Calcium and magnesium deficiency induces stress granule formation to maintain magnesium homeostasis

Abstract

Hypocalcemia and hypomagnesemia frequently occur under pathological conditions such as Crohn’s disease or during diuretic treatment. However, how the combined deficiency of Ca²⁺ and Mg²⁺ affects cellular physiology has remained unclear. In this study, we focused on this issue and found that Ca²⁺/Mg²⁺ deprivation is a potent driver of stress granule (SG) formation. When SG formation was inhibited by G3BP1/2 knockdown, Ca²⁺/Mg²⁺ deprivation caused a further decrease in intracellular Mg²⁺ levels and an increase in cell death, indicating that SGs function to mitigate Mg²⁺ loss and protect cells from death under cation-deficient conditions. Furthermore, we found that the expression of the Mg²⁺ transporter MAGT1 is upregulated in an SG-dependent manner, and that MAGT1 knockdown further decreases intracellular Mg²⁺ levels and increases cell death. Collectively, our results demonstrate that SG formation acts as an adaptive mechanism to maintain Mg²⁺ homeostasis during Ca²⁺/Mg²⁺ deficiency.

Key words: stress granule, MAGT1, magnesium, calcium

Graphical Abstract

Introduction

Intracellular Ca²⁺ and Mg²⁺ are essential divalent cations that function as enzymatic cofactors and structural stabilizers for a wide variety of biomolecules. Ca²⁺ plays indispensable roles in neurotransmission, muscle contraction, blood coagulation, cell–cell adhesion, and stimulus–secretion coupling (Ryan, 1998; Singh et al., 2019; Terrell et al., 2023). Mg²⁺ acts as a cofactor for more than 300 enzymes and is involved in ATP hydrolysis as well as the synthesis of ATP, nucleic acids, and proteins (Williams, 2000; Barbagallo et al., 2023). Owing to differences in their ionic radius, hydration shell, and charge density, Ca²⁺ and Mg²⁺ are distinguished by specific transporters, channels, and receptors, thereby being appropriately utilized in distinct cellular contexts (Tang et al., 2014; Takeda et al., 2014). The disruption or deficiency of Ca²⁺ and Mg²⁺ homeostasis thus impairs various cellular functions and contributes to the onset or exacerbation of multiple diseases.

Although Ca²⁺ and Mg²⁺ can each become deficient individually, their simultaneous deficiency also frequently occurs. For instance, in patients with Crohn’s disease or celiac disease, malabsorption in the gastrointestinal tract lowers the systemic levels of both ions (Zheng et al., 2023). Severe hypomagnesemia also suppresses parathyroid hormone secretion, leading to secondary hypocalcemia (Abate and Clarke, 2017). Combined Ca²⁺ and Mg²⁺ deficiency is associated with pathological conditions such as osteoporosis and cognitive impairment (Fouhy et al., 2023; Kravchenko et al., 2024). However, the cellular responses triggered by the concurrent depletion of these ions, and how these responses are associated with pathophysiological outcomes, remain largely uninvestigated to date. In this study, we addressed this question and found that the simultaneous deprivation of Ca²⁺ and Mg²⁺ rapidly induces the formation of stress granules (SGs), preceding other organellar stress responses. SGs are dynamic, membraneless cytoplasmic assemblies composed of RNA-binding proteins, 40S ribosomal subunits, and untranslated mRNAs, which form in response to various cellular stresses such as oxidative stress, heat shock, and viral infection, and function primarily to suppress protein translation (Anderson and Kedersha, 2009). Because SGs are driven by liquid–liquid phase separation, they assemble and disassemble rapidly in accordance with the cellular environment. Transiently formed SGs act in a cytoprotective manner by alleviating stress-induced damage, whereas persistent SG formation has been implicated in the induction of cell death.

SG assembly is known to occur via two major pathways, i.e., an eIF2α-dependent pathway and a 4E-BP1–dependent pathway (Martin et al., 2022). In the eIF2α pathway, stress-activated kinases phosphorylate eIF2α, thereby inhibiting the delivery of initiator tRNA to the ribosome and blocking translation initiation. In the 4E-BP1 pathway, nutrient or glucose deprivation leads to mTOR inactivation, resulting in the dephosphorylation of 4E-BP1, which inhibits formation of the eIF4F complex and suppresses cap-dependent translation initiation (Fu et al., 2016). In this study, we demonstrate for the first time that combined Ca²⁺ and Mg²⁺ deficiency strongly promotes SG formation through activation of both the eIF2α-dependent and 4E-BP1-dependent pathways. Moreover, we show that SGs induce the expression of the Mg²⁺ transporter MAGT1, thereby contributing to the maintenance of intracellular Mg²⁺ homeostasis and protection against cell death. These findings suggest that SGs activate as an adaptive response to Ca²⁺/Mg²⁺ deficiency, and may represent a novel therapeutic target for disorders associated with the dysregulation of these essential cations.

Materials and Methods

Antibodies and reagents

Antibodies for G3BP1 (ab56574), TIA1 (ab140595), KDEL (ab176333), and Rab9 (ab2810) were from Abcam. Antibodies for β-actin (4970S), eIF2a (9722S), phospho-eIF2a (Ser51, 9721S), mTOR (2983S), phospho-mTOR (Ser2448, 2971S), 4E-BP1 (9644S), phospho-4E-BP1 (Thr37/46, 2855S), AMPKα (2532S), phospho-AMPKα (Thr172, 2535S), PERK (3192S), GCN2 (3302S), phospho-GCN2 (Thr899, 94668T), phospho-PERK (3179S), Tom20 (42406S), CHOP (2895S), ATF4 (11815T), and STIM1 (5668S). Antibodies for HRI (20499-1-AP), PKR (18244-1-AP), MAGT1 (17430-1-AP), CNNM4 (14066-1-AP), and ORAI1 (66223-1-Ig) were from Proteintech. Antibodies for CNNM2 (PA5102013) and TRPM7 (MA527620) were from Invitrogen; Thermo Fisher Scientific. Antibodies for puromycin (MABE343), SLC41A1 (SAB2102214), and α-tubulin (T9206) were from Sigma-Aldrich. Antibodies for TRPM6 (sc-365536), LAMP1 (sc-20011), and TRPC1 (sc-133076) were from Santa Cruz Biotechnology. The antibody for GS28 (611184) was from BD Biosciences. The antibody for LC3 (CTB-LC3-2-IC) was from Comso bio.

Dulbecco’s Modified Eagle Medium (DMEM), bovine serum albumin (BSA) (general grade, pH 7.0), BSA (fatty acid free, pH 7.0), 100 mM sodium pyruvate, penicillin-streptomycin mixed solution, 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid (HEPES) buffer solution, MEM non-essential amino acids solution, 200 mM-L-glutamine stock solution, G418 disulfate aqueous solution, 2-mercaptoethanol, puromycin dihydrochloride, thapsigargin (Tg), 1,6-Hexanediol (Hex), and propidium iodide (PI) were from Nacalai Tesque. Lipofectamine^TM RNAiMAX reagent, Lipofectamine^TM 2000 reagent, and ProLong^TM Diamond Antifade Mountant with 4',6-diamidino-2-phenylindole (DAPI) were from Thermo Fisher Scientific. Opti-MEM^TM and MEM vitamin solution were from Gibco. Sodium arsenite (As) was from Sigma-Aldrich, cycloheximide (CHX) was from FUJIFILM Wako Pure Chemical Corporation, fetal bovine serum (FBS) was from Biowest, Metallo Assay Mg LS kit was from Metallogenics, and CellTiter-Blue^® Cell Viability Assay was from Promega.

Cell culture and DNA/small interfering RNA (siRNA) transfection

MEFs were cultured in DMEM supplemented with 1 mM sodium pyruvate, nonessential amino acids, 10 mM HEPES/NaOH (pH 7.4), 0.05 mM 2-mercaptoethanol, 100 U/mL penicillin, 100 μg/mL streptomycin, and 10% FBS. U2OS, HeLa, and HEK293T cells were grown in DMEM supplemented with 100 U/mL penicillin, 100 μg/mL streptomycin, and 10% FBS. For knockdown experiments, cells were transfected with siRNAs from Dharmacon (siGENOME SMART pool) using RNAiMAX according to the manufacturer’s protocol. To induce Ca²⁺ and Mg²⁺ deficiency, cells were cultured in Ca²⁺- and Mg²⁺-deficient DMEM (custom-made medium, Elab Science) supplemented with 0.1% fatty acid-free BSA for the indicated times. The medium contained 0.09 mM of Ca²⁺ and 0.04 mM of Mg²⁺.

Immunofluorescence analysis

Cells were fixed with 4% paraformaldehyde for 10 min, and subsequently permeabilized with 0.5% Triton X-100 for 5 min. Following permeabilization, cells were incubated with the indicated primary antibodies diluted in 3% BSA for 1 h at room temperature. After washing with phosphate buffered saline (PBS (–)), cells were incubated with the corresponding secondary antibodies prepared in 3% BSA, mounted with ProLong^TM Diamond Antifade Mountant containing DAPI, and imaged using a laser-scanning confocal microscope (FV3000, Olympus). Image acquisition and quantitative analysis were performed using ImageJ software.

Puromycin incorporation assay

Cells were exposed to the following stress conditions: 10 μM thapsigargin for 1 h, 500 μM sodium arsenite for 1 h, or Ca²⁺/Mg²⁺-deficient DMEM for 4 h or 8 h. During the final 10 min of each stress treatment, cells were incubated with 10 μg/mL puromycin. After the treatment, cells were washed three times with PBS and subsequently collected for western blotting analysis.

Magnesium quantification assay

Cells (3 × 10⁵) were lysed in 30 μL of lysis buffer (10 mM HEPES-KOH [pH 7.4], 150 mM NaCl, 0.1% sodium dodecyl sulfate (SDS) and 1% Triton X-100). The pH of the lysate was adjusted to 2 to 3 by adding HCl. The samples were vortexed briefly and incubated at room temperature for 10 min, followed by centrifugation at 9,000 × g for 10 min at 4°C. The supernatant was used for magnesium quantification using the Metallo Assay Mg LS kit (Metallogenics). Reagents were added according to the manufacturer’s instructions, and absorbance was measured by spectrophotometry using a GloMax^® Discover GM3000 microplate reader (Promega).

PI uptake assay

A total of 4 × 10⁴ cells were seeded onto each well of a 12-well plate. After incubation in Ca²⁺- and Mg²⁺-deficient DMEM, both floating cells and attached cells were collected into a 15 mL tube and centrifuged at 500 × g for 5 min. Cells were washed with PBS and centrifuged at 500 × g for 5 min. Cells were then incubated with 1 μM PI for 5 min, and subjected to flow cytometric analysis using BD FACSCanto II (BD Biosciences). PI fluorescence was detected in the phycoerythrin (PE) channel (excitation at 488 nm, emission at 585/42 nm). A total of 10,000 events per sample were acquired.

Cell viability assay

Cell viability was assessed using the CellTiter-Blue assay (Promega) according to the manufacturer’s instructions. A total of 2 × 10⁴ cells were seeded onto each well of a 24-well plate. After incubation in Ca²⁺- and Mg²⁺-deficient DMEM, supernatants were removed from each well. CTB solution was added to the cells and incubated for 100 min. The culture supernatant was collected, and fluorescence was measured using a GloMax^® Discover GM3000 microplate reader (Promega).

Statistical analysis

The values are expressed as the mean ± standard deviation (SD). Prism 10 (GraphPad) software was used to perform the unpaired Student t-test and one-way analysis of variance (ANOVA) with Dunnett’s post hoc test. Changes were considered to be statistically significant if the p-value was less than 0.05.

Results

Extracellular Ca²⁺/Mg²⁺ deprivation promotes SG formation

To investigate how cells respond to Ca²⁺ and Mg²⁺ deficiency, mouse embryonic fibroblasts (MEFs) were cultured in a medium containing approximately 1/20 of the physiological concentrations of these cations (0.09 mM Ca²⁺ and 0.04 mM Mg²⁺). Morphological analyses of various organelles using specific markers at an early time point (4 h) demonstrated no major structural alterations, except for the mild perinuclear clustering of mitochondria. At this time, no autophagosome formation was observed (Supplementary Fig. 1). Although Ca²⁺ depletion is a well-known inducer of endoplasmic reticulum (ER) stress, ER-stress markers were not increased under these conditions (Supplementary Fig. 2). We therefore analyzed the formation of membraneless organelles, namely, SGs, by immunostaining for the SG markers G3BP1 and TIA1. Small G3BP1/TIA1-positive foci appeared in a subset of cells as early as 2 h after deprivation (Fig. 1A), and by 8 h, approximately 70% of cells demonstrated prominent SG formation (Fig. 1A, B). Both the number and size of SGs increased over time (Fig. 1C, D). Similar results were obtained with human U2OS osteosarcoma cells (Supplementary Fig. 3), as well as with HeLa and HEK293 cells (Supplementary Fig. 4A).

Fig. 1

SG assembly by extracellular Ca²⁺/Mg²⁺ depletion

(A–D) G3bp1 and Tia1-positive puncta/stress granules (SGs) were induced by Ca²⁺/Mg²⁺ depletion. MEFs were incubated with Ca²⁺- and Mg²⁺-deficient DMEM (0.09 mM and 0.04 mM, respectively), and subsequently stained with antibodies against SGs markers: G3bp1(green) or Tia1 (red). In (A), representative images are shown. Magnified images of the areas within the squares are shown below each image. (B, C, D) Quantification of cells with SGs (n = 3), number of SGs per cell (n>30 cells in each experiment), and SG area (n = 30 cells in each experiment). (E) Puromycin incorporation assay. MEFs were treated with 10 μM thapsigargin (Tg) for 1 h, 500 μM sodium arsenite (As) for 1 h, or Ca²⁺/Mg²⁺-deficient DMEM for 4 h and 8 h. Puromycin (10 μg/mL) was added during the final 10 min of each stress treatment. Puromycin incorporation was assessed by western blotting using an anti-puromycin antibody. (F) Semiquantitative analysis of protein expression in (E) is shown. Data are shown as the mean ± SD (n = 3). (G, H) MEFs were incubated in Ca²⁺/Mg²⁺-deficient DMEM containing 100 μg/mL cycloheximide (CHX) for 8 h, and subsequently stained with anti-G3bp1 and anti-Tia1 antibodies. (G) Representative images are shown. (H) Quantification of cells with SGs (n = 3), number of SGs per cell (n>30 cells in each experiment), and SG area (n = 30 cells in each experiment). (I, J) MEFs were cultured in Ca²⁺/Mg²⁺-deficient DMEM for 8 h, and then treated with 5% 1,6-Hexanediol (Hex) for 5 min. (I) Representative images are shown. (J) Quantification of cells with SGs (n = 3), number of SGs per cell (n>30 cells in each experiment), and SG area (n = 30 cells in each experiment). (K, L) MEFs were cultured in Ca²⁺/Mg²⁺-deficient DMEM for 6 h, and then supplemented with 1.8 mM CaCl₂ and 0.81 mM MgCl₂. After the incubation for 0 to 6 h, the cells were immunostained with anti-G3bp1 and anti-Tia1 antibodies. (K) A schematic model (left) and representative images (right). (L) Analysis of cells with SGs (n = 3), number of SGs per cell (n>30 cells in each experiment), and SG area (n = 30 cells in each experiment). In panels (B, C, F, H[i, ii], J[i, ii], L[i, ii]), data are presented as the mean ± SD (n = 3). In (B–D, F, L), comparisons were performed by one-way ANOVA with Dunnett’s post hoc test. In (H[i,ii], J[i, ii]), comparisons were performed by the unpaired Student t-test. In (H[iii], J[iii]), comparisons were performed by the Welch’s t-test. *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001; NS, not significant; scale bars, 15 μm

To confirm translational repression—a hallmark of SG formation—we monitored puromycin incorporation, which reflects ongoing translation elongation. Treatment of MEFs with known SG inducers, such as thapsigargin (Tg) and arsenite (As) , markedly reduced puromycin labeling (Fig. 1E, F). Using the same assay, Ca²⁺/Mg²⁺ deprivation reduced translation activity to approximately 45% at 4 h and approximately 10% at 8 h (Fig. 1E, F). When cells were treated with cycloheximide (CHX), which stabilizes ribosomes performing translation and prevents SG assembly, the number and area of G3BP1/TIA1-positive foci were significantly decreased (Fig. 1G–Hiii). Likewise, the addition of 1,6-hexanediol (Hex), which disrupts hydrophobic interactions within liquid–liquid phase–separated structures, dissolved SG foci formed after 8 h of ion deprivation (Fig. 1I, J). Upon the readdition of 1.8 mM CaCl₂ and 0.8 mM MgCl₂ to the deprived medium after 6 h, the fraction of SG-positive cells decreased by approximately 50% within 6 h (Fig. 1K, L), indicating that SG formation under these conditions is reversible. Collectively, these results demonstrate that Ca²⁺/Mg²⁺ deprivation acts as a potent and reversible inducer of SG assembly.

We next investigated which ion plays the dominant role in the induction of SG formation. Supplementation of the deprivation medium with physiological concentrations of either CaCl₂ (1.8 mM), MgCl₂ (0.8 mM), or both completely abolished SG formation in MEFs (Fig. 2A, B), as well as in HeLa and HEK293 cells (Supplementary Fig. 4A). In U2OS cells, SGs were also markedly reduced by either ion, though mild residual SG formation persisted when only CaCl₂ was added (Supplementary Fig. 4B, C). These observations indicate that both ions are important for the induction of SG formation, with Mg²⁺ playing a slightly more crucial role in U2OS cells.

Fig. 2

Requirement of the depletion of both Ca²⁺ and Mg²⁺ for SG formation

MEFs were treated with Ca²⁺/Mg²⁺-deficient DMEM containing vehicle or 1.8 mM CaCl₂, 0.81 mM MgCl₂, or both CaCl₂ and MgCl₂ for the indicated times, and stained with anti-G3bp1 and anti-Tia1 antibodies. (A) Representative images are shown. (B) Quantification of cells with SGs (n = 3), number of SGs per cell (n>30 cells in each experiment), and SG area (n = 30 cells in each experiment). In (B), comparisons were performed by one-way ANOVA with the Dunnett’s post hoc test. *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001; NS, not significant; scale bars, 15 μm

Ca²⁺/Mg²⁺ deprivation induces SG formation via both the 4E-BP1 and eIF2α pathways

SG formation is driven by the accumulation of untranslated mRNPs that results from translation initiation arrest. The following two canonical signaling cascades mediate this process: (1) the mTOR–4E-BP1 pathway and (2) the eIF2α phosphorylation pathway. Under nutrient stress, mTOR inactivation causes 4E-BP1 dephosphorylation, which enhances its binding to eIF4E and suppresses cap-dependent translation. Conversely, the activation of stress-responsive kinases such as GCN2 phosphorylates eIF2α, which inhibits eIF2B and blocks ternary-complex formation. We first investigated mTOR signaling, and found a reduction in phosphorylated mTOR, indicating its inactivation (Fig. 3A, Bi). Total 4E-BP1 protein levels were increased, whereas its phosphorylated form decreased (Fig. 3A, Bii), consistent with accumulation of the dephosphorylated, translation-repressive form. Knockdown of 4E-BP1 in MEFs markedly suppressed the number of cells containing G3BP1/TIA1-positive foci, as well as the number and size of SGs per cell (Fig. 3C–Fiii). These results confirm that the mTOR–4E-BP1 axis mediates SG assembly under Ca²⁺/Mg²⁺-deficient conditions (Fig. 3G). In contrast, the AMPK pathway, which can also inhibit mTOR under energy stress, showed no significant activation (Supplementary Fig. 5).

Fig. 3

SG formation regulated by the 4E-BP1 pathway

(A, B) Cells were treated with Ca²⁺/Mg²⁺-deficient DMEM for 8 h. (A) Immunoblot analysis of upstream regulators of SG formation. (B) Quantification analysis of the protein expression in (A). Data are shown as the mean ± SD (n = 3). (C, D) MEFs were transfected with either control siRNA or 4E-bp1 siRNA for 48 h, and immunoblot analysis was performed (C). (D) Quantification analysis of the protein expression in (C). Data are shown as the mean ± SD (n = 3). (E, F) MEFs were transfected with either control siRNA or 4E-bp1 siRNA for 48 h, and subsequently treated with Ca²⁺/Mg²⁺-deficient DMEM for 6 h. (E) Immunostaining with anti-G3bp1 and anti-Tia1 antibodies. Representative images are shown. (F) Quantification of cells with SGs (n = 3), number of SGs per cell (n>30 cells in each experiment), and SG area (n = 30 cells in each experiment). (G) Schematic summary of the findings of the experiments in Fig. 3. In (B), comparisons were performed by one-way ANOVA with Dunnett’s post hoc test. In (D, F[i, ii]), comparisons were performed by the unpaired Student t-test. In (F[iii]), comparisons were performed by the Welch’s t-test. *p<0.05, **p<0.01, ***p<0.001; Scale bar, 15 μm

We next investigated the eIF2α pathway. Western blotting demonstrated a clear increase in eIF2α phosphorylation upon Ca²⁺/Mg²⁺ deprivation (Fig. 4A, B). Treatment with ISRIB, an inhibitor that counteracts of eIF2α-mediated translational arrest, reduced both the proportion of SG-positive cells and the size of individual SGs (Fig. 4C, D), indicating that the eIF2α pathway is also required. Among the four known eIF2α kinases, namely, HRI, PKR, PERK, and GCN2, knockdown experiments demonstrated that the phosphorylation of eIF2α was reduced upon the silencing of PKR, PERK, or GCN2 (Fig. 4E, F). Notably, only GCN2 knockdown markedly decreased the number of G3BP1/TIA1 foci (Fig. 4G, H). Furthermore, phosphorylation of GCN2 increased approximately seven-fold after 4 h of Ca²⁺/Mg²⁺ deprivation (Fig. 4I, J). Together, these findings indicate that Ca²⁺/Mg²⁺ deficiency activates both the mTOR–4E-BP1 and GCN2–eIF2α–eIF2B pathways to promote SG formation. Both pathways appear to contribute cooperatively, although their relative importance may vary depending on the cell type or deprivation duration (Fig. 4K).

Fig. 4

SG formation regulated by the GCN2-eIF2α pathway

(A, B) MEFs were treated with Ca²⁺/Mg²⁺-deficient DMEM for the indicated times. (A) Immunoblot analysis of upstream regulators of SG formation. (B) Quantification analysis of the protein expression in (A). Data are shown as the mean ± SD (n = 3). (C, D) MEFs were treated with Ca²⁺/Mg²⁺-deficient DMEM in the presence of ISRIB for 8 h, and subsequently immunostained with anti-G3bp1 and anti-Tia1 antibodies. (C) Representative images are shown. (D) Quantification of cells with SGs (n = 3), number of SGs per cell (n>30 cells in each experiment), and SG area (n = 30 cells in each experiment). (E–H) MEFs were transfected with either control siRNA or the indicated siRNAs for 48 h, then were treated with Ca²⁺/Mg²⁺-deficient DMEM for the indicated times. (E) Immunoblot analysis of the indicated proteins. (F) Quantification analysis of the protein expression in (E). Data are shown as the mean ± SD (n = 3). (G, H) At 6 h, MEFs were immunostained with anti-G3bp1 and anti-Tia1 antibodies, and representative images are shown in (G). In (H), quantification of cells with SGs (n = 3), number of SGs per cell (n>30 cells in each experiment), and SG area (n = 30 cells in each experiment) are shown. (I, J) MEFs were treated with Ca²⁺/Mg²⁺-deficient DMEM for the indicated times. (I) Immunoblot analysis of the indicated proteins. (J) Quantification analysis of the protein expression in (I). Data are shown as the mean ± SD (n = 3). (K) Schematic summary of the findings in Fig. 4. In (B, F, H, J), comparisons were performed by one-way ANOVA with Dunnett’s post hoc test. In (D [i, ii]), comparisons were performed by the unpaired Student t-test. In (D[iii]), comparisons were performed by the Welch’s t-test. *p<0.05, **p<0.01, ***p<0.001; NS, not significant; scale bars, 15 μm

SGs exert cytoprotective effects by upregulating MAGT1 expression

To elucidate the functional significance of SGs induced by Ca²⁺/Mg²⁺ deprivation, we compared G3BP1/2-knockdown MEFs with control cells (Fig. 5A–C). Because the effects of Ca²⁺/Mg²⁺ deprivation were expected to be observed upon the expression of Ca²⁺ and Mg²⁺ transporters localized at the plasma membrane, we first analyzed their expression levels. We found that the Mg²⁺ transporter MAGT1 was robustly and time-dependently induced by Ca²⁺/Mg²⁺ deprivation, and this induction was completely abolished in G3BP1/2-knockdown cells (Fig. 5B, C). Another Mg²⁺-associated transporter, CNNM2, also showed a modest increase upon Ca²⁺/Mg²⁺ deprivation, but its expression was unaffected by G3BP1/2 knockdown. The other Mg²⁺-associated transporters, CNNM4 and TRPM7, remained largely unchanged (Fig. 5B, C). Among the Ca²⁺ transporters located at the plasma membrane, Stim1 was slightly reduced by Ca²⁺/Mg²⁺ deprivation and G3BP1/2 knockdown individually, but no further changes were observed when both conditions were applied simultaneously. The other Ca²⁺-associated transporters, ORAI1 and TRPC1 showed no appreciable alterations (Supplementary Fig. 6). These results suggested that MAGT1 is most strongly affected by SGs. Considering that MAGT1 has been reported to mediate Mg²⁺ uptake at the plasma membrane, we hypothesized that SGs formed under Ca²⁺/Mg²⁺ deficiency enhance MAGT1 expression to help maintain intracellular Mg²⁺ levels. Notably, MAGT1 expression was observed under Mg²⁺ deficiency alone and Ca²⁺ deficiency alone, but its induction was much stronger when both ions were depleted simultaneously (Supplementary Fig. 7A, B). We next measured intracellular Mg²⁺ concentrations. Whereas Mg²⁺ levels were relatively preserved under either Ca²⁺ deficiency alone or Mg²⁺ deficiency alone (Supplementary Fig. 7C, D), they were decreased only when both ions were simultaneously depleted (Fig. 5D). The observation that intracellular Mg²⁺ levels were maintained under Mg²⁺ deficiency alone but decreased under combined Ca²⁺/Mg²⁺ deficiency is likely attributable to the impaired function of Ca²⁺-dependent Mg²⁺ transporters such as TRPM7. To test whether SG-mediated upregulation of MAGT1 contributes to maintaining intracellular Mg²⁺ levels, we measured Mg²⁺ concentrations in G3BP-knockdown cells and MAGT1-knockdown cells. Both knockdown cell lines showed a further reduction in intracellular Mg²⁺ levels (Fig. 5D–F). We then assessed cell death using propidium iodide (PI) staining, which labels the nucleotides of dead cells, and the CellTiter-Blue metabolic activity assay. Neither Ca²⁺ deficiency alone nor Mg²⁺ deficiency alone induced cell death (Supplementary Fig. 8), whereas combined Ca²⁺/Mg²⁺ deficiency led to progressive cell death (Fig. 5G–J). Importantly, G3BP1/2 knockdown and MAGT1 knockdown both resulted in markedly exacerbated cell death under these conditions (Fig. 5G–J). Taken together, these findings indicate that SGs formed under Ca²⁺/Mg²⁺ deprivation promote the induction of MAGT1, thereby maintaining intracellular Mg²⁺ homeostasis and suppressing cell death.

Fig. 5

Involvement of SGs in maintaining Mg²⁺ channels and transporter levels

(A–D) MEFs were transfected with either control siRNA or G3bp1/2 siRNA, and subsequently treated with Ca²⁺/Mg²⁺-deficient DMEM. (A) MEFs were immunostained with anti-G3bp1 and anti-Tia1 antibodies. Representative images are shown. (B) MEFs were treated as in (A) and lysates were subjected to western blotting. (C) Quantification of the protein expression in (B). (D) Intracellular Mg²⁺ levels were analyzed using the Metallo Assay Mg LS kit. (E, F) MEFs were transfected with either control siRNA or Magt1 siRNA, and the lysates were subjected to western blotting or intracellular Mg²⁺ quantification. (G–J) Cell death and cell viability were analyzed using the PI uptake assay and the CellTiter-Blue assay (CTB), respectively. In (C, D, F–J), comparisons were performed by the unpaired Student t-test. *p<0.05, **p<0.01, ***p<0.001; NS, not significant; scale bars, 15 μm

Discussion

The deficiency of serum Ca²⁺ and Mg²⁺ can occur under clinical conditions such as Crohn’s disease or during diuretic treatment (Quamme, 1981; Zheng et al., 2023). However, few studies have directly investigated the cellular and molecular mechanisms underlying the combined deficiency of Ca²⁺ and Mg²⁺. In this study, using in vitro systems including MEFs, we investigated cellular responses to Ca²⁺/Mg²⁺ deprivation. We found that simultaneous Ca²⁺/Mg²⁺ deficiency potently induces the formation of SGs, which in turn contribute to the maintenance of Mg²⁺ homeostasis and cell survival through the regulation of Mg²⁺ transporters. Although previous studies have shown that SGs can form in response to ionic imbalance or ion leakage (Duran et al., 2024), direct evidence that SGs actively participate in maintaining ion homeostasis has been lacking to date. Our findings therefore demonstrate a previously unrecognized role of SGs in the adaptive response to mineral deficiency.

SG formation is known to be regulated by both the eIF2α-dependent and 4E-BP1-dependent pathways, which were originally considered to function independently (Martin et al., 2022). In the present study, we demonstrated that both the GCN2–eIF2α and mTOR–4E-BP1 pathways cooperatively contribute to SG assembly under Ca²⁺/Mg²⁺-deficient conditions (Fig. 3, 4). How, then, does Ca²⁺/Mg²⁺ deprivation activate these pathways? Regarding the GCN2–eIF2α pathway, GCN2 is known to sense stress through binding to uncharged tRNAs or stalled ribosomes under amino acid deprivation or UV-induced stress (Misra et al., 2024; Paternoga et al., 2025). Because Mg²⁺ deficiency directly destabilizes ribosomes (Yu et al., 2023), GCN2 activation is likely triggered by its interaction with destabilized ribosomal complexes. In addition, Ca²⁺ deprivation can induce endoplasmic reticulum and mitochondrial stress (Pontisso et al., 2024), which may secondarily activate GCN2 through translational suppression and ribosome stalling. Regarding the mTOR–4E-BP1 pathway, Mg²⁺ plays a crucial role in the GTP binding required for the activation of Rag GTPases that localize mTOR to the lysosomal membrane (Vetter and Wittinghofer, 2001; Sancak et al., 2010). Mg²⁺ deficiency likely reduces Rag GTPase activity, thereby inactivating mTOR and leading to 4E-BP1 dephosphorylation. As Mg²⁺ is required both for ribosomal stability and for Rag GTPase function, Ca²⁺/Mg²⁺ deficiency may coordinately trigger the GCN2–eIF2α and mTOR–4E-BP1 pathways through a shared ionic stress mechanism. Consistently, previous reports have shown that Ca²⁺ chelation decreases 4E-BP1 phosphorylation (Li et al., 2016), suggesting that both ions are important for translational control through this pathway.

Several Mg²⁺ transporters have been identified, including TRPM6, TRPM7, and MAGT1 (Voets et al., 2004; Goytain and Quamme, 2005; Zhou and Clapham, 2009; Schmidt et al., 2022). In this study, we demonstrated using MEFs that Ca²⁺/Mg²⁺ deprivation increases MAGT1 protein levels, whereas Ca²⁺ transporters are largely unaffected. Thus, the cellular response to Ca²⁺/Mg²⁺ deprivation appears to be more prominently reflected in intracellular Mg²⁺ levels. Interestingly, intracellular Mg²⁺ concentrations decreased more markedly under combined Ca²⁺/Mg²⁺ deprivation than under Mg²⁺ deprivation alone. This may be explained by the Ca²⁺ dependence of plasma-membrane Mg²⁺ transporters such as TRPM7, which are likely impaired in their function when Ca²⁺ levels decrease. Because the inhibition of SG formation abolished MAGT1 induction, it is likely that MAGT1 upregulation is functionally linked to SG formation. However, it remains unclear whether MAGT1 lies directly downstream of SGs or whether both are regulated in parallel by a shared stress-response pathway. Indeed, the mild induction of MAGT1 was also observed under Mg²⁺ or Ca²⁺ deficiency alone, which are conditions that do not induce SG formation, suggesting that SG-independent regulatory mechanisms may also contribute. Nonetheless, knockdown of either G3BP1/2 or MAGT1 resulted in further reductions in intracellular Mg²⁺ levels and increased cell death, strongly indicating that SGs promote cell survival by inducing MAGT1 and thereby maintaining Mg²⁺ homeostasis.

Although our findings are based on in vitro experiments, they may have important implications in vivo under both physiological and pathological conditions. Clinical observations have demonstrated that patients with Crohn’s disease, celiac disease, or short bowel syndrome, or those undergoing cisplatin chemotherapy frequently develop hypocalcemia and hypomagnesemia. Our results suggest that SGs can sense and respond to Ca²⁺/Mg²⁺ deficiency by upregulating Mg²⁺ transporters such as MAGT1, thereby helping to maintain their enzymatic activity and their essential cellular functions under mineral-depleted conditions. In this context, SG formation may act as a protective cellular adaptation to ionic imbalances in epithelial, renal, and neuronal tissues, which are highly sensitive to divalent cation fluctuations. Furthermore, impaired SG dynamics caused by mutations in core SG components or chronic metabolic stress, may exacerbate the susceptibility to Ca²⁺/Mg²⁺ deficiency–associated disorders. Thus, the modulation of SG assembly or its downstream effectors may represent a novel therapeutic strategy to enhance cellular resilience in diseases associated with hypocalcemia, hypomagnesemia, or associated metabolic disorders.

Author Declaration Statements

Funding

This study was supported by JSPS KAKENHI Grants-in-Aid for Scientific Research (S) (23H05480) and (C) (24K11166, 22K07345, and 23K05748), and by MEXT KAKENHI Grants-in-Aid for Scientific Research on Innovative Areas (22H04639, 23H04773, and 24H01887). This study was also supported by AMED-CREST (JP23gm1410012), and by AMED under Grant Number JP21wm0525028, JP24ama221130, JP24ek0109770. This work was also supported by JST SPRING, Grant Number JPMJSP2120.

Conflict of Interest Statement

The authors declare that they have no competing interests associated with this manuscript.

Data Availability Statement

The datasets generated during the current study are available from the corresponding author upon reasonable request.

Author Contribution Statement

SS conceived and supervised the project. TS conducted all the experiments and analyzed the results with support from MT, SH, and ST. The work was carried out under the supervision of ST and SS. TS and SS wrote the manuscript. All authors read and approved the final version of manuscript.

Ethics Approval and Consent to Participate

Ethics approval is not required in this study.

Patient Consent for Publication

Not applicable.

Acknowledgments

We would like to thank Dr. Satoko Arakawa and all members of SHIMIZU Laboratory for helpful discussion.

References

•  Abate,  E.G. and  Clarke,  B.L. 2017. Review of hypoparathyroidism. Front. Endocrinol. (Lausanne), 7: 172.
•  Anderson,  P. and  Kedersha,  N. 2009. RNA granules: Post-transcriptional and epigenetic modulators of gene expression. Nat. Rev. Mol. Cell Biol., 10: 430–436.
•  Barbagallo,  M.,  Veronese,  N., and  Dominguez,  L.J. 2023. Magnesium—An ion with multiple invaluable actions, often insufficiently supplied: From in vitro to clinical research. Nutrients, 15: 3135.
•  Duran,  J.,  Salinas,  J.E.,  Wheaton,  R.P.,  Poolsup,  S.,  Allers,  L.,  Rosas-Lemus,  M.,  Chen,  L.,  Cheng,  Q.,  Pu,  J.,  Salemi,  M.,  Phinney,  B.,  Ivanov,  P.,  Lystad,  A.H.,  Bhaskar,  K.,  Rajaiya,  J.,  Perkins,  D.J., and  Jia,  J. 2024. Calcium signaling from damaged lysosomes induces cytoprotective stress granules. EMBO J., 43: 6410–6443.
•  Fouhy,  L.E.,  Mangano,  K.M.,  Zhang,  X.,  Hughes,  B.D.,  Tucker,  K.L., and  Noel,  S.E. 2023. Association between a calcium-to-magnesium ratio and osteoporosis among Puerto Rican adults. J. Nutr., 153: 2642–2650.
•  Fu,  X.,  Gao,  X.,  Ge,  L.,  Cui,  X.,  Su,  C.,  Yang,  W.,  Sun,  X.,  Zhang,  W.,  Yao,  Z.,  Yang,  X., and  Yang,  J. 2016. Malonate induces the assembly of cytoplasmic stress granules. Chen Q, editor. FEBS Lett., 590: 22–33.
•  Goytain,  A. and  Quamme,  G.A. 2005. Identification and characterization of a novel mammalian Mg²⁺ transporter with channel-like properties. BMC Genom., 6: 48.
•  Kravchenko,  G.,  Stephenson,  S.S.,  Gutowska,  A.,  Klimek,  K.,  Chrząstek,  Z.,  Pigłowska,  M.,  Kostka,  T., and  Sołtysik,  B.K. 2024. The concurrent association of magnesium and calcium deficiencies with cognitive function in older hospitalized adults. Nutrients, 16: 3756.
•  Li,  R.J.,  Xu,  J.,  Fu,  C.,  Zhang,  J.,  Zheng,  Y.G.,  Jia,  H., and  Liu,  J.O. 2016. Regulation of mTORC1 by lysosomal calcium and calmodulin. eLife, 5: e19360.
•  Martin,  J.L.,  Dawson,  S.J., and  Gale,  J.E. 2022. An emerging role for stress granules in neurodegenerative disease and hearing loss. Hear. Res., 426: 108634.
•  Misra,  J.,  Carlson,  K.R.,  Spandau,  D.F., and  Wek,  R.C. 2024. Multiple mechanisms activate GCN2 eIF2 kinase in response to diverse stress conditions. Nucleic Acids Res., 52: 1830–1846.
•  Paternoga,  H.,  Xia,  L.,  Dimitrova-Paternoga,  L.,  Li,  S.,  Yan,  L.L.,  Oestereich,  M.,  Kasvandik,  S.,  Nanjaraj Urs,  A.N.,  Beckert,  B.,  Tenson,  T.,  Zaher,  H.,  Inada,  T., and  Wilson,  D.N. 2025. Structure of a Gcn2 dimer in complex with the large 60S ribosomal subunit. Proc. Natl. Acad. Sci. USA, 122: e2415807122.
•  Pontisso,  I.,  Ornelas-Guevara,  R.,  Chevet,  E.,  Combettes,  L., and  Dupont,  G. 2024. Gradual ER calcium depletion induces a progressive and reversible UPR signaling. (Trejo J, ed.) PNAS Nexus, 3: pgae229.
•  Quamme,  G.A. 1981. Effect of furosemide on calcium and magnesium transport in the rat nephron. Am. J. Physiol. Renal. Physiol., 241: F340–347.
•  Ryan,  T.A. 1998. Probing a complex question: When are SNARE proteins ensnared? Nat. Neurosci., 1: 175–177.
•  Sancak,  Y.,  Bar-Peled,  L.,  Zoncu,  R.,  Markhard,  A.L.,  Nada,  S., and  Sabatini,  D.M. 2010. Ragulator-rag complex targets mTORC1 to the lysosomal surface and is necessary for its activation by amino acids. Cell, 141: 290–303.
•  Schmidt,  E.,  Narangoda,  C.,  Nörenberg,  W.,  Egawa,  M.,  Rössig,  A.,  Leonhardt,  M.,  Schaefer,  M.,  Zierler,  S.,  Kurnikova,  M.G.,  Gudermann,  T., and  Chubanov,  V. 2022. Structural mechanism of TRPM7 channel regulation by intracellular magnesium. Cell Mol. Life Sci., 79: 225.
•  Singh,  S.,  Dodt,  J.,  Volkers,  P.,  Hethershaw,  E.,  Philippou,  H.,  Ivaskevicius,  V.,  Imhof,  D.,  Oldenburg,  J., and  Biswas,  A. 2019. Structure functional insights into calcium binding during the activation of coagulation factor XIII A. Sci. Rep., 9: 11324.
•  Takeda,  H.,  Hattori,  M.,  Nishizawa,  T.,  Yamashita,  K.,  Shah,  S.T.A.,  Caffrey,  M.,  Maturana,  A.D.,  Ishitani,  R., and  Nureki,  O. 2014. Structural basis for ion selectivity revealed by high-resolution crystal structure of Mg²⁺ channel MgtE. Nat. Commun., 5: 5374.
•  Tang,  L.,  Gamal El-Din,  T.M.,  Payandeh,  J.,  Martinez,  G.Q.,  Heard,  T.M.,  Scheuer,  T.,  Zheng,  N., and  Catterall,  W.A. 2014. Structural basis for Ca²⁺ selectivity of a voltage-gated calcium channel. Nature, 505: 56–61.
•  Terrell,  K.,  Choi,  S., and  Choi,  S. 2023. Calcium’s role and signaling in aging muscle, cellular senescence, and mineral interactions. IJMS, 24: 17034.
•  Vetter,  I.R. and  Wittinghofer,  A. 2001. The guanine nucleotide-binding switch in three dimensions. Science, 294: 1299–1304.
•  Voets,  T.,  Nilius,  B.,  Hoefs,  S.,  van der Kemp,  A.W.C.M.,  Droogmans,  G.,  Bindels,  R.J.M., and  Hoenderop,  J.G.J. 2004. TRPM6 forms the Mg²⁺ influx channel involved in intestinal and renal Mg²⁺ absorption. J. Biol. Chem., 279: 19–25.
•  Williams,  N.H. 2000. Magnesium ion catalyzed ATP hydrolysis. J. Am. Chem. Soc., 122: 12023–12024.
•  Yu,  T.,  Jiang,  J.,  Yu,  Q.,  Li,  X., and  Zeng,  F. 2023. Structural insights into the distortion of the ribosomal small subunit at different magnesium concentrations. Biomolecules, 13: 566.
•  Zheng,  Y.,  Liao,  Y.,  Ouyang,  Y.,  Wu,  Z.,  Li,  Z.,  Lin,  J.,  Zhang,  K.,  Wang,  X.,  Cen,  Z.,  Ma,  W., and  Dai,  S. 2023. The effects and predictive value of calcium and magnesium concentrations on nutritional improvement, inflammatory response and diagnosis in patients with Crohn’s disease. J. Hum. Nutr. Diet., 36: 1649–1660.
•  Zhou,  H. and  Clapham,  D.E. 2009. Mammalian MagT1 and TUSC3 are required for cellular magnesium uptake and vertebrate embryonic development. Proc. Natl. Acad. Sci. USA, 106: 15750–15755.

Abbreviations

SG

stress granule

DMEM

Dulbecco’s Modified Eagle Medium

BSA

bovine serum albumin

Tg

thapsigargin

DAPI

4',6-diamidino-2-phenylindole

As

sodium arsenite

MEFs

mouse embryonic fibroblasts

ER

endoplasmic reticulum

CHX

cycloheximide

Hex

1,6-hexanediol

mRNPs

messenger ribonucleoproteins

ISRIB

integrated stress response inhibitor

PI

propidium iodide