A translational program that suppresses metabolism to shield the genome

Abstract

Translatome reprogramming is a primary determinant of protein levels during stimuli adaptation. This raises the question: what are the translatome remodelers that reprogram protein output to activate biochemical adaptations. Here, we identify a translational pathway that represses metabolism to safeguard genome integrity. A system-wide MATRIX survey identified the ancient eIF5A as a pH-regulated translation factor that responds to fermentation-induced acidosis. TMT-pulse-SILAC analysis identified several pH-dependent proteins, including the mTORC1 suppressor Tsc2 and the longevity regulator Sirt1. Sirt1 operates as a pH-sensor that deacetylates nuclear eIF5A during anaerobiosis, enabling the cytoplasmic export of eIF5A/Tsc2 mRNA complexes for translational engagement. Tsc2 induction inhibits mTORC1 to suppress cellular metabolism and prevent acidosis-induced DNA damage. Depletion of eIF5A or Tsc2 leads to metabolic re-initiation and proliferation, but at the expense of incurring substantial DNA damage. We suggest that eIF5A operates as a translatome remodeler that suppresses metabolism to shield the genome.

Fig. 1. MATRIX analysis reveals eIF5A-adaptive engagement during anaerobic acidosis.

a Workflow schematic of MATRIX, an unbiased, high-throughput platform that measures the activity of translational assets. MATRIX analysis of differential translation-factor utilization in human cells (U87MG) exposed to hypoxia acidosis (1% O₂, pH 6.0, 24 h) compared to b basal (21% O₂, pH 7.4, 24 h) and c hypoxia-neutral pH (1% O₂, pH 7.4, 24 h) conditions, using the ratio of heavy polysome to free abundance as the readout. Hypoxia acidosis-activated translation factors (dark-blue bars), basal-activated translation factors (red bars), and hypoxia-neutral pH-activated translation factors (light-blue bars). d Representative immunoblots of U87MG ribosome-density fractions from indicated conditions. Mono: monosome fraction. Poly: polysome fractions. f Representative immunoblots of U87MG were subjected to indicated treatments. e Representative images of eIF5A immunocytochemistry in human (WI-38, U87MG) and mouse (NIH/3T3) cell lines subjected to indicated treatments. Data represent mean ± SEM (n = 3). Scale bars: 20 μm.

Fig. 2. eIF5A enables metabolic depression during anaerobic acidosis.

a Relative ATP utilization, b transcriptional intensity, and c translational intensity in U87MG replete or depleted of eIF5A under hypoxia acidosis conditions. a, b NS nonsilencing. Data represent mean ± SEM (n = 3). An asterisk indicates p = 0.024, 0.028 (a, b) compared to NS siRNA, two-sided student’s t test. d Ki-67 and p21 immunocytochemistry, e DNA replication (BrdU staining), and f Congo red staining for A bodies in U87MG replete or depleted of eIF5A under hypoxia acidosis conditions (72 h) (n = 5). Quantitation in (d) and (e) represents mean ± SEM (n = 7, 5). Scale bars: 20 μm. g Effect of eIF5A knockdown on steady-state cell numbers in U87MG subjected to indicated treatments. NS nonsilencing. Data represent mean ± SEM (n = 3). An asterisk indicates p = 0.035 compared to NS siRNA, two-sided student’s t test. h Effect of eIF5A knockdown on steady-state cell numbers in human and mouse cell lines under hypoxia acidosis conditions. NS nonsilencing. Data represent mean ± SEM (n = 3). An asterisk indicates p = 0.001, 0.001, 0.047, and 0.025 (U87MG, MCF7, HCT116, and WI-38) compared to the corresponding NS siRNA, two-sided student’s t test. i Representative ultrasound images and j tumor volume measurements of mouse xenograft tumor-formation assays using MCF7 replete or depleted of eIF5A and pretreated with hypoxia acidosis for 72 h. Data represent mean ± SEM (n = 6). An asterisk indicates p = 0.021, 0.031, 0.004, and 0.014 (days 7, 14, 17, and 21) compared to NS siRNA, two-sided student’s t test.

Fig. 3. TMT-pSILAC identifies the translatome of anaerobic acidosis-dependent metabolic depression.

Proteomic output analysis by TMT-pSILAC followed by mass spectrometry cells subjected to hypoxia acidosis (HA) versus a basal normoxia-neutral pH (NN) and b hypoxia-neutral pH (HN) conditions. HA enrichment: ≥1.5× increase; NN/HN enrichment: ≤0.75× decrease. c Graph depicting HA-enriched proteins (≥1.5× increase), blue = proteins that are upregulated in HA compared to both conditions, yellow proteins that are upregulated in HA compared to HN but not NN, red = proteins that are upregulated in HA compared to NN but not HA. Representative d immunoblots and e immunocytochemistry images of hypoxia-acidosis (HA)-specific translatome targets in U87MG. NN normoxia-neutral pH, HN hypoxia-neutral pH, n = 5. Scale bars: 20 μm. f Representative images of Cyr61 (HA-specific marker, red) and HIF-1α (classic hypoxia marker, green) immunohistochemistry in embryonic day 12 (ED12) CD1 mouse embryo. Scale bars: 1000 μm. Enlarged images of indicated sections (top) are presented on bottom panels (n = 3). Scale bars: 100 μm. g Representative images of Pai1 (HA-specific marker, green) and HIF-1α (hypoxia marker, red) immunohistochemistry in human hypoxic tumor core sections (n = 3). Scale bars: 1000 μm.

Fig. 4. eIF5A suppresses global metabolism and proliferation through Tsc2/mTORC1.

a–c Representative immunoblots of U87MG replete or depleted of eIF5A under hypoxia-acidosis conditions (n = 3). d Representative immunoblots of U87MG replete or depleted of Tsc2 under hypoxia-acidosis conditions (n = 3). e Effects of eIF5A and Tsc2 knockdown on cellular ATP levels under hypoxia-acidosis conditions. NS nonsilencing. Data represent mean ± SEM (n = 3). An asterisk indicates p = 0.024 and 0.012 (eIF5A siRNA and Tsc2 siRNA) compared to NS siRNA, two-sided student’s t test. f Representative images of Ki-67 immunocytochemistry in U87MG (left panel) and MCF7 (right panel) replete or depleted of Tsc2 under hypoxia-acidosis conditions. NS nonsilencing. Data represent mean ± SEM (n = 5). Scale bars: 20 μm. Effect of Tsc2 knockdown on U87MG g cell number, h transcriptional intensity, and i Congo red staining for A bodies. Scale bars: 20 μm. NS nonsilencing. Data represent mean ± SEM (n = 3 (g, h); n = 5 (i)). An asterisk indicates p = 0.046 and 0.042, p = 0.007 and 0.002 (g, h eIF5A siRNA and Tsc2 siRNA) compared to NS siRNA, two-sided student’s t test. j Representative images of Ki-67 immunocytochemistry in U87MG replete or depleted of eIF5A and treated with mTORC1 inhibitors rapamycin and everolimus under hypoxia-acidosis conditions. NS nonsilencing. Data represent mean ± SEM (n = 5). Scale bars: 20 μm. k Effect of mTORC1 inhibition on eIF5A-replete or -depleted U87MG cell number under hypoxia-acidosis conditions. NS nonsilencing. Data represent mean ± SEM (n = 6). An asterisk indicates p = 0.026 and 0.022 (Rapamycin + eIF5A siRNA, Everolimus + eIF5A siRNA) compared to Vehicle + eIF5A siRNA, two-sided student’s t test.

Fig. 5. The eIF5A/Tsc2 axis averts DNA damage.

Assessment of the effect of eIF5A silencing on DNA damage by a alkaline comet analysis (n = 3) and b TUNEL measurements in U87MG subjected to indicated conditions. Data represent mean ± SEM (n = 5). Asterisk indicates p = 0.001 compared to NS siRNA, two-sided student’s t test. c (Top) Representative images of yH2AX foci in U87MG subjected to indicated conditions. Scale bars: 20 μm. (Bottom) analysis of yH2AX foci (n = 5), asterisk indicates p = 0.00001 and 0.00001 (NN eiF5A siRNA, HA eIF5A siRNA) compared to NS siRNA, two-sided Mann–Whitney U test. The top of the box denotes Q3, the bottom of the box represents Q1; middle line denotes median; X represents mean; bottom whisker denotes minimum: 1st quartile—(1.5*IQR); top whisker denotes maximum: 3rd quartile + (1.5*IQR). d TUNEL analyses of the effects of Tsc2 knockdown and mTORC1 inhibition (by Torin 1 and 2) on DNA damage in cells replete or depleted of eIF5A under hypoxia-acidosis conditions. Data represent mean ± SEM (n = 5). An asterisk indicates p = 0.005, 0.004, 0.002, and 0.027 (eIF5A siRNA + vehicle, Tsc2 siRNA + vehicle, and eIF5A siRNA + torin2) compared to NS siRNA + Vehicle. ^† indicates p = 0.006 eIF5A siRNA compared to Vehicle and Tsc2 siRNA + Vehicle, two-sided student’s t test. e Top panel: representative images of yH2AX foci in U87MG cells depleted of Tsc2, and in eIF5A-replete or -depleted cells treated with mTORC1 inhibitors (Torin 1 and 2) under hypoxia-acidosis conditions. Bottom panel: analysis of yH2AX (n = 5), an asterix indicates p = 0.0005 and 0.00001 (eIF5A siRNA, tsc2 siRNA) compared to NS siRNA + vehicle, two-sided Mann–Whitney U test. Top of the box denotes Q3, the bottom of the box represents Q1; middle line denotes median; X represents mean; bottom whisker denotes minimum: 1st quartile—(1.5*IQR); top whisker denotes maximum: 3rd quartile + (1.5*IQR).

Fig. 6. Mechanism of eIF5A-mediated Tsc2 protein induction.

a Co-immunoprecipitated mRNA levels of eIF5A-regulated and nonregulated mRNAs relative to IgG isotype control pulldown. Data represent mean ± SEM (n = 3). An asterisk indicates p = 0.005, 0.031, and 0.001 (Tsc2, c-Jun, and Sdc4) compared to IgG control, two-sided student’s t test. b Effect of eIF5A knockdown on mRNA subcellular localization of indicated mRNAs under hypoxia acidosis conditions. NS nonsilencing. Data represent mean ± SEM (n = 3). An asterisk indicates p = 0.007, 0.041, and 0.027 (Tsc2, c-Jun, and Sdc4) eIF5A siRNA compared to NS siRNA, two-sided student’s t test. c mRNA fluorescent in situ hybridization. NS nonsilencing. One-third exposure level was used for Tsc2 mRNA FISH under eIF5A siRNA conditions relative to all other conditions to avoid signal saturation (n = 5). Scale bars: 20 μm. d Effect of eIF5A knockdown on translation efficiencies of indicated mRNAs under hypoxia-acidosis conditions. NS nonsilencing. Data represent mean ± SEM (n = 3). An asterisk indicates p = 0.0.004, 0.002, and 0.006 (Tsc2, c-Jun, and Sdc4) eIF5A siRNA compared to NS siRNA, two-sided student’s t test. e Effect of leptomycin B treatment on eIF5A protein subcellular localization in U87MG under indicated conditions. Vehicle: DMSO (n = 5). Scale bars: 20 μm. f Effect of leptomycin B treatment on Tsc2 mRNA subcellular localization under hypoxia-acidosis conditions. Vehicle (Veh): DMSO. Data represent mean ± SEM (n = 5). An asterisk indicates p = 0.015 compared to DMSO vehicle, two-sided student’s t test. g Representative immunoblot of Tsc2 protein levels in U87MG treated with leptomycin B (n = 3).

Fig. 7. pH-sensing Sirt1 modulates eIF5A through deacetylation.

a Representative images of eIF5A immunocytochemistry showing eIF5A subcellular localization in U87MG subjected to indicated conditions. Data represent mean ± SEM (n = 5). Scale bars: 20 μm. b Representative immunoblots of U87MG subjected to indicated conditions. NN normoxia-neutral pH, HN hypoxia neutral, HA hypoxia acidosis (n = 3). Representative c immunoblots (n = 3) and e immunocytochemistry images (n = 5) of U87MG were treated with the indicated compounds under the indicated conditions. Scale bars: 20 μm. Ex-527: Sirt1 inhibitor; AGK2: Sirt2 inhibitor; DMSO: vehicle. d Representative immunoblots of U87MG depleted of Sirt1 (n = 3). f Representative immunoblots of U87MG ribosome-density fractions treated with the indicated compounds. Mono: light monosome fraction. Poly: heavy-polysome fractions (n = 3). g Effect of Sirt1 inhibition (using ex-527) on Tsc2 mRNA subcellular localization under hypoxia-acidosis conditions. NS nonsilencing. Data represent mean ± SEM (n = 3). An asterisk indicates p < 0.05 compared with DMSO vehicle, two-sided student’s t test. h Representative immunoblots of U87MG treated with the Sirt1 inhibitor ex-527 (n = 3). i Summary model of the Sirt1/eIF5A/Tsc2/mTORC1 pathway that enables metabolic depression and proliferative inhibition during anaerobic acidosis to prevent DNA damage.