The autophagic protein LC3 translocates to the nucleus and localizes in the nucleolus associated to NUFIP1 in response to cyclic mechanical stress

Abstract

The trabecular meshwork (TM) is a key regulatory tissue of intraocular pressure (IOP) in the anterior chamber of eye. Dysfunction of the TM causes resistance to outflow of aqueous humor, which in turn leads to elevated IOP, a main risk factor of glaucomatous neurodegeneration. Due to variations in IOP, TM cells are continuously exposed to mechanical deformations. We previously reported activation of macroautophagy/autophagy, as one of the physiological responses elicited in TM cells following mechanical strain application. By using biochemical fractionation analysis and imaging techniques, we demonstrate here for the first time the nuclear accumulation of the autophagic marker MAP1LC3/LC3 (microtubule associated protein1 light chain 3)-II, endogenous and exogenously added (AdGFP-LC3, AdtfLC3), in response to cyclic mechanical stress (CMS). Wheat germ agglutinin (WGA) and leptomycin B treatment suggest LC3 to enter the nucleus by passive diffusion, but to exit in an XPO1/CRM1 (exportin 1)-dependent manner in human TM (hTM) cells. While blockage of nuclear export leads to accumulation of LC3 with promyelocytic leukemia (PML) bodies, nuclear LC3 localizes in the nucleolus in cells under CMS. Moreover, nuclear LC3 co-immunoprecipitated with NUFIP1, a ribosome receptor for starvation-induced ribophagy. More interestingly, we further demonstrate that NUFIP1 translocates from the nucleus to LAMP2 (lysosomal associated membrane protein 2)-positive organelles in the stretched cells without triggering ribophagy, suggesting a more general role of NUFIP1 as a selective autophagy receptor for another yet-to-be-identified target in CMS and a surveillance role of nuclear LC3 against stretch-induced damage.

Abbreviation: AdGFP: adenovirus encoding GFP; ATG: autophagy-related; BSA: bovine serum albumin; CMS: cyclic mechanical stretch; Co-IP: coimmunoprecipitation; DAPI: 4',6-diamidino-2-phenylindole; DFCs: dense fibrillar components; EM: electron microscopy; FCs: fibrillar centers; GCs: granular components; GFP: green fluorescent protein; hTM: human trabecular meshwork; HBSS: Hanks balanced salt solution; IOP: intraocular pressure; LAMP1/2: lysosomal associated membrane protein 1/2; LepB: leptomycin B; MTOR: mechanistic target of rapamacyin kinase; NES: nuclear export signals; MAP1LC3/LC3: microtubule associated protein 1 light chain 3; NLS: nuclear localization signal; NPCs: nuclear pore complexes; NUFIP1: nuclear FMR1 interacting protein 1; NS: non-stretched; PBS: phosphate-buffered saline; PE: phosphatidylethanolamine; pfu: plaque-forming units; PML: promyelocytic leukemia; RFP: red fluorescent protein; RPS15A: ribosomal protein S15a; RPL26: ribosomal protein L26; rRNA: ribosomal RNA; SIRT1: sirtuin 1; SQSTM1/p62: sequestosome 1; tfLC3: mRFP-GFP tandem fluorescent-tagged LC3; TM: trabecular meshwork; WB: western blot; WDR36: WD repeat domain 36; WGA: wheat germ agglutinin; XPO1/CRM1: exportin 1.

Keywords: Glaucoma; LC3; NUFIP1; PML bodies; autophagy; mechanical stress; nuclear LC3; nucleolus; stretching; trabecular meshwork.

Figure 1.

Increased nuclear LC3-II levels in hTM cells under CMS. (A) Immunocytochemical analysis of endogenous LC3 (upper panels, red color) or tfLC3 fluorescence (lower panels, RFP and GFP signals) in non-stretched (NS) or stretched (CMS, 8% elongation, 24 h) hTM cells. LC3 puncta in the nucleus are indicated by arrows in each inset. Scale bars: 20 μm. Images were acquired with confocal microscopy from 3 independent experiments with different primary cultured hTM cells, processed and LC3 puncta per nucleus were quantified by using Fiji software. Data are shown as the mean ± S.D. (n = 30 and 29 nuclei number for endogenous LC3, and 14 and 13 for tfLC3 in NS vs CMS, respectively). ***, p < 0.001, (two-tailed unpaired Student’s t-test). (B) Protein expression levels of LC3 in cytosolic and nuclear fractions of hTM cells subjected to CMS for 1 or 24 h, evaluated by WB. Band densities were quantified by Image Lab™ touch software and normalized with TUBB and H2B for cytosolic and nuclear fractions, respectively (C). NS: non-stretch control, CMS: cyclic mechanical stretch. Data are shown as the mean ± S.D. (n = 3). *, p < 0.05, (two-tailed unpaired Student’s t-test).

Figure 2.

Nucleocytoplasmic shuttling of LC3 in hTM cells. (A) Representative immunocytochemical analyses of endogenous LC3 in hTM cells subjected to CMS for 24 h in the presence of Lep B (20 nM) or WGA (5 µg/mL). Nuclear LC3 dots resulting from LepB treatment are indicated by white arrows; yellow arrows indicate LC3 puncta formed by CMS. Asterisk indicate perinuclear autophagosome. (B) Time-lapse live cell imaging of AdGFP- or AdGFP-LC3-transduced hTM cells following LepB (20 nM) treatment. Upper panels represent images taken after 24-h treatment. Lower panels showed GFP-LC3 fluorescence chase through the duration of the treatment, using CELENA® S Digital Imaging System. Nucleus is outlined by ellipse with dot line. Each nuclear LC3 dot is indicated by red arrows. (C) Representative image of AdGFP-LC3-transduced cells treated simultaneously with LepB (20 nM) and WGA (5 µg/ml) for 24 h. (D) Time-lapse live cell imaging of GFP-LC3 nuclear dots after LepB removal. GFP-LC3 fluorescent signals were chased for 24 h. Nuclear LC3 dots are indicated by red arrows. (E) Protein levels of LC3-I, LC3-II and SQSTM1 in purified cytosolic and nuclear fractions from hTM cells subjected to CMS (8% elongation, 24 h) in the presence of LepB (20 nM) or WGA (5 µg/mL). Band densities were quantified by Image Lab™ touch software and normalized with H2B for nuclear fractions (F). NS: non-stretch control, CMS: cyclic mechanical stretch. Data are shown as the mean ± S.D. (n = 3). *, p < 0.01; ***, p < 0001, #, p < 0.05 (two-tailed unpaired Student’s t-test). Scale bars: 20 μm. * compares Lep and WGA treatment versus non-treated; # compares CMS versus NS.

Figure 3.

Ultrastructural appearance of hTM cells under CMS. Left panels: Representative lower magnification of NS or cells under CMS. Right panel: higher magnification of inset in stretched cells. Note that autophagosomes were not found within nuclei of NS or CMS cells. Nu, nucleus; V, vacuoles; AV, autophagic vacuole; ER, endoplasmic reticulum; M, mitochondria.

Figure 4.

LC3 interacts with PML nuclear bodies in LepB-treated cells, but not with CMS-induced ones. hTM cells were transduced with AdGFP-LC3 (5 pfu/cell) and either exposed to CMS (8% elongation) or LepB (20 nM) treatment for 24 h. Cells were fixed and immunostained with specific antibodies against nuclear speckels (A), Cajal bodies (B) and PML bodies (C). Fluorescent intensity of GFP-LC3 signal (green) and PML (red) in CMS (D) and LepB-treated cells (E) were quantified with Fiji software. DAPI was used to stain nuclei. Scale bars: 20 μm.

Figure 5.

Nuclear LC3 associates with the nucleolus with CMS. (A) Representative immunostaining of the nucleolar marker FBL in AdGFP-LC3-transduced cells exposed for 24 h to CMS (8% elongation) or LepB (20 nM) treatment. DAPI was used to stain nuclei. Scale bars: 20 μm. (B) Interactive 3D surface plot analysis visualizing the interaction between GFP-LC3 and the nucleolus. (C) Co-IP analysis of GFP-LC3 with FBL and SQSTM1 in purified nuclear fractions from hTM cells transduced with AdGFP or AdGFP-LC3. IP was performed using GFP-Trap; co-immunoprecipitated proteins were detected by WB, 5 µg of protein were loaded for input control. (D) Representative EM images depicting nucleolar structure and components in NS and cells under CMS. FC: Fibrillar Center; DFC: dense fibrillar component; GC: granular component. (E) Quantification of nucleolar number in NS and cells under CMS. Nucleoli were detected by immunostaining with FBL antibody (red). Y axis represents percentage of cells containing nucleoi number>4 per nucleus in total cells counted. Data are shown as the mean ± S.D. (n = 375 and 309 in CNT and CMS, respectively), two-tailed unpaired Student’s t-test.

Figure 6.

SIRT1 inhibition does not prevent CMS-induced autophagy in hTM cells. (A) Live cell imaging of AdtfLC3-transduced cells treated with EX527 (25 µM) for 24 h. Scale bars: 20 μm. (B) Effect of EX527 on CMS-induced autophagy. hTM cells were subjected to CMS for 24 h in the presence or absence of EX527 (25 or 50 µM). The expression levels of LC3 were measured in whole cell lysates by WB. ACTB is used as a loading control. (C) Quantification of LC3 expression in cytosolic and nuclear fractions of hTM cells subjected to CMS for 24 h in the presence or absence of EX527 (50 µM). s.e., short exposure; l.e., long exposure. LC3-I and LC3-II band intensities were normalized by TUBA4A and H2B in cytosolic and nuclear fractions, respectively (D). Data are shown as the mean ± S.D. (n = 3), *, p < 0.05, two-tailed unpaired Student’s t-test. ns: not significant, nd: not detected.

Figure 7.

Nuclear LC3 interacts with the ribophagy receptor NUFIP1 and promotes its nucleocytoplasmic distribution in hTM cells with CMS. Co-IP analysis in nuclear fraction of control cells transduced with (A) AdGFP or AdGFP-LC3 or (B) AdGFP-LC3 expressing cells with or without CMS. IP was performed using GFP-Trap; co-immunoprecipitated proteins were detected by WB, 5 µg of protein were loaded for input control. (C) WB blot analysis of NUFIP1 and LC3 in fractionated cytosolic, nuclear and membrane-bound organelle enriched fractions from NS and hTM cells under CMS. TUBA4A and H2B are used as a loading controls for cytosolic and nuclear fraction, respectively. LAMP1 and LAMP2 were used as markers for membrane-bound organelle enriched fractions containing lysosomes. (D) NUFIP1 band intensity was normalized and fold expression calculated. Data are shown as the mean ± S.D. (n = 3), **, p < 0.01, two-tailed unpaired Student’s t-test. ns: not significant. (E) Representative immunostaining of NUFIP1 (green) and LAMP2 (red) in NS and hTM cells subjected to CMS (8% elongation, 24 h). Yellow arrowheads indicate nuclear NUFIP1 staining; white arrows indicate co-localization of NUFIP1 and LAMP2.

Figure 8.

CMS does not trigger ribophagy in hTM cells. (A) hTM cells were transfected with siATG5,7 or siNC for 48 h and then subjected to CMS (8% elongation, 24 h). Protein levels of LC3, RPL26 and RPS15A were evaluated by WB. Cells cultured for 24 h in HBSS served as positive control. (B) Band intensities were normalized with ACTB and fold expression calculated. Data are shown as the mean ± S.D. (n = 3), two-tailed unpaired Student’s t-test. ns: not significant.

Figure 9.

Schematic diagram summarizing the results and working hypothesis. Under non-stretched control conditions, LC3 resides primarily in the cytosol, shuttling in-and-out of the nucleus. LC3 enters the nucleus by passive diffusion and exits in an XPO1-dependent manner. Blockage of active nuclear export by LepB leads to the accumulation of LC3 in the nucleus interacting with PML bodies. Mechanical stress triggers activation of autophagy and nuclear translocation of LC3 to the nucleolus, where it interacts with the autophagy receptor NUFIP1. We propose LC3-NUFIP1 complex to act as a surveillance mechanism that recognizes stretch-induced damaged nuclear proteins and facilitate their export, either via active nuclear transport or nuclear envelope budding, for autophagic degradation in the cytosol.