Introduction
Amyotrophic lateral sclerosis (ALS) is a progressive and fatal neurodegenerative disease, characterized by the loss of upper and lower motor neurons in the brain and spinal cord. Reference Brown and Al-Chalabi1 More than 50 fused in sarcoma (FUS) mutations have been reported, accounting for about 5.8% familial ALS (fALS) and 1.3% sporadic ALS (sALS) cases in China. Reference Zou, Zhou, Che, Liu, He and Huang2 The majority of FUS mutations occur in exons 5–6 and exons 13–15 of different protein domains, suggesting variable pathogenic mechanisms. Reference Chen, Ding, Akram, Xue and Luo3 For example, FUS-Arg244Cys, FUS-Arg514Ser, and FUS-Arg521Cys mutations cause DNA damage by impairing FUS interaction with histone deacetylase 1 (HDAC1) and reducing double-strand breaks repair efficiency, Reference Qiu, Lee and Shang4,Reference Shang and Huang5 FUS-Pro525Leu causes abnormal RNA metabolism by inhibiting splicing of minor introns, Reference Reber, Stettler and Filosa6 and C-terminal mutations usually cause cytoplasmic mislocalization of FUS. Reference Dormann, Rodde and Edbauer7
Pathogenic genetic mutations in the intron/exon flanking regions (e.g. classical splicing sites) may lead to complete exon skipping, intron retention, or the introduction of a new splice site within an exon or intron. Reference Baralle and Baralle8 Splice site mutations are DNA sequence changes that alter or abolish correct mRNA splicing during the process of precursor mRNA maturation. Reference Skjørringe, Tümer and Møller9 It has been estimated that about 15% of human disease-causing mutations may disrupt mRNA splicing. Reference Krawczak, Reiss and Cooper10 For example, a C/T change at position 6 in exon 7 of survival of motor neuron 2 (SMN2) gene causing frequent skipping of exon 7 that produces an inactive and unstable protein lacking the last 16 amino acids, caused spinal muscular atrophy. Reference Cooper, Wan and Dreyfuss11
So far, few splice site mutations in FUS have been reported in the ALSoD database with unclear pathogenicity. The FUS splicing site mutations (IVS13-2A > G, p. Gly466Valfs*14; c.1542-2A > C, p. Gly515Serfs*8; c.1542-1G > T, p. Gly515Serfs*8) at C-terminal were reported to cause exon skipping, truncated protein products, and cytosolic mislocalization. Reference DeJesus-Hernandez, Kocerha and Finch12–Reference Canosa, Lomartire and De Marco14 Previously, two independent studies reported the FUS c.1394-2delA variant in western ALS patients. Briefly, the heterozygous FUS c.1394-2delA variant was found in 1 out of 301 German ALS patients, Reference Müller, Brenner and Weydt15 and 1 out of 391 US ALS patients. Reference Cady, Allred and Bali16 However, the frequency of FUS c.1394-2delA variant in Chinese ALS patients is unclear, and its functional mechanisms are largely unknown.
In this study, we aimed to investigate FUS splice site mutations in Chinese ALS patients and reported one Chinese sALS patient carrying the FUS splice acceptor site mutation (c.1394-2delA, p. Gly466Valfs*14). We also characterized the molecular mechanisms of this splice site FUS mutation at mRNA and protein levels.
Materials and Methods
Participants
The included patients were diagnosed with ALS according to the revised EI Escorial criteria Reference Ludolph, Drory and Hardiman17 in the ALS clinic at Huashan Hospital (Shanghai, China). The parents of the FUS splice mutation carrier were also included. Age of onset was defined as the age of initial appearance of symptoms. The functional impairment of the proband was assessed according to the Revised Amyotrophic Lateral Sclerosis Functional Rating Scale (ALSFRS-R). Reference Cedarbaum, Stambler and Malta18 This study was approved by the ethical review boards of Huashan Hospital (Shanghai, China).
Genetic Analysis
We collected the peripheral blood using the EDTA-containing vacuum blood collection tube (BD Vacutainer, 367863, USA). Genomic DNA was extracted using the QIAGEN DNA extract kit (QIAGEN, 51206, Germany) according to the standard protocols. We amplified all coding regions of Cu/Zn superoxide dismutase (SOD1), exon 6 of TAR DNA binding protein (TARDBP) and selected FUS coding regions (exon 5, exon 6, exons 13–14, and exon 15) for polymerase chain reaction (PCR), followed by sequencing with the ABI-3730XL genetic analyzer (Applied Biosystems, USA). We used CodonCode Aligner software (V.10.0.2) to analyze the amplicon sequences.
The minor allele frequency was obtained from the gnomAD database (Version 2.1.1) (http://gnomad-sg.org) (n = 15,708 Genomes data) and the China Metabolic Analytics Project (ChinaMAP) database (http://www.mbiobank.com) (n = 10,588 Genomes data). We used the SpliceAI deep learning algorithm (https://github.com/Illumina/SpliceAI) Reference Jaganathan, Kyriazopoulou Panagiotopoulou and McRae19 to predict the effect of mutation on splicing. The software sets a cutoff of 0.8 to indicate a high-precision prediction result.
Prediction of Protein Structure
AlphaFold deep learning algorithm was used to predict protein structures of FUS carrying wild type or mutant variant (https://github.com/deepmind/alphafold#running-alphafold). Reference Jumper, Evans and Pritzel20 We used the ranked_0.pdb file for further analysis, which contained the prediction result with the highest confidence. The results were visualized by PyMol (Version 2.5.2) software.
RNA Sequencing and RT-PCR
Lymphocytes were isolated from 3 ml EDTA-containing venous blood using LymphoprepTM (Stem Cell, Germany) and density-gradient centrifugation. Total RNA extraction from lymphocytes was performed using the total RNA extraction kit (Aidlab, China). 1 μg RNA was used for RNA sequencing at BerryGenomics Co., Ltd. In brief, the Illumina Ribo-Zero Plus rRNA Depletion Kit (Illumina, 20037135, USA) was used to remove the ribosomal RNA and TruSeq Stranded mRNA kit (Illumina, RS-122-2001, USA). Strand-specific sequencing libraries were prepared and sequenced by the Illumina Novaseq 6000 platform (Illumina, San Diego, USA). The alternative splicing was analyzed by LeafCutter using the RNA sequencing data with the default setting, which is available on https://github.com/davidaknowles/leafcutter/. Reference Li, Knowles and Humphrey21 In short, the RNA sequencing reads were aligned by STAR (Version 2.7.10a). Reference Dobin, Davis and Schlesinger22 Reads that span exon–exon junctions and map with a minimum of 6 nt into each exon were extracted from the bam files with filter_cs.py script. Intron clustering was performed using the leafcutter_cluster.py script. We used the leafcutter_ds.R script to calculate the differential excision of FUS (wild type vs mutant).
To validate abnormal RNA splicing event, we used the PrimeScriptTM RT reagent Kit (Takara Bio, Japan) to convert RNA to cDNA. FUS-specific primers targeting the exon 14 were used for RT-PCR (Supplementary Figure 1). The PCR products were separated by 2% agarose gel electrophoresis and visualized using iBrightCL1000 Imager (Thermo Fisher Scientific, USA). We then cut the target gel band, extracted the cDNA by the gel extraction kit (TianGen, China) and sequenced it with the ABI-3730XL genetic analyzer (Applied Biosystems, USA).
Preparation of Plasmids and Cell Culture
The human full-length coding sequence of wild-type FUS (NM_004960.3) was amplified from pEGFP-N1-FUS/TLS-FLAGC (addgene:60362) and was cloned into pEGFP-C1 vector at the XhoI/BamHI restriction site. The cDNA of the patient was used as a template to generate the truncated mutant FUS sequences and cloned into the XhoI/BamHI site of the pEGFP-C1 vector. HEK293T cells (purchased from Cell Bank of Chinese Academic of Science) and SH-SY5Y cells (a gift from Dr. Jian Wang in Fudan University) were cultured in DMEM or DF12 supplemented with 10% fetal bovine serum (FBS, Gibco) and Glutamax (Gibco), MEM non-essential amino acids solution (Gibco), and 1% penicillin/streptomycin (Gibco) respectively.
Immunofluorescence
HEK293T and SH-SY5Y cells were plated on 20 mm Glass Bottom Cell Culture Dishes (NEST, 801001, China) and transfected with wild type or mutant FUS plasmids using Lipofectamine 2000 (Thermo Fisher Scientific, USA). Reference Dalby, Cates and Harris23 After 36 h of plasmids transfection, we washed cells with phosphate-buffered saline and fixed cells with 4% cold paraformaldehyde for 15 min at room temperature. We used the Hoechst 33342 (Thermo Fisher Scientific, H3570, USA) to label the nuclei and Nikon’s Eclipse Ti2 inverted microscope confocal system (Nikon, Japan) to acquire images at 60× magnification with a Z stack interval of 2 μm. ImageJ software was used to quantify the intensity of the GFP fluorescence (NIH, USA, https://imagej.nih.gov/ij/). Briefly, we selected the cells with the nucleus stained by Hoechst and GFP for quantifying the intensities. The cytoplasmic FUS intensities were calculated by the total GFP intensities minus the GFP intensities of the nucleus.
Real-Time Quantitative Polymerase Chain Reaction (RT-qPCR)
Total RNA was reverse transcribed to cDNA using the PrimeScriptTM RT reagent Kit (Takara Bio, Japan). RT-qPCR was performed in a 20-μl reaction volume using TB Green® Premix Ex TaqTM II (Tli RNaseH Plus) reagent (Takara Bio, Japan) and specific primers (Supplementary Table 1) in Bio-Rad CFX96 Real-Time PCR Detection System (Bio-rad, USA). The expression level of total FUS mRNA was estimated based on the 2−ΔΔCT method and was normalized to the expression level of GAPDH housekeeping gene to obtain a relative expression level.
Statistics
We used the Wilcoxon rank sum nonparametric test (R Version 4.0.4) to analyze the mean cytoplasmic/total FUS ratio between wild-type and mutant groups. P value < 0.05 was accepted as statistically significant.
Results
Identification of the FUS Splicing Mutation in Chinese ALS Patients
In 233 Chinese ALS patients (Table 1), we found that one sporadic case carried a heterozygous FUS splicing mutation (NM_004960.4, chr16:31202282, c.1394-2delA, p. Gly466Valfs*14, rs1555509569) (Figure 1(A) and (B)). Sanger sequencing excluded other mutations of SOD1 or TARDBP in the proband. The healthy parents were further analyzed and did not carry this variant, suggesting that the FUS c.1394-2delA is a de novo variant (Figure 1(B)). It was absent in the gnomAD and ChinMAP databases. The proband has a younger sister [II-2] (39 years old), who had no complaint of any neurological problems and refused any genetic tests.
Table 1: Characteristics of 233 Chinese ALS patients
Abbreviations: ALS: amyotrophic lateral sclerosis; fALS: familial ALS; sALS: sporadic ALS; N, number; M: male; SD: standard deviation.
Figure 1: A novel FUS mutation affected RNA splicing in a sporadic ALS patient. (A) The pedigree of the investigated family. Proband was indicated by the arrowhead, and the age of family members and age at onset (AAO) were indicated. FUS genotype status was indicated as WT (wild type), MT (mutant); (B) Chromatography of Sanger sequencing. It showed that the proband carried FUS c.1394-2delA variant; (C) Visualization of altered RNA splicing of FUS exon13-exon15. The RNA sequencing data of the proband and one healthy control sample was processed by the LeafCutter tool (https://github.com/davidaknowles/leafcutter/). Location of the mutation site was indicated; (D) Agarose gel electrophoresis of FUS cDNA product. The patient II-1 (lane 1) and the health parents (I-1 and I-2, lane 2–3) had different FUS mRNA transcripts; and the patient had two dominant cDNA bands, corresponding to the normal transcript (393 bp) and the abnormal transcript (∼245 bp); (E) The Sanger sequencing chromatogram of the smaller cDNA product (∼245 bp on the gel). It validated the skipping of FUS exon 14 in the proband.
Clinical Presentation
The pedigree of this family is presented in Figure 1(A). The proband (II-1) developed muscle weakness in the right upper limb at the age of 39. The symptoms spread quickly to the left upper limb. Seven months later, he developed dysarthria, dysphagia, and dyspnea, with fasciculation in four limbs and obvious atrophy in both upper limbs. Neurologic examination at 11 months after initial symptoms showed reduced muscle strength of both distal and proximal upper extremities (4/5). Deep tendon reflexes were reduced in upper limbs and brisk in lower limbs. Hoffmann and Babinski’s signs were not present. His ALSFRS-R score was 40/48 with normal cognition and significant weight loss. Electromyography showed acute and chronic denervation in the upper and lower limb, thoracic, and sternocleidomastoid muscles. Blood biochemical tests (routine blood test, thyroid function, retinal and liver function, and immune-related antibodies [anti-ANA, anti-ENA, anti-CCP, anti-ACA, and anti-ANCAs]) and related examinations (cerebral magnetic resonance imaging and cerebrospinal fluid basic analyses) were normal. His blood gas analysis showed the high pressure of carbon dioxide (PaCO2: 46.3 mmHg) which indicated decreased respiratory function. The ALSFRS-R score of the patient was 15/48 at 17 months after symptom onset. The patient underwent tracheostomy 19 months after symptom onset due to respiratory failure.
The FUS Splicing Mutation is Linked to Altered RNA Splicing and Upregulated Gene Expression
Functional mechanisms of the FUS splice site mutation are largely unclear. SpliceAI analysis suggested that the FUS mutation (c.1394-2delA) may alter the normal splicing at chr16:31202284 (hg19) (SpliceAI = T|FUS|0.15|1.00|0.00|0.00|37|3|5|29), suggesting that the variant may cause a loss of splice acceptor. Sequencing of lymphocyte RNA revealed aberrant splicing leading to the skipping of FUS exon 14 (Figure 1(C)). The agarose gel electrophoresis of FUS cDNA products from the proband (II-1) revealed two PCR bands corresponding to a 393 bp (containing exon 14) and a 245 bp band (Figure 1(D)). Sanger sequencing chromatogram of the mutant products showed the skipping of FUS exon 14 (148bp) (Figure 1(E)). In addition, the mutation carrier showed elevated total FUS mRNA expression level (Supplementary Figure 2) when compared to the FUS expression of lymphocytes of the parents and another independent neurologically healthy individuals carrying the wild-type allele. RNA-seq further confirmed the upregulation of FUS expression in the mutation carrier (Supplementary Figure 3).
The FUS Splicing Mutation Induces Prominent Cytosolic Mislocalization
The FUS c.1394-2delA mutation was predicted to cause the frameshift and premature termination in exon 15, leading to a truncated FUS (p. Gly466Valfs*14) (Supplementary Figure 4), which might cause the loss of the nuclear localization signal (NLS) domain (Figure 2(A)).
Figure 2: Mislocalization of FUS in cell lines. (A) Schematic diagram of the EGFP-FUS fusion protein of different FUS plasmids. The protein domains of FUS are indicated by different boxes. The G in the mutant plasmid represents the truncated last RGG domain; (B) In HEK293T cells, mutant FUS caused mislocalization of FUS into the cytoplasm. GFP: EGFP-FUS fusion protein, Hoechst: nuclear staining, Merge: colocalization of GFP signal with the nuclear Hoechst fluorescence; (C) The histogram of the mean ratio of cytoplasmic/total FUS intensities in HEK293T cells (mean ratio: 0.45 vs 0.03 in MT and WT groups); (D) In SH-SY5Y cells, mutant FUS caused mislocalization of FUS into the cytoplasm. GFP: EGFP-FUS fusion protein, Hoechst: nuclear staining, Merge: colocalization of GFP signal with the nuclear Hoechst fluorescence; (E) The histogram of the mean ratio of cytoplasmic/total FUS in SH-SY5Y cells (mean ratio: 0.50 vs 0.05 in MT and WT groups); Scale bar: 10 μm; p < 10−4.
Since the C-terminus of FUS is a non-classical Proline/Tyrosine-nuclear localization signal (PY-NLS) domain and essential for FUS nuclear import, we tested whether the novel FUS c.1394-2delA mutation may affect the FUS cytosolic/nucleus mislocalization. In both HEK293T kidney and SH-SY5Y neuronal cell lines, we found that cells with the mutant FUS (c.1394-2delA, p. Gly466Valfs*14) showed the mislocalization of nucleus FUS into the cytoplasm (p < 0.0001; mean ratio:0.45 vs 0.03 in HEK293T cells, 0.50 vs 0.05 in SH-SY5Y cells) (Figure 2).
Discussion
Our study investigated FUS splice site mutations in Chinese ALS patients and identified one Chinese sALS patient carrying the rare FUS splice site mutation (c.1394-2delA, p. Gly466Valfs*14), which was previously reported in two western ALS patients, Reference Müller, Brenner and Weydt15,Reference Cady, Allred and Bali16 with unclear functional mechanisms. In lymphocytes of the patient, we found that the splice site mutation caused the splice acceptor loss at intron 13, leading to exon 14 skipping and upregulation of FUS gene expression. The mutant allele was predicted to encode a truncated protein. In both kidney and neuronal cell lines, we demonstrated that the FUS splice site mutation significantly induced cytosolic mislocalization, which is a key feature of FUS NLS domain mutations. Reference Dormann, Rodde and Edbauer7,Reference An, Rabesahala de Meritens, Buchman and Shelkovnikova24 Our findings support the pathogenicity of the FUS c.1394-2delA mutation in ALS.
A previous study reported a different mutation (g.10747A > G; IVS13-2A > G, p. Gly466Valfs*14) at the same FUS genomic locus in an American female sALS patient. Reference DeJesus-Hernandez, Kocerha and Finch12 It also caused abnormal RNA splicing, truncated FUS product, and cytoplasmic mislocalization of FUS in a mouse neuroblastoma cell line (N2A cells). Reference DeJesus-Hernandez, Kocerha and Finch12 In addition, both FUS c.1394-2delA mutation and the previously reported mutation were absent in the large public datasets (the ChinaMAP and the gnomAD databases), supporting that it might be pathogenic.
FUS is a DNA/RNA-binding protein involved in RNA metabolism processes, including gene transcription regulation, RNA splicing, and transport. Reference Niu, Zhang and Gao25 Wild-type FUS predominantly resides in the nucleus and shuttles between the nucleus and cytoplasm. Reference Zinszner, Sok, Immanuel, Yin and Ron26 Our in vitro experiments support that the FUS splicing mutation may cause truncated FUS protein lacking the PY-NLS domain, disrupt the interaction between FUS and karyopherin β2 (Kapβ2), and cytoplasmic mislocalization (Figure 2(B)–(E)). The results are in line with the previous report that FUS C-terminal NLS domain is important for FUS nuclear transportation by binding with Kapβ2 in a Ran-sensitive manner, Reference Lee, Cansizoglu, Süel, Louis, Zhang and Chook27 which would form the large TDP-43 negative inclusions and incorporate into the stress granules. Reference Daigle, Lanson and Smith28 Splicing site mutations may cause complete skipping of the exon, retention of the intron, or the introduction of a new splice site within an exon or intron, which could affect the balance of isoforms produced by alternatively spliced exons and cause disease. Reference Krawczak, Reiss and Cooper10 ALS has been reported to be caused by other splicing mutations, such as annexin A11 (ANXA11) c.1086+1G>A by promoting cytoplasmic aggregation of the mutant protein, SOD1 c.240-7T>G by reducing its stability and enzymic activity, and TANK-binding kinase 1 (TBK1) c.1522-3T>G, and c.2066+4A>G by reducing gene expression. Reference Sainouchi, Hatano and Tada29–Reference Lu, Chen and Lin31 Our findings support that abnormal splicing is an important molecular mechanism of ALS.
Genetic variants at C-terminal may escape nonsense-mediated RNA decay (NMD), Reference Amrani, Sachs and Jacobson32 but it is unusual to increase the gene expression level. To maintain its gene expression level in vivo, FUS has a unique homeostasis mechanism, in which FUS protein bind to exon 7 and flanking introns of its pre-mRNAs promoting intron retention Reference Zhou, Liu, Liu, Oztürk and Hicks33,Reference Humphrey, Birsa and Milioto34 that is followed by the NMD-mediated RNA degradation. Reference Zhou, Liu, Liu, Oztürk and Hicks33,Reference Humphrey, Birsa and Milioto34 The FUS splicing mutation may disrupt this homeostasis mechanism, because FUS mislocalization reduced the nucleus FUS protein level and may disrupt the intron retention and NMD related degradation, which might be related to upregulated gene expression level observed in this study.
The current study was limited to a moderate sample set of Chinese ALS patients. Future studies should investigate FUS splice mutations in a larger sample set with different ethnical backgrounds to understand the mechanisms of splicing mutations in ALS etiology.
In conclusion, we reported a de novo FUS splice site mutation (c.1394-2delA, p. Gly466Valfs*14) in 1 out of 233 (0.4%) Chinese ALS patients. It caused abnormal RNA splicing, upregulated gene expression, truncated FUS translation, and cytosolic mislocalization. Our findings suggested the pathogenic mechanism of the FUS c.1394-2delA mutation in ALS.
Supplementary Material
To view supplementary material for this article, please visit https://doi.org/10.1017/cjn.2022.336.
Acknowledgements
We would like to thank the patients and their family members for participating in this study.
Funding
This work was supported by Shanghai Municipal Natural Science Foundation General Program (22ZR1466400) (MZ), the National Natural Science Foundation of China (82071430) (MZ), and the Fundamental Research Funds for the Central Universities (MZ).
Conflict of Interest
There is no conflict of interest to disclose.
Statement of Authorship
Conceptualization: MZ and YC; Methodology: WLY, YZ, and XLT; Clinical data collection: XC, YMS, YD, and HY; Writing–original draft preparation: WLY, MZ, and XC; Writing—review and editing: MZ and YC.
