Nucleus to Synapse Nesprin1 Railroad Tracks Direct Synapse Maturation through RNA Localization

Abstract

An important mechanism underlying synapse development and plasticity is the localization of mRNAs that travel from the nucleus to synaptic sites. Here we demonstrate that the giant nuclear-associated Nesprin1 (dNesp1) forms striated F-actin-based filaments, which we dubbed "railroad tracks," that span from muscle nuclei to postsynaptic sites at the neuromuscular junction in Drosophila. These railroad tracks specifically wrap around immature boutons formed during development and in response to electrical activity. In the absence of dNesp1, mRNAs normally localized at postsynaptic sites are lacking and synaptic maturation is inhibited. This dNesp1 function does not depend on direct association of dNesp1 isoforms with the nuclear envelope. We also show that dNesp1 functions with an unconventional myosin, Myo1D, and that both dNesp1 and Myo1D are mutually required for their localization to immature boutons. These studies unravel a novel pathway directing the transport of mRNAs from the nucleus to postsynaptic sites during synaptic maturation. VIDEO ABSTRACT.

Figure 1. magi and par6 mRNAs are depleted at the NMJ of dnesp1^sZ75 mutants (see also Fig.S1)

(A-D) Larval NMJs labeled with anti-HRP and FISH to (A,B) a par6 probe and (C,D) a magi probe in (A,C) wild type and (B,D) dnesp1^sZ75 mutants. (E-H) Larval NMJs labeled with anti-HRP and either (E, F) anti-Par6 or (G,H) anti-MAGI in the indicated genotypes. (I) Immunoprecipitation of dNesp1 in wild type and dnesp1^sZ75 mutants. (J-L) Quantification of postsynaptic signal intensity normalized to wild type control (see Methods) for (J) par6 and magi RNA, (K) Par6 and MAGI protein and (L) DLG RNA and protein, in wild type, dnesp1^sZ75, and dnesp1^sZ75 mutants containing 2 copies of a duplication of the dnesp1 locus (rescue). (M) Ratio of FISH signal (intensity:background) in the region surrounding the nucleus in the above genotypes. (N-P) Muscle nuclei labeled with Hoechst, and FISH to par6 in (N) wild type, (O) dnesp1^sZ75 mutants, and (P) rescue. (Q) qPCR of par6 mRNA from wild type and dnesp1^sZ75 mutants normalized to ef1α48D. Calibration scale(µm): A-H:4; N-P:6. Error bars represent ±SEM (*p<0.05, **p<0.001, ***p<0.0001). Number of samples (animals:arbors/muscles)= (J) 20:38,18:36,10:20,9:18,10:20,9:18; (K) 8:15,8:16,8:15,8:16,8:15,8:16; (L) 9:18,9:18,9:18,9:18; (M) 7:13,7:14,7:13,6:12,6:12,6:12; (Q) 3. Image panels correspond to single confocal slices.

Figure 2. dNesp1 railroad tracks specifically wrap around ghost boutons (see also Fig.S2)

(A-I) Body wall muscles and NMJs labeled with anti-dNesp1, anti-HRP and anti-DLG in (A,B,D-G, I) wild type, (C) dnesp1^ΔKASH and (H) dnesp1^sZ75 mutants. Arrowheads point to dNesp1 railroad tracks; z=Z-line; n=nucleus. (D,E) are high magnification views of the regions marked by the boxes in (A,I, respectively) showing the presence of ghost boutons (arrows), which are labeled with anti-dNesp1 and anti-HRP, but not with anti-DLG. (F) STED image of a ghost bouton, showing dNesp1 label wrapping around a ghost bouton. (J) Quantification of anti-DLG, anti-dNesp1, and anti-HRP signal at the bouton border (see Methods; N=6) normalized to HRP intensity. (K) Representative relative signal intensity across the midline of a mature and a ghost bouton. (L) Relative levels of par6 RNA immunoprecipitated with dNesp1 as measured by real time PCR (N=3 biological replicates). Calibration scale(µm): A-C,G-I:10; D,E:4; F:2. Number of samples in (J) is 6 boutons. Image panels represent single confocal slices. Error bars represent ±SEM (*p<0.05, **p<0.001, ***p<0.0001).

Figure 3. dNesp1 is localized to new ghost boutons induced by spaced stimulation (see Fig. 5 for quantification)

(A,B) Wild type NMJ arbors labeled with anti-dNesp1, anti-HRP and anti-DLG in (A) unstimulated samples (box in A1 is shown at higher magnification in A2), and (B) samples subjected to high K⁺ stimulation (boxes in B1 are shown at high magnification in B2 and B3), showing that activity-induced ghost boutons (arrows) are wrapped by dNesp1. Calibration scale(µm): A1,B1:50; A2,B2,B3:20. Image panels represent single confocal slices.

Figure 4. NMJ expansion and synaptic bouton maturation depend on dnesp1, but are independent from the KASH domain (see Fig.S3)

(A-E) Larval NMJs labeled with anti-HRP and anti-DLG shown at (A,C) low and (B,D,E) high magnification from (A,B) wild type and (C-E) dnesp1^sZ75 mutant. Arrows point to ghost boutons, which are devoid of DLG label. (F,G) Quantification of (F) synaptic bouton number and (G) ghost bouton number (divided by total bouton number and normalized to wild type controls) in the indicated genotypes. Calibration scale(µm): A,C:40; B,D,E:12. Error bars represent ±SEM (*p<0.05, **p<0.001, ***p<0.0001). Number of samples are (animals:arbors; from left to right) (F) 31:62,28:56,33:65,23:45,28:55,12:24,12:24,12:24,12:24,12:24. (G) 31:62,28:56,33:65,23:46,29:57,12:24,12:24,12:24,12:24,12:24. Image panels correspond to (A,C) a maximal intensity Z-stack projection and (B,D,E) single confocal slices.

Figure 5. Mutations in dnesp1 alter activity-dependent bouton formation and dNesp1 railroad tracks contain F-actin (see Fig.S4)

(A-D) NMJ arbors labeled with anti-HRP and anti-DLG from (A,C) unstimulated controls and (B,D) body wall muscles subjected to high K⁺ spaced stimulation, in (A,B) wild type and (C,D) dnesp1^sZ75 mutants. Arrows point to ghost boutons. (E,G,H) Larval body wall muscle labeled with anti-dNesp1, fluorescent phalloidin to mark F-actin, and anti-HRP showing that (E) long dNesp1 railroad tracks (arrow) spanning from the nucleus to the NMJ contain F-actin, and (G,H) that these railroad tracks are composed of staggered F-actin and dNesp1 striations. (F) Larval body wall muscle labeled with anti-tubulin and dNesp1, showing that dNesp1 and microtubules are non-overlapping. Box in F1 is shown at high magnification in F2. (I,J) Quantification of ghost bouton number (divided by total bouton number and normalized to wild type controls) from animals stimulated with spaced (I) high K⁺ or (J) Chr2 stimulation in the indicated genotypes showing that dnesp1^sZ75 mutants have reduced activity-dependent induction of ghost boutons and synaptic bouton maturation. Calibration scale(µm): A-D (panels 1,2):18; A-D (panel 3):6; E:7, F:10; and G,H:3. Error bars represent ±SEM (*p<0.05, **p<0.001, ***p<0.0001). Number of samples (animals:arbors; left to right)=(I) 31:62,33:65,11:21,11:22;(J) 30:59,33:66,28:56,30:59,19:37,28:56. Image panels correspond to (A–D) maximal intensity Z-stack projections and (E–F) single confocal slices.

Figure 6. Myo31DF colocalizes with dNesp1 at ghost boutons in a mutually dependent manner and mutations in myo31DF mimic dnesp1^sZ75 mutant phenotypes

(A-E, I) NMJs from stimulated (A,C,D) wild type and (B,E,I) myo31DF^K1/K2, labeled with: (A,B) anti-Myo31DF (Myo1), anti-HRP, and anti-DLG, showing (A) enrichment of Myo31DF at ghost boutons and (B) elimination of Myo31DF signal in myo31DF mutants; (C) anti-Myo31DF and anti-dNesp1 showing colocalization of both proteins at ghost boutons (insets in C4 show an area of colocalization between dNesp1 and Myo31DF at railroad tracks; (D,E) anti-dNesp1, anti-HRP and anti-DLG showing that while activity induced ghost boutons in (D) wild type are surrounded by dNesp1, (E)myo31DF mutants show no dNesp1 signal at these boutons; (F) anti-Myo31DF and anti-HRP showing that the localization of Myo31DF at the NMJ is disrupted in dnesp1^sZ75 mutants. (G-L) Quantification of (G) Pearson’s colocalization coefficient of Myo31DF and dNesp1 signal at different regions, (H,K) total bouton number and (I,J,L) ghost bouton number (divided by total bouton number and normalized to wild type controls) in the indicated phenotypes. Preparations in (J) were stimulated with high K⁺ saline. Calibration scale(µm): A(1–3):40; B-E(1–3), I:30, A-D(4), I2:12. Error bars represent ±SEM (*p<0.05, **p<0.001, ***p<0.0001). Number of samples (animals:arbors) (G) 6,6,6; (H) 31:62,33:65,27:54,26:51,5:10; (I) 31:62,33:65,27:54,19:37,5:10; (J) 30:59,11:21,13:25,19:37,5:9,5:10; (K,L) 31:62,28:56,27:54,18:37. Image panels represent single confocal slices.

Figure 7. Postsynaptic par-6 and magi mRNA and protein are decreased in myo31DF mutants

(A-D) Larval NMJs labeled with anti-HRP and FISH to (A,B) par6 and (C,D) magi transcripts in (A,C) wild type and (B,D) myo31DF^K2 mutants. (E-H) Larval NMJs labeled with anti-HRP and either (E, F) anti-Par6 or (G,H) anti-MAGI in (E,G) wild type and (F,H) myo31DF^K2 mutants. (I-J) Quantification of postsynaptic Par6 and MAGI (I) RNA and (J) protein signals in the indicated genotypes. Calibration scale: (A–H) 8µm. Error bars represent ±SEM (*p<0.05, **p<0.001, ***p<0.0001). Sample numbers (animals:arbors) (I) 10:20,11:21,7:25,8:20,10:21,15:25; (J) 6:15,6:28,6:21,6:15,6:22,6:20. Image panels represent single confocal slices.