Pappacena F et al. (2026), Local translation of endosome-associated lc3b m...

XB-ART-61946

EMBO Rep 2026 Jul 17; doi: 10.1038/s44319-026-00867-5.

Show Gene links Show Anatomy links

Local translation of endosome-associated lc3b mRNA in axons contributes to endosomal clearance.

Pappacena F, Caterino C, De Pretis S, Sironi C, Esposito A, Cesare G, Bruni M, Cioni JM.

???displayArticle.abstract???
RNA localization to organelles is emerging as a key mechanism for regulating protein expression at the subcellular level in neurons. Although certain transcripts associate with endosomes, the functional significance remains poorly understood. Using APEX-seq, we identify a broad set of mRNAs localized to endosomes. We focus on the autophagy-related lc3b mRNA and confirm its endosomal association in cultured cells and Xenopus neuronal axons. In axons, lc3b mRNA is translated at endosomes, where the resulting LC3B protein also colocalizes, suggesting a tight spatial coupling between transcript localization and protein function. Impairment of LC3B membrane insertion via expression of a mutant ATG7 leads to the accumulation of enlarged axonal endosomes. Moreover, RAB5 overactivation promotes the formation of dysfunctional endosomes in axons that are targeted and cleared by LC3B-mediated autophagy. Finally, chloroquine-induced damage to axonal endosomes triggers their targeting by LC3B in a translation-dependent manner. Collectively, our findings expand the catalog of endosome-associated transcripts and reveal a functional link between autophagy and endosomal turnover in axons.

???displayArticle.pubmedLink??? 42469500
???displayArticle.link??? EMBO Rep
???displayArticle.grants??? [+]

Species referenced: Xenopus laevis
Genes referenced: abl1 actb atg12 atg13 atg4c chmp4b cltc eea1 fnbp1l foxk2 fundc1 gabarapl2 gria1 lamp2 map1lc3a map1lc3b map2 npc1 rab1a rab7a tbc1d5 tollip tp53inp1 tp53inp2 usp10 usp13 vcp wipi1
???displayArticle.antibodies??? Eea1 Ab5 GFP Ab25 MAP1LC3B Ab2 Rab5 Ab5 Rab7 Ab2 Raba5 AB6

???attribute.lit??? ???displayArticles.show???

	Synopsis lc3b mRNA is associated with and translated on endosomes in axons. Moreover, LC3B promotes the clearance of damaged axonal endosomes, establishing a link between RNA localization and local organelle quality control. APEX-seq identifies a broad repertoire of endosome-associated mRNAs. Endosomes serve as translation sites for lc3b mRNA in axons. Disruption of LC3B function impairs endosomal turnover in axons. Endosome damage leads to recruitment of lc3b mRNA.
	Figure 1: Identification of endosome-associated transcriptome by APEXseq. (A) Representation of APEX2-mediated proximity biotinylation and subsequent isolation of biotinylated RNAs. (B) IF performed on SH-SY5Y cells showing APEX2 construct localization (in green) and biotinylation activity (in magenta) in presence of H₂O₂. Arrows indicate biotinylation activity matching APEX2 localization. (C) Streptavidin-biotin RNA dot blot following APEX2 activation (B.P., Biotin Phenol, 3 exp). (D) Scatter plot showing mRNAs enriched following pulldown in APEX2-RAB5A line treated with or without H₂O₂ (4 exp). (E) Scatter plot showing mRNAs enriched following pulldown in 2xFYVE-APEX2 line treated with or without H₂O₂ (4 exp). (F) Venn diagram showing the number and overlap between the enriched mRNAs. (G, H) Gene Ontology (GO) enrichment analysis for Cellular Components and Biological Process of the identified transcripts. The number of transcripts analyzed is indicated for each condition. (D, E) Wald test. (G, H) Hypergeometric test. Scale bars: (B) 10 μm. Source data are available online for this figure.
	Figure EV1: APEXseq identifies distinct transcriptomes in H₂O₂-treated and untreated samples. (A) Heatmap showing similarity and variance between RNA samples obtained from each APEX2 cell line treated with or without H₂O₂ in 4 independent biological replicates. Color scale indicates distances between samples. (B) Principal Component Analysis (PCA) showing the clustering of untreated samples from each APEX2 line together, and the separation between samples from different APEX2 cell lines treated with H₂O₂. (C) Gene Ontology Cellular Component analysis of mRNAs identified in both 2xFYVE-APEX2 and APEX2-RAB5A conditions, showing the top 20 enriched categories. (D) Gene Ontology Biological Process analysis of mRNAs identified in both 2xFYVE-APEX2 and APEX2-RAB5A conditions, showing the top 20 enriched categories. (C, D) Hypergeometric test.
	Figure 2: lc3b mRNA localizes to endosomes in neuronal-like cells and neuronal axons. (A) smFISH images of eea1, lc3b, and sqstm1 mRNAs (in magenta) stained against RAB5 (in green) in SH-SY5Y cells. Arrows indicate mRNA signal colocalizing with RAB5 signal. (B) Colocalization analysis. Dots represent cells (eea1:n = 59 cells, P = 0.0087, 3 exp; lc3b: n = 75 cells, P = 0.0029, 3 exp; sqstm1: n = 68 cells, P = 0.0067, 3 exp; atg12: n = 48 cells, P = 0.8976, 3 exp). (C) smFISH images of lc3b mRNA (in magenta) stained against EEA1 or RAB7 (in green). Arrows indicate colocalizing mRNA signal. (D) Colocalization analysis. Dots represent cells (EEA1: n = 70 cells, P = 0.0211, 3 exp; RAB7: n = 82 cells, P = 0.0121, 3 exp). (E) Scheme of rapalog-based heterodimerization and predicted redistribution of RAB5-endosomes upon treatment. (F) smFISH images of lc3b and atg12 mRNAs (in magenta) in SH-SY5Y cells expressing GFP-2xFKBP-RAB5 and HA-BICD-FRB treated with control (CT) or 100 nM rapalog (RAPA) for 2 h. Arrows indicate colocalizing mRNA signal. (G) Dots represent cells. Quantification of lc3b (CT: n = 45 cells, RAPA: n = 44 cells, P = 0.0010, 3 exp) and atg12 (n = 57 cells, P = 0.2805, 3 exp) mRNAs subcellular distribution. (H) smFISH images of lc3b mRNAs (in white) in axons expressing RFP-RAB5. (I) Colocalization analysis. Dots represent mean per experiment. RAB5 colocalizing with lc3b: n = 62 fields of view, n ≥8 explants, 4 exp. (J) RT-qPCR analysis of gfp expression (P = 0.7000, 3 exp). (K) smFISH images showing gfp-lc3b or gfp-lc3b-Δ3’UTR mRNA signal (in white) in axons expressing RFP-RAB5 (in red). Arrows indicate mRNA signal colocalizing with RAB5 signal. (L) Quantification of gfp mRNA puncta in distal axons (gfp-lc3b: n = 145 axons, gfp-lc3b-Δ3’UTR: n = 146 axons, P < 0.0001, n ≥6 explants, 3 exp). (M) Colocalization between mRNA puncta and RFP-RAB5 in axons (gfp-lc3b: n = 72 axons, gfp-lc3b-Δ3’UTR: n = 47 axons, P = 0.3076, n ≥8 explants, 4 exp). (B, D, G, J, L, M) Mann–Whitney test. Data is represented as mean ± SEM. Scale bars: (A, C, H, K) 3 µm, (K) 5 μm and (F) 10 µm. Source data are available online for this figure.
	Figure EV2: lc3b mRNA association with RAB5-positive endosomes is puromycin insensitive. (A) smFISH images of lc3b mRNA (in magenta) in SH-SY5Y cells treated with vehicle or 100 μM puromycin and stained against RAB5 (in green). Arrows indicate lc3b mRNA colocalizing with RAB5 signal. Scale bar: 3 μm. (B) Colocalization analysis. Dots represent cells (vehicle n = 30 cells, puromycin n = 33 cells, P = 0.2804, 3 exp). Mann–Whitney test. Mean ± SEM.
	Figure EV3: Autophagy-related mRNAs are present in Xenopus RGC axons. (A) Schematic representation of compartmentalized chambers (Boyden chambers) used to isolate somatodendritic and axonal compartments of Xenopus RGCs. (B) Heatmap showing the ranking of abundance (average FPKM) of autophagy-related mRNAs in Xenopus RGC axons (Shigeoka et al, 2019). Red indicates mRNA identified by APEXseq in Fig. 1. (C) RT-PCR analysis of cDNA synthesized from somatodendritic and axonal RNAs (n = 3 exp.). actb, glur1 and map2 mRNAs (on the left) show purity of axonal RNAs.
	Figure EV4: Alteration of RAB5 activity does not impair lc3b mRNA expression or its distribution in SH-SY5Y cells. (A) IF images of EEA1 signal (in magenta) in SH-SY5Y cells expressing the indicated constructs (in green). (B) Quantification of EEA1 signal density. Dots represent cells (GFP: n = 64 cells, RAB5WT: n = 74 cells; RAB5DN: 76 cells, n = 3 exp). (C) RT-qPCR analysis in SH-SY5Y cells expressing the indicated constructs (n = 4 exp). (D) lc3b mRNA distribution in SH-SY5Y cells expressing the indicated constructs. Quantification of the relative frequency distribution of mRNA distances from the nucleus relative to the maximum cell extension. Dots represent cells (GFP: n = 36 cells, RAB5WT: n = 43 cells; RAB5DN: 36 cells, n = 3 exp). Mean ± SEM. (B–D) Kruskal–Wallis test. Scale bar: (A) 10 μm.
	Figure 3: LC3B is synthesized on RAB5-positive endosomes in axons. (A) Image and kymograph (2 min) of time-lapse sequence showing Cy5-lc3b (in white) and RFP-RAB5 (in red) expression in axons. Arrows indicate mRNA puncta associated with RAB5. (B, C) Quantitative analysis of the motion-types or directionality of Cy5-lc3b mRNAs in axons expressing the indicated constructs (RFP: n = 79 axons; RAB5WT: n = 53 axons; RAB5DN: n = 71 axons). Dots represent mean/experiments (4 exp.). (D) Images showing LC3B Puro-PLA signal in axons and related quantification in the indicated conditions (CT (DMSO), n = 39 fields of view; CHX, n = 26 fields of view; P < 0.0001, n ≥6 explants, 3 exp). Arrows indicate Puro-PLA dots (red) in axons (white). (E) Images showing SunTag-LC3B foci (indicated by arrows) in axons. Quantification of SunTag-LC3B foci in axons in Control (CT, n = 101 axons) or treated with Puromycin (PURO, n = 88 axons, P < 0.0001) or Harringtonine (HARR., n = 48 axons, P = 0.0092), n ≥6 explants, 3 exp. (F) Image and time-lapse sequence showing SunTag-LC3B foci (indicated by arrow) in axon expressing RFP-RAB5. Time is indicated in seconds (s). (G) Colocalization between RFP-RAB5 and SunTag-LC3B foci in axons. n = 17 axons, n ≥4 explants, 2 exp. Mean ± SEM. (D) Mann–Whitney test. (B, C, E) Kruskal–Wallis test. Scale bars: (A) 2.5 μm, (D) 10 μm, (E) 4 μm, and (F) 5 μm. Source data are available online for this figure.
	Figure 4: LC3B regulates RAB5-positive endosomes in axons. (A) IF image of RAB5 (in magenta) and LC3B (in green) stainings in RGC axons. Arrows indicate colocalization. (B) Mander’s colocalization coefficient. Dots represent axons. (n = 37 axons, n≥6 explants, 3 exp.). (C) Images of TEM coupled with immunogold labeling of GFP-LC3B in axons. Arrowheads indicate positive signals along (a) microtubules, (b) in and on double-membrane (blue dashed line), and (c) in and on single-membrane vesicles (yellow dashed line). (D) Scheme illustrating the dominant negative effect of ATG7(C572S). (E) Images and quantification of GFP-LC3B puncta. Dots represent axons (n = 48 in CT, n = 31 in ATG7(C572S), P < 0.0001, n ≥4 explants, 2 exp). (F) Images of RAB5 IF signal in axons co-expressing GFP alone or GFP and ATG7(C572S). (G) Quantification of LC3B signal (mean pixel intensity). Dots represent axons. (n = 90 in CT, n = 78 in ATG7(C572S), n ≥6 explants, 3 exp), P < 0.0001, n ≥6 explants, 3 exp. (H) Quantification of RAB5 signal (mean pixel intensity). Dots represent axons (n = 90 in CT, n = 78 in ATG7(C572S), P < 0.0001, n ≥6 explants, 3 exp). (I) Area of RAB5 positive signal in axons. Dots represent vesicles. (n = 244 in CT, n = 273 in ATG7(C572S), P < 0.0001, n ≥6 explants, 3 exp). Mean ± SEM. (E, G, H, I) Mann–Whitney test. Scale bars: (A) 2 μm, (Ca) 2 μm, (Cb and Cc) 200 nm, (E) 3 μm, and (F) 5 μm. Source data are available online for this figure.
	Figure EV5: Expression of the ATG7(C572S) mutant does not affect LC3B synthesis in axons. (A) Images showing LC3B Puro-PLA signal in axons. Scale bar: 10 μm. (B) Quantification in the indicated conditions (CT, n = 13 fields of view; ATG7(C572S), n = 15 fields of view; n ≥6 explants, 3 exp). Mean ± SEM. (B) Unpaired t test.
	Figure EV6: RAB5 overactivation drives an increase in GFP-LC3B puncta in axons. (A) Images showing GFP-LC3B signal in axons expressing increased concentration of rab5a mRNA. (B) Quantification of GFP-LC3B puncta. Dots represent axons (CT: n = 30; RAB5 50 ng: n = 29; RAB5 200 ng: n = 26, *P = 0.0107, n ≥4 explants, 2 exp). Kruskal–Wallis test. Mean ± SEM. Scale bar: (A) 10 μm.
	Figure 5: Dysfunctional endosomes generated by RAB5 overexpression become LC3B-positive, and LC3B function regulates their clearance in axons. (A) Image and kymograph (2 min) showing BSA-647 signal (in magenta) in axons expressing GFP-LC3B (in green). Arrows indicate colocalizing signals. (B) Colocalization analysis. Dots represent axons (CT: n = 27; RAB5 100 ng: n = 14, *P = 0.0021, P = 0.0127, n ≥6 explants, 3 exp). (C) Representative image showing RFP-RAB5 and GFP-LC3B signals in axons. (D) Expanded views and intensity plot of the GFP-LC3B signal along the indicated line. (E) Colocalization analysis. Dots represent axons (n = 25, P = 0.0091, n ≥6 explants, 3 exp). (F) Time-lapse sequence showing an axon expressing RFP-RAB5 and GFP-LC3B. Arrows indicate nascent AP in the growth cone, in which RAB5 signal accumulates. (G) Time-lapse sequence of RFP-RAB5 (in blue), GFP-LC3B (in yellow) and Cy5-gfp-lc3b mRNA (in magenta) signals. Arrows indicate colocalization between signals. (H) Colocalization analysis. Dots represent axons. (CT: n = 38, CHX: n = 37, P = 0.0485, n ≥4 explants, 2 exp). (I) Colocalization analysis. Dots represent axons. (CT: n = 22, PURO: n = 29, P = 0.0011, n ≥6 explants, 3 exp). (J) Images and quantification of GFP-RAB5 signal. Arrows indicate ring-like endosomes. Dots represent axons (CT: n = 31, ATG7(C572S): n = 23, P = 0.0038, n ≥6 explants, 3 exp). Mean ± SEM. (B, E, H, I, J) Mann–Whitney test. Scale bars = (A, J) 5 μm, (G) 1 μm, and (C, F) 2 μm. Time is indicated as min:sec. Source data are available online for this figure.
	Figure EV7: RAB5 overexpression induces enlarged endosomes positive for SQSTM1. (A) Image showing GFP-SQSTM1 signal colocalizing with ring-like RFP-RAB5 positive endosome in axons. Arrow indicates GFP-SQSTM1 signal on RFP-RAB5-endosome. (B) Colocalization between RFP-RAB5 and GFP-SQSTM1. Dots represent axons (n = 31, n ≥6 explants, P = 0.0011, 3 exp). Mann–Whitney test. Mean ± SEM. (C) Time-lapse sequence on axons expressing RFP-RAB5 and GFP-SQSTM1. Arrows indicate a RFP-RAB5-endosome becoming GFP-SQSTM1 positive. Time is indicated as min:sec. Scale bar: (A) 5 μm, and (C) 2 μm.
	Figure 6: Chloroquine induces endosomal swelling and protein synthesis-dependent LC3B recruitment in axons.(A) IF images of RAB5 (in red) and LC3B (in white) stainings in RGC axons with indicated treatments. (B) Area of RAB5 positive signal in axons. Dots represent vesicles. (n = 774 in CT, n = 782 in CHQ and n = 757 in CHQ + CHX, **P < 0.0001, P = 0.0159, n ≥6 explants, 3 exp). (C) Percentage of RAB5-positive vesicles also positive for LC3B. Dots represent axons (CT: n = 39; CHQ: n = 37; CHQ + CHX: n = 31, *P = 0.0407, n ≥6 explants, 3 exp). (D) Images and quantification of RFP-RAB5 signal. Dots represent axons (CT: n = 23, CHQ: n = 21, P = 0.0053, n ≥4 explants, 2 exp). (E) Time-lapse sequence showing axon expressing RFP-RAB5 and GFP-LC3B treated with CHQ. Arrow indicates a RAB5 vesicle that transitions to LC3B-positive. Mean ± SEM. (B, C) Kruskal–Wallis test. (D) Mann–Whitney test. Scale bars = (A, E) 2 μm, and (D) 3 μm. Time is indicated as minutes (min). Source data are available online for this figure.
	Figure 7: lc3b mRNA is recruited on damaged organelles upon organelle-specific stress. (A) smFISH images of lc3b mRNAs (in magenta) stained against EEA1 (in white) in SH-SY5Y cells treated with control (CT) or 50 μM chloroquine (CHQ) for 20 min. Arrows indicate mRNA signal colocalizing with EEA1 signal. A magnified view of the highlighted region is shown in the inset; the inset image has been rotated by 90°. (B) Colocalization analysis. Dots represent cells (CT: n = 58 cells; CHQ: n = 55 cells, P = 0.0005, 3 exp). (C) Scheme of the MS2 reporter system. (D) Example of a cell expressing lc3b-MS2 reporter mRNA labeled by MCP-Halo (in magenta) and GFP-RAB5 (in green). Time-lapse sequence indicate association between the two signals (arrows). Time is indicated as min:sec:ms. A magnified view of the highlighted region is shown in the inset. (E) Quantification of the number of RAB5-lc3b mRNA reporter associations lasting for longer than 10 s. Dots represent cells (CT: n = 22 cells; CHQ: n = 21 cells, P = 0.0022, 3 exp.). (F) Example of a cell expressing lc3b-MS2 reporter mRNA labeled by MCP-Halo (in magenta) and GFP-RAB5 (in green) treated with CHQ. Time-lapse sequence indicate association between the two signals (arrows). Time is indicated as min:sec. A magnified view of the highlighted region is shown in the inset; the inset image has been rotated by 90°. (G) smFISH images of lc3b mRNAs (in magenta) stained against RAB7 (in white) in SH-SY5Y cells treated with control (CT) or 250 μM LLOMe for 20 min. Arrows indicate mRNA signal colocalizing with RAB7 signal. A magnified view of the highlighted region is shown in the inset; the inset image has been rotated by 90°. (H) Colocalization analysis. Dots represent cells (CT: n = 116 cells; LLOMe: n = 96 cells, P = 0.0013, 4 exp). (I) smFISH images of lc3b mRNAs (in magenta) stained against TOM20 (in white) in SH-SY5Y cells treated with control or with 1 μM oligomycin A plus 1 μM antimycin (O + A) for 20 min. Arrows indicate mRNA signal colocalizing with TOM20 signal. A magnified view of the highlighted region is shown in the inset. (J) Colocalization analysis. Dots represent cells (CT: n = 162 cells; O + A: n = 104 cells, P = 0.0019, 4 exp). Mean ± SEM. (B, E, H, J) Mann–Whitney test. Scale bar = (A, F, G, I) 4 μm, and (D) 5 μm. Source data are available online for this figure.

External Resources: E-MTAB-15783
biostudies