Profilin 1 (PFN1) is an actin monomer-binding protein essential for regulating cytoskeletal dynamics in all cell types. Recently, mutations in the PFN1 gene have been identified as a cause of familial amyotrophic lateral sclerosis (ALS). The co-aggregation of PFN1 bearing mutations that cause ALS with TDP-43 (a key molecule in both sporadic and some familial forms of ALS), together with the classical TDP-43 pathology detected in post-mortem tissues of patients with autosomal dominant PFN1 mutation, imply that gain-of-toxic-function of PFN1 mutants is associated with the onset of ALS. However, it remains unknown how PFN1 mutants cause ALS. We found mutant PFN1 that causes ALS formed cytoplasmic aggregates positive for ubiquitin and p62, and these aggregates sequestered endogenous TDP-43. In cells harboring PFN1 aggregates, formation of aggresome-like structures was inhibited in the presence of proteasome inhibitor, and conversion of LC3-I to LC3-II was suppressed in the presence of lysosome inhibitor. Further, insoluble TDP-43 was increased in both cases. Co-expression of ALS-linked mutant PFN1 and TDP-43 increased insoluble and phosphorylated TDP-43 levels. The C-terminal region of TDP-43, essential for aggregation of TDP-43, was also indispensable for the interaction with PFN1. Interestingly, insoluble fractions prepared from cells expressing ALS-linked mutant PFN1 functioned as a seed to induce accumulation and phosphorylation of TDP-43, indicating that TDP-43 accumulated in the presence of the PFN1 mutants is converted to prion-like species. These findings provide new insight into the mechanisms of neurodegeneration in ALS, suggesting that gain-of-toxic-function PFN1 gene mutation leads to conformational change of TDP-43.
Amyotrophic lateral sclerosis (ALS) is a devastating neurodegenerative disease characterized by muscle weakness due to the rapidly progressive degeneration of motor neurons. Although the molecular mechanisms of ALS pathogenesis remains unknown, it has been shown that the nuclear protein TAR DNA-binding protein of 43 kDa (TDP-43) accumulates as a major component in the ubiquitin-positive inclusions in the affected regions of the spinal cord and brain of patients with both sporadic and many familial forms of ALS (1,2). The identification of mutations in the TDP-43 gene (TARDBP) in familial forms of ALS and frontotemporal lobar degeneration (FTLD) indicates that TDP-43 has a pivotal role in the pathogenesis of these diseases (3,4). Biochemical and morphological analyses revealed that filamentous, detergent-insoluble, abnormally phosphorylated (Ser-403/404 and Ser-409/410), partially fragmented and ubiquitinated TDP-43 is accumulated in the brains of these patients (5,6). Furthermore, the pathological TDP-43 has been shown to have prion-like properties, i.e. it can self-propagate by converting normal TDP-43 to an abnormal form (7).
Recently, mutations in the profilin 1 (PFN1) gene have been identified as a cause of familial ALS (8). PFN1 is an actin monomer-binding protein essential for regulating cytoskeletal dynamics in all cell types (9). PFN1 also has other binding domains for phosphoinositides and proline-rich proteins, and is involved in various cellular functions (10,11). Cytoskeletal disruption and altered stress granule dynamics were detected in cells expressing ALS-linked PFN1 mutations associated with the loss of function (8,12). On the other hand, both wild-type (WT) PFN1 and ALS-linked PFN1 mutants were co-localized with TDP-43 (8), and classical TDP-43 pathology was detected in the post-mortem tissues of patients with autosomal dominant PFN1 mutation (13), suggesting that a gain-of-toxic-function by mutant PFN1 is associated with the onset of ALS. Therefore, we investigated the relationship between ALS-linked PFN1 mutations and formation of pathological TDP-43 using human neuroblastoma SH-SY5Y cells.
ALS-linked PFN1 mutants form cytoplasmic aggregates positive for p62 and ubiquitin in cultured cells
To examine the effect of the PFN1 mutations, WT or ALS-linked mutant PFN1 (C71G, M114T, E117G or G118V) was transiently expressed in SH-SY5Y cells, and cells were subjected to biochemical and immunocytochemical analyses at 3 days after transfection. In biochemical analyses of transfected cells, the levels of WT and E117G PFN1 were higher than these of PFN1 bearing mutations that cause ALS (C71G, M114T or G118V) (Fig. 1A). Similar results were obtained in Escherichia coli (Supplementary Material, Fig. S1). Most of the WT and E117G PFN1 was recovered in Sarkosyl (Sar)-soluble fraction (sup), while PFN1 with mutations that cause ALS was found in the Sar-insoluble fraction (ppt) (Fig. 1B). Confocal microscopic studies and biochemical analyses showed that WT and E117G PFN1 were diffusely localized predominantly in the cytoplasm, but aggregates were hardly detected (Fig. 1C–E and Supplementary Material, Fig. S2). On the other hand, C71G PFN1 was mainly localized in the cytoplasm and aggregates were detected in over 70% of the C71G-positive cells. M114T and G118V PFN1 showed mixed localization, either diffuse or concentrated in the cytoplasm, and aggregates were detected in ∼40% of the M114T- and G118V-positive cells. C71G and M114T PFN1 formed clear and distinct aggregates, while most G118V aggregates were smaller and subtler than those of C71G and M114T. These PFN1 aggregates were also positive for ubiquitin and p62 (Fig. 2A and B). These results indicate that PFN1 with mutations that cause ALS is more aggregation-prone than WT, suggesting that these PFN1 aggregates may have a gain-of-toxic-function.
PFN1 aggregates affect endogenous TDP-43 function
Co-immunoprecipitation study using cells transfected with WT PFN1 revealed that PFN1 interacts with endogenous TDP-43, but not with fused in sarcoma (FUS), another nuclear RNA-binding protein associated with ALS/FTLD (Fig. 3A). To assess the influence of the PFN1 mutations on TDP-43, cells were transfected with WT or ALS-linked mutant PFN1 and analyzed by immunostaining for PFN1 and TDP-43. PFN1 and TDP-43 were co-aggregated in the cytoplasm of some cells harboring PFN1 aggregates, and TDP-43 was depleted from the nucleus (Fig. 3B). Additionally, expression of HAS2, which is downstream of TDP-43 (14), was significantly decreased in cells transfected with PFN1 with mutations that cause ALS, compared with pcDNA3, WT or E117G PFN1 transfected cells (Fig. 3C). Red fluorescence from acridine orange staining showed that RNA and/or single-stranded DNA were sequestered into PFN1 aggregates (Supplementary Material, Fig. S3). These results suggest that PFN1 aggregates sequester endogenous TDP-43, and inhibit its function.
Insoluble TDP-43 in cells harboring PFN1 aggregates is increased by inhibition of proteasome or lysosome
In most neurodegenerative diseases, deficits in protein degradation systems, including the ubiquitin-proteasome and autophagy-lysosome systems, are suggested to facilitate formation of abnormal aggregates (15,16). Therefore, we investigated the association of ALS-linked PFN1 mutants with these degradation systems. We first checked the effect of MG132, a proteasome inhibitor, on the proteasomal degradation system using cells transfected with GFP-CL1, which is specifically degraded by proteasome (17). GFP-CL1 levels were increased in an MG132 concentration-dependent manner (Fig. 4A). Then, we investigated the effect of ALS-linked mutant PFN1 on the proteasomal degradation system using cells double-transfected with GFP-CL1 and pcDNA3, WT or ALS-linked mutant PFN1. GFP-CL1 levels were not increased in WT or ALS-linked mutant PFN1 transfectants (Fig. 4B). Additionally, we investigated the effect of MG132 on TDP-43 aggregation in cells transfected with WT or ALS-linked mutant PFN1. Although TDP-43 levels in the Sar-insoluble fraction were similar among groups in the absence of MG132, Sar-insoluble TDP-43 levels in cells expressing PFN1 with mutations that cause ALS were higher in the presence of MG132 (Fig. 4C). Immunostaining studies showed that ubiquitin and p62 immunoreactivities were increased, and aggresome-like structures positive for p62 and γ-tubulin were formed in the presence of MG132 (Supplementary Material, Fig. S4). However, formation of aggresome-like structures was inhibited in cells harboring PFN1 aggregates. On the other hand, PFN1 aggregates were co-localized with ubiquitin, p62 and TDP-43 (Fig. 5A and B).
We next investigated the effect of bafilomycin A1, a lysosome inhibitor, on the lysosomal degradation system. p62 is selectively degraded via the autophagy-lysosomal pathway, and conversion of LC3-I to LC3-II is a marker of autophagic activity. As shown in Figure 6A, p62 levels and conversion of LC3-I (over 15 kDa) to LC3-II (under 15 kDa) were increased in a bafilomycin A1 concentration-dependent manner. Then, we investigated the effect of ALS-linked mutant PFN1 on the autophagy-lysosomal system. p62 levels were mildly increased in cells expressing C71G and M114T. Increased conversion of LC3-I to LC3-II upon addition of bafilomycin A1 was suppressed in cells expressing PFN1 with mutations that cause ALS, possibly because LC3 was sequestered into PFN1 aggregates even in the absence of bafilomycin A1 (Fig. 6B and C). Then, we investigated the effect of bafilomycin A1 on insoluble TDP-43 level in cells transfected with WT or ALS-linked mutant PFN1. As shown in Figure 6D, Sar-insoluble TDP-43 levels in cells expressing C71G and M114T PFN1 were increased by bafilomycin A1 treatment. These results demonstrate that degradation is involved in the clearance of insoluble TDP-43 induced by PFN1 aggregates.
Insoluble TDP-43 and phosphorylated TDP-43 are increased in cells co-expressing ALS-linked mutant PFN1 and TDP-43
To further investigate the relationship between the gain-of-toxic-function of PFN1 and pathological alterations of TDP-43, we analyzed cells co-expressing PFN1 and TDP-43. As shown in Figure 7, double-immunostaining revealed that some TDP-43 aggregates co-localized with PFN1 aggregates were phosphorylated in cells co-expressing PFN1 with mutations that cause ALS and TDP-43. In immunoblot analyses, we also observed that the levels and smearing of TDP-43 in the Sar-insoluble fraction were increased in these cells. Furthermore, phosphorylated TDP-43 (pTDP-43) was also detected in the Sar-insoluble fractions of cells expressing C71G or M114T and TDP-43 (Fig. 8A). To address whether formation of these pTDP-43 aggregates was accelerated by the change in TDP-43 localization, we investigated pTDP-43 levels in cells transfected with PFN1 (WT, C71G or M114T) and a deletion mutant of TDP-43 lacking the nuclear localization signal (ΔNLS) (Supplementary Material, Fig. S5). Accumulation of pTDP-43 in Sar-insoluble fractions of cells expressing C71G or M114T and ΔNLS was higher than that in cells expressing C71G or M114T and normal TDP-43 (Fig. 8B), indicating that cytoplasmic TDP-43 was preferentially recruited into PFN1 aggregates and phosphorylated. These results indicate that expression of autosomal dominant ALS-linked mutant PFN1, C71G or M114T, triggers cytoplasmic pTDP-43 aggregate formation resembling the pathological hallmark of ALS.
We also checked the cytotoxicity in cells co-expressing PFN1 and TDP-43. At 3 days after transfection of constructs, cell death assay of transfected cells was performed by the trypan blue-exclusion method. Viability of cells expressing only ALS-linked mutant PFN1 was unaffected, but that of cells co-expressing TDP-43 and C71G, M114T or G118V PFN1 was significantly decreased, indicating that cells containing cytoplasmic TDP-43 aggregates were more damaged (Fig. 8C). Overall, these results indicate that the co-expression of ALS-linked mutant PFN1 with TDP-43 induces cytoplasmic formation of TDP-43 aggregates that are partially phosphorylated and might be more toxic than PFN1 deposits.
Carboxyl-terminal portion of TDP-43 is associated with the C71G PFN1 aggregates
To assess which domains are important for TDP-43 aggregate formation by PFN1 with mutations that cause ALS, we constructed truncated forms of TDP-43 tagged with GFP (Supplementary Material, Fig. S5), and co-transfected them together with PFN1 (WT or C71G). Immunoblot analyses of lysates from these cells showed that co-expression of a C-terminal fragment (162–414 residues: 162C) with C71G PFN1 increased the level of Sar-insoluble pTDP-43 pellets, but co-expression of an N-terminal fragment (1–273 residues: N273) did not (Fig. 9A–C). Next, we investigated the localization of these GFP-tagged TDP-43 constructs in cells co-transfected with WT or C71G PFN1 (Fig. 9D). FL was localized to the nucleus in cells co-expressing WT PFN1, but was localized to the cytoplasm and aggregated in cells co-expressing C71G PFN1. 162C was localized to the cytoplasm regardless of co-expression with WT or C71G PFN1, but it was co-aggregated in cells co-expressing C71G PFN1. N273 was localized in the nucleus even when it was co-expressed with C71G PFN1. It is noteworthy that C71G PFN1 aggregates were co-localized with endogenous TDP-43 (Fig. 4B), FL and 162C aggregates but not with N273. These results indicate that the C-terminal portion of TDP-43 interacts with C71G PFN1 aggregates.
PFN1-induced insoluble TDP-43 functions as a seed for intracellular aggregation of TDP-43
Recently, it has been shown that pTDP-43 accumulated in ALS or FTLD brains has prion-like properties and functions as a seed for intracellular TDP-43 aggregation in cultured cells (7). Thus, we investigated whether the insoluble TDP-43 induced by PFN1 aggregates has prion-like properties and induces intracellular seeded aggregation of TDP-43. Sar-insoluble fraction was prepared from cells co-transfected with TDP-43 and WT PFN1 (WT + T) or C71G PFN1 (71 + T), and introduced as a seed into cells expressing TDP-43. Time-course immunoblot analysis showed that pTDP-43 was predominantly deposited in the Sar-insoluble fraction from cells exposed to 71 + T as a seed for 2 days, but was not detected at 4 h or 1 day (Fig. 10A). Immunocytochemical analysis showed that only cells exposed to 71 + T as a seed formed inclusion-like structures positive for pTDP-43 (Fig. 10B). Furthermore, Sar-insoluble fraction prepared from cells transfected with pcDNA, WT PFN1 or ALS-linked mutant PFN1 was introduced into cells expressing TDP-43 as seeds. Sar-insoluble fraction from cells transfected with ALS-linked PFN1 induced accumulation of pTDP-43, implying that the fraction contained endogenous TDP-43 that can function as a seed for intracellular TDP-43 (Fig. 10C). These findings indicate that Sar-insoluble form of TDP-43 induced by ALS-linked mutant PFN1 expression can function as a seed for aggregation of intracellular TDP-43, and this function is enhanced by co-expression of ALS-linked mutant PFN1 and TDP-43. Co-expression of C71G PFN1 with TDP-43 increased the seeding function for aggregation of intracellular TDP-43 compared with the expression of C71G PFN1 only, suggesting that PFN1 aggregates themselves have little activity as a seed, but provide a scaffold for the formation of conformationally altered types of TDP-43.
The PFN1 gene was shown to be involved in familial ALS (8), but the mechanism through which ALS-linked PFN1 mutations cause ALS is poorly understood. The previous study showed that axonal outgrowth and growth cone size are decreased in cells transfected with ALS-linked mutant PFN1, probably because of decreased levels of binding between actin and the mutant PFN1 (8). We also found PFN1 with mutations that cause ALS did not increase globular actin compared with WT or E117G PFN1, although both WT and ALS-linked mutant PFN1 were co-localized with actin (Supplementary Material, Figs S6 and S7). A role of PFN1 in stress granule dynamics was also reported (12). These results suggest that ALS-linked PFN1 mutations cause onset of ALS via a loss-of-function mechanism. On the other hand, PFN1 hetero-deficient mice were viable, and showed normal longevity and a normal phenotype under standard mouse colony conditions. Filamentous actin levels and cytoskeletal architecture of cells from the hetero-deficient mice were also similar those from WT mice (18). Moreover, ALS-linked PFN1 mutations are autosomal-dominant. Therefore, a gain-of-toxic-function mechanism is associated with the cause of this disease.
Our present results show that PFN1 with mutations that cause ALS forms cytoplasmic aggregates positive for ubiquitin and p62, and the aggregates sequester endogenous TDP-43. TDP-43 inclusions positive for ubiquitin and p62 are a significant pathological feature in ALS/FTLD (19,20). Therefore, PFN1 aggregates might cause formation of TDP-43 inclusions in familial ALS associated with PFN1 gene mutations. Actually, in the presence of proteasome or lysosome inhibitor, Sar-insoluble TDP-43 levels in cells express PFN1 with mutations that cause ALS were higher than that in cells expressing WT PFN1. Then, Sar-insoluble TDP-43 levels in cells co-transfected with ALS-linked mutant PFN1 and TDP-43 were higher than those in cells co-transfected with WT PFN1 and TDP-43. Moreover, TDP-43 aggregation induced by PFN1 aggregates was accelerated in cells expressing TDP-43 localized to the cytoplasm. Although the mechanism through which TDP-43 is sequestered into PFN1 aggregates is largely unknown, TDP-43 might be captured in PFN1 aggregates due to its interaction with PFN1. Interestingly, pTDP-43 accumulation increased Sar-insoluble endogenous PFN1. Taken together, these results indicate that formation of aggregates of either PFN1 or TDP-43 triggers sequestration of the other protein. Co-aggregation between PFN1 and TDP-43 might accelerate the progression of TDP-43 proteinopathy.
We have shown that the C-terminal region of TDP-43 is associated with PFN1 aggregates. The C-terminal region of TDP-43 is essential for aggregation of TDP-43, and most TDP-43 mutations that cause ALS are found in the C-terminal region (21,22). Our data also demonstrate that the C-terminal region has a critical role in the aggregation process of TDP-43 caused by PFN1 aggregates.
The mechanism of neurodegeneration with TDP-43 accumulation in the cytoplasm involves loss-of-function of TDP-43 in the nucleus and/or cytotoxicity of TDP-43 aggregates in the cytoplasm (23). We observed that cell death in cells co-transfected with ALS-linked mutant PFN1 and TDP-43 was accelerated compared with that in cells co-transfected with WT PFN1 and TDP-43. Since pTDP-43 was deposited in cells co-transfected with PFN1 with mutations that cause ALS and TDP-43, the accumulation of pTDP-43 might play a key role in induction of cell death. We previously reported that sequestration of RNA polymerase II into pTDP-43 aggregates causes cell death via a non-apoptotic mechanism in SH-SY5Y cells and FTLD-TDP patients (24). Therefore, PFN1 aggregates might induce similar aberrations of RNA metabolism in ALS patients (25–27).
The Sar-insoluble fraction sampled from cells transfected with ALS-linked mutant PFN1 functioned as a seed to induce accumulation and phosphorylation of TDP-43. We previously reported that introduction of Sar-insoluble fraction from ALS or FTLD-TDP patients into SH-SY5Y cells transfected with TDP-43 induces aggregation of phosphorylated and ubiquitinated TDP-43 (7). TDP-43 sequestered into PFN1 aggregates might function as a seed, and promote TDP-43 aggregation in a self-template manner.
E117G mutation was detected in sporadic ALS and FTLD patients (28–32), and identified as a moderate risk factor (13,28). In the present study, the propensity of E117G PFN1 mutant to function as a seed was closer to that of WT PFN1 than to that of other ALS-linked mutant PFN1. Actually, the E117G mutation has little impact on the stability and structure of PFN1 (33). However, the Sar-insoluble fraction from cells expressing E117G PFN1 functioned as a seed for intracellular TDP-43. Therefore, this mutation might accelerate TDP-43 accumulation by promoting a conformational change in TDP-43 to an abnormal form (13).
Our results show that the gain-of-toxic-function mutation of PFN1 causes degenerative changes in association with increased TDP-43 aggregation. PFN1 mutants form cytoplasmic aggregates, sequester endogenous TDP-43, and convert TDP-43 from normal form to a prion-like abnormal form, which is able to function as a seed for the propagation of pathological TDP-43 (Fig. 11). Sequestration of TDP-43 into PFN1 aggregates is similar to that of normal TDP-43 into the inclusions in ALS/FTLD patients. The results of this study provide evidence that actin-binding protein PFN1 and TDP-43 interact with each other, and ALS-linked mutant PFN1 causes TDP-43 aggregation, suggesting that alteration of PFN1 might be involved in the pathogenesis of sporadic ALS and FTLD (Supplementary Material, Figs S8 and S9), and that PFN1 aggregates act as a scaffold at the early phase of TDP-43 aggregation. Our model should be useful to study the formation of pathological TDP-43 and to develop new therapies and pharmaceuticals for TDP-43 proteinopathy. This study highlights the importance of the formation of conformationally altered types of TDP-43 in ALS pathophysiology.
Materials and Methods
Antibodies and reagents
Anti-PFN1 antibody (ab50667) was purchased from Abcam. Monoclonal anti-TDP-43 antibodies (60019-2) were purchased from ProteinTech. Anti-GAPDH antibody (MAB374) and anti-ubiquitin antibody (MAB1510) were purchased from Millipore. A monoclonal anti-mouse p62 antibody (610833) was purchased from BD Biosciences. A polyclonal anti-guinea pig p62 antibody (GP62-C) was purchased from Progen. Anti-α-tubulin antibody (T5168), anti-FUS antibody (HPA008784) and anti-γ-tubulin antibody (T6557) were purchased from Sigma. A polyclonal anti-microtubule-associated protein light chain 3 (LC3) antibody (NB100-2220) was purchased from Novus Biologicals. A monoclonal anti-green fluorescent protein (GFP) antibody (M048-3) and anti-LC3 antibody (M152-3) were purchased from MBL. Monoclonal and polyclonal anti-pTDP-43 antibodies specific for Ser409/410 were prepared as described (34,35). Alexa Fluor 488 or 568-conjugated secondary antibodies were purchased from Life Technologies. MG132 was purchased from Peptide Institute. Bafilomycin A1 was purchased from Wako.
Construction of plasmids
To construct human PFN1-expressing plasmid, a cDNA encoding full-length human PFN1 was amplified from cDNA extracted from human neuroblastoma SH-SY5Y using the following primers (PFN1 forward; 5′-TTTGGATCCATGGCCGGGTGGAACGC-3′, PFN1 reverse; 5′-GGGGGAATTCTCAGTACTGGGAACGCCG-3′). The amplified fragment was digested with BamHI and EcoRI and cloned into the same cleavage sites of the vector [pc DNA3.1 (+), Invitrogen]. To construct mutant PFN1-expressing plasmids, we used a site-directed mutagenesis kit (Strategene) to substitute Cys71 to Gly (C71G), Met114 to Thr (M114T), Glu117 to Gly (E117G) and Gly118 to Val (G118V). All constructs were verified by DNA sequencing. Non-tagged TDP-43 expressing plasmid [full-length TDP-43 (FL) and deletion mutant of 78–84 residues (ΔNLS)] and GFP-tagged TDP-43 expressing plasmid [1–414 (FL), 1–273 (N273) and 162–414 (162C)] were constructed as previously described (36,37) (Supplementary Material, Fig. S5).
Cell culture and expression of plasmids
SH-SY5Y cells, not neuronally differentiated, were cultured in Dulbecco's modified Eagle's medium (DMEM)/F12 medium (Sigma) supplemented with 10% (v/v) fetal calf serum, penicillin–streptomycin–glutamine (Gibco) and MEM non-essential amino acids solution (Gibco). The cells were maintained at 37°C under a humidified atmosphere of 5% (v/v) CO2. They were grown to 50% confluence in collagen-coated six-well plates for transient expression and then transfected with plasmids using X-tremeGENE 9 DNA Transfection Reagent (Roche) according to the manufacturer's instructions. Under our experimental conditions, the efficiency of transfection was around 30%.
Extraction of proteins and immunoblotting
SH-SY5Y cells transfected transiently with the expression plasmids were grown in collagen-coated six-well plates for the indicated period. For sequential extraction of protein, cells were harvested and lysed in 1% (w/v) Sarkosyl in A68 [10 mm Tris–HCl buffer, pH 7.5, 0.8 M NaCl, 1 mm ethylene glycol bis(β-aminoethyl ether)-N,N,N,N-tetraacetic acid]. The lysates were centrifuged at 290 000g for 20 min at room temperature (RT). The supernatant was collected as the Sarkosyl-soluble fraction (sup). The remaining pellets were lysed in sodium dodecyl sulfate (SDS)-sample buffer (ppt). Samples were boiled for 5 min. Protein concentration was estimated using BCA Protein Assay Kit (Pierce). Each sample was separated by SDS–polyacrylamide gel electrophoresis (PAGE), and transferred onto polyvinylidene difluoride membrane (Millipore). The blots were blocked with 3% (w/v) gelatin and incubated overnight with the indicated primary antibody in 10% (v/v) calf serum at RT. The membranes were washed and incubated with a biotin-conjugated secondary antibody (Vector) or a horseradish peroxidase-labeled secondary antibody (Bio-Rad) for 2 h at RT. Signals were detected using the ABC staining kit (Vector) or ECL Prime Western Blotting Detection System (GE Healthcare).
Real-time polymerase chain reaction
Total RNA was isolated from cells with TRIzol (Invitrogen), and first-strand cDNA was synthesized with SuperScript II reverse transcriptase (Invitrogen). Polymerase chain reaction (PCR) reactions for Homo sapiens hyaluronan synthase 2 (HAS2, NM_005328.2, forward: 5′-CTCCGGGACCACACAGAC-3′, reverse: 5′-TCAGGATACATAGAAACCTCTCACAA-3′) and hypoxanthine phosphoribosyltransferase 1 (HPRT1; an internal standard, NM_000194.2, forward: 5′-TGACCTTGATTTATTTTGCATACC-3′, reverse: 5′-CGAGCAAGACGTTCAGTCCT-3′) were performed with Thunderbird SYBR qPCR MIX (Toyobo) and a LightCycler (Roche). The PCR reactions were carried out as follows: 1 min at 95°C for the initial denaturation, followed by 40 cycles of amplification at 95°C for 15 s and 60°C for 60 s.
SH-SY5Y cells were grown on collagen-coated coverslips and transfected as described above. After incubation for the indicated times, cells were fixed with 4% paraformaldehyde in phosphate-buffered saline (PBS) for 30 min. The coverslips were then incubated in 50 mm NH4Cl in PBS for 10 min and cell permeabilization was performed with 0.2% (v/v) Triton X-100 in PBS for 10 min. After blocking for 30 min in 5% (w/v) BSA in PBS, cells were incubated with the primary antibody for 2 h at 37°C. The cells were washed and further incubated with Alexa 488- or 568-conjugated secondary antibody, and then with TO-PRO-3 (Life Technologies) or Hoechst 33342 (Lonza) to counterstain nuclear DNA. Finally, they were analyzed using a LSM780 confocal laser microscope (Carl Zeiss).
SH-SY5Y cells were grown in collagen-coated six-well plates, and transfected as described above. After incubation for the indicated times, cells were harvested and lysed in radio-immunoprecipitation assay (RIPA) buffer [(25 mm Tris–HCl (pH 7.6), 150 mm NaCl, 1% NP-40 (Millipore), 1% sodium deoxycholate, 0.1% SDS and 0.1% protease inhibitor cocktail (Millipore)]. The lysates were centrifuged at 290 000g for 20 min at 4°C and the supernatant was collected and subjected to immunoprecipitation with anti-FUS or anti-TDP-43 (5 μl) and protein G-Sepharose (20 μl, Sigma). Bound proteins were washed with RIPA buffer and then eluted from the beads with SDS-sample buffer. Each sample was separated by SDS–PAGE and immunoblotted with anti-PFN1 antibody at RT, followed by incubation with HRP-conjugated secondary antibody. Signals were detected using the ECL prime Western Blotting Detection System (GE Healthcare).
Cell viability assay
Cell viability was examined by the trypan blue dye exclusion method using an automated cell counter TC10 (Bio-Rad).
Introduction of Sarkosyl-insoluble fraction as a seed to cells
Sarkosyl-insoluble fractions prepared as described above were suspended in 100 μl sterilized PBS by sonication. Then, 10 μl of PBS or insoluble fraction was mixed with 62.5 μl of Multifectam (Promega) for 30 min at RT. After that, 62.5 μl of Opti-MEM (Life Technologies) was added and incubation was continued for 5 min at RT. Then, the mixtures were added to cells expressing TDP-43 for 1 day and incubation was continued for 4 h in a CO2 incubator. After incubation, the medium was changed to fresh DMEM/F12 and culture was continued for 44 h. These samples were used for western blotting or immunocytochemistry.
The data were analyzed with an unpaired t-test or Tukey–Kramer test. Data are expressed as the mean ± SEM. Differences were considered statistically significant when P < 0.05.
Conflict of Interest statement. None declared.
This work was supported by Grant-in-Aid for Research on rare and intractable diseases, the Research Committee on Establishment of Novel Treatments for Amyotrophic Lateral Sclerosis, from Japan Agency for Medical Research and Development, AMED. This work was also supported by Japan Society for the Promotion of Science KAKENHI (grant numbers 15K18370 and 14J03307 to Y.T., 23228004 and 15H02356 to M.H.); Ministry of Education, Culture, Sports, Science and Technology in Japan KAKENHI (grant numbers 26117005 and 23240050 to M.H., 26111730 to T.N.) and Ministry of Health, Labour and Welfare (grant number 12946221 to M.H.).