Abstract
Amyotrophic lateral sclerosis (ALS) is a progressive neurodegenerative disorder characterized by motor neuron loss in the spinal cord and brain. Mutations in the superoxide dismutase 1 (SOD1) gene have been linked to familial ALS. To elucidate the role of SOD1 mutations in ALS, we investigated 14-3-3, a crucial regulator of cell death that was identified in patients with familial ALS. In a transgenic mouse model (SOD1-G93A) of ALS, 14-3-3 co-localized with mutant SOD1 aggregates and was more insoluble in the spinal cords of mutant SOD1 transgenic mice than in those of wild-type mice. Immunofluorescence and co-immunoprecipitation experiments showed that the 14-3-3ɛ and θ isoforms interact with mutant SOD1 aggregates in the juxtanuclear quality control compartment of N2a neuroblastoma cells. Fluorescence loss in photobleaching experiments revealed that movement of the isoforms of 14-3-3 was markedly reduced in SOD1 aggregates. Bax translocation into and cytochrome c release from the mitochondria were promoted by the sequestration of 14-3-3 into mutant SOD1 aggregates, increasing cell death. Mutant SOD1 aggregates were dissolved by the Hsp104 chaperone, which increased the interaction of 14-3-3 with Bax, reducing cell death. Our study demonstrates that mutant SOD1 inhibits 14-3-3-mediated cell survival. This information may contribute to the identification of a novel therapeutic target for ALS.
Introduction
Amyotrophic lateral sclerosis (ALS) is a common adult-onset, progressive neurodegenerative disease involving upper and lower motor neuron loss (1,2). It usually manifests as sporadic ALS (sALS), but in approximately 10% of the cases, it manifests as familial ALS (fALS). Many genes or loci associated with ALS, including TDP43, FUS, VCP, UBQLN, OPTN, DCTN1, PFN, and C9ORF72, have been identified or at least mapped to a specific region of a chromosome (3). Approximately, 20% of the fALS cases are linked to mutations in the Cu/Zn superoxide dismutase 1 (SOD1) gene. To date, more than 100 point mutations have been identified in the SOD1 gene (4). Many pathogenic mechanisms have been suggested for ALS, including genetic factors (5–9), oxidative stress (10), excitotoxicity (11,12), ER stress (13–15), proteasome inhibition (16,17), mitochondrial dysfunction (18–20), altered axonal transport (21–23), neuroinflammation (24–26), dysregulated RNA processing (27,28), and protein aggregation (29–31). However, the precise mechanism underlying the pathogenesis of ALS remains unclear. Protein aggregation is a hallmark of many neurodegenerative diseases, including Alzheimer’s disease (AD), Parkinson’s disease (PD), Huntington’s disease (HD), and ALS (32). Protein aggregates are detergent insoluble and have low mobility within the cell (33). Protein aggregates in patients with ALS contain a variety of cellular proteins, including SOD1 (34). In addition, they tend to attract cellular proteins and organelles such as the Golgi apparatus, mitochondria, and endoplasmic reticulum, disrupting the normal function of other cell proteins and organelles (16,35,36). It is still controversial whether protein aggregates confer a toxic or protective effect (37).
The 14-3-3 proteins are a family of protein chaperones that are abundant in the brain and comprise approximately 1% of whole brain proteins (38). The seven 14-3-3 isoforms (β, ɛ, γ, η, θ, σ, and ζ) (39) form homo- or heterodimers and bind to phosphorylation sites of target proteins (40,41). Furthermore, 14-3-3 proteins are involved in many cellular functions, including signaling, growth, division, adhesion, the cell cycle, neuronal plasticity, differentiation, and apoptosis. These functions may be regulated by the subcellular localization of 14-3-3-binding proteins and protein–protein interactions at the phosphorylation sites of interacting proteins (42). In particular, 14-3-3 proteins inhibit apoptosis by interacting with Bad or Bax, resulting in the cytosolic localization of these proteins. Bad and Bax are proapoptotic proteins that induce the release of cytochrome c into the cytosol by mitochondrial translocation (43,44). Moreover, 14-3-3 proteins have been linked to several neurodegenerative diseases, including AD, HD, PD, and spinocerebellar ataxia. They are also present in disease-specific protein aggregates in neurodegenerative brains (45). In addition, 14-3-3 proteins have been detected in the spinal cords of patients with ALS and in transgenic mouse models of the disease (46,47).
The juxtanuclear quality control compartment (JUNQ) is a distinct compartment in cells that contains protein quality control machineries such as chaperones and proteasomes. Sequestration of SOD1 aggregates to JUNQ inhibits proteasomal function and reduces cell viability (48).
We investigated the relationship between mutant SOD1 aggregates and 14-3-3 proteins. Our results showed that 14-3-3 isoforms co-localize with mutant SOD1 aggregates in JUNQ but not with wild-type SOD1. Moreover, the antiapoptotic function of 14-3-3 proteins was inhibited by mutant SOD1 aggregates, but this was reversed by the overexpression of 14-3-3 isoforms. Finally, Hsp104-mediated disaggregation of mutant SOD1 aggregates decreased cell toxicity.
Results
Solubility of 14-3-3 is significantly decreased in ALS transgenic mice
Previously, 14-3-3 aggregates have been detected in mice and human tissues associated with ALS and other neurodegenerative diseases (45–47). Using immunohistochemistry, we observed co-localization of mutant SOD1 aggregates and 14-3-3 proteins in the spinal cord of ALS transgenic mice (8, 12, and 16 weeks) but not control mice (Fig. 1A). These results suggest that this co-localization is important in the pathogenesis of ALS. Next, we investigated the expression levels of all 14-3-3 isoforms (β, γ, ɛ, θ, ζ, and η), except σ, in the spinal cord of ALS transgenic mice (8 weeks). We could not investigate 14-3-3 σ because the isoform was not expressed in the spinal cord (data not shown). The amounts of soluble 14-3-3 ɛ and θ were significantly lower in ALS transgenic mice than in control mice (14-3-3 ɛ, 54%; 14-3-3 θ, 60%) (Fig. 1B).
Soluble 14-3-3 proteins were significantly reduced in ALS transgenic mice. (A) Immunohistochemistry of the spinal cord in ALS transgenic mice (8, 12, and 16 weeks) and control mice. The spinal cord was double immunostained with SOD1 (green) and pan 14-3-3 (red) antibodies. SOD1 aggregates and pan 14-3-3 co-localized in ALS transgenic mice (8, 12 and 16 weeks). Scale bar = 5 μm. (B) Quantitation of soluble 14-3-3 isoforms from the spinal cords of 8-week-old ALS transgenic mice. Spinal cords of ALS transgenic and control mice were isolated, and soluble proteins were extracted from spinal cord homogenates and incubated with primary antibodies against pan 14-3-3, 14-3-3 β, 14-3-3 ɛ, 14-3-3 γ, 14-3-3 θ, 14-3-3 ζ, or 14-3-3 η overnight at 4 °C. 14-3-3 isoforms were quantified using SDS gel electrophoresis and western blotting. Results were normalized to GAPDH. n = 3 per group. Error bar indicates SD. *P < 0.05; **P < 0.01. Statistical analysis was performed using Student’s t-test. (C) Quantitation of soluble 14-3-3 and insoluble 14-3-3 in the spinal cord of 8-week-old ALS transgenic mice. Spinal cords of ALS transgenic mice and control mice were fractionated into soluble and insoluble fractions. The level of pan 14-3-3 protein was quantified by SDS gel electrophoresis and western blotting. Results were normalized to GAPDH. n = 3 per group. Error bar reflects SD. Statistical analysis was performed using Student’s t-test.
To confirm that the decrease in 14-3-3 protein levels in ALS transgenic mice was not caused by a decreased mRNA expression, we quantified 14-3-3 isoform ɛ and θ in ALS transgenic (8 weeks) and control mice by quantitative polymerase chain reaction (PCR). No difference was observed in the mRNA levels of 14-3-3 isoforms between ALS transgenic and control mice (Supplementary Material, Fig. S1). These results suggest that the reduced 14-3-3 protein level is explained by differences in the solubility of 14-3-3 proteins between ALS transgenic and control mice. Mutant SOD1 proteins are less soluble than wild-type SOD1 proteins (49). We conducted soluble–insoluble fraction experiments to compare the solubility of 14-3-3 proteins in ALS transgenic (8 weeks) and control mice. As expected, 14-3-3 protein was more insoluble (approximately three times) in ALS transgenic mice than in control mice; conversely, the solubility of 14-3-3 protein was lower in ALS transgenic mice than in control mice (Fig. 1C). These results indicate that the reduced 14-3-3 protein level in the spinal cord of ALS transgenic mice is related to changes in protein solubility and not mRNA levels.
Pan 14-3-3 sequestrates with mutant SOD1 aggregates
We investigated whether 14-3-3 proteins co-localize with mutant SOD1 aggregates. First, we performed immunofluorescence assays on N2a cells, which are derived from mouse brain tissue. GFP-tagged SOD1 wild-type, A4V, G85R, or G93A were transfected into N2a cells, and aggregates were visualized by GFP fluorescence. More than 100 ALS mutants have been reported; among these, for this study, we used the most common mutants A4V, G85R, and G93A (50,51). In particular, the A4V mutant is the most common in the U.S. (approximately 50% of SOD1-ALS patients) and induces rapid disease progression (52). We observed aggregates in cells containing mutant SOD1 (A4V, G85R, and G93A) but not wild-type SOD1 (Supplementary Material, Fig. S2A–D). Furthermore, 14-3-3 proteins co-localized with mutant SOD1 (A4V, G85R, and G93A) aggregates but were dispersed in the cytosol of SOD1 wild-type cells. We then performed immunoprecipitation experiments to confirm the interaction of these proteins. N2a cells were transfected with constructs containing either FLAG-tagged SOD1 WT, A4V, G85R, or G93A. The cell extracts were immunoprecipitated with pan 14-3-3 antibodies and then analysed by western blotting with anti-FLAG antibodies. Mutant SOD1 (A4V, G85R, and G93A) but not wild-type SOD1 immunoprecipitated with 14-3-3, indicating that mutant SOD1 proteins but not wild-type SOD1 interact with 14-3-3 (Supplementary Material, Fig. S2E). Soluble–insoluble protein fractionation experiments were performed to confirm the co-localization of insoluble mutant SOD1 aggregates and pan 14-3-3. Pan 14-3-3 was more soluble in wild-type SOD1 transfected cells and more insoluble in mutant SOD1 transfected cells (Supplementary Material, Fig. S2F). Taken together, these results demonstrate that mutant SOD1 aggregates sequester 14-3-3 by direct protein–protein interaction.
Both 14-3-3 ɛ and θ are sequestered with mutant SOD1 aggregates in JUNQ
There are seven known isoforms of 14-3-3 (β, ɛ, γ, η, θ, σ, and ζ) (39). We investigated which of the 14-3-3 isoforms interact with mutant SOD1 aggregates using antibodies against 14-3-3 β, γ, ɛ, θ, and ζ. In cells transfected with wild-type SOD1, no 14-3-3 isoforms co-localized with wild-type SOD1. However, in cells transfected with mutant SOD1, the isoforms ɛ and θ co-localized with mutant SOD1 aggregates (Fig. 2A and B). Furthermore, co-immunoprecipitation experiments showed that the isoforms ɛ and θ interact with mutant SOD1 (A4V) but not wild-type SOD1 (Fig. 2C). The isoforms β and γ did not co-localize with mutant SOD1 aggregates (Supplementary Material, Fig. S3A and B). Soluble–insoluble fractionation experiments confirmed that the isoforms ɛ and θ were more insoluble in mutant SOD1 (A4V) transfected cells (Fig. 2D) and that these isoforms interacted with mutant SOD1 (A4V) in both the insoluble and soluble fractions (Fig. 2E).
14-3-3 ɛ and 14-3-3 θ were sequestered with mutant SOD1 aggregates in JUNQ. N2a cells were transiently transfected with GFP-SOD1, GFP-A4V, GFP-G85R, or GFP-G93A. At 48 h after transfection, cells were immunostained with antibodies against 14-3-3 ɛ (A) or 14-3-3 θ (B). Both 14-3-3 ɛ and 14-3-3 θ co-localized with mutant SOD1 (A4V, G85R, G93A) aggregates but not with wild-type SOD1. Scale bar = 20 μm. (C) For immunoprecipitation assays, N2a cells were transiently transfected with FLAG-SOD1 or FLAG-A4V. At 48 h after transfection, each sample was immunoprecipitated with 14-3-3 ɛ or 14-3-3 θ antibodies and immunoblotted with FLAG antibody. 14-3-3 ɛ and 14-3-3 θ interacted with mutant SOD1 (A4V) but not with wild-type SOD1. (D) N2a cells were transiently transfected with GFP-SOD1 or GFP-A4V. At 48 h after transfection, proteins were fractionated using a solution containing 0.5% NP-40. 14-3-3 ɛ and 14-3-3 θ were more insoluble in cells transfected with mutant SOD1 (A4V) than wild-type SOD1. (E) For immunoprecipitation experiments using fractionated soluble and insoluble proteins, each sample was immunoprecipitated with 14-3-3 ɛ or 14-3-3 θ antibodies and immunoblotted with GFP antibody. 14-3-3 ɛ and 14-3-3 θ interacted with mutant SOD1 (A4V) in soluble and insoluble fractions. (F–I) For FLIP assays, N2a cells were transiently transfected with either GFP-SOD1 or GFP-A4V and either DsRed2-14-3-3 ɛ or DsRed2-14-3-3 θ. At 48 h after transfection, each sample was bleached, and the fluorescence intensity of 14-3-3 ɛ or 14-3-3 θ was measured. The bleached region lost 14-3-3 ɛ or 14-3-3 θ fluorescence faster than the unbleached region in GFP-A4V transfected cells but not GFP-SOD1 transfected cells. Scale bar = 10 μm. (J, K) N2a cells were transiently transfected with GFP-SOD1 or GFP-A4V. At 48 h after transfection, cells were immunostained with antibodies against either 14-3-3 ɛ and Hsp70 or 14-3-3 θ and Hsp70. Hsp70 co-localized with mutant SOD1 (A4V) aggregates and either 14-3-3 ɛ or 14-3-3 θ. Scale bar = 10 μm. (L, M) N2a cells were transiently transfected with GFP-SOD1 or GFP-A4V. At 48 h after transfection, cells were immunostained with antibodies against ubiquitin and either 14-3-3 ɛ or 14-3-3 θ. Ubiquitin co-localized with mutant SOD1 (A4V) aggregates and either 14-3-3 ɛ or 14-3-3 θ. Scale bar = 10 μm.
We previously distinguished mutant SOD1 aggregates from soluble SOD1 proteins using fluorescence loss in photobleaching (FLIP) (53). In this study, we used FLIP assays to investigate the movement of the isoforms 14-3-3 ɛ and θ. In cells overexpressing wild-type SOD1, DsRed2-conjugated 14-3-3 ɛ and θ showed similar loss of fluorescence in photobleached and unbleached regions (Fig. 2F and G), indicating that these isoforms moved in and out of the bleached spot. However, when mutant SOD1 aggregates were expressed in cells containing DsRed2-conjugated 14-3-3 ɛ and θ, the levels of fluorescence differed between the bleached and unbleached regions (Fig. 2H and I), indicating that the mobility of these isoforms was reduced. We hypothesize that these isoforms were sequestered with mutant SOD1 aggregates, reducing their mobility.
Mutant SOD1 is known to form aggregates in the JUNQ compartment (48). To examine whether sequestered 14-3-3 ɛ and θ localized to the JUNQ compartment, we conducted immunofluorescence assays. Hsp70 and ubiquitin are established makers of the JUNQ compartment (48); therefore, we tested the co-localization of Hsp70 and ubiquitin with mutant SOD1 and 14-3-3 aggregates after transfection of GFP-SOD1 or GFP-A4V into N2a cells. Pan 14-3-3 co-localized with Hsp70 and ubiquitin in the presence of mutant SOD1 aggregates in N2a cells (Supplementary Material, Fig. S4A and B). In addition, we confirmed the co-localization of 14-3-3 ɛ and θ with Hsp70 and ubiquitin (Fig. 2J–M). These results demonstrate that these isoforms sequester with mutant SOD1 aggregates in the JUNQ compartment.
Mutant SOD1 inhibits 14-3-3-mediated cell survival
Many studies have reported that mutant SOD1 causes cell death via diverse pathways (54). We performed propidium iodide cell viability assays in the presence of either wild-type or mutant SOD1. N2a cells were transiently transfected with GFP-SOD1 or GFP-A4V. N2a cells expressing mutant SOD1 (A4V) showed a three-fold greater propensity for cell death than cells expressing wild-type SOD1 (Supplementary Material, Fig. S5A). One of the important functions of 14-3-3 in cell survival is its interaction with Bax, which inhibits the movement of Bax into the mitochondria. Therefore, we examined the localization of Bax in mitochondrial and cytosolic fractions prepared from cells transfected with wild-type and mutant SOD1. Mutant SOD1 induced greater mitochondrial localization of Bax than wild-type SOD1 (Supplementary Material, Fig. S5B). Experiments in parallel showed that more cytochrome c was released into the cytosol of cells expressing mutant SOD1 than in those expressing wild-type SOD1 (Supplementary Material, Fig. S5C). Taken together, these results demonstrate that mutant SOD1 promotes cell death by reducing the cell survival function of 14-3-3 protein.
Overexpression of 14-3-3 ɛ and θ ameliorates cell toxicity induced by mutant SOD1
We hypothesized that the observed cell death might be caused by the reduction in 14-3-3 ɛ and θ levels due to sequestration with mutant SOD1. To examine whether the overexpression of these isoforms could protect cells from mutant SOD1-induced toxicity, 14-3-3 ɛ or θ was transiently co-transfected with mutant SOD1. Overexpression of 14-3-3 ɛ reduced cell toxicity by 20% and that of 14-3-3 θ reduced cell toxicity by 22% (Fig. 3A and B). Overexpression of 14-3-3 ɛ and θ in mutant SOD1 cells reduced both the mitochondrial localization of Bax (Fig. 3C) and the release of cytochrome c (Fig. 3D), which may ameliorate the cell toxicity caused by mutant SOD1.
Overexpression of 14-3-3 ɛ and 14-3-3 θ reduced cell toxicity induced by mutant SOD1. (A) For PI staining, N2a cells were transiently co-transfected with HA-14-3-3 ɛ and either GFP-SOD1 or GFP-A4V. Forty-eight hours after transfection, cells were stained with PI (200 ng/ml). Overexpression of 14-3-3 ɛ reduced cell death by approximately 20% (P < 0.05) compared with controls in the presence of mutant SOD1 (A4V). Results reflect three independent experiments. Error bar reflects SEM. (B) For PI staining, N2a cells were transiently co-transfected with HA-14-3-3 θ and either GFP-SOD1 or GFP-A4V. Forty-eight hours after transfection, cells were stained with PI (200 ng/ml). Overexpression of 14-3-3 θ reduced cell death by approximately 22% (P < 0.001) compared with controls in the presence of mutant SOD1 (A4V). Results reflect three independent experiments. Error bar reflects SEM. (C) For the mitochondrial fraction, co-transfected N2a cells were resuspended in buffer D. The cell lysates were centrifuged to obtain cytosol and mitochondria fractions. Overexpression of 14-3-3 ɛ or 14-3-3 θ reduced the mitochondrial translocation of Bax in cells expressing mutant SOD1 (A4V). (D) Overexpression of 14-3-3 ɛ or 14-3-3 θ reduced the release of cytochrome c into the cytosol of cells expressing mutant SOD1 (A4V).
Disaggregation of mutant SOD1 by Hsp104 reduces cell death
We investigated whether disaggregation of mutant SOD1 releases 14-3-3 proteins and reduces cellular toxicity. For these purposes, we employed the molecular chaperone Hsp104, which promotes the reactivation of heat-damaged proteins in yeast. Our previous experiments showed that Hsp104 dissolved mutant SOD1 aggregates into soluble proteins (53). To examine the disaggregation of 14-3-3 with Hsp104, we performed FLIP assays. The loss of fluorescence from fluorescent-conjugated 14-3-3 ɛ and θ differed between bleached and unbleached regions in mutant GFP-A4V cells (Fig. 4A and C). However, the co-expression of Hsp104 resulted in a similar loss of fluorescence from14-3-3 ɛ and θ in bleached and unbleached regions (Fig. 4B and D). In contrast, in cells overexpressing wild-type SOD1, loss of 14-3-3 ɛ and θ fluorescence was similar between bleached and unbleached regions in the presence and absence of Hsp104 (Supplementary Material, Fig. S6A–D). These results show that Hsp104 dissolves sequestered 14-3-3 protein and mutant SOD1 aggregates. Moreover, soluble–insoluble fractionation experiments confirmed that Hsp104 decreased the interaction of both 14-3-3 ɛ and θ with mutant SOD1 (A4V) in the insoluble fraction (Fig. 4E).
Disaggregation of mutant SOD1 by Hsp104 reduced cell death. (A, B) For the FLIP assay, N2a cells were transiently transfected with GFP-A4V, Cerulean-14-3-3 ɛ, and either DsRed2 or DsRed2-Hsp104. Forty-eight hours after transfection, each sample was bleached and the fluorescence intensity of 14-3-3 ɛ was measured. Scale bar = 10 μm. (C, D) For the FLIP assay, N2a cells were transiently transfected with GFP-A4V, Cerulean-14-3-3 θ and either DsRed2 or DsRed2-Hsp104. Forty-eight hours after transfection, each sample was bleached and the fluorescence intensity of 14-3-3 θ was measured. Scale bar = 10 μm. (E) For immunoprecipitation experiments using fractionations of soluble and insoluble proteins, each fractionated sample was immunoprecipitated with 14-3-3 ɛ or 14-3-3 θ antibodies and immunoblotted with GFP antibody. (F, G) N2a cells were transiently transfected with either GFP-SOD1 and DsRed2-Hsp104 or GFP-A4V and DsRed2-Hsp104. Forty-eight hours after transfection, cells were immunostained with an antibody against Hsp70. Mutant SOD1 (A4V) aggregates co-localized with Hsp70 but not in cells coexpressing GFP-A4V and Hsp104. Scale bar = 10 μm. (H) Cell viability assays were performed by counting cells that contained Hsp104 and either wild-type SOD1 or mutant SOD1 (A4V). At 48 h after transfection, cells with co-localization of Hsp70 and aggregates were counted. Overexpression of Hsp104 reduced cell death by approximately 20% (P < 0.05) compared with controls in the presence of mutant SOD1 (A4V). Results reflect three independent experiments. Error bar indicates SEM. *P < 0.05; **P < 0.0005. Statistical analysis was performed using Student’s t-test. (I) For the mitochondrial fraction, co-transfected N2a cells were resuspended in buffer D. The cell lysates were centrifuged to obtain cytosolic and mitochondrial fractions. Overexpression of Hsp104 reduced the mitochondrial translocation of Bax and release of cytochrome c into the cytosol in cells expressing mutant SOD1 (A4V).
To assess cell death after dissolution of mutant SOD1 aggregates by Hsp104, we quantified the co-localization of SOD1 aggregates and Hsp70. A previous study reported that the presence of mutant SOD1 aggregates in the JUNQ compartment induces cellular toxicity (48). We co-transfected Hsp104 and either GFP-SOD1 or GFP-A4V into N2a cells and performed immunofluorescence assays using Hsp70 antibody. Hsp104 dissolved the mutant SOD1 aggregates into soluble proteins and reduced the co-localization of mutant SOD1 with Hsp70 (Fig. 4F and G). Based on those figures, by measuring the number of co-localization of mutant SOD1 aggregates with Hsp70, we observed that the dissolution of aggregated protein by Hsp104 reduced cell death by about 20% (Fig. 4H). Moreover, Hsp104 treatment decreased both the mitochondrial translocation of Bax and the release of cytochrome c into the cytosol in cells expressing mutant SOD1 (A4V) (Fig 4I). Taken together, these findings show that Hsp104 dissolves 14-3-3 ɛ and θ sequestered in mutant SOD1 aggregates and ameliorates cell death.
Mutant SOD1 is disaggregated by Hsp104 within the JUNQ compartment
A previous study showed that Hsp104 was predominantly located in IPOD in yeast (55). Therefore, we were curious to know the location in which mutant SOD1 aggregates are disaggregated by yeast Hsp104. We used two Hsp104 mutants that do not have disaggregase activity: Ax-Hsp104, which cannot bind ATP, and Trap-Hsp104, which can bind but cannot hydrolyze ATP (53). N2a cells were transiently co-transfected with either GFP-A4V or wild-type SOD1 and either Ax-Hsp104 or Trap-Hsp104. Ax-Hsp104 and Trap-Hsp104 did not co-localize with wild-type SOD1 and ubiquitin (Fig. 5A and C) but did with mutant SOD1 aggregates and ubiquitin (Fig. 5B and D). These results imply that Ax-Hsp104 and Trap-Hsp104 mutants cannot dissolve aggregates but still bind to mutant SOD1 in the same JUNQ compartment as wild-type Hsp104. We conclude that mutant SOD1 aggregates are disaggregated by yeast Hsp104 proteins within the JUNQ compartment.
Disaggregation of mutant SOD1 by Hsp104 occurred within the JUNQ compartment. (A, B) N2a cells were transiently transfected with DsRed2-Ax-Hsp104 and either GFP-SOD1 or GFP-A4V. Forty-eight hours after transfection, Ax-Hsp104 co-localized with mutant SOD1 (A4V) aggregates and ubiquitin but not with wild-type SOD1. Scale bar = 10 μm. (C, D) N2a cells were transiently transfected with DsRed2-Trap-Hsp104 and either GFP-SOD1 or GFP-A4V. Forty-eight hours after transfection, Trap-Hsp104 co-localized with mutant SOD1 (A4V) aggregates and ubiquitin but not with wild-type SOD1. Scale bar = 10 μm.
Discussion
Although mutant SOD1 aggregates are a hallmark of ALS, it is unclear whether these aggregates confer a toxic or defensive effect on cells (32,37). Many mechanisms, including excitotoxicity, ER stress, proteasome inhibition, and mitochondrial dysfunction, are considered to underlie the pathogenesis of ALS. The focus of this study is the role of sequestration of 14-3-3, an important protein for cell survival, in mutant SOD1 aggregates in the pathogenesis of ALS.
We propose a mechanism by which mutant SOD1 aggregates contribute to ALS pathogenesis (Fig. 6). Soluble mutant SOD1 proteins interact with 14-3-3 ɛ and θ in the cell. Mutant SOD1 proteins form aggregates; therefore, 14-3-3 ɛ and θ become insoluble in the JUNQ compartment. As a consequence, the amount of functional 14-3-3 ɛ and θ in the cytosol decreases, promoting the translocation of Bax into the mitochondria and inducing the release of cytochrome c into the cytosol. Although cells gather mutant SOD1 aggregates to protect themselves, cellular toxicity is induced by mutant SOD1-mediated 14-3-3 dysfunction.
Model for sequestration of 14-3-3 by mutant SOD1 aggregates. 14-3-3 inhibits mitochondrial translocation of Bax and is sequestered into mutant SOD1 aggregates by protein–protein interaction between 14-3-3 and mutant SOD1. Mutant SOD1 aggregates reduce cytosolic 14-3-3 by sequestering 14-3-3 in the JUNQ compartment. The lower concentration of soluble 14-3-3 impairs the localization of Bax to the cytosol. The resulting translocation of Bax into mitochondria induces the release of cytochrome c and cell death.
In this study, we used G93A transgenic mice, which exhibit various ALS phenotypes, including abnormal motor neuron morphology, decreased number of motor neurons, and abnormal mitochondrion morphology (56). However, we think that further studies using G37R transgenic mice are needed because G93A and G37R transgenic mice differ with respect to protein aggregation propensity, protein stability, disease onset, and disease progression (57,58), although G37R and G93A transgenic mice have similar disease phenotypes, pathological features, and increased dismutase activity.
The level of soluble 14-3-3 proteins, particularly the 14-3-3 ɛ and θ isoforms, in the spinal cord of ALS G93A transgenic mice was lower than that in the spinal cord of control mice. This decrease in concentration impaired the function of 14-3-3 in mutant SOD1-expressing cells. Although the expression of 14-3-3 mRNAs was not altered, the amount of insoluble 14-3-3 protein was higher in ALS mice. Furthermore, mutant SOD1 aggregates co-localized with 14-3-3 proteins in the spinal cord of ALS mice (Fig. 1). Taken together, these results indicate that the decrease in 14-3-3 in the spinal cords of ALS transgenic mice is not caused by a change in the transcriptional level but rather a decrease in the soluble protein level caused by sequestration with mutant SOD1 aggregates. We observed co-localization of mutant SOD1 aggregates and 14-3-3 proteins in the spinal cords of ALS transgenic mice on postnatal day 56 (8 weeks) by immunofluorescence, protein expression, and fractionation experiments (Fig. 1). A previous study observed mutant SOD1 positive inclusions on postnatal day 30 (P30) in ALS-model motor neurons (59). Therefore, we assume that the sequestration of 14-3-3 proteins into aggregates begins at least before P30 in the spinal cords of ALS transgenic mice and that neuronal cell death is induced after P60. Misfolded mutant SOD1 proteins might be removed by protein quality control systems such as the proteasome system and the autophagy system in in vivo motor neurons at P30, which would diminish the effect of aggregates on cell death. However, as the age increases, these quality control systems might be inhibited by oligomeric or aggregated SOD1 proteins and sequestering of various autophagic components (60,61). Finally, impaired protein homeostasis in ALS transgenic mice appears to cause motor neuron death after P60. In contrast, in N2a cells overexpressing mutant SOD1, aggregates form 48 h after transfection, impairing protein homeostasis and rapidly inducing cell death.
The co-localization of 14-3-3 with mutant SOD1 aggregates has not been investigated during ALS pathogenesis. In this study, we confirm the co-localization and interaction of mutant SOD1 aggregates and 14-3-3 proteins in N2a cells. The amount of insoluble 14-3-3 protein increased when mutant SOD1 was expressed, which is a well-known phenomenon (62). We assumed that mutant SOD1 forms aggregates and sequesters 14-3-3 proteins through protein–protein interactions when still in the soluble state. As expected, mutant SOD1 aggregates co-localized and interacted with pan 14-3-3 (Supplementary Material, Fig. S2). Using several 14-3-3 isoform-specific antibodies, we found that 14-3-3 ɛ and θ bound to SOD1 aggregates in mutant SOD1-expressing cells and interacted with mutant SOD1 in soluble and insoluble fractions (Fig. 2). These results suggest that 14-3-3 proteins and soluble mutant SOD1 first interact in the early stage; an insoluble aggregation stage follows, during which 14-3-3 proteins are sequestered. Furthermore, 14-3-3 isoform proteins co-localized with mutant SOD1 aggregates in the JUNQ compartment (Fig. 2). Previous studies have shown that 14-3-3 ɛ and θ promote cell survival by interacting with Bax and Bad (43,44). Interestingly, we observed reduced levels of these isoforms in ALS transgenic mice (Fig. 1B). These findings suggest that sequestration of these isoforms in mutant SOD1 aggregates affects 14-3-3-mediated cell survival.
Although the cellular toxicity of mutant SOD1 is well-known (54), the mechanism underlying it has not been fully elucidated. We found that 14-3-3 ɛ and θ are associated with cell survival; therefore, we investigated the mitochondrial localization of Bax and the release of cytochrome c in cells expressing mutant SOD1. Mutant SOD1 induced mitochondrial localization of Bax and the release of cytochrome c into the cytosol (Supplementary Material, Fig. S5). This relationship between mutant SOD1 expression and intrinsic mitochondria-mediated cell-death pathways may play an important role in the pathogenesis of ALS. Mutant SOD1 is reported to interact with the dynein–dynactin complex, forming large inclusions (63). In addition, dynein-mediated retrograde transport is slower in ALS transgenic mice than in control mice (64). Given that misfolded proteins are moved to the JUNQ compartment by retrograde transport, it is possible that slowed retrograde transport caused by mutant SOD1 aggregates also contributes to SOD1 toxicity. Reduced levels of 14-3-3 ɛ and θ caused cell death, and the overexpression of 14-3-3 ɛ and θ reversed this effect. Moreover, mitochondrial localization of Bax and cytochrome c release into the cytosol were reduced (Fig. 3). These results suggest that although overexpressed 14-3-3 proteins are recruited into mutant SOD1 aggregates, such overexpression promotes neuron survival in ALS.
Hsp104 dissolves mutant SOD1 aggregates (53), but it remains unclear whether the disaggregation of mutant SOD1 affects cell toxicity. We showed that Hsp104 dissolves sequestrated 14-3-3 ɛ and θ as well as mutant SOD1 aggregates. Moreover, Hsp104 reduced mutant SOD1-induced cell death (Fig. 4). Hsp104 overexpression partially rescued cell toxicity, as shown in Figure 4H. We think that this partial rescue is related to the ability of Hsp104 to dissolve aggregates. We previously reported that Hsp104 dissolves mutant SOD1 aggregates to trimers but not dimers (53). Although some aggregates are dissolved and refolded into native proteins, some aggregates remain as oligomeric structures. Therefore, some sequestered 14-3-3 will remain in oligomers, allowing for only a partial rescue of cell toxicity. At present, we are further investigating the removal of mutant SOD1 aggregates using an Hsp104 variant. Although a previous study showed that Hsp104 was predominantly located in IPOD (55), we found Hsp104 dissolved mutant SOD1 aggregates located in the JUNQ compartment (Fig. 5). Hsp70 is located in the JUNQ compartment and is a cochaperone of Hsp104; this relationship may explain as to why we observed Hsp104 in the JUNQ compartment but not in IPOD in our study. Recovery of proteasome function and disaggregation might be crucial for cell survival.
We described the interactions between mutant SOD1 aggregates and 14-3-3 proteins. Mutant SOD1 aggregates may sequester other important proteins in addition to 14-3-3, leading to functional loss of these proteins. Given the cellular toxicity of mutant SOD1 aggregates, disaggregation would be one way to reduce mutant SOD1 toxicity. This strategy may contribute to the identification of novel therapeutics for ALS.
Materials and Methods
Cell culture and transfection
The mouse neuroblastoma cell line N2a was maintained in DMEM supplemented with 8% fetal bovine serum, streptomycin (100 μg/ml), and penicillin (100 U/ml). Transfections were performed using polyethyleneimine (Sigma-Aldrich, St. Louis, MO, USA), as previously described (65). DNA was combined with transfection reagent at a ratio of 1:3 and incubated in serum-free DMEM for 30 min. After incubation, DNA–reagent complexes were applied to the cells.
Plasmid constructs
SOD1-FLAG, SOD1-GFP, and Hsp104-DsRed2 constructs were described previously (53,66). The cDNA fragments for 14-3-3 ɛ and θ were amplified by PCR using pfu polymerase (Fermentas, Waltham, MA, USA) and first-strand human cDNAs as templates. HA-14-3-3 ɛ and HA-14-3-3 θ constructs were generated by inserting amplified HA-14-3-3 ɛ and HA-14-3-3 θ fragments into HA-pcDNA3.0 plasmids at the EcoRI and NotI restriction sites, respectively. Cerulean-14-3-3 ɛ and Cerulean-14-3-3 θ constructs were generated by inserting the amplified fragments into Cerulean-N3 plasmids at the NheI and BamHI restriction sites, respectively.
Animals
Transgenic SOD1-G93A mice used in this study were generated by breeding male hemizygous SOD1-G93A mice [B6SJL-Tg (SOD1-G93A) 1 Gur/J] with female B6SJLF1/J hybrids. Their non-transgenic offspring served as controls. The genotyping of SOD1-G93A mice was performed by PCR, as previously reported (67). Male hemizygous SOD1-G93A mice and female B6SJLF1/J hybrids were both purchased from Jackson Laboratories (Bar Harbor, ME, USA). All experiments with mice were permitted by the Institutional Animal Care and Use Committee at Korea University.
Tissue preparation
Mice were anesthetized and perfused transcardially with phosphate-buffered saline (PBS), followed by 3% paraformaldehyde. Spinal cords were isolated and post-fixed and stored in 30% sucrose before being frozen with dry ice. The frozen spinal cords were dissected into cervical, thoracic, and lumbar segments. The dissected lumbar spinal cord tissues were embedded in optimal cutting temperature (OCT) compound (Sakura, Torrance, CA, USA) and cut coronally at 30 μm using a cryostat (68).
Immunofluorescence
Cells were fixed in 4% formaldehyde for 15 min and permeabilized with 0.1% Triton X-100. After incubation in 2% BSA, cells were incubated overnight at 4 °C in one of the following primary antibodies: anti-pan 14-3-3, anti-14-3-3 β, anti-14-3-3 ɛ, anti-14-3-3 γ, anti-14-3-3 θ, anti-14-3-3 ζ, anti-14-3-3 η, anti-ubiquitin (1:100, Santa Cruz, CA, USA), or anti-Hsp70 (1:50, Enzo Life Sciences, Farmingdale, NY, USA). Cells were washed three times in PBS, then incubated in Alexa Fluor 594 (Life Technologies, Carlsbad CA, USA) for 1 h at room temperature. After washing three times with PBS, stained cells were mounted onto glass slides using fluorescent mounting medium (Dako, Santa Clara, CA, USA). For tissue staining, spinal cord sections were incubated in 3% H2O2 and 0.25% Triton X-100 for 10 min. After incubation in 10% horse serum for 1 h at room temperature, the sections were incubated overnight at 4 °C with anti-SOD1 (1:50, Santa Cruz) and anti-pan 14-3-3 (1:50, Santa Cruz) primary antibodies. After washing three times, the sections were incubated with Alexa Fluor 488 (Life Technologies) and Alexa Fluor 594 (Life Technologies) secondary antibodies for 2 h at room temperature, respectively. Fluorescent images were captured using an LSM-700 confocal microscope (Carl Zeiss, Oberkochen, Germany).
Flip
FLIP experiments were performed as described previously (53). At 48 h after transfection, photobleaching of live cells was performed using a 405-nm laser at 50–100% output for ten iterations. Fluorescent images were captured using an LSM-700 confocal microscope (Carl Zeiss).
Immunoprecipitation and western blotting
Immunoprecipitation experiments were performed as described previously (69). N2a cells were transiently transfected with GFP-tagged wild-type SOD1 or mutant SOD1. At 48 h after transfection, cells were harvested and lysed on ice for 60 min in 1% NP-40 lysis buffer supplemented with protease and phosphatase inhibitors. Cell lysates were centrifuged at 17,000 × g for 20 min at 4 °C, and the supernatants were incubated with anti-pan 14-3-3 or 14-3-3 ɛ or 14-3-3 θ antibodies (Santa Cruz) overnight at 4 °C. Protein G-Sepharose beads (GE Healthcare Life Sciences, Chicago, IL, USA) were added, and the bead-bound proteins were separated by sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE). Separated proteins were transferred to nitrocellulose membranes and probed with anti-GFP antibody (1:3000) (Santa Cruz). The immune complexes were detected using an enhanced chemiluminescence (ECL) immunoblotting system as described by the manufacturer (GE Healthcare Life Sciences).
Soluble–insoluble fraction
Detergent soluble–insoluble fraction experiments were performed as described previously (62). Cells were harvested in PBS and centrifuged. The cells were resuspended in 100 μl 1× TEN (10 mM Tris, 1 mM EDTA, and 100 mM NaCl), mixed with an equal volume of 2× extraction buffer 1 (10 mM Tris, 1 mM EDTA, 100 mM NaCl, 1% Nonidet P-40, and protease inhibitor mixture), and then sonicated. The cell lysate was centrifuged for 5 min at 100,000 × g in a Beckman TLR100.3 ultracentrifuge to separate the detergent-insoluble pellet (P1) from the supernatant (S1). The P1 was resuspended in 200 μl 1× extraction buffer 2 (10 mM Tris, 1 mM EDTA, 100 mM NaCl, 0.5% Nonidet P-40, and protease inhibitor mixture) and sonicated. The resuspended pellet was then centrifuged for 5 min at 100,000 × g in a Beckman TLR100.3 ultracentrifuge to separate the second pellet (P2) from the supernatant. P2 was resuspended in 200 μl buffer 3 (10 mM Tris, 1 mM EDTA, 100 mM NaCl, 0.5% Nonidet P-40, 0.25% SDS, 0.5% deoxycholic acid, and protease inhibitor mixture) by sonication. S1 indicates the soluble fraction and P2 indicates the insoluble fraction. Protein concentration was measured in S1 and P2 fractions using the Bradford assay (Bio-rad, Hercules, CA, USA).
Isolation of mitochondria
Mitochondria were isolated as described previously (70). Harvested cells were broken down using a G26 needle in buffer D (250 mM sucrose, 20 mM HEPES, 10 mM KCl, 5 mM MgCl2, 1 mM EDTA, 1 mM EGTA, and 10 mM Tris–Cl, pH 7.5) containing protease inhibitors and 0.02% digitonin. The cell lysates were centrifuged at 900 × g for 2 min at 4 °C. After centrifugation, the supernatants were further centrifuged at 17,000 × g for 30 min at 4 °C to separate pellets (mitochondrial fraction) and supernatants (cytosolic fraction).
PI staining
Cells were fixed with 4% formaldehyde in PBS for 15 min. After washing three times with PBS, the fixed cells were incubated with RNase A (100 μg/ml) for 20 min and incubated with PI (200 ng/ml). After washing three times with PBS, stained cells were counted using an Axio fluorescence microscope (Carl Zeiss).
Statistical analysis
Differences between the various experimental groups were calculated using Student’s two-tailed t-test. P < 0.05 was considered significant.
Supplementary Material
Supplementary Material is available at HMG online.
Conflict of Interest statement. None declared.
Funding
This work was supported by the Korea Healthcare Technology R&D Project, Ministry for Health, Welfare & Family Affairs, Republic of Korea (No. A120340), National R&D Program for Cancer Control, Ministry of Health & Welfare, Republic of Korea (No. 1320010), National Research Foundation of Korea Grants, Ministry of Science, ICT and Future Planning, Republic of Korea (NRF-2015R1A4A1041919), and National Research Foundation of Korea (NRF) grant (MEST) (NRF-2015R1A2A2A01003516).
