Chronic stress antagonizes formation of stress granules

Summary

Stress granules (SGs) are cytoplasmic ribonucleoprotein condensates formed in response to stress-induced inhibition of mRNA translation and polysome disassembly. Despite the broad interest in SG assembly and disassembly in response to acute stress, SG dynamics under chronic stress has not been extensively investigated. We show that cells pre-conditioned with low-dose chronic (24 h exposure) stresses of various natures fail to assemble SGs in response to acute stress. While protein synthesis is drastically decreased by acute stress in pre-conditioned cells, polysome profiling analysis reveals the partial preservation of polysomes. Mechanistically, chronic stress slows down the rate of mRNA translation at the elongation phase, and triggers phosphorylation of translation elongation factor eEF2. These events further promote ribosome stalling, which is distinct from ribosome collisions known to trigger ribosome-associated quality-control pathways. In summary, chronic stress triggers ribosome stalling, which prevents efficient polysome disassembly and SG formation by subsequent acute stress.

Graphical abstract

Highlights

• •Chronic stress inhibits formation of stress granules by subsequent acute stress
• •Chronic stress triggers phosphorylation of translation elongation factor eEF2
• •Chronic stress slows down the rate of mRNA translation at the elongation phase
• •Chronic stress induces ribosome stalling and blocks polysome disassembly

Biochemistry; Cell biology

Introduction

Chronic stress is the cellular response triggered by a long-term exposure to intracellular or extracellular stressors, which vary in their nature (e.g., biotic versus abiotic). Eukaryotic cells have evolved universal stress-response programs that allow them to survive under toxic environmental conditions.^1,2 Stress granule (SG) assembly, an early response to cellular stress, is triggered by stress-induced global translation inhibition and polysome disassembly.^3,4 Canonical SGs are membrane-less cytoplasmic RNA granules consisting of 40S ribosomal subunit-containing mRNPs, translation initiation factors, mRNAs, RNA-binding proteins, and signaling molecules. SG composition varies in a stress-specific manner and determines whether cells live or die.^5,6 For instance, while sodium arsenite (SA), commonly used to mimic oxidative stress, induces the assembly of canonical SGs that promote cell survival, nitric oxide induces the assembly of non-canonical SGs linked to decreased cell viability.⁷ SGs also contribute to the aggregation of disease-related proteins in the neurons of patients with neurodegenerative diseases such as amyotrophic lateral sclerosis.⁸ Both acute and chronic stress can contribute to disease pathogenesis.⁹ We and others have previously reported the effects of chronic stress on SG formation: chronic bisphenol A exposure suppresses SG formation, chronic proteasomal inhibition impairs SG formation, and chronic glucose starvation induces pro-death SGs.^10,11,12 However, the molecular mechanisms by which acute and chronic stress differentially modulate SG assembly and cell survival is poorly understood.

SA triggers phosphorylation of the α-subunit of eukaryotic initiation factor 2 (eIF2α) to inhibit translation initiation,⁶ which is a common initiator of the cellular stress response. eIF2 is a part of the ternary complex that delivers the initiator tRNA (tRNA_i^Met) to the 40S ribosomal subunit. Together with other initiation factors, the ternary complex becomes the 43S mRNA pre-initiation complex, which scans the 5′-untranslated region of an mRNA. When an optimal AUG codon is found, the 60S ribosomal subunit joins the 48S complex to form an 80S ribosome capable of translation elongation.¹³ Phosphorylation of eIF2α at serine 51 prevents efficient GDP/GTP exchange on eIF2 and ultimately inhibits rates of translation initiation. Because phosphorylated eIF2α inhibits mRNA translation initiation, this quickly causes polysome disassembly. In turn, the disassembly of polysomes promotes SG formation. Like SA, many other stress inducers (e.g., endoplasmic reticulum [ER] stress or starvation stress) activate eIF2α kinases, induce translation inhibition, and promote SG formation.¹⁴

eIF2α is the key molecule of the integrated stress response (ISR), a stress-induced pathway that regulates stress adaptation and contributes to the cell-survival and -death decisions. Upon eIF2α phosphorylation, global mRNA translation is decreased and production of house-keeping proteins is minimal. At the same time the translation of some pro-survival, anti-apoptotic transcripts, such as activating transcription factor 4 (ATF4) mRNA is paradoxically increased. The transcription factor ATF4 in turn activates expression of cytoprotective genes that help cells coping with stress. The ISR acts as an integrator pathway that senses difference in the nature of stresses and converges on eIF2α phosphorylation. Previous works suggested that pre-conditioning (the phenomenon where exposure to low levels of stress that does not induce any severe cell damage could later protect against subsequent exposure to more severe stress) through stresses that cause eIF2α phosphorylation activate pro-survival downstream signaling pathways and confer cytoprotection.^15,16 The mechanism of the cytoprotection likely relies on two processes: down-regulation of global translation to conserve cellular energy and production of cytoprotective molecules (such as ATF4) during pre-conditioning. Whether heterologous pre-conditioning (pre-conditioning through various unrelated stresses) via ISR also engages SGs in the cytoprotection is not well understood. However, Shelkovnikova et al. showed that stress-preconditioned or chronically stressed neurons failed to maintain SG assembly in response to proteasomal inhibition,¹¹ suggesting that SGs can be a part of the ISR that can be affected by pre-conditioning or chronic stress exposure.

SGs are dynamic entities in equilibrium with polysomes. Puromycin, a translation elongation inhibitor that causes premature termination of polypeptide synthesis, also inhibits translation and promotes SG formation.¹⁷ Puromycin forces 80S ribosomes to dissociate into 40S and 60S ribosomal subunits, effectively stripping ribosomes off of the mRNA. In contrast, translation elongation inhibitors such as anisomycin (ANS), cycloheximide, and emetine “freeze” ribosomes on the mRNA, preventing untranslated transcripts from assembling SGs.^17,18 While the above drugs are known to regulate SG dynamics, how impairment of the translation elongation under physiological (unrelated to pharmacological treatment) situations affects SG formation is largely understudied.

The elongation step of mRNA translation is also well regulated. The primary target is eukaryotic elongation factor 2 (eEF2), a GTP-dependent translocase responsible for the movement of nascent peptidyl-tRNAs from the A-site to the P-site of the ribosome. Its activity is regulated by phosphorylation on threonine 56 (T56) by eEF2 kinase (eEF2K), causing downregulation of its activity.^19,20 In turn, eEF2K itself is a phosphorylation target for diverse pathways to impact elongation, ranging from signaling from nutrient deprivation to changes in cell cycle. In addition, the rate of ribosome elongation can also be affected by mRNA secondary structures, insufficient supply of aminoacyl-tRNAs, or inefficient ribosome termination/recycling, which can cause ribosomes to stall and collide.^21,22,23

Ribosome stalling and collisions can also be induced by oxidative stress, starvation, ultraviolet irradiation, or stress-induced mRNA damage.^24,25,26,27 Collided ribosomes are translationally incompetent and quickly recognized and degraded by the ribosome-associated quality-control (RQC) pathways. Extensive ribosome collision induces the ribotoxic stress response (RSR) which is initiated by phosphorylation of ZAKα kinase to activate a pro-death signaling cascade.^27,28,29,30

SGs are in dynamic equilibrium with polysomes, and this equilibrium can be affected by acute and chronic stresses. While SGs contribute to cell survival or cell death, a connection between chronic stress, SG formation, and ribosome stalling under physiological conditions has not been previously investigated well. Here, we report that various chronic stressors (oxidative stress, ER stress, mitochondrial stress, and nutrient starvation) render cells incompetent for SG assembly in response to acute stress. The data on polysome profiling and the phosphorylation of eEF2 implicate incomplete polysome disassembly in this process. Specifically, chronic stress slows down mRNA translation elongation and promotes ribosome stalling, and thus, antagonizes efficient SG formation.

Results

Chronic stress conditioning inhibits SG formation

To understand the effects of chronic stress on SG formation, we incubated human osteosarcoma U2OS cells with various doses (10, 50, 100, and 500 μM) of SA, for different times (1, 4, 12, and 24 h) prior to quantifying SG assembly (Figure 1A). Prolonged SA treatment promotes cell death in a time-dependent and dose-dependent manner (Figure S1). We also pre-incubated cells with low dose of various stressors (including 10 μM SA, which did not induce SG assembly) for 24 h prior to treating with 100 μM SA (Figure 1B). In all cases, chronic stress pre-conditioning inhibited acute SA-induced SG assembly (Figures S2 and 1C–F). SA induces SG formation via phosphorylation of eIF2α, but some other SG inducers (e.g., rocaglamide A [Roc A]) are independent of eIF2α phosphorylation. To test whether chronic stress can also inhibit formation of SGs by eIF2α phosphorylation-independent stressors, we treated SA-pretreated U2OS cells with RocA. Like SA, RocA induced fewer SGs in chronic stress pre-conditioned cells (Figure S3), suggesting that chronic stress can inhibit SG formation independently of their trigger mechanism (eIF2α phosphorylation-dependent versus -independent). Typically, acute SA stress promotes SG formation as early as 30 min (Figure S4). We determined whether pre-treatment with low-dose SA simply shifts SG formation toward earlier time points (less than 30 min). We could not, however, detect SG formation at the earlier time points during chronic stress (Figure S4). Chronic stress further promoted the degree of cell death in a dose-dependent manner (Figure S5). Importantly, chronic stress antagonizes SG formation also in human embryonic kidney HEK293 cell line (Figure S6). These data suggest that chronic stress pre-conditioning interferes with acute stress-induced SG assembly irrespective of the mechanism by which translation initiation, eIF2α phosphorylation -dependent or -independent, is inhibited.

Figure 1.

Chronic stress pre-incubated cells fail to respond to form SGs by acute stress

(A) U2OS cells were subjected to treatment with SA (10, 50, 100, and 500 μM) at each time (0, 1, 4, 12, and 24 h). Cells were examined for the presence of the core SG markers G3BP1 (green), eIF4G (red), and eIF3b (blue). Scale bars represent 10 μm. All positive cells were quantified. p values were assessed using a one-way ANOVA (vs. NT; p∗∗∗∗ <0.0001). Results are mean ± S.E.M. (n = 3).

(B) Schematic illustration of the experimental timeline.

(C–F) U2OS cells were subjected to treatment with 100 μM SA for 1 h after pre-incubation with 10 μM SA, 30 μM Mena (menadione), 2 μM Drb (doxorubicin), 1 μM Tg (thapsigargin), 25 μg/mL Tun (tunicamycin), 25 μg/mL BA (brefeldin A), 60 μM CCCP (carbonyl cyanide m-chlorophenyl hydrazone), 60 mM 2DG (2-deoxy-D-glucose), or HBSS for 24 h. Unstressed cells (NT) were used as a control.

(C) Cells were examined for the presence of the core SG markers G3BP1, eIF4G, and eIF3B.

(D) Representative images of U2OS cells stained with G3BP1 (green), eIF4G (red), and eIF3B (blue) after the cells had been subjected to specific stresses. Scale bars represent 20 μm.

(E) Cells were examined for the presence of the core SG marker G3BP1 and poly (A) mRNAs (FISH using oligo[dT] probe).

(F) Representative images of U2OS cells stained with G3BP1 (green) and oligo(dT) (red) after the cells had been subjected to specific stresses. Scale bars represent 20 μm.

(C and E) p values were assessed using a one-way ANOVA (vs. 1 h SA; p∗∗∗∗ <0.0001). Results are mean ± S.E.M. (n = 3). All experiments were done for each of the three biological replicates (independent experiments).

Incomplete polysome disassembly in chronic stress-pre-conditioned cells

Since SG formation is mediated by global translation inhibition and SA induces phospho-eIF2α-dependent translational repression,³¹ we hypothesized that SA may not inhibit translation in chronic stress pre-conditioned cells. We compared several different types of stress (SA as an oxidative stress, thapsigargin as an ER stress, carbonyl cyanide m-chlorophenyl hydrazone as mitochondrial stress, and Hanks’ balanced salt solution [HBSS] as starvation stress) in these experiments. It is evident that an 1 h incubation with 100 μM SA triggers phosphorylation of eIF2α and inhibits protein synthesis (Figure 2A). Moreover, the expression levels of SG core proteins G3BP1, G3BP2, caprin1, and USP10 are not affected under these conditions (Figure 2A). m⁷GTP pull-downs reveal the cap-binding eIF4F complex to be intact under these conditions suggesting that mTOR signaling is not involved in the suppression of SG formation (Figure S7). Translation inhibition by SA rapidly disassembles polysomes (Figures 2B–2F, 1h SA), which triggers efficient SG assembly.³² In contrast, 1 h incubation of 100 μM SA did not completely disassemble polysomes in chronic stress-pre-conditioned cells (Figures 2B–2F).

Figure 2.

Chronic stress-pre-incubated cells fail to disassemble polysomes by acute stress

U2OS cells were subjected to treatment with 100 μM SA for 1 h after pre-incubation with 10 μM SA, 1 μM Tg (thapsigargin), 60 μM CCCP (carbonyl cyanide m-chlorophenyl hydrazone), or HBSS for 24 h. Unstressed cells (NT) were used as a control.

(A) Cells were pulsed with puromycin and emetine for 5 min and lysed. Cell lysates were subjected to western blotting using antibodies for puromycin, p-eIF2α, total eIF2α, G3BP1, G3BP2, caprin1, USP10, and β-actin. Three blots were taken for each of the three biological replicates (independent experiments). The same colors of the circle dot in the graph are from the same independent experiment. Results are mean ± S.E.M. (n = 3). p values were assessed using a one-way ANOVA (p∗ <0.05, p∗∗ <0.01, p∗∗∗ <0.001, p∗∗∗∗ <0.0001).

(B–F) Polysome profiles from U2OS cells. NT, black; 1 h SA, red; (B) 24 h pre-incubation of SA, blue; 1 h SA with 24 h pre-incubation of SA, green.

(C) 24 h pre-incubation of Tg, blue; 1 h SA with 24 h pre-incubation of Tg, green.

(D) 24 h pre-incubation of CCCP, blue; 1 h SA with 24 h pre-incubation of CCCP; green.

(E) 24 h pre-incubation of HBSS, blue; 1 h SA with 24 h pre-incubation of HBSS; green.

(F) The polysome/monosome ratio was calculated with normalization by each of NT as a control.

Because SG formation also relies on mRNA condensation besides proteins, low amounts or limited availability of mRNAs in the cells as a result of chronic stress conditions could reduce the efficiency of SG formation. Indeed, SGs are enriched in specific subsets of RNAs such as long mRNAs (e.g., AHNAK and DYNC1H1) or long non-cording RNAs such as NORAD.^33,34 We thus hypothesized chronic stress decreases a pool of mRNAs, and, in turn, this decrease negatively affects SGs formation. We pulled down mRNAs from total RNAs, and quantified and compared the amounts of mRNAs under different conditions. mRNA pull-down efficiency was also checked by RT-qPCR, and pulled-down samples analyzed for specific mRNAs, previously shown to be enriched in SGs (Figure S8). The amounts of pulled-down mRNA between unstressed cells (NT) and chronic SA (10 μM) were not significantly different (Figure 3A). Additionally, the relative levels of GAPDH, AHNAK, and DYNC1H1 mRNAs and NORAD RNA, which are examples of efficiently enriched RNAs in SGs,³¹ were not significantly different (Figure 3B), thus rejecting the hypothesis that chronic stress limits pool of available RNAs for SG assembly. Moreover, a much higher dose of SA (500 μM) induces SG assembly in chronic SA (10 μM)-pre-conditioned cells, consistent with efficient polysome disassembly, but interestingly, p-eIF2α level was not different between 100 and 500 μM of SA (Figures 3C–3E). These results imply that chronic stress pre-conditioned cells retain the ability to form SGs but disassemble polysomes less efficiently, which causes less efficient SG formation in pre-conditioned cells.

Figure 3.

Chronic stress pre-incubated cells can form SGs by excessive stress promoting polysome disassembly

(A and B) U2OS cells were incubated with 10 μM of SA for 24 h (24 SA), and total RNA was extracted.

(A) mRNA was pulled down from total RNA and both RNAs were measured. Results are mean ± S.E.M. (n = 3).

(B) The mRNA expression levels of GAPDH, AHNAK, DYNC1H1, and NORAD were determined by RT-qPCR (scandalized by BACTIN). Results are mean ± S.E.M. (n = 3).

(C and D) U2OS cells were subjected to treatment with 500 μM SA for 1 h after pre-incubation with 10 μM SA for 24 h.

(C) Cells were examined for the presence of the core SG markers G3BP1 (green), eIF4G (red), and eIF3b (blue). Scale bars represent 20 μm. All positive cells were quantified. Results are mean ± S.E.M. (n = 3).

(D) U2OS cells were subjected to treatment with 100 or 500 μM SA for 1 h after pre-incubation with 10 μM SA for 24 h. Cells were pulsed with puromycin and emetine for 5 min and lysed. Cell lysates were subjected to western blotting using antibodies for puromycin, p-eIF2α, total eIF2α, and β-actin.

(E) Polysome profiles from U2OS cells. NT, black; 1 h of 500 μM SA, red; 24 h 10 μM SA pre-incubation, blue; 1 h of 500 μM SA with 24 h 10 μM SA pre-incubation, green.

(A–C) All experiments were done for each of the three biological replicates (independent experiments).

Puromycin does not promote polysome disassembly and SG formation in chronic stress-pre-conditioned cells

Puromycin promotes SG assembly by inducing premature elongation termination and ribosomal subunit dissociation, while other translation elongation inhibitors, such as ANS, inhibit SG assembly by freezing 80S ribosomes on mRNA^17,18,35,36 (Figure 4A). Consistent with these mechanisms, puromycin promotes SG assembly by an intermediate dose of SA (50 μM) in control cells (Figure 4B). However, low-dose (10 μM) SA pre-conditioning inhibits puromycin-induced SG assembly (Figure 4B). Polysome profiling of 50 μM SA-treated cells show the expected partial disassembly of polysomes (with ∼40% of cells being SG-positive) and complete disassembly of polysomes during puromycin co-incubation (Figure 4C). But in 10 μM SA pre-conditioned cells, puromycin does not disassemble polysomes (Figures 4C and S9). It is known that translation elongation inhibitors such as cycloheximide freeze 80S ribosomes and inhibit SG assembly in the continued presence of stress (Figure 4A). Similar results were observed in cells treated with low-dose (1 μg/mL) ANS; SA did not induce SG formation or polysome disassembly in ANS-pre-incubated cells (Figures 4D–4E). Our data suggest that various chronic stresses inhibit translation but do not completely disassemble polysomes (Figure 2). Considering the data in both Figures 2 and 4, chronic stress may “freeze” or slow down some of the 80S ribosomes on mRNA, which inhibits ribosome dissociation and SG formation. Taken together, we hypothesized that chronic stress slows ribosome elongation.

Figure 4.

Puromycin treatment is not sufficient to promote SG formation in chronic stress pre-incubated cells

(A) Schematic illustration of the effects of puromycin and other translation inhibitors on SG formation. (1) Normal condition; 80S ribosomes are on mRNA and no SG. (2) Stressed condition; phosphorylated eIF2α (p-eIF2α) blockes translation initiation, which also promotes 80S ribosomes "run-off" mRNA and induces SG formation. (3) Puromycin makes ribosomes split on mRNA by causing premature polypeptide release during translation elongation, which releases mRNA from polysomes and induces SG formation. (4) Other translation elongation inhibitors "freeze" the translating ribosomes on mRNA, which inhibits polysome disassembly and SG formation. (5) Stress-induced eIF2α phosphorylation inhibits the initiation of translation, but translation elongation inhibitors (except puromycin) shown as four "frozen" translating ribosomes on mRNA inhibits polysome disassemblys, which induce no SG formation.

(B and C) U2OS cells were subjected to treatment with 50 μM SA for 1 h after pre-incubation with pre-incubation of 10 μM SA for 24 h and 20 μg/mL puromycin (puro) for the last 0.5 h.

(B) Cells were examined for the presence of the core SG markers G3BP1 (green), eIF4G (red), and eIF3b (blue). Scale bars represent 20 μm. All positive cells were quantified. Results are mean ± S.E.M. (n = 3). p values were assessed using a one-way ANOVA (p∗∗∗∗ <0.0001, N.S.: not significant).

(C) Polysome profiles from U2OS cells. NT, black; 1 h of 50 μM SA, red; 1 h of 50 μM SA + 0.5 h of puromycin, orange; 24 h 10 μM SA pre-incubation, blue; 1 h of 50 μM SA with 24 h SA pre-incubation, green; 1 h of 50 μM SA + 0.5 h of puromycin with 24 h SA pre-incubation, purple.

(D and E) U2OS cells were subjected to treatment with 100 μM SA for 1 h after pre-incubation with 1 μg/mL ANS (anisomycin) for 15 min.

(D) Cells were examined for the presence of the core SG markers G3BP1 (green), eIF4G (red), and eIF3b (blue). Scale bars represent 20 μm. All positive cells were quantified. Results are mean ± S.E.M. (n = 3).

(E) Polysome profiles from U2OS cells. NT, black; 1 h SA, red; 15 min ANS, pink; 1 h SA with 15 min ANS pre-incubation, dark blue. (B and D) All experiments were done for each of the three biological replicates (independent experiments).

Chronic stress induces ribosomal translocation failure and translational delay

To determine whether chronic stress slows down ribosome elongation, we treated pre-conditioned cells with harringtonine, which is an inhibitor of the initial step of translation elongation.^37,38 Polysome profiling showed disassembled polysomes by harringtonine incubation under normal control conditions, but harringtonine did not disassemble polysomes as efficiently (within the same time frame) in 10 μM SA chronic-stressed cells (Figure 5A). This indicates that while chronically stressed cells are still translating in a manner sufficient for cell survival, the rates at which ribosomes perform mRNA translation are likely down-regulated. Such elongation slow-down may also cause ribosome stalling and collisions.

Figure 5.

Chronic stress slows down translation at the elongation step

(A) Polysome profiles from U2OS cells. NT, black; 2 μM harringtonine incubation for 3 min, red; 5 min, blue; 8 min, green. For these treatments cell without (left) or with (right) 24 h of SA 10 μM pre-incubation were used. Two biological replicates (individual experiments) were done. The polysome/monosome ratio was calculated and shown on the right. The same colors of the circle dots in the graph are from the same independent experiment.

(B) Polysome profiles from RNase A (0.5 mg/mL)-digested lysates of U2OS cells. NT, black; 1 μg/mL ANS, pink; 10 μM SA incubation for 24 h, blue; 100 μM SA treatment for 1 h with 10 μM SA pre-incubation for 24 h, green. The polysome/monosome ratio was calculated and shown on the right.

(C) U2OS cells were subjected to treatment with 100 μM SA for 1 h after pre-incubation with 10 μM SA, or 1 μg/mL ANS for 15 min. Cell lysates were subjected to western blotting using antibodies for p-eEF2, total eEF2, and β-actin.

(D) U2OS cells were subjected to treatment with 10 μM SA, or 12.5 μM NFV (nelfinavir) for 24 h. Cell lysates were subjected to western blotting using antibodies for p-eEF2, total eEF2, and α-tubulin.

(E) U2OS cells were subjected to treatment with 100 μM SA for 1 h after pre-incubation with 12.5 μM NFV for 24 h. Cells were examined for the presence of the core SG markers G3BP1 (green), eIF3b (red), and DAPI (blue). Scale bars represent 20 μm. G3BP1- and eIF3b-positive cells were quantified (right graph). p values were assessed using a one-way ANOVA (p∗∗∗∗ <0.0001). Results are mean ± S.E.M. (n = 3).

(F) U2OS cells were subjected to treatment with 10 μM SA for 1, 4, 12, and 24 h or 1 μg/mL ANS for 15 min. Cell lysates were subjected to western blotting using antibodies for p-eEF2, total eEF2, and β-actin. (G) U2OS cells were subjected to treatment with 100 μM SA for 1 h after pre-incubation with 10 μM SA for 0, 1, 4, 12, and 24 h. Cells were examined for the presence of the core SG markers G3BP1 (green), eIF4G (red), and eIF3b (blue). Scale bars represent 20 μm. All positive cells were quantified. p values were assessed using a one-way ANOVA (vs. 1 h pre-incubation of SA +1 h SA; p∗∗∗∗ <0.0001). Results are mean ± S.E.M. (n = 3).

(E and G) All experiments were done for each of the three biological replicates (independent experiments).

While moderate ribosome collisions are efficiently rescued by the RQC response program, an excessive ribosome collision activates the stress-sensing MAP3 kinase ZAKα, which binds ribosomes and acts as one of the arms of RSR.^26,27,28,29 At low doses (1 μg/mL), ANS is known to induce ribosome collision, whereas higher doses "freeze" translation elongation without inducing collisions.²⁶ We used polysome profiling followed by RNase A treatment to quantify the ribosome fraction involved in stalling or collisions (manifested by disomes) (Figure 5B). Chronic stress pre-conditioning (24 h SA + NT and 24 h SA + 1 h SA) produces polysome peaks similar to those observed in cells treated with 1 μg/mL ANS for 15 min. We confirmed that the observed peaks are reminiscent of disomes of collided ribosomes (as in control ANS-treated sample), which are resolved to monosomes following exposure to a higher dose of RNase (Figure S10A).

To determine whether stalled ribosomes actually represent collided ribosomes, we used phos-tag SDS-PAGE to quantify ZAKα phosphorylation in chronic SA-treated cells (Figures S10B and S10C). Our data suggest that 24 h low-dose SA treatment only partially triggers ZAKα phosphorylation. In addition, we determined whether chronic SA treatment will promote ubiquitination of RPS10, a ribosomal protein that is specifically modified during excessive ribosome collisions such as promoted by ANS treatment^39,40 (Figure S10D). Our data suggest that stalled ribosomes either mostly do not represent bona fide collided ribosomes or only constitute their small non-dominant fraction, which are direct targets of RSR. Thus, chronic stress induces ribosome stalling that is different from collisions observed during excessive ribotoxic damage.

We further analyzed the eEF2 phosphorylation status. eEF2 regulates translation elongation by GTP-dependent translocation of peptidyl-tRNA, but phospho-eEF2 inhibits this step and mRNA translation, leading to ribosome stalling.²⁴ Indeed, low dose SA pre-conditioning induces phosphorylation of eEF2 while ANS does not (Figure 5C). As nelfinavir (NFV) is known to promote phosphorylation of eEF2,⁴¹ we incubated cells with NFV chronically for 24 h and checked cell ability to induce SGs by additional SA incubation. We observed that NFV induced the same level of phosphorylation of eEF2 as 24 h 10 μM SA, and significantly inhibited SG formation by additional acute stress (Figures 5D and 5E).

Next, we examined the profiles of phosphorylation of ZAKα and eEF2 in a time-dependent manner. While phosphorylated ZAKα was only partially observed at both 12 and 24 h (with phosphorylation stronger at 24 h than 12 h) (Figure S10C), eEF2 was only significantly phosphorylated at 24 h (Figure 5F). These results indicate the phospho-eEF2-induced translocation failure may slowly accumulate in form of stalled ribosomes which may lead to severe ribosome collision at much later stages, but it is not the main trigger of collisions itself. We also determined the ability of cells to form SGs (by an addition of 100 μM SA) at each time point. The percentage of SG-positive cells decreased in 12 h and 24 h pre-incubated cells, coinciding with levels of ZAKα phosphorylation (Figure 5G). It is important to note that various types of chronic stress induced eEF2 phosphorylation (Figure S10E). We also tested how cells react to repetitive stress where prolonged acute stress (4 h of 100 μM SA) was followed up with another acute stress (1 h of 100 μM SA or 500 μM SA) (Figure S11). Interestingly, SGs were disassembled after 4 h of 100 μM SA, with concurrent partial restoration of protein synthesis (Figures S11C and S11D), but following acute stress still could not promote efficient SG formation. Likewise, repetitive SA stresses (up to three times 100 μM SA with two 3 h washout in-between) also inhibited SG formation (Figure S12).

These data suggest that chronic stress pre-conditioning slows down mRNA translation and promotes ribosome stalling, which consequently delays polysome disassembly and SG formation in response to additional acute stress (Figure 6).

Figure 6.

Schematic illustration of SG formation failure by mRNA translation slow-down in chronic stress pre-incubated cells

Acute SA stress induces mRNA translation initiation inhibition thereby promoting polysome disassembly and release of mRNPs, which contribute to SG formation. Chronic stress slows down the rate of mRNA translation elongation causing ribosome stalling and ineffective release of mRNPs from the polysomes. Despite translation inhibition by additional acute stress, chronically stressed cells fail to promote effective SG formation.

Discussion

SG assembly and disassembly in response to acute stress has been extensively studied.⁴² SG dynamics following chronic stress has received less attention, despite SGs playing an important role in the pathophysiology of neurodegenerative disease, cancer, inflammatory diseases, and other chronic stress-related conditions.^8,14,43,44 While we showed that any regime of SA-induced chronic stress itself did not induce SG formation (Figure 1A), it has been reported that chronic incubation (30 h) of low-dose (15 μM) SA induces SG formation in induced pluripotent stem cell-derived motoneurons,⁴⁵ suggesting cell type-specific differences in response to stresses. Interestingly, aging is considered as a condition of complex chronic stress that induces cellular senescence, a stable cell-cycle arrest linked to various diseases.^46,47,48 Since it is known that cellular senescence promotes ribosome collision and translational repression,^28,49,50,51 our data suggest that senescent cells may also have a reduced propensity to form SGs, a phenotype that may be intrinsically linked to chronic stress-induced changes in gene expression.

Our studies indicate that most of chronic stress conditions inhibit both initiation (judged by eIF2α phosphorylation) and elongation (eEF2 phosphorylation) phases of protein synthesis while acute stress tends to target the initiation step. We think that slow-down of mRNA translation by eEF2 phosphorylation under conditions of chronic stress aims on the maintenance of cellular metabolism mostly by energy conservation, and by making some adjustments in gene expression. This is in contrast to extensive stress (e.g., causing ribosome collision and activation of RSR) or transient acute stress (causing abrupt global translation arrest coupled with sequestration of mRNPs into SGs), which aim on cell survival and make drastic changes in gene expression.

Mechanistically, G3BP1/2 (G3BP) proteins are SG nucleators that are required for SG formation under most of stresses, except osmotic. As G3BP needs to bind 40S ribosomal subunits to nucleate SGs, 80S ribosome dissociation and polysome disassembly both contribute to SG formation.³¹ Although our results suggest that ribosome stalling and inhibition of translation elongation inhibits SG disassembly in chronic stress pre-conditioned cells, distinct polysome profiles are observed in cells exposed to different types of stress (Figures 2B–2F, blue and green), whereas the percentage of SG-positive cells is almost the same in each case (Figures 1C–1F). One potential explanation is that the presence of a single 80S ribosome on mRNA is sufficient to prevent its recruitment into SGs, as suggested by.⁵² Puromycin is known to promote SG formation via polysome disassembly¹⁷ (Figure S5), but chronic stress pre-conditioned cells show only partial disassembly of polysomes in response to puromycin (Figure 4B). It should, however, be noted that high doses of SA are able to promote SG formation suggesting that SG components such as pool of mRNAs (Figure S9) and SGs nucleators (Figure 2A) are present in sufficient amounts to trigger SG formation per se. In agreement with this notion, we could not detect any significant differences in poly(A) mRNA levels or proteins playing roles in SG dynamics.

Our analysis of the effects of chronic stress on polysome formation suggests that stalled ribosomes contribute to changes in mRNA translation and SG assembly. We also showed that chronic stress-induced eEF2 phosphorylation may promote ribosome stalling via translocation failure, but does not trigger extensive collisions, at least at the earlier time points of stress exposure (Figures 5C, S10B and S10C). The exact contribution of chronic stress-induced ribosome collisions in the observed slow-down of translation is still unclear but there are several possibilities: an insufficient supply of aminoacyl-tRNAs, translation of damaged mRNAs, damaged ribosomes, etc.^21,22,23 Slowing down translation also can be beneficial for correct protein folding and oppose proteostasis, a hallmark of chronically stressed cells. Ribosome stalling and eEF2 phosphorylation are not increased abruptly but increased in a time-dependent manner (Figures 5C and S10), which provides RQC enough time to adjust and be efficient under chronic stress. Loss-of-function experiments of RQC proteins, which are observed under excessive stress, suggest that RQC failure is a trigger of cell death, which can contribute to some diseases.^53,54,55 Our data also indicate that the inability to promote SGs may also contribute to cell death.

In conclusion, as summarized in Figure 6, we propose that chronic stress-induced eEF2 phosphorylation slows down mRNA translation, causing ribosome stalling, which in turn inhibits SG formation. Inefficient disassociation of stalled 80S ribosomal subunits may directly contribute to this phenomenon. While canonical SGs are proposed to have a cytoprotective role,⁵ chronic stress-induced failure of SG formation together with other processes such as RQC mechanisms may activate cell death. Ribosome stalling and collisions are known to induce proteostasis that ultimately contributes to human diseases via cell death pathways.^27,28,30 Our data have uncovered a new functional connection between ribosome stalling and SG formation under chronic stress and provide a foundation for understanding and exploring acute stress in the context of chronic stress models such as aging.

Limitations of the study

Certainly, our studies have a number of limitations. First, these studies are done in the laboratory conditions and its translation to in vivo settings will be the key direction in the future. In fact, studies centered around pre-conditioning (e.g., dietary restrictions such as protein deprivation) showed that such regime is beneficial, preserving organ function by protecting against renal and hepatic ischemic injury and resulting in reduced inflammation.¹⁶ Mechanistically, such beneficial pre-conditioning requires GCN2 kinase and eIF2α phosphorylation.¹⁶ Whether preconditioning with other stresses, such as used in our study, will be also beneficial in in vivo studies is an open question. Also, whether SG response to chronic stress will be aberrant in vivo is unclear. Second, we admit here that our definitions of chronic stress (e.g., by choosing 24 h time span and various concentrations of stressors based solely on dosage used in acute stress studies) are not ideal and can be less representative as actual in vivo chronic conditions. Third, we have not dissected all the molecular details on relative contributions of ISR versus translation elongation changes in the regulation of SG dynamics under chronic stress. Finally, the nature of stalled ribosomes during stress and whether such staling may progress to more obvious translational arrest scenarios such as ribosome collision and trigger of RQC under more prolonged chronic stress conditions require further investigations.

Resource availability

Lead contact

Further information and requests for reagents and resources can be directed to, and will be fulfilled by, the lead contact, Pavel Ivanov (pivanov@bwh.harvard.edu).

Materials availability

This study did not create any new unpublished materials.

Data and code availability

• •This article does not report any original code.
• •Any additional information required to reanalyze the data reported in this article is available from the lead contact upon request.
• •No additional resources are reported.

Acknowledgments

We thank members of Ivanov and Anderson labs for the valuable discussion. The figures are created with BioRender.com. This work was supported by funds from the National Institutes of Health grant R35 GM126901 (P.J.A.), National Institutes of Health grant R01 GM126150 and R01 GM146997 (P.I.), Japan Society for the Promotion of Science (JSPS) Grants-in-Aid for Scientific Research (PS KAKENHI) 22KJ2354 (Y.A.), 20K21761 and 21H03359 (M.M.), 19H04053 and 23H03329 (Y.T.). Y.A. was supported by the JSPS Overseas Challenge Program for Young Researchers.

Author contributions

Y.A., P.I., and P.J.A. designed the research; Y.A. and A.M.W. performed the research; M.M. and Y.T. analyzed data and provided helpful discussion; Y.A., A.M.W., P.I., and P.J.A. wrote the manuscript.

Declaration of interests

Authors declare no competing interests.

STAR★Methods

Key resources table

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
Rabbit Polyclonal α-tubulin antibody	Proteintech	11224-1-AP; RRID: AB_2210206
Mouse Monoclonal β-actin antibody	Proteintech	66009-1-Ig; RRID: AB_2687938
Rabbit Polyclonal Caprin1 antibody	Proteintech	15112-1-AP; RRID: AB_2070016
Goat Polyclonal eIF3b antibody	Santa Cruz Biotechnology Inc.	sc-16377; RRID: AB_671941
Rabbit Polyclonal eIF3b antibody	Invitrogen	PA5-117928; RRID: AB_2902534
Mouse Monoclonal eIF4A antibody	Santa Cruz Biotechnology Inc.	sc-377315; RRID: AB_2868449
Rabbit Polyclonal eIF4E antibody	Cell Signaling Technology	Cat#9742; RRID: AB_823488
Rabbit Polyclonal eIF4G antibody	Santa Cruz Biotechnology Inc.	sc-11373; RRID: AB_2095750
Mouse Monoclonal G3BP1 antibody (TT-Y)	Santa Cruz Biotechnology Inc.	sc-81940; RRID: AB_1123055
Rabbit Polyclonal G3BP2 antibody	Bethyl Laboratories Inc.	A302-040A; RRID: AB_1576545
Rabbit Polyclonal Phospho-eEF2 (Thr56) antibody	Cell Signaling Technology	Cat#2331; RRID: AB_10015204
Rabbit Polyclonal Phospho-eIF2α (Ser51) antibody	Cell Signaling Technology	Cat#9721; RRID: AB_330951
Mouse Monoclonal Puromycin antibody	Sigma-Aldrich	MABE343; RRID: AB_2566826
Rabbit Monoclonal RPS10 antibody	Abcam	Ab151550; RRID: AB_2714147
Rabbit Polyclonal Total eEF2 antibody	Cell Signaling Technology	Cat#2332; RRID: AB_10693546
Rabbit Polyclonal Total eIF2α antibody	Cell Signaling Technology	Cat#9722; RRID: AB_2230924
Rabbit Polyclonal USP10 antibody	Bethyl Laboratories Inc.	A300-900A; RRID: AB_625312
Rabbit Polyclonal ZAKα antibody	Bethyl Laboratories Inc.	A301-933A; RRID: AB_1547954
Chemicals, peptides, and recombinant proteins
Sodium arsenite	Sigma-Aldrich	S7400
Rocaglamide A	Med Chem Express	HY-19356
Menadione	Sigma-Aldrich	M5625
Doxorubicin	AdooQ Bioscience	A14403
Thapsigargin	Invitrogen	T7458
Tunicamycin	Sigma-Aldrich	T7765
Brefeldin A	Sigma-Aldrich	B7651
Carbonyl cyanide 3-chlorophenylhydrazone	Sigma-Aldrich	C2759
2-deoxy-D-glucose	Sigma-Aldrich	D8375
Nelfinavir	Med Chem Express	HY-15287A
Puromycin	Sigma-Aldrich	P4512
Anisomycin	Med Chem Express	HY-18982
Emetine	Sigma-Aldrich	E2375
Cycloheximide	Sigma-Aldrich	C7698
PureLink RNase A	Invitrogen	12091021
RNasin Ribonuclease Inhibitor	Promega	N2111
HEPES	Gibco	15630080
Superscript IV first-strand synthesis kit for RT-qPCR	Invitrogen	18091050
iQ™ SYBR® Green Supermix	Bio-Rad	1708880
Critical commercial assays
Dynabeads mRNA DIRECT Purification Kit	Invitrogen	61012
Immobilized g-aminophenyle-m7gtp	Jena Bioscience	NU-870-1
Cell Titer Glo Luminescent Cell Viability Assay	Promega	G7570
Experimental models: Cell lines
U2OS	ATCC	Cat# HTB-96
HEK293	ATCC	Cat# CRL-1573
Oligonucleotides
5′-cy3-Oligo-d(T)40	IDT, This paper	N/A
Primers for Figures 3B and S8, see Table S1	IDT, This paper	N/A
Software and algorithms
Adobe Illustrator	Adobe	https://www.adobe.com/products/illustrator.html
Biocomp gradient station	Biocomp	https://biocompinstruments.com
BioRender	BioRender	https://www.biorender.com/
Image J	National Institutes of Health	https://imagej.net/ij/
Prism 10	Graph Pad	www.graphpad.com
NIS-Elements imaging software	Nikon	https://www.microscope.healthcare.nikon.com/products/software/nis-elements

Experimental model and study participant details

Cell culture and drug treatment

Human osteosarcoma U2OS cells and human embryo HEK293 cells were purchased from ATCC and routinely screened for mycoplasma contamination. Both cells were maintained at 37°C in a 5.0% CO2 in DMEM (Corning) containing 20 mM HEPES (Gibco), 10% FBS (Sigma), 100 U/ml penicillin, and 100 μg/ml streptomycin. For any experiments, cells were grown to ∼70% confluency and then treated as indicated in figure legends: sodium arsenite (SA, Sigma), rocaglamide A (RocA, MedChemExpress), menadione (Mena, Sigma), doxorubicin (Drb, AdooQ Bioscience), thapsigargin (Tg, Invitrogen), tunicamycin (Tun, Sigma), brefeldin A (BA, Sigma), carbonyl cyanide 3-chlorophenylhydrazone (CCCP, Sigma), 2-deoxy-D-glucose (2DG, Sigma), nelfinavir (NFV, MedChemExpress). Puromycin (puro, Sigma) treatment was performed 30 min before collecting the coverslips or as described in the text. Anisomycin (ANS, MedChemExpress) was performed 15 min before harvest or 15 min before the incubation with SA as indicated.

Method details

Western blotting

Following drug treatment, cells were washed with phosphate-buffered saline (PBS). Protein samples were heated to 95°C for 10 min in Laemmli sample buffer in the presence of 100 mM dithiothreitol (DTT) and subjected to SDS–PAGE. Samples were loaded on a 4–20% Tris-Glycine gel (BioRad) and transferred to nitrocellulose membrane. Membranes were blocked with Tris-buffered saline with 0.1% Tween-20 (TBS-T) with 5% milk for 1 h at room temperature. Antibodies were diluted in 5% normal horse serum in PBS. Primary antibodies were incubated overnight at 4°C and secondary antibodies for 1 h at room temperature. Antibody detection was performed using SuperSignal West Pico Chemiluminescent Substrate (Thermo Scientific).

RiboPuromycylation assay

Ribopuromycylation assay was described in Panas et al.⁵⁵ In brief, cells were unstressed or stressed as indicated. 5 min before harvest, puromycin and emetine were added to a final concentration of 9 and 91 μM, respectively, and the incubation continued for 5 min. Cells were then lysed and subjected to Western blotting using an anti-puromycin antibody.

M7GTP pull-down assay

Cells were lysed by lysis buffer (Tris–HCl pH7.4, 100 mM NaCl, 0.5% NP-40, protease inhibitor (Thermo Scientific)), and centrifuged for 15 min at 12,000×g at 4°C. The supernatant was mixed with Immobilized g-aminophenyle-m7gtp (Jena Bioscience) for 2 h at 4°C with rotation. The beads were washed extensively with the lysis buffer and cap-bound materials were eluted by boiling in Laemmli sample buffer supplemented with 100 mM DTT.

mRNA pull-down, cDNA synthesis, and qPCR

Total RNA was extracted by using Trizol (Invitrogen). mRNA was extracted by polyA+ purification with Dynabeads™ mRNA DIRECT™ Purification Kit (Invitrogen). All RNAs were measured by NanoDrop™ One (thermo fisher). Both 30 ng of total RNA and pulled-down mRNA were reverse transcribed with the Superscript IV first-strand synthesis kit for RT-qPCR (ThermoFisher). The qRT-PCR was performed by using iQ™ SYBR® Green Supermix (Bio-Rad), cDNA template, and gene-specific primer sets designed using the IDT primer design tool. No reverse transcriptase and no template controls were performed in parallel to check for DNA contamination and primer-dimer. The primer sets used for the study are given in Table S1. Threshold cycle (CT) values in qRT-PCR experiments were averaged across three biological replicates.

Immunofluorescence

Cells were plated into a 24-well plate seed with coverslips. The following day, cells were treated as indicated in figure legends. Then the cells were fixed with 4% paraformaldehyde for 15 min, permeabilized with −20°C methanol for 5 min, and blocked for 1 h with 5% normal horse serum (NHS; ThermoFisher) diluted in PBS. Primary antibodies were diluted in blocking solution and incubated for 1 h at room temperature or overnight at 4°C. Next, cells were washed three times and then incubated with secondary antibodies (Jackson Laboratories) and Hoechst 33258 (Sigma-Aldrich) for 1 h at room temperature and washed. Coverslips were mounted on glass slides with Vinol and imaged.

Fluorescence in situ hybridization (FISH)

For in situ hybridization, cells were fixed with 4% paraformaldehyde for 15 min and then permeabilized with −20°C methanol for 5 min. Cells were incubated overnight in 70% ethanol at 4°C. The following day, cells were washed twice with 2× saline-sodium citrate (SSC), blocked in hybridization buffer (Sigma) for 30 min, then hybridization was performed using a biotinylated oligo(dT40) probe (2 ng/μl) diluted in hybridization buffer at 37°C. After extensive washes with 2× SSC at 37°C the probe was revealed using Cy-conjugated streptavidin (Jackson Immunoresearch Laboratories), followed by immunostaining as described above.

Microscopy

Wide-field fluorescence microscopy was performed using an Eclipse E800 microscope (Nikon, Minato, Tokyo, Japan) equipped with epifluorescence optics and a digital camera (Spot Pursuit USB). Image acquisition was done with a 40× objective (PlanApo; Nikon, Minato, Tokyo, Japan). Images were merged using Adobe Photoshop. Eclipse Ni-E microscope (Nikon, Minato, Tokyo, Japan) and NIS-Elements imaging software (Nikon, Minato, Tokyo, Japan) were also used for the observation. Image acquisition was done with a 40× objective (PlanApo; Nikon, Minato, Tokyo, Japan).

Polysome profiling

Cells were treated with 100 mg/ml cycloheximide (Sigma) for 10 min, washed with PBS, and harvested into lysis buffer (10 mM Tris [pH 7.4], 150 mM NaCl, 5 mM MgCl2, 1 mM DTT, 100 μg/ml cycloheximide, 1% Triton-X100) supplemented with RNasin Plus inhibitor (Promega) and HALT phosphatase and protease inhibitors (Thermo Scientific). Lysates were rotated at 4°C for 10 min, and centrifuged for 5 min at 12,000×g. Supernatants were loaded onto 10 – 45 % sucrose gradients made in gradient buffer and centrifuged in a Beckman SW 40 Ti rotor for 3.5 h, 29,000xg at 4°C. Samples were eluted and its OD 254 was measured by a Biocomp gradient station attached to a syringe pump. With RNase experiments, supernatants of lysates were incubated with RNase A (Invitrogen) at 37°C for 15 min, then loaded onto 10 – 30% sucrose gradients.

Cell viability measurement

ATP measurement was done using CellTiter-Glo Luminescent Cell Viability Assay (Promega Madison, WI USA) following manufacturer instructions. Measuarements are made using the GloMax explorer plate reader (Promega Madison, WI USA).

Quantification and statistical analysis

Means were compared to each other using One-way ANOVA and both asterisks and P Values are defined as p∗ < 0.05, p∗∗ < 0.01, p∗∗∗ < 0.001, p∗∗∗∗ < 0.0001. Results are mean values ± standard deviation (SD) of at least three independent experiments, or results show one representative experiment of a minimum of three biological replicates. This biological replicates mean N. Statistical analyses were performed on all available data. Statistical tests were performed using GraphPad Prism software. SGs were visualized by IF with indicated markers. For quantifications, four fields were taken from selected samples, with three replicated independent experiments. Cells were considered SG-positive if they had at least three cytoplasmic foci.

Published: December 26, 2025