Dynamic m6A modification regulates local translation of mRNA in axons

Abstract

N6-methyladenosine (m6A) is a reversible modification in mRNA and has been shown to regulate processing, translation and decay of mRNA. However, the roles of m6A modification in neuronal development are still not known. Here, we found that the m6A eraser FTO is enriched in axons and can be locally translated. Axon-specific inhibition of FTO by rhein, or compartmentalized siRNA knockdown of Fto in axons led to increases of m6A levels. GAP-43 mRNA is modified by m6A and is a substrate of FTO in axons. Loss-of-function of this non-nuclear pool of FTO resulted in increased m6A modification and decreased local translation of axonal GAP-43 mRNA, which eventually repressed axon elongation. Mutation of a predicted m6A site in GAP-43 mRNA eliminated its m6A modification and exempted regulation of its local translation by axonal FTO. This work showed an example of dynamic internal m6A demethylation of non-nuclear localized mRNA by the demethylase FTO. Regulation of m6A modification of axonal mRNA by axonal FTO might be a general mechanism to control their local translation in neuronal development.

Figure 1.

FTO is enriched and locally translated in axons. E13 DRG neurons were cultured and subcellular localization of FTO was examined by immunostaining with a specific antibody. (A) Anti-FTO IF showed a strong signal in axons suggesting FTO protein is enriched in axons. GAP-43 is used for counter staining of neurons/axons. (B) Quantification of FTO IF signals in nucleus (n = 18), cytoplasm (n = 18) and axons (n = 18). (C) Fto mRNA was detected in axons. RT-PCR using RNA respectively from distal axon compartment and cell body compartment was carried out. Similar as β-actin mRNA which is a positive control for axonal mRNAs, Fto mRNA was detected in both axons and soma. The failure to detect H1f0 mRNA from axons confirmed that the axonal material was pure with no neuronal soma incorporation. (D) Axonal quantitaive-PCR (qPCR) analysis showed that axonal Fto mRNA has a higher level than β-actin mRNA, and is similar as Nrxn2 mRNA which is another positive control for axonal mRNAs (n = 4). (E) NGF induces local translation of FTO in axons. DRG explants grown in microfluidic chambers were starved with NGF and then axons were severed from explants to exclude the contribution of axonal FTO by soma. NGF treatment of severed axons resulted in an increase of FTO IF signals in axons, which could be blocked by co-application of cycloheximide (CHX, 10 μM). (F) Quantification of results in (E) showing axonal FTO IF intensities (per axonal area defined by Tau1 counter staining) of different treatments (vehicle, n = 15 axons; NGF, n = 16 axons; NGF + CHX, n = 10 axons; CHX, n = 12 axons). (G) Validation of siFto. DRG neurons were cultured and siRNA was transfected by bath application. Compared with a negative control siRNA (siNC), two siRNA against Fto (siFto-1, siFto-2) led to significant knockdown of Fto mRNA by qRT-PCR (n = 3). (H and I) Compartmentalized knockdown of Fto in axons. DRG neurons were grown in microfluidic chambers and siRNA were specifically transfected to axons. Compared with siNC, siFto-1 and 2 led to significant decreases of FTO IF signals (siNC, n = 10 axons; siFto-1, n = 9 axons; siFto-2, n = 11 axons). All data are mean ± s.e.m. Data of IF quantification (B, F and I) are represented as box and whisker plots: 25th–75th percentiles (boxes), minimum and maximum (whiskers) and medians (horizontal lines). For B: nucleus versus cytoplasm, **P = 0.00207; nucleus versus axon, *P = 0.04603. For F: NGF versus vehicle, ***P = 2.77E-4; NGF versus NGF + CHX, ***P = 2.45E-5. For I: siFto-1 versus siNC, **P = 0.00280; siFto-2 versus siNC, **P = 0.00318. qRT-PCR data (n = 4 for D and G) are represented as dot plots. For D: Fto versus β-actin, **P = 0.00474; Nrxn2 versus β-actin, *P = 0.01434. For G: siFto-1 versus siNC, ***P = 4.39E-5; siFto-2 versus siNC, ***P = 1.15E-4. All by ANOVA followed by Tukey’s multiple comparison test. Each experiment was performed in triplicate (A, C, E, G and H) or quadruplicate (D). Scale bars, 10 μm (A, E and H).

Figure 2.

Bath application of rhein increased m⁶A signals in axons and inhibited axon elongation. (A and B) m⁶A signals are much lower in axons than soma. IF was carried out using an m⁶A-specific antibody and m⁶A IF signals were quantified in axons and soma, respectively (n = 9 neurons). (C and D) Bath application of rhein in cultured DRG neurons led to a moderate (0.26-fold) increase of m⁶A signals in soma compared with vehicle treatment (vehicle, n = 20 neurons; rhein, n = 22 neurons). (E and F) Bath application of rhein in cultured DRG neurons resulted in a robust increase of m⁶A signals in axons compared with vehicle treatment (vehicle, n = 12 axons; rhein, n = 12 axons). This increase (0.95-fold increase) in axons is much bigger than in soma (C and D). (G and H) Bath application of rhein in cultured DRG neurons inhibited axon elongation by reducing average axon growth rate of 22.17 ± 1.74 μm/h with vehicle treatment to 11.20 ± 0.66 μm/h (vehicle, n = 16 neurons; rhein, n = 20 neurons). Data are mean ± s.e.m. All data are represented as box and whisker plots: 25th–75th percentiles (boxes), minimum and maximum (whiskers) and medians (horizontal lines). ***P = 9.88E-8 (B); ***P = 6.80E-5 (D); ***P = 3.26E-7 (F); ***P = 2.60E-7 (H); by unpaired t-test. Each experiment was performed in triplicate. Scale bars, 10 μm (A, C, E and G).

Figure 3.

Axon-specific loss-of-function of FTO increased m⁶A signals in axons and inhibited axon elongation. DRG explants were cultured in microfluidic chambers. After axons grew to the axonal compartment, different experiments were performed. (A and B) Axonal rhein treatment led to a significant increase of m⁶A signals in axons (vehicle, n = 12 axons; rhein, n = 10 axons). Before rhein treatment, axons were severed from explants to prevent any further transport of m⁶A modified RNA from soma. (C and D) Rhein treatment of intact axons resulted in a significant reduction of axon growth rate compared with vehicle treatment (vehicle, n = 16 axons; rhein, n = 17 axons). (E and F) Compartmentalized Fto knockdown in intact axons using two different siFto in axons led to significant increase of axonal m⁶A levels compared with siNC (siNC, n = 17 axons; siFto-1, n = 15 axons; siFto-2, n = 15 axons). (G) Compartmentalized Fto knockdown in intact axons significantly inhibited axon elongation (siNC, n = 8 axons; siFto-2, n = 10 axons). Data are mean ± s.e.m. All data are represented as box and whisker plots: 25th–75th percentiles (boxes), minimum and maximum (whiskers) and medians (horizontal lines). For B, D and G: *P = 0.0300 (B); ***P = 2.29E-6 (D); ***P = 1.089E-4 (G); by unpaired t-test. For F: *P = 0.038, ***P = 7.49E-4; by one-way ANOVA followed by Tukey’s multiple comparison test. Each experiment was performed in triplicate. Scale bars, 10 μm (A, C and E).

Figure 4.

Axonal FTO regulates local translation of GAP-43 in axons. (A and B) DRG neurons were cultured in vitro and bath application of rhein led to a decrease of GAP-43 IF signals in axons (Vehicle, n = 15 axons; Rhein, n = 14 axons). (C and D) Axon-specific treatment of rhein in severed axons resulted in a significant decrease of GAP-43 IF signals in axons (Vehicle, n = 10 axons; Rhein, n = 12 axons). (E and F) Loss-of-function of FTO by rhein (E) or siFto knockdown (F) did not change GAP-43 mRNA levels by qRT-PCR (n = 3). (G and H) Co-application of rhein with MG-132 in severed axons led to a significant decrease of GAP-43 IF signals in axons compared with MG-132 treatment alone (MG-132 + Vehicle, n = 11 axons; MG-132 + Rhein, n = 11 axons). Data are mean ± s.e.m. IF quantification data (B, D and H) are represented as box and whisker plots: 25th-75th percentiles (boxes), minimum and maximum (whiskers) and medians (horizontal lines). ***P = 2.76E-8 (B); ***P = 0.0002 (D); **P = 0.0026 (H); by unpaired t-test. qRT-PCR data (E and F) are represented as dot plots. n.s., not significant (unpaired t-test for E, one-way ANOVA followed by Tukey’s multiple comparison test for F). Each experiment was performed in triplicate (C, E, F and G) or quadruplicate (A). Scale bars, 10 μm (A, C and G).

Figure 5.

Axonal FTO de-(m⁶A)methylates GAP-43 mRNA in axons. (A and B) Modification of GAP-43 mRNA by m⁶A. Total RNA from mouse DRG (A) or spinal cord (B) was immunoprecipitated with a specific m⁶A antibody or an IgG control. In either of these two tissues, GAP-43 mRNA was pulled down and detected by RT-PCR. (C) Mutation of m⁶A site eliminated m⁶A modification in GAP-43 mRNA. Wild-type (‘WT’) GAP-43 or a mutated form which has the predicted m⁶A site mutated (‘MT^m6A’) were overexpressed in DRG neurons. RNA was purified after rhein treatment for 4 h. After anti-m⁶A IP, much less GAP-43 mRNA was detected from ‘MT^m6A’ neurons, compared with ‘WT’ neurons. (D) Rhein treatment led to increase of m⁶A modification in GAP-43 mRNA. More GAP-43 mRNA was pulled down by anti-m⁶A IP in rhein-treated DRG neurons compared with vehicle control. (E and F) Mutation of m⁶A site in GAP-43 mRNA could rescue inhibition of GAP-43 local translation and axon elongation by rhein. ‘WT’ or ‘MT^m6A’ GAP-43 was overexpressed in DRG neurons and rhein was specifically applied to axons in microfluidic chambers. Axon-specific treatment of rhein in ‘WT’ GAP-43-overexpressing DRG neurons led to significant decreases of GAP-43 protein levels (E) and axon growth rate (F) compared with vehicle control. These decreases were not observed with axonal rhein treatment of ‘MT^m6A’ GAP-43-overexpressing DRG neurons (E and F). For E: WT, n = 15 axons; WT + Rhein, n = 10 axons; MT^m6A + Rhein, n = 8 axons. For F: WT, n = 7 axons; WT + Rhein, n = 5 axons; MT^m6A + Rhein, n = 6 axons. Data are mean ± s.e.m. Data are represented as box and whisker plots: 25th–75th percentiles (boxes), minimum and maximum (whiskers), and medians (horizontal lines). For E: WT versus WT + Rhein, *P = 0.01006; WT + Rhein versus MT^m6A + Rhein, *P = 0.02977. For F: WT versus WT + Rhein, *P = 0.04229; WT + Rhein versus MT^m6A + Rhein, *P = 0.01899; by one-way ANOVA followed by Tukey’s multiple comparison test. Each experiment was performed in triplicate.

Figure 6.

Working model. (A) Under normal conditions, Fto mRNA is transported to axons and can be locally translated. These axonally derived FTO can de-(m⁶A)methylate GAP-43 mRNA. De-(m⁶A)methylated GAP-43 mRNA can be locally translated to promote axon elongation. (B) Compartmentalized siFto knockdown in axons will lead to failure of local translation of Fto in axons. Axon-specific treatment of rhein will inhibit FTO functions. Both of these two loss-of-function assays of axonal FTO will result in inhibiting de-(m⁶A)methylation of GAP-43 mRNA. Axonal GAP-43 mRNA with maintained m⁶A modifications cannot be locally translated, thus inhibiting axon elongation.