The LRRK2 G2385R variant is a partial loss-of-function mutation that affects synaptic vesicle trafficking through altered protein interactions

Mutations in the Leucine-rich repeat kinase 2 gene (LRRK2) are associated with familial Parkinson’s disease (PD). LRRK2 protein contains several functional domains, including protein-protein interaction domains at its N- and C-termini. In this study, we analyzed the functional features attributed to LRRK2 by its N- and C-terminal domains. We combined TIRF microscopy and synaptopHluorin assay to visualize synaptic vesicle trafficking. We found that N- and C-terminal domains have opposite impact on synaptic vesicle dynamics. Biochemical analysis demonstrated that different proteins are bound at the two extremities, namely β3-Cav2.1 at N-terminus part and β-Actin and Synapsin I at C-terminus domain. A sequence variant (G2385R) harboured within the C-terminal WD40 domain increases the risk for PD. Complementary biochemical and imaging approaches revealed that the G2385R variant alters strength and quality of LRRK2 interactions and increases fusion of synaptic vesicles. Our data suggest that the G2385R variant behaves like a loss-of-function mutation that mimics activity-driven events. Impaired scaffolding capabilities of mutant LRRK2 resulting in perturbed vesicular trafficking may arise as a common pathophysiological denominator through which different LRRK2 pathological mutations cause disease.

Parkinson's disease (PD) is the second most common age-related neurodegenerative disease clinically characterized by movement impairments such as bradykinesia, rigidity and resting tremor and pathologically by the progressive death of dopaminergic neurons in the substantia nigra and the deposition of protein aggregates known as Lewy bodies 1,2 . Although the majority of cases are idiopathic, mutations in the Leucine-rich repeat kinase 2 (LRRK2) gene (PARK8; OMIM 609007) have been robustly correlated to late-onset autosomal dominant PD. LRRK2 mutations are responsible for up to 13% of familial PD cases and have been identified also in 1 to 2% of idiopathic PD patients 3 . LRRK2 is a large protein with a dual enzymatic activity and two domains involved in protein-protein interaction 4 . Our previous studies suggest that LRRK2 acts at the synaptic site where it binds to synaptic vesicles (SV) 5,6 . In addition, LRRK2 rodent models show neurotransmission defects [7][8][9] and LRRK2 binds and phosphorylates several presynaptic proteins 10,11 . Since its initial description as PD gene, a major focus has been dedicated to LRRK2's GTPase and the kinase domain, linking pathological mutations to altered enzymatic activities 12 . Robust evidences show that kinase activity influences LRRK2 dimerization 13,14 , subcellular distribution 15 and regulates binding to 14-3-3 proteins 16,17 . Accordingly, our previous work suggests that LRRK2 kinase activity modulates its supra-molecular organization and ultimately LRRK2 function at the synaptic site 5 . LRRK2 N-terminal Armadillo, LRR and Ankyrin repeats as well as the C-terminal WD40 domain have been predicted to be involved in protein interactions 18,19 . In this study, we demonstrate that LRRK2 interacts with different partners that exert divergent functions in SV trafficking. Importantly, strong genetic association suggests that the missense substitution of glycine 2385 to arginine (G2385R) within the C-terminal WD40 domain is a pathologically relevant variant. The G2385R variant is associated with an increased risk of developing idiopathic PD in Chinese Han and Korean ethnicity [20][21][22] . We show here that the G2385R variant alters the LRRK2-associated protein interactome and reduces its binding to SV. By combining TIRF microscopy to synaptopHluorin detection, we noticed that LRRK2-G2385R expression as well as LRRK2 silencing resulted in an increase in SV fusion events, suggesting that the G2385R variant is a partial loss-of-function mutation.

Results
Domain-wise dissection of LRRK2 impact on vesicle trafficking. We contributed to accumulating literature describing the role played by LRRK2 in regulating SV fusion events 5,6,10,23 . To further explore the functional role of LRRK2, we combined synaptopHluorin assay to TIRF microscopy (TIRFM) in the neuroblastoma cell line SH-SY5Y.
SynaptopHluorin (sypHy) is a fluorescent reporter of vesicle fusion and recycling. It consists of a pH-sensitive GFP fused to the luminal side of the synaptophysin protein 24 . At the acidic pH inside transmitter vesicles, sypHy fluorescence is low. Upon vesicle release, sypHy is exposed to the neutral extracellular space and fluorescence abruptly increases. We used the sypHy assay with the SH-SY5Y line which expresses LRRK2 together with a panel of SV-associated proteins. We determined that this line shuffles synapto-pHluorin reporter (sypHy) to VAMP2 positive vesicles (Supplementary Figure 1A,B) and exposes sypHy on the membrane as a consequence of SV fusion events 25 . Consequently, we over-expressed a panel of LRRK2 derived expression constructs together with sypHy reporter in the SH-SY5Y line (Fig. 1A). Western-blotting analysis showed that LRRK2-derived constructs are expressed in a similar extent and do not significantly affect sypHy level compared to control empty vector transfected cells ( Fig. 1B and Supplementary Figure 1C,D). Next we analysed SV dynamics by TIRFM. Upon over-expression of full-length LRRK2 we did not notice any major impact on SV dynamics in terms of number of events, total fluorescence elicited ( Fig. 1C-E) or increase in peak fluorescence intensity (Supplementary Figure 1E). Our previous data showed that the C-terminal LRRK2 WD40 domain has a major role in controlling SV trafficking 6 . Therefore we investigated SV trafficking upon over-expression of the isolated LRRK2 WD40 domain. According to our previous data, we found that LRRK2 WD40 domain does not influence single peak intensity ( Fig. 1C and Supplementary Figure 1E) but severely reduces the number of fusion events (Fig. 1D,E). To further characterize the functional role of LRRK2 WD40 domain, we studied SV trafficking upon expression of LRRK2 construct lacking the WD40 domain (hereinafter LRRK2ΔWD40). Interestingly we noticed that the expression of LRRK2ΔWD40 increased the frequency of fusion events (Fig. 1C-E). All together, these findings suggest that the LRRK2 N-terminal fragment and WD40 domain have opposite impact on SV dynamics. LRRK2 interactome is domain specific and modulated by activity. Our previous data suggested that LRRK2 influences SV trafficking via binding to presynaptic proteins 5,6,10,23 . In particular, immunoprecipitation of endogenous LRRK2 protein from mouse forebrain lysates showed that LRRK2 interacts with β3-Ca V 2.1 channel subunit, synapsin I and β-actin ( Fig. 2A). To better characterize actin, synapsin and β3 Ca V 2.1 channel subunit as LRRK2 interactors, we evaluated their sub-cellular localization in cortical neurons. To this aim, DIV14 cortical neurons were fixed and stained with LRRK2, F-actin, synapsin and β3-Ca V 2.1 channel subunit specific antibodies (Fig. 2B). Interestingly, the three proteins showed a high degree of colocalization with LRRK2 as judged by Pearson's coefficient (Fig. 2C). LRRK2 harbours at least two major sites for protein interaction featured by its N-and C-terminus. Thus, we mapped the LRRK2 domain hosting the interaction with β3-Ca V 2.1 channel subunit, synapsin I and β-actin. To this aim, we expressed Strep-FLAG LRRK2 full-length and Strep-FLAG LRRK2ΔWD40 in N2A cells. Upon streptavidin-pull-down, we studied interacting proteins by western-blotting (Fig. 2D). Interestingly, we noticed that while the β3-Ca V 2.1 channel subunit was equally retained by full-length LRRK2 and LRRK2ΔWD40, the interaction with synapsin I and β-actin was severely reduced in the LRRK2ΔWD40 variant (Fig. 2E). In conclusion, these data suggest that LRRK2 binds the β3-Ca V 2.1 channel subunit via its N-terminus while β-actin and synapsin I binding is mainly mediated by the C-terminal WD40 domain. SV move along tubulin tracks, are tethered by actin and synapsin in a resting state and are released upon presynaptic increase of local [Ca 2+ ] 26,27 . Thus we triggered exocytosis by increasing bath [Ca 2+ ] to 10 mM for 30 minutes in DIV14 cortical cultures 28 and we investigated LRRK2 interacting proteins via immunoprecipitation. We noticed that LRRK2 interaction with β-actin and synapsin I was robustly reduced upon stimulation (Fig. 3A,B). This evidence suggests that LRRK2 interactome is shaped by neuronal activity.
The G2385R PD-risk variant is a partial loss-of-function mutation. Our previous findings suggest that LRRK2 binds and mobilizes SV through a specific set of protein-protein interactions taking place at the WD40 domain 6,29 . Thus we studied the impact on SV trafficking of the G2385R WD40 domain variant. To this aim, we analyzed by TIRFM the trafficking of sypHy vesicles upon over-expression of LRRK2 wild-type or G2385R variant (Fig. 4A). Western-blotting analysis showed that LRRK2-derived constructs are expressed in a similar extent and do not affect significantly sypHy level ( Fig. 4B and Supplementary Figure 2A,B). Notably, while single peak fluorescence remained mostly not affected (Supplementary Figure 2E), we reported an increase in the number of fusion events upon LRRK2 G2385R expression ( Fig. 4D-F). This result may arise from an increase in the number of sypHy-positive vesicles or instead from a redistribution of vesicles towards the membrane. To discriminate between these two scenarios, we revealed the total pool of sypHy vesicles via incubation with NH 4 Cl 2 . By epifluorescence microscopy we did not find any major difference in total vesicle number between    Figure 2F,G). Instead, by restricting our analysis to the TIRF zone (i.e within a range of up to 100 nm away from the plasma membrane) we identified significantly more sypHy clusters in cells expressing LRRK2 G2385R (Fig. 4G,H). We have previously shown that LRRK2 acute silencing correlates with an increase in SV fusion 29 . Western-blotting analysis showed that siLRRK2 transfection decreases endogenous LRRK2 level but does not affect significantly sypHy level ( Fig. 4C and Supplementary Figure 2C,D). Next we verified the functional impact of LRRK2 down-regulation by TIRFM. Interestingly, we noticed that upon silencing of endogenous LRRK2, cells exhibited an increase in fusion events ( Fig. 4D-F) as well as an increase in sypHy clusters within the TIRF zone ( Fig. 4G,H). Collectively, these findings suggest that either LRRK2 silencing or G2385R expression increases the number of fusion events by redistributing vesicles near the membrane. To further prove that G2385R variant behaves as a loss-of-function mutation, we tried to rescue its phenotype by co-expressing isolated WD40 domain. We analysed vesicle fusion in stable N2A line expressing either LRRK2 WT or G2385R transfected with sypHy reporter alone or together with RFP-WD40 construct (Fig. 5A). Western-blotting analysis showed that LRRK2-derived constructs are expressed in a similar extent and do not affect significantly sypHy level (

The G2385R PD-risk variant alters LRRK2-binding properties. The G2385R mutation within LRRK2
WD40 domain has been suggested to interfere with LRRK2 biochemical properties 6,30-32 . We therefore asked whether the G2385R variant impairs LRRK2 interaction with SV. To test this hypothesis, we incubated full-length LRRK2 wild-type or G2385R RFP-fusion protein with purified SV at a nanomolar concentration and tested the extent of binding in a high-speed sedimentation assay. Western blotting analysis using a RFP-specific antibody revealed that LRRK2 G2385R binds less efficiently to purified SV (Fig. 6A,B). The reduced stability of G2385R binding to SV might arise from slight differences between wild type and mutant proteins in terms of affinity/avidity or nature of interactions with SV proteins. In order to address this question we analysed the binding properties of the two LRRK2 WD40 variants coupling a domain-based GST pull down assay to LC-MS/MS. To obtain a relative quantification of the proteins bound to the two domains, we took advantage of mass spectrometry-based label-free quantification, a recent development which allows the identification of specific protein binders 33 . In the initial MS-analysis, 235 proteins were significantly enriched in the GST-WD40 bait condition, compared to the    We measured the extent of LRRK2 and SV binding by ultracentrifugation sedimentation assay. We incubated purified RFP-LRRK2 wild-type and G2385R variant with isolated SV (10 μg protein/sample). Bound RFP-LRRK2 was separated from free RFP-LRRK2 by high-speed centrifugation. We appreciated SV-bound LRRK2 by immunoblotting with anti-RFP antibody. The recovery of SV in the pellet was evaluated based on synaptophysin immunoreactivity. (B) The binding of RFP-LRRK2 wild-type and G2385R to SV was calculated as the ratio of total RFP-LRRK2 and expressed as mean ± SE; n = 6. *p < 0.05; Student's t-test. (C) GST-pull down approach was performed to explore the interactome associated to LRRK2 WD40 domain. (Upper panels) GST-fusion proteins corresponding to WD40 domain of LRRK2 wild type and LRRK2 G2385R were used to retain interactors from adult forebrain lysate. (Lower panel) Ponceau staining shows the abundance of GST fusion proteins used as bait. (D) We evaluated the extent of synapsin I, α-tubulin, β-Actin and 14-3-3 bound to WD40 G2385R domain. Data are expressed as ratio over WD40 wild-type domain. Graph reports mean ± S.E; n = 4. *p < 0.05, Student's T-test. (E) We isolated on streptavidin resin full-length FLAG-LRRK2 wild-type and FLAG-LRRK2 G2385R protein from N2A over-expressing cells. Interacting proteins were resolved by western-blotting. (F) We evaluated the extent of synapsin I, α-tubulin, β-Actin and 14-3-3 bound to the different LRRK2 variant. Data are expressed as the ratio over LRRK2 wild-type. Graphs report mean ± S.E; n = 4. *p < 0.05, Student's T-test.
GST-control. Among these proteins, 95 were significantly reduced in the GST-WD40 G2385R condition when compared to the wild-type. Interestingly, among these differentially bound proteins we found cytoskeletal elements such as tubulin and actin as well as various vesicular proteins, including Rabs, synapsin and 14-3-3 (Table 1  and Supplementary Table 1). By western-blotting, we confirmed that GST-WD40 G2385R domain binds to actin, synapsin, tubulin and 14-3-3 with reduced affinity (Fig. 6C,D). To further validate our findings, we investigated the protein complex associated with full-length LRRK2 wild-type and G2385R variants. To this aim, we purified full length Strep-FLAG LRRK2 wild-type and G2385R variants from N2A cell lines. Eluted proteins were resolved by western-blotting. This approach confirmed that G2385R variant reduces LRRK2 affinity towards synapsin, actin, tubulin and 14-3-3 (Fig. 6E,F). Taken together, these results suggest that structural alterations induced by the G2385R substitution functionally modify LRRK2 binding properties thereby diminishing its interaction with a specific subset of vesicular as well as cytoskeletal proteins.

G2385R variant mimics activity-driven events. Neuronal activity triggers vesicle release via increase
in presynaptic calcium 34 . Thus we triggered exocytosis by increasing bath [Ca 2+ ] to 10 mM in SH-5Y5Y cells expressing either LRRK2 wild-type or G2385R variant (Fig. 7A). Western-blotting analysis showed that in all the conditions studied, LRRK2-derived constructs and sypHy report are expressed in a similar extent (Fig. 7B and Supplementary Figure 3E,F). As expected, we noticed a significant increase in the number of fusion events, area under the curve (Fig. 7C-E) as well as single peak fluorescence intensity (Supplementary Figure 3G) in LRRK2 wild-type expressing cell upon 10 mM CaCl2 incubation. Instead, we did not observe any additional increase in the number of fusion events and area under the curve in cells expressing LRRK2 G2385R variant. All together these data suggest that G2385R variant mimics activity-driven events.

Discussion
By applying TIRFM to SH-SY5Y cells expressing synaptopHluorin we dissected the functional implication of the different LRRK2 domains in the context of SV fusion. We noticed a severe impairment of SV fusion events upon over-expression of a construct containing the isolated LRRK2 WD40 domain while, in contrast, the over-expression of LRRK2 variant lacking the WD40 domain increased the number of event. These data are in agreement with our previous finding describing the exo-endocytic activity in neurons expressing isolated WD40 domain or C-terminal truncated LRRK2 6  exocytosis. In fact, CaV2.1 channels allow the activity-driven increase in local Ca 2+ concentration triggering SV fusion 35,36 . Under resting conditions, SV are tethered by cross-linking via synapsin proteins to the actin cytoskeleton 27,37 . Thus at the presynaptic bouton LRRK2 brings together SV and proteins executing opposite functions: it binds proteins supporting SV fusion (such as β3 Ca V 2.1 channel subunit) as well as those hampering exocytosis (actin and synapsin). Interestingly, we found that LRRK2 interactome is dynamically regulated by activity: in fact, we reported a reduced LRRK2 affinity for synapsin, actin and 14-3-3 upon high calcium stimulation. Our biochemical investigation does not have the necessary temporal and spatial resolution to precisely dissect whether these events take place at the presynaptic bouton upon action potential arrival. However, it is tempting to speculate that LRRK2 acts as a molecular scaffold that integrates neuronal stimuli into defined subsets of protein complexes to modulate SV release. In the basal condition LRRK2 may organize a balanced equilibrium between pro-exocytic and hampering molecules, while, upon stimulation, it may release confining effectors such as actin and synapsin to allow efficient SV fusion. Thus, it is evident that, besides its enzymatic activities, the investigation into the dynamics of the protein complex associated to LRRK2 deserves further studies. Within this context, the C-terminal WD40 domain may play a crucial role. In fact, we have shown that the WD40 domain interacts with SV as well as several presynaptic proteins 6 . Furthermore, a LRRK2 deletion mutant, lacking the entire WD40 domain, fails to induce apoptosis and toxicity concomitant with a lack of kinase activity 30,31 21 . The functional consequences of the G2385R variant remain almost enigmatic so far. It has been shown that this variant has an impact on LRRK2 biochemical properties at different levels such as reduced LRRK2 kinase activity and autophosphorylation in vitro 30 . In silico and electron microscopy-based analyses suggest that the G2385R variant, that replaces a uncharged glycine residue with a charged arginine, induces structural stress via charge-charge repulsion between WD repeats 5 and 6 6 . These changes are predicted to alter the biochemical properties of LRRK2 C-terminal domain. We report here that the G2385R variant affects LRRK2-SV interaction, adding to our previous result on the isolated WD40 domain. By following the "guilt by association" hypothesis 38 we further investigated the G2385R-specific protein interactome. We demonstrated that the G2385R variant binds less efficiently to some key presynaptic proteins, including actin and synapsin. Thus, the G2385R variant impairs LRRK2 binding to these two major SV tethering elements as well as with SV themselves. Interestingly, we reported an increased number of sypHy-positive vesicles within the TIRFM zone, up to 100 nm away from the cell membrane. In the presynaptic bouton, SV are organized in distinct functional pool, namely ready-releasable pool and resting pool. Ready-releasable SV are those recruited first during synapse activity 39 and are distributed without a specific localization within the bouton in a range of 100 nm from the active zone 35 .
Although the SH-5Y5Y cellular model may not fully recapitulate the entire molecular complex organizing SV trafficking in the presynaptic bouton, we can speculate that the G2385R variant partially abrogates the function of LRRK2 in restraining vesicle movement. An altered binding of LRRK2 to synapsin and actin as well as to SV might allow the redistribution of vesicle towards the membrane and, eventually, promote a more efficient fusion. Accordingly, our data suggest that the G2385R variant mimics biochemically and functionally activity driven events. Finally, we observed that LRRK2 acute silencing and LRRK2 G2385R over-expression have a similar impact on the trafficking and distribution of synaptophluorin-positive vesicles in SH-SY5Y cells. In agreement with previous data from Rudenko and colleagues 30 , this finding suggests that the G2385R variant abolishes the physiological activity of LRRK2 and, in consequence, behaves as a partial loss of function mutation.

Conclusions
Our findings show that the nature of the macro-molecular complex formed by LRRK2 is tightly correlated to its physiological and pathological function. Furthermore, our data support the idea that the G2385R variant induces a loss of function rather than a gain-of-function phenotype. This suggests that the pathophysiology underlying the dominant nature of LRRK2-associated PD might be more complex and the functional impact of other PD mutations may need reconsideration.

Methods
Generation of in-house antibodies. For the generation of the novel anti-LRRK2 antibody, a His-GST fusion protein containing the ankyrin domain (amino acids 688-901) of human LRRK2 was generated by cloning of the corresponding cDNA into the pETM30 vector (provided by the EMBL) and recombinantly expressed in E. coli. The soluble protein fraction was purified via the N-terminal 6xHis tag using Ni-NTA Agarose (Qiagen) following standard protocols provided by the manufacturer. Immunization of rats and the generation stable hybridoma were performed in-house (Service Unit Monoclonal Antibodies, Helmholtz-Zentrum München) as described previously 40 . Finally, a rat monoclonal IgG2a clone (LANK/24D8) was established.
Plasmids and transfection. Human LRRK2 was subcloned into pDEST57 (N-terminal red fluorescent protein RFP tag, Invitrogen) and N-terminal Strep-FLAG (SF-TAP) plasmid 41 using the Gateway system (Invitrogen). Full-length RFP and Strep-FLAG G2385R variants were generated by site-directed mutagenesis using the QuikChange mutagenesis kit (Stratagene). GST-LRRK2 WD40 wild-type (WT) or GST-LRRK2 WD40 G2385R and RFP LRRK2 1-2141 (hereinafter termed as LRRK2ΔWD40) were described in 6 . LRRK2 silencing and controls shRNA (respectively siLRRK2 and siControl) were previously described 42 . PolIII and shRNA cassette were cut from the pLVTH vector and subcloned in the pSuper vector co-expressing DsRed fluorescent protein via EcoRI and ClaI digestion. SynaptopHluorin expressing vector was previously described 24 . N2A and SH-SY5Y cells were transfected with the different constructs for 48 hours using Lipofectamine 2000 (Invitrogen). Stable N2A clones expressing LRRK2 wild-type or G2385R variant were isolated upon Neomycin selection (1 mg/ml).
To appreciate co-immunoprecipitation efficiency, we normalized the intensity of the co-immunoprecipitated protein to the amount of LRRK2 variant immunoprecipitated. Pull-down efficiency was judged as amount of preys normalized to bait quantity as monitored by Ponceau staining. Secondary antibodies were prepared in GDB buffer and used for 2 h at room temperature; secondary antibodies include: goat anti-mouse Alexa Fluor 488 (Invitrogen), goat anti-rabbit Cy3 (Jackson Immunoresearch) and Alexa Fluor 555 Phalloidin (Molecular Probes, Life Technologies). Cover slips were mounted with prolonged reagent (Life Technologies) and observed with Zeiss Observer Z1 microscope equipped with an Apotome module. The obtained images provide an axial resolution comparable to confocal microscopy 44,45 . Images were acquired with AxioObserv Z1 microscope equipped with Apotome module using a plan-Apochromat 63×/1.40 Oil objective, pixel size 0,102 μm × 0.102 μm. Colocalization studies were performed on the single plan generated by optical sectioning elaborated by Apotome module. Acquired images were analyzed with ImageJ software using the colocalization analysis plugin to detect colocalization between LRRK2 and interactors.
Synaptic vesicle isolation and binding assay. Highly purified rat SV were obtained through controlledpore glass chromatography 46 . The binding of LRRK2 fusion proteins to SV was carried out using a high-speed sedimentation assay 47 . Briefly, SV (5-10 µmg total protein) were incubated for 1 hour at 0 °C with RFP-LRRK2 WT or RFP-LRRK2 G2385R protein in a buffer containing 220 mM glycine, 30 mM NaCl, 5 mM Tris/HCl, 4 mM Hepes (pH 7.4), 0.22 mM NaN 3 , 0.2 mM PMSF, 2 μg/ml pepstatin and 100 µg/mL of bovine serum albumin (BSA). After the incubation, LRRK2 bound to SV were separated by high-speed centrifugation (400,000 × g for 45 min). Aliquots of the resuspended pellets were subjected to SDS-PAGE and subsequent Western blotting with LRRK2-specific antibodies. The amount of LRRK2 was determined in function of optical density in comparison to known amount of fusion proteins. The recovery of SV, used to correct the amounts of LRRK2 bound to SV, was determined by Western blotting with anti-synaptophysin antibodies.
MS/MS identification. The mass spectrometric analysis was performed as described previously 6 . Each replicate was analyzed as a single run by LC-MSMS using a nano-flow HPLC system (Dionex Ultimate 3000 RSLC, Thermo Scientific) coupled to a Orbitrap XL mass spectrometer (Thermo-Scientific). The Raw data were directly analyzed by the MaxQuant Software ver. 1.5.2.8 (www.maxquant.org) allowing protein identification and label-free quantification 48 with the following Andomeda search-engine parameters: database: Swissprot database mouse subset (v2016-09, 16,811 entries), E. coli K12 subset (v2016-09, 4,252 entries), sequence of the bait (LRRK2) and fusion tag (GST). Carbamidomethylation was set as fixed modification as well as N-terminal protein acetylation as variable modification. As enzyme trypsin was set. The initial mass accuracy (MS) for mass recalibration was set to 20 ppm. For MS/MS spectra, the mass accuracy was set to 0.1 Da. The FDR (false discovery rate) threshold was set to 0.01. Razor-peptides and peptides with N-terminal acetylation or methionine oxidation were included into protein quantification. Only proteins identified by at least two peptides were quantified. Downstream label-free quantification and statistical analysis was carried out following established workflows in Perseus ver. 1.5.4.0 as described in 33 . Briefly, for the analysis, 4 independent samples per condition (GST-control, GST-WD40 and GST-WD40-G2385R) have been considered. Only IDs with at least 3 appearances in one condition have been included. Samples have been normalized by the label free-quantification (LFQ) values corresponding to the GST-tag. Missing values have been replaced by the build-in imputation function of Perseus. Specific interactors for the WD40 domain have been determined by the built-in two sample test function (Student's T-test), followed by a permutation-based FDR (FDR = 0.05, number of randomerizations = 250). For these interactors a comparison of WT v.s. mutant has been performed and the statistical significance has been determined with the same test as described above.
Time lapse microscopy by Total internal reflection fluorescence microscopy (TIRFM). 48 hours after transfection, transfected cells were imaged through a TIRF microscope (Carl Zeiss Inc.) equipped with an Argon laser at 25 °C using a 100 × 1.45 numerical aperture (NA) oil immersion objective as in 25 . Green fluorescence was excited using the 488-nm laser line and imaged through a band-pass filter (Zeiss) onto a Retiga SRV CCD camera. Single-cell imaging under TIRF illumination was carried out at 1 hz for a total of one minute, in a standard saline solution (KRH) containing in mM: 125 NaCl, 5 KCl, 2 CaCl2, 1.2 MgSO4, 1.2 KH2PO4, 5 Glucose, 25 4-(2-Hydroxyethyl)-piperazine-1-ethanesulfonic acid (HEPES) (buffered to pH 7.4) at room temperature (25 °C). Where indicated, cells were incubated in a KRH solution containing 10 mM [CaCl2] for 20 minutes, before recording. Up to ten cells were imaged on each coverslip in at least three independent experiments for each construct. TIRF images were analyzed using Image-Pro Plus Analyser Image Software (Media Cybernetics, Bethesda, MD, USA). A set of automated image processing macro/subroutines was developed on the basis of existing algorithms of the Image-Pro Plus Analyser software (High Pass Gaussian filtering, nearest neighbouring deconvolution). Finally, the images were background subtracted by using the averaged value within a user-defined background region and corrected for photobleaching. The resulting corrected images were then analyzed using an existing Image-Pro Plus plug-in (tracking object) for selection and quantification of fluorescent spots according to their shapes, size and intensity. The following criteria were used to include individual structures in the analysis: (1) mean area 0.039-1 μm 2 , (2) minimal pixel density 20, (3) aspect (major/minor axis) 1-3, (4) velocity limit search radius 1 (micron/frame). Fixed spots (spots present in the same position in at least 20 frames of the movie) were automatically excluded from the analysis. The obtained data were exported in Excel and further analysed. The fluorescence intensity (F) of each spot in the various frames was normalized to its respective initial intensity (F0) and plotted against time. The number of fusion events and the whole cell fluorescence changes were automatically evaluated by a custom written macro. Data, normalized to the cell area, come from at least 15 cells for each construct. To quantify the vesicle density in the TIRF zone, the cells were incubated for 5 minutes with 50 mM NH 4 Cl 2 in KRH solution, to label all synapto-pHluorin positive vesicles. Then, cells were fixed in paraformaldehyde and imaged by TIRFM or epifluorescence. The number of vesicles was quantified as described above.
Statistical analysis and guidelines. All data are expressed as mean ± standard error of the mean (SE).
Data were analyzed with an unpaired Student's t test (two classes) or ANOVA followed by Tuckey's post-hoc test (more than two classes). The indication of number of experiment (n) and level of significance (p) are indicated throughout the text. All methods were performed in accordance with the relevant guidelines and national regulations. All procedures involving animals were approved by Institutional and National Agencies (autorizzazione 793/2016-PR).