|
Figure 1. Visualization of β-actin mRNA trafficking in axons. (A) Fluorescent β-actin mRNA was electroporated into stage 24–27 Xenopus embryo eye primordia. The electroporated eye primordia were dissected and cultured for live imaging. (B,C) The synthetic β-actin mRNA was observed as fluorescent granules in cultured RGC axons and growth cones. (B) Live imaging of β-actin mRNA granules revealed different modes of motion and a broad range of speed (arrows and arrowhead). (C) Both anterograde (blue dots) and retrograde (yellow dots) movements were present. (D) OMX imaging of endogenous RNA labeled globally with Cy3-UTP (Cy3-RNA) and mitochondria. (E) Within the growth cone the β-actin mRNA granules tended to concentrate in the organelle-rich central domain, as indicated by MitoTracker Green. Scale bars, 5 μm for (B,D,E) and 2 μm for (C). See also Figure S1.
|
|
Figure 2. Dynamics of β-actin mRNA granules in axons. (A) An example of kymographs generated along axon shaft for kinetics analysis. (B) Population distribution of β-actin mRNA granule speeds for anterograde- and retrograde-moving granules in axon and growth cone. β-actin mRNA granule speeds showed biphasic distribution. One peak consisted of stationary and slow diffusive motion with speed < 0.5 μm/s. The more motile population showed a bell-shaped distribution. High speed (>2 μm/s) was observed for anterograde-moving granules along the axon shaft. (C) Proportion of anterograde-moving, retrograde-moving, and stationary β-actin mRNA granules in axons (n = 562 granules). (D) Anterograde-moving granules showed higher average speed than retrograde-moving granules. Nocodazole treatment greatly reduced kymographic tracks of fast-moving mRNA granules within 30 min, leaving mainly horizontal tracks from stationary mRNA granules. (F) Average instantaneous speed of all β-actin mRNA granules obtained from automated tracking is presented. The speed reduced to below 0.15 μm/sec upon 35-min nocodazole (Noco) treatment, but was unchanged upon 20-min cytochalasin D (CytoD) treatment. [F(2, 3643) = 146, p < 0.0001]. Error bars represent SEM. *p < 0.05, ***p < 0.001 (one-way ANOVA with Tukey multiple comparisons test for F).
|
|
Figure 3. Global netrin-1 stimulation changes the dynamics of β-actin mRNA granules. Axons were imaged at 1 frame per second from 1 min before bath application of netrin-1, and the imaging continued for up to 10 min after netrin application. (A) The numbers of β-actin mRNA granules passing three chosen locations (dashed lines) along the axon shaft were scored. The number of anterograde-moving granules increased upon netrin-1 treatment, reaching peak value at 5–6 min and returning to the base line by 8 min of netrin-1 treatment. The number of retrograde-moving granules showed moderate decrease with time upon netrin-1 treatment. (n = 9 axons) Anterograde vs. retrograde [F(7, 120) = 3.775, ###p = 0.001]; Vs. T = −1 min, **p < 0.01 (two-way ANOVA with Dunnett multiple-comparison test) (B–C) Global netrin-1 treatment enhanced β-actin mRNA localization to the growth cone periphery in a 3′UTR-dependent manner. (B) Quantification of mRNA granules localizing to the growth cone periphery. Netrin-1 treatment induced a marked increase in the number of full-length β-actin mRNA granules localizing to the growth cone periphery (n = 11 growth cones), a response that was abrogated for β-actin mRNA with truncated 3′UTR (n = 8 growth cones). Dotted lines represent least-squares fits to a Lorentzian function. [F(3, 178) = 22.01, ***p < 0.0001; extra sum-of-squares F test]. (C) Growth cones containing both full-length and Δ3′UTR β-actin mRNAs were fixed after 5-min bath netrin-1 treatment. The full-length mRNA granules localized further into the growth cone periphery (white arrows) while the Δ3′UTR granules remained in the central domain. Error bars represent SEM. See also Figure S2.
|
|
Figure 4. Netrin-1 gradient induces asymmetric localization of β-actin mRNA in growth cone. (A) Growth cones were imaged at 1 frame per second in the presence a netrin-1 gradient set up from a micropipette perpendicular to the growth cone. (B) The mean x-coordinate of all β-actin mRNA granules at each timepoint was compared to the mean x-coordinate at time 0. With the source of netrin-1 on the left, a negative shift value indicates a shift of β -actin mRNA granules toward the netrin-1 source. (C) Relative granule centroid shift values are plotted. (n = 15 growth cones for full length 3′UTR at 1fps for 5–15 min; n = 9 growth cones for Δ3′UTR at 1fps for 5–15 min) Black dotted lines represent least-square fits to a linear function. [F(1, 407) = 98.27, ***p < 0.0001; extra sum-of-squares F test] (D) The distribution of full-length β-actin mRNA granules are binned and show a rapid shift toward the source of netrin-1. The bar chart shows the averages from 3 growth cones. (E) β-actin mRNA with truncated 3′UTR did not show asymmetric distribution toward netrin-1 gradient. The bar chart shows the averages from 6 growth cones. (F) Time lapse images showing a prominent shift of full-length β-actin mRNA toward the source of netrin gradient (toward the left as denoted by green arrows; midline is denoted by dashed red line. (F; far right) Photobleaching-corrected, denoised time lapses images were stacked and maximally projected. The temporal-color coded image reveals the full-length β-actin mRNA shift toward netrin gradient is both asymmetrical and fast. Scale bars, 5 μm. Error bars represent SEM.
|
|
Figure 5. Netrin-1 induced RNA movement to growth cone periphery depends on dynamic microtubules and not F-actin. (A) Fluorescence micrographs show the relative distributions of F-actin (phalloidin staining) and dynamic microtubules (tyryosinated-tubulin) in distal axon. (B–C) Dynamic microtubule and F-actin were selectively disrupted to test which cytoskeletal element was required for facilitating netrin-1 induced peripheral mRNA localization. Unlike with intact cytoskeleton (B, Netrin-1 only; n = 17 growth cones), disrupting dynamic microtubules with colchicine (C, Colchicine + Netrin-1; n = 4 growth cones) abolished the netrin-1-induced increase in peripheral mRNA localization. In the presence of cytochalasin D (C, Cytochalasin D + Netrin-1; n = 9 growth cones), which disrupts F-actin, mRNA granules still exhibited increased peripheral localization in response to netrin-1 stimulation. Each panel in (B,C) shows the same growth cone before and after netrin-1 treatment. (D) Quantification of the number of peripherally localized mRNA granules. Dotted lines represent least-squares fits to a Lorentzian function. (Netrin vs. Netrin + Cytochalasin D: F(3, 243) = 2.405, p = 0.068; Netrin vs. Netrin + Colchicine: F(3, 222) = 25.75, ***p < 0.0001; Netrin + Cytochalasin D vs. Netrin + Colchicine: F(3, 119) = 57.09, ###p < 0.0001; extra sum-of-squares F test) (E) The peripheral distribution of Cy3- β-actin mRNA granules in live, netrin-1 bath treated growth cones (left) was compared to that of dynamic microtubules (right), which was revealed by tyrosinated tubulin immunostaining after immediate fixation of the live growth cones. The distance by which dynamic microtubules probed into growth cone periphery was comparable to that reached by netrin-1-recruited peripheral mRNA granules. (F) Live imaging of Cy3-β-actin mRNA granules with GFP-tagged EB1 captured outward-moving mRNA granules in filopodia trailing behind EB1-GFP comets that tipped the dynamic microtubules (image sequence of region in square). (G) Dynamic microtubules in growth cone periphery do not show asymmetric distribution concurrent with or prior to β-actin mRNA polarization upon netrin-1 gradient stimulation. (n = 4 growth cones) Scale bars, 10 μm. Error bars represent SEM.
|
|
Figure 7. Endogenous RNA worms in vitro. (A–E) Endogenous RNA was labeled by fluorescent-UTP. Elongated RNA species, which we termed “RNA worms,” were sporadically observed during live imaging. (A,B) An RNA worm showing anterograde movement in the axon shaft (A) and more subtle twisting-turning dynamics in the distal axon/growth cone (B). (C,D) RNA worms are not labeled by MitoTracker, as shown in both axon shaft (C) and growth cone (D). (E) RNA worms are not labeled by ER-Tracker. Scale bars, 10 μm for (A,B) and 5 μm for (C–E). See also Figures S3–5.
|
|
Figure 8. Endogenous RNA worms in vivo. (A) Endogenous RNA was labeled by fluorescent-UTP and the RGC axons were labeled by mGFP for visualization in vivo. The example shows the anterograde movement of the RNA worm toward the growth cone, followed by instantaneous condensation and fission into at least two separate RNA granules with globular morphologies. (B) Quantification of the speed of the RNA worm movement for (A). Scale bars, 5 μm for upper panel and 1 μm for lower panel of (A).
|
|
Figure 6. Super-resolution microscopy shows direct association between β-actin mRNA and dynamic microtubules. (A–E) Fluorescence in situ hybridization (FISH) for β-actin mRNA was followed by immunostaining for tyrosinated-tubulin in growth cone. (B–E) FISH puncta overlap with dynamic microtubules and 3-D rendering supports direct association. 64.08 ± 6.44% of β-actin mRNA colocalized with dynamic microtubule (n = 8 growth cones; 280 identified β-actin mRNA granules). Scale bar, 4 μm.
|
