A TFEB nuclear export signal integrates amino acid supply and glucose availability

Abstract

How cells coordinate the response to fluctuating carbon and nitrogen availability required to maintain effective homeostasis is a key issue. Amino acid limitation that inactivates mTORC1 promotes de-phosphorylation and nuclear translocation of Transcription Factor EB (TFEB), a key transcriptional regulator of lysosome biogenesis and autophagy that is deregulated in cancer and neurodegeneration. Beyond its cytoplasmic sequestration, how TFEB phosphorylation regulates its nuclear-cytoplasmic shuttling, and whether TFEB can coordinate amino acid supply with glucose availability is poorly understood. Here we show that TFEB phosphorylation on S142 primes for GSK3β phosphorylation on S138, and that phosphorylation of both sites but not either alone activates a previously unrecognized nuclear export signal (NES). Importantly, GSK3β is inactivated by AKT in response to mTORC2 signaling triggered by glucose limitation. Remarkably therefore, the TFEB NES integrates carbon (glucose) and nitrogen (amino acid) availability by controlling TFEB flux through a nuclear import-export cycle.

Fig. 1

TFEB is subject to nuclear export. a Immunofluorescence with indicated antibodies using control MCF7 cells or those treated with Torin 1 (250 nM, 1 h). n > 30 cells per condition. b Real time imaging of MCF7 stable cell lines expressing doxycycline-inducible TFEB-GFP (iTFEB-GFP) imaged at indicated intervals in the absence of doxycycline. Images derived from Supplementary Movie 1. c Live cell imaging of MCF7 iTFEB-GFP cells before and after 1 h Torin 1 (250 nM) or 2 h U0126 (20 μM) treatment or after removal of the drug (upper panels) as indicated, or 30′ and 1 h after removal of Torin 1 (lower panels). n > 30 cells per condition. d Imaging of MCF7 iTFEB-GFP or iGFP cells before and after LMB (2 h; 20 nM) treatment. Quantification is derived from one experiment that is representative of 2 independent experiments. n > 30 cells per condition. e Immunofluorescence using MCF7 or HT29 cells and indicated antibodies with or without treatment with Torin 1 (1 h; 250 nM), LMB (2 h; 20 nM), and U0126 (3 h; 20 μM). n > 500 cells per condition. Details of all quantifications including number of replicates and cells imaged are provided in Supplementary Table 1 and Methods. Scale bars = 20 μM. Error bars = SD. Student’s t-test; *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001, n.s. not significant

Fig. 2

Identification of a TFEB nuclear export signal. a Live cell fluorescence images of indicated WT and mutant TFEB-GFP proteins in stably expressing MCF7 cells. Cells were treated with DMSO, LMB (20 nM; 2 h) or Torin 1 (20 nM; 1 h). n > 30 cells per condition. b Alignment of TFEB amino acids 129–152 and equivalent TFE3 residues. Amino acids highlighted in blue conform to the consensus for an NES. c Live cell fluorescence imaging of MCF7 cells stable expressing indicated fusion proteins imaged before or after treatment with LMB (2 h; 20 nM), Torin 1 (1 h; 250 nM), or U0126 (3 h; 20 μM). Imaging was performed after fixation. n > 50 cells per condition. Error bars = SD. Student’s t-test; ***P < 0.001, ****P < 0.0001, n.s. not significant. d Real time fluorescence imaging of MCF7 cells stably expressing the indicated fluorescent TFEB (129–152) cargo vectors before or after treatment with LMB (20 nM; 2 h). Scale bars = 20 μM

Fig. 3

TFEB is subject to CRM1-mediated nuclear export. a Live cell steady state images of indicated cargo vectors stably expressed in MCF7 cells. n > 120 cells per condition. b Western blot of HT29 cells expressing doxycycline-inducible WT or indicated mutant TFEB-GFP (−Dox). c Fluorescence images of HT29 cells stably expressing indicated WT and mutant TFEB-GFP treated with Torin 1 (25 nM) for 1 h before cells were washed and placed in Torin 1-free medium for 15 or 60 min. n > 300 cells per condition. d Western blot using indicated antibodies to detect CRM1 and RAN after pull down of immobilized HIS-tagged TFEB (aas 1–200). All proteins were bacterially expressed and purified. Purified TFEB was visualized by Coomassie staining (lower panel). e Schematic depicting the role of CRM1 and RAN in nuclear export using TFEB as a potential substrate. Top: RAN-WT; Bottom, RAN-Q69L. See text for details. f Fluorescence images of steady state subcellular localization of MCF7 cells stably expressing the TFEB 1-159-NLS-cargo together with either mCherry-LacR or mCherry RAN-Q69L. Cells were treated with DMSO, Torin 1 (250 nM) alone or with LMB (20 nM) as indicated. Subcellular localization of TFEB-GFP was quantified for cells expressing mCherry. n > 100 cells per condition. In the RanQ experiment, only mCherry positive cells were counted. Scale bars = 20 μM. Error bars = SD. Student’s t-test; ****P < 0.0001, n.s. not significant

Fig. 4

Phosphorylation at S142 primes for GSK3β phosphorylation at S138. a Fluorescence images of cells expressing indicated WT and mutant cargo vectors in the presence and absence of BIO (10 μM). n > 50 cells per condition. Error bars = SD. Scale bar = 20 μM. ****p < 0.0001, n.s. not significant. b TFEB amino acids 129–152 showing the GSK3β consensus sequence. The S138 GSK3β site is highlighted in blue, the ERK/mTOR phosphorylation site S142 is boxed, as are three additional potential GSK3β sites that could be primed by phosphorylation at S138. Peptides 1–12 were spotted onto a membrane (see key corresponding to peptides in lower left panel) and subject to GSK3β phosphorylation in vitro. Peptides highlighted in red were phosphorylated by GSK3β. Lower panels: Peptide key corresponding to peptides above spotted onto a membrane (left) and visualized by UV (middle). The result of the GSK3β kinase assay is shown in the right panel and the red boxes correspond to phosphorylated peptides highlighted in red above. c Coomassie-stained gel of bacterially expressed and purified 6xHIS-tagged TFEB amino acids 1–200 (left) and associated Mass Spec spectrum (right). d Extracted ion chromatograms of tryptic peptides covering S138 and S142 in their differential phosphorylation states after indicated kinase treatment. A peptide with a single phosphorylation on S138 was detected at low intensity in all samples, including those not treated with GSK3β, while a peptide with a phosphorylation event on S142 was only detected in the ERK and GSKβ + ERK treated samples. Peptides dually phosphorylated on both S138 and on S142 were detected only after GSKβ + ERK treatment. Numbers represent maximum ion count for each peptide/phospho-peptide. e Quantification of TFEB in vitro kinase/Mass Spec data. The pie-charts illustrate the signal contribution of each (phospho-)peptide variant to total signal intensity covering the analyzed phosphorylation locus after kinase treatments. Assuming that each peptide variant has identical ionization characteristics, the charts reflect the stoichiometry of the different phospho-proteoforms of TFEB after kinase treatment

Fig. 5

S138 and S142 control nuclear export. a Live cell imaging of MCF7 cell lines stably expressing the indicated TFEB-cargo vectors. Cells were treated with Torin 1 (20 nM; 1 h) before the medium was changed and cells imaged at indicated times. n > 200 cells per condition. b Immunoprecipitation assay of extracts of cells transfected with FLAG-TFEB WT or S138A mutant. After immunoprecipitation, western blots were probed with anti-FLAG and anti-14-3-3 antibody. c Immunofluorescence of HT29 cells expressing stable doxycycline-inducible TFEB WT or S138A mutant treated with Torin 1 (1 h; 250 nM) showing co-localization with lysosomes marked using anti-LAMP2. Scale bar 10 μM. For lysosome co-localization quantification, lysosomes were stained by anti-LAMP2 antibody and their total numbers were counted. The percentage of TFEB lysosome co-localization was defined by the percentage of lysosomes with positive TFEB-GFP co-localization. N > 50 cells per condition. d Western blot of HT29 or MCF7 cells stably expressing indicated WT or mutant TFEB probed with anti-TFEB antibody or an antibody recognizing specifically phospho-S142. Actin is a loading control. e Fluorescence assays for MCF7 cells stably expressing doxycycline-inducible WT TFEB-GFP and S138A and S142A mutants in the absence of doxycycline transfected with mCherry RAN-Q69L. Quantification is for mCherry-RAN Q69L-expressing cells. n > 40 cells per condition. Student’s t-test; *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001. Error bars = SD. Scale bars = 20 μM

Fig. 6

TFEB nuclear export is regulated by Glucose. a Immunofluorescence of HT29 cells grown under indicated conditions. n > 100 cells per condition. b Immunofluorescence of MCF7 cells grown in DMEM, minus glucose, or +Torin 1 (250 nM) or HBSS. n > 30 cells per condition. c Live cell imaging of MCF7 iTFEB-GFP cells before or after glucose deprivation or Torin 1 (250 nM) treatment. d Left: western blot of HT29 cells expressing control shRNA or shTFEB grown in the presence or absence of Bafilomycin A (10 μM). Right: real time qRT PCR of CDKN1A (p21) and RRAGD mRNAs from shControl or shTFEB cells grown with or without glucose (3 h) with or without shTFEB. n = 3 individual experiments. Error bars = SD. Two-way ANOVA, *P < 0.05, ***P < 0.001, ****P < 0.0001, n.s. not significant. e Immunofluorescence of HT29 cells grown under indicated conditions. n > 100 cells per condition. f Western blot of fractionated HT29 cells grown under indicated conditions. g Immunofluorescence of HT29 cells grown in DMEM, or starved of glucose 1 h before refeeding with glucose alone or with LMB (20 nM), BIO (10 μM) or cycloheximide (50 μg ml⁻¹). n > 500 cells per condition. h Western blot of HT29 cells treated with Toirn 1 (250 nM), LMB (2 nM) or BIO (10 μM) as indicated. i Fluorescence assay of TFEB-GFP WT and indicated mutants stable expressed in HT29 cells. Cells were grown in DMEM, starved of glucose for 2 h, then refed with glucose for 15 or 30 min and fixed before imaging. j Western blot of HT29 cells stably expressing doxycycline-inducible kinase dead (K.D.) or constitutively active (C.A.) HA-GSK3β and induced with 0, 10 and 100 ng ml⁻¹ doxycycline for 20 h. k Immunofluorescence using anti-TFEB or anti-HA antibodies of HT29 cells expressing doxycycline-inducible constitutively active (C.A) or kinase dead (K.D.) GSK3β or empty vector grown in the presence of doxycycline (100 ng ml⁻¹; 20 h) in the presence of glucose or deprived of glucose for 1 h. n > 200 cells per condition. Scale bars = 20 μM. Error bars = SD. Student’s t-test; *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001

Fig. 7

mTORC2 regulates GSK3β in response to glucose limitation. a Schematic showing the mTORC1 and mTORC2 complexes and downstream targets. b–d Western blots using indicated antibodies of HT29 cells grown under indicated conditions. Unless otherwise indicated cells were placed in HBSS for 2 h, or treated with Torin 1 (250 nM) or starved of glucose for 2 h and where indicted refed glucose for 1 h. e Western blot using indicated antibodies of HT29 cells grown in DMEM or starved of glucose for 4 h and treated with DMSO, MK2206 (3 μM), or LY333531 (5 μM). f Western blot using indicated antibodies of HT29 cells. 48 h after transfection with RICTOR-specific or control siRNAs cells were glucose starved for 2 h, as indicated. g Western blot using indicated antibodies of HT29 cells transfected with control or RICTOR-specific siRNA for 48 h prior to glucose starvation for indicated times. h Western blot using indicated antibodies of HT29 cells grown in the presence of absence of glucose (4 h) of cells treated with Rapamycin (1 μM) or Torin 1 (250 nM). i Western blot using indicated antibodies of HT29 cells following glucose starvation for indicated times

Fig. 8

Model to explain regulation of TFEB subcellular localization. a Summary of key regulatory phosphorylation events and their impact on TFEB subcellular location. b Model to explain the potential regulation of TFEB when amino acids or glucose are limiting. Under these conditions increased phosphatase activity targeting S211, and likely other phosphorylation sites, will release TFEB from its cytoplasmic anchor. Whether TFEB is retained in the nucleus will be dependent on the phosphorylation status of S138 and S142. Absence of phosphorylation of S138 or S142 will lead to nuclear retention. Note however that in this model dephosphorylation of S142 or S138 is not a necessary pre-requisite for nuclear entry