α-Synuclein oligomers mediate the aberrant form of spike-induced calcium release from IP3 receptor

Emerging evidence implicates α-synuclein oligomers as potential culprits in the pathogenesis of Lewy body disease (LBD). Soluble oligomeric α-synuclein accumulation in cytoplasm is believed to modify neuronal activities and intraneural Ca2+ dynamics, which augment the metabolic burden in central neurons vulnerable to LBD, although this hypothesis remains to be fully tested. We evaluated how intracellular α-synuclein oligomers affect the neuronal excitabilities and Ca2+ dynamics of pyramidal neurons in neocortical slices from mice. Intracellular application of α-synuclein containing stable higher-order oligomers (αSNo) significantly reduced spike frequency during current injection, elongated the duration of spike afterhyperpolarization (AHP), and enlarged AHP current charge in comparison with that of α-synuclein without higher-order oligomers. This αSNo-mediated alteration was triggered by spike-induced Ca2+ release from inositol trisphosphate receptors (IP3R) functionally coupled with L-type Ca2+ channels and SK-type K+ channels. Further electrophysiological and immunochemical observations revealed that α-synuclein oligomers greater than 100 kDa were directly associated with calcium-binding protein 1, which is responsible for regulating IP3R gating. They also block Ca2+-dependent inactivation of IP3R, and trigger Ca2+-induced Ca2+ release from IP3R during multiple spikes. This aberrant machinery may result in intraneural Ca2+ dyshomeostasis and may be the molecular basis for the vulnerability of neurons in LBD brains.

A growing body of evidence implicates α-synuclein oligomers are potential culprits in the pathogenesis of Lewy body dementia (LBD), which refers to dementia with Lewy bodies and Parkinson's disease with dementia 1 . The presence of α-synuclein oligomers has been demonstrated in LBD brains 2,3 , not only in the neuropil, but also in the soma of LBD vulnerable neurons 4 . The aggregation of α-synuclein is upregulated by either mutations to α-synuclein or exposure to dopamine 1 . α-Synuclein oligomers mediate toxicity that occurs via several intracellular mechanism such as mitochondrial and endoplasmic reticulum (ER) stress and an impaired autophagy-lysosomal pathway 5,6 . Therefore, the α-synuclein oligomer is a key molecule in respect to the toxicity to LBD vulnerable neurons.
Dysregulated Ca 2+ homeostasis has emerged as an underlying pathological mechanism in LBD; it triggers the formation of α-synuclein oligomers, mitochondrial and ER stress, and the inhibition of autophagy and lysosomal pathways, thereby prompting neurodegeneration 5,7 . Epidemiological studies indicate that L-type VDCC (L-VDCC) blockers diminish the risk of Parkinson's disease (PD) 8,9 . In general, LBD vulnerable neurons such as neurons in the substantia nigra pars compacta (SNc), locus coeruleus, raphe nuclei and the nucleus basalis of Meynert, have a common physiological phenotype; an autonomous pacemaker, broad and slow spiking, or lower expression of Ca 2+ -binding proteins. These physiological characteristics lead to increased cytosolic Ca 2+ and augment the metabolic burden in these neurons critical for selective neuronal degeneration 7 . In LBD, Lewy bodies appear in neocortical pyramidal neurons and contribute to dementia 10,11 . This raises a question on how intraneural oligomeric α-synuclein can pathologically modify neuronal activity and intracellular Ca 2+ dynamics in neocortical neurons; this question remains to be answered.
We previously demonstrated how Ca 2+ or K + channels are involved in the regulation or pathophysiological alteration of neocortical pyramidal cell excitability and Ca 2+ dynamics. This was performed by using intracellular Specimen recordings of action potentials during positive current pulses (300 ms, 0.3 nA and 0.5 A) in neurons injected by vehicle solution (Control), dopamine incubated without α-synuclein (DA), α-synuclein incubated without dopamine (αSN), and α-synuclein incubated with dopamine (αSNo). The interspike interval was prolonged and the spike frequency was reduced in αSNo-administered neurons. Calibration: 100 ms, 10 mV. (c) Average spike frequencies elicited by varying positive current steps (0.1-0.5 nA) in neurons of layer II/III in the frontal cortex. The frequency was significantly lower in neurons infused with αSNo (*<0.05, **<0.01, One way ANOVA) or A53T α-synuclein incubated with dopamine (αSN53o, + <0.05, ++ <0.01, One way ANOVA), than in neurons with Control at 0.3, 0.4, and 0.5 nA current steps. (d) Mean spike frequencies elicited by varying steps of depolarizing current in neurons of layer V in frontal cortex. The frequency was significantly lower in neurons injected with αSNo than in neurons with Control at 0.2, 0.3, 0.4, and 0.5 nA current steps. *<0.05, **<0.01 (One way ANOVA) (e). Numeric data of average spike frequencies elicited by 0.1-0.5 nA current steps question, the spike frequency and I AHP charge in αSNo-injected or αSN-injected neurons were examined under the application of blockers for the channels or receptors responsible for intraneural Ca 2+ dynamics (Fig. 3).
Previous studies have established that a Ca 2+ -dependent functional triad consisting of VDCC, IP 3 R and SK channel is linked to spike-triggered Ca 2+ inflow and Ca 2+ release from IP 3 R in neurons of the neocortex and amygdala [12][13][14][21][22][23][24] . Therefore, our findings strongly suggest that, via this channel coupling, α-synuclein oligomers mediate Ca 2+ -induced Ca 2+ release (CICR) from IP 3 R, which are triggered by Ca 2+ influx via L-VDCC during multiple spikes, followed by the elongation of SK channel opening, the prolongation of I AHP , and reductions in spike frequency. Consequently, in neocortical pyramidal neurons, we can detect the occurrence of this mode of CICR by observing the enlargement of I AHP charge and the reduction in spike frequency. α-Synuclein oligomers target the regulation of ip 3 R gating and mediate an aberrant form of cicR from ip 3 R during multiple spikes. Which player is the direct target of αSNo mediation of CICR from IP 3 R? IP 3 R has two separate binding sites for Ca 2+ and IP 3 , with these being regulated allosterically by these two ligands, with binding of one ligand facilitating additional binding of the other 25,26 . Under this positively cooperative mechanism, IP 3 R responds to the increase in neuronal cytosolic Ca 2+ and IP 3 , and effectively opens, releasing Ca 2+ from the ER in an activity-dependent manner [12][13][14]21,27,28 . Accordingly, there are two candidates for the target mechanism by which αSNo causes CICR from IP 3 R: (1) the elevation of IP 3 turnover; (2) the regulation of IP 3 R gating.
αSNo can directly upregulate IP 3 R and cause aberrant CICR from IP 3 R without elevating IP 3 turnover during repetitive spikes, which would not take place under physiological conditions in neocortical pyramidal neuron. In this scenario, intracellular application of IP 3 will mimic and occlude the action of αSNo. Indeed, I AHP charge and the spike frequency in neurons with D-IP 3 and αSN (7.5 ± 1.1 pC and 19.3 ± 1.6 Hz, n = 5) were at the same level as those in neurons with D-IP 3 and αSNo (7.3 ± 0.8 pC and 19.4 ± 1.0 Hz, n = 6; Fig. 4a-d) and αSNo alone (6.2 ± 0.7 pC and 21.1 ± 0.8 Hz, n = 6; αSNo, no drug), and were significantly different from those in neurons with αSN alone (3.5 ± 0.7 pC, p = 0.019, and 31.1 ± 2.8 Hz, n = 6, p = 0.005; αSN, no drug). In neurons with D-IP 3 and αSNo, I AHP charge was also significantly larger (p = 0.005), while the spike frequency was significantly smaller (p = 0.002) than in those with αSN alone (αSN, no drug). In contrast to D-IP 3 , the application of L-IP 3 , a negative analog of IP 3 , had no effect on the αSNo-mediated alteration of I AHP charge (αSNo, 6.7 ± 1.1 pC, n = 4 vs αSN, 2.5 ± 0.5 pC, n = 4, p = 0.025; Fig. 4a,b) and spike frequency (αSNo, 21.7 ± 1.0 Hz vs αSN, 28.3 ± 1.1 Hz, p = 0.018; Fig. 4c,d). In combination, our findings indicate that α-synuclein oligomers target the regulation of IP 3 R gating, and mediate the aberrant form of CICR from IP 3 R during repetitive spikes, without enhancing Ca 2+ influx or IP 3 production in neocortical neurons.
in neurons of layer II/III (c) and V (d). (f) Resting membrane potential (RMP), spike half-width, and medium afterhyperpolarization (mAHP) in pyramidal neurons of layer II/III and V from frontal cortex slices. (2019) 9:15977 | https://doi.org/10.1038/s41598-019-52135-3 www.nature.com/scientificreports www.nature.com/scientificreports/ the association of α-synuclein oligomers with CaBP1 allows aberrant CICR from IP 3 R by suppressing CaBP1-mediated inactivation of IP 3 R. The gating of IP 3 R is not only regulated by IP 3 binding; it is also modulated by Ca 2+ and a variety of proteins 29,30 . Given that αSNo directly targets IP 3 R gating without enhancing Ca 2+ influx or IP 3 turnover, αSNo could be associated with the protein that directly binds and regulates IP 3 R in central neurons. To determine the site of action of αSNo and the mechanism by which αSNo mediates CICR from IP 3 R, we bibliographically searched for a protein that meets the conditions, and focused on Ca 2+ -binding protein 1 (CaBP1) amongst the binding partners of IP 3 R, because CaBP1 is (1) a Ca 2+ -binding www.nature.com/scientificreports www.nature.com/scientificreports/ protein distributed in the cytosol of rodent and human central neurons [31][32][33] , (2) a preferential interacting protein with α-synuclein oligomers 34 , and (3) a binding partner and negative regulator of IP 3 R under high intraneural Ca 2+ by means of Ca 2+ -dependent inactivation [35][36][37] . If αSNo captures CaBP1 and pulls it away from IP 3 R, thus preventing IP 3 R from Ca 2+ -dependent inactivation, an aberrant CICR from IP 3 R could occur, without reinforcing Ca 2+ influx or cytosolic IP 3 level.
To determine whether higher-order oligomeric α-synuclein actually binds to CaBP1, we conducted an immunoprecipitation (IP) experiment. The input solution containing αSNo and GST-CaBP1 was immunoprecipitated with anti-CaBP1 antibody, followed by IB with antibodies against α-synuclein (Fig. 6a) and anti-GST (Fig. 6b). This experiment demonstrated that anti-CaBP1 antibodies were sufficient for IP, but too weak to detect CaBP1 for IB; we therefore used antibodies against GST tagging CaBP1, instead of anti-CaBP1 antibodies, for IB. The results shown in Fig. 6a demonstrate that α-synuclein oligomers larger than 100 kDa and aggregates were present in anti-CaBP1-precipitated samples. We also confirmed that CaBP1 was present in the same batch of the precipitated sample (Fig. 6b). These results indicate the direct binding of higher-order α-synuclein oligomers larger than 100 kDa with CaBP1. In combination, our findings demonstrate that the aberrant CICR occurred only by higher order α-synuclein oligomer larger than 100 kDa.

Discussion
The present study revealed that intracellularly injected α-synuclein oligomers mediate activity-dependent CICR from IP 3 R, as indicated by the resulting prolonged I AHP and decreased spike frequency in neocortical pyramidal neurons. α-Synuclein oligomers capture CaBP1, and prevent IP 3 R from causing Ca 2+ -dependent inactivation For the 'no drug' example in (b,d), the same data as shown in Fig. 3 are reproduced for clarity. *p < 0.01, **p < 0.02 (αSNo vs αSN, t-tests), + p < 0.01 (D-IP 3 vs no drug, αSN, t-tests).
during multiple spikes, thereby releasing Ca 2+ from ER Ca 2+ store via IP 3 R without increasing Ca 2+ influx or IP 3 turnover. This aberrant form of activity-dependent Ca 2+ release is mediated only by higher order α-synuclein oligomers larger than 100 kDa, but not by α-synuclein species less than 100 kDa.
In previous studies reporting the effect of intraneural α-synuclein on cytoplasmic Ca 2+ dynamics and neuronal excitability, transgenic α-synuclein mice exhibited augmented long-lasting Ca 2+ transients in response to repetitive stimulation in vivo 42 , and a reduction in neocortical pyramidal cell excitability was observed by injecting α-synuclein oligomers prepared without dopamine 43 . The present study is consistent with these studies and the first to search in detail for the mechanism of intracellular oligomeric α-synuclein modifying Ca 2+ handling, and to identify the site of action.
The coupling of the spike-induced Ca 2+ entry via VDCC and CICR from IP 3 R with the enhancement of the SK channel is a well-documented mechanism in the somatodendritic area of neurons in the neocortex and amygdala; it contributes to the regulation of neuronal excitability and synaptic plasticity [12][13][14][21][22][23][24] . In contrast with previous reports emphasizing that the physiological upregulation of IP 3 turnover finely tuned by synaptic stimulation or neuromodulation, is necessary for spike-induced or IP 3 -induced Ca 2+ release from IP 3 R in central neurons [12][13][14]21,22,27,44,45 (Fig. 6c i,ii), our observation revealed that oligomeric α-synuclein-mediated CICR from IP 3 R was independent of the elevation of IP 3 production, because the PLC blocker failed to inhibit it (Fig. 4). Such an unusual mode of CICR provoked by highly frequent neuronal activity, independent of IP 3 turnover, does not usually take place in central neurons, as the regulation of IP 3 R gating displays bell-shaped dependence on cytosolic Ca 2+ concentration 46 . This mode of CICR can therefore be reasonably considered as pathological, imposing an excess Ca 2+ burden on neurons (Fig. 6c iii).
The reason why we used dopamine is to obtain stabilized α-synuclein oligomers. Co-existence of α-synuclein with dopamine results in the formation of SDS-resistant stable soluble oligomers due to dopamine quinones, which contribute to the inhibition of fibrillization by stabilizing α-synuclein oligomers 18,19 . However, the www.nature.com/scientificreports www.nature.com/scientificreports/ present αSNo-mediated action is attributable to α-synuclein oligomers per se, but not to dopamine or dopamine quinones, and the possibility is also excluded that the intracellular presence of dopamine or dopamine quinones may cause some additional artifactual effects on neuronal properties as follows. First, the injection of α-synuclein oligomers produced without dopamine also results in a spike reduction similar to our findings 43 . Second, the application of DA failed to alter neuronal excitability (Fig. 1). Third, the possibility that monomeric α-synuclein-dopamine adducts may be part of the overall effect is unlikely because monomeric α-synuclein fails to bind CABP1 (Fig. 6) and does not mediate the aberrant CICR.
Neuronal Ca 2+ -binding proteins (CaBPs), a sub-branch of the calmodulin superfamily, are Ca 2+ -sensor proteins, and regulate various Ca 2+ channel targets 39,47 . CaBP1, a splice variant of CaBPs, is distributed in the cytosol of central neurons [31][32][33] , and is a preferential α-synuclein oligomer interacting protein, as shown by a co-immunoprecipitation study 34 . As CaPB1 has four EF-hand Ca 2+ -binding motifs, and can bind and regulate IP 3 R under high intraneural Ca 2+ 35-37 , the inhibition of interaction between CaBP1 and IP 3 R can result in the aberrant activity-dependent CICR from IP 3 R without increasing IP 3 production (Fig. 6c iii). We identified the target of α-synuclein oligomers as CaBP1 by electrophysiological recordings (Fig. 5a-d), and confirmed the direct association of α-synuclein oligomers greater than 100 KDa and CaBP1 by IP (Fig. 6a).
Previous reports demonstrated that IP 3 and CaBP1 have opposing effects on the IP 3 R channel under high intracellular Ca 2+ concentrations [35][36][37] . IP 3 disrupts the inter-subunit interaction of IP 3 R and promotes IP 3 R channel opening (Fig. 6c ii), while CaBP1 binds IP 3 R and clamps the inter-subunit interaction of IP 3 R when the IP 3 level is low, thereby inhibiting IP 3 R channel opening in a Ca 2+ -dependent manner (Fig. 6c i). These mechanisms explain why either IP 3 or CaBP1 Ab mimics and occludes and either heparin or CaBP1 blocks the effect of electrophysiological recordings. Electrophysiological recordings were performed as described previously 15,16 . Briefly, slices were placed in a recording chamber on the stage of an upright microscope (BHWI; Olympus, Tokyo, Japan) with a 40× water-immersion objective (LUMPlan FI/IR 40x/0.80 W). The chamber was continuously perfused with bathing medium (25 °C) bubbled with a mixture of 95% O 2 and 5% CO 2 . For recording, patch pipettes (resistance, 5-10 MΩ) filled with a solution (pH 7.3) containing (in mM) 7 KCl, 144 K-gluconate, 10 KOH, 10 HEPES, 4 Na 2 ATP, and 0.4 Na 2 GTP were used. Whole-cell recordings were made at the soma from layer II/III or V pyramidal neurons in frontal cortex slices. Capacitance was compensated to 70-80%. Neurons that had sufficiently negative resting membrane potentials (RMPs; more negative than −60 mV) without spontaneous action potentials were selected.
Recombinant α-synuclein [Wild type (WT) or A53T mutant (A53T), Sigma] was dissolved with sterile water (10 µM) and co-incubated with 100 µM dopamine hydrochloride (Sigma) at 37 o C for 3 days (αSNo or αSN53o). For comparison, 10 µM recombinant α-synuclein (WT or A53T) solution was incubated without dopamine at 37 °C for 3 days (αSN or αSN53). The αSNo or αSN53o included higher-order oligomers of α-synuclein, which were absent in αSN or αSN53 (Fig. 1a). Dopamine (100 µM) without α-synuclein was also prepared, and incubated at 37 °C for 3 days (DA). All these solutions were filtered through a PVDF filter (millex-HV SLHV004SL; Merck Millipore, Darmstadt, Germany), which removed α-synuclein aggregates, including fibrils more than 0.45 μm. These solutions were diluted for the patch pipette internal solution, so that the final concentration of α-synuclein and DA were 1 µM and 10 µM, respectively. Present experiments used the solutions containing α-synuclein oligomers made without phosphate buffered salts (PBS), which are generally used to get a better control of ionic strength and PH 54 , to avoid the possibility that PBS could unexpectedly affect spike property and Ca 2+ dynamics when injected intracellularly.
These solutions or the vehicle solution (Control) were distributed into the cell by diffusion (infusion) for at least 5 min after whole-cell break in, and before the recording session commenced. Membrane potentials were recorded in the current clamp mode (Axopatch 200B; Axon Instruments, CA, USA) and digitized at 10 kHz (Digidata 1440 and pCLAMP10, Axon Instruments). Depolarizing currents (0.1-0.5 nA for 300 ms) were injected through the patch pipette to assess the membrane excitability of the recorded neurons. A single action potential or trains of five spikes at 30 Hz were evoked by a 3-5 ms depolarization current pulse (0.7-0.9 nA) or five depolarization current pulses, respectively. The width of the evoked spikes was measured at 50% of the peak. The afterhyperpolarization (AHP) current (I AHP ) was recorded in voltage clamp mode. I AHP charge was defined as the charge transfer carried by the current elicited by five depolarization pulses that would produce AHP under current clamp. I AHP were integrated after these step depolarization pulses to calculate the charge transfer (I AHP charge) representing the AHP. www.nature.com/scientificreports www.nature.com/scientificreports/ membrane was incubated with anti-α-synuclein antibody (Sigma) and horseradish peroxidase-conjugated secondary antibody (GE Healthcare, Little Chalfont, UK). Immunodetection was performed using the ECL Western blotting detection system (GE Healthcare). immunoprecipitation (ip). GST-CaBP1 protein (100 ng/µL) was added in equal amounts to αSNo containing wild type recombinant α-synuclein (10 µM) and dopamine (100 µM) and incubated for 1 h at 37 °C. The mixture was incubated with or without anti-CaBP1 antibody (Sigma) for 1 h at 37 °C followed by incubation with Protein G sepharose 4 Fast Flow (GE Healthcare) for 1 h at 4 °C with gentle shaking. The beads were precipitated by centrifugation and washed four times with an excess volume of Tris-buffered saline containing 0.1% Triton X-100. Proteins bound to beads were eluted by boiling in a sample-loading buffer. Western blotting was performed as described above, except that anti-α-synuclein antibody (Sigma) and anti-GST antibody (Nacalai Tesque, Kyoto, Japan) were used.
Experimental design and statistical analysis. As was the case for the electrophysiological recordings, experimental data were obtained from four to nine cells in neocortical slices of brains from mice of either sex. Data are expressed as mean ± SEM. "One way ANOVA followed" by post hoc Turkey HSD tests or Games-Howell tests, and paired and unpaired t-tests were used for statistics (SPSS v22, Japan IBM Ltd, Tokyo, Japan).