1. Introduction
Myelin sheaths are derived from the differentiated plasma membranes of oligodendroglial cells (oligodendrocytes) in the central nervous system (CNS) or Schwann cells in the peripheral nervous system (PNS). These cells wrap multiple layers of myelin, which then become the myelin sheaths, around the neuronal axons. These myelin sheaths grow dynamically to be more than 100 times larger than the collective surface area of the premyelinated plasma membranes [
1,
2,
3,
4]. Myelin sheaths play an indispensable role in the propagation of saltatory conduction. They also protect axons from physical and physiological stresses [
1,
2,
3,
4].
It is thought that defective myelin formation causes serious diseases. Hypomyelinating leukodystrophies (HLDs) are a recently classified group of hereditary neuropathies, primarily linked to oligodendrocytes, that affects 1 out of every 200,000 to 500,000 people. Pelizaeus-Merzbacher disease (PMD), now renamed HLD1, is a prototypic HLD [
5,
6,
7,
8]. Recent advances in powerful nucleotide sequencing technologies have enabled us to identify several unexpected genes that are responsible for other HLDs as well as other HLD-related neuropathies [
5,
6,
7,
8,
9,
10]. For example, HLD13 is caused by the Val54-to-Leu (V54L) mutation of cytoplasmic- and nuclear-localized C11ORF73 (also called HIKESHI) [
10]. It is thought that C11ORF73 normally supports the heat-shock-induced nuclear transport of 70 kDa heat-shock proteins [
11].
Furthermore, the infantile leukoencephalopathy-associated Cys4-to-Ser (C4S) mutation of C11ORF73 is known to decrease the protein expression levels, although C11ORF73 mutant proteins are still expressed at low levels in the nervous tissues [
12]. To date, it has remained unclear whether or how these residual mutant proteins cause their pathological effects at the molecular and cellular levels.
Here, we report that the Cys4-to-Ser (C4S) mutant proteins of C11ORF73 are localized and aggregated in the lysosome in the mouse oligodendroglial FBD-102b cells that we have used as model oligodendrocytes. For comparison purposes, it is noteworthy that C11ORF73 Val54-to-Leu (V54L) mutant proteins are also associated with aggregation. Cells expressing the C4S mutant proteins exhibit decreased lysosomal mammalian target of Rapamycin (mTOR) signaling, including S6 and 4E-BP1 phosphorylation, which is essential for oligodendrocyte differentiation and myelination [
1,
2,
3,
4]. Additionally, the C4S mutant proteins, but not the wild type ones, interact with Filamin A, which anchors transmembrane proteins to the actin cytoskeleton and connects the actin filaments [
13]. Both C4S mutant proteins and Filamin A are present in the lysosome. As a result, they may inhibit morphological differentiation. These results suggest a possible basis for the molecular and cellular pathological mechanisms that underlie infantile leukoencephalopathy.
2. Material and Methods
2.1. Primary and Secondary Antibodies
The following antibodies were purchased: mouse monoclonal antibody against a KDEL-containing peptide of the endoplasmic reticulum (ER)-resident glucose-regulated protein (GRP78) (Cat. No. M181-3; immunoblotting (IB), 1/10,000; and immunofluorescence (IF), 1/200), mouse monoclonal anti-actin (for the ß type; Cat. No. M177-3; IB, 1/40,000), mouse monoclonal anti-DDDDK antigen (also called FLAG antigen, Cat. No. M185-3; IB, 1/20,000), and mouse monoclonal anti-GFP (Cat. No. M048-3; and IB, 1/1000) from MBL (Aichi, Japan); mouse monoclonal anti-Golgi matrix protein of 130 kDa (GM130) (Cat. No. 610822; IB, 1/500; and IF, 1/200) from BD Biosciences (Franklin Lakes, NJ, USA); rabbit polyclonal anti-late endosomal Rab7 (Cat. No. 9367S; IF, 1/100) from Cell Signaling Technology (Danvers, MA, USA); goat polyclonal anti-lysosome-specific Catalase (Cat. No. AF3398; IF, 1/200) from Bio-Techne (Minneapolis, MN, USA); mouse monoclonal anti-lysosomal-associated membrane protein 1 (LAMP1) (Cat. No. ab25630; IB, 1/10,000; and IF, 1/100), rabbit monoclonal anti-Cathepsin D (Cat. No. ab75852; IF, 1/100), rabbit monoclonal anti-Rab9 (Cat. No. ab179815; IF, 1/100), rabbit monoclonal anti-Filamin A (Cat. No. ab 76289; immunoprecipitation (IP), 0.25 μg/500 μg of proteins; IB, 1/40,000; and IF, 1/100), mouse monoclonal anti-filamentous actin (F-actin, Cat. No. ab205; IF, 1/100), rabbit monoclonal anti-(pSer240/244) S6 protein/RPS6, which recognizes a phosphorylated state by p70S6K (Cat. No. ab215214; IB, 1/20,000), rabbit polyclonal anti-S6 protein/RPS6 (Cat. No. ab70227; IB, 1/250), rabbit monoclonal anti-(pThr37) 4E-BP1/eIF4E-BP1, which recognizes a phosphorylated state by mTOR kinase (Cat. No. ab75767; IB, 1/2500), rabbit monoclonal anti-4E-BP1 (Cat. No. ab32024; IB, 1/5000), and rabbit polyclonal anti-lysosome-specific transmembrane protein SLC38A9 (Cat. No. ab130398; IP, 1 μg/500 μg of proteins; and IB, 1/500) from Abcam (Cambridgeshire, UK); rabbit polyclonal anti-proteolipid protein 1 (PLP1) (Cat. No. HPA004128; IB, 1/500) from Atlas Antibodies (Bromma, Sweden); and rabbit polyclonal anti-C11ORF73 (Cat. No. 20524-1-AP; IF, 1/100) from Proteintech (Rosemont, IL, USA).
The following secondary antibodies were purchased: anti-rabbit or mouse IgG F(ab’) conjugated with horseradish peroxidase (Cat. Nos. 458 or 330; IB, 1/5000) from MBL; and anti-rabbit or mouse IgG (H+L) conjugated with Alexa Fluor 488 (Cat. Nos. A-11008 or A-11001; IF, 1/500) and anti-rabbit or mouse IgG (H+L) conjugated with Alexa Fluor 594 (Cat. Nos. A-11012 or A-11005; IF, 1/500) from Thermo Fisher Scientific (Waltham, MA, USA).
2.2. Plasmid Constructions
The plasmid encoding the human full-length c11orf73 (GenBank Acc. No. NM_016401) gene was amplified from SuperScript III reverse transcriptase (Thermo Fisher Scientific)-mediated human brain cDNAs (human RNA originally from Nippon Gene Co. Ltd., Tokyo, Japan) using Gflex DNA polymerase (Takara Bio, Shiga, Japan), according to the manufacturer’s instructions. The amplified fragments were ligated with the GFP-expressing pEGFP-C1 vector (Takara Bio).
The Cys4-to-Ser (C4S; 11G-to-C in the nucleotide level) and Val54-to-Leu (V54L; 160G-to-C in the nucleotide level; OMIN ID 616881) mutations were produced from pEGFP-C1-C11ORF73 as the template using a site-directed mutagenesis kit (Takara Bio), in accordance with the manufacturer’s instructions. All DNA sequences were confirmed by sequencing (Fasmac, Kanagawa, Japan).
2.3. Cell culture, Differentiation, Transfection, and Isolation of Stable Clones
African green monkey kidney epithelial cell-like COS-7 and human glial T98G cells (Human Health Science Research Resource Bank, Osaka, Japan) were cultured on cell culture dishes (Greiner, Oberösterreich, Germany) in Dulbecco’s modified Eagle medium (DMEM) containing 10% heat-inactivated FBS and PenStrep (Thermo Fisher Scientific) in 5% CO2 at 37 °C.
Mouse brain oligodendroglial FBD-102b cells were cultured on cell culture dishes in DMEM/Nutrient Mixture F-12 containing 10% heat-inactivated FBS and PenStrep in 5% CO
2 at 37 °C. To induce differentiation, FBD-102b cells were cultured on cell culture dishes (Greiner) with advanced TC polymer modification in culture medium without FBS for several days in 5% CO
2 at 37 °C. Cells with multiple processes from the cell bodies and myelin web-like membrane structures were identified as differentiated [
14]. Cells with a differentiated phenotype were considered to be those of cellular areas with a field of 25 square micrometers or more. FBD-102b cells were kindly provided by Dr. Y. Tomo-oka (Tokyo University of Science, Chiba, Japan).
Cells were transfected with the respective plasmids using a ScreenFect A or ScreenFect A Plus transfection kit (Fujifilm, Tokyo, Japan) according to the manufacturer’s instructions. The medium was replaced four hours after transfection and was generally used for experiments 48 h after transfection. Transfection efficiencies of COS-7 cells were approximately 75% as described in the manufacturer’s instructions. Transfection efficiencies of FBD-102b cells are 25 ± 5.0%, 27 ± 2.7%, and 26 ± 2.2% for the wild type, C4S, and V54L C11ORF73, respectively. Thus, COS-7 cells but not FBD-102b cells are suitable for biochemical experiments such as analyzing protein properties that need high level expression of proteins. To confirm the viability of COS-7 cells and FBD-102b cells under each experimental condition, we verified that attached trypan-blue-incorporating cells made up less than 5% of all cells in each culture [
14].
To collect FBD-102b cells stably harboring the wild type or the mutant constructs of C11ORF73, cells were transfected with the respective plasmids in a 3.5-cm cell culture dish. Growth medium containing 0.1250 mg/mL G418 (Nacalai Tesque, Kyoto, Japan), was changed every two or three days. After approximately 14 days, G418-resistant colonies were collected and their phenotypes were compared with those of their parental or control cells.
2.4. Fluorescence Images
Cells on a coverslip were fixed with 4% paraformaldehyde or 100% cold methanol. They were blocked with Blocking One reagent (Nacalai Tesque) and incubated with a primary antibody and an Alexa Fluor-conjugated secondary antibody (Thermo Fisher Scientific). For F-actin staining, cells were incubated with Phalloidin-iFluor594 reagent (Cat. No. ab176757, Abcam; IF 1/1000). The coverslips on the slide glass were mounted with Vectashield reagent (Vector Laboratories, Burlingame, CA, USA). TIFF images were collected with a microscope system equipped with a laser-scanning Fluoview apparatus (FV1000D or FV1200, Olympus, Tokyo, Japan) using Fluoview software (Olympus). The resulting color images were measured with the line and region analyses using Image J software for line plots and merged percentages, respectively. Each image in the figures is the representative of three independent experimental results (n = 25 cells in total).
2.5. Polyacrylamide Gel Electrophoresis and Immunoblotting
Cells were lysed in lysis buffer A (50 mM HEPES-NaOH, pH 7.5, 150 mM NaCl, 20 mM MgCl
2, 1 mM phenylmethane sulfonylfluoride, 1 μg/mL leupeptin, 1 mM EDTA, 1 mM Na
3VO
4, 10 mM NaF, and 0.5% NP-40) [
15,
16]. For non-denatured or denatured conditions, the supernatants were incubated with non-denaturing or denaturing sample buffer (Nacalai Tesque), respectively. The samples were separated on pre-made non-denatured (native) or denatured polyacrylamide gels (Nacalai Tesque). The electrophoretically separated proteins were transferred to polyvinylidene difluoride membranes (Merck Millipore, Darmstadt, Germany), blocked with Blocking One reagent, and immunoblotted using first primary antibodies, then horseradish peroxidase (HRP)-conjugated secondary antibodies. The bound antibodies were detected by X-ray film exposure using ImmunoStar Zeta reagent (Fujifilm). The films were captured as TIFF image files using a Canon LiDE 400 scanner (Canon, Tokyo, Japan) and its driver software (Canon). The band pixels were measured with the segment analysis using UN-SCAN-IT software. The pixel values of immunoreactive bands were reported as percentages and were compared with the control values. Each image in the figures is representative of three independent experimental results.
2.6. Immunoprecipitation for the Intended Protein or the Lysosome
Supernatants of the cell lysates in buffer A were used for immunoprecipitation of the purposed proteins [
15,
16]. The supernatants were mixed with protein G resin (GE Healthcare, Fairfield Easton, CT, USA) that had been absorbed with an antibody. The immunoprecipitates in supernatants of the cell lysate were denatured, subjected to polyacrylamide gel electrophoresis, and blotted onto polyvinylidene difluoride membranes for immunoblotting.
For immunoprecipitation of the lysosome, we used buffer B (50 mM HEPES-NaOH, pH 7.5, 150 mM NaCl, 5 mM MgCl
2, 1 mM phenylmethane sulfonylfluoride, 1 μg/mL leupeptin, 1 mM EDTA, 1 mM Na
3VO
4, 10 mM NaF) and homogenized cell lysates with Potter-Elvehjem homogenizer. The homogenates were centrifuged at 150×
g for 10 min using a tabletop centrifuge. The supernatants were gently mixed with an anti-SLC38A9 antibody-absorbed protein G resin [
17]. The immunoprecipitates were denatured, subjected to polyacrylamide gel electrophoresis, and blotted onto membranes for immunoblotting.
2.7. Mass Spectrometry (MS)-Based Identification of Affinity-Precipitated C11ORF73-Binding Proteins
Human glial cell line T98G was infected with retroviruses harboring either the gene encoding C11ORF73 or its mutant (C4S or V54L) tagged with FLAG and hexa-histidine in tandem and puromycin-resistant gene. These experiments were virus-mediated transfections, since their efficiency almost reaches 100%. After puromycin selection, cells were harvested and lysed with RIPA Buffer (50 mM sodium phosphate, pH 7.4, 1 mM MgCl2, 150 mM NaCl, 1% Nonidet P-40, 0.8% deoxycholate, 100 mM Na3VO4, 10 mM NaF, and protease inhibitors; Fujifilm). Using cell lysates, protein complexes including C11ORF73 or one of its mutants were purified through sequential affinity chromatography including M2-conjugated (Sigma-Aldrich, St. Louis, MO, USA) and Ni-NTA agarose (Qiagen, Venlo, Netherlands). Purified fractions were concentrated according to the 20% TCA precipitation method and degraded with 1 U/mL Trypsin Gold (Promega, Madison, WI, USA) overnight. To prevent re-formation of the disulfide bond between degraded peptides, thiol groups were masked by treatment with iodoacetamide. Trypsin-degraded peptides were separated by L-column 2 ODS (CERI, Saitama, Japan) by gradient with acetonitrile: 0 to 60% in 0.1 M trifluoroacetic acid. The fractions were mixed with α-cyano-4-hydroxy cinnamic acid and spotted onto MTP Anchor Chip 384 (Bruker Daltonics, Billerica, MA, USA). Peptides were analyzed in an Autoflex speed analyzer (Bruker Daltonics) for MALDI-TOF/TOF analysis, and proteins were identified using a Mascot server (Matrix Science, Tokyo, Japan).
2.8. Statistical Analysis
Values are means ± SD from separate experiments. Intergroup comparisons were performed according to Student’s t test using Microsoft Excel (Redmond, WA, USA). A one-way analysis of variance (ANOVA) was followed by a Fisher’s protected least significant difference test as a post hoc comparison using AnalystSoft StatPlus software. Differences were considered significant when p < 0.05.
2.9. Ethics Statement
Gene recombination techniques were performed in vitro and in vivo in accordance with a protocol approved by both the Tokyo University of Pharmacy and Life Sciences Gene and Animal Care Committee and the Japanese National Research Institute for Child Health and Development Gene and Animal Care Committee (L18-04 (data of approval, 4/1/2018), L19-04 (data of approval, 4/1/2019), and L20-04 (data of approval, 4/1/2000)).
4. Discussion
C11ORF73 mediates the heat-shock-induced nuclear uptake of 70 kDa heat-shock proteins through interactions with nucleoporins. In response to heat-shock, C11ORF73 forms a complex with ATP-bound 70 kDa heat-shock proteins and supports the transport of heat-shock proteins into the nucleus [
11]. Recent studies have illustrated that two types of C11ORF73 mutations, C4S and V54L, are responsible for the oligodendrocyte myelin-related disorders [
10,
12]. Although the C4S mutation is known to cause infantile leukoencephalopathy, to date, it has remained unclear how this mutation induces its pathological effects on oligodendrocytes at the molecular and cellular levels. Here, we show that the C4S mutant proteins aggregate and are abundantly localized in the lysosome, where they downregulate lysosomal signaling through S6 and 4E-BP1 phosphorylation. C4S mutant proteins, but not wild type proteins, specifically interact with Filamin A. Filamin A, which is also detected in the lysosome fraction, plays a key role in mediating the formation of a cytoplasmic actin network among actin filaments or between actin filaments and other molecules, and also contributes to cell morphological maintenance and change [
22]. While cells harboring the wild type constructs normally exhibit differentiated phenotypes, cells harboring the C4S mutant ones fail to do so. These findings suggest that the C4S mutation can lead to defective morphological differentiation in oligodendrocytes, possibly by generating a de novo interactive ability with Filamin A and accumulating Filamin A in the lysosome.
Some nuclear proteins related to myelin-related disorders have been identified previously. The mutations of the genes encoding the RNA polymerase III subunits POLR3A and POLR3B are responsible for HLD7 and HLD8, for example [
23]. RNA polymerase III functions as transcripting noncoding RNA and sometimes as doing tRNA. HLD7 is an autosomal recessive disorder characterized by childhood onset of progressive motor decline. The features involve spasticity, ataxia, and tremor. Other features may include hypodontia or oligodontia and hypogonadotropic hypogonadism [
23]. HLD8 is an autosomal recessive disorder characterized by cerebellar ataxia and intellectual disabilities with oligodontia or hypogonadotropic hypogonadism [
23]. It is likely that loss of function of RNA polymerase III activities causes HLD7 and HLD8 [
23], but it remains unclear whether the observed disease features are actually due to downregulation of the transcription of noncoding RNAs. Alternatively, it is thought that mutations of RNA polymerase III subunit A (POLR3A) or RNA polymerase III subunit B (POLR3B) may induce aggregation of POLR3A and POLR3B proteins, leading to proteopathies (also called proteinopathies) associated with gain of function in POLR3A or POLR3B mutants. Indeed, the infantile leukoencephalopathy-associated C4S mutation of nuclear and cytoplasmic C11ORF73 also causes C11ORF73 protein aggregation, but it has remained unclear to date whether this protein aggregation was directly related to the leukoencephalopathy phenotypes observed in vivo.
The C4S mutant proteins were accumulated in the lysosome, since they might not be degraded; in contrast, the wild type proteins were not present in the lysosome. Alternatively, disease-associated mutation might change properties of organelle localization of proteins. In either case, evidence shows that signaling through mTOR is involved in the regulation of lysosome function and vice versa [
24]. mTOR is present in two structurally distinct forms: mTOR complex 1 (mTORC1) and mTOR complex 2 (mTORC2). mTORC1 and mTORC2 specifically control distinct groups of the downstream effectors. mTORC1 links energy metabolism and amino acid balance to cell growth and proliferation. It is also involved in the regulation of cell volumes by molecular events such as protein synthesis including S6 and 4E-BP1 phosphorylation and in the utilization of energy stores. For example, when amino acid levels are high, mTORC1 is recruited to the lysosomal surface, where it is mediated by the Ras family small GTPase Rheb [
24]. The translocation of mTOR to the lysosome participates in protein synthesis through S6 and 4E-BP1 phosphorylation. These effectors are associated with maintaining the functional lysosome, which can lead to the determination of cell fate through a possible regulation linking to transcription factor EB (TFEB) [
25]. In this study, cells harboring wild type C11ORF73 constructs have the ability to promote the lysosome-related S6 and 4E-BP1 phosphorylation, whereas cells harboring C4S mutant proteins lack this ability. Furthermore, mTOR signaling through S6 and 4E-BP1 phosphorylation is important for oligodendrocyte differentiation and myelination [
19,
20]. Thus, it is conceivable that a decrease in phosphorylation level may be associated with inhibitory differentiation in FBD-102b cells.
We found that the C4S mutation of C11ORF73 gives C11ORF73 proteins the capacity for de novo interaction with Filamin A. This interaction inhibits F-actin formation in cells, perhaps because mutated aggregated C11ORF73 proteins cause Filamin A to accumulate in the lysosome. Cycles of F-actin organization and disorganization are necessary for the cell morphological changes underlying oligodendrocyte differentiation and myelination [
22]. Filamin A is a 280 kDa actin-binding protein that binds to cell adhesion molecules and receptors and regulates actin cytoskeletal reorganization [
13]. Filamin A is one of the central molecules linking upstream signaling to actin cytoskeletal changes [
13]. In fact, mutations in Filamin A are responsible for a variety of genetic diseases related to cell morphological changes. These diseases are likely caused by loss-of-function of Filamin A. It is possible that, in infantile leukoencephalopathy tissues and cells, the C4S mutant proteins’ interaction with Filamin A inhibits F-actin formation and thereby inhibits cell morphological changes.