Expression of a Truncated Form of the Endoplasmic Reticulum Chaperone Protein, σ1 Receptor, Promotes Mitochondrial Energy Depletion and Apoptosis

Background: An ER-associated chaperone protein, σ1 receptor (σ₁R), regulates ER/mitochondrial Ca²⁺ mobilization through the IP₃ receptor.

Results: We identify a novel short splicing variant of σ₁R, termed σ₁SR, and demonstrate its dominant negative function.

Conclusion: σ₁SR interferes with σ₁R function in mitochondrial Ca²⁺ mobilization and ATP production under ER stress conditions.

Significance: In contrast to σ₁R function, σ₁SR has detrimental effects on cell survival.

Abstract

The σ1 receptor (σ₁R) regulates endoplasmic reticulum (ER)/mitochondrial interorganellar Ca²⁺ mobilization through the inositol 1,4,5-trisphosphate receptor (IP₃R). Here, we observed that expression of a novel splice variant of σ₁R, termed short form σ₁R (σ₁SR), has a detrimental effect on mitochondrial energy production and cell survival. σ₁SR mRNA lacks 47 ribonucleotides encoding exon 2, resulting in a frameshift and formation of a truncated receptor. σ₁SR localizes primarily in the ER at perinuclear regions and forms a complex with σ₁R but not with IP₃R in the mitochondrion-associated ER membrane. Overexpression of both σ₁R and the truncated isoform promotes mitochondrial elongation with increased ER mitochondrial contact surface. σ₁R overexpression increases the efficiency of mitochondrial Ca²⁺ uptake in response to IP₃R-driven stimuli, whereas σ₁SR overexpression reduces it. Most importantly, σ₁R promotes ATP production via increased mitochondrial Ca²⁺ uptake, promoting cell survival in the presence of ER stress. By contrast, σ₁SR suppresses ATP production following ER stress, enhancing cell death. Taken together, the newly identified σ₁SR isoform interferes with σ₁R function relevant to mitochondrial energy production under ER stress conditions, promoting cellular apoptosis.

Introduction

Endoplasmic reticulum (ER)²/mitochondrial Ca²⁺ transport contributes to many cellular processes, including ATP generation and cell survival (1, 2). ER inositol 1,4,5-trisphosphate receptors (IP₃Rs) are localized in the mitochondrion-associated ER membrane (MAM) (3, 4), where the IP₃R plays critical roles in mitochondrial Ca²⁺ transport. Ca²⁺ overload through IP₃R promotes apoptosis under pathological conditions (5). By contrast, mitochondrial Ca²⁺ uptake through IP₃R is crucial for basal mitochondrial ATP production required to maintain normal cellular biogenesis (6).

Mitochondrial Ca²⁺ uptake derived from IP₃R-mediated Ca²⁺ release is facilitated by interaction between IP₃R and voltage-dependent anion channels in the MAM (4, 7, 8), and several IP₃R-binding molecular chaperones regulate mitochondrial Ca²⁺ influx from the ER (9). Among them, the σ1 receptor (σ₁R), which was cloned by Hanner et al. (10), was recently identified as an ER-associated chaperone protein (11). σ₁R can also translocate from the ER to the MAM or plasma membrane to modulate diverse cellular activities, including lipid metabolism (12) and N-methyl-d-aspartate receptor activity (13). Functional analyses in Chinese hamster ovary (CHO) cells reveal that the σ₁R stabilizes the conformation of the MAM-associated IP₃R type-3 in the ER, positively regulating Ca²⁺ influx into mitochondria. In addition, σ₁R knockdown in CHO cells facilitates ER stress-induced cell death (11).

Ca²⁺ transport between the ER and mitochondria plays important roles in neurodegenerative diseases, such as Alzheimer, Parkinson, and Huntington disease (14). Indeed, σ₁R ligands are considered therapeutic targets for several psychiatric and neurodegenerative diseases (15–17). However, how these ligands mediate neuroprotective effects remains unclear. More recently, σ₁R mutations have been observed in patients with dementia attributable to frontotemporal lobar degeneration (18) or juvenile amyotrophic lateral sclerosis (19). Interestingly, the latter apparently promotes nuclear transport of mutant forms of σ₁R.

In this study, we observed a novel σ₁R splicing variant in mouse brain, namely a short form (σ₁SR) lacking 47 bp of exon 2. σ₁SR protein is primarily expressed in the ER, where it interacts with σ₁R. Interestingly, σ₁SR overexpression decreased mitochondrial Ca²⁺ uptake in response to IP₃R-mediated stimulation, indicating that it antagonizes σ₁R activity. Moreover, σ₁SR overexpression promoted autophagic apoptosis, consistent with IP₃R destabilization and decreased ATP production attributable to reduced mitochondrial Ca²⁺ uptake. Our results show for the first time that σ₁SR interferes with mitochondrial ATP biogenesis by inhibiting σ₁R function following ER stress.

EXPERIMENTAL PROCEDURES

Sequencing and Construction of Expression Vectors Encoding σ₁R Isoforms

Total RNAs were prepared from mouse brain hippocampus using TRIzol LS reagent (Invitrogen) according to the manufacturer's protocol. mRNA was reverse-transcribed into single-stranded cDNA using an oligo(dT) primer (Promega, Madison, WI) and Moloney murine leukemia virus-reverse transcriptase (Invitrogen). DNA sequences of σ₁R isoforms amplified by PCR using specific 5′- and 3′-primers and primer sequences are shown in supplemental Table 1. All DNA sequences were determined at the Fasmac DNA Sequence Service (FASMAC Co., Ltd., Atsugi, Japan). To construct expression vectors, PCR amplification products were digested with XhoI and BamHI and ligated with purified XhoI- and BamHI-digested pmCherry-N1 or pEGFP-N1 vector (Clontech) in the sense orientation.

Cell Culture, Transfection, and Production of Stably Transfected Cell Lines

Neuro-2a cells from a mouse neuroblastoma C1300 tumor were obtained from the Human Science Research Resources Bank (IFO50081) (Osaka, Japan). Neuro-2a cells were grown in Dulbecco's minimal essential medium (DMEM) supplemented with 10% heat-inactivated fetal bovine serum (FBS) and penicillin/streptomycin (100 units/100 μg/ml) in a 5% CO₂ incubator at 37 °C. Cells were transfected with expression vectors using Lipofectamine 2000 (Invitrogen), and experiments were performed 48 h later as described (20). σ₁R small interfering (si)RNA (sense, 5′-ACACGTGGATGGTGGAGTA-3′, and antisense, 5′-TACTCCACCATCCACGTGT-3′) was purchased from Exigen (Tokyo, Japan). Transfections were performed with 100 nm σ₁Rs siRNA according to the methods of Ref. 21. Stably transfected Neuro-2a cells were prepared as described (22). Briefly, cells were transfected with σ₁R or σ₁SR cDNA in pmCherry-N1 expression vectors as described above. Transfected cells were plated in medium containing 1000 μg G418/ml, and G418-resistant colonies were isolated.

Immunoprecipitation and Immunoblotting

Immunoprecipitation and immunoblotting analysis were performed as described (20). 12-Week-old male C57BL/6J mouse brains were immediately removed from euthanized mice and perfused in ice-cold buffer for 3 min (0.32 m sucrose, 20 mm Tris-HCl, pH 7.4), and olfactory bulb, cortex, hippocampus, striatum, and brainstem were dissected. These samples and Neuro-2a cells were homogenized in buffer containing 50 mm Tris-HCl, pH 7.4, 0.5% Triton X-100, 0.5 m NaCl, 4 mm EDTA, 4 mm EGTA, 1 mm Na₃VO₄, 50 mm NaF, 1 mm DTT, 2 μg/ml pepstatin A, and 1 μg/ml leupeptin, then treated with Laemmli sample solution, and boiled for 3 min. Antibodies included the following: rabbit polyclonal antibodies against the N-terminal cytosolic domain (52–69 amino acids) (1:1000) and C-terminal lumenal domain (143–165 amino acids) of σ₁R (1:1000; a kind gift of Dr. Teruo Hayashi, NIDA, Baltimore); GFP (1:2000; Clontech); calcineurin (1:1000) (23); voltage-dependent anion channel (1:1000; Cell Signaling Technology, Beverly, MA); CREB-2 (1:200; Santa Cruz Biotechnology, Santa Cruz, CA); LC-3 (PM036, 1:1000; MBL, Nagoya, Japan); rabbit monoclonal antibody against phospho-PERK (1:1000; Cell Signaling Technology); mouse monoclonal antibodies against C/EBP homologous protein (1:200; Santa Cruz Biotechnology); β-tubulin (1:1000; Sigma); and pan-IP₃R (1:500; Millipore, Bedford, MA). For subcellular fractionation, cells were washed twice with ice-cold phosphate-buffered saline (PBS), separated into cytosolic, ER, and mitochondrial membranes and nuclear fractions using the subcellular protein fractionation kit (Thermo Fisher Scientific Inc., Waltham, MA) following the manufacturer's protocol, and analyzed by Western blot analysis.

Immunohistochemistry

Fluorescence immunohistochemical studies were performed as described (24). Cells were fixed in 4% paraformaldehyde in phosphate buffer for 30 min at room temperature and washed in PBS. Cells were blocked with 3% bovine serum albumin in PBS for 1 h. First antibodies included mouse monoclonal antibodies against GM130 (1:1000; BD Biosciences) and rabbit polyclonal antibody against LC-3 (PM036, 1:1000; MBL). After thorough washes in PBS, sections were incubated with secondary antibodies in blocking solution at 20 °C for 24 h. Antibodies included Alexa 594-labeled anti-mouse IgG and Alexa 448-labeled anti-rabbit IgG (1:500; Invitrogen) in blocking solution at 20 °C for 3 h. After several washes in PBS, sections were mounted on slides with Vectashield (Vector Laboratories Inc., Burlingame, CA). For nuclear staining, sections were incubated with DAPI (Vector Laboratories, Burlingame, CA). Immunofluorescent images were analyzed using a confocal laser scanning microscope (LSM700; Zeiss, Thornwood, NY). To detect ER and mitochondrial distribution, Neuro-2a cells were transfected with erRFP or mitochondrial GFP (mtGFP) (Bacman 2.0, Invitrogen) and analyzed using confocal microscopy of fixed cells.

Morphometric and Contact Analysis

Morphometric and contact analysis was performed as described previously (25) with modifications. For morphometric analysis of mitochondria, the major axis length of each identified object was calculated. Cells were scored as showing elongated mitochondria when >50% of the objects in the image displayed a major axis longer than 3 μm. For morphometric analysis of ER, major axis length and elongation index of each identified object were calculated. Cells were scored as showing reticular ER when the major axis was longer than 5 μm, and the elongation index exceeded 4 in more than 50% of the identified objects. For mitochondrial network and ER co-localization, one focal plane was analyzed. Images were deconvoluted, and background was subtracted using ImageJ software. Co-localization of organelles was quantified using the co-localization coefficient of Manders et al. (26).

Quantification of Neurite Sprouting

Neurite sprouting by Neuro-2a cells was assessed by staining with βIII-tubulin (1:1000, Promega, Madison, WI) and use of the VECTASTAIN ABC kit (Vector Laboratories, Burlingame, CA), according to the manufacturer's protocol. Images were acquired using the 40× objective of a microscope (BX51WI, Olympus, Tokyo, Japan) equipped with a digital camera (Micropublisher 5.0, QIMAGING, Burnaby, British Columbia, Canada). Sprouting was quantified as the percentage of cells with neurites longer than twice the body diameter. Neurite length was determined using ImageJ software. Six fields (100 cells per field) in each condition were chosen randomly and photographed.

Fluorometric [Ca²⁺]_c and [Ca²⁺]_mt Measurements

Neuro-2a cells were grown on 0.01% poly-l-lysine (Sigma)-coated glass-bottom dishes. For [Ca²⁺]_c measurement, cells were rinsed in Krebs buffer (135 mm NaCl, 6 mm KCl, 1.2 mm MgCl₂, 12 mm glucose, 1.5 mm CaCl₂, 12 mm HEPES, pH 7.3), loaded with 2.5 μm Fura-2/AM (Sigma) for 15 min at 37 °C in darkness, and washed with Krebs buffer for 15 min. Measurements were performed in Ca²⁺-free Krebs buffer including 2 mm EGTA. Ratio measurements were performed every 3 s by excitation at 340 and 380 nm and recording of the emission at 530 nm. Ratio values were derived by averaging fluorescence intensity from the entire cytosolic area. To measure [Ca²⁺]_mt, cells were transfected with the ratiometric pericam (a kind gift of Dr. Atsushi Miyawaki, RIKEN Brain Science Institute, Wako-City, Japan) targeted to the mitochondrial matrix. Two days later, cells were exposed to externally applied 10 μm ATP in Ca²⁺-free Krebs buffer and imaged. Cells were permeabilized by exposure to intracellular like medium (ICM) (125 mm KCl, 19 mm NaCl, 10 mm HEPES-KOH, pH 7.3, 1 mm EGTA) and appropriate concentrations of CaCl₂ (330 μm CaCl₂ for 50 nm free Ca²⁺) (27) containing 250 μg/ml (w/v) saponin (MP Biomedicals, Ohio) for 3 min. Permeabilized cells were washed with ICM and then exposed to ICM containing 1 μm inositol 1,4,5-trisphosphate (IP₃) (Biomol, Plymouth Meeting, PA). Dual excitation imaging with ratiometric pericam-mt required two filters (excitation 482/35, dichroic mirror 506, emission 536/40 and excitation 414/46, dichroic mirror 510, emission 527/20). [Ca²⁺]_c and [Ca²⁺]_mt, which were monitored during the assay on an inverted microscope (Leica DM IRB, Japan), were equipped with CCD cameras (ORCA-ER; Hamamatsu, Japan). Captured images were analyzed using the Metafluor imaging system (Molecular Devices, Sunnyvale, CA).

Drug Treatments

ER stress was induced by treatment of cells with 2 μg/ml tunicamycin (Sigma) for 4 or 24 h. For cell viability experiments, tunicamycin treatment continued for 48 h. Mitochondrial Ca²⁺ uptake was inhibited by 10 μm Ru360 (Calbiochem).

ATP Measurement

Cells were plated in 6-cm plates, and ATP content was determined using luciferin and a luciferase assay kit (Toyo B-net, Tokyo, Japan) following the manufacturer's protocol.

TUNEL Staining

DNA fragmentation and apoptotic bodies were detected by the TUNEL method using an in situ apoptosis detection kit (Takara Bio Inc., Shiga, Japan), as described (28). One hundred cells from 13 randomly selected fields were counted in each experiment.

Quantification of mRNA by Real Time PCR

Real time PCR analysis was performed as described (24) in 48-well plates (Mini Opticon Real Time PCR System, Bio-Rad) using iQ SYBR Green Supermix 2× (Bio-Rad). Primer sequences are shown in supplemental Table 1. Relative quantities of target mRNAs were determined by the comparative threshold cycle (ΔCT) method and normalized to GAPDH quantity. Product purity and specificity were confirmed by omitting the template and performing a standard melting curve analysis.

Statistical Evaluation

All values were expressed as means ± S.E. Comparison between two experimental groups was made using the unpaired Student's t test. Statistical significance for differences among groups was tested by one-way analysis of variance, followed by multiple comparisons between control and other groups using Dunnett's multiple comparison test. p < 0.05 was considered significant.

RESULTS

Identification of a Novel σ₁R Splicing Variant

σ₁R, a nonopioid receptor (29, 30), was previously cloned from guinea pigs (10), humans (31), and mice (32). All forms showed greater than 80% amino acid homology but no structural homology with any other receptor family. We screened a mouse hippocampal cDNA library by PCR using primer sets derived from the coding region of mouse Sigmar1 (Fig. 1A, p1 and p2) and identified the Sigmar1-coding sequence fragment (Fig. 1B, lane 1) and a novel smaller fragment (Fig. 1B, lane 2). When we amplified what would be an ∼200-bp fragment spanning σ₁R exon 2 to exon 3 (Fig. 1A, p3 and p4) from the Sigmar1-coding sequence fragment, we detected the expected 200-bp fragment (Fig. 1B, lane 3) and also an ∼150-bp fragment (Fig. 1B, lane 4). Sequencing analysis revealed that the short fragment lacks 47 bp from exon 2 of the full-length gene. We subsequently identified a novel 231-bp splicing variant of σ₁R and called it σ₁SR (Fig. 1B, lane 5). A human sequence corresponding to mouse σ₁SR had been deposited in GenBank^TM under accession number BC007839.2. (Fig. 1C, left), but protein expression and function had not been evaluated. Our analysis indicated that the novel splice conformed to the G(T/A)G rule (Fig 1D) and that splicing of the 47-bp fragment caused a frameshift starting at amino acid 103 and translation termination due to a newly generated stop codon. The predicted primary structure of σ₁SR contains 106 rather than the 223 amino acids seen in σ₁R. The first 102 amino acids of both proteins are identical, but the C-terminal four residues of σ₁SR differ from the longer form, because they are derived from translation of the frame-shifted σ₁R exon 3 (Fig. 1, C and D).

FIGURE 1.

σ₁R alternative splicing. A, position of PCR primers relative to receptor sequences (transmembrane domain). B, PCR of an adult mouse brain hippocampus cDNA library was performed using the indicated oligonucleotide primers. C, sequence of coding region of human SIGMAR1 (left) and mouse Sigmar1 (right) cDNA. The deduced amino acid sequence is presented using the one-letter code. Differences between homologues are shown in bold. Two putative transmembrane domains (TM1 and TM2) are shown in gray underlines, and the 47-bp of sequences deleted from full-length σ₁R are shown in black underlines. Their deletion results in a frameshift, giving rise to four novel amino acids in σ₁SR (boxed). D, proposed splicing mechanism of σ₁Rs. Exons are outlined, and coding sequence is denoted by capital letters. Intronic sequence is denoted by lowercase letters.

Tissue Distribution and Localization of Both σ₁R Isoforms

To investigate tissue distribution and relative expression levels of σ₁R and σ₁SR, we performed immunoblot analysis in adult mouse brain using antibodies against the σ₁R N-terminal cytosolic domain (amino acids 52–69) or the C-terminal lumenal domain (amino acids 143–165). To confirm endogenous expression of σ₁R and σ₁SR, we observed immunoreactive bands in cell extracts from neuroblastoma Neuro-2a cells transfected with both cDNAs (Fig. 2A, lanes 3–8). Using an N-terminal cytosolic domain-specific antibody, we observed marked variation in the ratio of endogenous σ₁SR (seen as a 12-kDa band) to σ₁R (seen as a 26-kDa band) protein in the mouse brain regions tested (Fig. 2A, upper panel, lanes 9–13). An unidentified protein of the 19-kDa protein was observed in the hippocampus and brainstem lysates. As expected, a C-terminal lumenal domain antibody detected an immunoreactive σ₁R band but not σ₁SR (Fig. 2A, lower panel, lanes 9–13). σ₁SR protein levels were high in cortex, hippocampus, and striatum (σ₁SR/σ₁R ratio of 0.4) but barely detectable in olfactory bulb and brainstem (Fig. 2B). To further investigate differences in subcellular localization, we fractionated lysates of Neuro-2a cells transfected with 3′ eGFP-tagged σ₁R (σ₁R-eGFP) or σ₁SR (σ₁SR-eGFP) cDNA into cytosolic, ER, and mitochondrial membrane (including MAM) and nuclear fractions. High levels of σ₁R-eGFP were found in ER and mitochondrial membrane fractions, and relatively low levels were seen in other fractions, whereas σ₁SR-eGFP was detected in membrane and nuclear fractions (Fig. 2C). To determine subcellular localization of these proteins, we undertook confocal microscopy of fluorescence-tagged σ₁R and σ₁SR in Neuro-2a cells. σ₁R-eGFP and σ₁SR-eGFP were primarily observed in perinuclear regions, although low levels of fluorescence were seen in the nuclei of σ₁SR-eGFP-expressing cells (Fig. 2D). A punctate perinuclear staining pattern of erRFP, which targets and serves as a marker of endoplasmic reticulum, largely co-localized with σ₁R-eGFP (Fig. 2E) and σ₁SR-eGFP (Fig. 2F). Interestingly, the Golgi marker GM130 showed considerable co-localization with σ₁SR-eGFP (Fig. 2H) but not with σ₁R-eGFP (Fig. 2G). Both mCherry-tagged σ₁R isoforms were detected in some overlays with the mitochondrial marker mtGFP, suggesting that σ₁SR protein also localizes on MAMs, as reported for σ₁R (Fig. 2, I and J) (11). These results suggest that the subcellular distribution of both σ₁R isoforms is similar, although σ₁SR is also seen in the Golgi apparatus and nucleus.

FIGURE 2.

Tissue distribution and localization of σ₁Rs. A, representative immunoblots probed with antibodies against the σ₁R N-terminal cytosolic domain (upper panel) or the C-terminal lumenal domain (lower panel). As control bands, extracts from Neuro-2a cells transfected with σ₁R and σ₁SR constructs (lanes 1–8) are shown. B, quantitative densitometry shows the ratio of σ₁SR to σ₁R protein expression among different brain regions. OB, olfactory bulb; CX, cortex; HP, hippocampus; ST, striatum; BS, brainstem. n = 4 in each group. C, cytosolic (Cyto), ER, mitochondrial membrane (Mem), and nuclear (Nuc) fractions from eGFP tagged-σ₁R isoform-transfected Neuro-2a cells were blotted with antibodies against GFP, calcineurin (CaN, cytosolic marker), voltage-dependent anion channel (ER and mitochondrial membrane marker), and CREB-2 (nuclear marker). D–I, σ₁R localization in Neuro-2a cells. Confocal images show co-localization of fluorescence (eGFP or mCherry)-tagged-σ₁R isoforms and markers of ER (erRFP), Golgi apparatus (GM130), and mitochondria (mtGFP). D, σ₁R-eGFP (left) and σ₁SR-eGFP (right) are mainly expressed in perinuclear regions. A small number of GFP aggregates are detected in nuclei of σ₁SR-eGFP-expressing cells (shown by arrow). Right panels are high magnification images. E and F, immunoreactivity of both σ₁R-eGFP isoforms (green) and erRFP (red) almost completely merge. G and H, GM130 (green) immunoreactivity co-localizes with σ₁SR-mCherry (H) (red) but not with σ₁R-mCherry (G) (red). I and J, both mCherry-tagged σ₁R isoforms (red) were detected in a few overlays with the mitochondrial marker mtGFP (green). Lower panels in I and J are high magnification images. Scale bars, 10 μm.

Overexpression of Either σ₁R Isoform Promotes Mitochondrial Elongation and Increases ER/Mitochondrial Contact Space

Based on their subcellular location, we analyzed the role of both σ₁R isoforms in ER and mitochondrial morphology and in ER/mitochondria interaction. σ₁R or σ₁SR overexpression did not alter the reticular ER structure of erRFP-infected Neuro-2a cells (Fig. 3B; 89.1 ± 2.8% in control, 88.7 ± 2.9% in σ₁R-, and 83.6 ± 3.6% in σ₁SR-transfected cells). However, expression of either isoform alone promoted mitochondrial elongation, as shown by three-dimensional reconstruction of mitochondrion-targeted green fluorescent protein (Fig. 3C; 41.6 ± 7.5% in control, 77.4 ± 7.0% in σ₁R-, and 67.3 ± 8.9% in σ₁SR-transfected cells). In addition, confocal semiquantitative analysis of ER/mitochondria juxtaposition showed that tethering of ER and mitochondria increased either isoform expressed in Neuro-2a cells (Fig. 3D; 2.6 ± 0.3% in controls, 4.6 ± 0.9% in σ₁R-, and 3.8 ± 0.7% in σ₁SR-transfected cells).

FIGURE 3.

Both σ₁R isoforms regulate mitochondrial morphology and juxtaposition of the ER to mitochondria. A, representative images of three-dimensional reconstructions of ER (erRFP; left), mitochondria (mtGFP; middle), and merged images showing ER and mitochondria co-localization (yellow; right). B, morphometric analysis of ER shape in cells transfected with erRFP (n = 3, 18 cells per experiment). C, quantitative analysis of mitochondrial shape in cells transfected with mtGFP (n = 4, 12 cells per experiment). D, quantitative analysis of co-localization of ER and mitochondria (as a percentage of total mitochondrial volume) in cells co-transfected with erRFP and mtGFP (n = 4, 20 cells per experiment). Each bar represents the mean ± S.E. *, p < 0.05; **, p < 0.01 versus control cells. Scale bar, 10 μm.

σ₁SR Forms a Complex with σ₁R but Not with IP₃Rs

Because both σ₁R isoforms are similarly distributed, we asked whether either directly interacted with IP₃Rs. To do so, we performed immunoprecipitation of extracts from σ₁SR-mCherry-expressing cells co-expressing σ₁R-eGFP using an eGFP antibody followed by immunoblotting with an anti-σ₁R N-terminal cytosolic domain-specific antibody. We observed an immunoreactive σ₁SR-mCherry band (41 kDa) (Fig. 4A, upper panel, lane 9) not seen in immunoprecipitates from extracts from eGFP-, σ₁R-eGFP-, or σ₁SR-mCherry-overexpressing cells (Fig. 4A, upper panel, lanes 6–8). Immunoreactive bands with molecular masses corresponding to two fluorescence-tagged σ₁R isoforms of 55 and 41 kDa were detected in each cell extract (Fig. 4A, lanes 2–4). To confirm the specificity of fluorescence-tagged σ₁SR bands, we used an anti-σ₁R C-terminal lumenal domain-specific antibody. σ₁SR-eGFP (37 kDa) and σ₁SR-mCherry (41 kDa) bands were not detected (Fig. 4A, lower panel).

FIGURE 4.

σ₁SR forms a complex with and co-localizes with σ₁R. A, co-immunoprecipitation of fluorescence-tagged σ₁SR and σ₁R in Neuro-2a cells. Extracts were immunoprecipitated (IP) with anti-eGFP antibody, and immunoprecipitates were immunoblotted (WB) with anti-σ₁R N-terminal cytosolic domain (upper panel) or with the C-terminal lumenal domain (lower panel) antibody. Cell extracts (Input) from Neuro-2a cells transfected with eGFP (lane 1), σ₁R-eGFP (lane 2), σ₁R-mCherry (lane 3), σ₁SR-eGFP (lane 4), or σ₁SR-mCherry (lane 5) constructs are shown as positive controls for σ₁R isoform immunoreactive bands. B and C, co-immunoprecipitation of σ₁Rs and IP₃Rs in Neuro-2a cells. Extracts were immunoprecipitated with anti-IP₃Rs (B) or anti-σ₁R N-terminal cytosolic domain (C) antibody, and immunoprecipitates were immunoblotted (WB) with anti-IP₃Rs (B) or anti-σ₁R N-terminal (C) antibody. As a positive control for σ₁Rs and IP₃Rs immunoreactive bands, cell extracts (Input) are shown from intact (lane 1) or σ₁R- (lane 2) or σ₁SR (lane 3)-transfected cells. D, confocal images showing co-localization of σ₁R-eGFP (green) and σ₁SR-mCherry (red) in Neuro-2a cells. Scale bar, 10 μm.

Because σ₁R forms a complex with IP₃R type-3 in CHO cells (11), we examined potential binding of σ₁R and σ₁SR with IP₃R in Neuro-2a cells. After immunoprecipitation with an IP₃R antibody, we observed only a 26-kDa σ₁R immunoreactive band in σ₁R- or σ₁SR-overexpressing cells (Fig. 4B, lanes 4–6). Next, after σ₁R immunoprecipitation with an anti-σ₁R N-terminal specific antibody, we undertook immunoblotting using an anti-IP₃Rs antibody. As expected, increased levels of an IP₃R immunoreactive band were seen in σ₁R-overexpressing (Fig. 4C, lane 5) but not σ₁SR-overexpressing (Fig. 4C, lanes 4 and 6) cells relative to controls. No significant differences in IP₃R protein expression were observed between control cells and cells expressing σ₁R isoforms (Fig. 4C, lanes 1–3). Confocal microscopy of Neuro-2a cells indicated that σ₁SR-mCherry coincided with σ₁R-GFP immunofluorescence in the ER (Fig. 4D).

σ₁SR Expression Reduces Mitochondrial Ca²⁺ Uptake in Response to IP₃R-driven Stimuli

IP₃Rs are important for ER-mitochondrial Ca²⁺ transport, which is regulated by σ₁R in the MAM of CHO cells (11). Because σ₁SR forms a complex with σ₁R, we assessed a potential role for σ₁SR in mitochondrial Ca²⁺ uptake. To assay mitochondrial Ca²⁺ uptake in Neuro-2a cell lines stably expressing σ₁R-mCherry (σ₁R-mCh cells) or σ₁SR-mCherry (σ₁SR-mCh cells), we conducted Ca²⁺ imaging using ratiometric pericam-mt Ca²⁺ probes, which localize to mitochondria (Fig. 5A) (33). We confirmed that σ₁R- and σ₁SR-mCherry mRNA levels in each line were equivalent, and we also found that levels showed an approximate 20-fold increase over parental Neuro-2a cells (supplemental Fig. 1A).

FIGURE 5.

σ₁SR expression reduces mitochondrial Ca²⁺ uptake in response to IP₃R-driven stimuli. A, pseudo-colored images of Neuro-2a cells expressing the mitochondrial Ca²⁺ indicator, ratiometric pericam-mt, at 30 s (left) and 60 s (right), corresponding to a control cell in B. Regions of interest for [Ca²⁺]_mt measurement is indicated by the circles in A. B, C, and E, relative fluorescence intensity changes following treatment with 10 μm ATP in Neuro-2a cells (B and E) or 1 μm IP₃ in permeabilized Neuro-2a cells (C) after transfection of a mitochondrial pericam probe. D and F, relative Fura-2/AM fluorescence intensity changes following 10 μm ATP stimulation. Top line indicates incubation period. The ratio of the ratiometric pericam-mt and Fura-2 fluorescence peak value of the release phase relative to the preceding base line is shown on the right. Each bar represents the mean ± S.E. *, p < 0.05; **, p < 0.01 versus control cells. #, p < 0.05; ##, p < 0.01 versus σ₁SR-mCh cells. n = 12 cells per experiment, performed in triplicate from four preparations. Scale bar, 10 μm.

We then asked whether mitochondrial Ca²⁺ elevation elicited by ATP, acting on receptors coupled with G_q protein to stimulate IP₃ production. For this experiment, Ca²⁺-free Krebs buffer containing 2 mm EGTA was used to assess intracellular Ca²⁺ mobilization and mitochondrial Ca²⁺ elevation through the ER IP₃R. As shown in Fig. 5B, ATP stimulation caused a rapid rise in [Ca²⁺]_mt, followed by a gradual decline with a sustained plateau phase in control cells (black line). In σ₁R-mCh cells (Fig. 5B, red line), the [Ca²⁺]_mt increase was markedly enhanced compared with that seen in control cells, although it was significantly decreased in σ₁SR-mCh cells (blue line). Interestingly, decreased [Ca²⁺]_mt mobilization seen in σ₁SR-mCh cells was significantly rescued to levels comparable with those seen in control cells by σ₁R co-expression (Fig. 5B, green line). Peak amplitudes of [Ca²⁺]_mt mobilization are summarized (0.18 ± 0.02 in control, 0.3 ± 0.01 in σ₁R-mCh, 0.09 ± 0.02 in σ₁SR-mCh, and 0.14 ± 0.01 in σ₁SR-mCh plus σ₁R expression).

To confirm that IP₃Rs function in [Ca²⁺]_mt mobilization, saponin-permeabilized cells were treated with IP₃, and [Ca²⁺]_mt was monitored in the presence of ICM buffer. Consistent with results seen after ATP application, IP₃-mediated [Ca²⁺]_mt mobilization was significantly enhanced in σ₁R-mCh cells and decreased in σ₁SR-mCh cells relative to control cells. In addition, [Ca²⁺]_mt decreases seen in σ₁SR-mCh cells were rescued by σ₁R co-expression (Fig. 5C; peak amplitude 0.17 ± 0.04 in control cells, 0.22 ± 0.04 in σ₁R-mCh cells, 0.13 ± 0.03 in σ₁SR-mCh cells, and 0.19 ± 0.06 in σ₁SR-mCh plus σ₁R expression).

We next examined intracellular mobilization [Ca²⁺]_c following ATP stimulation using the ratiometric indicator Fura-2/AM in Ca²⁺-free Krebs buffer. Interestingly, ATP-induced [Ca²⁺]_c mobilization in σ₁R-mCh cells was significantly reduced compared with controls. By contrast, [Ca²⁺]_c mobilization in σ₁SR-mCh cells was greater than that seen in control cells and that increase was suppressed by σ₁R co-expression (Fig. 5D; peak amplitude 0.13 ± 0.02 in control cells, 0.12 ± 0.08 in σ₁R-mCh cells, 0.16 ± 0.06 in σ₁SR-mCh cells, and 0.1 ± 0.03 in σ₁SR-mCh plus σ₁R cells). This evidence strongly suggests that increased IP₃-mediated [Ca²⁺]_mt mobilization stimulated by σ₁R overexpression decreases [Ca²⁺]_c mobilization, although IP₃-mediated [Ca²⁺]_mt mobilization due to σ₁SR overexpression enhances it.

We also confirmed that ATP-induced [Ca²⁺]_mt mobilization was mediated by σ₁Rs using siRNA knockdown of both σ₁R isoforms. Expression levels of both proteins were down-regulated ∼70% by siRNA treatment (supplemental Fig. 1B). σ₁R isoform knockdown significantly decreased ATP-induced [Ca²⁺]_mt mobilization (Fig. 5E; peak amplitude 0.18 ± 0.03 in control cells and 0.1 ± 0.02 in σ₁R isoform siRNA cells); conversely, knockdown of both σ₁R and σ₁SR protein elevated ATP-induced [Ca²⁺]_c mobilization (Fig. 5F; peak amplitude 0.15 ± 0.05 in control cells and 0.19 ± 0.08 in σ₁R isoform siRNA cells).

Overexpression of σ₁R-mCh and σ₁SR-mCh may alter the capacity of ER Ca²⁺ stores. To assess this possibility, we stimulated cells with thapsigargin, an inhibitor of the sarcoplasmic reticulum calcium ATPase. Thapsigargin treatment promoted a transient increase in [Ca²⁺]_c by depleting Ca²⁺ stores in the absence of extracellular Ca²⁺, and subsequent addition of extracellular Ca²⁺ increased [Ca²⁺]_c through capacitative Ca²⁺ entry. Both the capacity of ER Ca²⁺ stores and capacitative Ca²⁺ entry were comparable in cells with or without σ₁R and σ₁SR overexpression (supplemental Fig. 2). These results suggest that both σ₁R isoforms affect the efficiency of mitochondrial Ca²⁺ uptake in response to IP₃R-driven stimuli without changing the capacity of ER Ca²⁺ stores.

σ₁SR Acts Antagonistically to σ₁R following ER Stress

Because σ₁SR expression suppresses IP₃-mediated [Ca²⁺]_mt mobilization and enhances receptor-mediated [Ca²⁺]_c mobilization through the ER, we hypothesized that σ₁SR overexpression increases the vulnerability of a cell to ER stress. To test this, we evaluated the effect of the ER stressor tunicamycin on overexpression of σ₁R isoforms. Tunicamycin treatment promoted expression of both endogenous σ₁R and σ₁SR protein in σ₁R-mCh but not control cells. Conversely, tunicamycin treatment reduced expression of both endogenous σ₁R and σ₁SR in σ₁SR-mCh cells (Fig. 6B).

FIGURE 6.

σ₁SR enhances IP₃R destabilization following ER stress. A, representative immunoblots probed with various antibodies are shown in control, σ₁R-mCh, and σ₁SR-mCh cells at 0, 4, and 24 h after 2 μg/ml tunicamycin treatment. B, quantitative densitometry analyses are shown. n = 6 per experiment. Each bar represents the mean ± S.E. *, p < 0.05; **, p < 0.01 versus control cells.

Tunicamycin treatment also promoted significant degradation of IP₃R proteins in control but not σ₁R-mCh cells, suggesting a chaperone activity for σ₁R. More importantly, σ₁SR-mCh overexpression enhanced degradation of IP₃R proteins relative to that seen in control cells (Fig. 6B; 50.8 ± 14.1% in control cells, 77 ± 24.1% in σ₁R-mCh cells, and 16.3 ± 8.8% in σ₁SR-mCh cells at 24 h after tunicamycin treatment). Because σ₁SR-mCh overexpression promoted protein degradation, we examined the effects of σ₁SR-mCh overexpression on ER stress using stress markers, such as phosphorylation of PERK and its downstream target C/EBP homologous protein. As unexpected, neither PERK phosphorylation nor C/EBP homologous protein levels were enhanced by σ₁SR-mCh overexpression (Fig. 6B). However, when we evaluated autophagic responses based on production of the autophagy marker LC3-II, the ratio of LC3-II to total (LC3-I plus LC3-II) LC3 was markedly increased by 24 h after tunicamycin treatment (Fig. 6B).

σ₁SR Enhances Autophagosome Formation following ER Stress

Given that LC3-II production was enhanced in σ₁SR-mCh cells (Fig. 6B), we determined the diameters of vacuoles containing LC3-II (34). Twenty four hours after tunicamycin treatment, diameters of ∼80% of LC3-positive vacuoles were smaller than the 1.0 μm seen in control cells (Fig. 7, A and D). σ₁R-mCh overexpression decreased the number of LC3-positive vacuoles compared with control without changing the size distribution (Fig. 7, B, D and E). By contrast, σ₁SR-mCh overexpression induced formation of larger autophagosomes than those seen in control cells, significantly shifting the size distribution profile. Over 50% of vacuoles were between 1.0 and 1.5 μm at 24 h after tunicamycin treatment (Fig. 7,C and D). The total number of LC3-positive particles also markedly increased in σ₁SR-mCh cells following tunicamycin treatment (Fig. 7E; 98.3 ± 11.4 in controls, 66.3 ± 7.2 in σ₁R-mCh cells, and 195 ± 16.7 in σ₁SR-mCh cells). The size of LC3-positive vacuoles did not change in control, σ₁R-mCh, and σ₁SR-mCh cells not treated with tunicamycin. Taken together, σ₁R overexpression inhibited formation of tunicamycin-induced autophagosomes, although σ₁SR overexpression promoted autophagosome formation under ER stress conditions.

FIGURE 7.

σ₁SR enhances autophagosome formation following ER stress. A–C, LC3 immunostaining in control (A), σ₁R-mCh (B), and σ₁SR-mCh (C) cells 24 h after 2 μg/ml tunicamycin treatment. Enlarged images indicate the boxed areas. Scale bars, 30 μm in low magnification and 10 μm in high magnification images. D and E, quantitative analysis of particle size (D) and total number (E) of LC3-immunoreactive autophagosomes 24 h after tunicamycin treatment. Ten fields (30 cells per field) in each condition were chosen randomly and photographed. Each bar represents the mean ± S.E. *, p < 0.05; **, p < 0.01 versus control cells.

σ₁SR Suppresses Mitochondrial ATP Production and Enhances Apoptosis following ER Stress

Finally, we asked how σ₁SR overexpression enhances the autophagic response. Because IP₃R-mediated Ca²⁺ transport into mitochondrial Ca²⁺ promotes oxidative phosphorylation, respiration, and ATP production by activating the tricarboxylic acid cycle (2, 35), we speculated that ATP production would be suppressed by σ₁SR overexpression. Therefore, we measured changes in cellular ATP production with or without tunicamycin-induced ER stress. Unexpectedly, overexpression of σ₁R or σ₁SR markedly enhanced ATP production without tunicamycin treatment. In control cells, tunicamycin treatment significantly increased ATP production at 24 h but markedly suppressed it after 48 h of treatment. σ₁R overexpression significantly promoted ATP production both at 24 and 48 h after tunicamycin treatment (Fig. 8A). ATP production in σ₁SR-mCh cells was significantly suppressed at both the 24- and 48-h time points compared with control cells (Fig. 8A; 41.9 ± 1.1% in controls, 86.4 ± 5.3% in σ₁R-mCh cells, and 14.5 ± 0.24% in σ₁SR-mCh cells at 48 h compared with control cells at a resting state). To confirm that changes in ATP production are related to mitochondrial Ca²⁺ mobilization, we examined cellular ATP levels with or without tunicamycin treatment in the presence or absence of Ru360, a mitochondrial Ca²⁺ uptake blocker. ATP production enhanced by σ₁R or σ₁SR overexpression or by tunicamycin treatment was completely inhibited by Ru360 treatment (supplemental Fig. 3).

FIGURE 8.

σ₁SR suppresses mitochondrial ATP production and enhances apoptosis following ER stress. A, intracellular ATP levels were measured in untreated cells or cells treated with 2 μg/ml tunicamycin at 24 and 48 h. B, TUNEL-positive cells were counted in treated cells at 24 and 48 h. One hundred cells from 13 randomly selected fields were counted in each experiment. C, real time PCR analysis showed that cytochrome c and ATP5A1 mRNA expression showed no significant differences between each group. n = 6 per experiment. Each bar represents the mean ± S.E. **, p < 0.01 versus control at no stimulation. #, p < 0.05; ##, p < 0.01 versus control at each time point.

Because ATP production was suppressed 48 h after tunicamycin treatment and treatment could promote apoptosis, we evaluated apoptosis using TUNEL staining. σ₁SR-mCh overexpression significantly enhanced the number of TUNEL-positive cells 24 h after tunicamycin treatment, although σ₁R-mCh overexpression significantly suppressed tunicamycin-induced apoptosis. By contrast, σ₁SR-mCh overexpression markedly enhanced tunicamycin-induced apoptosis at 48 h (Fig. 8B; percentages of TUNEL-positive cells were 43.5 ± 8.4% in control, 19.1 ± 4.4% in σ₁R-mCh, and 70.4 ± 5.4% in σ₁SR-mCh).

To confirm that σ₁R-mCh and σ₁SR-mCh overexpression does not alter mitochondrial protein levels, we examined the expression of cytochrome c and the α subunit of ATP synthase (ATP5A1), subunits of the respiratory chain complex. σ₁R-mCh and σ₁SR-mCh cells showed no change in mRNA levels of these factors (Fig. 8C), indicating that σ₁R-mCh and σ₁SR-mCh overexpression does not interfere with mitochondrial structure or directly regulate IP₃R-mediated ATP production.

DISCUSSION

Here, we identified σ₁SR, a novel truncated isoform of the σ₁R. Using Neuro-2a cells, we showed that under physiological conditions the σ₁SR overexpression slightly stimulates ATP-induced cytosolic Ca²⁺ mobilization, in contrast with a small reduction promoted by σ₁R overexpression (Fig. 5D). These changes are likely due to increased and decreased Ca²⁺ mobilization into mitochondria induced by σ₁SR and σ₁R overexpression, respectively. Unexpectedly, σ₁SR overexpression increased ATP production in the absence of stress stimuli, such as ER stress (Fig. 8A), an activity not associated with reduction in ATP-induced mitochondrial Ca²⁺ mobilization. Thus, although ATP-induced mitochondrial Ca²⁺ mobilization is significantly reduced by σ₁SR overexpression, mitochondrial Ca²⁺ levels may be moderately increased with mitochondrial elongation. In support of this idea, increased ATP production seen following σ₁SR overexpression was eliminated by treating cells with the mitochondrial Ca²⁺ transport blocker Ru360 (supplemental Fig. 3).

Because basal ATP levels increase following overexpression of either isoform (Fig. 8A) and σ₁R overexpression in PC12 cells reportedly stimulates neurite sprouting in response to nerve growth factor (36), we hypothesized that σ₁SR overexpression might stimulate Neuro-2a cell differentiation. To evaluate this possibility, we measured morphological changes in both σ₁R- and σ₁SR-overexpressing Neuro-2a cells. Consistent with the findings of Ref. 36, σ₁R overexpression significantly stimulated neurite extension as compared with control cells. Similarly, like σ₁R overexpression, σ₁SR overexpression significantly stimulated neurite extension (supplemental Fig. 4; 3.2 ± 1.3% in control, 45.5 ± 4.3% in σ₁R, and 46 ± 7.4% in σ₁SR-transfected cells).

Mitochondrial elongation and extension of gap junctions between ER and mitochondria promoted by σ₁SR and σ₁R overexpression likely account for enhanced mitochondrial ATP production. Elongated mitochondria express higher levels of the dimeric form of ATPase, which is associated with more efficient ATP production (37). Increased contact space between the ER and mitochondria also likely enhances mitochondrial Ca²⁺ transport (38, 39). An ER-associated sorting protein, phosphofurin acidic cluster sorting protein 2 (PACS-2), is required for association of mitochondria with the ER. PACS-2 depletion induces mitochondrial fragmentation and uncoupling with the ER, resulting in aggravated ER stress (40). Mitofusin-2 (Mfn2) in the ER is also required for adhesion between mitochondrial and ER membranes. Depletion of Mfn2 impairs ER-mitochondrial Ca²⁺ transport (25) and induces cell death in cerebellar granule neurons (41). Taken together, σ₁R overexpression causes elongation of mitochondria and enhances IP₃R-induced Ca²⁺ transport, thereby promoting mitochondrial ATP production. Thus, elevated mitochondrial energy production likely promotes cell survival in the presence of ER stress.

σ₁SR Expression Promotes Mitochondrial ATP Depletion and Autophagy Only under ER Stress Conditions

The σ₁R reportedly stabilizes IP₃Rs to maintain Ca²⁺ transport from the ER into mitochondria in CHO cells (11). We confirmed that σ₁R overexpression enhances IP₃-induced mitochondrial Ca²⁺ transport and ATP production, whereas σ₁SR did not bind to IP₃Rs, and its overexpression did not enhance IP₃-induced mitochondrial Ca²⁺ transport. Under ER stress conditions, σ₁SR expression had detrimental effects on mitochondrial Ca²⁺ transport through the ER (Fig. 5B), promoting mitochondrial ATP depletion (Fig. 8). σ₁SR overexpression also destabilized IP₃Rs (Fig. 6B), which may account for decreased mitochondrial Ca²⁺ transport. Finally, σ₁SR overexpression promoted an autophagic response to ER stress following tunicamycin treatment without altering PERK activity (Figs. 6B and 7). Autophagy functions to recycle energy and nutrients in nutrient starvation and ER stress conditions (42), but the balance between autophagy and cell death is highly dependent on intracellular Ca²⁺ levels (43). IP₃R protein levels and activity are critical to inhibit autophagy (44). Indeed, autophagy is inhibited and promoted by IP₃R agonists (such as IP₃) and antagonists (such as xestospongins), respectively (45–47). In addition, IP₃R is required to inhibit autophagy under physiological conditions (6). The lack of mitochondrial Ca²⁺ transport by IP₃R depletion inhibits pyruvate dehydrogenase and increases the AMP/ATP ratio, thereby aggravating autophagy via AMP-activated protein kinase (6). IP₃Rs also inhibit autophagy through binding with Bcl-2 and Beclin-1 (known as autophagy-related gene 6) in HeLa cells (48). Inhibition of IP₃Rs by xestospongin B promotes disruption of complexes formed by IP₃R, Bcl-2, and Beclin-1, activating autophagy. Taken together, IP₃R dysfunction through σ₁R down-regulation accounts for autophagic mechanisms in σ₁SR-expressed cells.

Physiological Relevance of Interaction between σ₁R and σ₁SR and Their Ligands

As shown in Fig. 1, the σ₁R is composed two transmembrane domains (TM1, amino acids 11–29; TM2, amino acids 91–109), an extracellular loop, and intracellular N and C termini with the C-terminal region including a large soluble domain (49, 50). Pharmacological studies indicate that numerous compounds, including benzomorphans (SKF-10047, pentazocine), antipsychotics (haloperidol), antidepressants (fluvoxamine), steroids (dehydroepiandrosterone, progesterone), and drugs of abuse (methamphetamine, cocaine) can bind to the σ₁R, primarily through the TM2 and C-terminal regions (50). The σ₁R agonist (+)-pentazocine positively modulates σ₁R/IP₃R association and stabilizes the complex at ER-mitochondria contact sites under ER stress. As a result, IP₃ binding to the IP₃R increases, and Ca²⁺ efflux is enhanced (11). In addition, ligand-induced regulation of this function apparently resides largely in the N terminus, contributing to functional coupling of C- and N-terminal σ₁R fragments. Wu and Bowen (51) reported that agonist binding to the σ₁R may change its conformation such that the N-terminal segment dissociates from the C terminus, which in turn can interact more avidly with IP₃R. Here, we showed that σ₁R binds to the IP₃R via its C terminus (Fig. 4). σ₁SR displays the opposite function, suppressing mitochondrial ATP production and promoting cell death following tunicamycin-induced ER stress. We suggest that σ₁SR, whose sequence is almost identical to the σ₁R N terminus (Fig. 1), interacts with the σ₁R C terminus, inhibiting σ₁R-mediated IP₃R-derived ER-mitochondrial Ca²⁺ transport.

Lymphocytes also express another σ₁R splice variant, which is replaced at Ala-13, Leu-28, and Ala-86 to Thr-13, Pro-28, and Val-86, respectively, and lacks 31 amino acids corresponding to residues 119–149 in the σ₁R protein (52). The σ₁R isoform lacks ligand binding sites. In NG108-15 cells, the full-length σ₁R forms a trimeric complex with the cytoskeletal adaptor protein ankyrin B and IP₃R in the ER membrane, and σ₁R agonists cause dissociation of ankyrin B from the IP₃R, promoting IP₃R activation (53). Indeed, the C-terminal σ₁R peptides (amino acids 102–223) transfected into MCF-7 breast tumor cells promote decreased levels of ankyrin B associated with the IP₃R compared with untransfected cells, enhancing IP₃R activation (51). In addition, the N-terminal σ₁R peptides (amino acids 1–100) expressed in MCF-7 cells weakly associate with ankyrin B and IP₃R complexes, but have little capacity to enhance IP₃R activity (51). In addition, a glutathione S-transferase (GST) fusion form of σ₁R (amino acids 116–223) has chaperone activity, blocking aggregation of denatured citrate synthase in vitro. However, GST-σ₁R (amino acids 29–92) lacks chaperone activity (11). Taken together with these studies, σ₁SR as defined here likely lacks chaperone activity and ligand binding ability and acts instead as a dominant negative form of σ₁R by blocking C-terminal chaperone activity under ER stress conditions or disrupting IP₃R/σ₁R interaction. However, under normal conditions, overexpression of either isoform promotes similar phenotypes relevant to ATP production, mitochondrial elongation, and neurite extension, suggesting that endogenous σ₁SR has an alternate unique function in mitochondrial elongation and adhesion with the ER.

We showed that ATP-induced [Ca²⁺]_c mobilization in σ₁SR-mCh cells was greater than that seen in control cells in Ca²⁺-free Krebs buffer (Fig. 5D). σ₁SR expression possibly destabilizes IP₃Rs by inhibiting chaperone activity of endogenous σ₁R, thereby reducing IP₃R-derived ER-mitochondrial Ca²⁺ transport (Fig. 9). This activity may elicit enhanced [Ca²⁺]_c mobilization following ATP stimulation of σ₁SR-mCh cells. However, extensive studies are required to define how IP₃R/σ₁R interaction is disrupted by σ₁SR overexpression.

FIGURE 9.

Schematic representation of altered mitochondrial ATP production and cell death in the presence of two σ₁R isoforms following ER stress. Upper, in ER stress conditions, σ₁R stabilizes IP₃Rs and sustains mitochondrial Ca²⁺ uptake-derived ATP production, enhancing survival. Lower, by contrast, σ₁SR expression destabilizes IP₃Rs and promotes dysfunction of IP₃R-derived ER-mitochondrial Ca²⁺ transfer through functional loss of σ₁R, resulting in ATP depletion and autophagic apoptosis. In resting conditions, both σ₁R isoforms positively regulate mitochondrial biogenesis.

Among the three IP₃R isoforms, type-3 IP₃R plays a critical role in induction of apoptosis by preferentially transmitting Ca²⁺ signals into mitochondria (54). siRNA-based knockdown of type-3 IP₃R significantly decreases IP₃-induced mitochondrial Ca²⁺ concentrations in CHO cells, whereas knockdown of type-1 IP₃R reduces cytosolic Ca²⁺ concentration (54). In addition, type-3 IP₃Rs are particularly enriched at the MAM, whereas type-1 IP₃R is homogeneously expressed in ER membranes in neurons (9). σ₁Rs form a trimeric complex with ankyrin-B and the type-3 IP₃Rs but not with type-1 IP₃Rs in NG108-15 cells (53). Taken together, σ₁R acts as a specific chaperone for type-3 IP₃R at the MAM, regulating [Ca²⁺]_mt mobilization in neurons. Notably, σ₁R agonists, including (+)-pentazocine, increase [Ca²⁺]_c from the ER following bradykinin stimulation of NG108-15 cells (55). In addition, σ₁R (amino acids 102–223)-transfected MCF-7 cells also show increased [Ca²⁺]_c induced by bradykinin stimulation (51), suggesting that other σ₁R-binding proteins regulate intracellular Ca²⁺ mobilization. In the future, we will investigate whether these binding proteins regulate [Ca²⁺]_c mobilization after stimulation with σ₁R ligands in neurons and define σ₁SR function in [Ca²⁺] mobilization.

Pathophysiological Relevance of σ₁SR

Autophagy is activated in several neurodegenerative disorders, although its significance in neuronal death and survival remains to be defined (56). Al-Saif et al. (19) reported a SIGMAR1 missense mutation in exon 2 associated with juvenile amyotrophic lateral sclerosis (c.304G→C) that substitutes glutamine for glutamic acid at residue 102 (E102Q), which occurs at the σ₁SR splice site (Fig. 1). Expression of the E102Q mutant enhances apoptosis in the mouse motor neuron-like cell line NSC34 (19). Further studies are needed to reveal functions of σ₁SR in neurodegenerative disorders.

In conclusion, we have identified a novel σ₁R isoform, σ₁SR, as a key molecule in ER stress-related cellular apoptosis through regulation of ER-mediated mitochondrial biogenesis. Importantly, σ₁SR does not bind IP₃Rs but inhibits the ability of σ₁R to promote mitochondrial ATP production through IP₃R-mediated Ca²⁺ transport. Future studies should address pathological conditions in which σ₁SR is up-regulated and potential interference with σ₁R and mitochondrial function.