Repeat-element RNAs integrate a neuronal growth circuit

Abstract

Neuronal growth and regeneration are regulated by local translation of mRNAs in axons. We examined RNA polyadenylation changes upon sensory neuron injury and found upregulation of a subset of polyadenylated B2-SINE repeat elements, hereby termed GI-SINEs (growth-inducing B2-SINEs). GI-SINEs are induced from ATF3 and other AP-1 promoter-associated extragenic loci in injured sensory neurons but are not upregulated in lesioned retinal ganglion neurons. Exogenous GI-SINE expression elicited axonal growth in injured sensory, retinal, and corticospinal tract neurons. GI-SINEs interact with ribosomal proteins and nucleolin, an axon-growth-regulating RNA-binding protein, to regulate translation in neuronal cytoplasm. Finally, antisense oligos against GI-SINEs perturb sensory neuron outgrowth and nucleolin-ribosome interactions. Thus, a specific subfamily of transposable elements is integral to a physiological circuit linking AP-1 transcription with localized RNA translation.

Keywords: RNA localization; Short Interspersed Nuclear Elements; axon growth; axonal transport; local translation; nerve injury; neuronal length sensing; non-coding RNA.

Graphical abstract

Figure S1

Polyadenylated B2-SINEs in DRG neurons and N2a cells, related to Figure 1 (A) Model of intrinsic neuron growth regulation by motor-driven RNA localization. Anterograde transport of mRNAs on nucleolin (Nucl) and kinesin (K), calcium-induced local translation of nucleolin-associated mRNAs, including mTor and importin β1 (β), and retrograde transport of the locally synthesized proteins on dynein (D) and importin α (α), regulates protein synthesis and growth. (B) Volcano plots of sciatic-nerve-injured/naive (72 h) differential expression of B2-SINE RNAs in L4/L5 DRG neurons (data from Figure 1A). n = 19,279 elements. Red elements are p < 0.05 (marked by dotted line). (C) Examples of polyadenylation site (PAS) mapping reads in L4/L5 DRG neurons. (1) Genome browser tracks showing reads mapping to 3′ end of Mrpl37. Injury-induced PAS reads were also detected in a Mrpl37 intron, mapping on a SINE. (2) A chromosome 14 intergenic locus with a highly regulated SINE-PAS (closed triangle) alongside a modestly induced SINE-PAS (open triangle). (D) In vitro removal of B2 RNA polyadenylated tails. Total Neuro2a cell RNA was treated with RNase H in the presence of (dT)₂₀, degrading A-rich RNA RNA-DNA duplexes, collapsing the smeared B2 signal to a single band. 7SL was blotted and shown as loading control. (E) Volcano plots of optic nerve injury/naive differential expression of B2-SINE RNAs at 72 h after injury in RGC (data from Figure 1F). n = 8,028 elements. Red elements are p < 0.05 (marked by dotted line).

Figure 1

B2-SINEs are upregulated in DRG but not RGC neurons after nerve injury (A) Polyadenylated B2-SINE RNA expression determined by Quantseq REV, 75 bp single-end read RNA-seq from L4/L5 DRG neurons 24 and 72 h after sciatic nerve injury. Injury-dependent log₂ fold changes (log₂FC) of uniquely mapped B2 elements. Boxplot midline shows medians, and whiskers indicate 10th and 90th percentiles. RE (non-B2)—all repeat-element-mapped RNAs excluding B2-SINEs. ^∗∗∗∗p < 0.0001, Mann-Whitney test, n = 23,212, 13,686, 19,279, and 9,121 elements for B2 (24 h), RE (24 h), B2 (72 h) and RE (72 h), respectively. (B) Northern blot with consensus B2-SINE probe on total L4/L5 DRG RNA 72 h after sciatic nerve injury. Representative blot shown on the left, 7SL RNA probed as loading control. Right: B2/7SL ratios from quantification of all bands visualized relative to the mean of naive controls. n = 4 mice; ^∗p < 0.05, t test, means ± SEM. (C) qRT-PCR on the same samples as (B) with primers recognizing consensus mouse B2-SINE sequence. 18S RNA was used for normalization, data are relative to the mean of naive controls. n = 4 mice; ^∗∗p < 0.01, t test, means ± SEM. (D) Analysis of B2-SINE RNA expression by whole transcriptome sequencing of DRG neurons from an independent sciatic nerve injury time course experiment. Data presentation as in (A), n = 18,905. (E) All non-B2 repeat-element RNAs from the experiment described in (D), n = 7,034. (F) B2-SINE RNA expression by whole transcriptome sequencing of retinal ganglion cells (RGCs) after optic nerve injury. Data presentation as in (A), n = 8,028. (G) All non-B2 repeat-element RNAs from the experiment described in (F), n = 12,319. (H) Statistical comparison of time courses shown in (D)–(G). Fold changes were modeled with a linear mixed effects model, with time, group and set as fixed factors, and a random intercept per element. ^∗∗∗p < 0.001 between DRG B2 and RGC B2. (I) Proportions of reads mapping to different repeat-element classes in DRG and RGC datasets shown in (E) and (G) (summing reads from all data points). All repeat-element-mapping reads were taken as 100%. See also Figure S1 and Table S1.

Figure S2

Validation of exogenous B2-SINE expression, related to Figure 2 (A and B) B2-SINE consensus northern blot on human HEK293 cells transfected with AAV-B2 vs. untransfected HEK293 and mouse N2a cells. RNA size marker on left. No B2-SINE signal in untransfected HEK293, whereas transfection generates both non-polyadenylated and polyadenylated B2-SINE. (C) FISH of exogenous B2-SINE in retinal ganglion cells (RGCs) transduced with AAV-B2. B2-SINE FISH in white, YFP co-expressed from AAV-B2 in magenta. (D) Quantification of (C), n = 7 (control), 13 (B2) transduced retinas (72 and 189 cells), means ± SEM, ^∗∗p < 0.01, Mann-Whitney test. (E and F) Representative images of whole retinas injected with AAV-B2, AAV-control, or not injected, confirming specific B2-SINE overexpression. Scale bars, 100 μm.

Figure 2

Exogenous B2-SINE expression promotes axonal growth (A) DRG neuron spot cultures transduced with B2-SINE AAV (AAV-B2) or control vector (AAV-control) underwent axotomy and axon growth was tracked by live imaging of AAV-expressed YFP (yellow fluorescent protein). Representative images of axon outgrowth at the site of axotomy at indicated time points following axotomy. Scale bar, 100 μm. (B) Processive growth rates across imaging sequences for individual axons are shown based on distance extended in perpendicular to the axotomy line. n = 23 (AAV-control) or 26 (AAV-B2) axons from three independent cultures after ROUT (Robust regression and Outlier removal) outlier analysis with Q = 1%; mean ± SEM, ^∗∗∗∗p < 0.0001, Mann-Whitney test. (C) Experiment timeline for transduction of AAV2 for exogenous expression of consensus B2-SINE in mouse retinal ganglion cells (RGCs). Crush lesions of the optic nerve were done 2 weeks after intraocular injection with AAV-B2 or AAV-control. 12 days later, mice received intraocular injections with the CTB-555 axonal tracer (cholera toxin B subunit) and 2 days later animals were sacrificed for histology. (D) Representative images of axonal growth after optic nerve crush injury, visualized with CTB-555. Scale bar, 100 μm. (E) RGC axon numbers at indicated distances from the crush site, normalized to control at 400 μm, n = 9 mice. Mean ± SEM, ^∗∗p < 0.01, ^∗∗∗p < 0.001, t test. (F) Experiment timeline in corticospinal tract injury model. AAVs were injected into the mouse motor cortex 1 week before dorsal column spinal cord lesions at T1 level (SC lesion). Two weeks later an anterograde AAV2 axonal tracer expressing tdTomato was injected into the motor cortex and 2 weeks after that animals were sacrificed for histology. (G) Representative tdTomato tracer images of axons growing into gray matter above the lesion. Horizontal lines mark dorsal-to-ventral distance from tract-gray matter border. Scale bar, 100 μm. (H) Axonal growth was normalized to axon numbers in main tract and calculated as percent of mean control axon growth. Data points show mean percentages at 100 μm intervals over 0–500 μm from the border. Results are from 3 (AAV-control) or 4 (AAV-B2) mice. ^∗∗∗p < 0.001 for effect of AAV-treatment on outgrowth by two-way ANOVA. See also Figure S2 and Videos S1 and S2.

Figure 3

GI-SINES are a specifically regulated subset of B2 elements (A) Enrichment of B2-SINE subtypes in naive and injured DRG at 24 and 72 h after sciatic nerve crush. Paired-end, 150 bp reads total RNA-seq was carried out and B2-SINEs were mapped using RSEM (RNA-seq by Expectation Maximization) algorithm for injury/naive differential expression analysis. Boxplot whiskers show 10th and 90th percentiles. n = 4,205, 2,282, 2,559, and 886 elements for B2mm1, B2mm2, B3, and B3A subfamilies, respectively. (B) Genomic loci distribution of B2 elements significantly upregulated with fold change > 2 at 72 h after injury (GI-SINEs). n = 447 individual loci. (C) Profile of GI-SINEs association with Tn5-accessible loci by ATAC-seq analysis. Loci of GI-SINEs and non-significantly regulated B2-SINEs expressed 72 h after injury were aligned to ATAC-seq datasets of either sciatic-nerve-crushed or sham-operated DRG neurons (NeuN+). Mean ± SEM are shown for ATAC-seq read densities across the span of the B2 metagene loci for each paired alignment. (D) Transcription-factor-binding site (TFBS) predictive analysis using TOBIAS to identify enrichment of TFBS for GI-SINE loci. Data points in green and red are TFBS significantly enriched in injury and sham datasets, respectively. Yellow points depict a cluster of AP-1 TFBS enriched in GI-SINE loci after injury. (E) B2-SINE expression in cultured DRG neurons transduced with the AP-1 dominant-negative A-Fos, GFP in PHP.S AAV, or left untransfected (Unt.) were taken from differential expression analysis of RNA-seq (150 bp reads, paired-end sequencing). Fold change ratios of B2-SINE RNA that match GI-SINE loci (GI-SINE RNA), plotted in boxplots, as described in (A). n = 73 GI-SINE loci, from 9 independent cultures. ^∗∗∗p < 0.0001, Kruskal-Wallis with Dunn’s multiple comparison test. (F) Differential expression of GI-SINE RNA in sciatic-nerve-injured vs. naive DRG in sensory neuron conditional knockout of ATF3 (Advillin-Cre⁺/ATF3-exon3-flox^+/+) vs. wild-type (WT) mice. RNA was collected 72 h after injury for RNA-seq and expression analysis was as described in (A). B2-SINEs with injury/naive fold change > 2, p < 0.05 in ATF3-wt mice were annotated as GI-SINEs. Sciatic-injured and naive DRG RNA were each taken from 9 (ATF3-wt) and 8 mice (ATF3-cKO). Data are plotted in boxplots, as in (A), n = 311 GI-SINE loci, ^∗∗∗∗p < 0.0001, Mann-Whitney test. (G) Volcano plots of injury/naive differential expression of GI-SINE RNA. Data are from the experiment described in (F). Red data points are elements with p < 0.05 (marked by dotted line). Dashed line marks injury/naive fold change = 2 (fold change criteria for GI-SINE), n = 311 elements. See also Figure S3 and Table S2.

Figure S3

Supporting data for specific regulation of GI-SINES, related to Figure 3 (A) Volcano plot of B2-SINE RNAs expression in L4/L5 DRG 72 h after injury, as described in Figure 3A. Red elements are p < 0.05 (dotted line). Dashed line is 2× threshold for GI-SINE definition. (B) Genomic region annotation of B2-SINE loci not regulated in injured DRG neurons (left, 1,887 elements) and distribution of annotation categories across mm10 mouse genome (right). (C) Correlation of 72 h injury/naive expression changes from intragenic GI-SINEs with their associated coding genes (n = 144). Pearson’s correlation linear fit with 95% confidence intervals. (D) Injury/naive fold changes of Pol III-transcribed ncRNA genes and SINE subfamilies (group mean) in DRG after sciatic nerve injury. (E) RGC ATAC-seq association profiles of GI-SINEs identified in injured DRG. Loci of significantly upregulated GI-SINEs and non-significantly regulated B2-SINEs from DRG data were aligned to ATAC-seq datasets of RGC 3 days following optic nerve injury or sham operation, as previously described. Note the lack of association of SINE loci with accessible chromatin in DRG. (F) Mean expression fold changes of injury-induced and regeneration-associated genes, n = 9 for AAV-PHP.S-A-Fos dominant negative, control GFP transduced, and untransduced control (A-Fos/GFP and GFP/Unt.), n = 4 sciatic-nerve-injured mice (injury/naive 24 h and 72 h). (G and H) Mean expression fold changes and p values of mapped B2-SINE loci from Figure 3E, compared between A-Fos vs. GFP or GFP vs. untransduced. n = 9. Red elements are p < 0.05 (dotted line). (I) Mean expression fold change of injury-induced and regeneration-associated genes between ATF3-wt and sensory neuron ATF3-cKO (Advillin-Cre⁺/ATF3-exon3-fl^+/+) mice. n = 8 (ATF3-cKO) and n = 9 (ATF3-wt).

Figure 4

B2-SINE RNA interacts with the protein translation machinery (A) Schematic of B2-SINE interactome analysis (consensus structure model based on Espinoza et al. and Ponicsan et al.⁴⁵), blue and red dashed lines delineate the two biotin-conjugated constructs used for pull-downs. (B) Top 20 enriched proteins co-precipitated with B2-SINE baits. Nucleolin (Ncl) is highlighted in magenta and ribosomal proteins are in green. Enrichment scores over mock (biotin only) pull-downs were calculated for each protein hit. Heatmap depicts mean log₂FC (vs. biotin-only control) of three independent pull-down experiments for B2_5′ and B2_3′ pull-downs. (C) GO term analysis of cellular processes and subcellular component enrichment in B2-SINE pull-downs (n = 3). GO terms presented are terminal nodes with enrichment FDR (false discovery rate) adjusted p < 0.05. (D) Heatmap of B2-SINE-bound proteins enriched in the GO term “ribosome cellular component.” Mean log₂ of PSM fold change over control, n = 3 independent pull-downs. (E) Western blot validation of B2-SINE binding to nucleolin in axoplasm. Representative blot from three independent experiments. (F) Nucleolin-B2-SINE interaction requires the nucleolin GAR domain. DRG extracts were taken from nucleolin-GAR^+/Δ mice for B2-SINE RNA bait pull-downs. Representative nucleolin western blots of pull-down and input samples from three independent experiments are shown. Arrowheads mark bands of nucleolin WT (full length) and nucleolin-ΔGAR. (G) RT-qPCR analysis of RNA co-immunoprecipitation (RIP) of endogenous B2-SINE and nucleolin in DRG extracts. Nucleolin IP/input ratio for each transcript is normalized to mean IgG IP/input ratio. Mean ± SEM from 3 RIP experiments on extracts from 2 mice each. (H) B2-SINE upregulation in ribosome-associated RNA from DRG neurons 72 h after sciatic nerve injury. Advillin-RiboTag mouse DRG were analyzed by Quantseq REV 75 bp read paired-end RNA-seq of RiboTag HA-RPL22 immunoprecipitation (IP) and input RNA extracts. B2-SINEs were mapped in injured and naive samples. Boxplot midlines show median fold change, and whiskers indicate 10th and 90th percentile ranges of injury/naive of IP and naive DRG from three independent sciatic nerve crush experiments. n = 10,644 B2 elements. (I) Enhanced upregulation of B2-SINE in RiboTag-IP-bound RNA compared with total RNA (input). RiboTag RPL22 IP vs. input of significantly regulated B2-SINEs injury/naive expression (log₂FC injury/naive in IP or input with p < 0.05). Diagonal line depicts the trend of equal IP/input fold change ratio. n = 379 elements. See also Figure S4 and Table S3.

Figure S4

B2-SINE vs. U1 RNA-interacting proteins from sciatic nerve axoplasm, related to Figure 4 (A) Re-analysis of RNA bait pull-downs from a previous dataset, comparing protein spectral counts. Top 20 enriched hits (mean RNA/biotin ratio) with p < 0.05, two-tailed t test from three independent pull-downs. Nucleolin (Ncl) in magenta and ribosomal proteins in green. (B) Proteins co-precipitated from sciatic nerve axoplasm with biotinylated B2_5′ (nt1–75) vs. U1 (SL3 + 4) baits. Protein hits identified by TMT mass spectrometry before filtering based on enrichment in either B2 or U1 against biotin-only control at p < 0.05 from four independent pull-down experiments. The plot shows mean ratio of B2/U1 protein abundance fold change (log₂ scale). Hits with negative x-axis values are de-enriched for B2, whereas those with positive values are enriched in B2. (C) Mean B2/U1 enrichment for the top 20 B2_5′ enriched and de-enriched proteins, ordered by ratio. (D) Mean B2/U1 enrichment of ribosome-associated protein interactors, filtered by p < 0.05 in either B2 or U1 sample vs. biotin control. (E) Western blots for RBP interactors with B2_5′, U1 (SL3 + 4) and Actb baits from sciatic nerve axoplasm. n = 6 (B2, U1, and biotin-only), n = 2 (Actb). (F) Quantified normalized mean integrated density of nucleolin and PurB bands from the experiment of (E), mean ± SEM, n = 6. ^∗∗p < 0.01, ^∗∗∗∗p < 0.0001, two-tailed t test. (G) Nucleolin-B2-SINE interaction requires the nucleolin GAR domain. B2-SINE RNA bait pull-downs from HEK cell lysates transfected with the indicated HA-nucleolin constructs, representative HA blots from four independent experiments.

Figure 5

B2-SINE affects protein synthesis in DRG neurons Neurons were cultured from mouse L4/L5 DRG 2 weeks after intrathecal injection of either PHP.S-AAV-B2 or PHP.S-AAV-control vectors. (A) Representative images of YFP-positive (transduced) neurons from 48 h cultures, fixed and processed for proximity ligation assay (PLA) between nucleolin (Ncl) and RPL11, with indicated markers. Scale bar, 20 μm. (B) Quantification of PLA signal density in the nucleus and somatic cytosol of AAV-control (n = 49) vs. AAV-B2-transduced (n = 57) neurons. Mean ± SEM, ^∗p < 0.05, Mann-Whitney test. (C) Cultures pulsed with puromycin for 10 min and processed as indicated. Puromycin and merged channels masked to show signal only in YFP-positive (AAV-transduced) regions. Scale bar, 50 μm. (D) High-contrast zoomed-in images of puromycin signal in axon tips. Yellow lines show axon tip boundaries based on YFP channel (AAV transduction reporter). Scale bar, 5 μm. (E) Quantification of puromycin intensities, mean ± SEM, n = 26 (soma, AAV-control), 57 (soma, AAV-B2), 101 (axon tips, AAV-control), 226 (axon tips, AAV-B2) from 3 (AAV-control) or 4 (AAV-B2) independent cultures, after ROUT outlier analysis with Q = 0.1%. ^∗p < 0.05, Mann-Whitney test. See also Figure S5.

Figure S5

Supporting data for puromycin and proximity ligation assay, related to Figure 5 (A) Untransduced DRG neurons cultured for 48 h before imaging as indicated. Scale bar, 10 μm. (B) Quantification of proximity ligation assay (PLA) signal density in somata, as indicated. Mean ± SEM, n = 22 (PLA) or 21 (negative control) somata, ^∗∗∗∗p < 0.0001, Mann-Whitney test. (C) Representative NFH-masked images of untransduced DRG neurons pulsed with either puromycin alone or pre-incubated with anisomycin followed by addition of puromycin. Scale bar, 50 μm. (D) Comparison of puromycin intensity in somata and axon tips between puromycin-only and anisomycin + puromycin-treated DRG neurons. Mean ± S.E.E., Soma n = 29 or 40, Axon tips n = 115 or 162 for puromycin or anisomycin+puromycin, respectively, ^∗∗p < 0.01, ^∗∗∗∗p < 0.0001, Mann-Whitney test.

Figure 6

Disrupting GI-SINE inhibits nucleolin-ribosome complexes and axon growth (A) GI-SINE consensus sequence obtained by motif analysis. Colored highlights show motifs 1 and 2 by STREME motif search that have the highest GI-SINE specificity. Black boxed sites mark conserved box A and box B domains of the internal RNA Pol III promoter. Red dashed boxes and black dashed underline show GI-SINE ASO mix and GI-SINE RNA-FISH probe target sites, respectively. (B) GI-SINE RNA-FISH in LNA-ASO-transfected DRG neurons. Neurons were transfected with ASO mix (5 ASOs) targeting GI-SINE or with ASO control and probed with GI-SINE RNAscope probe or negative control probe (bacterial DapB mRNA) and counterstained with NFH (neurofilament heavy chain). Scale bar, 10 μm. (C) Mean RNA-FISH probe intensity in DRG neuronal somata. Mean ± SEM of n = 28, 31, and 38 somata for control probe, ASO control and ASO GI-SINE, respectively, from 3 independent cultures. ^∗∗∗∗p < 0.0001, one-way ANOVA, Tukey’s multiple comparison test. (D) Representative images of axonal RNA-FISH with NFH counterstain from the experiment described in (B). (E) RNA-FISH mean signal in axon segments. Mean ± SEM of n = 58, 135, and 150 axon segments for control probe, ASO control and ASO GI-SINE, respectively, from 2 independent cultures. Scale bar, 5 μm, ^∗∗∗∗p < 0.0001, one-way ANOVA, Tukey’s multiple comparison test. (F) Heatmap of B2-SINEs expression ratio in GI-SINE-ASO-transfected DRG neurons. Data are based on differential expression analysis from RNA-seq (150 bp reads, paired-end sequencing) of GI-SINE-ASO and control-ASO-transfected DRG cultures extracted 2 days after transfection. Fold-change expression ratios of significantly regulated (fold change p < 0.05) GI-SINE-matched elements and all other B2-SINE are shown (n = 21 GI-SINE and 323 other B2-SINE elements). (G) Representative images of LNA-ASO-transfected and untransfected (Unt.) neurons labeled with NFH and imaged for axon outgrowth analysis, scale bar, 50 μm. (H) Ratio of axon-growing vs. non-growing NFH-positive neurons. Neurons with longest axons < 2× their soma diameters were classified as non-growing. Lines depict mean ± SEM of n = 4 independent culture experiments. ^∗p < 0.05, one-way ANOVA and Tukey’s multiple comparisons test. (I) Total axonal outgrowth per NFH-positive neuron, mean ± SEM of n = 3,730, 3,632, 3,044, and 3,432 neurons in untransfected (Unt.), ASO control, GI-SINE ASO, and ASO B2_146–166 groups, respectively, pooled from 4 separate culture experiments. ^∗∗∗∗p < 0.0001, one-way ANOVA, Kruskal-Wallis with Dunn’s multiple comparisons test. (J) Cultured DRG neurons transfected with either GI-SINE ASO mix, control ASO or untransfected. Neurons were fixed 48 h after transfection and nucleolin-RPL11 PLA was carried out, followed by NFH and DAPI staining. Representative z-projected confocal images. Scale bar, 10 μm. (K) Density of PLA spots was quantified in the cytosol. Data shown are mean ± SEM of n = 21, 29, 31, and 37 neurons in untransfected (Unt.), ASO control, GI-SINE ASO, and ASO B2_146–166 groups, respectively, pooled from 3 independent culture experiments. ^∗∗∗p < 0.001, one-way ANOVA, Kruskal-Wallis with Dunn’s multiple comparisons test. See also Figures S6 and S7.

Figure S6

ASOs targeting GI-SINE expression, related to Figure 6 (A) Top two GI-SINE-specific motifs from STREME motif search results, mapping on 391 GI-SINEs vs. 6,665 non-regulated B2-SINEs. (B) STREME motif search results for top 8 GI-SINE sequence motifs used to assemble the consensus motif shown in Figure 6A. (C) GI-SINE RNA-FISH on DRG neurons transfected with LNA-ASO mix targeted to a GI-SINE sequence stretch distinct from that recognized by the FISH probe. Neurons were transfected with ASO control or with GI-SINE ASOs targeting GI-SINE consensus nucleotides 10–30. FISH was conducted with GI-SINE RNAscope probe matching nucleotides 39–78 (Figure 6) or a negative control bacterial DapB mRNA probe. Scale bar, 10 μm. (D) RNA-FISH probe intensity in DRG neuronal somata, mean ± SEM, n = 71, 31, and 39 somata for control, ASO control, and ASO GI-SINE, respectively. ^∗∗∗∗p < 0.0001, one-way ANOVA, Tukey’s multiple comparison test. (E) As in (C) for axon segments. Scale bar, 5 μm. (F) RNA-FISH probe intensity in axon segments, mean ± SEM of n = 71, 107, and 100 axon segments for control, ASO control and ASO GI-SINE, respectively. ^∗∗∗∗p < 0.0001, one-way ANOVA, Tukey’s multiple comparison test. (G) Droplet digital PCR (ddPCR) analysis of GI-SINE RNA expression in cultured DRG neurons transfected with GI-SINE ASO (nt10–30) compared with ASO control. GI-SINE RNA expression is normalized to Hmgb1. n = 3, mean ± SEM, ^∗∗p < 0.01, two-tailed t test.

Figure S7

Characterization of B2-SINE ASOs, related to Figure 6 (A) L4/L5 DRG neurons cultured 72 h after sciatic nerve crush. ASO transfections and outgrowth analyses as for Figures 6G–6K. Scale bar, 50 μm. (B) Longest axon ratios > 2× soma diameter from all NFH-positive somata, n = 5. (C) Total outgrowth, mean ± SEM, n = 1,577 (ASO control), 1,000 (ASO GI-SINE). ^∗∗∗∗p < 0.0001, one-way ANOVA, Kruskal-Wallis with Dunn’s multiple comparison test. (D) Ncl-RPL11 PLA in cultured DRG after conditioning sciatic nerve injury and ASO transfections as in Figure 6J. Scale bar, 10 μm. (E) PLA spot density in cytosol, mean ± SEM, n = 60 (ASO control), 61 (ASO GI-SINE), ^∗∗∗∗p < 0.0001, one-way ANOVA, Kruskal-Wallis with Dunn’s multiple comparisons test. (F) B2-SINE secondary structure model,^, blue and red dashed lines delineate the two biotin-conjugated constructs used for pull-downs. Overlaying lines mark B2-antisense oligonucleotide (ASO) position. (G) Nucleolin western blot after biotinylated-B2_5′ (nt1–75) and -B2_3′ (nt75–166) RNA pull-down from sciatic nerve axoplasm. Biotin-B2 RNA segments pre-annealed with DNA ASO targeting different B2-RNA sequence sites to test effects on B2-nucleolin interaction. (H) Nucleolin integrated density in pull-down samples of B2_3′ pre-annealed with ASO-146-166 normalized to B2_3′ alone. Mean ± SEM, n = 3.

Figure 7

GI-SINEs are intrinsic to a physiological neuronal growth circuit Upper: Axon-soma communication in neuronal growth is mediated by kinesin (K)-dependent anterograde transport of mRNAs on nucleolin (Nucl), including mTor and importin β1 (β), and retrograde transport of their encoded proteins in a complex with dynein (D) and an importin α (α) after local protein synthesis. The GI-SINEs are induced by a retrograde injury signal (S) through AP-1 transcription and then regulate nucleolin-ribosome interactions and local translation in the cytoplasm. Lower: GI-SINEs are embedded in a physiological circuit that responds to a canonical transcription factor complex to impact protein synthesis. Such circuits may control other functions beyond neuronal growth, and functionally analogous circuits may exist for related repeat elements in other species.