C-terminal calcium binding of α-synuclein modulates synaptic vesicle interaction

Alpha-synuclein is known to bind to small unilamellar vesicles (SUVs) via its N terminus, which forms an amphipathic alpha-helix upon membrane interaction. Here we show that calcium binds to the C terminus of alpha-synuclein, therewith increasing its lipid-binding capacity. Using CEST-NMR, we reveal that alpha-synuclein interacts with isolated synaptic vesicles with two regions, the N terminus, already known from studies on SUVs, and additionally via its C terminus, which is regulated by the binding of calcium. Indeed, dSTORM on synaptosomes shows that calcium mediates the localization of alpha-synuclein at the pre-synaptic terminal, and an imbalance in calcium or alpha-synuclein can cause synaptic vesicle clustering, as seen ex vivo and in vitro. This study provides a new view on the binding of alpha-synuclein to synaptic vesicles, which might also affect our understanding of synucleinopathies.

In the present work, the authors investigated the binding properties of the C-terminus of αsynuclein (αS) to lipids. They estimated the affinity of calcium binding to αS, tested whether the binding of calcium by aS differ in the absence and presence of SVs, and studied the interaction of aS and SVs with and without calcium. They used NMR, (CEST) NMR, mass spectrometry, immunofluorescence and superresolution microscopy techniques, and used isolated synaptic vesicles, synaptosomes, and dopaminergic neurons in culture. They main findings are that the calcium affinity for αS is 21 μM, what is within the range of physiological intracellular calcium transients, and that calcium increases the binding of αS to lipids and SVs. They propose that αS is a modulator of SVs homeostasis and that the clustering of synaptic vesicles is regulated by calcium binding to αS. The manuscript is clear and the experiments are well designed and performed. The results could be of interest for understanding the physiological and pathological properties of αS.

Major concerns:
Of interest is that the Ca-dependent clustering of αS was reversed in the presence of a calcium chelator (EGTA), while VAMP2 distribution was not affected by calcium ( Fig. 2C & D). Were 2.5 mM calcium or 1 mM EGTA present at the extracellular medium? In this case, is the synaptosome's membrane permeable to EGTA? Did you used EGTA-AM? How was calcium entry stimulated?. In experiments illustrated in Figure 3, superresolution of isolated synaptic vesicles (SVs), the calcium concentration used was far away from physiological cytosolic values. Do you have the same effect with 100-200 μM calcium?
Finally, I am a little confused after having compared the present conclusions with those in a previous paper of the authors in the same journal (Fusco et al., 2016. Nat Com). In the 2016's paper, they also proposed that aS induces the clustering of SVs by a double-anchor mechanism consisting in the binding of its N-terminus to one SV and its lipophilic central segment (NAC region) to another SV. Here, however, it seems that the N-and the Cterminus are the binding regions. Is the discrepancy methodological (synthetic lipid vesicles versus synaptic vesicles)? And if not, should be concluded that all three main regions of aS can physiologically bind to lipids? This point should be clarified in the Discussion.

Minor point:
In the graph axis of several figures, it is confusing to see the '/' symbol before the units. In this study Schierle and collaborators describe a novel calcium-binding feature of alphasynuclein. They suggest that calcium binds to the C-terminus of alpha-synuclein, and that this region of the molecule also binds to synaptic vesicles. Finally, they suggest that this interaction is involved in the regulation of the localization of alpha-synuclein, in the regulation of synaptic vesicle clustering, and possibly also in synucleinopathies.
As the interaction of alpha-synuclein with synaptic vesicles and other synaptic elements has been analyzed in the past in detail, as the authors acknowledge, the current manuscript needs to be highly convincing, in order to be considered a useful addition to the field. A number of issues, however, are still open, and therefore render the manuscript unsuitable for publication at this time.
Here are the major points: 1) The binding of calcium in vitro seems to require millimolar concentrations. Is this the case? If so, the interaction described here has limited relevance, as such concentrations will only be encountered in neurons with a severely damaged membrane, in which physiological processes would be shut off.
2) In figure 1d, the effect of calcium does not appear to be significantly different from that of magnesium. The blot shown is very different from the quantification shown, and is therefore misleading. 3) Figure 2c contradicts strongly the main claims of the manuscript. It shows that alphasynuclein colocalizes very poorly with the synaptic vesicle marker VAMP2. In fact, it seems to avoid the synaptic vesicles in these images, which implies that the vesicle-binding reaction that the authors describe in vitro is not particularly strong in real synapses. Also, the localization of synuclein seems to be affected by calcium, while that of the vesicles is not. This again suggests that the interaction described here is not particularly strong, and/or that it requires extremely high calcium levels, which are not encountered in living synapses. 4) In Figure 3, alpha-synuclein interacts with synaptic vesicles in vitro, and causes their clustering in the absence of calcium. The clustering actually appears to be reduced in presence of calcium, at least according to the typical images provided, which again contradicts the main statements of the authors. 5) The neuron experiments presented in Figure 4 do not seem to add substantially to the manuscript. The effects of a calcium channel blocker are extremely complex in this context, since calcium entry has a wide range of effects, in addition to the potential effects on alphasynuclein. 6) The effects presented in Figure 5 are not convincing, especially as the experiment has only been repeated twice. The fact that no vesicles are actually seen in the preparation made in presence of synaptic vesicles is confusing.

Reviewer #3 (Remarks to the Author):
The central question addressed by this manuscript is how does calcium modulate the physiological function of alpha-synuclein? The authors show that in the presence of Ca2+ synuclein binds isolated synaptic vesicles not only though its N-terminal region, as previously shown, but also through its C-terminus. The latter interaction is modulated by Ca2+ binding to synuclein and is a key determinant of synaptic vesicle clustering. This manuscript represents a major advancement relative to the previous model of synuclein -lipid vesicle interactions (i.e. the "double anchor mechanism") and one of the major strengths of this manuscript is that it focuses on physiological synaptic vesicles isolated from rat brains as opposed to the simpler lipid vesicles used in previous studies.
Considering that large calcium fluctuations occur in vivo at the pre-synapsis during neuron depolarization, this manuscript is highly relevant for the physiology and pathology of synuclein in vivo. Another notable strength of this ms is the excellent integration of orthogonal experimental techniques (NMR, MS, immunofluorescence, TEM, STED, dSTORM, Tht and SEC assays).
Minor points: -The authors mention that the binding of calcium to synuclein was analyzed in terms of a "multiple ligand model". This term is vague and should be explained better. For example, is this multiple ligand model a Scatchard model where all calcium binding sites are equivalent and independent of each other? Or is this model a Hill-like model in which calcium binding occurs without intermediates in a highly cooperative manner? The latter scenario seems more consistent with the 2nd equation in p. 34, however in this case the concentration of free calcium in the 1st equation in p.34 should appear at the power of "L". It is possible that the effective affinity may not change after these ambiguities about the model are clarified, but this needs to be addressed so that the first two equations of p. 34 are reconciled. -In page 8, the endogenous levels of alpha-synuclein should be specified to let the reader appreciate the concertation difference relative to what is called high synuclein concertation (i.e. 50 uM) -P. 35: was the reference offset for CEST only at -100 kHz or at both +/-100 kHz? -In p. 41, the exp is missing the last parenthesis ")"

Reviewer #1
We would like to thank reviewer 1 for his/her very constructive criticism. There were three major points the reviewer wanted us to address: Clarification of the experimental set-up described in Figure 2c&d Both reviewer 1 and 2 commented on this and we apologize that this has not been clear. We provided more details on the experimental procedures in the text and added new experiments to make this point clearer.
(ii) Does synaptic vesicle clustering also occur at lower, physiological calcium concentrations?
This is a very valid point and also reviewer 2 wanted this point to be addressed. We thus performed the experiments described in Fig. 3 at lower, physiological calcium concentrations and show that synaptic vesicle clustering already occurs at much lower calcium concentrations than shown before.
(iii) How do our current results compare to our previous results published in 2016? (Fusco et al. 2016 Nat Commun) We agree with the reviewer that this point has not been very clear and we have now significantly amended the text to clarify this.

Point1
• The reviewer asked to specify the conditions used for the incubation of synaptosomes (Figure 2c and d). We thank for this comment, as it helped us to make this point clearer in the manuscript.
The synaptosomes were either incubated with an extracellular solution containing 2.5 mM calcium which represents normal physiological conditions (i.e. low calcium concentrations in the intracellular space and high calcium concentrations in the extracellular space) or with an extracellular solution containing 1 mM EGTA. We clarified this point in the text pointing out that the alpha-synuclein clustering is seen under physiological conditions (page 7). Furthermore, we termed the EGTA condition "calcium starved/depleted", as we think this probably describes the treatment better, as the reviewer is right we did not directly add EGTA into the synaptosomes (via permeabilization or using EGTA-AM) for the reasons that direct intracellular manipulation could affect other pathways in intact synaptosomes.
• In addition to this, the reviewer asked how we had increased the calcium concentration in the intracellular space.

vesicles. We referred to this N-terminus and NAC-region binding of alphasynuclein to SUVs as double anchor mechanism. In our current study we have characterised the binding of alpha-synuclein to synaptic vesicles isolated from rat brain using CEST NMR, showing again that the strongest binding occurs at the Nterminus and revealing an intermediate level of synaptic vesicle interaction of the NAC-region and the C-terminus. Upon addition of calcium this binding via the Cterminus, and also for certain residues of the NAC-region, was found to be increased.
We refer to this now as extended double anchor mechanism since the double anchor has been extended to the C-terminus of alpha-synuclein.
We acknowledge that the first version of the paper was not entirely clear on this aspect and have revised the manuscript accordingly (page 6/7 and 12/13).
We agree with the reviewer and have changed the axes labelling in Figure 1b and Figure 4

Reviewer #2
We would like to thank the reviewer for his/her very constructive criticism. The major points from reviewer #2 refer to the physiological relevance of our findings and the repeatability of our results. To address these points we

Point-to-Point discussion
Point 1 • Reviewer 2 wondered whether our results in vitro are physiologically relevant as they require mM concentrations. Llinás et al., 1992;Schneggenburger and Neher, 2000). For NMR studies high protein concentrations are required (200 µM), thus higher concentrations of calcium had to be used for stoichiometric reasons. The titration was done up to 18 equivalents of [alpha-synuclein] which are ranges of ligand commonly found in titrations of other calcium binding proteins using NMR (Prebil et al., 2016;Ubach et al., 1998).

Our NMR-based calcium titration experiment determined an alpha-synuclein calcium binding constant of 21 µM, which is well within the range of physiological concentrations such as occur during synaptic stimulation (up to several 100 µM,
However, for the in vitro data described in Fig. 3 we had used millimolar calcium concentrations. This was also a concern that was raised by reviewer 1 (see comment to reviewer 1 point 2). We thus performed new experiments and studied synaptic vesicle clustering at 200 µM rather than at 2.5 mM calcium. We now show that synaptic vesicle clustering already occurs at physiological calcium concentrations (Figure 3a, page 8/9). synaptic vesicle clustering already occurs at physiological calcium concentrations reached during synaptic activity (Figure 3a, page 8/9).

Point 2
• The reviewer asked for the quantification of the Western blot shown in Figure 1d.
We agree that the Western blot shown in Figure 1d did not represent the statistics shown. We have now repeated the experiments and show a more representative image in Fig. 1d. For further transparency, we also added our Western blot repeats to the supplement (Supplementary figure 3).

Point 3
• The reviewer asked on the interpretation of our results on synaptosomes regarding the distribution of alpha-synuclein and VAMP2 shown in Fig. 2c&d. In particular, the reviewer thinks that the vesicle-binding reaction that we describe in vitro is not particularly strong in real synapses. We apologize and we hope our new data and our clarifications in the text will help to understand our results better.
We have added new data to show that synaptic vesicle clustering already takes place at 200 µM calcium which is well within a physiological range (see also comment to point 1 above). Furthermore, our synaptosome experiments were done under normal physiological conditions (normal physiological calcium in the intracellular space and high extracellular calcium). We have now clarified this in the text and apologize if this has not been clear. In order to address whether increased intracellular calcium concentrations have a further effect on synaptic vesicle clustering in synaptosomes we added new experimental data for which we stimulated synaptosomes with 70 mM KCl (affecting membrane potential and leading to calcium entry via voltage-gated calcium channels; see also point 1 of reviewer1). We did not observe any further clustering of alpha-synuclein confirming that already normal physiological calcium concentrations can lead to alphasynuclein clustering.
The reviewer is correct, that in (Figure 2c), under physiological conditions (normal intracellular calcium and high (2.5 mM) extracellular calcium), we see that alphasynuclein staining does not overlap with VAMP2 staining. We agree that this could be interpreted as if alpha-synuclein does not bind to synaptic vesicles at all. We do, however, not believe that this is the case for two main reasons: First, all our in vitro data as well as data from others (Georgieva et al. 2008, Jao et al. 2008, Bodner et al. 2009, Middleton et al. 2010, Trexler et al. 2010, Ulmer et al. 2005 (Fusco et al. 2016).
Point 5 • The reviewer indicated that the experiments done with isradipine on neurons do not add substantially to the manuscript.
We disagree with the reviewer on this point. Isradipine is a voltage-gated calcium channel blocker and thus reduces the amount of calcium that can enter upon neuronal stimulation such as upon dopamine treatment. We show that synaptic vesicle clustering occurs upon dopamine treatment which is blocked by isradipine and we also show that toxicity is blocked by the addition of isradipine. We thus provide further evidence that controlling the calcium levels, such as by using a voltage gated calcium channel blocker, may be a protective strategy. We thus feel that these experiments are relevant and support our hypothesis that increased calcium concentrations may enhance toxicity.

Point 6
• The reviewer pointed out that the experiment in Figure 5 was only repeated twice and that synaptic vesicles should be visible in the TEM images shown in Fig.5c.  (Supplementary figure 10). We added a paragraph to the manuscript (page 11), outlining that we do not see synaptic vesicles present at the end of the aggregation reaction and make reference to the supplementary figure.

Reviewer #3
We would like to thank reviewer 3 for highlighting the advancement of our approach to previous approaches and for mentioning that our manuscript is highly relevant for the physiology and pathology of synuclein. There were three main points reviewer #3 wanted us to address:

(i) Explain "multiple ligand model"
We addressed this in the text and added the respective reference. We corrected and defined this in the text.
(iii) Improve Figure 2a As suggested we added a smoothed average trend line to Fig. 2a.

Point-to-Point discussion
Point 1 • Define the multiple ligand model. We agree with the reviewer and we have now provided more detail on our choice of the multiple ligand model.
We used a model that has been previously used to characterize the binding of alpha synuclein to lipids (Galvagnion et al Nat Chem Bio, 2015). The advantage of this model is that it is based on the total concentration of ligands, which we know accurately, whereas other models such as the Hill equation for multiple binding sites requires the knowledge of the free ligand concentration, which is quite challenging to measure since calcium is in fast exchange with alpha-synuclein. In particular, as in our case the concentration of the protein is close to the KD we cannot approximate free [L] with total [L]. We have now provided a more detailed set of equations to explain the basis of our model and added a reference for further details (Galvagnion et al Nat Chem Bio, 2015) (see page 34/35 and 5).
Point 2 • The referee pointed out that there are discrepancies between Figure 1a and Supplementary figure 1, which we addressed accordingly.
We thank the referee for this point and have now corrected the legend of figure 1a.
We also confirm that the units of Fig. S1 are in ppm (these have now been added to the plot). The peak changes in figure 1a are not as high as the maximum changes in the supplementary figure 1 (3.6mM calcium) as Figure 1a shows an intermediate point in the titration (at a calcium concentration of 1.6 mM). We have amended the legend of figure 1a to clarify this.

Point 3
• We thank the reviewer for the comment on figure 2a as the new figure helps the reader to see the differences between the two measurements better.
We included the CEST Figure 2a with a smoothed average trend line.

Point 4
• The reviewer wanted us to compare the alpha-synuclein concentration used for the experiment shown in Figure 3a with the endogenous alpha-synuclein concentration.
The value estimated for the endogenous presynaptic alpha-synuclein concentration is 43 µM (Takamori et al., 2006), thus the concentration used for our experiment will be equivalent to twice the endogenous concentration (as seen in disease upon gene duplication). We have now added the alpha-synuclein concentration on page 8.

Point 5
• The reviewer asked about the offset for the CEST measurements.
We employed -100 kHz to measure the reference in the CEST experiment as for our previous experiments published by Fusco et al. 2016.