The Calmodulin Binding Region of the Synaptic Vesicle Protein Mover Is Required for Homomeric Interaction and Presynaptic Targeting

Abstract

Neurotransmitter release is mediated by an evolutionarily conserved machinery. The synaptic vesicle (SV) associated protein Mover/TPRGL/SVAP30 does not occur in all species and all synapses. Little is known about its molecular properties and how it may interact with the conserved components of the presynaptic machinery. Here, we show by deletion analysis that regions required for homomeric interaction of Mover are distributed across the entire molecule, including N-terminal, central and C-terminal regions. The same regions are also required for the accumulation of Mover in presynaptic terminals of cultured neurons. Mutating two phosphorylation sites in N-terminal regions did not affect these properties. In contrast, a point mutation in the predicted Calmodulin (CaM) binding sequence of Mover abolished both homomeric interaction and presynaptic targeting. We show that this sequence indeed binds Calmodulin, and that recombinant Mover increases Calmodulin signaling upon heterologous expression. Our data suggest that presynaptic accumulation of Mover requires homomeric interaction mediated by regions distributed across large areas of the protein, and corroborate the hypothesis that Mover functionally interacts with Calmodulin signaling.

Keywords: TPRGL; calmodulin; mover; presynaptic; synaptic vesicle.

Figure 1

Subcellular distribution of recombinant Mover variants in cultured hippocampal neurons. (A) Schematic depicting the proteins tested. The numbers indicate amino acids in the primary structure of rat Mover. The black bar indicates the location of the Calmodulin (CaM) binding region. All proteins are tagged with monomeric EGFP (mGFP). (B–G) Inverted gray level epifluorescence images representing the localization of GFP-tagged rat Mover constructs in DIV14 rat hippocampal neurons. The square panels on the right represent zooms of the boxes. Mover-mGFP and 52-266-mGFP are known to accumulate at synapses and produce the corresponding punctate staining pattern. All other variants are homogeneously distributed throughout the cytoplasm, indicating that they fail to accumulate at synapses. A 20× objective was used for the overviews on the left, in order to capture an entire neuron. The exposure time was set to visualize fluorescence in the neurites. Using these setting, nuclear and somatic fluorescence is saturated, because these compartments have much bigger volume than neurites. The distribution of each construct was determined in more than 30 neurons on a total of six coverslips from three independent experiments.

Figure 2

Mover variants that fail to target fail to undergo homomeric interaction with full-length Mover. The figure shows Western blots analyzing the results of immunoprecipitation of Mover variants expressed in HEK293 cells. (A–H) Each GFP-tagged Mover variant (indicated on top of each panel) was separately co-expressed with myc-tagged Mover. Sepharose-coupled antibodies against the myc epitope were used to pull down protein complexes. The 30 kDa region of the blot was probed with anti-Mover, to verify immunoprecipitation of Mover-myc, which runs as a double band. Panels (A,C) are controls where the GFP-tagged constructs were expressed without Mover-myc, showing that the GFP tagged constructs expressing amino acids 1-266 and 52-266 display only residual binding to Sepharose beads. The arrowheads indicate degradation products of the GFP-Mover constructs detected by the Mover antibody. “Lysate” indicates samples obtained before adding anti-myc antibodies, “IP” indicates samples of the immunoprecipitation pellet. The regions of the blot corresponding to the molecular weights of the GFP-tagged proteins were probed with anti-GFP. Only full-length Mover-mGFP (“1-266”) and the variant lacking the amino terminal 51 amino acids (“52-266”) co-immunoprecipitate with Mover-myc. All others fail to co-immunoprecipitate. The data represent two independent experiments, i.e., two HEK293 cell transfections followed by lysis and co-immunoprecipitation.

Figure 3

Homomeric interaction of recombinant Mover in cultured hippocampal neurons revealed by FRET analysis. (A) Schematic representation of the constructs and the result of FRET analysis. A construct encoding amino acids 52-266 of Mover was tagged either at its N-terminus or at its C-terminus with the fluorescent proteins indicated. (B) Representative intensity images (top row, gray scale) and FRET images overlayed with the corresponding intensity images (bottom row, false color). (C) FRET distribution per condition. n = 5 regions of interest (ROIs), N = 2 independent experiments, two-sided student’s t-test, ***p ≤ 0.001. Scale bar is 10 μm.

Figure 4

CaM-binding properties of wildtype (WT) and mutated Mover peptides. (A) Amino acid sequences of Mover(203-221), the photoreactive variant Bpa-Mover(203-221), the point mutant R-Mover(203-221) carrying an Arg residue instead of the hydrophobic anchor residue Phe, and the CaM-binding deficient R/E-Mover(203-221) carrying multiple mutations of basic residues within the CaM-binding sequence in addition. The sequence positions for amino acid exchanges are boxed. f, p-benzoylphenylalanine (Bpa). For photoaffinity labeling (PAL)-based competition assays, 5 μM CaM was incubated with 5 μM Bpa-Mover(203-221). Photoreactions were analyzed by gel electrophoresis (B) and MALDI-TOF-MS (C) to monitor the formation of covalent photoadducts (PA) in the mass range of 19 kDa (PA1, 1:1 peptide/CaM complex) to 21 kDa (PA2, 2:1 peptide/CaM complex). Photoreactions were performed in the absence (−) and presence (+) of 100 μM Ca²⁺, and with increasing concentrations of Mover(203-221), R-Mover(203-221), and R/E-Mover(203-221) as competitors. For the mutant peptides, only the highest concentration [50-fold molar excess over Bpa-Mover(203-221)] is shown. As can be followed most clearly by means of the intensity of the signal for free CaM, Mover(203-221) effectively suppressed photoadduct formation, while the mutant competitors R-Mover(203-221) and R/E-Mover(203-221) showed reduced or no affinity to CaM, respectively. Note that also a 50-fold molar excess of Mover(203-221) did not lead to full suppression of photoadduct formation. This was most likely due to the known positive correlation between bulkiness/hydrophobicity of N-terminal anchor positions in amphipathic CaM-binding peptides and their affinity for CaM (O’Neill et al., ; Dimova et al., 2006). Accordingly, Bpa-Mover(203-221) with a Bpa anchor residue (two phenyl moieties) binds CaM with a higher affinity than Mover(203-221) with a Phe anchor residue (one phenyl moiety). Two independent experiments were performed.

Figure 5

NFAT1-GFP translocation is enhanced by overexpression of Mover in HeLa cells. (A) The dynamics of nuclear NFAT1-GFP fluorescence was imaged upon histamine stimulation of live HeLa cells expressing either recombinant full-length Mover (red color) or DsRed as a control (black color). Expressing Mover increases the histamine-induced accumulation of NFAT1-GFP in the nucleus. (B) Quantification of the nuclear NFAT1-GFP fluorescence as fold change at the end of the experiment vs. the beginning (p < 0.0001). (C,D) Full-length Mover carrying the F206R point mutation still increases histamine-induced NFAT1-GFP nuclear translocation (p = 0.004). (E,F) The F206R/K207E/K215E/K219E quadruple mutant, called Mover 4mut, fails to increase the histamine induced NFAT1-GFP nuclear translocation (p < 0.0001). Two independent experiments were performed, Mann–Whitney test was used, n is indicated in the panels. **p < 0.01,****p < 0.0001.

Figure 6

Targeting and self-interaction properties of Mover constructs containing mutations in the CaM binding domain. (A) F206R-mGFP, corresponding to full-length Mover carrying one point mutation, and 4-mut-mGFP, corresponding to full-length Mover carrying four point mutations (F206R, K207E, K215E and K219E) are diffusely distributed in rat cultured hippocampal neurons. Their distribution resembled that of soluble mGFP alone. The overviews (left) show inverted gray level epifluorescence images of GFP-immunofluorescence. The zooms on the right represent triple-fluorescence images of the boxes, showing that axonal areas (MAP2 negative processes) of the transfected neurons do not contain local accumulation of the constructs even where the axon contacts a dendrite (MAP2 positive process), i.e., at a location where synapses can be formed. Synaptophysin (S-physin) is a marker for presynaptic SV clusters. The distribution of each construct was analyzed in more than 30 neurons from three independent experiments. (B) Full-length Mover (1-266) co-immunoprecipitates with Mover-myc in transfected HEK293 cells, but the F206R mutant and the 4-mut mutant do not, indicating that these point mutations disrupt homomeric interaction of Mover. The data represent two independent experiments, i.e., two HEK293 cell transfections followed by lysis and co-immunoprecipitation.

Figure 7

Targeting and self-interaction properties of phospho-deficient mutants of Mover. (A) The T13A mutant and the T64A mutant of Mover co-immunoprecipitates with Mover-myc in transfected HEK293 cells, indicating that they undergo homomeric interaction. The data represent two independent experiments. (B,C) The two phospho-deficient mutants are targeted to presynaptic sites in rat cultured hippocampal neurons. This is indicated both by the punctate staining pattern (inverted gray level images) and the colocalization with the synapse marker Synaptophysin (small panels, representing zoomed version of the boxes). The distribution of each construct was analyzed in more than 15 neurons from three independent experiments.

Figure 8

Targeting of Mover constructs in Mover knockout neurons. (A) Western blot demonstrating the absence of Mover in a brain homogenate from Mover knockout mice. (B,C) Immunofluorescence of cultured hippocampal neurons from WT mice (B) and Mover knockout mice (C) indicating the absence of Mover immunofluorescence in the knockout cultures. (D) Inverted gray level images showing the distribution of the indicated constructs in transfected Mover knockout cultures. All constructs produce a punctate staining characteristic of presynaptic targeting. The small panels represent zooms of the boxes. Arrowheads show examples of punctate colocalization of the constructs with the synapse marker Synapsin. (E) Quantification of colocalization with the synapse marker Bassoon. The number of punctate GFP-signals colocalizing with Bassoon is shown as percentage of the total number of punctate GFP-signals (n = 18 regions of interest; N = 6 independent experiments; student’s t-test). (F) Enrichment of the constructs at synapses. The average fluorescence intensity of punctate GFP-signals colocalizing with a synapse marker and the average fluorescence intensity of nearby diffuse GFP-signals (i.e., non-synaptic signals) was determined. The ratio of synaptic vs. non-synaptic fluorescence signals is shown (n = 18 regions of interest, including 540 synapses total; N = 6 independent experiments; student’s t-test).