Evidence that electrostatic interactions between vesicle-associated membrane protein 2 and acidic phospholipids may modulate the fusion of transport vesicles with the plasma membrane

Abstract

The juxtamembrane domain of vesicle-associated membrane protein (VAMP) 2 (also known as synaptobrevin2) contains a conserved cluster of basic/hydrophobic residues that may play an important role in membrane fusion. Our measurements on peptides corresponding to this domain determine the electrostatic and hydrophobic energies by which this domain of VAMP2 could bind to the adjacent lipid bilayer in an insulin granule or other transport vesicle. Mutation of residues within the juxtamembrane domain that reduce the VAMP2 net positive charge, and thus its interaction with membranes, inhibits secretion of insulin granules in beta cells. Increasing salt concentration in permeabilized cells, which reduces electrostatic interactions, also results in an inhibition of insulin secretion. Similarly, amphipathic weak bases (e.g., sphingosine) that reverse the negative electrostatic surface potential of a bilayer reverse membrane binding of the positively charged juxtamembrane domain of a reconstituted VAMP2 protein and inhibit membrane fusion. We propose a model in which the positively charged VAMP and syntaxin juxtamembrane regions facilitate fusion by bridging the negatively charged vesicle and plasma membrane leaflets.

Figure 1.

The VAMP2 JMD binds to acidic phospholipids through electrostatic interactions. (A) Amino acid sequence alignment of the human VAMP2 JMD residues (77-96) with the human VAMP1 and VAMP3 proteins. Conserved basic residues (+) are in blue, acidic residues (−) are in red, and hydrophobic residues (Φ) are in green. The mutation of the human VAMP2 used in this study is underlined. (B) A peptide corresponding to the VAMP2 JMD, VAMP2(79-95), was labeled with [³H]NEM and incubated with sucrose-loaded LUVs containing the indicated mole ratios of PC/PS. After separation by centrifugation the percentage of bound peptide was determined directly from the radioactivity in the supernatant and the pellet. The molar partition coefficient, K, was determined by plotting the percentage of bound peptide versus concentration of lipid accessible to the peptide (1/2 total lipid concentration because the membrane-impermeable peptide is added to the preformed vesicles). The curves illustrate the best fit of the equation in Materials and Methods to the data: note K is the reciprocal of the lipid concentration that binds 50% of the peptide. (C) The value of K obtained from Figure 1B was plotted as a function of the mole % PS in the PC/PS LUVs. Note the binding of the positively charged (6 basic residues) peptide increases exponentially with the mole fraction of acidic lipid PS in the membrane, as predicted theoretically for electrostatic interactions. (D) The binding of the [³H]NEM-labeled VAMP2(79-95) to PC/PIP₂ (99:1) or PC/PS/PIP₂ (74:25:1) vesicles was determined as described above. Data for 3:1 PC/PS vesicles reproduced from B to facilitate comparison.

Figure 2.

The VAMP2 JMD tandem Trp residues penetrate inside the low dielectric acyl chain region of the lipid bilayer. (A) The VAMP2(79-95) peptide (1 μM) in an aqueous environment (reconstitution buffer) was excited at 274 nm and the emission spectra scanned from 310 to 370 nm (dotted line). The VAMP2 (79-95) peptide was incubated with 100 μM (accessible lipid concentration = 1/2 total) PC (dashed line) or 3:1 PC/PS LUVs (solid line), and Trp fluorescence was determined as described in Materials and Methods. The blue shift and enhanced fluorescence observed when the peptide binds to the PC/PS vesicles suggest the Trp residues insert into the low dielectric interior of the bilayer. (B) Fluorescence of ∼1 μM full-length VAMP2 WT protein reconstituted into PC or 3:1 PC/PS LUVs (∼50 μM lipid) were scanned from 310 to 370 nm (excitation, 274 nm). Maximal peak of the Trp signal was normalized to 100 arbitrary units to illustrate the slightly greater blue shift observed with PC/PS versus PC membranes.

Figure 3.

Reversing the net negative charge on a phospholipid vesicle reverses the membrane binding of the positively charged JMD of reconstituted VAMP2. The full-length VAMP2-WT protein (∼1 μM) was reconstituted into 3:1 PC/PS LUVs and incubated with the indicated concentrations of the amphipathic weak base sphingosine. The lipid concentration of the reconstituted LUVs, after purification, was ∼50 μM, which corresponds to ∼15 μM PS. (A) The samples were excited at 274 nm and emission spectra scanned from 310 to 370 nm. Trp fluorescence was determined by subtraction of fluorescence from full-length VAMP2-WW/AA reconstituted into LUVs. (B and C) The Trp fluorescence emission spectra obtained in panel A were deconvoluted into two components representing the hydrophobic (B) emission spectra (335 nm maximal emission peak) and the hydrophilic (C) emission spectra (355 nm maximal emission peak). (D) Percentage of the Trp inserted in the low dielectric acyl chain region of the membrane was determined by calculating the mass contribution (half width of the maximum peak method) of the 335 and 355 nm maximal peak of the deconvoluted spectra.

Figure 4.

Decreasing the positive charge of the VAMP2 JMD results in inhibition of calcium-dependent fusion of insulin granules. Min6B1cells were cotransfected with 6 μg of the VAMP2-WT or VAMP2-KR/ED mutant expression plasmids and with 2 μg of the hGH expression plasmid. (A) Twenty-four hours later, the cells were either left untreated (open bars) or stimulated (filled bars) with 30 mM KCl for 10 min. The cell supernatants and total cell extracts were then assayed for hGH and the release of hGH was determined as the percentage of total hGH content. These data are the average ± SD for three independent experiments, *p < 0.05. (B) The cells were fixed and subjected to confocal microscopy. White bar, 5 μm (c). This is a representative image showing the colocalization of the coexpressed VAMP2-WT and VAMP2-KR/ED mutant.

Figure 5.

Insulin secretion is dependent upon the net positive charge of the VAMP2 JMD but is independent of the specific amino acid residues that are mutated. Min6B1cells were cotransfected with 2 μg of the hGH expression plasmid plus either 6 μg of empty vector, VAMP2-WT (+6), VAMP2-RK/AA (+4), VAMP2-KNLK/ANLA (+4), VAMP2-KRK/AAA (+3), VAMP2-KLKRK/ALAAA (+2), or VAMP2-KR/ED (+2) mutants that have net JMD charges as indicated in the brackets and in the figure. Twenty-four hours later, the cells were either left untreated (open bars) or stimulated (closed bars) with 30 mM KCl for 10 min. The cell supernatants and total cell extracts were then assayed for the release of hGH as the percentage of total cellular hGH content. These data are the average ± SD from three independent experiments, *p < 0.05.

Figure 6.

Increased intracellular salt concentrations, which screen electrostatic interactions inhibit insulin secretion. Min6B1cells were treated briefly with a low concentration of the detergent digitonin to permeabilize the plasma membrane. The cells were then incubated with a buffer containing the indicated concentrations of potassium glutamate. The cells were then treated with or without 10 μM free Ca²⁺ in solutions containing the indicated concentrations of potassium glutamate. After 15 min of stimulation, Ca²⁺-dependent secretion was determined using an insulin ELISA assay and insulin release was reported as the percentage of total cellular insulin content. These data are the average ± SD from three independent experiments.

Figure 7.

Weak bases that neutralize the electrostatic surface potential of phospholipid membranes inhibit calcium-dependent fusion of insulin granules. (A) Min6B1 cells were treated with vehicle, 100 μM sphingosine, 100 μM W13, or 100 μM W12 for 20 min followed by stimulation with 30 mM KCl for 10 min. The cell supernatants were then collected and assayed for the presence of released insulin. These data are the average ± SD for two to five independent experiments, *p < 0.05. (B) Min6B1 cells were treated with vehicle or the indicated concentrations of W12, W13, or sphingosine for 20 min followed by stimulation with 30 mM KCl for 10 min. The supernatants and total cell extracts were then collected and insulin release was determined as the percentage of total cellular insulin content. These data are the average ± SD for two independent experiments each performed in duplicate.

Figure 8.

Weak bases that neutralize the electrostatic surface potential of membranes inhibit GLUT4 vesicle fusion with the plasma membrane. Fully differentiated 3T3L1 adipocytes were cotransfected with the myc-GLUT4-GFP reporter construct and the constitutively active Myr-Akt construct as described in Materials and Methods. Twenty-four hours later, cells were treated with vehicle, 100 μM sphingosine, or 100 μM W13 for 20 or 60 min. Cells were then fixed and subjected to confocal microscopy for the localization of the myc-GLUT4-GFP reporter. (A) Representative images from each treatment condition. White bar, 10 μm (t). (B) The percentage of GLUT4 vesicle fusion was calculated by a single-cell measurement of the ratio of exposed exofacial myc label (Alexa 594) to total GLUT4 (GFP label). These data represent the analysis of 15 cells/experiment from three independent experiments, *p < 0.05 compared with vehicle controls. (C) The percentage of GLUT4 translocation was calculated by a single-cell measurement of the ratio of plasma membrane GFP label to total GFP label. These data represent the analysis of 15 cells/experiment from three independent experiments, *p < 0.05 compared with vehicle controls. Error bars represent SD.

Figure 9.

Cartoon of how the SNARE JMDs may interact with the cytoplasmic leaflets of the plasma and vesicle membranes. Top, the positively charged JMD of VAMP2 is strongly attracted by the negative electrostatic potential produced by acidic phospholipids in the GLUT4 cargo vesicles; the two hydrophobic Trp residues penetrate into the bilayer. Similarly, the positively charged JMD of syntaxin 4 is strongly attracted by the negative electrostatic potential produced by the acidic phospholipids in the plasma membrane. Bottom, after the assembly of the SNARE complex, the opposing membranes come in proximity and the two phospholipid leaflets share the two basic clusters, which glue the bilayers together through nonspecific electrostatic interactions. Only a single JMD for VAMP2 and syntaxin are shown for clarity. See text for details.