Plasticity of the mammalian integrated stress response

Cell lines, cell culture and shRNA treatments

MEFs and HEK293T cells were grown in high-glucose Dulbecco’s modified Eagle’s medium (DMEM, Gibco number 11960044) supplemented with 10% defined fetal bovine serum (FBS; Gibco number 26140079), 2 mM l-glutamine, 100 units ml⁻¹ penicillin and 100 μg ml⁻¹ streptomycin (Gibco number 10378016). Mouse ES cells were grown in Iscove’s modified Dulbecco’s medium (Gibco number 12440053) supplemented with 0.1 mM 2-mercaptoethanol (Gibco number 21985023), 1× MEM non-essential amino acids solution (Gibco number 11140050), 50 units ml⁻¹ penicillin and 50 µg ml⁻¹ of streptomycin (Gibco number 15070063), 1,000 units ml⁻¹ ESGRO leukaemia inhibitory factor supplement for mouse ES cell culture (Sigma ESG1106) and 20% Oneshot ES cell heat-inactivated FBS (Gibco number 16141079). Herceptin-sensitive parental (BT474-P) and Herceptin-resistant (BT474-R) human breast cancer cells⁴⁵ were grown in high-glucose DMEM (Gibco number 11960044) supplemented with 10% heat-inactivated FBS (Gibco number 10082147), 2 mM l-glutamine (Gibco number 10378016), 100 units ml⁻¹ penicillin, 100 μg ml⁻¹ streptomycin and 1 mM sodium pyruvate (Gibco number 11360070). Mouse induced pluripotent stem cells (iPSCs) carrying the eIF2Bε(R132H) alteration (targeting G2723A in the eIF2B5 gene) in the eIF2Bε subunit of eIF2B, alongside WT iPSCs, were differentiated into oligodendrocyte precursor cells (OPCs) as previously described^46,47,48. iPSC-derived OPCs were grown in DMEM/F12 (Thermo Fisher Scientific number 11320082), 1× N2 supplement (R&D Systems AR009), 1× B-27 without vitamin A supplement (Thermo Fisher Scientific number 12587010) and 1× Glutamax, supplemented with 20 ng ml⁻¹ fibroblast growth factor 2 (R&D Systems 233-FB) and 20 ng ml⁻¹ platelet-derived growth factor-AA (R&D Systems 221-AA). Mouse ES cells containing the Eif2b5^R191H/R191H mutation or NIH 3T3 cells containing a point mutation in ΔuORF2 (ATG to ATA) were generated through CRISPR–Cas9 genome editing technology in the Case Western Transgenic and Targeting Facility. NIH 3T3 cells containing a point mutation in ΔuORF2 (ATG to ATA) were generated through CRISPR–Cas9 genome editing technology in SYNTHEGO.

For glucose-limitation experiments, cells were grown in no-glucose DMEM (Gibco number 11966025) supplemented with 10% FBS (Gibco number 26140079), 100 units ml⁻¹ penicillin, 100 μg ml⁻¹ streptomycin (Gibco number 15070063) and indicated glucose concentrations (Sigma G8644). All cells were maintained at 37 °C with 5% CO₂ for all experiments before specific treatments. For shRNA knockdown experiments, lentiviral particles expressing shRNA against target mRNAs were prepared and propagated in HEK293T cells as described previously^49,50 using the second-generation pLKO.1, psPAX2 and pMD2.G vectors. After two rounds of lentiviral infection, cells were selected under puromycin (30 μg ml⁻¹ in MEFs, 2 μg ml⁻¹ in mouse ES cells and 1.5 μg ml⁻¹ in BT474) for 3 days. The last-selection-day cells were passaged for experimentation. Day 4 denotes 1 day after passage of the cells in the presence of puromycin. Puromycin was removed 3 h before collection or treatment of the cells. Proliferation was monitored by cell counting with trypan blue exclusion of dead cells as described previously⁵¹. In addition, proliferation was estimated using CellTiter-Glo Luminescent Cell Viability Assay kit (Promega G7572) according to the manufacturer’s instructions.

shRNAs and RT–qPCR primers

Plasmid expressing shRNA against Eif2b5 (TRCN0000109990), Eif4e (TRCN0000077474), Upf1 (number 1: TRCN0000009663; number 2: TRCN0000274486), Eif2s2 (TRCN0000096876), Ddx3x (TRCN0000287239) and MISSION pLKO.1-puro Non-Target shRNA Control Plasmid DNA (Sigma-Aldrich, SHC016) were purchased from Sigma-Aldrich.

For RT–qPCR analysis, we used the following primer sets: ATF4 (forward (−) GTTTGACTTCGATGCTCTGTTTC; reverse (+) GGGCTCCTTATTAGTCTCTTGG); GADD34 (forward (−) TACCCCTGTCTCTGGTAACCT; reverse (+) TGGCTTTGCATTGTACTCATCA); IBTKα (forward (−) CCACCGTCTGCAGGATTATT, reverse (+) CTCGACCTTATCCGAATGGA); ATF5 (forward (−) AAGCTTGTAAGGCCCCCTGT, reverse (+) GTGCGCTTGATGTAGGGATT); BiP (forward (−) ACTTGGGGACCACCTATTCCT, reverse (+) ATCGCCAATCAGACGCTCC); α-tubulin (forward (−) CACTTACCACGGAGATAGCGA, reverse (+) ACCTTCTGTGTAGTGCCCCTT); GAPDH (forward (−) CGCCTGGAGAAACCTGCCAAGTATG, reverse (+) GGTGGAAGAGTGGGAGTTGCTGTTG); CHOP (forward (−) CTGGAAGCCTGGTATGAGGAT, reverse (+) CAGGGTCAAGAGTAGTGAAGGT); XBP1s (forward (−) GAGTCCGCAGCAGGTG, reverse (+) CTGGGAGTTCCTCCAGACTA); β-actin (forward (−) CTGGCACCACACCTTCTACAATG, reverse (+) GGTCATCTTTTCACGGTTGGC); GADD45a (forward (−) GAGGAATTCTCGGCTGCAGA, reverse (+) CACGTTATCGGGGTCTACGT).

Chemicals, reagents and antibodies

Chemicals used in this study: Tg (400 nM, Sigma-Aldrich T9033); sodium arsenite (1 mM, Sigma S7400); CPA (100 μM (BT474) and 200 μM (MEFs and mouse ES cells) Tocris number 1235); actinomycin D (10 μg ml⁻¹, Sigma-Aldrich A9415); cycloheximide (100 μg ml⁻¹, Sigma C7698); salubrinal (15 μM, Tocris number 3657); Torin 1 (250 nM, Tocris number 4247); LiCl (10 mM, Sigma); Herceptin (20 μg ml⁻¹, Genentech); 4EGI-1 (200 μM, Med Chem Express HY-19831); SGC-CK2-1 (5 μM, Cayman number 34103).

Antibodies used in this study: anti-PERK (1:1,000, Cell Signaling Technology number 3192); anti-eIF4E (1:1,000, Cell Signaling Technology number 9742); anti-eIF2Bε (1:1,000, Cell Signaling Technology number 3595); anti-eIF2α (1:1,000, Cell Signaling Technology number 9722); anti-eIF2α-phospho(Ser51) (1:3,000, Abcam ab32157); anti-ATF4 (1:1,000, Cell Signaling Technology number 11815); anti-α-tubulin (1:4,000, Sigma T9026); anti-citrate synthase (1:1,000, Sino Biological 14083-T46); anti-GADD34 (1:3,000, Proteintech 10449-1-AP); anti-BiP (1:1,000, Cell Signaling Technology number 3177); anti-DDX3X (1:1,000, Cell Signaling Technology number 2635); anti-CHOP (1:1,000, Cell Signaling Technology number 2895); anti-PCK2 (1:1,000, Cell Signaling Technology number 6924); anti-UPF1 (1:1,000, Cell Signaling Technology number 12040); anti-S6 ribosomal protein (1:1,000, Cell Signaling Technology number 2217); anti-phospho-S6 ribosomal protein (Ser240/244) (1:1,000, Cell Signaling Technology number 5364); anti-GSK3β (1:1,000, Cell Signaling Technology number 9315); anti-phospho-GSK3β (Ser9) (1:1,000, Cell Signaling Technology number 9323); anti-4E-BP1 (1:1,000, Cell Signaling Technology number 9644), anti-phospho-4E-BP1 (Ser65) (1:1,000, Cell Signaling Technology number 9451) and anti-eIF2β (1:1,000, Santa Cruz sc-9978). For immunostaining: anti-G3BP1 (1:200, Santa Cruz sc-81940); anti-DDX3X (1:200, Bethyl A300-474A); anti-O1 (1:100, CCF Hybridoma Core Facility); anti-MBP (1:100, Abcam ab7349); anti-beta actin (1:4,000, Abcam ab6276). Hoechst (Thermo Fisher Scientific number 3570) was used to detect nuclei in immunostaining experiments.

Cell extract preparation for western blotting

Cells were washed twice with ice-cooled 1× PBS before lysis. Ice-cooled (4 °C) lysis buffer (50 mM Tris-HCl pH 7.5, 150 mM NaCl, 2 mM EDTA, 1% NP-40, 0.1% SDS, 0.5% sodium deoxycholate), supplemented with EDTA-free protease inhibitor (Sigma number 34693159001) and PhosSTOP phosphatase inhibitor (Sigma number 4906837001) was added to cells. Cells were scraped off, collected and sonicated on ice. Protein lysates were centrifuged for 5 min at 10,000g and 4 °C. Supernatant was collected and quantified using the DC Protein Assay Kit (Bio-Rad number 5000112). Lysate was diluted to 1 μg μl⁻¹ using lysis buffer. The diluted lysates were mixed with 5× sample loading buffer (300 mM Tris-HCl pH 6.8, 50% glycerol, 10% (v/v) β-mercaptoethanol, 10% (w/v) SDS and 50 mg bromophenol blue) for western blot analysis. Protein lysates were separated by SDS–PAGE before electrotransfer to Immobilon-P PVDF membrane (Sigma-Aldrich). When possible, membranes were stripped and re-probed, but in the cases in which this was not feasible (for example, antibodies from the same species in different dynamic ranges), the same lysates were simultaneously resolved on duplicate gels. Representative corresponding loading controls are shown.

Measuring in vitro GEF activity of eIF2B

eIF2B activity was measured as previously described⁶. In brief, cells were washed and scraped off in homogenization buffer (45 mM HEPES-KOH pH 7.4, 0.375 mM MaOAc, 75 mM EDTA, 95 mM KOAc, 10% glycerol, 1 mM dithiothreitol (DTT), 2.5 mg ml⁻¹ digitonin), supplemented with EDTA-free protease inhibitor (Sigma number 4693159001) and PhosSTOP phosphatase inhibitor (Sigma number 4906837001). Cell lysates were homogenized and quantified for protein concentration. eIF2B activity was calculated as the rate of exchange from [³H]eIF2α GDP to non-radioactive GDP at each time point.

Measurement of global protein synthesis

Protein synthesis rates were measured as previously described⁶. In brief, cells were treated with designated chemicals for the indicated durations. At the end of treatments, [³⁵S]Met and Cys (30 μCi ml⁻¹ EXPRE³⁵S Protein Labeling Mix (PerkinElmer NEG072002MC) was added to the cells for an additional 30 min. After labelling, cells were washed and lysed, and the radioactivity incorporated into proteins was determined by liquid scintillation counter. The protein synthesis rate was calculated as the rate of [³⁵S]Met and Cys incorporation to total cellular protein from the same lysate.

Polysome profile analysis and mRNA distribution

Cells were seeded in 150-mm culture dishes and grown up to 70% confluence (about 1.0 × 10⁷ cells). Cells were washed twice with cold PBS containing CHX (100 μg ml⁻¹), scraped off and pelleted at 4,000 r.p.m. for 10 min. The cell pellets were suspended in 500 μl of lysis buffer (10 mM HEPES-KOH at pH 7.4, 2.5 mM MgCl₂, 100 mM KCl, 0.25% NP-40, 100 μg ml⁻¹ CHX, 1 mM DTT), 200 units ml⁻¹ of RNase inhibitor (NEB number M0314) and EDTA-free protease inhibitor (Sigma number 693159001), kept on ice for 20 min and then passed 15 times through a 23-gauge needle. Lysates were cleared at 14,000 r.p.m. for 15 min, and supernatants (cell extracts) were collected and measured at absorbance of 260 nm. An equal amount (approximately 500 μg of lysate) was layered over 10–60% of cold sucrose gradients prepared in buffer (10 mM HEPES-KOH at pH 7.4, 2.5 mM MgCl₂, 100 mM KCl). Gradients were centrifuged at 35,000 r.p.m. in a Beckman SW41Ti rotor for 3 h at 4 °C. After centrifugation, 12 equal-sized fractions (1 ml per fraction) were collected. RNA from each fraction was isolated using TRIzol LS reagent (Invitrogen number 10296028) and an equal volume of RNA from each fraction was cDNA-synthesized using the SuperScript III First-Strand Synthesis SuperMix (Thermo Fisher Scientific number 18080044). The relative quantity of specific mRNAs was measured by RT–qPCR using the VeriQuest SYBR Green qPCR Master Mix (Thermo Fisher Scientific 756002000RXN) with the StepOnePlus Real-Time PCR System (Applied Biosystem). For conventional measurement of total RNA levels, cells were seeded in 60-mm culture dishes and grown up to 70% confluence (1.0–1.5 × 10⁶ cells) before treatment. Following indicated treatments, total intracellular RNA was isolated using TRIzol reagent (Invitrogen number 15596018). cDNAs were synthesized and relative RNA levels were measured by RT–qPCR as described above.

RNA-sequencing data preprocessing and quality control

RNA-sequencing libraries were prepared according to the TruSeq Stranded Total RNA protocol (Illumina) following the manufacturer’s instructions, and paired-end reads were obtained using a HiSeq2500 system (Illumina). The quality of sequencing reads was confirmed using FastQC (v0.11.4; http://www.bioinformatics.babraham.ac.uk/projects/fastqc/). For the removal of Illumina TruSeq adaptor sequences and low-quality base calls, BBmap (v36.59; https://www.osti.gov/servlets/purl/1241166) was used with the following parameters: k = 13, ktrim = n, useshortkmers = t, mink = 5, qtrim = t, trimq = 10, minlength = 25. Subsequently, resulting reads were mapped to the mm10 genome assembly using HISAT (v2.0.4, in addition to default parameters, –no-mixed and –no-discordant were applied)⁵². The aligned reads were summarized using htseq-count⁵³. Data quality was assessed using principal component analysis on trimmed mean of M values (TMM)-log₂-normalized counts using the PCAtools R package (v2.4.0; https://github.com/kevinblighe/PCAtools; parameters removeVar = 0.75 and scale = T).

Raw fastq files for the dataset from ref. ²⁴ were obtained from the National Center for Biotechnology Information Gene Expression Omnibus repository (GSE128092) and prepared the same way as described above, with a difference that the aligned reads were summarized using the featureCounts function of the RSubread (v2.6.4) R/Bioconductor package⁵⁴.

Analysis of differential translation in datasets using anota2seq

Genes with 0 mapped RNA-sequencing reads in one or more samples were discarded. The data were TMM-log₂-normalized and analysed using anota2seq^19,55 (v1.14.0, parameters: minSlopeTranslation = −1, minSlopeBuffering = −1, maxSlopeTranslation = 1.5, maxSlopeBuffering = 1.5, deltaPT = deltaTP = deltaP = deltaT = log₂(1.5) and FDR < 0.05). To classify genes into translation, offsetting or mRNA abundance gene expression modes, the anota2seqRegModes function within anota2seq was used.

Analysis of differentially expressed genes in the Eif2b5
^R191H/R191H, ΔORF1 mouse ES cells and ref. ²⁴ datasets

The datasets were analysed using DESeq2 (ref. ²⁶; v1.38.2). To identify differentially expressed genes, an FDR threshold of <0.05 was applied to the Eif2b5^R191H/R191H data and the datasets (all comparisons) from ref. ²⁴. Additionally, changes with an absolute fold change of <log₂(1.2) were filtered out in the Eif2b5^R191H/R191H dataset. In the ΔORF1 dataset, an FDR threshold of <0.01 and an absolute fold change of >log₂(1.5) were applied.

GO analysis

GO analysis was performed using GOstats (v2.68.0) using a hypergeometric test for categories from biological process ontology terms⁵⁶.

Bioenergetic analysis

OCR and ECAR were measured using a Seahorse XFe24 analyser (Agilent Technologies), as previously shown⁵⁷. In brief, mouse ES cells were fed with growth medium for 1 h, and then trypsinized and resuspended in growth medium. Cells were washed twice in Seahorse XF medium (10 mM glucose, 2 mM glutamine, 1 mM sodium pyruvate, pH 7.4) following centrifugation. Cells were subsequently seeded at a density of 100,000 cells in a volume of 200 μl per well on plates coated with 22.4 μg ml⁻¹ Cell-TAK (Corning number 354240). The plate was centrifuged at 200g for 1 min to allow cells to adhere and subsequently placed in a 37 °C non-CO₂ incubator for 1 h. Three measurements of OCR and ECAR were recorded under basal conditions, and following the injection of each compound (25 μl per injection): oligomycin (1 μM; Sigma O4876-25MG), FCCP (1.5 μM; Sigma C2920-10MG), rotenone (1 μM; Sigma R8875-1G) and antimycin A (1 μM; Sigma A8674-25MG), and monensin (20 μM; Sigma M5273-1G). Rates of oxidative ATP production (J ATP ox) and glycolytic ATP production (J ATP glyc), as well as bioenergetic capacity, were calculated as previously described⁵⁸. Bioenergetic capacity is defined by the maximum values of J ATP glyc and J ATP ox in cells⁵⁸. Buffering power of the Seahorse XF medium was measured as previously described⁵⁹. All values were normalized to protein content.

Immunofluorescent staining

Cells were plated on glass microscope coverslips (Thermo Fisher) in 6-cm culture dishes and were allowed to grow for 48 h. After the designated treatments, cells were washed twice with ice-cooled 1× PBS on ice. Cells were fixed with 4% paraformaldehyde for 10 min at room temperature on a shaker at 50 r.p.m. Fixed cells were washed twice with ice-cooled 1× PBS and incubated in PBST (1× PBS + 0.02% Triton X-100) for 15 min, PBST with 10% FBS for 30 min, and PBST with 10% FBS and primary antibodies at 4 °C for 16 h. After being washed with ice-cooled PBST twice, cells were incubated in PBST with 10% FBS and secondary antibodies for 2 h in the dark. This was followed by washing with ice-cooled PBST twice and nuclei staining with Hoechst 33342 for 5 min in the dark. After being washed with ice-cooled PBST twice, cells were mounted in Fluoromount-G (Electronic Microscope Sciences) and sealed with clear nail polish on microscope slides. The images were captured using a Leica SP8 confocal microscope. Imaging areas were randomly selected in a single-blind manner by a microscope specialist. Three areas were imaged in each condition, and a representative image is shown.

Metabolic labelling

Cells were plated onto 12-well plates and cultured in the cell growth medium. Cells were starved for 6 h with DMEM (Thermo Fisher A1443001) without glucose, glutamine and pyruvate. Next, cells were washed with 1× PBS and incubated with labelling medium containing 4 mM glutamine, 2× non-essential amino acids, 20% ES cell FBS, 1× penicillin–streptomycin and 4 mM [3-¹³C]pyruvate for 6 h. Finally, labelling medium was removed, and cells were washed with cold saline twice. Metabolites were quenched by the addition of 80% methanol/water (v/v), chilled on dry ice, and stored at −80 °C until processing.

For metabolite extraction, samples were thawed on ice and mixed well, followed by centrifugation at 4 °C for 10 min at 14,000 r.p.m. to pellet protein. Supernatant was removed into gas chromatography–mass spectrometry (GC–MS) vials. Pellets were extracted again with cold 80% methanol in water (v/v), and supernatants were combined and derivatized as described elsewhere⁶⁰. In brief, 10 µl of 1 N NaOH was added to methanol supernatants and vortexed. Next, 15 µl of NaB₂H₄ (10 mg ml⁻¹ in 50 mM NaOH) was added to reduce keto bonds and convert them into their respective deuterated hydroxyl groups. Next, samples were vortexed for 20 s and allowed to react at room temperature for 1 h. Reaction was stopped by the addition of 1 N HCl, and samples were evaporated to dryness. To remove boric acid, 50 µl of methanol was added, and samples were left to stand. After methanol was evaporated to dryness, samples were further derivatized by the addition of 60 µl of N-(tert-butyldimethylsilyl)-N-methyltrifluoroacetamide (MTBSTFA) to form the tri-tert-butyldimethylsilyl (t-BDMS), and the reactions were performed for 1 h at 60 °C. Derivatized samples were transferred to inserts and 1 µl of each sample was analysed by GC–MS.

GC–MS analyses were carried out on an Agilent 5973 mass spectrometer equipped with a 6890 Gas Chromatograph. A DB17-MS capillary column (30 m × 0.25 mm × 0.25 μm) was used in all assays with a helium flow of 1 ml min⁻¹. Oven temperature was set to 100 °C, held for 1 min, and ramped at 7.5 °C min⁻¹ until 260 °C, then 30 °C min⁻¹ until 300 °C, and held for 10 min. Injector temperature was set at 250 °C and detector temperature at 280 °C. Samples were analysed in selected ion monitoring mode using electron impact ionization. Ion dwell time was set to 10 ms. Metabolite abundances were collected, and natural abundance was normalized using matrix analyses. Fractional enrichment was determined by dividing the abundance of labelled species by the sum of all of the molecular species for each particular metabolite.

Differentiation and imaging of OPCs

To differentiate WT and eIF2Bε(R132H) OPCs, cells were plated and grown for 72 h in differentiation medium that consisted of DMEM/F12, 1× N2 supplement and 1× B-27 without vitamin A supplement, supplemented with 100 ng ml⁻¹ noggin (R&D Systems 3344-NG), 10 ng ml⁻¹ neurotrophin-3 (NT-3; R&D Systems 267-N3), 50 μM cAMP (Sigma D0260), 100 ng ml⁻¹ insulin-like growth factor 1 (R&D Systems 291-G1) NT-3 and 40 ng ml⁻¹ triiodothyronine (thyroid hormone; Sigma T-6397).

After 72 h of differentiation, cells were live-stained with mouse anti-O1 antibody, then fixed with 4% paraformaldehyde (PFA) and immunostained using rat anti-MBP antibody, secondary antibodies conjugated to an Alexa Fluor (4 μg ml⁻¹, Thermo Fisher) and the nuclear stain DAPI (Sigma, 1 μg ml⁻¹). Cells were then imaged using the Operetta High Content imaging system (PerkinElmer) and analysed using automated scripts in Columbus v2.8.1.141347 software (PerkinElmer). Both outlier tests and P value calculation were performed in GraphPad Prism v10.4.1. Outliers were statistically identified and removed using ROUT test with Q = 1%, and P values were calculated using a Welch’s unpaired t-test.

Quantification and statistical analysis

For quantification of band intensities from western blot, ImageJ software was used (https://imagej.net/ij/). For statistical analysis, unless noted otherwise, data shown in this study are presented as the mean of triplicate determinations ± s.e.m. Unless otherwise indicated in the figure legends, statistical significance between groups was evaluated using Student’s t-test. In the figure legends (n = X) indicates the number of independent experiments.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.