Synapsins are abundant SV (synaptic vesicle)-associated phosphoproteins that regulate synapse formation and function. The highly conserved C-terminal domain E was shown to contribute to several synapsin functions, ranging from formation of the SV reserve pool to regulation of the kinetics of exocytosis and SV cycling, although the molecular mechanisms underlying these effects are unknown. In the present study, we used a synthetic 25-mer peptide encompassing the most conserved region of domain E (Pep-E) to analyse the role of domain E in regulating the interactions between synapsin I and liposomes mimicking the phospholipid composition of SVs (SV–liposomes) and other pre-synaptic protein partners. In affinity-chromatography and cross-linking assays, Pep-E bound to endogenous and purified exogenous synapsin I and strongly inhibited synapsin dimerization, indicating a role in synapsin oligomerization. Consistently, Pep-E (but not its scrambled version) counteracted the ability of holo-synapsin I to bind and coat phospholipid membranes, as analysed by AFM (atomic force microscopy) topographical scanning, and significantly decreased the clustering of SV–liposomes induced by holo-synapsin I in FRET (Förster resonance energy transfer) assays, suggesting a causal relationship between synapsin oligomerization and vesicle clustering. Either Pep-E or a peptide derived from domain C was necessary and sufficient to inhibit both dimerization and vesicle clustering, indicating the participation of both domains in these activities of synapsin I. The results provide a molecular explanation for the effects of domain E in nerve terminal physiology and suggest that its effects on the size and integrity of SV pools are contributed by the regulation of synapsin dimerization and SV clustering.
Small SVs (synaptic vesicles) are the central organelle at the synapse. SVs contain quanta of NT (neurotransmitter) to be released into the synaptic cleft, undergo cycles of exo-endocytosis and are distributed into at least two functionally distinguishable pools within the axon terminals, namely the RRP (readily releasable pool), made of SVs actively involved in exocytosis, and the RP (reserve pool), which represents a reserve of NT to sustain transmission during periods of intense activity . Several proteins participate in the regulation of SV cycling at the synapse, and the clarification of the molecular mechanisms underlying SV trafficking are fundamental for the understanding of synaptic function and plasticity.
The proteome of SVs is known in great detail and consists of several copies of a surprising diversity of trafficking proteins, some of which are entirely SV-specific . A family of SV-associated phosphoproteins playing a key role in the maintenance of SV pools and in the regulation of NT release are the Syns (synapsins; [3,4]). Syns bind reversibly to the cytoplasmic surface of SVs through interactions with both SV protein and lipid components [5–8] and are believed to tether SVs to each other and to actin filaments [9–11]. Syns are encoded by three distinct genes (SYN1, SYN2 and SYN3) in mammals, and alternative splicing of these genes gives rise to multiple different isoforms. Syns are major substrates for several kinases, including CaMK (Ca2+/calmodulin-dependent protein kinase) I/II/IV, PKA (protein kinase A), ERK (extracellular-regulated protein kinase), Cdk5 (cyclin-dependent kinase 5) and Src [12,13]. Phosphorylation differently affects the ability of Syns to bind SV and actin [11,14–18] and serves as an important functional switch to regulate SV trafficking.
In addition to phosphorylation, oligomerization appears to be fundamental in mediating Syn functions. Syns have been reported to form homo- and hetero-dimers and tetramers via a dimerization surface which is not overlapping with the phospholipid-binding domains and is modulated by ATP binding [8,19–21], and dimerization appears to be required for correct targeting of some Syn isoforms to the axon terminal . Moreover, Syns have been proposed to maintain SV integrity and uniform size by coating the cytoplasmic surface of SVs by virtue of their high surface activity [23,24]. Direct visualization of the Syn coat on planar lipid bilayers and the demonstration of stabilizing activity of Syn on SV membranes were reported using AFM (atomic force microscopy) and phosphorus NMR [10,25–27].
Syns are composed of individual and common domains (Figure 1). The N-terminal portion of all Syns is highly conserved and consists of domains A to C, whereas the C-terminal portion is more variable and consists of a combination of domains D to J. Among the latter domains, domain E is the most conserved and is shared by all a-type Syn isoforms [3,4]. Domain E has been reported to play important roles in Syn function. Injection of a specific antibody against domain E in the lamprey reticulo-spinal synapses , as well as injection of a synthetic 25-mer peptide encompassing the most conserved region of domain E (Pep-E) in the squid giant synapse [29,30], or overexpression of the same peptide in cerebellar Purkinje cells  dramatically decreased the number of SVs in the RP, leaving the size of the RRP substantially unaffected. Moreover, Syn Pep-E also seems to participate in post-docking functions by regulating the kinetics of NT release and synaptic depression [29,31]. In hippocampal neurons, domain E, in tandem with domain C, is a positive determinant for the correct synaptic targeting of Syn Ia , and biochemical studies revealed that Pep-E selectively competes for the binding of Syn I to F-actin .
Schematic representation of the domain structure of the mammalian Syn family
To understand the molecular mechanisms underlying these effects, we analysed the role of domain E in the interactions between Syn I and liposomes mimicking the phospholipid composition of SVs (SV–liposomes), as well as between Syn I and other pre-synaptic protein partners, by using affinity chromatography, cross-linking, FRET (Förster resonance energy transfer) and AFM using synthetic Pep-E. The results demonstrate that domain E participates in the formation of Syn I dimers and that inhibition of domain E-mediated oligomerization disrupts the ability of Syn to coat and aggregate SV–liposomes.
Antibodies to VAMP2 (vesicle-associated membrane protein 2), SNAP25 (25 kDa synaptosome-associated protein), Rab3A, syntaxin and synaptotagmin were obtained from Synaptic Systems. Antibodies against Syn I-(10–22) and Syn II-(19–21) were raised in our laboratory. The antibody against Syn III (RU-486) was a gift from Dr H.T. Kao (Division of Biology and Medicine, Brown University, Providence, RI, U.S.A.). Bovine brain PC (phosphatidylcholine), bovine brain PS (phosphatidylserine), bovine brain PE (phosphatidylethanolamine), bovine liver PI (phosphatidylinositol), NBD-PE [N-(4-nitrobenzo-2-oxa-1,3-diazole)-L-α-phosphatidylethanolamine] and LRh-PE [N-(lissamine rhodamine B sulfonyl)-L-α-phosphatidylethanolamine] were obtained from Avanti Polar Lipids, stored at −20 °C in the dark and used within 3 months. Streptavidin–Sepharose beads were from GE Healthcare Life Sciences. All other chemicals were from Sigma.
The peptide SLSQDEVKAETIRSLRKSFASLFSD [rat Syn Ia-(680–704); Pep-E], its scrambled version STESLKVEDSLSAKFRAISRFLQSD (ScrPep-E), the rat peptide corresponding to the squid C5 peptide  AYMRTSVSGNWKTNTGSAMLEQI [rat Syn I-(325–347); Pep-C] and its scrambled version STWTEAINLSGMYTMKGVNQRAS (ScrPep-C) were manually synthesized using the standard method of solid-phase peptide synthesis, which follows the Fmoc (fluoren-9-ylmethoxycarbonyl) strategy with minor modifications . To obtain biotinylated peptides (BiotPep-E and BiotScrPep-E respectively), either Pep-E or SrcPep-E was labelled by direct coupling of biotin on the solid support to the N-terminal of the protected peptide sequences during the synthesis. Synthesized compounds were purified by RP-HPLC (reversed-phase HPLC) on a Shimadzu LC-9A preparative HPLC equipped with a Phenomenex C18 Luna column (21.20 mm×250 mm). Peptide quality and purity were confirmed by HPLC–MS/MS (tandem MS) analysis (results not shown).
Preparation of extracts of purified rat brain synaptosomes
Subcellular fractionation of the rat cerebral cortex was performed as described previously . Highly purified synaptosomes were obtained by centrifugation of the resuspended crude synaptosomal pellets (P2 fraction) at 32500 g for 5 min on a discontinuous four-step 3–23% Percoll gradient and by collecting the third and fourth interfaces, as described previously . The synaptosomal fraction was then resuspended in Binding Buffer [150 mM NaCl, 50 mM Tris/HCl (pH 7.4), 0.1% SDS, 1% Nonidet P40, 0.5% sodium deoxycholate, 200 mM PMSF, 2 μg/ml pepstatin and 1 μg/ml leupeptin] for 1 h on ice and was extracted by centrifugation at 400000 g for 1 h in a Beckman TL-100 ultracentrifuge. The soluble fraction was then immediately subjected to affinity chromatography (see below).
Affinity chromatography with Biot Pep-E
Either BiotPep-E or BiotScrPep-E (250 μg) was immobilized with 400 μl of settled pre-equilibrated Streptavidin–Sepharose beads. After an overnight adsorption at 4 °C under gentle rotation in small columns, the samples were incubated with freshly prepared synaptosomal extracts (2–3 mg of protein). After an overnight incubation at 4 °C, columns were packed and extensively washed with Binding Buffer, followed by a final wash with detergent-free Binding Buffer. Elution of the bound proteins was performed with 5 mM biotin in detergent-free Binding Buffer. Samples were then separated by SDS/PAGE and analysed by silver staining, followed by MALDI (matrix-assisted laser-desorption ionization)–TOF (time-of-flight)-MS or by immunoblotting.
MALDI–TOF MS analysis
The excised bands were reduced, alkylated and digested overnight as described previously . Proteins were identified by MALDI–TOF peptide mass mapping. A portion (1 μl) of the supernatant of the digestion was loaded on to the MALDI target using the dried-droplet technique and α-cyano-4-hydroxycinnamic acid as the matrix. MALDI MS measurements were performed on a Voyager-DE STR TOF mass spectrometer (Applied Biosystems). Spectra were accumulated over a mass range of 750–400 Da with a mean resolution of approx. 15000. Spectra were internally calibrated using trypsin autolysis products and processed via Data Explorer software version 22.214.171.124 (Applied Biosystems). Alkylation of cysteine by carbamidomethylation and oxidation of methionine were considered as fixed and variable modifications respectively. Two missed cleavages per peptide were allowed, and a mass tolerance of 50 p.p.m. was used. Peptides with masses corresponding with those of trypsin and the matrix were excluded from the peak list. Proteins were identified by searching against MSDB [MS protein sequence database; MSDB 20050701 restricted to mammalia (292819 sequences)] using the MASCOT algorithm . Probability-based protein identification by searching sequence databases using MS data was performed .
Syn I cross-linking assays
The formation of Syn I dimers was assessed by cross-linking experiments using the catalytic oxidation of sulfhydryl groups by the o-phenantroline/copper complex . Briefly, 1 μg of purified Syn I solubilized in 200 mM NaCl and 20 mM NaPO4 (pH 7.4) was incubated with 0.5 mM CuSO4 and 1 mM o-phenanthroline for 40 min at room temperature (22 °C) in the presence or absence of Pep-E, ScrPep-E, Pep-C or ScrPep-C in various combinations. At the end of the incubation period, the reaction was stopped by the addition of 50 mM EDTA followed by 2-mercaptoethanol-free SDS sample buffer. The cross-linked complexes were separated by SDS/PAGE on 6% (w/v) polyacrylamide gels under non-reducing conditions, transferred on to nitrocellulose membranes and detected by immunoblotting with an anti-Syn I antibody.
SV–liposomes aggregation assays
Phospholipid vesicles mimicking the phospholipid composition of SVs (PC/PE/PS/PI/cholesterol, 40:32:12:5:10, by vol.) were made by sonication as described previously . SV–liposomes were resuspended in buffer A [150 mM NaCl, 3 mM NaN3 and 10 mM Hepes/NaOH (pH 7.4)]. Fluorescently labelled SV–liposomes had the same lipid composition with the addition of the appropriate amounts (2% of the total lipid, w/w) of NBD-PE or LRh-PE either alone (single-labelled liposomes) or in combination (double-labelled liposomes). Changes in Syn-induced vesicle aggregation in the presence of either Pep-E or ScrPep-E were followed by analysing FRET between the energy donor NBD-PE and the acceptor LRh-PE, as described previously . Syn I (100 nM)-induced aggregation in the presence or absence of various concentrations of Pep-E or ScrPep-E was followed by monitoring FRET at 22 °C using a PerkinElmer Life Sciences LS-50 spectrofluorometer equipped with a thermostated sample holder. In the first assay (aggregation/fusion assay), the fluorescence donor and acceptor were incorporated into separate vesicle populations then the two populations of vesicles (100 μg of phospholipid for each population) were mixed, and FRET was measured by exciting the donor at 470 nm and following either the decrease in NBD emission at 520 nm (NBD quenching) or the increase in LRh emission at 590 nm (excitation and emission slits of 2.5 and 5 nm respectively). In the second assay (fusion assay), one population of vesicles containing the same amount of both fluorophores (50 μg of phospholipid, 2% labelled phospholipids) was mixed with unlabelled vesicles (150 μg of phospholipid). Membrane fusion, leading to intermixing of labelled and unlabelled membrane components, results in a decrease in the surface density of donor and acceptor fluorophores and, thereby, in a decrease in FRET between NBD and LRh. Under these conditions, pure aggregation of vesicles not accompanied by fusion is silent, but can be evaluated by an enhancement in the rate and extent of vesicle fusion induced by the addition of Ca2+ (3 mM). Alternatively, aggregation of unlabelled liposomes was evaluated by a light scattering/turbidity assay as described previously . Samples containing 20 μg of phospholipids were incubated in 150 μl of buffer A in the absence or presence of Syn I (200 nM) and/or the various peptides. Turbidity was measured by reading the absorbance at 325 nm as a function of time at 22 °C in a DU-8 spectrophotometer (Beckman Coulter) equipped with a thermostated sample holder.
All AFM measurements were taken using an Agilent Technologies 5400 system equipped with a 90 μm×90 μm scanner and a low-coherence laser to avoid interference of the laser beam during force–distance curves. Images were acquired in contact mode (512 pixels×512 pixels) using triangular silicon nitride cantilevers (MLCT, Microlever type; Veeco) with a nominal spring constant of 0.06 N/m. AFM samples were prepared by using a mica substrate. The solution containing the phospholipids or the Syn/peptide was gently added directly into the AFM liquid cell without removing the mica substrate by using a syringe and small Teflon tubes, while keeping the same imaged area during addition. For lipid bilayer preparation, 200 μl of a 10 mg/ml solution of phospholipid vesicles (prepared as for the vesicle aggregation assay in the absence of labelled compounds) in saline buffer were deposited on to freshly cleaved mica. After an overnight incubation at room temperature (under a water-saturated atmosphere to prevent evaporation of the solution), the cell was rinsed with buffer solution to eliminate excess phospholipids. The presence of a uniform bilayer on the mica surface was assessed by AFM imaging and indentation. As reported previously , the AFM image of lipid bilayers showed a flat uniform surface which could be pierced by the AFM tip in the course of force–distance curves, resulting in a highly reproducible and characteristic jump of few nanometres (corresponding to the thickness of the double layer) at a force of approx. 10–20 nN. Each sample was checked for the presence of the double lipid layer before proceeding with the addition of Syn and/or peptides.
Syn I was purified from bovine brain as described previously . Protein concentration assay, SDS/PAGE, electrophoretic transfer of proteins on to nitrocellulose membranes (Schleicher & Schuell) and immunoblotting were carried out according to standard procedures, with an ECL (enhanced chemiluminescence) detection system being used.
Pep-E binds to Syns
To understand the molecular mechanisms by which domain E exerts its activity, affinity-chromatography experiments from nerve terminal extracts were performed to reveal putative interacting synaptic proteins using a peptide (Pep-E) encompassing the most conserved region of domain E. Detergent extracts of Percoll-purified rat forebrain synaptosomes were loaded on to BiotPep-E columns, and bound proteins were eluted with biotin and analysed by SDS/PAGE. Silver staining showed that several proteins bound specifically to BiotPep-E columns, but not to the control columns in which ScrPep-E was immobilized (Figure 2A). MALDI–TOF-MS revealed that Syn Ia was bound to Pep-E columns (indicated by arrow 1 in Figure 2A and corresponding to gi|6686305|sp|P09951|SYN1_RAT Syn I; matched peptides eight out of 18 measured peptides, sequence coverage 18%, score 73). Other proteins were detected in the eluates by MS, including tubulin β 2B chain and ATP synthase α chain (Figure 2A). However, except for Syn Ia, the other bound proteins were neither synaptic nor enriched in brain tissue and their relevance remains elusive. To confirm the interaction between Pep-E and Syn I and investigate other possible interactions of domain E with synaptic proteins which might have escaped detection by silver staining, we performed immunoblot analysis on affinity-chromatography samples using antibodies against specific Syn isoforms, SNARE (soluble N-ethylmaleimide-sensitive fusion protein-attachment protein receptor) proteins and Rab3A. Both Syn Ia/Ib and Syn IIa/IIb bound to the Pep-E, but not to the control column, in contrast with Syn III or the other tested pre-synaptic proteins, which did not bind to immobilized Pep-E (Figure 2B).
Pep-E interacts with Syns I and II
Pep-E inhibits Syn dimerization
Syn I is known to form homodimers or heterodimers with Syn II that are believed to play a role in SV clustering and trafficking [10,20,38]. Domain E has been reported to interfere with the ability of Syn I to bind and bundle actin filaments , and bundling is believed to be contributed by self-association of Syn molecules . Therefore we analysed the role of domain E in Syn dimerization using the copper/phenantroline method to covalently link Syn dimers  in the presence of Pep-E or ScrPep-E. Syn monomers and dimers were separated by non-reducing SDS/PAGE and visualized by immunoblotting (Figure 3A, upper panel). Upon exposure to the cross-linking agent, a significant fraction of Syn I formed dimers of 160–170 kDa molecular mass. Interestingly, Pep-E significantly inhibited Syn dimer formation in a concentration-dependent manner with an IC50 of 180±9 μM, whereas the same concentrations of ScrPep-E were virtually ineffective (Figure 3A, lower panel). As structural studies performed on recombinant Syn C domain or ABC domains demonstrated that the formation of Syn dimers and tetramers is mediated by stretches of domain C [19,21], we investigated whether peptides belonging to the C and E domains (Pep-C and Pep-E respectively) have a synergistic action in inhibiting Syn dimerization. Similar to Pep-E, Pep-C dose-dependently inhibited the formation of Syn dimers, with a higher potency (IC50 of 61±5 μM), whereas its scrambled version was ineffective. However, when Pep-C was tested in the presence of a Pep-E concentration corresponding to the IC50, the association of the two peptides did not further increase the inhibition of Syn dimerization with respect to the effect of either peptide (Figure 3B).
Either Pep-E (A) or Pep-C (B) is sufficient to inhibit the formation of Syn dimers
Pep-E inhibits the formation of Syn islands on phospholipid bilayers
We used AFM to directly observe the effect of the domain E peptide on the oligomerization of Syn I bound to membranes mimicking the phospholipid composition of native SVs. Consistent with previous observations , a time-dependent formation of protein clusters on top of the phospholipid surface was observed after the injection of Syn I. A representative AFM image taken 5 min after the in situ addition of 10 μM Syn I (Figures 4A and 4C) shows the formation of extended (1–2 μm) ‘islands’ of approx. 2.1 nm in height (Figures 4D and 4E). Force–distance curves performed on top of the Syn I ‘islands’ did not show the characteristic jump into contact, suggesting that Syn I makes the underlying lipid bilayer more stable than the bare bilayer, and prevents the tip from penetrating the phospholipid membrane. The effect of Pep-E on the Syn I aggregation properties was tested by adding increasing concentrations of either the peptide or its scrambled version immediately before the addition of Syn I. A typical AFM image obtained 5 min after the in situ addition of 10 μM Syn on a bilayer sample pre-treated with 250 μM Pep-E is shown in Figure 4(B). The islands formed in the presence of Pep-E were markedly smaller and with different shapes, while their height (2.1±0.4 and 1.9±0.4 nm for Syn I and Syn I+Pep-E respectively) was not affected (Figures 4F and 4G). When the concentration of Syn I was increased to 25 μM, a very large and confluent protein plateau was formed that appeared almost homogeneous over a 3 μm×3 μm scale (Figure 5A). However, when Pep-E was added immediately before Syn I (Figure 5B), it significantly altered the ability of Syn I to form confluent ‘islands’, whereas ScrPep-E had only a small effect (Figure 5C). When a standard flooding algorithm was used to quantify the percentage of ‘rised’ regions, the area of Syn bilayer coverage obtained at 25 μM Syn I was strongly decreased in the presence of Pep-E (250 μM), but was affected to a lesser extent by the scrambled version of the peptide (Figure 5D).
Pep-E alters the Syn I coverage of phospholipid bilayers
Pep-E inhibits the formation of Syn islands on phospholipid bilayers
Pep-E inhibits the Syn-induced aggregation of phospholipid vesicles
Syn I was shown to cluster phospholipid vesicles mimicking the composition of SVs . Thus we investigated whether Pep-E was able to interfere with the Syn-induced SV–liposome clustering by using fluorometric assays sensitive to either vesicle aggregation (followed or not by fusion) or fusion only (see the Experimental section). Both assays confirmed that Syn I induces vesicle clustering without triggering fusion, as the addition of Syn I readily increased FRET in the aggregation assay (Figure 6), whereas it potentiated the fusogenic effect of Ca2+ in the fusion assay (Figure 7). Pep-E and ScrPep-E were devoid of any effect on vesicle aggregation and fusion by themselves (Figures 6B and 7B). Interestingly, Pep-E virtually abolished the effect of Syn I on vesicle aggregation, while its scrambled version was ineffective (Figures 6A and 6B). Consistent results were obtained in the fusion assay, in which Pep-E, but not its scrambled version, blocked the potentiating effect of Syn I on the fusogenic effect of Ca2+ in a concentration-dependent manner (Figures 7A and 7B). We also studied whether Pep-C and Pep-E have a synergistic action when tested in the inhibition of phospholipid vesicle clustering evaluated by a turbidity/light scattering assay . As observed in the dimerization experiments, either peptide inhibited the Syn I-induced vesicle clustering (IC50≅40 and 62 μM for Pep-C and Pep-E respectively). Moreover, when increasing concentrations of Pep-C were incubated in the presence of a fixed concentration of Pep-E (200 μM), the two peptides had an occlusive effect, suggesting that both C and E domains participate in vesicle clustering (Figure 8).
Pep-E inhibits the Syn-induced aggregation of phospholipid vesicles
Pep-E inhibits the Syn-induced potentiation of Ca2+-triggered vesicle fusion
Either Pep-E or peptide C are able to inhibit the Syn I-induced phospholipid vesicle aggregation
Syn domain E has been implicated in mediating key Syn functions such as maintenance of the reserve pool and regulation of the kinetics of exocytosis in distinct experimental systems. Injection of anti-domain E antibodies into the lamprey reticolo-spinal axons disrupted the RP of SVs upon activity . On the other hand, injection of Pep-E into the terminal of the squid giant axon or overexpression of Pep-E in cerebellar Purkinje cells of L7-transgenic mice dramatically decreased the RP of SVs and perturbed the kinetics of exocytosis [29,31]. However, the molecular mechanisms by which domain E exerts its function remain largely unresolved.
Previous findings have suggested that the physiological effects of Syn domain E are generated through a perturbation of the interactions of Syn I with F-actin. Actin is known to be a fundamental player linking the regulation of SV cycling and trafficking between functional SV pools to synaptic efficacy . It has been demonstrated that Pep-E, as well as a highly conserved peptide from domain C characterized by similar physiological effects, competitively inhibit the Syn–actin interactions . Thus their physiological effects could be ascribed to an inhibition of SV tethering to the actin cytoskeleton which is believed to regulate size and maintenance of the RP [11,41–43]. However, the physiological role of actin in the regulation of SV pool size and trafficking was recently questioned . Actin filaments were shown to be less abundant inside the SV clusters than near the plasma membrane at the endocytic periactive zone and around the SV clusters [45–48]. Moreover, at the giant lamprey synapse disruption of the actin cytoskeleton led to a depletion of the RP of SVs upon activity, by impairing the transport of SVs from the endocytic periactive zone to the SV cluster and decreasing the cluster size upon stimulation , whereas at hippocampal synapses it had no significant effects on the SV pool size . The characteristic distribution of F-actin at pre-synaptic sites, around rather than within the SV cluster, suggests a scaffolding function of the actin meshwork, rather than a direct role in clustering . Interestingly, Syns have been reported to disperse from the SV cluster to the endocytic zone during synaptic activity [45,49]. Taken together, these data suggest that Syn–actin interactions may be more important for the regulation of SV trafficking between distinct pools than for the maintenance of the RP and for SV clustering. Accordingly, the Syn domain E–actin interaction may not be the sole mechanism by which domain E exerts its functions.
In the present paper, we addressed the possibility that domain E contributes to SV clustering through additional interactions within the nerve terminal. Our data indicate that the Syn domain E: (i) binds to Syn I and Syn II; (ii) mediates Syn dimerization and participates in the stabilizing effect of Syn on the phospholipid bilayer; and (iii) is involved in Syn-mediated SV clustering.
Syn oligomerization has been demonstrated in living cells expressing distinct Syn isoforms where Syn I/II and Syn II/III heterodimers, in addition to homodimers, were detected . The interaction of domain E with Syn I and Syn II, but not Syn III, detected by affinity chromatography in the present paper is in full agreement with the yeast two-hybrid screens which found Syn I and Syn II in the large majority of the specifically isolated prey clones using a domain E-bearing Syn IIa bait . The fact that Syn III was excluded from the interaction with Pep-E is in line with the low expression levels of this isoform in mature nerve tissue  and with previous data demonstrating that Syn III is not able to heterodimerize with Syn I when overexpressed in COS7 cells . The involvement of domain E in the formation of Syn dimers also implicates this domain in the specific targeting of Syns to synaptic terminals. Domain E was identified as a positive determinant of nerve terminal targeting and dimerization was reported to be crucial for bringing to synaptic terminals Syn isoforms with weak targeting potential, such as Syn Ib .
The peptide competition experiments indicate that domain E-mediated Syn dimerization plays a key role in the ability of Syn to coat phospholipid bilayers and to maintain their uniform size and structural integrity [10,25]. As Syn I represents 4% of SV protein, it is thought to cover large portions of the SV surface [23,25] and by this means to have a role in maintaining the uniform shape of SV, as well as in preventing random fusion events. The antagonizing effect of Pep-E on the Syn-mediated stabilizing effect of the phospholipid bilayer and on the ability of Syn to dimerize and cluster SVs suggests that the maintenance of SV integrity and the prevention of random fusion events are causally linked to Syn oligomerization on the SV membrane.
Syn dimerization is thought to play a key role in SV clustering and stabilization as the result of the binding of Syn homo- or hetero-dimers to adjacent SVs. The crystal structure of Syn domain C suggests the presence of a dimerization surface  which does not overlap with the membrane-binding domains , and is therefore a good candidate for mediating SV clustering. Several observations indicate that peptides encompassing critical regions of domain C and domain E (Pep-C and Pep-E) alter synaptic function and SV clustering in a similar fashion, suggesting that these two domains directly or indirectly mediate the same interactions of Syns. In fact, both C and E peptides, injected into the pre-terminal region of the squid giant axon disassemble SV clusters, slow down the kinetics of release and enhance synaptic depression in response to high-frequency stimulation [29,30]. Indeed, our data support a concomitant involvement of domain E and domain C in the dimerization and vesicle-clustering processes. The occlusive effects displayed by the combination of the two peptides, and the observation that each peptide is able to disrupt Syn dimers and vesicle clusters indicate that, under native conditions, stretches belonging to both domain C and domain E are involved in the oligomerization of holo-Syn I and that both interaction sites are required to stabilize the Syn dimer and form the Syn lattice on phospholipid membranes. It is tempting to speculate that either peptide may lead to destabilization of the Syn conformation involved in the formation of oligomers. Analysis of holo-Syn crystals formed in the presence of the peptides would be required to prove this hypothesis. It is also possible that, although the maximal effect reached in the presence of a single peptide does not exceed the effect obtained in the presence of both C and E peptides, the presence of both C and E domains may allow a more rapid kinetics of dimerization and clustering in vivo.
In conclusion, although unidentified protein partners and/or regulation of Syn–actin interactions may potentially contribute to its physiological effects, the Syn domain E appears to be directly involved, together with domain C, in the formation of Syn oligomers and in the ability of Syns to form and maintain the SV clusters of the RP. Thus it is tempting to speculate that the dramatic depletion of the distal SV pool observed in lamprey reticulo-spinal synapses after injection of anti-domain E antibody , in squid giant synapse after injection of Pep-E  or in Purkinje cell terminals overexpressing Pep-E  can be ascribed to a major impairment of Syn oligomerization caused by interference with the interactions mediated by the endogenous domain E.
atomic force microscopy
Förster resonance energy transfer
N-(lissamine rhodamine B sulfonyl)
matrix-assisted laser-desorption ionization
readily releasable pool
vesicle-associated membrane protein
Ilaria Monaldi performed most of the affinity-chromatography, dimerization and liposome aggregation/fusion experiments. Massimo Vassalli and Roberto Raiteri performed all of the AFM experiments. Angela Bachi performed the mass spectrometry analysis. Silvia Giovedì performed the dimerization experiments with peptide C. Enrico Millo synthesized all of the peptides used in the study. Flavia Valtorta contributed to data interpretation and preparation of the manuscript. Fabio Benfenati supervised and financially supported the whole work, and was mainly involved in the analysis and interpretation of the results and in the writing of the paper. Anna Fassio supervised the experiments, analysed the experimental data, wrote the paper and provided partial financial support for the work.
We thank Dr Paul Greengard (The Rockefeller University, New York, NY, U.S.A.), Dr Hung-Teh Kao (Brown University, Providence, RI, U.S.A.) for the gift of anti-Syn III antibodies and invaluable discussions, Michele Zoli (University of Modena and Reggio Emilia, Modena, Italy) for statistical analysis and helpful discussions, and Fabia Filipello, Franco Onofri and Annalisa Furlan (University of Genova and Italian Institute of Technology, Genova, Italy) for their valuable help.
This work was supported by the Ministry of the University and Research [grant number PRIN 2006 to A.F. and F.B] and Compagnia di San Paolo (to F.B, F.V. and A.F.). The support of Telethon-Italy [grant numbers GCP05134 and GCP09134], Cariplo Foundation and Fondazione Pierfranco and Luisa Mariani grants (to F.B. and F.V.) is also acknowledged.