
==== Front
Sci Adv
Sci Adv
sciadv
advances
Science Advances
2375-2548
American Association for the Advancement of Science

adj0347
10.1126/sciadv.adj0347
Research Article
Neuroscience
SciAdv r-articles
Biochemistry
Neuroscience
Biochemistry
Sulfated disaccharide protects membrane and DNA damages from arginine-rich dipeptide repeats in ALS
Disaccharide rescues Arg-rich DPRs toxicity in ALS
https://orcid.org/0009-0008-8787-0661
Chang Yu-Jen Conceptualization Formal analysis Investigation Methodology Project administration Resources Supervision Validation Visualization Writing - original draft Writing - review & editing 1 2
Lin Kai-Tai Conceptualization Data curation Formal analysis Investigation Methodology Resources Validation Writing - review & editing 3 4
https://orcid.org/0000-0002-0610-6856
Shih Orion Data curation Formal analysis Investigation Methodology Software Validation 3
https://orcid.org/0000-0001-9598-8936
Yang Chi-Hua Formal analysis Investigation Methodology Resources Visualization 1
https://orcid.org/0000-0002-9146-8965
Chuang Ching-Yu Resources 1
Fang Ming-Han Methodology Resources 5
https://orcid.org/0009-0002-1139-7923
Lai Wei-Bin Resources 1
https://orcid.org/0000-0003-0102-164X
Lee Yi-Chung Resources Writing - review & editing 6 7
Kuo Hung-Chih Methodology Resources Validation Writing - review & editing 8
https://orcid.org/0000-0002-8797-729X
Hung Shang-Cheng Investigation Methodology Resources Writing - review & editing 1
https://orcid.org/0000-0003-0977-4347
Yao Chi-Kuang Conceptualization Investigation Methodology Resources Validation Visualization Writing - review & editing 5 9
https://orcid.org/0000-0002-2247-5061
Jeng U-Ser Investigation Methodology Validation Visualization Writing - original draft 3 4
https://orcid.org/0000-0002-6596-6338
Chen Yun-Ru Conceptualization Data curation Funding acquisition Methodology Project administration Resources Supervision Validation Visualization Writing - original draft Writing - review & editing 1 2 *
1 Genomics Research Center, Academia Sinica, Taipei 115, Taiwan.
2 Taiwan International Graduate Program in Interdisciplinary Neuroscience, National Taiwan University and Academia Sinica, Taipei 115, Taiwan.
3 National Synchrotron Radiation Research Center, Hsinchu 300, Taiwan.
4 Department of Chemical Engineering, National Tsing Hua University, Hsinchu 300, Taiwan.
5 Institute of Biochemical Sciences, College of Life Science, National Taiwan University, Taipei 106, Taiwan.
6 Department of Neurology, Taipei Veterans General Hospital, Taipei 112, Taiwan.
7 Department of Neurology, National Yang Ming Chiao Tung University School of Medicine, Taipei, Taiwan.
8 Institute of Cellular and Organismic Biology, Academia Sinica, Taipei 115, Taiwan.
9 Institute of Biological Chemistry, Academia Sinica, Taipei 115, Taiwan.
* Corresponding author. Email: yrchen@gate.sinica.edu.tw
23 2 2024
23 2 2024
10 8 eadj034712 6 2023
22 1 2024
Copyright © 2024 The Authors, some rights reserved; exclusive licensee American Association for the Advancement of Science. No claim to original U.S. Government Works. Distributed under a Creative Commons Attribution NonCommercial License 4.0 (CC BY-NC).
2024
The Authors
https://creativecommons.org/licenses/by-nc/4.0/ This is an open-access article distributed under the terms of the Creative Commons Attribution-NonCommercial license, which permits use, distribution, and reproduction in any medium, so long as the resultant use is not for commercial advantage and provided the original work is properly cited.

Hexanucleotide repeat expansion in C9ORF72 (C9) is the most prevalent mutation among amyotrophic lateral sclerosis (ALS) patients. The patients carry over ~30 to hundreds or thousands of repeats translated to dipeptide repeats (DPRs) where poly-glycine-arginine (GR) and poly-proline-arginine (PR) are most toxic. The structure-function relationship is still unknown. Here, we examined the minimal neurotoxic repeat number of poly-GR and found that extension of the repeat number led to a loose helical structure disrupting plasma and nuclear membrane. Poly-GR/PR bound to nucleotides and interfered with transcription. We screened and identified a sulfated disaccharide that bound to poly-GR/PR and rescued poly-GR/PR–induced toxicity in neuroblastoma and C9-ALS-iPSC–derived motor neurons. The compound rescued the shortened life span and defective locomotion in poly-GR/PR expressing Drosophila model and improved motor behavior in poly-GR–injected mouse model. Overall, our results reveal structural and toxicity mechanisms for poly-GR/PR and facilitate therapeutic development for C9-ALS.

Sulfated disaccharide rescues membrane and DNA damages induced by Arg-rich DPRs in ALS animal models.

http://dx.doi.org/10.13039/501100001869 Academia Sinica AS - TP - 109 - LM – 08 http://dx.doi.org/10.13039/501100004663 Ministry of Science and Technology, Taiwan MOST 111-2113-M-001-024 Academia Sinica Core Facility and Innovative Instrument Project AS-CFII-111-204 CopyeditorEunice Ann Alesin
==== Body
pmcINTRODUCTION

Hexanucleotide GGGGCC (G4C2) repeat expansion in the noncoding region of C9ORF72 gene is the most prevalent mutation in frontotemporal lobar dementia (FTLD) and amyotrophic lateral sclerosis (ALS) (1, 2). FTLD is a neurodegenerative disease characterized by frontotemporal dementia, behavioral deficits, and progressive aphasia. ALS is a motor neuron (MN) disease that affects upper and lower MNs of patients. People with ALS are paralyzed with an average survival time of 3 to 5 years. These two diseases share some pathologies and gene mutations.

The pathogenic mechanisms of C9ORF72 mutation include loss of function (3) and gain of function, where both RNA and protein toxicity have been reported (4, 5). Five dipeptide repeats (DPRs) are generated from the hexanucleotide repeat expansion through repeat-associated non-ATG initiated translation, including poly-glycine-alanine, poly-glycine-arginine (GR), poly-glycine-proline (GP), poly-proline-arginine (PR), and poly-proline-alanine. These DPRs from sense and antisense RNA are identified in hippocampus, frontal and motor cortices, and spinal cord of patients with FTLD or ALS (6). Among DPRs, arginine-rich DPRs, poly-GR and poly-PR, have been observed as the most toxic species (5, 7). The toxicity mechanism of poly-GR and poly-PR is still under intensive investigation including compromised RNA maturation (7), the translation inhibition (8, 9), and inhibition of nucleocytoplasmic transportation (10–13). The repeat number of G4C2 for healthy individuals is from 2 to 24 repeats, whereas it varies from over ~30 to hundreds or even thousands among patients (14). The exact pathological threshold of G4C2 repeat number is still ambiguous. The toxicity boundary of peptide length of DPRs is also unknown.

To investigate the disease mechanism of poly-GR/PR, we used synthetic poly-GR peptides in this study to examine the minimal length required for poly-GR toxicity. We provided the structural information of poly-GR/PR by small-angle x-ray scattering (SAXS) combined with molecular simulation. We examined the membrane integrity after the treatment of poly-GR/PR with liposome leakage and calcium influx assays and further examined their interaction with nucleotides and interference with RNA transcription. Last, we identified a sulfated disaccharide from chemical library screening to rescue poly-GR/PR–induced cytotoxicity in C9ORF72 induced pluripotent stem cells (C9-iPS)–derived MNs, transgenic poly-GR/PR fly models, and a poly-GR–injected mouse model.

RESULTS

Poly-GR toxicity is length and dose dependent

To elucidate the minimal length required for poly-GR toxicity, we treated mouse neuroblastoma N2a (Fig. 1, A to D) and rat primary cortical neurons (Fig. 1, E to H) with synthetic poly-GR peptides from 0 to 8 or 10 μM in different repeat numbers ranging from 5, 10, 15, 20, 25, to 30 (denoted as GR5, GR10, GR15, GR20, GR25, and GR30) and subjected them to cellular assays. In 3-(4,5-Dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay (Fig. 1A) using 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide to detect cell proliferation, we found that only GR30 and GR25 showed dose-dependent toxicity to N2a cells above 4 μM compared with the buffer control. Poly-GR peptides shorter than 20 repeats did not show significant toxicity. Similar results were observed in the lactate dehydrogenase (LDH), reactive oxygen species (ROS), and caspase-3 activity assays (Fig. 1, B to D). We further used primary rat cortical neurons (Fig. 1, E to H) and showed that GR30 and GR25 were significantly toxic above 2 μM in all assays except for the ROS assay, and GR20 were significantly toxic above 4 μM in all assays. GR15, but not GR10 and GR5, was also toxic at high concentration in the MTT, ROS, and caspase-3 assays. The result showed a repeat threshold for cytotoxicity. Besides poly-GR peptides, PR30 also contained significant toxicity while PR10 and GP30 were not toxic even at 8 μM (fig. S1, A to D). The length dependence for cytotoxicity was also found in poly-GR overexpression in MN-like NSC34 cells transfected with enhanced green fluorescent protein (EGFP), EGFP-GR15, and EGFP-GR30 by propidium iodide (PI) staining (fig. S1E).

Fig. 1. Cytotoxicity of poly-GR coordinates with the gradually shaped loose helical conformation that enhanced membrane penetration.

N2a cells (A to D) and primary rat cortical neurons in day in vitro (DIV) 7 (E to H) were treated with poly-GRs and subjected to MTT for detection of cell proliferation, LDH for detection of cell membrane integrity, ROS for detection of oxidative stress, and caspase-3 assays for detection of cell apoptosis. Data were represented as mean ± SD, and the statistical analysis for each group compared with the buffer control was performed by two-way ANOVA with Bonferroni post hoc test (n = 6; *P < 0.05, **P < 0.01, and ***P < 0.001). (I) The structural models of poly-GRs obtained from SAXS and molecular simulation. Backbones, blue; side chains, red. Scale bar, 10 Å. (J) Liposome leakage assay after treatment of DPRs. Quenched fluorescent dye strongly emits after leaking from liposome. Data were represented as mean ± SD (n = 3). (K) Calcium influx assay after treatment of DPRs. Data were represented as mean ± SD and analyzed by one-way analysis of variance (ANOVA) with Tukey post hoc test comparing the buffer control group (n = 6; **P < 0.01 and ***P < 0.001).

Poly-GR peptides form loose helical conformation in a length-dependent manner

First, we used circular dichroism spectroscopy to examine the secondary structure of poly-GR peptides and found longer poly-GR peptides, more than 20 repeats, and performed the spectra with a small positive peak at 215 nm, which is similar to the minor positive peak around 210 nm found in poly-glycine type II helix (fig. S2A) (15). We also found that this feature disappeared upon melting and recovered upon cooling indicating the folding reversibility of the conformation (fig. S2B). Since our circular dichroism spectra cannot sufficiently provide detailed information of the secondary structures, we further examined the solution structural features of the poly-GR peptides using SAXS combined with molecular structural simulations (Fig. 1I). The SAXS data with radius of the gyration, Rg values, are shown in fig. S2C. The structural models obtained from the analysis protocol demonstrated that GR5, GR10, and GR15 had gradually increased Rg value from 5.95 ± 0.04, 9.2 ± 0.08, to 11.3 ± 0.08 Å, respectively (fig. S2D). Furthermore, when the repeat number increased near or above 20, the poly-GR peptides, i.e., GR20, GR25, and GR30, gradually formed and developed a loose helical structure, thus resulting in correspondingly increased Rg values to 13.7 ± 0.08, 16.5 ± 0.22, and 18.5 ± 0.6 Å, respectively, which were deviating from the theoretical Rg values according to their length (fig. S2E). However, the helical structure in this model is not rigid comparing to typical α helix, which could be a molten globule conformation. Our result indicated that the increasingly prominent helical structure formed with a repeat number above 20 was correlated with the increasingly higher toxicity observed in the toxicity assays. We also examined the solution structures of poly-PR and nontoxic poly-GP and found a relatively extended coil rod–like conformation in PR30 with a fitted Rg, 25.4 ± 0.3 Å (fig. S2F), in PR10 with an Rg of 12.2 ± 0.3 Å (fig. S2G), and in GP30 with an Rg of 22.0 Å (fig. S2H). Our results demonstrated that poly-GR and poly-PR were structurally distinct in solution where the loose helical structure formed in poly-GR had more than 20 repeats. The structural differences might shed light on tracing the origin of their respective toxic behaviors.

Poly-GR peptides disrupt liposomes and plasma/nuclear membranes in a length-dependent manner

Since helical structures of peptides could facilitate membrane penetration (16), we performed liposome leakage assay by artificial lipid membrane and calcium influx in cells to examine the membrane penetration ability of poly-GRs. We found that the fluorescence of GR30-treated liposome increased to ~30% of the totally lysed liposome (Fig. 1J). GR20- and GR10-treated liposomes increased to ~20 and ~5%, respectively. PR30 showed ~8% leakage, and GP30 showed no leakage. When we examined plasma membrane penetration in NSC34 cells (Fig. 1K), we found that GR30 and GR20 induced significant calcium influx higher than the control with a ~104 and ~17% increase, respectively. PR30-treated cells also showed significantly increased calcium intensity that was ~43% higher than the control. In contrast, GR10 and GP30 treatment showed no significant calcium influx.

To elucidate the potential penetration of nuclear membrane by DPRs, we isolated the nuclei of NSC34 cells and directly treated them with DPRs. The isolated nuclei (fig. S3A) were incubated with 20 μM DPRs followed by nuclear membrane protein Lamin staining (Fig. 2A). The results showed that the buffer control treatment did not affect nuclear membrane. GR30-treated cells presented an extreme effect with no Lamin signal and with smeared 4′,6-diamidino-2-phenylindole (DAPI) signal indicating complete disruption of the nuclear membrane. The nuclei treated with GR10 and GP30 were still morphologically intact, where GR20 and PR30 treatment reduced the number of intact nuclei. After Lamin distribution in cross section of the nuclei was analyzed and quantified (Fig. 2B), about ~70% Lamin signal was located on the nuclear membrane in the buffer control, whereas the signal was completely undetectable after GR30 treatment and significantly reduced to ~40% in GR20 and PR30. We also expressed EGFP-conjugated poly-GR in N2a cells and found that nuclear membrane was significantly damaged by the expression of 150 repeats of poly-GR (fig. S3D). The results again indicated that poly-GR was capable of damaging nuclear membrane. We hypothesized that both synthetic and overexpressed poly-GR localized to nucleus. We treated Alexa Fluor 647-labeled GR30 (AF647-GR30) with N2a neuroblastoma and found that most AF647-GR30 was detected in the nuclei with a minor amount of cytosolic AF647-GR30 (fig. S4A). We further examined the nuclear location for Atto488-GR25 by immunostaining with a nucleoli marker nucleolin antibody (fig. S4B) and found that Atto488-GR25 colocalized with nucleolin. We also transfected EGFP-GR30 and EGFP-PR30 to N2a and found them mostly condensed in the nucleoli, while EGFP alone distributed diffusely (fig. S4, C to E). We further used immuno-transmission electron microscopy (immuno-TEM) to examine the location of EGFP-GR30 expressed in N2a. The fixed cells were stained with GFP antibody and gold nanoparticles (AuNP)–conjugated secondary antibody (fig. S4, F to H). The results showed that most AuNPs were in the electron-dense area in the nuclei, likely heterochromatin and nucleoli, while the distribution of AuNPs in EGFP-expressed cells was diffused.

Fig. 2. Poly-GR and poly-PR damage nuclear membrane.

(A) Fluorescent images of the isolated nuclei after DPR treatment. DAPI, blue; Lamin, green. Scale bars, 10 μm. RFU, relative fluorescence units. The quantified data by cross-sectional intensity of Lamin versus the distance from the nucleus center were shown. Data were represented as mean ± SD (n = 15 to 25). (B) Quantification of nuclear membrane after DPRs treatment. Data were represented as mean ± SD and analyzed by one-way ANOVA with Tukey post hoc test comparing the buffer control (n = 15 to 25; ***P < 0.001).

Poly-GR has stronger binding affinity to DNA than RNA

As poly-GR had repeated positively charged residues, we suspected that poly-GR binds to nucleic acids. We performed isothermal calorimetry (ITC) to differentiate the affinity of GR25 to different nucleotides (Fig. 3A and fig. S5, A to D). We first titrated GR25 to 12-base single-strand DNA (ssDNA) composed of four single nucleotides, i.e., poly-A, poly-T, poly-G, and poly-C (fig. S5, A to D). We found that GR25 bound to poly-T with dissociation constant (Kd) at 0.3 μM and poly-C at 0.15 μM, better than poly-G at 1.2 μM. It did not favor poly-A binding with the lowest binding affinity (Kd = 6.7 μM). The data suggested that poly-GR has binding preference to pyrimidine than purine. Next, we examined ssDNA, single-strand RNA (ssRNA), and double-strand DNA (dsDNA) (Fig. 3A) with mixed pyrimidine and purine focused on AC-rich and TG-rich to avoid self-complementary (ssDNA: ACACACACACAC, denoted as AC6; ssDNA: GTGTGTGTGTGT, denoted as GT6; ssRNA ACACACACACAC denoted as rAC6). We also designed a 12-bp dsDNA by annealing AC6 and GT6 denoted as AC/GT6. We titrated GR25 to AC6, GT6, rAC6, or AC/GT6. The ITC data showed that the affinity of GR25 to ssDNA AC6, Kd at 61 nM, was similar to GT6, at 64 nM. However, GR25 had a twofold higher affinity to ssDNA AC6 than to ssRNA rAC6, which was 144 nM. The affinity to dsDNA AC6/GT6 was approximately 119 nM (Fig. 3B). Furthermore, we examined GR10 at 150 μM to AC6 in ITC and found no binding (fig. S5E). When we examined GR10 at 450 μM containing equal amount of arginine to GR30 at 150 μM, there was also no binding (fig. S5F), while GR30 at 150 μM strongly bound to AC6, Kd at 30 nM (fig. S5G). PR30 at 150 μM also bound to AC6 but with a ~19-fold weaker affinity, Kd at 571 nM (fig. S5H). These results demonstrated that nucleic acid interaction with longer poly-GR is not solely due to electrostatic interactions. The poly-GR conformation and types of nucleic acids also play important roles. To further validate whether poly-GR binds to DNA in cell, we expressed EGFP-GR15, EGFP-GR30, and EGFP-GR80 in N2a cells, performed crosslinking by 1% paraformaldehyde, isolated the crossed-linked DNA-protein, and detected the DNA-bound EGFP and poly-GR in cell (Fig. 3C). The result showed strong GFP and GR signals in the cells expressing EGFP-GR30 and EGFP-GR80 but weak signals in EGFP-GR15 transfected cells, which may be due to less DNA binding. No signal was detected in the EGFP control. The result demonstrated that poly-GR binds to DNA in cell.

Fig. 3. Poly-GR binds to nucleic acids and poly-GR/PR inhibit transcription in vitro and in cell.

(A) Isothermal titration of GR25 to nucleic acids. (B) Binding affinity of GR25 to nucleic acids. (C) Crosslinking assay for DNA-protein complex. (D) In vitro transcription assay in the presence of DPRs. (E) Newly synthesized RNA in NSC34 cells after the treatment of poly-GRs. Nascent RNA, red; DAPI, blue. Scale bars, 10 μm. (F) Quantification of the newly synthesized RNA in cell treated with DPRs. Data were represented as mean ± SD and analyzed by one-way ANOVA with Tukey post hoc test comparing the buffer control (n = 8000, triplicate; **P < 0.01 and ***P < 0.001). (G) Quantification of the newly synthesized RNA in cells by flow cytometry treated with poly-GRs. (H) R-loop staining of primary cortical neurons. Representative images of primary neurons treated with PBS, GR30, GR10, and PR30 were shown. R-loop, green; nuclei, blue. Scale bars, 10 μm. (I) R-loop staining of primary cortical neurons with mCherry-GR150 overexpressed. Nuclei, blue; R-loop detected by S9.6 antibody, green; mCherry-GR150, red. Scale bars, 10 μm.

Poly-GR/PR peptides inhibit replication and transcription

As poly-GR interacted with DNA and located in nucleus, we wondered whether poly-GR impairs DNA functions. We first used polymerase chain reaction (PCR) to test DNA replication affected by poly-GR peptides (fig. S6). Upon the presence of poly-GR, the PCR products were reduced in a GR length–dependent manner. We also used in vitro DNA replication assay (fig. S7) to check newly synthesized dsDNA monitored by EvaGreen fluorescence with the addition of poly-GR peptides. The result again showed a length- and dose-dependent manner for poly-GR inhibiting DNA replication (fig. S7, A to F). The inhibition is also observed under PR30 treatment (fig. S7G). Next, we performed in vitro RNA transcription assay to examine the potential of poly-GR inhibiting RNA transcription (Fig. 3D). The in vitro RNA transcription in the presence of poly-GRs, PR30, and GP30 was examined. The results showed that GR30, GR25, GR20, and PR30 strongly inhibited RNA synthesis above 2.5 μM, and GR15 and GR10 inhibited RNA synthesis above 5 and 7.5 μM, respectively. GR5 and GP30 did not inhibit RNA synthesis. We also applied NSC34 cells with 5-ethynyl-UTP to incorporate into nascent RNA in cell after 4-hour treatment of poly-GR peptides. Newly synthesized RNA mainly localized in the nucleus and colocalized with DAPI staining (Fig. 3E). After quantification (Fig. 3F), the results showed that GR30 and PR30 significantly reduced the nascent RNA level, whereas GR20, GR10, and GP30 had similar fluorescence intensity to the buffer control. We also conducted flow cytometry to quantify the cells with nascent RNA and confirmed that the population of GR30-treated cells had fewer fluorescent cells than phosphate-buffered saline (PBS)–, GR20-, and GR10-treated cells (Fig. 3G). RNA polymerase inhibitor, actinomycin D, was used as a positive control that showed the complete inhibition of the transcription. The level of active RNA polymerase II in NSC34 cells was checked by Western blot and showed no difference in both active (by phospho-ser2/5 Rpb1 antibody) and total (by Rpb1 antibody, clone D8L4Y) RNA polymerase II (fig. S8, A and B). Therefore, the poly-GR/PR inhibition of RNA synthesis was not attributed from the down-regulation of active RNA polymerase II. As the consequence of poly-GR binding to ssDNA and inhibition of transcription, we examined R-loop formation in primary neurons treated with poly-GR peptides. Representative images of enhanced ssDNA-RNA complex by both GR30 and PR30 treatment were shown (Fig. 3H and fig. S8C). We also found that poly-GR colocalized with R-loop in primary cortical neurons expressed mCherry-GR150 (Fig. 3I).

To further investigate DNA-related toxicity induced by poly-GR/PR, we examined the DNA damage in N2a cells after 4-hour treatment of poly-GR/PR using anti-γH2AX (Fig. 4A) which corresponds DNA double-strand breaks (17) or single-strand DNA breaks during replication stress (18). Quantification of anti-γH2AX–positive particles showed that GR30- and PR30-treated cells had more DNA damage than the buffer-treated cells, while GR10 did not show significant difference (Fig. 4B). The results revealed that longer poly-GR/PR peptides also significantly induced DNA damage besides inhibiting transcription. Meanwhile, we asked whether poly-GR peptides compromised protein translation (fig. S9). After 4-hour of DPR treatment to NSC34 cells, newly synthesized peptides were incorporated with O-propargyl-puromycin, fluorescently labeled by click reaction, and located in both nuclei and cytosol. The cells with the florescent signal were quantified. The results showed no significant effect of poly-GR treatment to protein translation after 4-hour of DPR treatment, while translation was reported to be compromised in later stage (19, 20).

Fig. 4. Treatment of poly-GR/PR peptides induces DNA damage.

(A) Representative fluorescence images of anti-γH2AX staining. N2a cells were treated with PBS, GR30, PR30, and GR10 for 4 hours. Nuclei, blue; anti-γH2AX, green. Scale bars, 10 μm. (B) Quantification of anti-γH2AX–positive signals in N2a nucleus. The number, total area, and accumulated intensity of anti-γH2AX–positive particles in N2a nucleus was quantified. N2a cells were treated with PBS (○, black), or 5 μM GR30 (●, red), PR30 ((◆, magenta), and GR10 (▲, brown) for 4 hours. Then, the result was quantified from ≥250 N2a cells and represented as mean ± SD. The statistical analysis was performed by one-way ANOVA with Tukey post hoc test comparing the PBS-treated group (****P < 0.0001). ns, not significant.

Poly-GR/PR blocks nucleocytoplasmic transport at later stage

We further investigated if nucleocytoplasmic transportation was compromised by the treatment of poly-GR/PR peptides which was assayed in many groups (10–13, 21–24). We constructed a reporter plasmid containing double EGFP conjugated with a nuclear export sequence (NES) and a nuclear localization sequence (NLS) (fig. S10A). The N2a cells were transfected with the EGFP NES-NLS reporter and treated with PBS, GR30, PR30, or GR10 for 4, 8, and 24 hours. If EGFP fluorescence equally distributed in the nucleus and cytoplasm, then the cell was classified as N = C. If EGFP intensity exclusively distributed or had stronger signal in cytoplasm, then the cell was classified as N < C. If EGFP intensity preferentially located in the nucleus, then the cell was classified as N > C. In the PBS-treated group, ~70% of reporter-expressed N2a cells were N = C, while ~30% of the cells were N < C (fig. S10, B and F). In the GR30- and PR30-treated groups, the signal distribution did not differ from the PBS-treated control after 4-hour incubation (fig. S10, C and D) when the transcription inhibition had been observed (Fig. 3, E to G) and membrane damages (Fig. 1K). After 8-hour incubation, GR30- and PR30-treated cells showed a significant reduction in the percentage of cells with equally distributed EGFP signal to 65%. After 24-hour incubation, the percentage reduced to 55 and 50% for GR30- and PR30-treated groups, respectively. In contrast, there was no difference in the GR10 treatment group (fig. S10E). The images were quantified and categorized on the basis of the nucleus and cytoplasm distribution (fig. S10F). The result indicated that poly-GR/PR–elicited impairment of nucleocytoplasmic transport occurred at a later time after the poly-GR/PR–induced DNA and membrane damage.

Negatively charged disaccharides rescue cytotoxicity of GR30

Since poly-GR contains highly positive charges, electrostatic complementarity may neutralize poly-GR. Glycosaminoglycans (GAGs) are polymorphous in negatively charged sulfation at multiple positions, and GAGs with low molecular weight can penetrate blood-brain barrier (BBB) (25, 26). To test whether GAGs could rescue GR30 toxicity, we first examined two types of GAGs, heparan sulfate (HS) and chondroitin sulfate (CS), by mixing heparin and CS with GR30 in MTT assay (Fig. 5A). The results showed that both heparin and CS rescued GR30 toxicity, while heparin performed better than CS. Then, we used a comprehensive HS disaccharide library to screen the effective compounds. The basic unit of HS can be resolved into disaccharides with various modifications. The disaccharide units are mainly composed of N-acetyl-D-glucosamine (GlcNAc) and D-glucuronic acid denoted as NG or composed of GlcNAc and L-iduronic acid denoted as NI. We chemically synthesized a comprehensive HS disaccharide library containing 48 disaccharides with all possible sulfation (27) and a fully sulfated disaccharide sodium sucrose octasulfate (SOS). The representative HS disaccharide structure is illustrated (Fig. 5B), and the comprehensive chemical library is shown in fig. S11. NSC34 cells were incubated with 5 μM GR30 mixed with a single disaccharide from the library at 3, 10, and 30 μM for 24 hours, and the cell viability was evaluated by MTT assay (fig. S12). Four compounds, NG21, NG24, NI11, and NI12, which showed 50% rescue and dose dependency, were selected and subjected to a second dose-dependent study. The cells were treated with GR30 mixed with candidates ranging from 0 to 100 μM, and the normalized cell viability was plotted (Fig. 5C). The median effective concentration (EC50) of the candidates was calculated and shown. The most effective disaccharide was SOS with EC50 at 0.46 μM, and the second best were NG21 and NI11 with EC50 at 6.01 and 8.38 μM, respectively. Among the candidates from the HS library, the highest sulfated chemicals were not always the most potent ones, thus indicating that the position of sulfation played a role in rescuing poly-GR toxicity. We also tested whether SOS rescues PR30 toxicity in the MTT assay (Fig. 5C). We found that SOS could also rescue PR30 toxicity, but the EC50 of PR30 was higher than GR30, that is, 2.42 and 0.46 μM, respectively. To verify SOS diminishing the toxicity of GR30 and PR30, we measured the binding affinity of SOS to GR30 or PR30 using ITC. We found that SOS had a sevenfold higher binding affinity to GR30 than to PR30 (fig. S13), which were 26.3 ± 4.7 and 173.8 ± 22 nM, respectively. The result supported a better rescue ability of SOS to GR30 than PR30.

Fig. 5. A specific sulfated disaccharide reduces cytotoxicity, rescues membrane disruption, and reverses the transcriptional inhibition induced by poly-GR/PR.

(A) Sulfated glycans rescued cytotoxicity induced by GR30. Data were represented as mean ± SD and analyzed by one-way ANOVA with Tukey post hoc test comparing the buffer control (n ≥ 3, ***P < 0.001). (B) The representative structure of the HS disaccharide library. Possible sulfated sites of HS disaccharides are illustrated. (C) Effective concentration of the disaccharide candidates for poly-GR and poly-PR–induced toxicity. Data were represented as mean ± SD (n = 4). (D) Liposome leakage assay under the treatment of poly GR and PR coincubated with SOS. Data were represented as mean ± SD (n = 4). (E) Calcium influx assay for NSC34 cells. Data were represented as mean ± SD and analyzed by two-way ANOVA with Bonferroni post hoc test (n = 4; *P < 0.05, **P < 0.01, and ***P < 0.001). (F) Cross-sectional nuclear membrane profiles for NSC34 nuclei under treatment of poly-GR and poly-PR with and without SOS. Data were represented as mean ± SD (n > 30). Representative images of nuclei treated with SOS (green) and without SOS (magenta) were shown. Scale bars, 10 μm. (G) Quantification of Lamin on nuclear membrane. Data were represented as mean ± SD and analyzed by two-way ANOVA with Bonferroni post hoc test (n > 30; **P < 0.01 and ***P < 0.001). (H) Newly synthesized RNA in NSC34 cells after GR30 or PR30 treatment with and without SOS. Fluorescence was generated by fluorescently labeling ethynyl uridine anchored in nascent RNA. DAPI, blue; labeled RNA, red. Scale bars, 20 μm. (I) Quantification of labeled RNA. Data were represented as mean ± SD and analyzed by two-way ANOVA with Bonferroni post hoc test (n > 420; **P < 0.01 and ***P < 0.001).

A fully sulfated disaccharide, SOS, rescues membrane penetration and transcription inhibition induced by poly-GRs

Next, we performed liposome leakage assay to validate whether the membrane penetration of poly-GR peptides could be rescued by SOS (Fig. 5D). In GR30 at 2 μM, the addition of 2 μM SOS reduced ~30% of the leakage compared with GR30 alone. The addition of 4 μM SOS showed a nearly complete elimination of the leakage. In GR20 at 2 μM, the addition of 6 μM SOS showed the complete elimination of the liposome leakage. In contrast, SOS showed no effective reduction of the liposome leakage for PR30. We also performed calcium influx assay to test whether SOS could reverse the elevated intracellular calcium induced by GR30 and PR30 (Fig. 5E). Without SOS, GR30 at 10 μM induced an approximately threefold calcium influx compared with PBS treatment. With the addition of 10 μM SOS to GR30, the calcium influx was significantly reduced compared with GR30 only. When 40 μM SOS was mixed with GR30, no calcium influx was observed. Meanwhile, PR30 treatment increased ~40% calcium influx compared with PBS-treated cells. When PR30 was mixed with 10 and 40 μM SOS, no calcium influx was observed. Furthermore, nuclear membrane integrity assay was conducted to examine the rescue effect of SOS (fig. S14 and Fig. 5, F and G). The cross-sectional profiles showed that the nuclei incubated with GR30 or PR30 at 10 μM lost the clear boundary shown by Lamin immunostaining (Fig. 5F) compared with the buffer control. When the nuclei were incubated with 40 μM SOS with 10 μM GR30 or PR30, integrity of the nuclear membrane was fully restored in GR30 treatment and partially in PR30 treatment (Fig. 5G). Moreover, treatment of SOS can also reduce the toxicity and damaged cell membrane induced by expression of EGFP-conjugated poly-GR by PI staining (fig. S15A) and calcium influx (fig. S15B), respectively. The results indicated that SOS could rescue the nuclear membrane damage induced by arginine-rich dipeptides.

Next, we wondered whether SOS could rescue transcription inhibition induced by poly-GR. We first used ITC to perform a competitive binding assay. We titrated 400 μM GT6 to the premixed sample of 40 μM GR30 with 120 μM SOS (fig. S16A). The ITC result showed no binding response, demonstrating that GT6 did not affect GR30/SOS complex. In contrast, we titrated 400 μM SOS to the premixed sample of 40 μM GR30 with 120 μM GT6, a binding response representing the displacement of GT6 by SOS was observed (fig. S16B). The data showed that SOS has a higher binding affinity than GT6 to poly-GR. Next, we performed in vitro transcription assay as previously described (fig. S17). While GR30 alone at 10 μM completely inhibited the transcription, coincubation with 40 μM SOS to GR30 fully rescued the transcript level and the effect reduced in a dose-dependent manner. We further used 40 μM SOS to evaluate the effect on nascent RNA synthesis in NSC34 cells. NSC34 cells were incubated with PBS, 10 μM GR30, and 10 μM PR30 with and without 40 μM SOS, and the nascent RNA was labeled (Fig. 5H). After quantification, we found that transcription inhibited by GR30 can be rescued after treatment of fourfold SOS (Fig. 5I), but transcription inhibited by PR30 cannot be rescued, which might be due to lower affinity of SDS to PR30. Also, SOS can rescue nucleocytoplasmic transportation defect caused by treatment of GR30 and PR30 (fig. S18). Together with the nuclear membrane integrity studies, we found that the SOS rescue effect is through direct poly-GR/PR binding and is more potent for GR than PR.

SOS rescues neuronal toxicity in iPSC-derived MNs from a patient with C9-ALS

To examine whether SOS was capable of rescuing toxicity from endogenous poly-GR/PR, we used iPSC-derived MNs from a patient with C9-ALS (C9-iPS-MNs) to perform cytotoxicity assay. C9-iPS-MNs were incubated with SOS for 48 hours. Dead cells were probed by PI staining (Fig. 6A) and quantified (Fig. 6B). The results revealed that the mortality of C9-iPS-MNs was significantly reduced ~40% after treatment with 30 μM SOS, while the control iPS-derived MNs from a healthy individual had no such effect (fig. S19). We further identified the apoptotic MNs by examining the colocalization of the MN marker, neurofilament H, by SMI32 antibody, and active caspase-3 (Fig. 6C). After quantification, we found that active caspase-3 was significantly reduced under the treatment of SOS at 15 and 30 μM. Furthermore, the colocalization of SMI32 and active caspase-3 was reduced by ~8 and ~15%, respectively, which indicated that C9-iPS–derived MNs were rescued by SOS treatment (Fig. 6D). Expression of poly-GR in C9-iPS-MNs was examined by immunocytochemistry staining compared with control-iPS-MNs, and poly-GR was mainly located in nucleus (Fig. 6E). We also performed DNA-protein complex crosslinking using both control- and C9-iPS-MNs incubated with SOS and found SOS significantly reduced poly-GR binding to DNA in C9-iPS-MNs (Fig. 6F). Hence, SOS was able to rescue cell death in C9-iPS derived MNs without decreasing poly-GR expression (Fig. 6, E and F).

Fig. 6. SOS treatment enhances the survival of C9-iPS-MNs.

(A) Survival of C9-iPS-MNs with and without SOS treatment. The dead C9-iPS-MNs were probed by PI staining after the treatment of buffer and 30 μM SOS. Scale bars, 10 μm. (B) Quantification of PI-positive C9-iPS-MNs with 0, 15, and 30 μM SOS. Data were represented as mean ± SD and analyzed by one-way ANOVA with Tukey post hoc test comparing the buffer control (n ≥ 4; **P < 0.01). (C) Representative immunocytochemistry images of C9-iPS-MNs. C9-iPS-MNs were coincubated with and without SOS. Active caspase-3, red; MN, green; nuclei, blue. Scale bars, 10 μm. (D) Quantification of the C9-iPS-MNs under apoptosis. Neurons treated with and without SOS were performed. Data were represented as mean ± SD and analyzed by one-way ANOVA with Tukey post hoc test comparing the buffer control (n > 50; *P < 0.05, **P < 0.01, and ***P < 0.001). (E) Representative immunocytochemistry images of control- and C9-iPS-MNs with and without SOS-treatment. Control- and C9-iPS-MNs were stained with GR antibody. Poly-GR, red; nuclei, blue. Scale bars, 10 μm. (F) Crosslinking assay for DNA-protein complex. C9-iPS-MNs were treated with SOS, and DNA was blotted on the membrane and immunostained with GR antibody. Lysate from C9-iPS-MNs was also immunoblotted with GR antibody to identify the poly-GR level.

SOS prolongs the life span of Drosophila expressing poly-GR and poly-PR with longer repeat numbers

To validate whether SOS could rescue DPR-induced toxicity in vivo, we measured the life span of Drosophila melanogaster harboring elav-GeneSwitch driver for expressing inducible poly-GR with 100 or 36 repeats, denoted as GR100 or GR36, respectively (Fig. 7A and fig. S20A) or poly-PR with 100 or 36 repeats, denoted as PR100 or PR36, respectively (Fig. 7B and fig. S20B) in adult neurons. The flies with 100 repeats, but not with 36 repeats, were previously shown to have disease phenotypes after induction (5). Without induction, the life span of flies was over 50 days. Poly-GR and poly-PR expressions were induced by mifepristone, and the expression of GR100 and PR100 greatly shortened the median life span to 15 and 13 days, respectively. After the addition of 5 μM SOS in the food tube, the median life span significantly increased by ~26 and ~23% to 19 and 16 days for GR100- and PR100-expressing flies, respectively. In contrast, the expression of GR36 and PR36 shorten the median life span to 29 and 30 days, respectively, and the treatment of SOS did not affect the life span. Locomotion was also evaluated after SOS-feeding by monitoring Drosophila climbing to a set threshold in the climbing test. After 5 days of induction when lethality was not substantially distinguishable, both strains of flies with or without SOS feeding performed equally. Only less than 5% of flies were unable to reach the threshold. After 13 days of induction, percentage of GR100 and PR100 flies without SOS feeding increased to 50 and 51% (Fig. 7, C and D) and reduced to 9 and 22% with SOS feeding, respectively. However, treatment of SOS did not affect the locomotion of both GR36 and PR36 flies (fig. S20, C and D). The results demonstrated that SOS can rescue poly-GR/PR–induced toxicity in the C9-FTLD/ALS fly models since the GR36/PR36-expressed flies did not show disease phenotype. Although GR36/PR36 were closer to our peptide model, we found SOS was less effective in the GR36/PR36-expressed flies than in GR100/PR100-expressed flies. The inconsistency may come from the different affinity or toxicity mechanisms between 36 and 100 repeats of arginine-rich DPRs. However, hexanucleotide G4C2 expansion is reported to be more than hundreds of repeats in patients with C9-ALS.

Fig. 7. SOS extends the life span and improves the locomotion of poly-GR– and poly-PR–expressing Drosophila and rescues the motor impairment in the WT mice intracranially injected with GR30.

Survival rate of GR100 (A)– and PR100 (B)–expressing transgenic flies. The survival rate was calculated and analyzed by the log-rank test (n = 120; ****P < 0.0001). The survivorship of Drosophila without DPR expression was shown in black lines. The locomotion of GR100 (C)– and PR100 (D)–expressed flies was examined after 5 and 13 days. Data were represented as mean ± SD and analyzed by one-way ANOVA with Tukey post hoc test comparing the buffer control (n ≥ 3; ****P < 0.0001). (E) Experimental flow chart of the injection mice study. Pretest rotarod task was performed 1 week before motor cortex injection. Tests 1 and 2 of the rotarod task were performed after recovery in the second and third week. Mice were sacrificed after the fourth week. (F) Motor function of the mice performed in rotarod task. The latency was measured in three phases. Data were represented as mean ± SEM. The statistical analysis was performed by two-way ANOVA with Holm-Sidak’s multiple comparisons test [n ≥ 9, treatment, F(3, 35) = 4.26, P < 0.05; time effect, F(2, 70) = 4.55, P < 0.05]. (G) Rotarod performance in test 1. Data were represented as mean ± SEM. The statistical analysis was performed by one-way ANOVA with Holm-Sidak’s multiple comparisons test [n ≥ 9, treatment, F(3, 35) = 4.11, P = 0.01]. (H) Representative NeuN staining images from motor cortex of sample-injected mice. Scale bars, 100 μm. (I) Quantification of NeuN-positive cells from the motor cortex region. Data were represented as mean ± SEM. The statistical analysis was performed by two-way ANOVA with Holm-Sidak’s multiple comparisons test (n = 3; *P < 0.05 and **P < 0.01).

SOS rescues GR30-induced toxicity in a GR30-injected mouse model

To further evaluate the SOS rescuing effect in vertebrates, we performed GR30 injection into the primary motor cortex of wild-type C57BL/6 mice and subjected them to rotarod tests for motor behavior. The mice were tested once before the surgery and twice weekly after the surgery. The experimental flow chart is shown (Fig. 7E). The result showed that the GR30-injected mice had motor function deficit by rotarod assay (Fig. 7F). In contrast, the mice injected with GR30/SOS or SOS-only did not show significant difference from those injected with the buffer. When comparing the performance in test 1 (Fig. 7G), the latency of GR30-treated mice was significantly shortened showing that SOS rescued GR30-induced toxicity. However, in test 2, the GR30-treated mice recovered from the motor deficit, which might have been due to the clearance of the injected peptide or compensatory response to neurotoxicity in the mouse brain. We also examined the neuronal loss in motor cortex of GR30-injected mice using neuronal nuclear protein (NeuN) antibody (Fig. 7H) and found significant reduction of one-third neurons, while SOS prevented the neuronal death from GR30 (Fig. 7I). Overall, the results indicated that motor cortex injection of GR30 caused motor dysfunction, which could be rescued by the cotreatment of SOS.

DISCUSSION

In the present study, we provided evidences to explain the critical number of hexanucleotide expansion in C9ORF72 pathology with a structure-function relationship, in which poly-GRs with over 20 repeats form helices and penetrate plasma/nuclear membranes. This number correlated with the critical number of G4C2 repeats observed from C9 pathology (28, 29). We provided the binding isotherm showing binding affinity of poly-GR to nucleic acids and poly-GR and poly-PR inhibited global transcription in vitro and in cell. Affinity of poly-GR to ssDNA implied that poly-GR may bind to ssDNA and elongate the R-loop during transcription. The elevated R-loop was reported in patients with C9-FTLD/ALS (30). Inhibition of DNA replication in non-neuron cell types could also be possible. It was shown the DPR pathology is present in skeletal muscle of patients with C9-ALS (31).

Previously, compromised RNA maturation was reported as one of poly-GR/poly-PR toxicities through binding to the low-complexity domain of RNA-binding proteins (7). Proteomic approaches were used to investigate the protein interactors of poly-GR/poly-PR either through applying PR peptide (32) or transfecting GFP-conjugated poly-GR/poly-PR (33). Here, we proposed that poly-GR/poly-PR also directly interacted with nucleotides through the strong electrostatic force and interfered with normal functions of nucleotides. The results expand the list of interactors for poly-GR/poly-PR. About the cell penetration mechanism, although poly-PR did not form helical conformation in solution, poly-PR is still able to penetrate membrane and enter nucleus with a lesser degree comparing to poly-GR. This might be due to conformational changes of poly-PR in membrane or via a different cell entry mechanism as indicated previously by clatherin-dependent mechanism (34). In addition, compromised nucleocytoplasmic transportation was reported as an important toxicity mechanism induced by poly-GR/poly-PR (10–13). Genes involving nucleocytoplasmic transportation were identified from yeast (22) or Drosophila (21) screening which can modify poly-GR/poly-PR toxicity. Poly-PR was also found to directly binding to nuclear pores of Xenopus laevis oocyte (23) and β-karyopherin family (24). Beside proteins, RNA transportation is also compromised in Drosophila-expressed G4C2 expansions (21). Here, we found that poly-GR/poly-PR caused nucleocytoplasmic transport defect in a time-dependent manner occurred later than membrane and DNA damages.

As only a few US Food and Drug Administration–approved drugs, such as Riluzole and Edaravone, with modest effects are available, we identified sulfated disaccharides that bind to poly-GR to reduce its membrane penetration ability and rescue cytotoxicity in MN-like cells, primary neurons, C9-patient iPSC-derived MNs, transgenic flies, and a mouse model. Although we used longer GR/PR repeats in the transgenic fly model to demonstrate the rescue effect by SOS, which is different from the in vitro and the peptide-injected mouse studies, it is due to lacking of recovered life spans by SOS-treatment in the GR36/PR36 expressing flies. This could be attributed from the protecting effects in vivo. In this study, we proposed a common mechanism shared by poly-GR and poly-PR for the SOS rescue effect. Our results indicated that SOS protects the membrane penetration by poly-GR/poly-PR and hinders the nucleic acid binding of poly-GR/poly-PR most likely through electrostatic interactions between SOS and poly-GR/poly-PR. Heparan sulfate derivatives can penetrate the BBB in vivo (26). Hence, potential HSs considering the length, degree and position of sulfation, and BBB penetration could be examined in the near future. Together, our study provided a structure-function relationship for poly-GR/PR–induced toxicity via membrane binding and DNA interaction and provided a potential compound for C9-ALS therapeutic development.

MATERIALS AND METHODS

Peptide synthesis and preparation

Poly-GR peptides with 15, 20, 25, and 30 repeats were synthesized through Fmoc solid state synthesis in the Genomics Research Center, Academia Sinica, Taiwan. Poly-GR peptides with 5 and 10 repeats were purchased from Kelowna International Scientific, Taiwan. The peptides were directly dissolved in 10 mM phosphate buffer [150 mM NaCl (pH 7.4)] (PBS150). The stock solution was filtrated through a 0.2-μm filter membrane.

Cell viability and cytotoxicity assays

Mouse neuroblastoma N2a cells were maintained in minimum essential medium with 10% fetal bovine serum (FBS) while assayed in Dulbecco’s modified eagle medium (DMEM), phenol red-free, with 2% FBS. Primary rat cortical neurons were cultured in neurobasal medium with 2% B-27 supplement and penicillin/streptomycin (100 U/ml; Invitrogen). Then, they were replenished every 2 days. Toxicity assays were performed with 20,000 N2a neuroblastoma cells per well seeded in a 96-well plate and incubated for 24 hours. For primary neurons, 40,000 cells per well were assayed in day in vitro (DIV) 7. Cells were treated with DPRs at indicated concentrations. In the MTT assay, the samples were incubated with the cells for 24 hours followed by the addition of MTT solution for 3 hours. The media were discarded, and the violet crystals were dissolved by dimethyl sulfoxide (DMSO). The absorbance at 570 nm was measured in a microplate reader SpectraMax M5 (Molecular Devices) and subtracted to the background at 690 nm. The data were averaged (n = 6), normalized by the buffer control, and represented as mean ± SD. In the LDH assay, the cells were treated with GR peptides for 5 hours, and an equal volume of LDH reagent was added following the manufacturer’s protocol (Promega). The fluorescence was measured at 590 nm, while the excitation was at 560 nm. The LDH activity was calculated from the slope of ascending fluorescence during the first 10 min right after addition of LDH reagent. The slopes were averaged (n = 6) and normalized by the buffer control. In the ROS assay, the cells were treated with 10 μM of ROS indicator, CM-H2DCFDA, for 1 hour followed by treating poly-GR peptides for 15 hours. The fluorescence at 535 nm was recorded, while the excitation was at 490 nm. The data were averaged (n = 6), normalized to the buffer control, and represented as mean ± SD. In the caspase-3 activity assay, the cells were treated with GR peptides for 24 hours followed by the removal of the medium and cell lysis. Caspase-3 activity was monitored by applying 50 μM caspase-3 substrate z-DEVD-AFC. The slope of ascending fluorescence was recorded at 505 nm, while the excitation was at 400 nm. The data were averaged (n = 6), normalized to the buffer control, and represented as mean ± SD. The statistical analysis for each group compared with the buffer control was performed by two-way analysis of variance (ANOVA) with Bonferroni post hoc test (*P < 0.05, **P < 0.01, and ***P < 0.001).

SAXS measurement

Poly-GR samples with different peptide lengths were prepared in PBS150. The concentration for GR5, GR10, and PR10 was at 40 mg/ml and for GR15, GR20, GR25, GR30, and GP30 at 20 mg/ml. Each sample was filtrated through a 0.2-μm filter membrane. The samples were measured at 288 K either at the Taiwan Photon Source (TPS) 13A BioSWAXS or at the Taiwan Light Source (TLS) 23A SWAXS end station, National Synchrotron Radiation Research Center (Hsinchu, Taiwan) (35–37). One hundred microliters of each sample was injected into a high-performance liquid chromatography (HPLC) system (Agilent chromatographic system 1260 series) with a flow rate of 0.35 ml/min. This HPLC system was integrated with SAXS/ultraviolet-visible (UV-Vis)/refractive index (RI) detections as described in a previous report (36–38). The column-bypass mode of the HPLC system was selected because the unwanted strong interaction between poly-GR dipeptides and silica gel inside the size exclusion column would lead to sample entrapment inside the column. A 15-keV x-ray beam irradiated on the sample flow within a thin-wall (30 μm) quartz capillary cell (2.0 mm in diameter) for SAXS measurements with simultaneous online UV-Vis absorption data acquisition, followed by downstream RI detection. The sample-to-detector distance of 2.52 m together with the TLS 23A Pilatus 1 M-F area detector covered an x-ray scattering vector q-range of 0.01 to 0.35 Å−1, where q = 4πλ-1sinθ was defined by the scattering angle 2θ and x-ray wavelength λ (=0.8266 Å used); the data measured at the TPS 13A with an Eiger X 9M detector could cover the q-range up to 0.7 Å−1. The 2D SAXS patterns were radially averaged, background-subtracted, and scaled to the absolute scattering intensity in units of cm−1 via the scattering intensity of water at the same temperature. Primary SAXS data reduction was performed using an in-house data reduction kit. I(0) and Rg were evaluated using the Guinier approximation, with the limit qRg < 1.3 (39). UV-Vis absorption at 216 nm (from amide bonds) was used to monitor the elution of samples due to the lack of aromatic residues in poly-GR.

I-TASSER/Rosetta structure calculations and implementation of SAXS constraint

The initial models for poly-GR peptides were constructed using I-TASSER protein structure, function prediction server, or AlphaFold (40–43). The program generated three-dimensional atomic models from multiple threading alignments and iterative structural assembly simulations. For poly-GR peptides, the Protein Data Bank library had no structurally related proteins. Consequently, the structure was built from scratch by ab initio modeling (44). The theoretical scattering curves of the output five models with the lowest free energy were computed using the program CRYSOL (45), which fit the experimental scattering curve using only two parameters of the average displaced solvent volume per atomic group and the contrast of the hydration layer. The model with the lowest χ2 (goodness of fit) was selected for further energy minimization with SAXS constraint using Rosetta modeling suite (46–48). SAXS constraint was applied by adding the fastsaxs term in Rosetta default scoring function Talaris2014 (49). The fastsaxs term is additive to the overall score and is based on the χ2 value between the experimental SAXS data and the modeled SAXS profile during energy minimization, which guides the model search toward a structure that matches reasonably well with the SAXS data. The two-step minimization procedure proceeded with the backbone minimization stage first, followed by the side chain minimization.

Liposome leakage assay

Liposomes were prepared from 1,2-dipetroselenoyl-sn-glycero-3-phosphocholine [catalog no. 850374P, 18:1 (Δ6-Cis) PC, Avanti Polar Lipids Inc.], 1,2-dioleoyl-sn-glycero-3-phospho-(1′-rac-glycerol) [catalog no. 840475P, 18:1 (Δ9-Cis) PG) (Avanti Polar Lipids Inc.], and cholesterol (catalog no. C8523, Sigma-Aldrich) at the molar ratio of 3.5:3.5:3. Each lipid powder was dissolved in chloroform and mixed. Then, the solvent was evaporated in a 100-ml round-bottom flask. Residual chloroform was eliminated in a vacuum for an additional 12 hours. Fifty millimolars of 5(6)-carboxyfluorescein (Sigma-Aldrich), a fluorescent dye used to indicate leaked liposome, was prepared in 10 mM phosphate buffer [100 mM NaCl (pH 7.4)]. The oversaturated dye powder was fully dissolved by titrating 6 M NaOH, and the pH was neutralized. The dye solution was mixed with the lipid film and fully dissolved by sonication at 25°C for 1 hour. The solution was further homogenized through Avanti Mini Extruder seven times using a membrane with 100-nm pore size. Liposome was separated from the unpacked dye using a PD-10 desalting column (GE Healthcare Life Sciences, IL, USA). Before being assayed, the liposome was diluted in 10 mM phosphate buffer [100 mM NaCl (pH 7.4)]. Then, it was mixed with the 2.5 μM DPR samples, and the liposome leakage was monitored for 1 hour. The fluorescence of 5(6)-carboxyfluorescein was measured (excitation/emission = 492/517 nm), averaged (n ≥ 3), and normalized to buffer-treated and 1% Triton X-100–treated groups. In the test of SOS rescuing the leakage, liposome was treated with 2 μM DPR samples mixed with 0, 2, 4, 6, 8, and 10 μM SOS. The assay was monitored for 2 hours. Data were represented as mean ± SD (n = 4).

Calcium influx assay

The fluorescent probe, Fluo-4 acetoxymethyl (AM) ester(Invitrogen), was used to evaluate the elevation of intracellular calcium. A total of 30,000 NSC34 cells per well were seeded in a 96-well plate and incubated for 24 hours. Fluo-4 AM was dissolved in DMSO, diluted into the final concentration at 5 μM in each well, and incubated with NSC34 cells for 1 hour. Peptides at 5 μM were added into each well, and the cells were further incubated at 37°C for 3 hours. Elevated intracellular calcium was quantified using the fluorescein isothiocyanate (FITC) channel of an Cellomics ArrayScan VTI HCS reader (Thermo Fisher Scientific, MA, USA) with a 20× objective lens (n ≥ 4, scanning of 45 fields per well). In the test of SOS rescuing calcium influx, NSC34 cells were treated with PBS, 10 μM GR30, 10 μM GR30 with 10 μM SOS, 10 μM GR30 with 40 μM SOS, 10 μM PR30, 10 μM PR30 with 10 μM SOS, and 10 μM PR30 with 40 μM SOS. Data were represented as mean ± SD. The statistical analysis was performed by two-way ANOVA with Bonferroni post hoc test (*P < 0.05, ***P < 0.001, and ****P < 0.0001). In the test of SOS rescuing calcium influx of N2a cells expressed EGFP, EGFP-GR30, and EGFP-GR150, cells were first transfected with the indicated construct using Lipofectamine 3000 reagent following the manufacturer’s procedures for 24 hours. N2a cells were treated with and without 100 μM SOS for another 48 hours. Elevated intracellular calcium was quantified using the FITC channel of an Cellomics ArrayScan VTI HCS reader (Thermo Fisher Scientific, MA, USA) with a 20× objective lens (n ≥ 4, scanning of 45 fields per well).

Nucleus isolation

NSC34 cells from a 10-cm dish with 80% confluence were washed by PBS twice and collected in 1.5 ml of hypotonic buffer [10 mM Hepes (pH 7.6), 10 mM KCl, and 1 mM EDTA]. Cells were incubated on ice for 15 min. Then, 150 μl of 5% NP-40 was added, and the cells were shortly vortexed twice. Nuclei were transferred to a 12-well plate with coverslips and attached to the coverslips by centrifugation at 400g and at 4°C for 6 min. The coverslips were washed with PBS twice. The samples were diluted in PBS at 20 μM and coincubated with the nuclei at 37°C for 24 hours. For SOS experiments, the peptides at 10 μM with and without SOS at 40 μM were coincubated with the nuclei. After incubation, the nuclei were fixed and stained with Lamin a/c immunoglobulin G (IgG) at the ratio of 1:500 (Cell Signaling) and Alexa Fluor 488-conjugated anti-mouse IgG at the ratio of 1:1000. Then, they were labeled with DAPI and observed under a Leica TCS-SP5-MP-SMD confocal microscope with HC PL APO 100×/1.40 oil CS2 objective lens. Quantification was first performed by cross-sectional intensity of Lamin versus the distance from the nucleus center (n = 15 to 25 for Fig. 2; n > 30 for SOS experiments). Data were represented as mean ± SD on the profiles. Then quantification was also performed by calculating the percentage of Lamin on the nuclear membrane. Lamin signals were analyzed at the outer concentric ring of the nucleus (n = 15 to 25 for Fig. 2; n > 30 for SOS experiments) and normalized to total Lamin intensity. Data were represented as mean ± SD. The statistical analysis was performed by one-way ANOVA with Tukey post hoc test in Fig. 2B (***P < 0.001) and two-way ANOVA with Bonferroni post hoc test in Fig. 6G (***P < 0.001 and ****P < 0.0001).

Isothermal titration calorimetry

Titration was performed at 37°C using MicroCal iTC200 (Malvern analytical, Malvern, UK). The synthetic poly-GR peptides at 200 μM were loaded into the syringe and were titrated to 20 μM single-stranded DNA/RNA, 10 μM double-stranded DNA, or other nucleic acids with indicated concentrations. In the test of AC6 binding to GR30 or PR30, 150 μM GR30 or PR30 was titrated to 10 μM AC6. In the test of SOS binding to GR30 or PR30, 100 μM GR30 or PR30 was titrated to 40 μM SOS. In the test of SOS competing single-strand DNA binding to GR30, 40 μM GR30 was premixed with 120 μM SOS, and 400 μM GT6 was titrated to the mixture. For the reversed condition, 40 μM GR30 was premixed with 120 μM GT6, and 400 μM SOS was titrated to the mixture. All samples were prepared in 10 mM phosphate buffer [150 mM NaCl (pH 7.4)]. The volume of each injection was 2 μl for Fig. 3A and 1.5 μl for fig. S5 except for an initial injection of 1 μl. The total injected volume of the titrant was less than 39 μl. Calorimetry was collected, analyzed using the software Origin 7.0 provided by MicroCal, and fitted with a one-site model (50, 51). The averaged Kd and SD are shown (n = 3).

DNA-protein complex crosslinking

The assay was performed either on N2a neuroblastoma cells transfected with EGFP, EGFP-GR15, EGFP-GR30, and EGFP-GR80 using Lipofectamine 3000 reagent following the manufacturer’s procedures or on control- and C9-iPS-MNs. For total lysate of C9-iPS-MNs, cells were directly lysed using radioimmunoprecipitation assay buffer without DNA-protein crosslinking, and the supernatant was collected. For DNA-protein crosslinking, cells were fixed by 1% paraformaldehyde for 10 min, neutralized by glycine at 0.125 M for 5 min, washed by PBS, and harvested. Cytosol was removed by using cytosol lysis buffer containing 5 mM Hepes at pH 8.0, 85 mM KCl, 0.5% NP-40, 1 mM dithiothreitol (DTT), and protease inhibitor cocktails. Nuclei were collected by centrifugation at 9000g for 5 min and resuspended in nucleus lysis buffer containing 50 mM tris-Cl (pH 8.1), 10 mM EDTA, 1% SDS, 1 mM DTT, and protease inhibitor cocktails. DNA was fragmented by sonication for 5 min using the Misonix sonicator 3000 (power level 7, pulse rate with 3 s on and 3 s off). DNA-protein complex was collected from the supernatant after centrifugation at 9000g for 5 min and precipitated using the isopropanol method described below. The supernatant was diluted in 20-fold volume of buffer containing 50 mM tris-Cl (pH 8.5) and 1.25 M NaCl, and an equal volume of cold isopropanol was added to the mixture followed by centrifugation at 13,000g for 30 min at 4°C. Pellet was rinsed by 2 ml of 70% ethanol, and ethanol was removed by centrifugation at 13,000g for 15 min at 4°C. DNA-protein complex was recovered by dissolving the pellet in 10 mM tris-Cl (pH 8.5) and quantified using the NanoDrop ND-1000 (Thermo Fisher Scientific). Twenty micrograms of DNA from transfected N2a neuroblastoma and 1 μg of DNA from control- and C9-iPS-MNs were loaded to a nitrocellulose membrane with 0.45-μm pore-size. GFP antibody (ab290, Abcam) and GR antibody (MABN778, Millipore) were used to identify the protein cross-linked to DNA at a ratio of 1 to 4000 diluted in the 5% milk followed by probing with horseradish peroxidase (HRP)–conjugated anti-rabbit and anti-rat IgG, respectively.

Transcription assay

Transcription assay in vitro was conducted using Riboprobe Systems (catalog no. P1440, Promega, WI, USA) following the protocol provided by the manufacturer. The active reagent was prepared by mixing the pGEM express template with the reaction mixture containing 5× optimized buffer, DTT, ribonuclease inhibitor, ribonucleoside triphosphates, and T7 polymerase. Transcripts containing two bands of 2346 and 1065 bases were analyzed by agarose electrophoresis. Each reaction was mixed with serially diluted DPRs in PBS150 at final concentrations of 0, 2.5, 5, 7.5, and 10 μM. For SOS experiments, GR30 was mixed with various concentrations of SOS before subjected to the mixed reagent. The volume of GR30/SOS mixture was one-tenth to the mixed reagent. Mixtures were incubated at 37°C for 2 hours and loaded into a 1% agarose gel for electrophoresis. The gel was premixed with a nucleic acid dye, SafeView Classic (Applied Biological Materials, Richmond, BC, Canada), and images were captured by an ImageQuant LAS 4000 imaging system (GE Healthcare Life Sciences, IL, USA).

RNA labeling and flow cytometry

NSC34 cells were seeded in 12-well plates with coverslips for 24 hours followed by the treatment of DPRs at 5 μM for 4 hours. The RNA labeling was performed following the protocol provided by the manufacturer of the RNA labeling kit (catalog no. K718, BioVision, CA, USA) to incorporate 5-ethynyl-UTP into nascent RNA. After the treatment, cells were incubated with media-diluted EZClick RNA Label solution (diluted from 100× to 1×) for 1 hour, washed twice by PBS, fixed for 15 min, washed twice again, perforated by permeabilization buffer for 15 min at room temperature in the dark, and washed. RNA labeling by click reaction was done by mixing the cocktails containing the reducing agent, copper, and the fluorescent dye followed by incubating the cells in the dark for 30 min. Extra dye was removed, and cells were stained with DAPI, mounted, and observed under a Leica TCS-SP5-MP-SMD confocal microscope. To quantify nascent RNA from the fluorescence images, 40,000 NSC34 cells were seeded in a 96-well plate for 24 hours and treated with DPRs for 4 hours. The nascent RNA was labeled as described above and stained with DAPI. Images were obtained by a Cellomics ArrayScan VTI HCS reader (Thermo Fisher Scientific). The images were quantified from three replicates with 8000 cells per replicate. Data were represented as mean ± SD. The statistical analysis was performed by one-way ANOVA with Tukey post hoc test comparing the buffer control (***P < 0.001). For SOS experiments, after the treatment of PBS, 10 μM GR30, and 10 μM PR30, with and without 40 μM SOS, labeled RNA was imaged using a Leica TCS-SP5-MP-SMD confocal microscope with HC PL APO 100×/1.40 oil CS2 objective lens. Images were analyzed by open-source software ImageJ [National Institutes of Health (NIH), USA], and the statistical analysis was performed by two-way ANOVA with Bonferroni post hoc test (n ≥ 400; **P < 0.01 and ****P < 0.0001).

To quantify the nascent RNA by flow cytometry, NSC34 cells were seeded in a 6-cm plate for 24 hours, treated with buffer, GR30, GR20, GR10, and actinomycin D for 4 hours, harvested on ice, and collected by centrifugation at 500g for 5 min. Then, the staining procedures were also performed. Stained cells were reconstituted in 1 ml of PBS in the dark, applied to a FACSCanto II flow cytometry system (BD Biosciences, NJ, USA), and analyzed using the software FlowJo (BD Biosciences, NJ, USA). The cell counts versus fluorescence intensity was plotted.

R-loop staining

Total number of 300,000 primary cortical neurons was seeded on poly-l-lysine–coated coverslips placed in a 12-well plate. Primary cortical neurons were cultured in neurobasal medium containing 1× B27, 1× penicillin/streptomycin, and 1× GlutaMAX starting from DIV0. Half volume of the medium was replenished every 3 days. Neurons at DIV7 were treated with DPR peptides at 2.5 μM for 24 hours. For expression of mCherry-GR150, primary cortical neurons were transfected at DIV6 using Lipofectamine 3000 reagent following the manufacturer’s procedures and incubated for 2 days. Samples were fixed by 4% paraformaldehyde for 15 min, washed, permeabilized by 0.3% Triton X-100 for another 15 min, and washed again. Blocking was performed by 3% bovine serum albumin (BSA) for 1 hour. R-loop staining was performed by using DNA/RNA hybrid antibody, S9.6, (catalog no. ENH001, Kerafast, Boston, MA, USA) at the ratio of 1 to 100 followed by staining with Alexa Fluor 488-conjugated anti-mouse secondary antibody (Thermo Fisher Scientific, MA, USA). Nuclei were stained by Hoechst dye, and samples were mounted with ProlongGold mounting buffer (Thermo Fisher Scientific, MA, USA). Images were captured under a Leica TCS-SP5-MP-SMD confocal microscope with HC PL APO 100×/1.40 oil CS2 objective lens.

γH2AX staining

N2a cells were seeded in 12-well plates with coverslips for 24 hours followed by the treatment of DPRs at 5 μM for 4 hours. The cells were fixed by 4% paraformaldehyde for 15 min and stained with phospho-H2A.X antibody (catalog no. 2577, Cell Signaling Technology, MA, USA) at the ratio of 1:100 and Alexa Fluor 488-conjugated anti-rabbit IgG (Thermo Fisher Scientific, MA, USA) at the ratio of 1:1000. Then, the cells were mounted by ProlongGold mounting buffer (Thermo Fisher Scientific, MA, USA). Images were captured under a Leica TCS-SP5-MP-SMD confocal microscope with HC PL APO 100×/1.40 oil CS2 objective lens, and the total number, area, and intensity of γH2AX positive particles in each nucleus was analyzed by open-source software ImageJ (NIH, USA).

Drug screening for rescuing GR30 toxicity

Chondroitin sulfate and heparin were purchased from Sigma-Aldrich, MO, USA. The toxicity was examined using MTT assay. Polysaccharides and pure disaccharides were prepared and diluted in MilliQ water. A total of 40,000 NSC34 cells were seeded in the 96-well plate for 24 hours. In the test of commercial polysaccharides rescuing effect, the cells were cotreated with GR30 at 5 μM and polysaccharides at 200 ng/ml for 24 hours. Data from more than three replicates were averaged, normalized, and represented as mean ± SD. The statistical analysis was performed by one-way ANOVA with Tukey post hoc test (***P < 0.001). In the first disaccharides screen, the cells were added with GR30 at 5 μM coincubated with disaccharides at 10 μM. Data from three replicates were averaged, normalized, and represented as mean ± SD. In further examination, GR30 or PR30 at 5 μM coincubated with the selected candidates at 100, 30, 10, 5, 2.5, 1, 0.25, 0.1, and 0.05 μM were applied to NSC34 cells. The data from four replicates were averaged and normalized to the buffer-treated group as 1 and the GR30- or PR30-treated group as 0. The data were fitted, and EC50 was calculated.

Motor neuronal differentiation

The procedure for motor neuronal differentiation was performed as previously described (52). Briefly, iPSC colonies from an ALS patient with C9ORF72 expansion (male, 49 years old) and from a healthy individual (male, 19 years old) were differentiated in a neural medium consisting of DMEM/F12 and neurobasal medium at 1:1 ratio, 0.5× N-2, 0.5× B27, 1× GlutaMAX, 1× penicillin/streptomycin (all from Thermo Fisher Scientific, MA, USA), and 0.1 mM ascorbic acid (Sigma-Aldrich, MO, USA). Three micromolars of CHIR99021 (Torcris, Bristol, UK), 2 μM DMH1 (Torcris, Bristol, UK), and 2 μM SB431542 (Sigma-Aldrich, MO, USA) were added to the medium. On day 6, the cells were split with dispase (1 mg/ml) and cultured in the same medium described above. Retinoic acid at 0.1 μM (Sigma-Aldrich) and purmorphamine (Stemgent, MD, USA) at 0.5 μM were added in combination with 1 μM CHIR99021, 2 μM DMH1, and 2 μM SB431542. On day 12, the cells were split again with dispase and culture in suspension with the same neural medium containing 0.5 μM RA and 0.1 μM Purmorphamine. On day 18, cells were then dissociated with Accutase (Thermo Fisher Scientific, MA, USA) into single cells and plated on Matrigel-coated plates in the neural medium with 0.5 μM RA, 0.1 μM purmorphamine, 0.1 μM compound E (Sigma-Aldrich, MO, USA), 1 μM cyclic adenosine 3′,5′-monophosphate (Sigma-Aldrich), brain-derived neurotrophic factor, cerebral dopamine neurotrophic factor, insulin-like growth factor, and ciliary neurotrophic factor (10 ng/ml; all from PeproTech, NJ, USA) for another 6 days.

PI staining and immunostaining of control- and C9-iPSC derived MNs

For PI staining, differentiated cells were incubated with 15 and 30 μM SOS for 48 hours and stained with 0.1 mM PI for 15 min. Dead cells were monitored and calculated using the tetramethyl rhodamine isothiocyanate channel of Cellomics ArrayScan VTI HCS reader (n ≥ 4, scanning of 30 fields per well). Data were represented as mean ± SD. The statistical analysis was performed by one-way ANOVA with Tukey post hoc test comparing the buffer control (**P < 0.01). For apoptosis detection, the neurons treated with 0, 15, and 30 μM SOS were stained by active caspase-3 antibody, 5A1E, at the ratio of 1 to 100 and neurofilament antibody, SMI-32 at the ratio of 1 to 500 followed by staining of respective secondary antibodies. Quantification of colocalized active caspase-3 signal and MN marker was calculated from more than 50 images. Data were represented as mean ± SD. The statistical analysis was performed by one-way ANOVA with Tukey post hoc test comparing the buffer control (**P < 0.01). For poly-GR detection, GR antibody (MABN778, Millipore) was applied at the ratio of 1 to 200 to both control- and C9-iPS-MNs.

Drosophila climbing and longevity assay

Flies (elav-GS/UAS-GR100, elav-GS/UAS-PR100, elav-GS/UAS-GR36, and elav-GS/UAS-GR36) were reared in regular food at 25°C. Five strains of flies, elav-GS (BDSC #43642), UAS-poly-GR.po-100 (BDSC #58696), UAS-poly-PR.po-100 (BDSC #58698), UAS-poly-GR.po-36 (BDSC #58692), and UAS-poly-PR.po-36 (BDSC #58694), were adopted. Sixty male and female adult flies with eclosion within 5 days were used. Then, 1 mM RU486 was used to induce the expression of poly-GR/PR from the adult stage. For pharmacological treatment, food was mixed with 5 μM SOS. Food with RU486 or RU486 combined with SOS was replaced every 3 days. For the longevity analysis, dead flies were counted and removed when new food was exchanged. The survival rate was calculated and analyzed by the log-rank test (****P < 0.0001). To analyze the climbing ability, flies were tapped onto the bottom of an empty new vial. A height of 3 cm was used to measure the climbing activity of the flies. During this time, the flies were video recorded. The flies that failed to go over this height within 10 s were counted as defective locomotion. Data were represented as mean ± SD (n ≥ 3). The statistical analysis was performed by one-way ANOVA with Tukey post hoc test comparing respective DPR-induced group (****P < 0.0001).

Animals

Male C57BL/6 mice were obtained from the National Laboratory Animal Center, Taiwan and were used in accordance with the NIH Guideline for Animal Research (Guide for the Care and Use of Laboratory Animals). Five mice were housed in each cage and fed ad libitum in a controlled environment with a 12-hour light-dark cycle.

Surgical procedures

Mice were anaesthetized by the intraperitoneal administration of a mixture of Zoleti 50 (70 mg/kg) and Rompun (653 μg/kg). In addition, they were placed in a stereotaxic apparatus. The skull was bilaterally drilled over the primary motor cortex (M1; coordinates: anteroposterior = −0.1 mm; mediolateral = ±1 mm; dorsoventral = −1 mm form skull). The mice received a coinfusion of GR30 (400 mg/ml), SOS (588 mg/ml), GR30 with SOS (molar ratio, ~1 to 16), or buffer control. The infusion solution was prepared in a 1:1 volume ratio. Then, 1 μl of each mixed liquid was injected into the motor cortex over 10 min by using a 26G-Hamilton microsyringe connected to a microinfusion pump. After injection, the needle was left in place for 5 min to avoid reflux. The scalp was closed with chromic catgut, and all subjects were given a recovery period of at least 11 days.

Rotarod task

The MK-660D rotarod treadmill (Muromachi Kikai Co. Ltd.) was used to evaluate the motor function by measuring the latency on the rod. The rotarod task was divided into three phases: pretraining, test 1, and test 2. One week before the surgery, the mice underwent a pretraining phase, during which the behavioral baseline scores were collected. After the surgery and the recovery period, each mouse received one rotarod task per week for two consecutive weeks. The mice were placed on a rotating axle with a constant speed at 4.5 rpm for 1 min to habituate. Then, the rod speed was gradually accelerated from 4.5 to 45 rpm over 300 s. The latency to fall from the rod was recorded, and the mice were examined three times per day with a trial interval of 25 min. Data were averaged (n ≥ 9) from three trials and represented as mean ± SEM. The statistical analysis was performed by two-way ANOVA with Holm-Sidak’s multiple comparisons test [treatment, F(3, 35) = 4.26, P < 0.05; time effect, F(2, 70) = 4.55, P < 0.05]. In the analysis of test 1, data were represented as mean ± SEM. The statistical analysis was performed by one-way ANOVA with Holm-Sidak’s multiple comparisons [treatment, F(3, 35) = 4.11, P = 0.01].

Brain tissue preparation

All mice were received an intraperitoneal injection containing the mixture of tranquilizer (Zoleti 50, 0.1 μg/g body weight, Vibrac, Amherst, MA, USA) and analgesics (Rompun, 1 μl/g body weight, Bayer, Toronto, Canada). After perfusion with ice-cold PBS, mouse brains were collected immediately. One hemisphere was postfixed in 10% formalin, and the other hemisphere was frozen and stored in −80°C. The formalin-fixed brains were dehydrated in sucrose solution and cut to 20-μm coronal frozen sections for immunohistochemistry.

Immunohistochemistry staining

The brain sections were mounted on an adhesive glass slide (Matsunami, Japan), and antigen retrieval was done by incubating in citric acid buffer [10 mM citric acid with 0.05% Tween 20 (pH 6.0)] at 90°C for 30 min. To remove the effect of endogenous peroxidase and reduce nonspecific binding, the slides were immersed in 1% hydrogen peroxidase for 30 min followed by blocking in 3% BSA, which is prepared in PBS containing 0.1% Triton X-100 for 1 hour. The sections were incubated at room temperature with NeuN (1:500; GeneTex, GTX132974, Taiwan). The signal was developed by an immunohistochemistry kit (Mouse/Rabbit PolyDetector Plus DAB HRP Brown Detection System, Bio SB, Goleta, CA, USA), and immunoreactivity was detected by 3,3′-diaminobenzidine provided in the kit. Slides were scanned through an Aperio GT 450 scanner (Leica Biosystems, Buffalo Grove, Illinois, USA). NeuN staining was further analyzed by open-source software ImageJ (NIH, USA) for counting survived neurons.

Statistics

The statistical analysis in the study was done with GraphPad Prism (version 9.3.1) adopting either one-way or two-way ANOVA (analysis of variance) with correction for the recommended multiple comparison tests as described. Analysis of Drosophila longevity was done with log-rank test as described.

Acknowledgments

We thank the imaging core facility, the Institute of Cellular and Organismic Biology, and the peptide synthesis core, the Genomics Research Center, Academia Sinica, Taiwan for assisting us with the immuno-TEM and peptide synthesis, respectively. We thank K. F. Liao and Y. Q. Yeh for the assistance on bioSAXS measurements. We thank T.-F. Lee for helping in fly collection and maintenance. We thank the Academia Sinica SPF Animal Facility for mouse housing and animal support.

Funding: The study is funded to Y.-R.C. by Academia Sinica (AS-TP-109-LM-08) and Ministry of Science and Technology, Taiwan (MOST 111-2113-M-001-024). The animal core was funded by Academia Sinica Core Facility and Innovative Instrument Project (AS-CFII-111-204).

Author contributions: Conceptualization: Y.-R.C. Investigation: Y.-J.C., K.-T.L., O.S. (SAXS), and C.-H.Y. (mouse). Resources: C.-Y.C. and H.-C.K. (iPSC-derived MNs), M.-H.F. and C.-K.Y. (adult flies), W.-B.L. (primary culture), Y.-C.L. (clinical specimens of C9 patients), and S.-C.H. (disaccharide library). Visualization: Y.-J.C. and C.-H.Y. Supervision: Y.-R.C. Writing—original draft: Y.-J.C., U.S.J., C.-H.Y., C.-K.Y., and Y.-R.C. Writing—review and editing: Y.-J.C., U.S.J., and Y.-R.C.

Competing interests: Y.-R.C., Y.-J.C., and S.-C.H. are inventors on provisional patent application submitted by Academia Sinica that covers “Saccharides and uses thereof in treating neurological diseases.” All other authors declare that they have no competing interests.

Data and materials availability: All data needed to evaluate the conclusions in the paper are present in the paper and/or the Supplementary Materials.

Supplementary Materials

This PDF file includes:

Supplementary Methods

Figs. S1 to S20
==== Refs
REFERENCES AND NOTES

1 M. DeJesus-Hernandez, I. R. Mackenzie, B. F. Boeve, A. L. Boxer, M. Baker, N. J. Rutherford, A. M. Nicholson, N. A. Finch, H. Flynn, J. Adamson, N. Kouri, A. Wojtas, P. Sengdy, G.-Y. R. Hsiung, A. Karydas, W. W. Seeley, K. A. Josephs, G. Coppola, D. H. Geschwind, Z. K. Wszolek, H. Feldman, D. S. Knopman, R. C. Petersen, B. L. Miller, D. W. Dickson, K. B. Boylan, N. R. Graff-Radford, R. Rademakers, Expanded GGGGCC hexanucleotide repeat in noncoding region of C9ORF72 causes chromosome 9p-linked FTD and ALS. Neuron 72 , 245–256 (2011).21944778
2 A. E. Renton, E. Majounie, A. Waite, J. Simón-Sánchez, S. Rollinson, J. R. Gibbs, J. C. Schymick, H. Laaksovirta, J. C. van Swieten, L. Myllykangas, H. Kalimo, A. Paetau, Y. Abramzon, A. M. Remes, A. Kaganovich, S. W. Scholz, J. Duckworth, J. Ding, D. W. Harmer, D. G. Hernandez, J. O. Johnson, K. Mok, M. Ryten, D. Trabzuni, R. J. Guerreiro, R. W. Orrell, J. Neal, A. Murray, J. Pearson, I. E. Jansen, D. Sondervan, H. Seelaar, D. Blake, K. Young, N. Halliwell, J. B. Callister, G. Toulson, A. Richardson, A. Gerhard, J. Snowden, D. Mann, D. Neary, M. A. Nalls, T. Peuralinna, L. Jansson, V.-M. Isoviita, A.-L. Kaivorinne, M. Hölttä-Vuori, E. Ikonen, R. Sulkava, M. Benatar, J. Wuu, A. Chiò, G. Restagno, G. Borghero, M. Sabatelli; ITALSGEN Consortium, D. Heckerman, E. Rogaeva, L. Zinman, J. D. Rothstein, M. Sendtner, C. Drepper, E. E. Eichler, C. Alkan, Z. Abdullaev, S. D. Pack, A. Dutra, E. Pak, J. Hardy, A. Singleton, N. M. Williams, P. Heutink, S. Pickering-Brown, H. R. Morris, P. J. Tienari, B. J. Traynor, A hexanucleotide repeat expansion in C9ORF72 is the cause of chromosome 9p21-linked ALS-FTD. Neuron 72 , 257–268 (2011).21944779
3 Q. Shao, C. Liang, Q. Chang, W. Zhang, M. Yang, J.-F. Chen, C9orf72 deficiency promotes motor deficits of a C9ALS/FTD mouse model in a dose-dependent manner. Acta Neuropathol. Commun. 7 , 32 (2019).30832726
4 Y.-B. Lee, H.-J. Chen, J. N. Peres, J. Gomez-Deza, J. Attig, M. Štalekar, C. Troakes, A. L. Nishimura, E. L. Scotter, C. Vance, Y. Adachi, V. Sardone, J. W. Miller, B. N. Smith, J.-M. Gallo, J. Ule, F. Hirth, B. Rogelj, C. Houart, C. E. Shaw, Hexanucleotide repeats in ALS/FTD form length-dependent RNA foci, sequester RNA binding proteins, and are neurotoxic. Cell Rep. 5 , 1178–1186 (2013).24290757
5 S. Mizielinska, S. Grönke, T. Niccoli, C. E. Ridler, E. L. Clayton, A. Devoy, T. Moens, F. E. Norona, I. O. C. Woollacott, J. Pietrzyk, K. Cleverley, A. J. Nicoll, S. Pickering-Brown, J. Dols, M. Cabecinha, O. Hendrich, P. Fratta, E. M. C. Fisher, L. Partridge, A. M. Isaacs, C9orf72 repeat expansions cause neurodegeneration in Drosophila through arginine-rich proteins. Science 345 , 1192–1194 (2014).25103406
6 K. Mori, S.-M. Weng, T. Arzberger, S. May, K. Rentzsch, E. Kremmer, B. Schmid, H. A. Kretzschmar, M. Cruts, C. Van Broeckhoven, C. Haass, D. Edbauer, The C9orf72 GGGGCC repeat is translated into aggregating dipeptide-repeat proteins in FTLD/ALS. Science 339 , 1335–1338 (2013).23393093
7 I. Kwon, S. Xiang, M. Kato, L. Wu, P. Theodoropoulos, T. Wang, J. Kim, J. Yun, Y. Xie, S. L. McKnight, Poly-dipeptides encoded by the C9orf72 repeats bind nucleoli, impede RNA biogenesis, and kill cells. Science 345 , 1139–1145 (2014).25081482
8 K. Kanekura, Y. Harada, M. Fujimoto, T. Yagi, Y. Hayamizu, K. Nagaoka, M. Kuroda, Characterization of membrane penetration and cytotoxicity of C9orf72-encoding arginine-rich dipeptides. Sci. Rep. 8 , 12740 (2018).30143685
9 Y.-J. Zhang, T. F. Gendron, M. T. W. Ebbert, A. D. O'Raw, M. Yue, K. Jansen-West, X. Zhang, M. Prudencio, J. Chew, C. N. Cook, L. M. Daughrity, J. Tong, Y. Song, S. R. Pickles, M. Castanedes-Casey, A. Kurti, R. Rademakers, B. Oskarsson, D. W. Dickson, W. Hu, A. D. Gitler, J. D. Fryer, L. Petrucelli, Poly(GR) impairs protein translation and stress granule dynamics in C9orf72-associated frontotemporal dementia and amyotrophic lateral sclerosis. Nat. Med. 24 , 1136–1142 (2018).29942091
10 H. Nanaura, H. Kawamukai, A. Fujiwara, T. Uehara, Y. Aiba, M. Nakanishi, T. Shiota, M. Hibino, P. Wiriyasermkul, S. Kikuchi, R. Nagata, M. Matsubayashi, Y. Shinkai, T. Niwa, T. Mannen, N. Morikawa, N. Iguchi, T. Kiriyama, K. Morishima, R. Inoue, M. Sugiyama, T. Oda, N. Kodera, S. Toma-Fukai, M. Sato, H. Taguchi, S. Nagamori, O. Shoji, K. Ishimori, H. Matsumura, K. Sugie, T. Saio, T. Yoshizawa, E. Mori, C9orf72-derived arginine-rich poly-dipeptides impede phase modifiers. Nat. Commun. 12 , 5301 (2021).34489423
11 S. Hutten, S. Usluer, B. Bourgeois, F. Simonetti, H. M. Odeh, C. M. Fare, M. Czuppa, M. Hruska-Plochan, M. Hofweber, M. Polymenidou, J. Shorter, D. Edbauer, T. Madl, D. Dormann, Nuclear import receptors directly bind to arginine-rich dipeptide repeat proteins and suppress their pathological interactions. Cell Rep. 33 , 108538 (2020).33357437
12 L. R. Hayes, L. Duan, K. Bowen, P. Kalab, J. D. Rothstein, C9orf72 arginine-rich dipeptide repeat proteins disrupt karyopherin-mediated nuclear import. eLife 9 , e51685 (2020).32119645
13 K. Zhang, C. J. Donnelly, A. R. Haeusler, J. C. Grima, J. B. Machamer, P. Steinwald, E. L. Daley, S. J. Miller, K. M. Cunningham, S. Vidensky, S. Gupta, M. A. Thomas, I. Hong, S. L. Chiu, R. L. Huganir, L. W. Ostrow, M. J. Matunis, J. Wang, R. Sattler, T. E. Lloyd, J. D. Rothstein, The C9orf72 repeat expansion disrupts nucleocytoplasmic transport. Nature 525 , 56–61 (2015).26308891
14 A. Iacoangeli, A. Al Khleifat, A. R. Jones, W. Sproviero, A. Shatunov, S. Opie-Martin, K. E. Morrison, P. J. Shaw, C. E. Shaw, I. Fogh, R. J. Dobson, S. J. Newhouse, A. Al-Chalabi, C9orf72 intermediate expansions of 24-30 repeats are associated with ALS. Acta Neuropathol. Commun. 7 , 115 (2019).31315673
15 W. B. Rippon, A. G. Walton, Optical properties of the polyglycine II helix. Biopolymers 10 , 1207–1212 (1971).5095092
16 H. Yamashita, T. Kato, M. Oba, T. Misawa, T. Hattori, N. Ohoka, M. Tanaka, M. Naito, M. Kurihara, Y. Demizu, Development of a cell-penetrating peptide that exhibits responsive changes in its secondary structure in the cellular environment. Sci. Rep. 6 , 33003 (2016).27609319
17 E. P. Rogakou, D. R. Pilch, A. H. Orr, V. S. Ivanova, W. M. Bonner, DNA double-stranded breaks induce histone H2AX phosphorylation on serine 139. J. Biol. Chem. 273 , 5858–5868 (1998).9488723
18 I. M. Ward, J. Chen, Histone H2AX is phosphorylated in an ATR-dependent manner in response to replicational stress. J. Biol. Chem. 276 , 47759–47762 (2001).11673449
19 K. Kanekura, T. Yagi, A. J. Cammack, J. Mahadevan, M. Kuroda, M. B. Harms, T. M. Miller, F. Urano, Poly-dipeptides encoded by the C9ORF72 repeats block global protein translation. Hum. Mol. Genet. 25 , 1803–1813 (2016).26931465
20 A. B. Loveland, E. Svidritskiy, D. Susorov, S. Lee, A. Park, S. Zvornicanin, G. Demo, F. B. Gao, A. A. Korostelev, Ribosome inhibition by C9ORF72-ALS/FTD-associated poly-PR and poly-GR proteins revealed by cryo-EM. Nat. Commun. 13 , 2776 (2022).35589706
21 B. D. Freibaum, Y. Lu, R. Lopez-Gonzalez, N. C. Kim, S. Almeida, K. H. Lee, N. Badders, M. Valentine, B. L. Miller, P. C. Wong, L. Petrucelli, H. J. Kim, F. B. Gao, J. P. Taylor, GGGGCC repeat expansion in C9orf72 compromises nucleocytoplasmic transport. Nature 525 , 129–133 (2015).26308899
22 A. Jovičić, J. Mertens, S. Boeynaems, E. Bogaert, N. Chai, S. B. Yamada, J. W. Paul III, S. Sun, J. R. Herdy, G. Bieri, N. J. Kramer, F. H. Gage, L. Van Den Bosch, W. Robberecht, A. D. Gitler, Modifiers of C9orf72 dipeptide repeat toxicity connect nucleocytoplasmic transport defects to FTD/ALS. Nat. Neurosci. 18 , 1226–1229 (2015).26308983
23 K. Y. Shi, E. Mori, Z. F. Nizami, Y. Lin, M. Kato, S. Xiang, L. C. Wu, M. Ding, Y. Yu, J. G. Gall, S. L. McKnight, Toxic PRn poly-dipeptides encoded by the C9orf72 repeat expansion block nuclear import and export. Proc. Natl. Acad. Sci. U.S.A. 114 , E1111–E1117 (2017).28069952
24 H. Jafarinia, E. Van der Giessen, P. R. Onck, Molecular basis of C9orf72 poly-PR interference with the β-karyopherin family of nuclear transport receptors. Sci. Rep. 12 , 21324 (2022).36494425
25 B. Leveugle, W. Ding, F. Laurence, M. P. Dehouck, A. Scanameo, R. Cecchelli, H. Fillit, Heparin oligosaccharides that pass the blood-brain barrier inhibit β-amyloid precursor protein secretion and heparin binding to β-amyloid peptide. J. Neurochem. 70 , 736–744 (1998).9453569
26 Q. Ma, B. Dudas, M. Hejna, U. Cornelli, J. M. Lee, S. Lorens, R. Mervis, I. Hanin, J. Fareed, The blood-brain barrier accessibility of a heparin-derived oligosaccharides C3. Thromb. Res. 105 , 447–453 (2002).12062547
27 Y.-P. Hu, Y.-Q. Zhong, Z.-G. Chen, C.-Y. Chen, Z. Shi, M. M. L. Zulueta, C.-C. Ku, P.-Y. Lee, C.-C. Wang, S.-C. Hung, Divergent synthesis of 48 heparan sulfate-based disaccharides and probing the specific sugar-fibroblast growth factor-1 interaction. J. Am. Chem. Soc. 134 , 20722–20727 (2012).23240683
28 S. Byrne, M. Elamin, P. Bede, A. Shatunov, C. Walsh, B. Corr, M. Heverin, N. Jordan, K. Kenna, C. Lynch, R. L. McLaughlin, P. M. Iyer, C. O'Brien, J. Phukan, B. Wynne, A. L. Bokde, D. G. Bradley, N. Pender, A. Al-Chalabi, O. Hardiman, Cognitive and clinical characteristics of patients with amyotrophic lateral sclerosis carrying a C9orf72 repeat expansion: A population-based cohort study. Lancet Neurol. 11 , 232–240 (2012).22305801
29 S. Millecamps, S. Boillée, I. Le Ber, D. Seilhean, E. Teyssou, M. Giraudeau, C. Moigneu, N. Vandenberghe, V. Danel-Brunaud, P. Corcia, P.-F. Pradat, N. Le Forestier, L. Lacomblez, G. Bruneteau, W. Camu, A. Brice, C. Cazeneuve, E. LeGuern, V. Meininger, F. Salachas, Phenotype difference between ALS patients with expanded repeats in C9ORF72 and patients with mutations in other ALS-related genes. J. Med. Genet. 49 , 258–263 (2012).22499346
30 C. Walker, S. Herranz-Martin, E. Karyka, C. Liao, K. Lewis, W. Elsayed, V. Lukashchuk, S.-C. Chiang, S. Ray, P. J. Mulcahy, M. Jurga, I. Tsagakis, T. Iannitti, J. Chandran, I. Coldicott, K. J. De Vos, M. K. Hassan, A. Higginbottom, P. J. Shaw, G. M. Hautbergue, M. Azzouz, S. F. El-Khamisy, C9orf72 expansion disrupts ATM-mediated chromosomal break repair. Nat. Neurosci. 20 , 1225–1235 (2017).28714954
31 M. D. Cykowski, D. W. Dickson, S. Z. Powell, A. S. Arumanayagam, A. L. Rivera, S. H. Appel, Dipeptide repeat (DPR) pathology in the skeletal muscle of ALS patients with C9ORF72 repeat expansion. Acta Neuropathol. 138 , 667–670 (2019).31375896
32 Y. Lin, E. Mori, M. Kato, S. Xiang, L. Wu, I. Kwon, S. L. McKnight, Toxic PR poly-dipeptides encoded by the C9orf72 repeat expansion target LC domain polymers. Cell 167 , 789–802.e12 (2016).27768897
33 K. H. Lee, P. Zhang, H. J. Kim, D. M. Mitrea, M. Sarkar, B. D. Freibaum, J. Cika, M. Coughlin, J. Messing, A. Molliex, B. A. Maxwell, N. C. Kim, J. Temirov, J. Moore, R. M. Kolaitis, T. I. Shaw, B. Bai, J. Peng, R. W. Kriwacki, J. P. Taylor, C9orf72 dipeptide repeats impair the assembly, dynamics, and function of membrane-less organelles. Cell 167 , 774–788.e17 (2016).27768896
34 R. Wang, X. Xu, Z. Hao, S. Zhang, D. Wu, H. Sun, C. Mu, H. Ren, G. Wang, Poly-PR in C9ORF72-related amyotrophic lateral sclerosis/frontotemporal dementia causes neurotoxicity by clathrin-dependent endocytosis. Neurosci. Bull. 35 , 889–900 (2019).31148094
35 U. S. Jeng, C. H. Su, C.-J. Su, K.-F. Liao, W.-T. Chuang, Y.-H. Lai, J.-W. Chang, Y.-J. Chen, Y.-S. Huang, M.-T. Lee, K.-L. Yu, J.-M. Lin, D.-G. Liu, C.-F. Chang, C.-Y. Liu, C.-H. Chang, K. S. Liang, A small/wide-angle X-ray scattering instrument for structural characterization of air–Liquid interfaces, thin films and bulk specimens. J. Appl. Cryst. 43 , 110–121 (2010).
36 D.-G. Liu, C.-H. Chang, L.-C. Chiang, M.-H. Lee, C.-F. Chang, C.-Y. Lin, C.-C. Liang, T.-H. Lee, S.-W. Lin, C.-Y. Liu, C.-S. Hwang, J.-C. Huang, C.-K. Kuan, H.-S. Wang, Y.-C. Liu, F.-H. Tseng, J.-Y. Chuang, W.-R. Liao, H.-C. Li, C.-J. Su, K.-F. Liao, Y.-Q. Yeh, O. Shih, W.-R. Wu, C.-A. Wang, U. Jeng, Optical design and performance of the biological small-angle X-ray scattering beamline at the Taiwan Photon Source. J. Synchrotron Radiat. 28 , 1954–1965 (2021).34738951
37 O. Shih, K. F. Liao, Y. Q. Yeh, C. J. Su, C. A. Wang, J. W. Chang, W. R. Wu, C. C. Liang, C. Y. Lin, T. H. Lee, C. H. Chang, L. C. Chiang, C. F. Chang, D. G. Liu, M. H. Lee, C. Y. Liu, T. W. Hsu, B. Mansel, M. C. Ho, C. Y. Shu, F. Lee, E. Yen, T. C. Lin, U. Jeng, Performance of the new biological small- and wide-angle X-ray scattering beamline 13A at the Taiwan Photon Source. J. Appl. Cryst. 55 , 340–352 (2022).35497659
38 O. Shih, Y.-Q. Yeh, K.-F. Liao, C.-J. Su, P.-H. Wu, R. K. Heenan, T.-Y. Yu, U. S. Jeng, Membrane charging and swelling upon calcium adsorption as revealed by phospholipid nanodiscs. J. Phys. Chem. Lett. 9 , 4287–4293 (2018).29999328
39 A. J. A. P. Guinier, La diffraction des rayons X aux très petits angles: Application à l'étude de phénomènes ultramicroscopiques. Ann. Phys. 11 , 161–237 (1939).
40 Y. Zhang, I-TASSER server for protein 3D structure prediction. BMC Bioinformatics 9 , 40 (2008).18215316
41 A. Roy, A. Kucukural, Y. Zhang, I-TASSER: A unified platform for automated protein structure and function prediction. Nat. Protoc. 5 , 725–738 (2010).20360767
42 J. Yang, Y. Zhang, I-TASSER server: New development for protein structure and function predictions. Nucleic Acids Res. 43 , W174–W181 (2015).25883148
43 J. Yang, R. Yan, A. Roy, D. Xu, J. Poisson, Y. Zhang, The I-TASSER Suite: Protein structure and function prediction. Nat. Methods 12 , 7–8 (2015).
44 S. Wu, J. Skolnick, Y. Zhang, Ab initio modeling of small proteins by iterative TASSER simulations. BMC Biol. 5 , 17 (2007).17488521
45 D. Svergun, C. Barberato, M. H. J. Koch, CRYSOL– A program to evaluate X-ray solution scattering of biological macromolecules from atomic coordinates. J. Appl. Cryst. 28 , 768–773 (1995).
46 R. Das, D. Baker, Macromolecular modeling with Rosetta. Annu. Rev. Biochem. 77 , 363–382 (2008).18410248
47 K. W. Kaufmann, G. H. Lemmon, S. L. DeLuca, J. H. Sheehan, J. Meiler, Practically useful: What the rosetta protein modeling suite can do for you. Biochemistry 49 , 2987–2998 (2010).20235548
48 B. J. Bender, A. Cisneros III, A. M. Duran, J. A. Finn, D. Fu, A. D. Lokits, B. K. Mueller, A. K. Sangha, M. F. Sauer, A. M. Sevy, G. Sliwoski, J. H. Sheehan, F. DiMaio, J. Meiler, R. Moretti, Protocols for molecular modeling with Rosetta3 and RosettaScripts. Biochemistry 55 , 4748–4763 (2016).27490953
49 K. Stovgaard, C. Andreetta, J. Ferkinghoff-Borg, T. Hamelryck, Calculation of accurate small angle X-ray scattering curves from coarse-grained protein models. BMC Bioinformatics 11 , 429–429 (2010).20718956
50 B. Samatanga, A. Cléry, P. Barraud, F. H.-T. Allain, I. Jelesarov, Comparative analyses of the thermodynamic RNA binding signatures of different types of RNA recognition motifs. Nucleic Acids Res. 45 , 6037–6050 (2017).28334819
51 A. Loregian, E. Sinigalia, B. Mercorelli, G. Palù, D. M. Coen, Binding parameters and thermodynamics of the interaction of the human cytomegalovirus DNA polymerase accessory protein, UL44, with DNA: Implications for the processivity mechanism. Nucleic Acids Res. 35 , 4779–4791 (2007).17617644
52 Z.-W. Du, H. Chen, H. Liu, J. Lu, K. Qian, C.-L. Huang, X. Zhong, F. Fan, S.-C. Zhang, Generation and expansion of highly pure motor neuron progenitors from human pluripotent stem cells. Nat. Commun. 6 , 6626 (2015).25806427
