Figure 1.
Acute spaced stimulation induces the formation of undifferentiated ghost boutons.
(A) A representative Canton S third-instar NMJ at larval muscle 6/7 in abdominal segment 3 that was double stained with antibodies against the presynaptic membrane marker, horseradish peroxidase (HRP; red) and postsynaptic discs large (DLG; green) after being subjected to 0x (control) or 5x high K+ (90 mM) spaced depolarization. Arrows point to ghost boutons (HRP+ DLG-). Each inset corresponds to the indicated region (dashed box) and is blown up to help visualize ghost boutons (red HRP+ presynaptic extensions). Scale bar = 20 µM. (B) Quantification of the number of ghost boutons per NMJ in Canton S (high K+ stimulated) or animals expressing the transgenic light-gated ion channel, ChR2, in motor neurons (light stimulated; the UAS-ChR2 line was crossed to C380-Gal4). These data are highly consistent with published results [12], [14], [15]. 1–2 unpaired NMJs per larval treatment group. Error bars indicate the mean ± SEM. STATISTICS: Student’s t-test. ** p<0.01 and **** p<0.0001.
Figure 2.
miRNA expression profile of the Drosophila larval CNS.
(A) Flowchart showing miRNA microarray analysis. The expression of 79 mature miRNAs was detected in the Drosophila larval CNS. (B) Heat map showing fold-change of miRNAs in the CNS of w1118 (control or “C”), ChR2 mock stimulated (unstimulated or “U”), and ChR2 light stimulated (stimulated or “S”) larvae. No neuronal miRNA exhibited a greater than 2-fold change in expression levels following ChR2 light stimulation. The largest differences observed were between the w1118 and C380-Gal4 x UAS-ChR2 genotypes.
Figure 3.
miRs-1, -8, -289, -314, and -958 are rapidly downregulated by activity.
(A) Flowchart showing RT-qPCR analysis. 62 of 79 mature neuronal miRNAs were screened by RT-qPCR (see methods). (B) Five miRNAs were determined to exhibit either a 2-fold or statistically significant downregulation in expression levels following spaced high K+ depolarization. Error bars indicate the mean ± SEM. STATISTICS: One-way ANOVA with a Tukey’s post-hoc test (n = 3 replicates). * p<0.05. (C) Table showing fold downregulation and p-values for each activity-regulated miRNA. While not statistically significant, miRs-12 and -304 show a 1.7-fold activity-dependent downregulation. miRs-12 and -304 are both located in the miR-12/304/283 cluster.
Figure 4.
miRs-8, -289, and -958 are required for activity-dependent ghost bouton formation at the larval NMJ.
(A) Our working model for miRNA-mediated control of activity-dependent synaptic growth at the larval NMJ. In an unstimulated motor neuron, key activity-regulated miRNAs negatively regulate the expression of target mRNAs involved in the control of activity-dependent synaptic growth. In contrast, our data suggests that acute spaced stimulation results in the rapid downregulation of mature miRNA levels. This would result in the increased translation of target mRNAs and subsequent rapid activity-dependent synaptic growth. (B) Transgenic pri-miRNAs for miRs-1, -8, -289, -314, or -958 were misexpressed in larval motor neurons using the C380-Gal4 driver (e.g. genotype C380-Gal4/+; UAS-pri-miR/+). Also examined were miR-8 knockdown (C380-Gal4/+; UAS-miR-8SP/+) or loss-of-function (mir-8Δ1/Δ2) lines. Ghost boutons per NMJ were quantified in indicated genotypes under conditions of mock (0x), intermediate (3x) or high (5x) K+ stimulation as indicated. Compared to their matched 0x unstimulated controls, the misexpression of miRs-8, -289, and -958 did not result in a significant increase in ghost bouton numbers following 5x high K+ stimulation. In contrast, miR-8 knockdown (UAS-miR-8SP) and deletion (mir-8Δ1/Δ2) larvae showed an enhanced ability to respond to intermediate stimulation. (C) Quantification of the combined number of synaptic boutons (stimulated and unstimulated) at the same NMJs assayed in (B). No significant difference in total bouton number was observed between unstimulated and stimulated NMJ within any miRNA genotype (data not shown). Note that only miR-8 has a significant effect on total bouton numbers when compared to controls (C380-Gal4/+). This suggests that miR-8 has a function in both ghost bouton maturation and development. (B-C) OE = overexpression; ROF = reduction-in-function; LOF = loss-of-function. The numbers located at the bottom of each column indicate the number of NMJs analyzed for each genotype and/or treatment group. Except in C380-Gal4/+ there are 2 paired NMJs per treatment group. Error bars indicate the mean ± SEM. (B-C) STATISTICS: Kruskal-Wallis multiple comparison analysis with a Dunn’s post-hoc test. n.s. = not significant * p<0.05 ** p<0.01 *** p<0.001 **** p<0.0001.
Figure 5.
Predicted targets of miRs-8 and -289 are found in neuron-related enriched functional annotation clusters.
(A) Venn diagram showing predicted mRNA targets for activity-regulated miRNAs. Notably, 33 mRNAs have putative binding sites for miRs-8, -289, and -958. 282 mRNAs have putative binding sites for miRs-8 and -289. (B-C) Functional annotation cluster analysis for predicted targets of miRs-8, -289 and both miRs-8 and -289. Only clusters enriched with targets significantly enriched in clusters involved in the control of neuronal processes are shown here. Note that the 304 predicted targets for miR-958 were not enriched in these clusters. Enrichment scores for each cluster are indicated [29], [30]. Genes within each category are indicated as a percentage of total genes (gray bar) and fold enrichment (blue bar) over expected number of genes in that category in the Drosophila genome. Statistical significance (Benjamini corrected p-values) for each category is indicated to right of gray columns.
Table 1.
Putative mRNA targets for co-regulation by miRs-8 and -289 that also map to a neuron-related functional annotation cluster (GO term enrichment).
Table 2.
Putative mRNA targets for co-regulation by miRs-8, -289, and -958 (miRBase).
Figure 6.
Drosophila lar and wg are authentic targets for translational repression by activity-regulated miRNAs.
(A and B) Reporter plasmids constitutively expressing firefly luciferase (Fluc) flanked by the 3′UTRs of lar and wg were cotransfected into S2 cells with plasmids expressing the indicated miRNA primary transcripts as indicated. Renilla luciferase (Rluc) was included as a transfection control. Fluc activity was normalized to Rluc activity in three independent experiments. Normalized Fluc activities in the absence of miRNA-expressing vectors (emply vector controls) were set to 1. (A) The Drosophila lar 3′UTR has one binding site each for miRs-1, -8, -289, and -958. miR-1 was not tested because it did not negatively regulate activity-dependent synaptic growth. miRs-8, -289, and -958 all significantly repress lar Fluc reporter activity. In contrast, a miRNA with no predicted binding site (miR-9a) has no effect on Fluc activity. (B) The wg 3′UTR has three predicted binding sites for miR-289 and two for miR-8. Interestingly, only miR-8 is capable of repressing wg Fluc reporter activity. miR-289, -958 nor -9a (not predicted to bind) had an effect of Fluc activity. In both (A) and (B) error bars indicate the mean ± SEM (n = 3). (C) Transgenic hairpin RNAi constructs targeting lar and wg were misexpressed in larval motor neurons using the C380-Gal4 driver (e.g. genotype C380-Gal4/+; UAS-RNAihairpin/+). Note that both hairpin constructs completely prevented activity-dependent ghost bouton formation. (D) Quantification of the total number of synaptic boutons at the same NMJs assayed in above (C). Neither hairpin construct had a significant effect on total bouton numbers compared to controls (C380-Gal4/+) suggesting that activity-dependent processes may be more sensitive to disruption of genes involved in synaptic growth pathways. STATISTICS: (A-B; D) One-way ANOVA with a Dunnett’s post-hoc test. (C) Kruskal-Wallis multiple comparison analysis with a Dunn’s post-hoc test. n.s. = not significant ** p<0.01 **** p<0.0001.