibits Endocytosis at Mammalian Central Nerve Terminals

Abstract

α-Synuclein (α-syn) missense and multiplication mutations have been suggested to cause neurodegenerative diseases, including Parkinson's disease (PD) and dementia with Lewy bodies. Before causing the progressive neuronal loss, α-syn mutations impair exocytosis, which may contribute to eventual neurodegeneration. To understand how α-syn mutations impair exocytosis, we developed a mouse model that selectively expressed PD-related human α-syn A53T (h-α-syn_A53T) mutation at the calyx of Held terminals, where release mechanisms can be dissected with a patch-clamping technique. With capacitance measurement of endocytosis, we reported that h-α-syn_A53T, either expressed transgenically or dialyzed in the short term in calyces, inhibited two of the most common forms of endocytosis, the slow and rapid vesicle endocytosis at mammalian central synapses. The expression of h-α-syn_A53T in calyces also inhibited vesicle replenishment to the readily releasable pool. These findings may help to understand how α-syn mutations impair neurotransmission before neurodegeneration.

SIGNIFICANCE STATEMENT α-Synuclein (α-syn) missense or multiplication mutations may cause neurodegenerative diseases, such as Parkinson's disease and dementia with Lewy bodies. The initial impact of α-syn mutations before neuronal loss is impairment of exocytosis, which may contribute to eventual neurodegeneration. The mechanism underlying impairment of exocytosis is poorly understood. Here we report that an α-syn mutant, the human α-syn A53T, inhibited two of the most commonly observed forms of endocytosis, slow and rapid endocytosis, at a mammalian central synapse. We also found that α-syn A53T inhibited vesicle replenishment to the readily releasable pool. These results may contribute to accounting for the widely observed early synaptic impairment caused by α-syn mutations in the progression toward neurodegeneration.

Introduction

Mutations of α-synuclein (α-syn), such as A53T, A30P, E46K, H50Q, G51D, and A53E missense mutations, and multiplication mutations have been suggested to cause neurodegenerative diseases, such as Parkinson's disease (PD) and dementia with Lewy bodies (Petrucci et al., 2016). Before causing the progressive neuronal loss, α-syn mutations suppress the release of neurotransmitters and hormones at synapses and neuroendocrine cells (Kurz et al., 2010; Nemani et al., 2010; Scott et al., 2010; Lundblad et al., 2012; Janezic et al., 2013). Impairment of exocytosis represents the earliest defects in the progression toward PD (Nemani et al., 2010; Bendor et al., 2013), which may contribute to eventual neurodegeneration. However, how α-syn mutations impair exocytosis remains poorly understood.

At cultured hippocampal synapses, α-syn mutations reduce the recycling vesicle pool, which is attributed to impairment of vesicle reclustering after endocytosis (Nemani et al., 2010). However, direct evidence showing that vesicle reclustering controls exocytosis is still needed to prove this hypothesis. Although α-syn overexpression has been implicated to inhibit vesicle priming in PC12 and chromaffin cells (Larsen et al., 2006), whether this mechanism applies to neurons is unclear. Injection of mutant or wild-type α-syn into lamprey nerve terminals reduces the number of synaptic vesicles and expands terminal membrane, implying that elevated α-syn impairs vesicle endocytosis (Busch et al., 2014). However, direct evidence showing inhibition of the endocytosis time course is missing, and whether the results apply to mammalian central synapses is unclear.

The difficulty in studying presynaptic mechanisms is often due to limited tools in small nerve terminals. Here, we developed a mouse model with human α-syn A53T (h-α-syn_A53T) expressed specifically in the large calyx of Held nerve terminal that can be patch clamped (Borst and Soria van Hoeve, 2012). We found that h-α-syn_A53T overexpression or short-term elevation inhibited two of the most commonly observed forms of endocytosis, slow and rapid endocytosis. The expression of h-α-syn_A53T also inhibited vesicle replenishment to the readily releasable pool (RRP). These results may help to understand the mechanisms underlying the early impairment of transmitter release caused by α-syn mutations.

Materials and Methods

Animals.

Animal care and use were performed following procedures approved by the Animal Care and Use Committees of the National Institutes of Health and the Medical College of Georgia. Math5^Cre/Cre knock-in mice generated previously (Yang et al., 2003) were kept as homozygous. ROSA:LNL:tTA transgenic mice (Wang et al., 2008) were purchased from The Jackson Laboratory. tetO-A53T mice, which express h-α-syn_A53T when tetO is activated by tetracycline transactivator protein (tTA), were generated previously (Lin et al., 2012). We bred Math5^Cre/Cre knock-in mice with ROSA:LNL:tTA transgenic mice to generate ROSA:LNL:tTA/Math5^Cre mice, in which tTA is turned on specifically at calyces. We then bred ROSA:LNL:tTA/Math5^Cre mice with tetO-A53T mice to generate tetO-A53T/ROSA:LNL:tTA/Math5^Cre (h-α-syn_{A53T_calyx}) mice, in which tTA activates tetO to express h-α-syn_A53T at calyces. Mouse genotypes were determined by PCR with primers described previously (Yang et al., 2003; Wang et al., 2008; Lin et al., 2012).

Slice preparation and capacitance recordings.

Parasagittal brainstem slices (200 μm thick) containing the medial nucleus of the trapezoid body were prepared from 7- to 10-d-old male or female mice or Wistar rats using a vibratome (Wu et al., 2009). Calyces were randomly selected for whole-cell patch-clamp capacitance measurements using the EPC-9 amplifier and the software lock-in amplifier (Wu et al., 2009). The sinusoidal stimulus frequency was 1 kHz, and the peak-to-peak voltage was 60 mV.

We pharmacologically isolated presynaptic Ca²⁺ currents with a bath solution (22–24°C) containing the following (in mm): 105 NaCl, 20 TEA-Cl, 2.5 KCl, 1 MgCl₂, 2 CaCl₂, 25 NaHCO₃, 1.25 NaH₂PO₄, 25 glucose, 0.4 ascorbic acid, 3 myo-inositol, 2 sodium pyruvate, 0.001 tetrodotoxin, and 0.1 3,4-diaminopyridine, pH 7.4 when bubbled with 95% O₂ and 5% CO₂. A presynaptic pipette contained the following (in mm): 125 Cs-gluconate, 20 CsCl, 4 MgATP, 10 Na₂-phosphocreatine, 0.3 GTP, 10 HEPES, and 0.05 BAPTA, pH 7.2, adjusted with CsOH. h-α-Syn, h-α-syn_A53T, and h-α-syn_A30P were purchased from Sigma-Aldrich in powder form, and were dissolved in pipette solution.

For EPSC recordings, transverse brainstem slices were prepared. EPSCs were induced by an electrical stimulus (0.2 ms, 8–20 V) via a bipolar electrode positioned at the midline of the trapezoid body where presynaptic axon fibers pass through. Pipettes (2–3 MΩ) for EPSC recordings contained the following (in mm): 125 K-gluconate, 20 KCl, 4 MgATP, 10 Na₂-phosphocreatine, 0.3 GTP, 10 HEPES, and 0.5 EGTA, pH 7.2, adjusted with KOH. The series resistance (<15 MΩ) was compensated by 95%.

Immunohistochemistry and native gel electrophoresis.

Postnatal day 9 (P9) mice were anesthetized using Nembutal and transcardially perfused with 4% paraformaldehyde (Electron Microscopy Sciences). The brain was postfixed in 4% paraformaldehyde overnight and infiltrated with 30% sucrose for another 48 h. An OCT (Electron Microscopy Sciences)-embedded brain was sectioned using a cryostat (model CM3050S, Leica) at 30 μm thickness. Target proteins at calyces were identified using a rabbit antibody against GFP (1:400; Invitrogen), a guinea pig antibody against vGluT₁ (1:5000; Millipore), a rabbit antibody against ZsGreen (1:400; Clontech), and a mouse antibody against h-α-syn_A53T (syn211; 1:500; Santa Cruz Biotechnology) or against both h-α-syn_A53T and mouse α-syn (3H2897; 1:500; Santa Cruz Biotechnology). DyLight-488 donkey anti-mouse antibody, DyLight-488 donkey anti-rabbit antibody, and rhodamine-conjugated donkey anti-guinea pig antibody (1:200; Jackson ImmunoResearch Laboratories) were used as secondary antibodies. Images were collected by Zeiss LSM510 confocal microscopy (40×, 1.3 numerical aperture).

For the detection of α-syn on native gels, 4 μm α-syn recombinant proteins and non-denaturing NativeMark molecular mass markers were loaded on a 4–16% Bis-Tris native PAGE gel (Invitrogen) and run in the running buffer following the protocol of the manufacturer. After gel transferring, the membrane was fixed and then immunoblotted with the antibody to α-syn (C20; 1:1000; Santa Cruz Biotechnology). Bands were visualized using the Odyssey infrared imaging system (LI-COR Biosciences).

Data collection and analysis.

As in previous studies (Wu et al., 2009), recordings were made within 10 min after break in, and the capacitance jump (ΔC_m) was measured at 0.5 s after depolarization (depol) to avoid artifacts. The initial decay rate (Rate_decay) at calyces was measured within 4 s after depol_20ms, and 1.5 s after depol_20ms×10. For statistical analysis a t test was used. Data were presented as the mean ± SEM.

Results

To express h-α-syn_A53T, the first genetic mutation found in familial PD (Polymeropoulos et al., 1997), specifically in mouse calyces, we combined the tet-off expression system (Lin et al., 2012) with Cre-loxP strategy (Fig. 1). The Math5 Cre (Math5^Cre) knock-in mouse drives the expression of Cre from embryonic day 12.5 to adult in calyx-containing neurons in the ventral cochlear nucleus (Saul et al., 2008). However, it is unclear whether the Cre is expressed in most calyces that we randomly selected for patching. To quantify Cre expression in calyces, we bred Math5^Cre mice with the following two Cre reporter mouse lines: (1) mT/mGFP mice that expressed membrane-targeted tandem dimer Tomato (mT) before Cre-mediated excision and membrane-targeted GFP (mGFP) after excision (Muzumdar et al., 2007); and (2) ZsGreen mice that expressed cytosolic ZsGreen after Cre excision. The expression of Math5^Cre, labeled with an antibody against GFP or ZsGreen in Math5^Cre/+; mT/mGFP^+/− mice (Fig. 1A) or Math5^Cre/+; ZsGreen^+/− mice (data not shown), was restricted to central auditory brainstem. Coimmunostaining with vesicular glutamate transporter 1 (vGluT₁) that labeled calyces showed that mGFP overlapped with vGluT₁ in 86 ± 3% calyces (from three Math5^Cre/+; mT/mGFP^+/− mice), but not postsynaptic neurons (Fig. 1B). Similar overlap between ZsGreen and vGluT₁ was observed (Fig. 1C), indicating that Math5^Cre is expressed in most calyces.

Figure 1.

Identification of a Cre mouse line specific for calyces and overexpression of h-α-syn_A53T in calyces. A, Horizontal brainstem section showing limited expression of Math5^Cre in the auditory system. Math5^Cre expression was detected with antibody staining of GFP in Math5^Cre/+; mT/mGFP^+/− mice. VCN, Ventral cochlear nucleus, containing globular bushy cells giving rise to calyces; MNTB, medial nucleus of the trapezoid body, containing calyces. B, Immunostaining in MNTB: Math5^Cre expression (green: GFP antibody staining, left) overlaps with most vGluT₁ antibody staining (red, pseudocolor, middle) that labels calyces in Math5^Cre/+; mT/mGFP^+/− mice (left and middle panel superimposed in the right). C, Math5^Cre expression, detected from antibody staining of ZsGreen (green, left), overlaps with most vGluT₁ staining (red, labeling calyces, middle) in MNTB of Math5^Cre/+; ZsGreen^+/− mouse (left and middle panel superimposed in the right). D, Breeding scheme for the generation of a Math5^Cre/tetO-A53T/ROSA:LNL:tTA (h-α-syn_{A53T_calyx}) triple-transgenic mouse. E, h-α-syn_A53T (green, left) and vGluT₁ staining (red, labeling calyces, middle) in MNTB of P9 h-α-syn_{A53T_calyx} mouse (P9 A53T, top), P9 Math5^Cre mouse (P9 Ctrl, middle top), and P5 (middle bottom) and P18 (bottom) h-α-syn_{A53T_calyx} mouse (green and red images are also superimposed in the right). F, Left, Immunostaining intensity (arbitrary unit, AU; mean + SEM) of h-α-syn_A53T at calyces in P9 Math5^Cre (Ctrl, 210 calyces, 3 mice) and h-α-syn_{A53T_calyx} (A53T, 164 calyces, 3 mice) mice. Right, Immunostaining intensity of h-α-syn_A53T at calyces in P5 (74 calyces, 3 mice), P9 (164 calyces, 3 mice), and P18 (228 calyces, 3 mice) h-α-syn_{A53T_calyx} mice. G, Top, Sampled immunostaining with an antibody recognizing both mouse α-syn and h-α-syn_A53T (green, α-syn_m+h) and vGluT₁ staining (red) in P9 h-α-syn_{A53T_calyx} mouse (A53T, top) and Math5^Cre mouse (Ctrl, bottom). Bottom, α-syn_m+h immunostaining intensity (in AU) in P9 Math5^Cre mouse (Ctrl, 83 calyces, 3 mice) and h-α-syn_{A53T_calyx} mouse (A53T, 129 calyces, 3 mice). H, Sampled EPSCs (left) and the amplitude of EPSCs (right, mean ± SEM) induced by a presynaptic fiber stimulation in h-α-syn_{A53T_calyx} mice (A53T, 7 mice, 7 neurons) and Ctrl mice (including Math5^Cre and wild-type mice, 7 mice, 7 neurons). *p < 0.05.

Next, we crossbred Math5^Cre with ROSA:LNL:tTA transgenic mice (Wang et al., 2008) to generate Math5^Cre/ROSA:LNL:tTA bigenic mice, in which tTA was turned on specifically at calyces (Fig. 1D). We then crossbred Math5^Cre/ROSA:LNL:tTA mice with tetO-A53T mice (expressing h-α-syn_A53T when tetO was activated by tTA; Lin et al., 2012) to generate Math5^Cre/tetO-A53T/ROSA:LNL:tTA (h-α-syn_{A53T_calyx}) triple-transgenic mice, in which tTA activated tetO to express h-α-syn_A53T at calyces (Fig. 1D). Coimmunostaining of vGluT₁ and h-α-syn_A53T confirmed the presence of h-α-syn_A53T in 94% of vGluT₁-containing calyces (220 of 234 calyces, two mice) in h-α-syn_{A53T_calyx} mice, but not in control mice (three Math5^Cre mice; Fig. 1E,F). The h-α-syn_A53T expression level in calyces increased slightly from P5 to P18 (Fig. 1E,F), which is consistent with the finding that Math5 promoter is turned on as early as embryonic day 12.5, and remains active postnatally in young and adult calyx-containing neurons (Saul et al., 2008). Immunostaining with an antibody recognizing both mouse α-syn and h-α-syn_A53T showed that at P7–P10 the staining intensity at calyces in h-α-syn_{A53T_calyx} mice was ∼12 times (n = 129 calyces, 3 mice) that of Math5^Cre mice (n = 83 calyces, 3 mice; Fig. 1G), which is within the range (2.5–30-fold increase) of overexpression level in various α-syn transgenic mouse models (Fernagut and Chesselet, 2004).

An axonal fiber stimulus, which induced a presynaptic action potential (Xu and Wu, 2005), evoked a mean EPSC of 7.2 ± 0.5 nA in P7–P10 control mice (n = 7 calyces, 7 mice), but a reduced EPSC of 4.6 ± 0.9 nA in p7–10 h-α-syn_{A53T_calyx} mice (n = 7 calyces, 7 mice; p = 0.02, Fig. 1H). This result is consistent with a previous finding that α-syn mutations suppress transmitter release (Nemani et al., 2010; Lundblad et al., 2012; Janezic et al., 2013).

Next, we examined how h-α-syn_A53T overexpression affects presynaptic mechanisms at P7–P10 calyces, at which neurodegeneration is too early to develop (Petrucci et al., 2016). In control mice, including Math5^Cre and wild-type mice, a 20 ms depolarization from −80 to +10 mV (depol_20ms) induced a calcium current (ICa; 1.7 ± 0.1 nA) and exocytosis (ΔC_m; 379 ± 14 fF, n = 22 calyces, 8 mice; Fig. 2A). C_m decayed monoexponentially with a τ of ∼10–25 s and a Rate_decay of 26.8 ± 2.1 fF/s (n = 22 calyces, 8 mice; Fig. 2A), which reflect slow endocytosis (Wu et al., 2009). In h-α-syn_{A53T_calyx} mice, depol_20ms induced an ICa and a ΔC_m similar to control (Fig. 2B,C), but a significantly smaller Rate_decay (17 ± 2 fF/s, n = 26 calyces, 9 mice, p < 0.01; Fig. 2B,C), suggesting that h-α-syn_A53T inhibits slow endocytosis.

Figure 2.

Inhibition of slow endocytosis at calyces in h-α-syn_{A53T_calyx} mice. A, B, Sampled ICa (top) and capacitance changes (C_m, bottom) induced by depol_20ms (arrow) in control (Ctrl, A) and h-α-syn_{A53T_calyx} (A53T, B) mice. C, ICa, Rate_decay, and ΔC_m (mean ± SEM) induced by depol_20ms in Ctrl (22 calyces, 8 mice) and A53T mice (26 calyces, 9 mice). **p < 0.01.

To study rapid endocytosis, we applied 10 depol_20ms at 10 Hz (depol_20ms×10) to induce rapid endocytosis. In control mice, depol_20ms×10 induced a ΔC_m of 1160 ± 71 fF, which decayed biexponentially with a rapid and a slow τ of 1–3.5 and 10–30 s, respectively (31 calyces, 11 mice; Fig. 3A). The Rate_decay was 163 ± 16 fF/s (n = 31 calyces, 11 mice; Fig. 3A), which reflected mostly (>80%) the rapid component of endocytosis (Wu et al., 2005, 2009; Sun et al., 2010). In h-α-syn_{A53T_calyx} mice, depol_20ms×10 induced a ΔC_m (898 ± 68 fF, n = 30 calyces, 10 mice) and a Rate_decay (78 ± 9 fF/s, n = 30 calyces) that were both significantly smaller than control (p < 0.01; Fig. 3A–C), suggesting that h-α-syn_A53T inhibited rapid endocytosis. The decrease in ΔC_m after depol_20ms×10, but not after depol_20ms, reflected a reduced replenishment of the RRP, because depol_20ms depletes the RRP, whereas depol_20ms×10 causes repeated RRP depletion and replenishment (Wu et al., 2009).

Figure 3.

Inhibition of rapid endocytosis at calyces in h-α-syn_{A53T_calyx} mice. A, B, Sampled capacitance changes (C_m, bottom) induced by depol_20ms×10 (arrow) in control (A) and h-α-syn_{A53T_calyx} (B) mice. C, Rate_decay and ΔC_m (mean ± SEM) induced by depol_20ms×10 in Ctrl (31 calyces, 11 mice) and A53T mice (30 calyces, 10 mice). *p < 0.05; **p < 0.01.

Inhibition of rapid and slow endocytosis in h-α-syn_{A53T_calyx} mice was not due to ΔC_m reduction, because the decrease of Rate_decay after depol_20ms was not accompanied by a significant ΔC_m decrease (Fig. 2). The ΔC_m induced by depol_20ms×10 decreased by only ∼23%, which was insufficient to account for an ∼52% reduction in Rate_decay (Fig. 3). Furthermore, exocytosis reduction has not been shown to prolong endocytosis time course or to reduce Rate_decay (Wu et al., 2005, 2009, 2014; Renden and von Gersdorff, 2007).

In h-α-syn_{A53T_calyx} mice, the h-α-syn_A53T level was elevated for days. In the following, we determined whether short-term whole-cell dialysis of h-α-syn_A53T, wild-type h-α-syn, and h-α-syn_A30P inhibits endocytosis in P7–P10 rat calyces. The concentration we used was 4 μm, because endogenous α-syn is ∼2–5 μm (Westphal and Chandra, 2013). Native gel electrophoresis showed two major bands of ∼50 and ∼100 kDa for these three proteins dissolved in the pipette solution (Fig. 4A), suggesting that α-syn (monomer, ∼14 kDa) proteins are most likely tetramers and octamers, consistent with α-syn tetramers found in physiological conditions (Bartels et al., 2011).

Figure 4.

Inhibition of slow and rapid endocytosis by dialysis of h-α-syn_A53T and h-α-syn, but not of h-α-syn_A30P. A, Native gel electrophoresis fractionates h-α-syn (Syn, 4 μm), h-α-syn_A30P (A30P, 4 μm), and h-α-syn_A53T (A53T, 4 μm) dissolved in the pipette solution. Arrows point to two major bands stained by an anti-α-syn antibody (C20). B, Sampled C_m changes induced by depol_20ms (arrow) in calyces dialyzed with a control solution (Ctrl), h-α-syn_A53T (A53T, 4 μm), h-α-syn (Syn, 4 μm), or h-α-syn_A30P A30P, 4 μm). C, ICa, Rate_decay, and ΔC_m (mean ± SEM) induced by depol_20ms in calyces dialyzed with a control solution (Ctrl, 11 calyces), h-α-syn_A53T (A53T, 4 μm, 8 calyces), h-α-syn (Syn, 4 μm, 11 calyces), or h-α-syn_A30P (A30P, 4 μm, 10 calyces). D, E, Similar arrangement as in B and C, respectively, except that depol_20ms was replaced with depol_20ms×10. Ctrl, 10 calyces; A53T, 10 calyces; Syn, 16 calyces; A30P, 9 calyces. F, G, Rate_decay induced by depol_20ms (F) and depol_20ms×10 (G) in rat calyces dialyzed with h-α-syn at 0 μm (11 calyces in F, 10 calyces in G), 1 μm (9 calyces in F, 8 calyces in G), 4 μm (11 calyces in F, 16 calyces in G), and 12 μm (8 calyces in F, 9 calyces in G).

At 5–10 min after whole-cell break-in, h-α-syn_A53T dialysis decreased Rate_decay induced by depol_20ms from 31 ± 2 fF/s (n = 11) in control to 18.4 ± 4 fF/s (n = 8, p < 0.05) without significantly affecting ICa or ΔC_m (Fig. 4B,C). h-α-syn_A53T dialysis also decreased Rate_decay induced by depol_20ms×10 from 188 ± 15 fF/s (n = 10) in control to 75 ± 14 fF/s (n = 10 calyces, p < 0.01; Fig. 4D,E). Similar reduction was observed at calyces dialyzed by h-α-syn (11 calyces for depol_20ms; 16 calyces for depol_20ms×10; p < 0.01), but not by h-α-syn_A30P (4 μm, n = 10 and 9 calyces, respectively; Fig. 4B–E). The effect of h-α-syn on Rate_decay induced by depol_20ms and depol_20ms×10 was dose dependent, and approached saturation at ∼4 μm (Fig. 4F–G). Thus, like chronically expressed h-α-syn_A53T, short-term dialysis of h-α-syn_A53T and h-α-syn inhibited both slow and rapid endocytosis, while the dialysis of h-α-syn_A30P did not. Unlike long-term h-α-syn_A53T expression, h-α-syn_A53T dialysis did not affect ΔC_m induced by depol_20ms×10 (Fig. 4B–E), probably due to shorter exposure.

Discussion

We reported for the first time that a prolonged increase of h-α-syn_A53T impairs synaptic transmission at a mammalian nerve terminal, the calyx of Held, by inhibiting both slow and rapid endocytosis, and the RRP replenishment (Figs. 2, 3). Short-term dialysis of h-α-syn_A53T or h-α-syn into calyces similarly inhibited slow and rapid endocytosis (Fig. 4). These defects may contribute to the synaptic impairment observed before the overt neurodegeneration in α-syn-linked neurodegenerative disorders, such as PD, dementia with Lewy bodies, and multiple-systems atrophy. They might be a sign of early neurodegeneration.

Inhibition of RRP replenishment by prolonged elevation of the levels of h-α-syn_A53T in h-α-syn_{A53T_calyx} mice could result from long-term inhibition of endocytosis, because endocytosis recycles vesicle membrane as well as facilitates the RRP replenishment by clearance of the active zone (for review, see Wu et al., 2014). However, we could not exclude the possibility that the prolonged elevation of h-α-syn_A53T levels inhibits RRP replenishment by a mechanism independent of endocytosis.

A recent study (Busch et al., 2014) shows that short-term injection of h-α-syn_A53T into lamprey nerve terminals leads to expansion of the plasma membrane and reduction of vesicle numbers, implying a defect in vesicle recycling. Our work advances over this previous study in three aspects. First, by real-time endocytosis measurements, we provided direct evidence for the inhibitory effects of h-α-syn_A53T on the two most common forms of endocytosis, slow and rapid endocytosis. Second, we found that endocytosis was inhibited not only by acute elevation, but also by prolonged elevation of h-α-syn_A53T levels in h-α-syn_{A53T_calyx} mice that mimics the constant presence of α-syn mutant proteins in α-syn-linked neurodegeneration. Third, our observations were from mammalian central synapses.

α-syn_A53T and α-syn, but not α-syn_A30P, reduce synaptic vesicle number and expand plasma membrane area at lamprey synapses (Busch et al., 2014), and inhibit transmitter release at cultured hippocampal synapses (Nemani et al., 2010). Consistently, we found that h-α-syn_A53T and h-α-syn, but not h-α-syn_A30P, inhibited endocytosis. The lack of inhibitory effects of α-syn_A30P might be due to its weaker membrane binding affinity than α-syn or α-syn_A53T (Jo et al., 2002; Outeiro and Lindquist, 2003; Bussell and Eliezer, 2003). We could not exclude the possibility that the longer exposure of α-syn_A30P in older mice inhibits endocytosis and exocytosis, considering that the α-syn_A30P mutation is associated with late PD onset and incomplete penetrance (Krüger et al., 2001).

α-Syn participates in membrane curvature formation and sensing (Pranke et al., 2011; Westphal and Chandra, 2013). Excess supply of α-syn or α-syn_A53T might thus inhibit endocytosis by affecting curvature formation. Since triple knockout of α-syn, β-syn, and γ-syn also inhibits endocytosis at hippocampal synapses (Vargas et al., 2014), we suggest that the right balance of the amount of synuclein in nerve terminals is critical for endocytosis.

Several possible mechanisms might explain how elevated levels of α-syn_A53T or α-syn inhibits endocytosis. α-Syn can interact with SNARE (Chandra et al., 2005; Burré et al., 2010; Darios et al., 2010) and cysteine string protein (Miller et al., 2003), which are known to regulate endocytosis (Deák et al., 2004; Hosoi et al., 2009; Rozas et al., 2012; Zhang et al., 2012, 2013; Xu et al., 2013). Alternatively, α-syn can bind to neuronal membrane and form nonselective cation pores (Mironov, 2015), permitting calcium influx to elevate intracellular calcium levels. If the increased calcium level is global, prolonged, and small, it may slow down endocytosis (von Gersdorff and Matthews, 1994; Wu and Wu, 2014) and thus explain why elevated an α-syn level inhibits endocytosis. Identifying the mechanism by which h-α-syn_A53T and h-α-syn inhibit endocytosis will be an interesting future research direction.

In addition to the study of endocytosis, the present work provided the first mouse line, the Math5^Cre mouse, for generating gene knockout or expression specifically in calyces but not in postsynaptic neurons (Fig. 1). The calyx is a preparation of choice for examining biophysical properties of nerve terminals, such as calcium channels, calcium affinity to exocytosis, exocytosis and endocytosis modes, and short-term synaptic plasticity (Borst and Soria van Hoeve, 2012). However, molecular studies of these processes at calyces have been limited partly because global gene knockout is often lethal at birth, before calyces are formed. This problem can now be overcome by making conditional knockout using the Math5^Cre mouse line. The combination of this mouse line with various disease mouse models will open the door to study presynaptic mechanisms of neurological disorders.