miR-8 controls synapse structure by repression of the actin regulator enabled

Abstract

MicroRNAs (miRNAs) are post-transcriptional regulators of gene expression that play important roles in nervous system development and physiology. However, our understanding of the strategies by which miRNAs control synapse development is limited. We find that the highly conserved miRNA miR-8 regulates the morphology of presynaptic arbors at the Drosophila neuromuscular junction (NMJ) through a postsynaptic mechanism. Developmental analysis shows that miR-8 is required for presynaptic expansion that occurs in response to larval growth of the postsynaptic muscle targets. With an in vivo sensor, we confirm our hypothesis that the founding member of the conserved Ena/VASP (Enabled/Vasodilator Activated Protein) family is regulated by miR-8 through a conserved site in the Ena 3' untranslated region (UTR). Synaptic marker analysis and localization studies suggest that Ena functions within the subsynaptic reticulum (SSR) surrounding presynaptic terminals. Transgenic lines that express forms of a conserved mammalian Ena ortholog further suggest that this localization and function of postsynaptic Ena/VASP family protein is dependent on conserved C-terminal domains known to mediate actin binding and assembly while antagonizing actin-capping proteins. Ultrastructural analysis demonstrates that miR-8 is required for SSR morphogenesis. As predicted by our model, we find that Ena is both sufficient and necessary to account for miR-8-mediated regulation of SSR architecture, consistent with its localization in this compartment. Finally, electrophysiological analysis shows that miR-8 is important for spontaneous neurotransmitter release frequency and quantal content. However, unlike the structural phenotypes, increased expression of Ena fails to mimic the functional defects observed in miR-8-null animals. Together, these findings suggest that miR-8 limits the expansion of presynaptic terminals during larval synapse development through regulation of postsynaptic actin assembly that is independent of changes in synapse physiology.

Keywords: Drosophila; Enabled; NMJ; SSR; Synapse; miRNA.

Fig. 1.

miR-8 activity is predominantly required at L2 and L3 larval stages to promote neuromuscular junction development. (A-D) Images of w¹¹¹⁸ (A,C) control and miR-8^Δ/Δ (B,D) Drosophila 6/7 NMJ at 35 h (mid-L1 stage; A,B) and 60 h (mid-L2 stage; C,D) after egg laying (AEL). Scale bars: 10 μm. Quantification of synaptic boutons (E) and axonal branches (F) shows a significant deficit in NMJ development at the L2 and L3 stages. Error bars indicate s.e.m. *P≤0.0001 relative to control animals (two-tailed Student's t-test). n≥12 for all genotypes.

Fig. 2.

miR-8 mutation disrupts synaptic specialization through regulation of Ena. (A-E) Analysis of 6/7 NMJ boutons immunostained with postsynaptic marker Disc large (Dlg, green, left panels), presynaptic marker HRP (red, middle panels) and merged Dlg/HRP channel image (right panels). Control w¹¹¹⁸ (n=38) (A,E) display a low frequency of ‘naked’ synaptic boutons, as observed by the consistent apposition of HRP and Dlg signal. miR-8^Δ/Δ (n=38) (B,E) and muscle-specific overexpression of Ena (UAS-Ena) driven by how^24B-Gal4 (n=22) (C,E) NMJs causes significant (4.60- and 4.06-fold) increases, respectively, in ‘naked’ synaptic boutons. Expression of UAS-FP4-mito using the how^24B-Gal4 driver in a miR-8 homozygous mutant background (n=41) (D,E) significantly rescues the increase in naked bouton number in miR-8^Δ/Δ animals. Expression of UAS-FP4-mito using the how^24B-Gal4 driver (n=21) in a wild-type background shows no quantifiable difference from control. Scale bars: 2 μm. Error bars indicate s.e.m. **P≤0.01 relative to w¹¹¹⁸, *P≤0.05 relative to miR8^Δ/Δ (two-tailed Student's t-test).

Fig. 3.

Ena is enriched in SSR, and conserved actin-associated domains are necessary for synaptic localization. (A-B) Images of synaptic boutons immunostained with postsynaptic markers Dlg (A) or Cactus (B), and Ena (A′,B′). Merged images of Ena/Dlg (A″) and Ena/Cactus (B″) show substantial colocalization (yellow). Scale bars: 5 μm. (C-H′′) Images of 6/7 NMJ boutons expressing wild-type or mutant UAS-EGFP-mouse Ena (Mena) transgenes using the how^24B-Gal4 driver. Immunostaining of Dlg (left panel), EGFP (middle panel) and Dlg/EGFP merge (right panel). Control UAS-EGFP-Mena^WT boutons (C-C″) display colocalization between postsynaptic EGFP-Mena and Dlg, analogous to Ena immunostaining. UAS-EGFP-Mena^ΔPRR (D-D″) and UAS-EGFP-Mena^ΔFAB (E-E″) transgenes show EGFP-Mena staining pattern that is indistinguishable from wild-type control. Expression of UAS-EGFP-Mena^ΔGAB (F-F″) and UAS-EGFP-Mena^ΔCC (G-G″) demonstrate a marked deficiency in EGFP-Mena recruitment to the postsynaptic space. Expression of the N-terminal EVH2 domain (H-H″), which contains the GAB, FAB and CC motifs, shows localization that is indistinguishable from wild type at the synapse. Scale bars: 2 μm. (I) The UAS mammalian Ena (Mena) domain mutant transgenes (Loureiro et al., 2002) used to determine the structural requirements of Ena localization and function at the synapse. (J) Western blot analysis of UAS-EGFP-Mena transgenes driven by the how^24B-Gal4 driver show stable and comparable levels when probed with anti-EGFP (upper panel) and anti-tubulin as loading control (lower panel). Two whole animals were used per larval extract.

Fig. 4.

Conserved actin-associated domains are required for Mena function at the synapse. (A-H) Images of Drosophila 6/7 NMJ expressing wild-type or mutant UAS-EGFP-Mena transgenes using the pan-muscle how^24B-Gal4 driver. Scale bars: 20 μm. Muscle expression of UAS-EGFP-Mena^WT (C), UAS-EGFP-Mena^ΔPRR (D) and UAS-EGFP-EVH2 (H) show significantly disrupted NMJ morphogenesis. (I) Quantification of synaptic boutons and NMJ expansion. Expression of UAS-EGFP-Mena^ΔGAB (E), UAS-EGFP-Mena^ΔFAB (F) and UAS-EGFP-Mena^ΔCC (G) is indistinguishable from wild-type control NMJs. Error bars indicate s.e.m. *P≤0.01, **P≤0.001 relative to how^24B-Gal4/+ animals (two-tailed Student's t-test). n≥17 for all genotypes and parameters.

Fig. 5.

Disruption of actin polymers reduces synaptic localization of Mena, and Ena can increase actin-associated protein abundance postsynaptically. (A-B″) Images of synaptic boutons immunostained with HRP and EGFP. Scale bars: 2 μm. Muscle expression of UAS-EGFP-Mena^WT and UAS-EnaRNAi show peri-synaptic bouton localization similar to animals expressing endogenous Ena in the background (A-A″). EGFP-Mena^WT localization is disrupted by LatA treatment (B-B″). Quantitative immunohistochemical (QIHC) analysis shows a significant decrease in EGFP-Mena^WT intensity in the peri-synaptic bouton area (C). (D-F″) Images of the 6/7 NMJ immunostained for HRP and α-Spectrin: w¹¹¹⁸ control (D-D″), miR-8^Δ/Δ (E-E″) and UAS-Ena/how^24B-G4. Scale bars: 10 μm (D,E,F) and 5 μm (D′,D″,E′,E″,F′,F″). Quantitative immunohistochemistry shows a significant increase in α-Spectrin levels in miR-8^Δ/Δ and UAS-Ena/how^24B-Gal4 NMJs (G). Error bars indicate s.e.m. *P≤0.05, relative to control animals (two-tailed Student's t-test). n≥9 for all genotypes.

Fig. 6.

Capping protein β genetically interacts with miR-8 to antagonize Ena activity at the synapse. (A-D) Images of the 6/7 NMJ showing miR-8^Δ/+ (A), cpb^F44/+ (B), miR-8^Δ/+; cpb^F44/+(C) and how^24B-Gal4/UAS-cpbRNAi (D). Scale bars: 20 μm. (E) Quantification of synaptic boutons and NMJ expansion. Error bars indicate s.e.m. *P≤0.05, **P≤0.001, relative to control animals (two-tailed Student's t-test). n≥12 for all genotypes.

Fig. 7.

miR-8 mediates SSR elaboration through repression of Ena activity. (A-E) Analysis of electron micrographs of type Ib synaptic boutons at the 6/7 NMJ. (A-D) SSR (pseudocolored red) and presynaptic bouton (pseudocolored yellow). Representative image of a control w¹¹¹⁸ (A) synaptic bouton with surrounding SSR. Both miR-8^Δ/Δ (B) and how^24B-Gal4>UAS-Ena (C) display significant reduction in SSR area relative to control (A). Expression of how^24B-Gal4>UAS-FP₄-mito in a miR-8^Δ/Δ background significantly rescues all synaptic bouton elaboration defects observed in miR-8 homozygous mutants presynaptically (green) and postsynaptically (red, D). Scale bars: 500 nm. SSR thickness and area of w¹¹¹⁸ (n=23), miR-8^Δ/Δ (n=31), how^24B-Gal4>UAS-Ena (n=23) and miR-8^Δ/Δ; how^24B-Gal4>UAS-FP₄-mito (n=21). Error bars indicate s.e.m. ***P≤0.0001, **P≤0.001, relative to w¹¹¹⁸ control (two-tailed Student's t-test).

Fig. 8.

Abnormal synaptic transmission of miR-8 mutant NMJs. (A-C) Representative traces of evoked excitatory junction potentials (EJPs) recorded from muscle 6 of the NMJ: (A) w¹¹¹⁸, (B) miR-8^Δ2/Δ2 and (C) how^24B-Gal4/UAS-Ena. (D-G) Quantitative histograms show a significant reduction in evoked EJP amplitude (mV), mini frequency (s⁻¹) and mean quantal content of miR-8^Δ2/Δ2, but not how^24B-Gal4/UAS-Ena, relative to w¹¹¹⁸ control. ***P<10⁻⁴, **P<0.003 (two-tailed Student's t-test).