A KDELR-mediated ER-retrieval system modulates mitochondrial functions via the unfolded protein response in fission yeast

Abstract

KDELR (Erd2 [ER retention defective 2] in yeasts) is a receptor protein that retrieves endoplasmic reticulum (ER)–resident proteins from the Golgi apparatus. However, the role of the KDELR-mediated ER-retrieval system in regulating cellular homeostasis remains elusive. Here, we show that the absence of Erd2 triggers the unfolded protein response (UPR) and enhances mitochondrial respiration and reactive oxygen species in an UPR-dependent manner in the fission yeast Schizosaccharomyces pombe. Moreover, we perform transcriptomic analysis and find that the expression of genes related to mitochondrial respiration and the tricarboxylic acid cycle is upregulated in a UPR-dependent manner in cells lacking Erd2. The increased mitochondrial respiration and reactive oxygen species production is required for cell survival in the absence of Erd2. Therefore, our findings reveal a novel role of the KDELR–Erd2-mediated ER-retrieval system in modulating mitochondrial functions and highlight its importance for cellular homeostasis in the fission yeast.

The maintenance of endoplasmic reticulum (ER) homeostasis is a fundamental process of the cell. ER-luminal proteins, including the ER-stress response chaperones, glucose-regulated protein-78 (GRP78) and glucose-regulated protein-94 (GRP94), and protein disulfide isomerase (PDI), are essential for the maintenance of ER homeostasis. These proteins carry the characteristic motif of xDEL (x, lysine, histidine, or arginine; D, aspartic acid; E, glutamic acid; and L, leucine) at their C-terminal ends (1, 2). The characteristic motif enables the retrieval of escaped ER-luminal proteins from the Golgi apparatus by the evolutionarily conserved receptor proteins, that is, KDELR1/2/3 (KDEL receptors 1, 2, and 3) in humans and Erd2 (ER retention defective 2) in yeasts (2). Therefore, the KDELR–Erd2 retrieval system plays a crucial role in maintaining ER homeostasis. Although much progress has been made in delineating the mechanism of action of KDELR–Erd2 (1, 3), how the KDELR–Erd2-mediated retrieval system regulates cellular homeostasis remains unclear.

KDELR–Erd2 is a seven-transmembrane transporter-like protein that cycles between the Golgi apparatus and the ER to retrieve the xDEL-containing ER-resident proteins (1, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17). It has been demonstrated that a polar luminal cavity is present on the surface of KDELR–Erd2, which is responsible for recognizing and selecting xDEL-containing ER-resident proteins (1, 3).

Erd2 in the budding yeast Saccharomyces cerevisiae is essential, and Erd2-deficient cells accumulate intercellular membranes and lipid droplets and have impaired secretory pathways (4). Despite the essentiality of Erd2, the Erd2-mediated retrieval system does not appear to be essential for cell viability (4, 18). These paradoxical findings are interpreted to be due to the existence of the Ire1 (inositol-requiring enzyme 1)-mediated mechanism that works in parallel with the Erd2-retrieval system to maintain ER homeostasis (19). Ire1 is an ER transmembrane sensor responsible for activating the unfolded protein response (UPR) signaling pathway (20). Impairment of Erd2 may also induce UPR, given its role in ER quality control. However, this hypothesis has not been tested for yeasts. Nonetheless, impairment of the KDELR–Erd2 retrieval system in mammalian cells enhances UPR induced chemically (20). The cellular responses caused by the impairment of KDELR–Erd2 and the related physiological significance await further investigation.

Impairment of the KDELR–Erd2 retrieval system has pathophysiological consequences. KDELR1 mutations incapable of binding ligands reduce the number of naive T cells (21) or cause lymphopenia (22). In addition, mice carrying the KDELR1 mutant D193N, capable of ligand recognition but defective in distribution to the ER, display accumulation of misfolded proteins in the ER and develop dilated cardiomyopathy (23). These pathophysiological consequences appear to be due to stress signaling responses, including the integrated stress responses and/or UPR (21, 23, 24), implying that a profound metabolic remodeling occurs upon impairment of the KDELR–Erd2 retrieval system. However, how the KDELR–Erd2 retrieval system mobilizes organelles, including the ER and mitochondria, to remodel metabolism remains unclear.

Using the fission yeast Schizosaccharomyces pombe as a model organism (25, 26, 27), we performed a microscopy-based screen and identified Emr1 as a crucial protein regulating the contact between mitochondria and the ER (28). In the screen, we also identified Erd2, the KDELR counterpart, as a protein regulating mitochondrial function. Although Erd2 has not been well characterized in fission yeasts, it has been shown that erd2 is not essential and that cells lacking Erd2 are sensitive to hydrogen peroxide, caffeine, and tunicamycin (29, 30). In the present work, we demonstrate that the absence of Erd2 induces UPR and increases mitochondrial respiration and reactive oxygen species (ROS) in an UPR-dependent manner. The absence of Erd2 affects mitochondrial functions by increasing expression of the genes involved in mitochondrial respiration and the tricarboxylic acid (TCA) cycle. Increases in mitochondrial respiration and ROS are required for maintaining cell viability. Hence, our findings reveal an uncharacterized role of the KDELR–Erd2 retrieval system in maintaining cellular homeostasis by guiding the signaling pathway of UPR.

Results

Erd2 is required for maintaining mitochondrial tubular structures and regulating mitochondrial respiration and ROS

Although Erd2 resides mainly at the Golgi apparatus and cycles between the Golgi apparatus and the ER (31), the canonical function of Erd2 is to retrieve ER-resident proteins that escape from the ER (2). Intriguingly, we observed that the absence of erd2 caused mitochondrial fragmentation (Fig. 1, A and B). This raises the question of how a protein that does not appear to localize to mitochondria is involved in regulating mitochondrial functions.

Figure 1.

The absence of Erd2 caused mitochondrial fragmentation and increased mitochondrial respiration and reactive oxygen species (ROS).A, maximum projection images of WT and erd2Δ cells cultured in EMM. Mitochondria were stained with MitoTracker Green. Note that mitochondria became fragmented in the absence of Erd2. Scale bar represents 10 μm. B, quantification of mitochondrial morphology for the cells in (A), and the number of cells observed is indicated. C, oxygen consumption rates (OCRs) of the indicated cells. A Strathkelvin oxygen respirometer was used to measure OCR. A representative result from three independent experiments was shown (also see Fig. S1A for all data), and five repeats (n) were performed for each type of strains. The height of the column is the mean, and error bars indicate SD. The p values were calculated by one-way ANOVA with a post hoc Tukey's honestly significant difference test. Antimycin A (0.15 μg/ml) was used to inhibit mitochondrial respiration. Note that the absence of Erd2 increased OCR. D, assessment of mitochondrial ROS of WT and erd2Δ cells by staining with MitoSOX Red. Scale bar represents 10 μm. E, quantification of average fluorescent intensity of mitochondrial staining indicated in (C). Three independent experiments were performed (indicated by R1–3), and the number of cells analyzed is indicated. The p values were calculated by the Wilcoxon–Mann–Whitney rank sum test or Student’s t test (indicated by asterisks). Note that the absence of Erd2 increased ROS. DIC, differential interference contrast; EMM, Edinburgh minimal medium; Erd2, ER retention defective 2.

To address this question, we examined mitochondrial morphology and functions in prototrophic WT and erd2-deleted (erd2Δ) strains. Staining the prototrophic strains with MitoTracker Green showed that mitochondria became highly fragmented in erd2Δ cells cultured in the Edinburgh minimal medium (EMM) (Fig. 1, A and B). Moreover, the oxygen consumption rate of erd2Δ cells was higher than that of WT cells (Figs. 1C and S1A), suggesting that mitochondrial respiration was increased in erd2Δ cells. In addition, MitoSOX staining was employed to assess ROS within mitochondria. As shown in Figure 1, D and E, mitochondrial ROS was higher in erd2Δ cells than in WT cells. Collectively, these results suggest that Erd2 is involved in regulating mitochondrial morphology, mitochondrial respiration, and mitochondrial ROS.

Erd2 guides the signaling pathway of UPR

Since Erd2 has a crucial role in maintaining ER homeostasis, we hypothesized that Erd2 may regulate mitochondria via the ER. To test this hypothesis, we assessed the UPR signaling pathway. Reduced expression of gas2 and yop1 is a hallmark of UPR in fission yeast (32). Therefore, we monitored the expression of gas2 and yop1 by RT–quantitative PCR (qPCR). As shown in Figure 2A, the absence of Erd2, but not the absence of the UPR activator Ire1, reduced the expression of both gas2 and yop1, similar to the effect by treatment of cells with the UPR inducer tunicamycin (0.5 μg/ml for 1 h). The absence of both Erd2 and Ire1, however, did not reduce the expression of gas2 and yop1. These results suggest that the absence of Erd2 induces UPR.

Figure 2.

The absence of Erd2 activated UPR, and UPR increased mitochondrial respiration and ROS.A, the expression of gas2 and yop1 in the indicated cells analyzed by real-time quantitative PCR (RT–qPCR). Shown are fold changes of the indicated cells over wildtype control. The number (n) of the independent measurements is indicated. The height of each bar represents the mean, and error bars indicate confidence intervals (alpha = 0.05). Single-group Student’s t test was used to calculate the p values. B, cell growth assays. The indicated cells, by 10-fold serial dilutions, were spotted on EMM plates containing tunicamycin (Tm) or DMSO. Images were taken after 3 days of culture at 30 °C. C, maximum projection images of WT cells treated with DMSO (solvent) or Tm (0.5 μg/ml) for 1 h or 2 h. Mitochondria were stained with MitoTracker Green. Note that, in cells treated with Tm, mitochondria remained tubular. Quantification of mitochondrial morphology is shown on the right of the images, and the number of cells observed is indicated. Scale bar represents 10 μm. D, oxygen consumption rates of the indicated cells. Shown is a representative result from three independent experiments (also see Fig. S1B for all data), and five repeats were performed for each type of treatment. The height of the column is the mean, and error bars indicate SD. The p values were calculated by one-way ANOVA with a post hoc Tukey's honestly significantly difference test. Antimycin A was used to inhibit mitochondrial respiration. Note that Tm increased oxygen consumption rate (OCR). E, assessment of mitochondrial ROS of DMSO- or Tm-treated cells by staining with MitoSOX Red. Scale bar represents 10 μm. F, quantification of average fluorescent intensity of mitochondrial staining indicated in (E). Three independent experiments were performed (indicated by R1–3), and the number of cells analyzed is indicated. The p values were calculated by a Wilcoxon–Mann–Whitney rank sum test. Note that Tm increased ROS. DIC, differential interference contrast; DMSO, dimethyl sulfoxide; Erd2, ER retention defective 2; ROS, reactive oxygen species; UPR, unfolded protein response.

Cells defective in UPR are sensitive to tunicamycin (32). Consistently, ire1Δ cells failed to grow on EMM plates containing tunicamycin (Fig. 2B). Paradoxically, erd2Δ cells, in which UPR was activated (Fig. 2A), also grew poorly on EMM plates containing tunicamycin. Considering the canonical role of Erd2 in retrieving ER-resident proteins required for maintaining ER homeostasis (2), we favored the interpretation that poor cell growth was due to impairment of ER protein–folding capacity caused by failure in the retrieval of ER-resident proteins in erd2Δ cells. Under unstressed conditions, that is, when cells were grown on EMM plates, ire1Δ and erd2Δ had no noticeable effect on cell growth, but ire1Δerd2Δ impaired cell growth (Fig. 2B). Collectively, these results indicate that when the KDELR–Erd2 retrieval system is impaired under unstressed conditions, UPR becomes necessary for maintaining cellular homeostasis.

UPR increases mitochondrial respiration and mitochondrial ROS

To investigate the effect of UPR on mitochondria, we treated cells with 0.5 μg/ml tunicamycin. MitoTracker Green staining revealed that tunicamycin did not alter mitochondrial tubular structures during a 1 h or 2 h treatment with tunicamycin (Fig. 2C). However, tunicamycin increased the oxygen consumption rate of cells (Figs. 2D and S1B), which was similar to the effect caused by the absence of Erd2 (Fig. 1C). In addition, tunicamycin increased ROS within mitochondria (Fig. 2, E and F), which was similar to the effect caused by the absence of Erd2 (Fig. 1, D and E). Therefore, these results suggest that UPR increases mitochondrial respiration and ROS but does not affect mitochondrial morphology.

Erd2 regulates mitochondrial respiration and ROS production via UPR

To test whether the mitochondrial alteration caused by the absence of Erd2 depends on UPR, we compared mitochondrial morphology, mitochondrial respiration, and mitochondrial ROS in WT, erd2Δ, ire1Δ, and erd2Δire1Δ cells. As shown in Figure 3, A and B, mitochondria were fragmented in erd2Δ cells but not in WT, ire1Δ, or erd2Δire1Δ cells. This result suggests that the absence of Erd2 caused mitochondrial fragmentation in an Ire1-dependent manner. Of note, the absence of Erd2 resulted in UPR activation (Fig. 2A), but induction of UPR by tunicamycin did not cause mitochondrial fragmentation (Fig. 2C). Therefore, mitochondrial fragmentation caused by the absence of Erd2 requires UPR, but UPR alone is not sufficient to induce mitochondrial fragmentation.

Figure 3.

Increases in mitochondrial respiration and ROS caused by the absence of Erd2 depended on Ire1.A, maximum projection images of WT, erd2Δ, ire1Δ, and erd2Δire1Δ cells stained with MitoTracker Green. Note that mitochondria were fragmented in erd2Δ cells and became tubular in erd2Δire1Δ cells. Scale bar represents 10 μm. B, quantification of mitochondrial morphology for the cells in (A), and the number of cells observed is indicated. C, oxygen consumption rates (OCRs) of the indicated cells. Shown is a representative result from three independent experiments (also see Fig. S3C for all data), and five repeats were performed for each type of strain. The height of the column is the mean, and error bars indicate SD. The p values were calculated by one-way ANOVA with a post hoc Tukey's honestly significant difference test. Antimycin A was used to inhibit mitochondrial respiration. Note that the absence of Erd2, but not the absence of both Erd2 and Ire1, increased OCR. D, assessment of mitochondrial ROS of the indicated cells by staining with MitoSOX Red. Scale bar represents 10 μm. E, quantification of average fluorescent intensity of mitochondrial staining indicated in (C). Three independent experiments were performed (indicated by R1–3), and the number of cells analyzed is indicated. The p values were calculated by a Wilcoxon–Mann–Whitney rank sum test or by Student’s t test (indicated by asterisks). Note that the absence of Erd2, but not the absence of both Erd2 and Ire1, increased ROS. F, summary of the effects of erd2Δ, ire1Δ, ire1Δerd2Δ, and tunicamycin on mitochondrial morphology, respiration, and ROS. Red arrows indicate an increase, blue X indicates no response of UPR, and green bars indicate no significant change. DIC, differential interference contrast; Erd2, ER retention defective 2; ROS, reactive oxygen species; UPR, unfolded protein response.

We further measured the oxygen consumption rates of WT, erd2Δ, ire1Δ, and erd2Δire1Δ cells (Figs. 3C and S1C). The results showed that the oxygen consumption rate of erd2Δ cells, but not ire1Δ or erd2Δire1Δ cells, was increased. Of note, induction of UPR by tunicamycin also increased the oxygen consumption rate (Fig. 2D). Therefore, the absence of Erd2 increased mitochondrial respiration in an UPR/Ire1-dependent manner. Similarly, the absence of Erd2 increased mitochondrial ROS in an UPR/Ire1-dependent manner (Fig. 3, D and E). The alterations of mitochondria by these mutants and tunicamycin are summarized in Figure 3F. In conclusion, our findings suggest that the absence of Erd2 increases mitochondrial respiration and ROS in an UPR-dependent manner.

Erd2 regulates the expression of the genes involved in the mitochondrial electron transport chain and the TCA cycle

To investigate how the absence of Erd2 increases mitochondrial respiration and mitochondrial ROS, we performed RNA-Seq for WT, erd2Δ, ire1Δ, and erd2Δire1Δ cells (samples in triplicate, see Supporting RNA-Seq data). Gene Ontology (GO) analysis of the differentially expressed genes (erd2Δ versus WT) revealed that genes related to the mitochondrial electron transport chain and the TCA cycle were enriched (Fig. S2). Examination of the GO-enriched genes revealed that the absence of Erd2 increased (with p values [adjusted] < 0.05 and fold change >1.5) the expression of 1, 2, 8, 6, and 13 genes encoding proteins in the yeast counterpart of the NADH dehydrogenase (complex I), succinate dehydrogenase (complex II), oxidoreductase (complex III), cytochrome c oxidase (complex IV), and the ATP synthase (complex V), respectively (Fig. 4A). Therefore, expression of ∼80% of the genes in the mitochondrial respiratory chain was increased, consistent with the enhancement of mitochondrial respiration in erd2Δ cells (Fig. 3C). Moreover, expression of most genes encoding proteins responsible for the TCA cycle was also increased (Fig. 4A). Such increased expression of genes regulating the mitochondrial respiratory chain and the TCA cycle was not found in ire1Δ (Fig. S3) and erd2Δire1Δ cells (Fig. S4). RT–qPCR was conducted to assess the expression of representative genes atp1 (complex V), atp2 (complex V), cox13 (complex IV), and rip1 (complex III) in WT, erd2Δ, ire1Δ, and erd2Δire1Δ cells, confirming our RNA-Seq findings (Fig. 4B). These results suggest that the absence of Erd2 increases expression of genes regulating mitochondrial respiration and the TCA cycle in a UPR/Ire1-dependent manner.

Figure 4.

The absence of Erd2 increased expression of the genes involved in mitochondrial respiration and the tricarboxylic acid (TCA) cycle in an ire1-dependent manner.A, diagram illustrating glycolysis, the mitochondrial electron transport chain (ETC), and the TCA cycle. Indicated, in red, orange, blue, and green colors, are the genes that exhibit differential expression in erd2Δ cells (versus WT, adjusted p values [padj] < 0.05). The genes of the mitochondrial ETC and the ATP synthetase are listed. For the comparisons between erd2Δ and erd2Δire1Δ and WT, see Figs. S3 and S4. B, RT–quantitative PCR (qPCR) was conducted to analyze the expression of the representative ETC genes atp1, atp2, cox13, and rip1 in the indicated cells. Shown are fold changes of the tested genes in the indicated cells (versus WT). The number (n) of experimental replicates is indicated. The height of each bar represents the mean, and error bars indicate confidence intervals (alpha = 0.05). A single-group Student’s t test was used to calculate the p values. C, RT–qPCR was conducted to analyze the expression of the genes encoding the CCAAT-binding transcription factors (i.e., php2, php3, php4, and php5) in the indicated cells. Shown are fold changes of the tested genes in the indicated cells over a wildtype control. The number (n) of experimental replicates is indicated. The height of each bar represents the mean, and error bars indicate confidence intervals (alpha = 0.05). A single-group Student’s t test was used to calculate the p values. Tm indicates cells treated with tunicamycin to induce UPR. D, RT–qPCR was conducted to analyze the expression of the representative ETC genes atp1, atp2, cox13, and rip1 in the indicated cells. Shown are fold changes of the tested genes in the indicated cells (versus WT). The number (n) of experimental replicates is indicated. The height of each bar represents the mean, and error bars indicate confidence intervals (alpha = 0.05). A single-group Student’s t test was used to calculate the p values. Erd2, ER retention defective 2; UPR, unfolded protein response.

In the budding yeast S. cerevisiae, the HAP transcriptional complex, which binds to the CCAAT-binding site, promotes the expression of genes involved in the mitochondrial respiratory chain and the TCA cycle (33). Similarly, in the fission yeast S. pombe, the HAP component Php2 is involved in regulating expression of subunits in the mitochondrial respiratory chain (34). To test whether the absence of Erd2 alters the expression of the HAP complex (Php2–Php3–Php4–Php5) in fission yeast, we conducted RT–qPCR. Our results showed that the absence of Erd2 increased the expression of php2, php3, php4, and php5 (Fig. 4C). Induction of UPR by treatment of cells with 0.5 μg/ml tunicamycin also increased the expression of php2, php3, php4, and php5 (Fig. 4C). Moreover, increased expression of php2, php3, php4, and php5 was not found in ire1Δ or erd2Δire1Δ cells (Fig. 4C). We further tested whether the increase in expression of genes involved in the mitochondrial respiratory chain depend on the HAP complex by RT–qPCR (Fig. 4D). The result confirmed that the increase in expression of genes involved in the mitochondrial respiratory chain in cells lacking Erd2 depends on the HAP complex (Fig. 4D). The HAP complex also promotes expression of the genes involved in the TCA cycle (33). Consistently, our transcriptomic data showed that the absence of Erd2 increased expression of many genes involved in the TCA cycle (Fig. 4A). Collectively, these results suggest that the absence of Erd2 increases expression of genes encoding the CCAAT-binding transcription complex in an UPR/Ire1- and HAP-dependent manner.

Given that the absence of Erd2 induces UPR (Fig. 2) and increases mitochondrial respiration and ROS in an UPR/Ire1-dependent manner (Fig. 3), we propose a model presented in Figure 6. The absence of Erd2 induces UPR and consequently increases the expression of genes encoding the CCAAT-binding transcription complex. The increased expression of CCAAT-binding transcription complex then promotes expression of genes involved in the mitochondrial respiratory chain and the TCA cycle. As a result, mitochondrial respiration and ROS are enhanced in cells lacking Erd2.

Figure 6.

A model illustrating the alterations of the UPR signaling pathway and the responses of mitochondria by the absence of KDELR–Erd2. The absence of Erd2 impairs the retrograde transport of the xDEL-containing ER-resident proteins, inducing UPR. As a result, mitochondrial respiration and the production of mitochondrial ROS are enhanced by a transcriptional program that is controlled by UPR. Both UPR induction and the increases in mitochondrial respiration and mitochondrial ROS production are required for maintaining cell viability of cells lacking Erd2. ER, endoplasmic reticulum; Erd2, ER retention defective 2; KDELR, KDEL receptor; UPR, unfolded protein response.

Increases in mitochondrial respiration and ROS are required to maintain viability of cells lacking Erd2

To understand the physiological significance of the increased mitochondrial respiration and ROS in cells lacking Erd2, we assessed cell viability by testing colony formation (Fig. 5A). Specifically, cells at the exponential phase were treated with 1 mM Trolox (an antioxidant used to clean ROS), 0.15 μg/ml antimycin (an inhibitor of the mitochondrial respiratory chain), or the solvent dimethyl sulfoxide for 4 h or 8 h, followed by single-cell dissection on EMM plates with a tetrad-dissection microscope (Fig. 5A). Colony formation on the EMM plates was examined after 3 days. As shown in Figure 5, B and D, treatment of WT cells with Trolox did not affect cell viability, whereas treatment of erd2Δ cells with Trolox, for either 4 h or 8 h, impaired cell viability. Similar results were found when cells were treated with antimycin for either 4 h or 8 h (Fig. 5, C and E). Thus, in cells lacking Erd2, the increases in mitochondrial respiration and ROS are required for maintaining cell viability.

Figure 5.

Increases in mitochondrial respiration and mitochondrial ROS production were required for maintaining viability of erd2Δ cells.A, diagram illustrating the procedure for assaying cell viability by tetrad-dissection analysis. The capability of forming a colony from a single cell after treatment with the indicated chemicals was examined on EMM plates. B and C, single-cell colony arrays. WT and erd2Δ cells were treated with 1 mM Trolox (B), 0.15 μg/ml antimycin (C), or DMSO (solvent) for 4 h or 8 h. Cells without treatment are controls. Dashed circles indicated the absence of colonies, that is, cell death. Note that for the convenient of comparison, the DMSO control groups are identical and shown in both (B) and (C). D and E, quantification of the percentage of cell death (the absence of colonies). The top of the column is the mean, and error bars indicate SD. Circles indicate the value measured from each experiment, and five independent experiments were performed. Student’s t test was used to calculate the p values. Note that the DMSO control groups are identical and shown in both (D) and (E). DMSO, dimethyl sulfoxide; EMM, Edinburgh minimal medium; ROS, reactive oxygen species.

Discussion

The KDELR–Erd2 retrieval system is involved in maintaining protein homeostasis in the ER by retrieving the xDEL-containing ER-resident proteins that escape from the ER (2, 3). In this study, we report that Erd2 plays a crucial role in modulating mitochondrial function by guiding the UPR signaling pathway (Fig. 6).

Erd2 collaborates with Ire1, the ER transmembrane sensor responsible for activating UPR, to maintain protein homeostasis in the ER of the budding yeast S. cerevisiae (19). However, it had not been directly tested whether the absence of Erd2 induces UPR. In our study, we provided evidence that UPR is induced by the absence of Erd2 (Fig. 2A). The induction of UPR by the absence of Erd2 is likely because of the defective retention of the ER-stress response chaperones GRP78/Bip within the ER because GRP78/Bip is a typical cargo protein of KDELR–Erd2 (2, 4, 35) and functions to inhibit Ire1 in the ER (36). In addition to GRP78/Bip, other cargo proteins of KDELR–Erd2, including protein disulfide isomerases, are required for promoting protein folding in the ER (37). Therefore, it is also possible that the induction of UPR by the absence of Erd2 is due to the defective retention of the ER-resident proteins required for promoting protein folding. In mammalian cells, expression of the KDELR mutant that fails to bind cargos enhances the responses of UPR induced with tunicamycin or dithiothreitol (24). Hence, induction of UPR by malfunction of the KDELR–Erd2 retrieval system may be an evolutionarily conserved mechanism.

Our study also establishes that Erd2 regulates mitochondrial function in an UPR-dependent manner. Specifically, in cells lacking Erd2, the induction of UPR elevates the expression of genes involved in the mitochondrial electron transport chain and the mitochondrial ATP synthase (Figs. 4, A and B and S2–S4). The elevation in gene expression is due to the increased expression of the transcription HAP complex. The reasons are as follows. It has been established that the transcriptional HAP complex binds to the CCAAT box within gene promoters (38) and promotes expression of genes involved in the mitochondrial electron transport and the TCA cycle (33, 34). Our findings revealed that the absence of Erd2 increased, in an Ire1/UPR-dependent manner, expression of the transcription HAP complex (Fig. 4C). Moreover, tunicamycin, the UPR inducer, promotes expression of the transcription HAP complex (Fig. 4C). Notably, the absence of Erd2 induces UPR (Fig. 2A). Therefore, our results support a model in which Erd2 alters mitochondrial function by guiding the expression of the UPR-promoted genes involved in the mitochondrial electron transport chain and the mitochondrial ATP synthase (Fig. 6).

Consistent with the finding that the absence of Erd2 increased expression of genes involved in mitochondrial electron transport chain and the mitochondrial ATP synthase, both the absence of Erd2 and the chemical induction of UPR by tunicamycin increased mitochondrial respiration and mitochondrial ROS (Figs. 1, C–E and 2, D–F). Therefore, these results further support the model in which activation of UPR boosts mitochondrial function and Erd2 regulates mitochondrial function via UPR.

Our study revealed that the UPR-induced increases in mitochondrial respiration and ROS in cells lacking Erd2 are required for maintaining cell viability (Fig. 5). Given the role of the KDELR–Erd2 retrieval system in maintaining protein homeostasis in the ER, we hypothesized that the absence of Erd2 may increase energy demands required for restoring protein homeostasis. To meet these energy demands, mitochondria functions may be enhanced. This hypothesis predicts that defects in promoting mitochondrial functions are detrimental to cells losing the KDELR–Erd2 retrieval system. Indeed, clearance of mitochondrial ROS (Fig. 5, B and D) or inhibition of mitochondrial respiration (Fig. 5, C and E) in cells lacking Erd2 caused cell death. Since the absence of Erd2 increases mitochondrial respiration and ROS in an UPR-dependent manner, it is also predicted that defects in UPR should impair cell growth. Consistent with this prediction, under unstressed conditions, erd2Δire1Δ cells, in which both mitochondrial respiration and mitochondrial ROS were diminished (Fig. 3, C–F), grew poorly compared with WT and erd2Δ (Fig. 2B).

Since Erd2 resides mainly at the Golgi apparatus (31), the Golgi apparatus was also examined in cells lacking Erd2. The results showed that the localization of the cis-Golgi protein Anp1-GFP to Golgi was impaired by the absence of Erd2 (Fig. S5A), whereas the localization of the trans-Golgi protein Sec72-GFP to Golgi was enhanced (Fig. S5B). However, the expression levels of Anp1-GFP and Sec72-GFP in WT and erd2Δ cells were comparable (Fig. S5, C–F). These findings indicate that the absence of Erd2 may affect the structure/organization of the Golgi apparatus and/or the proteins residing at the Golgi apparatus. This awaits further characterization.

In conclusion, our study revealed a previously uncharacterized role of KDELR–Erd2 in guiding the signaling pathway of UPR to maintain cellular homeostasis. Since KDELR–Erd2 have pathophysiological significance, our work provides insights into developing strategies for intervening in physiological disorders caused by malfunction of KDELR–Erd2.

Experimental procedures

Yeast strains

Yeast strains were created either by random spore digestion or tetrad dissection. To create strains carrying deletion of a gene, the PCR-based homologous recombination method using the pFA6a series of plasmids was employed. Yeast strains used in this study are listed in Table S1.

Yeast cell culture

Yeast cells were cultured in EMM supplemented with 2% glucose and five amino acids (adenine, leucine, uracil, histidine, and lysine, 0.225 g/l each) (referred to as EMM) at 30 °C, and cells at the exponential phase were collected for analysis. All culture media were purchased from Formedium (www.formedium.com).

Measurement of the rate of oxygen consumption

The rate of oxygen consumption was measured with a Strathkelvin respirometer (model 782) (www.strathkelvin.com), according to the manufacturer’s instructions. First, yeast cells were cultured in EMM until the exponential phase (absorbance at 600 nm = 0.6–0.8) was reached. The measurements were then made at room temperature. Antimycin A (www.enzolifesciences.com) at a working concentration of 0.15 μg/ml was used to block mitochondrial respiration, which was performed in parallel and served as a control. Three independent experiments were performed.

Measurement of mitochondrial ROS

To visualize and quantify mitochondrial ROS of fission yeast cells, MitoSOX Red (www.invitrogen.com) was used at a working concentration of 5 μM. First, yeast cells were cultured in EMM until the exponential phase (absorbance at 600 nm = 0.6–0.8) was reached; MitoSOX was then added to the culture, followed by culture in the dark at 30 °C for 30 min. In the case where ER stress was induced, tunicamycin (www.casmart.com.cn) at a working concentration of 0.5 μg/ml was added, and the tunicamycin-treated cells were cultured for 1 h. To measure mitochondrial ROS of these cells, MitoSOX was added to the culture at 30 min before the end of tunicamycin treatment. Finally, yeast cells were collected and washed with fresh EMM, followed by observation with a spinning-disk microscope. Three independent experiments were performed.

Serial dilution growth assays

Yeast cells were cultured in liquid EMM until the exponential phase (absorbance at 600 nm = 0.6–0.8) was reached. Serial dilutions (10-fold) of yeast cells were made, and then diluted cells were spotted onto solid EMM plates or EMM plates containing 0.5 μg/ml tunicamycin. The plates were incubated at 30 °C for 3 days. Three independent repeats were performed.

Cell viability test

Cell viability was determined by tetrad analysis of single cells. Specifically, yeast cells were cultured in EMM until the exponential phase (absorbance at 600 nm = 0.6–0.8) was reached. Before the drug treatment was conducted, yeast cells were collected, and single cells were separated and aligned on EMM plates (i.e., experiment at 0 h). To clear ROS, Trolox at a working concentration of 1 mM was used to treat yeast cells for 4 h or 8 h at 30 °C. To inhibit mitochondrial respiration, antimycin A at a working concentration of 0.15 μg/ml was used to treat yeast cells for 4 h or 8 h at 30 °C. After the treatments, single yeast cells were separated and grown on EMM plates by tetrad analysis. All plates were then incubated at 30 °C for 3 days. Five independent experiments were performed.

RNA extraction and real-time qPCR

Yeast cells (1–5 × 10⁷) from 1 to 1.5 ml cultures were collected, and total RNA was extracted with a genomic DNA purification kit (R1002; www.zymoresearch.com). RT–PCR with a kit (HiScript III RT SuperMix; H6201080; www.yeasen.com) was then conducted to generate complementary DNAs (cDNAs).

Real-time qPCR was accomplished with the primers listed in Table S2, using a kit purchased from YEASEN (Vazyme Univer Blue qPCR SYBR Master Mix, www.yeasen.com) on a Roche instrument (LightCycle 96, www.lifescience.roche.com). Five independent repeats were performed.

RNA-Seq and analysis

RNA was extracted as described previously. The concentration of RNA was determined by the NanoDrop (www.thermofisher.com) procedure. Three replicates were prepared for each type of strains. A total amount of 2 μg RNA per sample was used, and RNA-Seq was accomplished by Novogene. Extracted RNAs were fragmented and then used for generating cDNAs with random primers. Subsequently, PCR was performed to amplify cDNAs with adaptor primers. An Illumina HiSeq instrument was used for sequencing.

Raw RNA-Seq data were first processed by Trimmomatic (version 0.39) (39) to trim off adaptors and eliminate reads that were in low quality. The cleaned RNA-Seq data were aligned to the S. pombe reference genome GCF_000002945.1_ASM294v2.fa (40) by HISAT2 (version 2.2.0, http://www.ccb.jhu.edu/software/hisat/) (41). Gene expression levels were calculated by use of HTSeq (http://www-huber.embl.de/HTSeq) (42), with gene models obtained from the reference file (GCF_000002945.1_ASM294v2_genomic.gff) derived from the National Center for Biotechnology Information. Afterward, analysis of differential gene expression was performed using a DESeq2 (version 1.26.0) package in R (see Supporting excel file). GO enrichment analysis on the differentially expressed genes identified by DESeq2 were conducted using Annotation Hub package (version 1.26.0) in R. In RNA-Seq, three independent replicates were performed for each type of strains.

Western blotting analysis

Cell lysis were prepared, following the method described previously (43). Protein samples were analyzed by SDS-PAGE and Western blotting with antibodies against GFP: anti-GFP (Rockland Immunochemicals; catalog no.: 600-101-215; 1:3000 dilution) and antitubulin (BioAcademia; catalog no.: 63-160; 1:10,000 dilution). The secondary antibodies used in this study are anti–goat IgG (Abclonal; catalog no.: AS029; 1:10,000 dilution) and anti–rabbit IgG (Bio-Rad; catalog no.: 170-546, 1:10,000 dilution).

Microscopy and data analysis

Live-cell imaging was accomplished using a PerkinElmer Ultraview spinning disk microscope equipped with a Nikon Apochromat TIRF 100X 1.49 numerical aperture objective and a Hamamatsu C9100-23B EMCCD camera, and 405, 488, and 561 nm lasers were used in the experiments (www.perkinelmer.com). Images were acquired by Volocity (www.perkinelmer.com). To observe yeast cells, stack images containing 11 planes at 0.5 μm spacing were acquired. Yeast cells were observed on slides coated with EMM agarose (3%) pads.

Microscopic data were analyzed with Fiji ImageJ 1.52 (www.imagej.nih.gov) and MetaMorph 7.7 (www.moleculardevices.com). Graphs were created by KaleidaGraph 4.5 (www.synergy.com). Normality of data was determined by OriginPro (version 2021b) (www.originlab.com), and statistical analysis was performed by KaleidaGraph 4.5. The illustration graph in Figure 6 was created using graphs generated with BioRender.com.

Data availability

All data are contained within the article.

Supporting information

This article contains supporting information (44).

Conflict of interest

The authors declare that they have no conflicts of interest with the contents of this article.

Reviewed by members of the JBC Editorial Board. Edited by Phyllis Hanson