Mover is a homomeric phospho-protein present on synaptic vesicles

Abstract

With remarkably few exceptions, the molecules mediating synaptic vesicle exocytosis at active zones are structurally and functionally conserved between vertebrates and invertebrates. Mover was found in a yeast-2-hybrid assay using the vertebrate-specific active zone scaffolding protein bassoon as a bait. Peptides of Mover have been reported in proteomics screens for self-interacting proteins, phosphorylated proteins, and synaptic vesicle proteins, respectively. Here, we tested the predictions arising from these screens. Using flotation assays, carbonate stripping of peripheral membrane proteins, mass spectrometry, immunogold labelling of purified synaptic vesicles, and immuno-organelle isolation, we found that Mover is indeed a peripheral synaptic vesicle membrane protein. In addition, by generating an antibody against phosphorylated Mover and Western blot analysis of fractionated rat brain, we found that Mover is a bona fide phospho-protein. The localization of Mover to synaptic vesicles is phosphorylation dependent; treatment with a phosphatase caused Mover to dissociate from synaptic vesicles. A yeast-2-hybrid screen, co-immunoprecipitation and cell-based optical assays of homomerization revealed that Mover undergoes homophilic interaction, and regions within both the N- and C- terminus of the protein are required for this interaction. Deleting a region required for homomeric interaction abolished presynaptic targeting of recombinant Mover in cultured neurons. Together, these data prove that Mover is associated with synaptic vesicles, and implicate phosphorylation and multimerization in targeting of Mover to synaptic vesicles and presynaptic sites.

Figure 1. Developmental expression of Mover protein in brain.

Rat brain homogenates from the indicated developmental stages between embryonic day 14 (E14) and adult were separated by molecular weight via SDS-PAGE and analyzed by Western blot using affinity-purified Mover antibodies. Tubulin and synaptophysin served as controls. N = 2 experiments.

Figure 2. Mover associates with synaptic vesicle fractions in a membrane floatation assay.

SV fractions (LP2) with and without the addition of 1% Triton X-100 to solubilize membranes were overlaid with a 25–52.5% continuous sucrose gradient and centrifuged. Without detergent, proteins associated with membranes float upwards in the gradient. With the addition of 1% Triton, only proteins attached to detergent-resistant raft-like SV membranes, such as synaptophysin, float into the gradient. Floating of Mover when membranes are intact, and co-floating of Mover with synaptophysin in 1% Triton X-100 treated SV membranes, indicate that Mover is a membrane-associated protein. N = 3 experiments.

Figure 3. Mover is peripherally associated with synaptic vesicle membranes.

Carbonate stripping by treatment of SVs with 100 mM sodium carbonate was used to determine if Mover is integrally or peripherally associated with SV membranes. Synapsin, a peripheral synaptic-vesicle membrane protein served as a positive control and was partially removed from SVs by carbonate treatment, while synaptobrevin2 (syb2), an integral SV protein, was not. Rab3a is membrane-associated but is resistant to carbonate stripping . Mover was partially removed from SV membranes by carbonate stripping, at similar levels as synapsin, indicating that it is a peripheral SV membrane associated protein. N = 2 experiments.

Figure 4. Mover is present on synaptic vesicles.

(A) Mover is enriched in CPG-purified SV fractions; synaptophysin (syp), which is thought to be present exclusively on SVs, serves as a positive control (n = 2 experiments). (B) Mover, along with the positive control synaptophysin, is found in CPG-purified SV fractions analyzed by mass spectrometry (N = 1 experiment). (C) Immunogold labeling of Mover (n = 4 vesicle preparations) and synaptophysin (n = 3 vesicle preparations) in CPG-purified SV fractions imaged by electron microscopy. Scale bars are 50 nm. (D) LS1 fractions (supernatant following hypotonic lysis of synaptosomes) subjected to immunoisolation with antibodies against Mover or synaptophysin. Antibodies conjugated to Protein A beads are indicated above each lane; antibodies used for Western blotting are indicated on the left. The majority of synaptophysin-containing organelles (SVs) are immunoisolated with anti-Mover conjugated beads, and vice versa. The soluble LS1 components lactate dehydrogenas (LDH) and Rab-GDP-dissociation inhibitor (Rab-GDI) are only detected in the supernatant. Control beads without conjugated antibodies bind only trace amounts of SVs (n = 3 experiments).

Figure 5. Mover is present in phosphorylated and non-phosphorylated forms on synaptic vesicles.

(A) Crude synaptosomal fractions were treated with the lambda-protein phosphatase, with and without the addition of the phosphatase inhibitor sodium orthovanadate. In control conditions and in the presence of the phosphatase inhibitor, the anti-phospho-Mover antibody detects a protein band with identical molecular weight to that of non-phosphorylated Mover. This phospho-Mover protein band is reduced in intensity following phosphatase treatment (n = 3 experiments). (B) Phosphorylated Mover, similar to total Mover, is predominantly found in the SV fraction LP2. Indicated fractions are homogenate (H), crude synaptosomal fraction (P2) and corresponding supernatant (S2), synaptosomal fraction (LP1) and corresponding supernatant (LS1), SVs (LP2) and supernatant (LS2), synaptosomal membranes (SPM) and synaptosomal cytosol (Scyt). N = 2 experiments.

Figure 6. Dephosphorylated Mover dissociates from synaptic vesicles.

To analyze a potential influence of phosphorylation on the localization of Mover to SV membranes, SVs were treated with lambda-protein phosphatase and then analyzed for protein content by Western blot. Dephosphorylation caused a shift of Mover immunoreactivity from the pellet to the supernatant, while the immunoreactivity for synaptobrevin and synapsin was unchanged (N = 2 experiments).

Figure 7. Mover does not dissociate from synaptic vesicles in response to depolarization.

(A) Assay of glutamate release from synaptosomes using the fluorescence-based NADPH assay to verify that SVs in synaptosomes undergo calcium-dependent fusion and exocytosis to release glutamate. Depolarization with 50 mM KCl induces glutamate release in the presence of CaCl2, but not of EGTA. (B) Synaptosomal preparations were incubated for 10 min at 37°C in control conditions, in the presence of 1 mM EGTA to chelate calcium, in 1 µM okadaic acid to phosphorylate proteins, or in depolarizing conditions. Following treatment, each synaptosomal fraction was further fractionated to obtain a crude SV fraction. Equal volumes of the crude SV fractions were then subjected to Western blot analysis to test for Mover and synapsin protein levels associated with vesicles. Mover did not dissociate from vesicles in response to depolarization, whereas synapsin did. N = 2 experiments.

Figure 8. Mover is a homomeric protein.

(A) Yeast 2-hybrid assay using full length Mover (full-l.) as bait. A plasmid containing only the LexA interaction domain acted as a control. Beta-galactosidase expression was only induced if full-length Mover was used as both prey (rows) and bait (columns). All deletion constructs used as prey exhibited no interaction with full-length Mover. (B) Immunoprecipitation of GFP-Mover with Mover-myc co-expressed in HEK293 cells. Sepharose-coupled antibodies against the myc epitope were used to pull down protein complexes. As in the yeast 2-hybrid assay, only full-length GFP-Mover was pulled down from the HEK cell extracts, whereas GFP and GFP-Mover deletion constructs were not. N = 2 experiments.

Figure 9. Homomeric interaction of full length Mover in Vero cells.

(A–C) Mover-myc, Mover-GFP and GFP-Mover, each representing full-length versions of Mover, are either diffusely distributed or – in the case of GFP-Mover - as uniformly distributed aggregates, presumably due to some degree of homomerization. Rhodamine-Phalloidin, which stains F-actin associated with cytoplasmic actin-bundles as well as F-actin associated with the sub-plasmalemmal cell cortex, was used to delineate the cell periphery (red color in the merged images and the graphs). (D–F) Upon co-expression with the palmitoylated construct palm-Mover-flag, the constructs are recruited to the plasma membrane. Constructs were immunostained using antibodies against the tags, i.e. myc and GFP (green color in the merged images and the graphs) and flag (red color in the merged images and the graphs). The merged images also show DAPI staining in blue. For line scan analysis (right panels) a bar-shaped region of interest was placed in the image as shown (white bar, representing 5 µm×0.83 µm), and the average fluorescence occurring along its length was plotted in the graph, where 0 µm denotes the end of the bar placed in the extracellular area. Diffusely distributed constructs were characterized by a gradual decline of fluorecence from the cell interior towards the cell periphery (A–C), recruited constructs were characterized by a peak of fluorescence in the cell periphery (D–F). Scale bars are 10 µm. Zooms represent twofold magnification. N = 3 independent cultures.

Figure 10. Localization of Mover deletion constructs in Vero cells.

Upon expression in Vero cells, all variants of GFP-Mover carrying a deletion were diffusely distributed. Immunostaining, Phalloidin staining and line scan analysis were performed as in Figure 9. The line scan fluorescence profiles indicate that all constructs were characterized by a gradual decline of fluorescence towards the cell periphery. Scale bars are 10 µm. Boxes represent twofold magnification. N = 3 independent cultures.

Figure 11. Localization of Mover deletion constructs in Vero cells expressing membrane targeted Mover.

(A,B) Co-expression of GFP-Mover 34–266 or GFP-Mover 52–266 together with palm-Mover-flag recruited the GFP-constructs to the cell membrane. (C–F) Constructs with further truncations were not recruited to the membrane by palm-Mover-flag and thus are deficient in homomeric interaction. Immunostaining and line scan analysis were performed as in Figure 9. The line scan fluorescence profiles show that the fluoresence intensities for GFP-Mover 34–366 and GFP-Mover 52–266, but not of any other construct, co-peak with palm-Mover-flag immunofluorescence in the cell periphery. Scale bars are 10 µm. Boxes represent twofold magnification. N = 3 independent cultures.

Figure 12. A Mover mutant deficient in homomeric interaction does not target to presynaptic sites.

(A,B) Representative images showing recruitment of Mover 52–266-GFP, but not of Mover 91–266-GFP, to the plasma membrane in Vero cells expressing palm-Mover-flag. Scale bars are 5 µm, small zooms represent threefold magnification, large zooms represent sixfold magnification. (C–N) Expression of Mover-GFP, Mover 52–266-GFP and Mover 91–266-GFP in DIV 14 cultured hippocampal neurons. Mover-GFP and Mover 52–266-GFP produced a punctate fluorescence pattern, while Mover 91–266 was diffusely distributed in transfected cells (C,G,K). Note that these images were taken using a 20x objective and displayed using identical brightness settings. The diffusely labelled axons of Mover 91–266-GFP expressing cells were barely visible using these settings. The right box in K shows a brightness-enhanced and mirror-imaged copy of the left box. The boxes in C,G,K indicate areas imaged using a 40x objective and displayed at higher magnification in D–F, H–J and L–N. Immunostaining revealed extensive colocalization of punctate Mover-GFP and punctate Mover 52–266-GFP fluorescence with the synaptic vesicle marker synaptophysin (D–F and H–J). In contrast, Mover 91–266-GFP (L–N) was diffusely distributed. Axonal fluorescence was not enriched at axon-dendrite crossings (axons identified as MAP2-negative, dendrites as MAP2-positive processes. L and M show three axon-dendrite crossings). Scale bars are 20 µm in C, G, K and 10 µm in D, E, F, H, I, J, L, M and N. N = 3 independent cultures.