Phosphorylation of VAChT on Ser-480 in COS cells. A, Alignment of the sequences surrounding the dileucine motifs (underlined) in the COOH termini of VMAT1, VMAT2, and VAChT. All three transporters show a glutamate at position −4 relative to the dileucine motif. The VMATs also contain a glutamate at the −5 position. In contrast, VAChT contains a serine at the equivalent position, serine 480 (Ser-480), surrounded by several arginines that form a consensus sequence for phosphorylation by PKC and other basic residue-directed kinases (Pearson and Kemp 1991). B, Phosphorylation of VAChT in COS cells. Using COS cells transiently transfected with the cDNA encoding VAChT (+) or no DNA (−), the left shows Western analysis with a rabbit antibody to VAChT. In the right, cells metabolically labeled with ³²P_i were immunoprecipitated with the same VAChT antiserum. Of the ∼70-kD (black bracket), ∼50-kD (arrowhead), and ∼35-kD immunoreactive species (gray bracket), VAChT undergoes phosphorylation predominantly on the ∼70-kD form. The ∼50-kD species and the ∼35-kD doublet appear to undergo less prominent labeling. In the case of the ∼35-kD species, the reduced labeling results at least partly from the inefficiency of immunoprecipitation (data not shown). C, Phosphoamino acid analysis of VAChT. Immunoprecipitated VAChT was excised from the gel, hydrolyzed in hydrochloric acid, and the hydrolysate separated by thin-layer electrophoresis, using pH 1.9 for the first dimension and pH 3.5 for the second dimension. Autoradiography shows radiolabeled material comigrating with the phosphoserine (P-Ser) standard, but not with the phosphothreonine (P-Thr) or phosphotyrosine (P-Tyr) standards. Phosphopeptides resulting from incomplete hydrolysis migrate as a smear in the second dimension. Free ³²P_i migrates at the position shown (P_i), and the origin is circled. D, VAChT undergoes phosphorylation on Ser-480. COS cells transiently transfected with cDNAs encoding wild-type VAChT (wt), with mutant cDNAs containing alanine replacements at Ser-478, Ser-480, or with no DNA (−) were metabolically labeled with ³²P_i, immunoprecipitated with the VAChT antibody, and subjected to SDS-PAGE, followed by autoradiography. Mutation of Ser-480, but not Ser-478, reduced phosphorylation of the mature form of VAChT (black bracket) and the COOH-terminal fragment (gray bracket). Molecular weight markers (kD) are shown to the left in B and D.

In vitro phosphorylation of VAChT on Ser-480. A, Phosphorylation of wild-type and mutant VAChT bacterial fusion proteins. Fusion proteins containing GST joined in frame to the COOH terminus of wild-type VAChT (wt) or VAChT mutants with alanine replacements at Ser-478, Ser-480, or both Ser-478 and Ser-480 (478/480) were incubated with γ[³²P]ATP and a postnuclear COS cell supernatant for 20 min at 30°C. The proteins were then separated by electrophoresis and submitted to autoradiography. The fusion protein containing wild-type VAChT showed robust phosphorylation. Mutation of Ser-480 or both Ser-478/480, but not Ser-478 alone eliminates phosphorylation. Staining with Coomassie blue showed similar levels of fusion protein in each lane (data not shown). B, Calcium increases VAChT phosphorylation. Wild-type VAChT fusion proteins were incubated as in A with and without 2 mM calcium chloride, and 2.4 μM calmodulin. The addition of calcium potentiates VAChT phosphorylation, implicating a calcium-activated kinase. C, PKC inhibitors reduce VAChT phosphorylation. Wild-type VAChT fusion proteins were incubated as in A, but with the concentration of ATP lowered to 1 μM to reduce competition with kinase inhibitors. The PKC-selective inhibitor H7 (final concentration 6 or 60 μM), and the more PKC-specific BIS (final concentration 0.1 or 1 μM) or vehicle (−) were added before the lysate and γ[³²P]ATP. Coomassie blue staining confirmed the presence of similar amounts of fusion protein in each lane (bottom). D, Quantitative analysis of PKC inhibition. Autoradiographs of the kind shown in D were optically scanned, the digital images quantitated, and the values expressed as a percentage of samples treatment with vehicle alone (percent control). The average ± standard error of two independent experiments are shown. E, Phosphorylation of VAChT bacterial fusion proteins using PC12 cell lysate. Fusion proteins containing the COOH terminus of wild-type VAChT (lanes 1–7) or the VAChT mutant with an alanine replacement at Ser-480 (lane 8) were incubated with γ[³²P]ATP and a postnuclear PC12 cell supernatant for 20 min at 30°C with or without addition of the PKC inhibitors H7 or BIS, as described in C (lanes 1–5), or 2 mM calcium chloride (lanes 6 and 7). Coomassie blue staining confirmed the presence of similar amounts of fusion protein (bottom). F, Phosphorylation of VAChT bacterial fusion proteins by purified PKC. Fusion proteins containing wild-type VAChT (wt) or VAChT mutants with alanine replacements at Ser-480 or both Ser-478 and Ser-480 (478/480), or histone H1 as a positive control were incubated with γ[³²P]ATP and the catalytic fragment of PKC for 20 min at 30°C. Similar to histone H1, the fusion protein containing wild-type VAChT shows robust phosphorylation. Mutation of Ser-480, or both Ser-478 and Ser-480, eliminates phosphorylation by PKC.

Mutation of Ser-480 influences localization of VAChT to endosomes and SLMVs. PC12 cells stably expressing HA-tagged wild-type VAChT, the S480A and S480E mutants were double-stained with a mouse mAb to HA (a, d, and g), and a rabbit polyclonal antibody to synaptophysin (syn; b, e, and h), a marker for endosomes and synaptic-like microvesicles (SLMVs). Visualized with the appropriate secondary anti-mouse antibodies conjugated to FITC (a, d, and g) and anti-rabbit antibodies conjugated to Texas red (b, e, and h), the cells were examined by confocal laser microscopy. The overlays of FITC and Texas red channels are shown in c, f, and i. Wild-type VAChT (a–c) and the S480A mutant (d–f) extensively, but incompletely colocalize with synaptophysin in cell bodies and processes. In contrast, the S480E mutant (g–i) resides primarily in processes. Bar, 10 μm.

Density gradient fractionation of wild-type and mutant VAChT. PNSs of PC12 cells stably expressing HA-tagged wild-type VAChT (A), and mutants S480A (B) and S480E (C), were separated by equilibrium gradient centrifugation through 0.65–1.55 M sucrose. The fractions were submitted to Western analysis with the mAb to HA for VAChT (top of immunoblots in A, B, and C), a polyclonal antibody to the LDCV marker SgII (middle), and an mAb to the endosome/SLMV marker synaptophysin (syn; bottom). The gradients shown in A–C were digitized and quantified. For each fraction, the amount of HA-tagged VAChT, synaptophysin, and SgII was expressed as a percentage of the total immunoreactivity shown to the left of the immunoblots. Consistent with previous results (Liu and Edwards 1997a), wild-type VAChT partially cofractionates with synaptophysin as a broad peak that also overlaps with fractions containing SgII. In contrast, S480A cofractionates well with synaptophysin. The mutant S480E mutant, on the other hand, fractionates with synaptophysin, but also shows a second peak that cofractionates with SgII in heavy fractions. D, The ratio of HA-tagged VAChT localized to light fractions 6–8 (left) or heavy fractions 11 and 12 (right) were expressed as a percentage of total HA immunoreactivity. The mean ± standard error (n = 3) were derived from the data shown in A–C, together with additional experiments using two independently derived cell lines for each mutant. E, Immunoisolation of synaptophysin-containing vesicles. PNSs from cell lines expressing wild-type VAChT, and the S480A and S480E mutants, were incubated with magnetic beads conjugated with an mAb to synaptophysin. Western blots of immunoisolated synaptophysin positive (syn⁺) vesicles and starting lysate (start) were then immunostained using a polyclonal antiserum to VAChT. The amount of immunoisolated VAChT relative to the starting material was determined from digitized images and expressed as syn⁺/start. The average ± SEM were determined from two independent experiments. Similar proportions of wild-type of S480A localize to syn⁺ membranes, whereas the S480E mutant localizes approximately half as well, indicating the importance of a neutral residue at Ser-480 for sorting to light vesicles.

The E478A/E479A mutation redistributes VMAT2 away from LDCVs. PNSs of PC12 cells stably expressing HA-tagged wild-type VMAT2 and the E478A/E479A (EE/AA) mutant were separated by equilibrium sedimentation through 0.6–1.6 M sucrose. The fractions were analyzed by Western blot using a monoclonal HA antibody for VMAT2 (top of immunoblots in A and B), a polyclonal antibody to SgII (middle), and an mAb to synaptophysin (bottom). The immunoblots shown in A and B were digitized, quantified, and the amounts of HA-tagged VMAT2, synaptophysin, and SgII was expressed as a percentage of total immunoreactivity. Wild-type VMAT2 cofractionates more strongly with SgII than with synaptophysin. In contrast, the EE/AA mutation shifts VMAT2 into lighter membrane fractions. D, The proportion of VMAT2 localized to light fractions 6–8 (left) or heavy fractions 11–12 was expressed as a percentage of total VMAT2 immunoreactivity across the gradients, using the data shown in A–C and two additional independently isolated cell lines for each mutant. The bar graph shows the mean ± standard error (n = 3).

Phosphorylation of VAChT by PKC in PC12 cells. A, Untransfected PC12 cells (−), cells stably expressing HA-tagged wild-type VAChT (wt), and cells expressing mutant VAChT with an alanine replacement at Ser-480 (S480A) were metabolically labeled with ³²P_i, immunoprecipitated with an antiserum against VAChT, and subjected to SDS-PAGE, followed by autoradiography (right), or blotted to nitrocellulose and immunostained with an mAb to the HA tag (left). The mature ∼70-kD and the ∼35-kD doublet forms of VAChT undergo phosphorylation (see arrowheads). Substitution of Ser-480 by alanine reduces phosphorylation of the ∼70-kD form and eliminates phosphorylation of the ∼35-kD doublet without affecting the levels of the mature VAChT protein (left). Note that the mAb to HA does not recognize the ∼35-kD doublet as seen in the immunoblot, presumably because these species are COOH-terminal fragments that do not contain the more NH₂-terminal HA tag. B, Regulation of VAChT phosphorylation. PC12 cells stably expressing wild-type HA-tagged VAChT were metabolically labeled for 4 h with ³²P_i, treated for 15 min with vehicle alone (DMSO; lane 1), 10 μM A23187 (lane 2), 3 μM phorbol diacetate (PDA; lane 3), or 1 mM 8-bromo-cyclic AMP (cAMP) with 1 mM of the phosphodiesterase inhibitor 3-isobutyl-1-methylxanthine (lane 4), and the extracts either immunoprecipitated for VAChT (top), or transferred to nitrocellulose and immunostained for HA (bottom). VAChT shows a robust increase in phosphorylation with PDA, and a smaller increase in phosphorylation with A23187. A small decrease in phosphorylation occurs with increased cAMP (lane 4). C, Regulation of Ser-480 phosphorylation. PC12 cells stably expressing wild-type HA-tagged VAChT (lanes 1 and 2) and cells expressing mutant VAChT with an alanine replacement at Ser-480 (S480A; lanes 3 and 4) were metabolically labeled with ³²P_i, treated for 15 min with vehicle alone (−; lanes 1 and 3) or 3 μM phorbol diacetate (PDA; lanes 2 and 4) and immunoprecipitated as in B. Replacement of Ser-480 with alanine greatly reduces PDA-induced phosphorylation of VAChT.

Immunoelectron microscopy of PC12 cells expressing wild-type and mutant VAChT. Immunogold-labeled cryosections from PC12 cells expressing wild-type VAChT (A–C), S480A (D), or S480E (E) were examined by EM. Labeled LDCVs are indicated by arrowheads, whereas VAChT-positive small vesicular profiles are indicated with an asterisk. A, VAChT (10-nm gold) is found in the Golgi complex (G), as well as many neighboring vesicles (asterisk). LDCVs (arrowheads) are positive, but no label is found on lysosomes (L). B, Example of VAChT label (10-nm gold) in small vesicles close to an early endosomal vacuole (EE). C, Double immunogold-labeling shows that synaptophysin (Sphy, 10-nm gold) colocalizes to a considerable extent with VAChT (15-nm gold) in small vesicular profiles (asterisk). D, S480A (10-nm gold) localizes preferentially to small vesicles. E, By contrast, S480E (10-nm gold) is found preferentially in LDCVs. The small arrow points to a clathrin-coated pit at the plasma membrane (P). LE, late endosome. Bars, 200 nm.

Mutation of Ser-480 alters the localization of VAChT to process tips. PC12 cells stably expressing wild-type VAChT, S480A, and S480E mutants were double-stained with a mouse mAb to the epitope tag (HA; a, d, and g), and a rabbit polyclonal antiserum to SgII (b, e, and h), a marker for LDCVs. Visualized with the appropriate secondary anti-mouse antibodies conjugated to FITC (a, d, and g) and anti-rabbit antibodies conjugated to Texas red (b, e, and h), the cells were examined by confocal microscopy. The overlays of FITC and Texas red channels are shown in c, f, and i. Wild-type VAChT (a–c) and S480A (d–f) partially colocalize with SgII at the tips of processes, but also occur at high levels in cell bodies. In contrast, S480E (g–i) colocalizes extensively with SgII in the tips of processes, and like SgII, shows less expression in cell bodies. Bar, 10 μm.

Neutralization of glutamates-478 and -479 in VMAT2. To eliminate the negative charge upstream of the dileucine motif in VMAT2, Glu-478 and -479 were replaced by alanine. PC12 cells stably expressing wild-type VMAT2 and the E478A/E479A mutant (EE/AA) were double-stained with a mouse mAb to the HA epitope (a, d, g, and j) and either a rabbit polyclonal antiserum to SgII (b and e), or a rabbit antibody to synaptophysin (syn; h and k). Secondary anti-mouse antibodies conjugated to Cy3 (a, d, g, and j) or anti-rabbit antibodies conjugated to Cy5 (b, e, h, and k) were used to visualize the proteins, and the cells were examined by confocal microscopy. The overlay of Cy3 and Cy5 images are shown in c, f, i, and l. Wild-type VMAT2 colocalizes with both SgII and synaptophysin at the tips of cell processes, but does not colocalize with synaptophysin at cell bodies. In contrast, the EE/AA mutant resides at high levels in cell bodies and exhibits greater colocalization with synaptophysin than wild-type VMAT2. Bars, 10 μm.

Localization of wild-type and mutant VAChT to LDCVs. PNSs of PC12 cells stably expressing wild-type VAChT (wt), and S480A and S480E mutants preloaded with [³H]serotonin were separated on a sucrose velocity gradient, the peak LDCV fractions determined by scintillation counting, pooled, and separated again by equilibrium sedimentation through sucrose. The fractions were immunoblotted with an mAb to the HA tag and the amount of HA immunoreactivity in each fraction of the second gradient was quantified and expressed as a proportion of immunoreactivity in the starting material (PNS) loaded onto the first gradient (A). The peak of HA immunoreactivity on the second density gradient coincides with that of the LDCV marker SgII (B), and with the peak of [³H]serotonin (not shown). The average ± SEM was determined from two independently isolated cell lines for each mutant (C). S480E localizes to LDCVs ∼2.3-fold and threefold better than wild-type VAChT and the S480A mutant, respectively.