A Novel Neuron-Specific Regulator of the V-ATPase in Drosophila

Abstract

The V-ATPase is a highly conserved enzymatic complex that ensures appropriate levels of organelle acidification in virtually all eukaryotic cells. While the general mechanisms of this proton pump have been well studied, little is known about the specific regulations of neuronal V-ATPase. Here, we studied CG31030, a previously uncharacterized Drosophila protein predicted from its sequence homology to be part of the V-ATPase family. In contrast to its ortholog ATP6AP1/VhaAC45 which is ubiquitous, we observed that CG31030 expression is apparently restricted to all neurons, and using CRISPR/Cas9-mediated gene tagging, that it is mainly addressed to synaptic terminals. In addition, we observed that CG31030 is essential for fly survival and that this protein co-immunoprecipitates with identified V-ATPase subunits, and in particular ATP6AP2. Using a genetically-encoded pH probe (VMAT-pHluorin) and electrophysiological recordings at the larval neuromuscular junction, we show that CG31030 knock-down induces a major defect in synaptic vesicle acidification and a decrease in quantal size, which is the amplitude of the postsynaptic response to the release of a single synaptic vesicle. These defects were associated with severe locomotor impairments. Overall, our data indicate that CG31030, which we renamed VhaAC45-related protein (VhaAC45RP), is a specific regulator of neuronal V-ATPase in Drosophila that is required for proper synaptic vesicle acidification and neurotransmitter release.

Keywords: ATP6AP1L/ATP6AP1; CG31030/VhaAC45RP; Drosophila melanogaster; neuronal V-ATPase; quantal size; synaptic vesicle acidification.

Figure 1.

Structure of CG31030 compared with its paralog VhaAC45 and construction of CG31030^V5. A, Localization of the primers used for RT-PCR analysis of the two predicted CG31030 transcripts (CG31030-RA and CG31030-RB) in Drosophila heads. B, Primers detecting both CG31030-RA (F1-R1, expected amplicon size 288 bp) and CG31030-RB (F1-R1, expected amplicon size 233 bp) only amplify a band corresponding to CG31030-RA (lane 1). CG31030-RA-specific primers (F2-R2) also amplify a fragment of the expected size (485 bp, lane 2), whereas CG31030-RB-specific primers (F3-R3) did not amplify any cDNA fragment (lane 3). C, VhaAC45 and CG31030 both have a signal peptide (SP; orange bars), a transmembrane domain (TM; blue bars) and predicted N-glycosylation sites according to NetNGlyc (http://www.cbs.dtu.dk/services/NetNGlyc/; green diamonds). They also present the pair of cysteine residues characteristic of the AC45 family (red triangles). In addition, VhaAC45 has a predicted furin cleavage site (yellow bar) which is absent in CG31030 according to ProP 1.0 (http://www.cbs.dtu.dk/services/ProP/). D, Construction of the CG31030^V5 mutant strain. A 14-amino acid V5 tag (in red) was fused to the C-terminal end of the CG31030 protein by inserting the V5 coding sequence ended by a new stop codon (in blue) in place of the original stop codon (in green) in the CG31030 gene using the CRISPR-Cas9 technology. E, A Western blotting probed with anti-V5 antibody detected two protein bands at an apparent molecular weight of ∼32 kDa in head extracts of CG31030^V5 flies (lane 2) and none in the w¹¹¹⁸ control sample (lane 1).

Figure 2.

CG31030 is expressed specifically in neurons. A, Diagram representing expression levels of CG31030 and VhaAC45 in different tissues relative to the whole fly, according to FlyAtlas data (Chintapalli et al., 2007). CG31030 appears to be markedly enriched in the nervous system of larva and adult fly, whereas in contrast VhaAC45 seems to be uniformly expressed in all tissues in these two stages. B, In both males and females, CG31030 mRNA abundance follows the localization of the CNS (shown in blue on the fly sketch), with the highest expression in the head. Br, brain; VNC, ventral nerve cord. Results of three independent experiments. C, Knock-down of CG31030 expression in neurons decreases adult longevity. Pan-neuronal expression of CG31030^RNAi3 and Dcr-2 with the elav-Gal4 driver led to a marked shortening of the lifespan of adult flies. Experiment conducted with 105–110 females per genotype. D, Expression of CG31030^RNAi3 with Dcr-2 in all neurons using elav-Gal4 decreased CG31030 mRNA level in head by >80%, while expression of the RNAi construct and Dcr-2 in all adult glial cells with repo-Gal4 had not effect. Results of three independent experiments. Mean values with SD are reported on the graphs.

Figure 3.

CG31030 is addressed to synaptic areas. A, B, Anti-V5 immunostaining in the CG31030^V5 strain revealed that CG31030 is mainly addressed to synaptic areas in the adult brain (A) and larval CNS (B), as indicated by its colocalization with the presynaptic marker CadN. Green fluorescence outside synaptic areas reveal that cell bodies are also faintly marked by the V5 antibody, with a few of them showing a bright signal (arrowheads). OL, optic lobe; AL, antennal lobe; SO, subesophageal ganglion. C, CadN and V5 immunostaining of adult brain (left panels) and larval CNS (right panels) from w¹¹¹⁸ control flies. No signal was detected with the anti-V5 antibody in the absence of V5-tagged protein. D, Anti-V5 antibody also labels the neuromuscular junction of CG31030^V5 larvae, where it co-localizes with anti-HRP immunostaining that specifically marks neuronal membranes, confirming CG31030 synaptic localization. Scale bars: 100 μm (A–C) and 10 μm (D).

Figure 4.

CG31030 co-immunoprecipitates with V-ATPase subunits. A, Scatter plot of proteins identified by nano LC-MS/MS in three independent co-immunoprecipitation experiments using anti-V5 antibodies. R₁, R₂ and R₃ represent the abundance ratio of proteins identified in adult head extracts from CG31030^V5 over w¹¹¹⁸ control in experiments 1, 2 and 3, respectively. Log₂(R) = 1 corresponds to a 2-fold abundance difference. A total of 12 proteins were found to be at least twice as abundant in CG31030^V5 as in the control in all three experiments (red or blue dots on the graph). One of them is the immunoprecipitation target CG31030 (yellow dot) and three of these proteins belong to the V-ATPase complex (red dots). A list of these 12 proteins with their Log₂(R) values is provided in Table 3. B, Standard model of the Drosophila V-ATPase complex showing structure of the V1 and V0 domains and the predicted localization of the three subunits that co-immunoprecipitated with CG31030. C, Co-localization of CG31030-V5 with Vha100-1 at the larval neuromuscular junction. Larval muscles were labeled with an anti-HRP antibody to mark neuronal membranes. CG31030 has a similar localization to that of Vha100-1, the synaptic isoform of subunit a of the V₀ domain. D, In w¹¹¹⁸ control flies, no specific signal could be detected at the neuromuscular junction with the anti-V5 antibody. Scale bars: 10 μm (C, D).

Figure 5.

Change in CG31030 expression does not alter V-ATPase subunit transcript levels. A, B, CG31030 upregulation (A) or knock-down (B) in vivo did not significantly affect adult head transcript levels of several V-ATPase V₀ subunits. Results of three independent RT-qPCR experiments. Means and SD are displayed on the graph.

Figure 6.

CG31030 knock-down larvae have a synaptic vesicle acidification defect. A, Schematic representation of the protocols used to assess relative acidity levels of synaptic vesicles at the larval neuromuscular junction. Top center diagram, In control conditions, fluorescence can be emitted by VMAT-pHluorin in the presynaptic membrane but not in synaptic vesicles since their lumen is acidified. In case of defective synaptic vesicle acidification, both the presynaptic membrane and synaptic vesicles should emit fluorescence. Left diagram, The fluorescence emitted by VMAT-pHluorin in the presynaptic membrane can be quenched by replacing the extracellular medium with an acidic HL3 solution. This quenching should result in an almost complete extinction of the signal in control flies, in which synaptic vesicles are normally acidified, while a residual signal is expected to be visible in flies having a synaptic vesicle acidification defect. Right diagram, Replacement of 50 mm NaCl by 50 mm NH₄Cl in the neutral HL3 solution should lead to a collapse of the pH gradients because of the free diffusion of NH3 in membranes, so that fluorescence will be emitted both by the presynaptic membrane and synaptic vesicles both in control and mutant conditions. B, Representative pictures showing the effect of perfusing an acidic HL3 solution on VMAT-pHluorin fluorescence in control and CG31030 knock-down larvae. C, D, Quantification of the ratio of the fluorescence level at pH 5.5 over the original signal at pH 7.6. Whereas the low pH extinguished fluorescence in control flies, ∼37% of the signal persisted after quenching in CG31030 knock-down larvae using two different RNAi constructs. E, Representative pictures showing the effect of collapsing the synaptic vesicle pH by perfusing HL3-NH₄Cl in control and CG31030 knock-down larvae. F, G, Quantification of the ratio of the signal in HL3-NH₄Cl over the original signal in HL3 showed that fluorescence increased ∼3-fold in controls while it only rose by 10–20% depending on the RNAi in CG31030 knock-down larvae. Results of four independent experiments, with three to five larvae analyzed per genotype in each experiment. Unpaired Student’s t test; *p < 0.05, ***p < 0.001. Mean values with 95% confidence intervals are reported on the graphs. Scale bar: 15 μm (B).

Figure 7.

CG31030 downregulation in motoneurons decreases larval locomotor performance. A, CG31030 knock-down in motoneurons did not significantly alter larval size. Larval length is defined as the spine length from head to tail, while larval width is the diameter of the mid-spine circle. Average larval length and width were not significantly different between animals expressing CG31030 RNAi1 (left panels) and RNAi2 (right panels) in motoneurons and controls. B, Schematic representation of the successive phases of larval locomotion. Stride size is defined as the distance crawled during one peristaltic wave of muscle contraction, while stride duration is the time necessary for the completion of one peristaltic wave. C, Locomotor trails of individual larvae recorded over a period of 2 min. Larvae expressing CG31030 RNAi in motoneurons show reduced spontaneous movements (middle panel) compared with the driver (left panel) and effector (right panel) controls, respectively. Scale bar: 25 mm. D, Quantification of the traveled distances confirmed that knocking down CG31030 in motoneurons with two different RNAi induced significant locomotor defects. E, F, Stride size (E) appears to be less affected than stride duration (F) in the knocked-down larvae. Result of five independent experiments, with three to four larvae analyzed per genotype in each experiment. One-way ANOVA with Dunnett’s post hoc test for multiple comparisons; **p < 0.01, ***p < 0,001. Mean values with 95% confidence intervals are reported on the graphs.

Figure 8.

Lack of effect of CG31030 knock-down on the synaptic vesicle cycle at the larval neuromuscular junction. A, B, Representative images of mCD8::GFP-labeled neuromuscular junctions of a control (A) and a knock-down (B) larva after loading and unloading of the fluorescent FM4-64 dye. C, D, Nerve terminals from larvae expressing CG31030 RNAi1 (C) or RNAi2 (D) were able to both load and unload the fluorescent FM4-64 dye as efficiently as controls. No significant difference was observed between controls and knock-down animals during the loading and the unloading phases. Fluorescence was normalized relative to the level in control larva synapses after 5-min loading. E, F, The percentage of previously loaded FM4-64 dye released during the unloading phase was similar for control and knock-down larvae when using either RNAi1 (E) or RNAi2 (F); **p < 0.01, ***p 0.001, ns: not significant.

Figure 9.

Synaptic quanta size is reduced in CG31030 knock-down larvae. A, Schematic representation of a dissected larval fillet. Spontaneous miniature EPSPs (mEPSPs) were recorded intracellularly from the ventral longitudinal abdominal muscle 6 in segment A3. B, Expression of membrane-associated mCD8::GFP with the glutamatergic driver OK371-Gal4 strongly labels the presynaptic nerve endings at the neuromuscular junction of muscles 6–7 in segment A3. The scaffolding protein Dlg was used as a postsynaptic marker. Scale bar: 20 μm. C, Representative distributions of spontaneous mEPSPs recorded in a control larva (top panel, in gray) and a CG31030 RNAi knock-down larva (bottom panel, in orange). The dotted lines represent actually recorded amplitudes and the plain lines the computed theoretical distributions. Genotypes and quantal size (q) are indicated on each graph. D, Quantal analysis of recorded events showed that knock-down larvae have a significantly reduced synaptic quanta size compared with controls, both with CG31030 RNAi1 (top panel, in orange) or RNAi2 (bottom panel, in blue). E, Although both RNAi1 (left panel) and RNAi2 (right panel) larvae had a lower mean EPSP frequency than controls, this difference was not statistically significant. Six to seven cells recorded per genotype. One-way ANOVA with Dunnett’s post hoc test for multiple comparisons; *p < 0.05, **p < 0.01, ***p < 0.001. Mean values with 95% confidence intervals are reported on the graphs.

Figure 10.

Effect of CG31030 knock-down on synapse morphology at the larval neuromuscular junction. A–D, No significant difference was observed in the mean synaptic bouton diameter (A, B) or the average number of boutons per synapse (C, D) between larvae expressing CG31030 RNAi1 (A, C) or RNAi2 (B, D) in motoneurons and the controls. E, F, Knock-down larvae showed a smaller percentage of boutons with a diameter larger than 4 μm for both RNAi1 [E; but not significant (ns) compared with the effector control] and RNAi2 (F), n = 6–8 synapses per genotype, two-way ANOVA with Dunnett’s correction for multiple comparisons; *p < 0.05, **p < 0.01, ***p < 0.001, ns: not significant.