Transcription arrest induces formation of RNA granules in mitochondria

Cell culture

U2-OS cells (ATCC) were cultured in McCoy’s 5A media (#30-2007; ATCC) supplemented with 10% FBS. HeLa cells were cultured in DMEM supplemented with 10% FBS. For cell passaging, cells were rinsed with 1x PBS and treated with Gibco TrypLE Express Enzyme (12605010; Gibco) for 15 min (U2-OS cells) or 5 min (HeLa cells). Cells were seeded onto plates with fresh media.

Drug treatments of cells

Transcription was inhibited with either 2 μg/ml ethidium bromide, 10 μM IMT1B, or 100 μM 2′-CMA. The drugs were diluted to the respective concentration in medium and added to the cells for the indicated time points. As controls, 1% water was used for the ethidium bromide treatment and 0.1% DMSO for IMT1B and 2′-CMA. 100 μM antimycin A in EtOH or 100 μM bortezomib in DMSO was used in Fig 1D. As controls, cells were treated with 2% EtOH or 2% DMSO. Treatment time was 5 h. For treatments of cells with 100 nM antimycin A or 200 μg/ml chloramphenicol was performed for 24 h, 0.04% EtOH served as a control (Fig S1C).

siRNA treatment against SUV3

siGENOME Human SUPV3L1 (SUV3) siRNA was ordered in SMARTpool format from Horizon. For siRNA treatment, cells were grown until 70% confluency. Afterward, they were transfected using the Lipofectamine RNAiMAX kit (#13778075; Invitrogen). Therefore, 9 μl Lipofectamine was premixed with 350 μl Opti-MEM (#31985062; Gibco). 1.5 μl of 20 μM siRNA pools (siSUV3 and untargeted as a control) was mixed with 350 μl Opti-MEM. Both mixtures were combined and incubated for 5 min. In the meantime, cells were supplemented with fresh media. 250 μl of the mixed siRNAs was added to the respective wells. After 24 h of siRNA treatment, the media were replaced with fresh media lacking siRNA. After an additional 48 h, cells were harvested. Some cells were seeded into glass chambers to grow overnight and could be used for SABER-FISH. A fraction of cells were frozen down to be used for Western blot analysis.

Probe design

RNA-FISH probes were designed according to Beliveau et al (2018). Because the mitochondrial genes overlap with pseudogenes, the probe candidates were aligned against the mitochondrial DNA alone. If the whole genome was used, no specific probes could be selected, because of pseudogenes in the nuclear DNA. No K-mer filtering was used. Where necessary, probe sequences were converted to the reverse complement sequence.

After OligoMiner filtered suitable oligo sequences, they were manually extended by primer sequences to be able to do nucleotide exchange reactions with hairpins to synthesize the final oligos. All oligo sequences are listed in Table S1. Hairpin sequences were taken from Kishi et al (2019). For bicistronic transcripts (MT-ATP6/8, MT-ND4/L), probe pools were designed that detect the full bi-cistronic transcripts. All sequences used for SABER-FISH can be found in Table S1.

Probe synthesis

Probe synthesis was performed as published previously (Kishi et al, 2019). A reaction mix without oligos was prepared on ice and incubated for 15 min at 37°C to remove any free guanines in the mix. One reaction mix contained 0.5 μM (HP27 and HP28)–1 μM hairpins (all other hairpins), 1x PBS buffer, 300 nM dNTPs without guanines, 100 nM Clean.G oligo, 10 mM MgSO₄, and 8 U of Bst LF polymerase (#M1213-200; BioVision) or 100 U Bst LF polymerase (#BPL-200; McLab). After the 15-min preincubation, 1 μM of oligo pool mix was added. The reaction was incubated for 80 min at 37°C followed by 20 min at 80°C to inactivate the polymerase and cooled down to 4°C. Next, the elongation was tested on 1.3% agarose gels. If an elongation of at least 100 bp was observed, probes were purified using the MinElute PCR kit from Qiagen. The ssDNA probes were eluted using 30 μl of RNase-free water, and their concentration was measured by NanoDrop.

Coating of slides

150 μl of poly-L-lysine was added to each well and incubated for 10 min at RT. After three rinses with 1x PBS, slides were left to dry for 1 h and could be used for cell seeding.

SABER-FISH

The SABER-FISH protocol was adapted from Kishi et al (2019). Cells were grown in poly-L-lysine–coated ibidi glass-bottom well chambers (#80807; ibidi) until 50–70% confluency. For drug treatments, cells were grown overnight in normal media and at the next day treated as described in the drug treatment section. After the treatment, cells were rinsed 2x with 1x PBS and fixed with 4% freshly prepared formaldehyde (Pierce 16% Formaldehyde [wt/vol], methanol free, #28906; Sigma-Aldrich) for 1 h at RT. Fixation time for Figs 1B and D, S1A–C, and S6A and B has been 20 min. After fixation, cells were rinsed 3x with 1x PBS. Cells would be either stored at 4°C overnight to be processed the next day or directly subjected to permeabilization. In case of SABER-FISH data used for quantification, cells were subjected to WGA staining right before permeabilization. Therefore, cells were treated with 5 μg/ml WGA-CF405S (#29027-1; Biotium) for 30 min in the dark at RT. The cells were washed three times with 1xPBS and could be permeabilized afterward. To this end, cells were incubated in 1x PBS and 0.1% TX-100 for 20 min at RT. Afterward, cells were rinsed 2x with 1x PBS, followed by a 5-min wash in 1x PBS at RT. A wash for 5 min in 1x PBS and 0.1% Tween-20 was followed. Cells were equilibrated for 15 min at 43°C in prewarmed Whyb LD buffer (2xSSC, 40% formamide, 0.1% Tween-20). During this incubation, the hybridization mix was prepared. 1 μg of each probe was mixed with the 1.4x hybridization buffer (2.8x SSC, 56% formamide, 0.14% Tween-20, 14% dextran). The ratio of hybridization buffer to water including probes was 85–35 μl per well. The mix was prewarmed to 43°C, and the Whyb was replaced with 120 μl of the hybridization mix. The cells were incubated at 43°C overnight. The next day, cells were first quickly rinsed with prewarmed Whyb and washed two times for 30 min at 43°C with Whyb buffer. Two washes at 43°C for 5 min with 2x SSC and 0.1% Tween-20 were followed. Cells were rinsed for 1 min with 1x PBS buffer at RT. They were kept in fresh 1x PBS buffer until the hybridization oven cooled down to 37°C. In the meantime, the imager probes were mixed for fluorescence detection. From here, all incubations were done in the dark to avoid bleaching of the fluorophores. The imager mix contained 1 μM of each imager probe and 1x PBS. Cells were incubated either for 30 or 60 min with imagers at 37°C. The cells were washed with prewarmed 1x PBS once for 5 min and two times for 2 min. Except for WGA-stained samples, DAPI staining was followed. Therefore, cells were rinsed once with 1x PBS at RT and incubated for 10 min in DAPI (5 μg/ml in 1xPBS and 20 U/μl of SUPER RNasin) at RT, and washed three times for 5 min with 1x PBS. 150 μl mounting media (80% glycerol, 1xPBS, 20 mM Tris, pH 8, 2.5 mg/ml n-propyl gallate, 20 U/μl SUPER RNasin) were added, and the cells were imaged either the same or the next day. For Figs 1A and D and S1A–C, the Hyb buffer and Whyb contained 1% Tween-20.

Immunofluorescence

For immunofluorescence after SABER-FISH treatments, mounting media were removed by two quick rinses with PBST buffer (1xPBS, 0.1% TX-100). Cells were washed three to four times for 5 min with PBST followed by three washes of 5 min with displacement buffer (50% formamide, 1xPBS) to remove fluorescent probes. Next, cells were washed three times for 2 min with PBST and rinsed three times with 1xPBS. Cells were blocked for 1 h at room temperature with blocking buffer (5% goat serum in 1xPBS and 0.1% TX-100). Cells were incubated overnight at 4°C with the primary antibodies. Unbound antibodies were removed by three washes for 5 min with PBST at RT. Cells were incubated with a secondary antibody harboring the fluorophores diluted in blocking buffer for 1 h at RT and washed again three times for 5 min with PBST at RT. After staining nuclei by DAPI, cells were covered with mounting media and were stored at 4°C until imaging.

For immunofluorescence directly after drug treatments, cells were fixed and permeabilized as described in the SABER-FISH section. Afterward, cells were incubated with blocking buffer and treated as outlined above.

Antibodies

Listed are antibodies used in this study with information of dilution and where they are purchased from. All antibodies were diluted in blocking buffer, if not stated otherwise: TOM20 (#sc-17764; Santa Cruz) used in a 1:200 to 1:500 dilution, GRSF1 (#ab205531; Abcam) used in a 1:500 to 1:1,000 dilution, FASTKD2 (#17464-1-AP; ProteinTech) used in a 1:500 dilution, ds/ssDNA (#61014; PROGEN) used in a 1:100 dilution, LRPPRC/GP130 (#ab97505; Abcam) used in a 1:200 dilution, SUV3L (#ab127909; Abcam) used in a 1:1,000 dilution in milk for Western blot, ACTB (#3700; Cell Signaling) used in a 1:5,000 dilution in milk for Western blots, HRP-conjugated anti-rabbit IgG (cat. no. 7074S; Cell Signaling) used in a 1:10,000 dilution in 5% milk, HRP-conjugated anti-mouse IgG (cat. no. 70765; Cell Signaling) used in a 1:10,000 dilution in 5% milk, goat anti-rabbit IgG Alexa Fluor 647 (cat. no. A-21245; Thermo Fisher Scientific) used in a 1:1,000 dilution, goat anti-mouse IgG2a Alexa Fluor 488 (cat. no. A-21131; Thermo Fisher Scientific) used in a 1:1,000 dilution, and goat anti-mouse IgM Alexa Fluor 555 (cat. no. A-21426; Thermo Fisher Scientific) used in a 1:1,000 dilution.

Western blotting

Cells were lysed in RIPA buffer and treated with Benzonase for 30 min and afterward mixed with loading buffer (95 mM Tris–HCl, pH 6.8, 7.5% glycerol, 2% SDS, 0.5 mg/ml bromophenol blue, 50 mM DTT). Proteins were separated on 10% NuPAGE Novex Bis-Tris gels (Thermo Fisher Scientific) and transferred to a nitrocellulose membrane. Membranes were incubated with primary antibodies overnight (SUV3L and ACTB). For visualization, membranes were incubated with secondary antibodies conjugated with horseradish peroxidase for chemiluminescence detection.

Imaging

Samples were imaged on a Nikon Ti inverted microscope with a W1 Yokogawa Spinning disk with 50-μm pinhole disk, a Nikon motorized stage, a Physik Instrument Piezo Z motor, an Andor Zyla 4.2-Plus sCMOS monochrome camera VSC-06522 or VSC-10150 (used for quantification images of two replicates of IMT1B treatment and MT-ND5 of EtBr treatment), and OKO Lab Heated enclosure with CO₂ control at room temperature. For imaging, Nikon Elements Acquisition Software AR 5.02 was used. All files were saved in .ND2 format. Before each imaging, slides were cleaned first with Sparkle and water to remove salts and finally with 70% EtOH to remove any oils. Samples for quantifications were imaged with a Plan Apo λ 60x/1.4 Oil DIC objective in combination with immersion oil (Cargille, nondrying immersion oil Type 37, #16237) and with a laser device of Digital Mirror Device lines (Aura/LED) 350, 380, 488, and 561 with emission filter 455/50, 480/40, 525/36, 605/52, 630/75, and 705/72 or with laser lines 405, 488, 561, 640 (Nikon Instruments, Laser Unit model Lun-F) and emission filters (455/50, 480/40, 525/36, 605/52, 630/75, 705/72). For each image set, 10 fields of view of an empty ibidi glass chamber well were imaged with the same laser settings used for sample acquisition for flat-field correction. Z-stacks of 37 planes with a size of 200 nm were imaged, with no binning and dual-gain ¼ setting. For all images, z-stacks were imaged in the following order: first, all laser lines were imaged before the next z-plane. The final pixel size was 0.108 μm × 0.108 μm. Exposure times (usually between 500 and 700 ms) and laser power varied between experiments because of different probe sets used in SABER-FISH. If not stated differently, images of the residual experiment were taken with no binning and dual-gain ¼ settings, and using a silicon oil objective with correction collar (Plan Apo λS SR HP 100xC/1.45 Silicon DIC) in combination with silicon immersion oil 30cc (#MXA22179; Nikon). For collar correction of the silicon oil objective, fluorescent beads were prepared as described in Isaac et al (2024). Z-stacks with varying plain number-to-image full volumes of cells and sizes of 200 nm or 300 nm were taken. The final pixel size was 0.065 μm × 0.065 μm. RNase- and DNase-treated samples were imaged with a Plan Fluor 40x Oil DIC H/N2 objective in combination with immersion oil, 12-bit, and no binning settings. The final pixel size was 332.8 μm × 332.8 μm.

Image analysis

Image manipulations

To represent images in figures, brightness and contrast were adjusted for the whole fields of view in Fiji to represent the full spatial distribution of RNAs. This was necessary because of high differences in intensities between control and sample and in an image itself. Thus, images do not represent differences in intensities. Images were then cropped for single cells and zoom-ins using regions of interest. For figures showing WGA staining, a mean filter size 1 was used for the WGA channel. All other channels were only adjusted for brightness and contrast. TEM images were just cropped to show regions of interest of different cells.

TEM

Cells were treated for 24 h with IMT1B, DMSO or for 5 h with EtBr. Cells were fixed in 2.5% glutaraldehyde, 1.25% PFA, and 0.03% picric acid in 0.1 M sodium cacodylate buffer, pH 7.4, for 1 h. The cells were then postfixed for 30 min in 1% osmium tetroxide (OsO4)/1.5% potassium ferrocyanide (KFeCN6), washed in water 3x, and incubated in 1% aqueous uranyl acetate for 30 min followed by 2 washes in water and subsequent dehydration in grades of alcohol (5 min each; 50%, 70%, 95%, 2 × 100%). Cells were removed from the dish in propylene oxide, pelleted at 7,000g for 3 min, and infiltrated for 2 h ON in a 1:1 mixture of propylene oxide and TAAB Epon (TAAB Laboratories Equipment Ltd, https://taab.co.uk). The samples were subsequently embedded in TAAB Epon and polymerized at 60°C for 48 h.

Ultrathin sections (about 80 nm) were cut on a Reichert Ultracut-S microtome, picked up onto copper grids stained with lead citrate, and examined in a JEOL 1200EX transmission electron microscope or a Tecnai G2 Spirit BioTWIN, and images were recorded with an AMT 2k CCD camera.

Quantifying mitochondrial phenotypes of TEM data

Mitochondria of 10 images for EtBr and IMT1B samples and 11 images of DMSO were quantified by hand by two people in a double-blind fashion, and the data were analyzed in R.

qRT-PCR to measure RNA abundance after EtBr treatment

HeLa S3 cells were grown until 65–70% confluency, washed with prewarmed 1xPBS, and treated for 5 h with 0, 0.5, 1, or 2 μg/ml ethidium bromide for 5 h. Cells were scraped, spun down, and finally lysed in 750 μl Quizol and frozen for later use. RNA was isolated as described below. The RNA still containing DNA was resuspended in 30 μl water. To remove DNA, DNA-free DNA Removal Kit from Thermo Fisher Scientific (cat. no. AM1906) was used. In brief, samples were treated with DNase I for 30 min at 37°C and the DNase was inactivated via the addition of the DNase inactivation reagent. Samples were centrifuged at 10,000g for 90 s, and the supernatant was transferred to a new tube. The SuperScript III First-Strand Synthesis System from Thermo Fisher Scientific (cat. no. 18080051) was used to synthesize cDNA with the addition of random hexamers according to the instructions. cDNA was then used for qRT-PCR. 1 ng/μl cDNA was used per reaction. cDNA was mixed with the SsoFast EvaGreen Supermix from Bio-Rad (cat. no. 1725201) and 0.4 μM primers (MT-CO1 For: 5′-CTATGATATCAATTGGCTTCCTAG-3′, MT-CO1 Rev: 5′-TGGTAGCGGAGGTGAAATATGCTCGTGTGT-3′, MT-ND1 For: 5′-TGCCATCATGACCCTTGGCCAT-3′, MT-ND1 Rev: 5′-GCCTGAGACTAGTTCGGACTCCCCTTCGGC-3′, NDUFA7 For: 5′-CATGACCTGCAGGGGAAGCT-3′, NDUF7A Rev: 5′-CCACAGGGAGCTTGGGAGGA-3′, GAPDH For: 5′-GTCTTCACCACCATGGAGAAGG-3′, and GAPDH Rev: 5′-ATGATCTTGAGGCTGTTGTCAT-3′). GAPDH primer sequences were taken by Nagao et al (2008). The qRT-PCR was started in Bio-Rad CFX384 Touch Real-Time PCR Detection System. The following program was used: 30 s at 95°C followed by 40 cycles of 95°C for 5 s, 55°C for 10 s, and a melting curve starting at 65°C until 95°C in 0.5°C for 5 s.

Transient transcription analysis

To measure transient transcription during recovery, cells were first treated for 5 h with IMT1B as described above. The cells were washed three times with 1xPBS and grown in normal media for recovery. To label nascent RNAs, 4-thiouridine (4sU, T4509; Sigma-Aldrich) was directly added to the cells to a final concentration of 500 μM at the last 10 min of each time point. Samples were harvested after 0, 1, 2, 3, 4, and 5 h of recovery. Therefore, the media were removed and cells placed on ice. They were scraped directly into Quizol, and the lysates were transferred into tubes. 2 mM DTT was added to each tube, and the samples were heated for 3 min at 65°C and stored at −80°C for RNA extraction, followed by biotinylation and enrichment of labeled RNAs to be analyzed by MitoStrings.

4sU labeling to distinguish between old and new RNAs during recovery

To measure RNA abundance of old RNAs after stress release, cells were first treated with EtBr for 5 h as described above. Afterward, cells were washed three times with 1x PBS. Cells were then incubated in media supplemented with 500 μM 4sU for 2, 3, 4, or 5 h. For time point 0 h, cells were harvested directly after EtBr treatment. Harvesting of cells was performed as described in the transient transcription section. The lysates could be used for RNA extraction and biotinylation. The labeled RNAs were bound onto streptavidin beads, and the flow-through fraction was used for analysis by MitoStrings.

RNA extraction

RNAs of long-term treated cells and for RNA-seq were extracted as follows. After the respective treatment times, plates were put on ice. Cells were washed with ice-cold 1xPBS and scraped into Quizol. Lysates were stored at −80°C until RNA extraction. Lysates were thawed at RT. 200 μl chloroform was added to 1 ml of sample, mixed by vortexing, and incubated for 2 min at RT. Lysates were centrifuged for 15 min at 4°C and at 12,000g. The aqueous phase was transferred into a fresh tube without disturbing the interphase. RNAs were precipitated with 500 μl isopropanol; in case of qRT-PCR samples, 1 μl Invitrogen GlycoBlue from Invitrogen (AM9515) was added, and incubated either at RT for 10 min or at −80°C overnight. After a second spin at 4°C and 12,000g for 10 min, the supernatant was discarded. The pellet was washed with 75% EtOH and pelleted again by centrifugation at 4°C and 7,500g for 5 min. The supernatant was discarded and the pellet dried for 5–10 min at RT. The RNAs were resuspended in 32 μl of RNase-free water, and the concentration was measured by Qubit and NanoDrop. 40 ng of RNAs was used in MitoStrings measurements. For samples used to study the fate of old RNAs after stress release, 1 ng/μl unlabeled in vitro synthesized ERCC-00048 spike-in was added to the thawed cell lysates. In case of transient transcription analysis, 0.25 ng/μl of in vitro synthesized 4sU labeled ERCC-00136 spike-in (as described in McShane et al [2024]) was added additionally. RNAs were extracted as described previously, and in these cases, RNA was resuspended in 50 μl of RNase-free water.

RNA biotinylation and purification

A maximum of 40 μg of RNA per sample was used for the biotinylation reaction. Samples were brought to a volume of 70 μl and incubated at 60°C with 800 rpm in a thermoblock with rotation function for 10 min. Afterward, samples were immediately incubated for not longer than 2 min on ice. RNAs were biotinylated in a reaction containing 1x labeling buffer (20 mM Hepes, pH 7.4, 1 mM EDTA) and 0.5 mg/ml biotin-MTS (#90066; Biotium) in 20% dimethylformamide (Sigma-Aldrich). The reaction mixture was incubated in a thermoblock with rotation function at 800 rpm and 24°C for 30 min in the dark. Next, free biotin was removed. To this end, 2-ml phase-lock heavy gel tubes (#2302830; 5Prime) were prepared by centrifugation at 14,000g for 1 min. The biotinylated RNA was diluted with water, transferred into the phase-lock tube, and mixed with 1 volume of chloroform:isoamyl alcohol (24:1) by manually shaking for 15 s. A centrifugation step for 5 min at RT and 16,000g was followed. The upper phase was transferred into a fresh tube, and the RNA was precipitated by the addition of 0.1 volume of 5 M NaCl and 1 volume of isopropanol. The mixture was centrifuged for 40 min at 4°C and 20,000g. The supernatant was removed and the pellet washed with 75% EtOH, followed by a 5-min spin at 4°C and 7,500g. The supernatant was removed, and the pellet was dried for 5 min at RT. The RNA was resuspended in 40 μl of RNase-free water. Samples could be stored at −80°C for enrichment. The biotinylated RNA was purified with streptavidin beads of the uMACS Streptavidin kit (#130-133-282; Miltenyi Biotec). 3 μg RNA was mixed with streptavidin beads, incubated for 15 min at RT, and subsequently loaded onto a uMACS column. The flow-through was loaded again onto the columns to increase biotin binding. The flow-through contained the unlabeled RNA fraction. The beads were washed three times with prewarmed (65°C) washing buffer (100 mM Tris, pH 7.5, 10 mM EDTA, 1M NaCl, and 0.1% Tween-20) and three times with washing buffer at RT. RNAs were eluted twice by incubation at 65°C in elution buffer (550 mM Tris, pH 9, 10 mM EDTA, 150 mM TCEP). For transient transcription analysis, the elution fraction, and for the analysis of old RNA, the flow-through fraction were purified using the miRNeasy Micro kit following the manufacturer’s instructions including the DNase treatment (#217084; QIAGEN). 30 ng RNA was used for MitoStrings analysis.

MitoStrings

MitoStrings is a mitochondrial-tailored NanoStrings measurement (Wolf & Mootha, 2014). According to the manufacturer’s protocol, RNAs (30–40 ng) were hybridized for 16 h at 67°C with the XT Tagset-24 (NanoString Technologies) and with DNA probes specific for mitochondrial RNA (MitoStrings probes were modified as in Isaac et al [2024]) in hybridization buffer (NanoString Technologies). Afterward, samples were loaded onto a nCounter Sprint cartridge and quantified by the use of the nCounter SPRINT Profiler (NanoString Technologies) of the Bauer core facility of Harvard or at the Manning Lab (Harvard T.H. Chan School of Public Health).

Total RNA sequencing

Expression analysis

To analyze expression changes of OXPHOS genes, read counts were normalized to length- and library size-normalized read counts. Read counts were summed over protein-coding regions with the help of the R package featureCounts (Liao et al, 2019) with the parameters: isGTFAnnotationFile = TRUE, useMetaFeatures = TRUE, countMultiMappingReads = FALSE, strandSpecific = 1, isPairedEnd = TRUE, nthreads = 4. GTF files used in this study are described here (Soto et al, 2022). Table S6 contains the RNA-seq data.

Half-life calculations

To calculate half-lives, RNA abundance was measured in long-term transcription inhibition treatments by MitoStrings. RNA levels were measured after 0, 2, 5, 8, and 24 h of transcription inhibition. The mean of three biological replicates (EtBr-treated cells) or two biological replicates (IMT1B-treated cells) of GAPDH-normalized RNA counts was used for calculating half-lives. 0-, 2-, and 5-h measurements were used in linear regression with log transformation, and time point 0 h was set as 100%. The slope was used to calculate the half-lives: (log₁₀(0.5)–log₁₀(1))/slope. The same procedure was performed for late half-lives. Here, the 5-h measurements were set as 100%. For NDUFA7, we used for early half-lives only 0 and 5 h as the 2-h time point increased and led to negative half-lives. All NanoStrings data collected can be found in Table S7.

Correlation analysis

Pearson’s correlation analyses were performed using the smplot2 package in R with the following command: sm_statCorr (r2 = TRUE, color = “black,” size = 0.8).

Use of AI

ChatGPT was used to help formulate R code for half-life analysis, to extract statistics, and to write functions for image analysis. Claude was used to edit and improve the text.