METHOD DETAILS

For ER stress, cells were treated with 1 μM thapsigargin (Tg; Sigma, T9033; 1 mM stock in DMSO was used) for 1.5 hrs. To confirm SG-specific effects, 2 μg/ml of cycloheximide (CHX; Sigma, C7698), 10 μg/ml of emetine (EMT; Sigma, E2375) or 10 μg/ml of puromycin (PUR; Sigma, P8833) was co-treated with thapsigargin. DMSO (1:1000)-treated cells were used as a control for the ER experiment. For heat-shock (HS) stress, water (1:1000)-treated cells were incubated at 43 °C/5% CO₂ for 45 min. For arsenite (AS) stress, cells were treated with 0.5 mM NaAsO₂ (Sigma, S7400; 0.5 M stock in water was used) for 1 hr. For both HS and AS experiments, water (1:1000)-treated cells incubated at 37 °C/5% CO₂ were used as a control.

After appropriate treatments, cells were placed on ice and washed twice with 10 ml ice-cold PBS. This washing step was done quickly, within 10 seconds. After washing, the cells were harvested in 1.5 ml of ice-cold buffer L (50 mM Tris pH 7.6, 50 mM NaCl, 5 mM MgCl₂, 0.1% NP-40, 1 mM β-mercaptoethanol, 1x EDTA-free protease inhibitor cocktail (Roche, 05 892 791 001), 0.4 U/μL RNase inhibitor (Thermo Scientific, AM2696)) on ice. All subsequent steps were conducted on ice or at 4 °C. Cell suspensions were given 30 strokes (1 mL, in 30 seconds) in a Dounce homogenizer, followed by centrifugation at 2,000 g for 2 min. The supernatant (Cyt, cytosolic fraction) was collected without disrupting nuclei pellet (Nuc). 10 μl of Cyt was taken for protein quantitation by Bradford assay. 125 μl of Cyt was mixed with 375 μl of Trizol LS (Thermo Fisher, 10296010) for Cyt RNA extraction (Cyt-Trizol). 1 ml of Cyt was centrifuged at 10,000 g for 10 min to separate the soluble fraction (Sol) from the insoluble pellet fraction (RG). The RG was resuspended in 500 μl of Trizol (Thermo Fisher, 15596018) for RG RNA extraction (RG-Trizol).

Then, the same amount of spike-in RNA control, determined proportionally to the Cyt protein concentration (measured from the 10 μl aliquot as described above), was added to Cyt-Trizol and RG-Trizol. For example, when Cyt protein concentration was measured as 1 mg/ml, 10 μl of 1x spike-in control RNA mix (see below about how it was prepared) was used for RNA-seq experiments, and 0.5 μg of Drosophila total RNA was used for quantitative RT-PCR experiments. After Trizol extraction, performed according to the manufacturer’s instructions (Thermo Fisher, 15596018), RNA was precipitated in isopropanol and used for subsequent applications.

Performing the initial cell washing step at 37°C or on ice did not significantly alter basal RG content of ER cluster 1 transcripts, including Xiap, Ago3, Creb1 and Norad, as well as a ER cluster 4 transcript BiP (Figure S2I). Replacing 5 mM MgCl₂ in buffer L with 30 mM EDTA did not significantly change the stress-induced RG targeting of specific transcripts, such as Xiap, Ago3, Creb1, Rictor, Brca1 and Norad (Figure S2H, lower panel). Non-targeted transcripts, such as Actb, Gapdh, BiP and Rpl7, did not show stress-induced RG targeting in both MgCl₂ and EDTA conditions (Figure S2H, lower panel).

The pRFP-G3BP vector was a kind gift from Drs. B.L. Wolozin (Boston University) and N. Kedersha (Harvard) (Liu-Yesucevitz et al., 2014). HEK293 cells transfected with pRFP-G3BP were subjected to RG preparation as described above in the “Preparation of RNA samples for RG RNA-Seq” section. Isolated RG was resuspended in mounting media (H-1000, Vector laboratories) and spread on microscope slides. For whole cell imaging, cells were grown on coverslips, briefly fixed with 4% formaldehyde (Polysciences Inc., 18814), stained with DAPI, and mounted in Vectashield anti-fade mounting media (H-1000, Vector laboratories). Both whole cell and RG images were obtained under a Leica SP5X confocal microscope.

Drosophila total RNA, used as a spike-in RNA control for quantitative RT-PCR experiments, was prepared from w¹¹¹⁸ adult flies using the Trizol method (Thermo Fisher, 15596018). Spike-in control RNA mix, used for RNA-seq experiments, is a mixture of three in vitro transcribed Drosophila genes (dRpS7 30 pg/μl, dGapdh1 3 pg/μl, dAct5C 0.3 pg/μl). To prepare the spike-in control RNA mix, Drosophila cDNA was prepared from Drosophila total RNA using MMLV-reverse transcriptase (Thermo Fisher, 28025013) and random hexamers (Thermo Fisher, N8080127). For in vitro transcription, ~500 bp of dRpS7, dGapdh1 and dAct5C cDNA sequences, attached to a T7 minimal promoter sequence (5’-TAATACGACTCACTATAGGGAGA-3’), were amplified from the Drosophila cDNA using the primers listed in Table S5.

In vitro transcription reactions were conducted using the amplified DNA and the T7 MAXIscript kit (Thermo Scientific, AM1312). The transcription product was treated with DNase I (Thermo Fisher, 18068015) then subjected to phenol-chloroform extraction and isopropanol precipitation. After quantification, RNA products were diluted in DEPC-treated water and mixed to constitute the spike-in control RNA mix.

1 μg of DNase I-treated total RNA purified from Cyt and RG was subjected to rRNA depletion by Ribo-Zero rRNA removal Kit (Illumina, MRZH1116). rRNA-depleted RNA was precipitated in 100% ethanol and reconstituted in DEPC-treated water. The RNA-seq library was constructed using the NEXTflex™ Rapid Directional qRNA-seq™ Kit (Bioo scientific, 5130–01D), according to the manufacturer’s instructions. RNA-seq was performed at the University of Michigan DNA Sequencing Core Facility with Illumina HiSeq2000. For analyses of sequencing results, see the QUANTIFICATION AND STATISTICAL ANALYSIS section.

0.5 μg of DNase I-treated total RNA purified from Cyt and RG was reverse transcribed using MMLV reverse transcriptase (Thermo Fisher, 28025013) and random hexamers (Thermo Fisher, N8080127). The cDNA products were diluted 10-fold with nuclease-free water, and 1 μl of the diluted cDNA was used for real-time quantitative PCR. iQ SYBR Green Supermix (Bio-rad, 1708884). List of primers used in quantitative RT-PCR is provided in Table S5. Quantitative PCR was performed using the StepOnePlus Real Time PCR system (Applied Biosystems). The averages of dRpL32 and dGapdh1 measurements were used as a spike-in control to calculate relative gene expression.

The ribosome footprint strategy was designed according to a previous study (Ingolia et al., 2012). The method was partially modified for the generation of a matching RNA-seq library. In brief, NIH3T3 cells were cultured in DMEM with 10% (v/v) fetal bovine serum (FBS) in a humidified 5% CO₂ atmosphere at 37°C. 7 × 10⁵ cells were seeded in a 10-cm dish, followed by DMSO or thapsigargin (1 μM for 1.5 hr) treatment after 48 h. Cultured cells were washed twice with ice-cold 1x PBS and lysed in lysis buffer (20 mM Tris-Cl (pH7.4), 150 mM NaCl, 5 mM MgCl₂, 1 mM DTT, 100 μg/ml cycloheximide, 1% (v/v) Triton X-100, 25 U/ml Turbo DNase I) by triturating the cells ten times through a 26-gauge needle. After centrifuging the cell lysate at 20,000g in 4°C for 10 min, the supernatant was divided into two groups for generating the ribosome profiling library and mRNA expression profiling library.

For the ribosome profiling library, lysates were treated with RNase I (100 U/μl) and incubated at room temperature for 45 min. After nuclease digestion, ribosome footprinted RNAs were purified by sedimentation through a 1 M sucrose cushion and excising the urea gel between 26–34 nucleotides. The 3’ end was dephosphorylated by adding PNK and ligated by adding 40 pmol of an RA3 adaptor. The 5’ end was phosphorylated by adding PNK and ATP and ligated by adding 30 pmol of an RA5 adaptor and 50 pmol of RTP primer. After reverse transcription using Superscript II reverse transcriptase (Invitrogen), rRNA products were depleted by hybridization to biotinylated sense-strand oligonucleotides, followed by removal of the duplexes through streptavidin bead affinity. Finally, through PCR and gel extraction, the library was generated using the size range 150–160 bp.

For the mRNA-seq library, poly(A) RNA was isolated using Dynabeads® mRNA purification kit (Ambion), according to the manufacturer’s protocol. After RNA fragmentation with 1 N NaOH at 37°C for 3 min, the final library ranging from 150 to 200 bp was generated, followed by the same process for producing the ribosome profiling library from the step of 3’ dephosphorylation, excluding rRNA depletion.

Translation efficiency (TE) was defined as the amount of ribosomal footprint normalized to underlying mRNA abundance and expressed in a log₁₀ scale. For TE calculations, transcripts with raw read counts less than 50 in any of the sequencing reads were removed from the analysis. The TE values were normalized to have the same median values across replicates of the same experiment. ΔTE-ER was defined as the ER stress-induced change in TE, expressed in a log₂ scale.

Whole cell lysates were prepared in cell lysis buffer (Tris pH 7.5, 150 mM NaCl, 1% NP-40, 1 mM NaF, 0.1 mM NaVO₃, 1x EDTA-free protease inhibitor cocktail (Roche, 05 892 791 001)). Nuc, Cyt, RG and Sol fractions were prepared through the same method as described above in the “Preparation of RNA samples for RG RNA-Seq” section. Nuclear fraction (Nuc) and the RG were reconstituted in 8 M Urea buffer (8 M urea, 50 mM Tris pH 8.0, 300 mM NaCl, 0.5% NP-40, 50 mM Na₂HPO₄). After normalizing protein concentration by the Bradford assay, protein samples were boiled in SDS sample buffer for 5 min, resolved by SDS-PAGE, transferred to PVDF membrane filters, probed with primary antibodies (1:200 for Santa Cruz antibodies, and 1:1000 for all other antibodies), then probed with horseradish peroxidase-coupled secondary antibodies (1:5000; anti-goat (sc-2020) antibody was from Santa Cruz Biotechnology, and anti-rabbit (1706515) and anti-mouse (1706516) antibodies were from Bio-rad). Chemiluminescence images were acquired using LAS4000 (GE) system.

The pGFP-MS2 vector was a kind gift from Dr. J. L. Goodier (Johns Hopkins) (Goodier et al., 2010). The pSL-MS2(6x) vector was obtained from Addgene (27118, originated from Dr. Robert Singer). The MS2(6x) binding sequence was subcloned into a lentiviral expression vector (pLU-CMV) to create a pLU-CMV-MS2(6x) vector. Mouse cDNA was prepared from NIH3T3 total RNA using MMLV-reverse transcriptase (Thermo Fisher, 28025013) and random hexamers (Thermo Fisher, N8080127). 3’-UTR sequences of Brca1, Ago3 and Xiap were amplified from the mouse cDNA using the following primers; Brca1 3’UTR F: 5’-TTTAAGAAATGGTCTTCCATG-3’, Brca1 3’UTR R: 5’-GTCAAAATGTTTGCTATTTAGTTTAT-3’, Ago3 3’UTR F: 5’-CAAGTCCAAGTTTATTCTCCGTG-3’, Ago3 3’UTR R: 5’-CAGGGGGTTGGCAAAGTC-3’, Xiap 3’UTR F: 5’-TGGGGCACCACATGTTAT-3’, Xiap 3’UTR R: 5’-AACATTTTAAAGAATAGTGCTTTATTG-3’. HEK293 cells were transiently co-transfected with both pGFP-MS2 and pLU-CMV-3’UTR-MS2(6x) using polyethylenimine (Polysciences Inc., 23966).

NIH3T3 cells or GFP-MS2-transfected HEK293 cells were grown on coverslips for 48 hrs. After DMSO or Tg treatment, cells were fixed with 4% formaldehyde (Polysciences Inc., 18814) and permeabilized with 0.2% Triton X-100. Cells were incubated with 3% bovine serum albumin (Sigma, A8806) in PBS then with primary antibodies (1:1000 for P-body markers and 1:200 for all others) in PBS for 2 hrs. After washing, cells were incubated with Alexa fluor-conjugated secondary antibodies (1:1000; anti-goat (A11055, A11058), anti-mouse (A21202) and anti-rabbit (A21207) antibodies from Thermo Fisher) for 1 hr, washed with PBS, counterstained with DAPI, and mounted in Vectashield anti-fade mounting media (H-1000, Vector laboratories).

For detection of single molecule RNA in cells, the Stellaris® RNA fluorescence in situ hybridization (FISH) method was used. In brief, each mRNA in cells was stained with 30 different Quasar-570 labelled FISH probes. The FISH probes used for in situ detection of Brca1, Ago3, Xiap, Norad and Actb genes are listed in Table S5. For probing the Gapdh gene, the ShipReady probe set (SMF-3002–1) was used. All these probes were designed and produced by LGC Bioresearch Technologies, Inc. (Petaluma, CA).

The cells were processed as described above in the “Immunostaining” section. After the secondary antibody incubation and the PBS washing step, cells were fixed again with 4% formaldehyde (Polysciences Inc., 18814) and incubated with wash buffer A (10% formamide, 2x SSC) for 5 min. Cells were then incubated with the probes (0.25 μM for Actb, 1 μM for the others) in Hybridization buffer (1% dextran sulfate, 10% formamide, 2x SSC) at 37 °C for 12 hrs, incubated with wash buffer A, containing DAPI for 30 min, washed twice with PBS, and mounted in an anti-fade mounting buffer (100 mM Tris pH 8.0, 0.4% glucose, 2x SSC, 90% glycerol, 0.29 mg/ml catalase (Sigma, C3515), 0.04 mg/ml glucose oxidase (Sigma, G2133)). Images were captured under a Leica SP5X confocal microscope. For colocalization analysis between RNA and protein, Manders’ coefficients using threshold values were calculated from the images as described previously (Bolte and Cordelieres, 2006) (n=3).