The RNA helicase DHX36-G4R1 modulates C9orf72 GGGGCC hexanucleotide repeat-associated translation

Abstract

GGGGCC (G₄C₂) hexanucleotide repeat expansions in the endosomal trafficking gene C9orf72 are the most common genetic cause of ALS and frontotemporal dementia. Repeat-associated non-AUG (RAN) translation of this expansion through near-cognate initiation codon usage and internal ribosomal entry generates toxic proteins that accumulate in patients' brains and contribute to disease pathogenesis. The helicase protein DEAH-box helicase 36 (DHX36-G4R1) plays active roles in RNA and DNA G-quadruplex (G4) resolution in cells. As G₄C₂ repeats are known to form G4 structures in vitro, we sought to determine the impact of manipulating DHX36 expression on repeat transcription and RAN translation. Using a series of luciferase reporter assays both in cells and in vitro, we found that DHX36 depletion suppresses RAN translation in a repeat length-dependent manner, whereas overexpression of DHX36 enhances RAN translation from G₄C₂ reporter RNAs. Moreover, upregulation of RAN translation that is typically triggered by integrated stress response activation is prevented by loss of DHX36. These results suggest that DHX36 is active in regulating G₄C₂ repeat translation, providing potential implications for therapeutic development in nucleotide repeat expansion disorders.

Keywords: ALS (Lou Gehrig's disease); C9orf72; DHX36–G4R1–RHAU; DNA helicase; G-quadruplex; RNA helicase; dipeptide repeat proteins; fragile X; repeat-associated non-AUG translation.

Figure 1

DHX36 binds and enhances transcription of C9-DNA in vitro.A, representative EMSA image. C9-repeat DNA oligonucleotides were heated and cooled in the presence (lanes 1–5) and absence (lanes 6 and 7) of KCl to induce or prevent G4 formation, respectively. DNA was incubated with increasing concentrations of recombinant DHX36, analyzed with nondenaturing PAGE, and imaged. B, representative EMSA image. Scrambled control DNA oligonucleotides were heated and cooled in the presence (lanes 1–5) and absence (lanes 6 and 7) of KCl. DNA was incubated with increasing concentrations of recombinant DHX36, analyzed with nondenaturing PAGE, and imaged. C, densiometric quantification of panels A and B. The percent bound for each lane was graphed versus the concentration of DHX36. Data are presented as mean ± SD, n = 3. Multiple t tests for each concentration of protein, ∗p ≤ 0.05, ∗∗p < 0.01, and ∗∗∗p < 0.001. D, T7 polymerase transcript products generated from equal amounts of linearized NLuc (lanes 1 and 2) or (G₄C₂)₇₀ plasmids (lanes 3 and 4) resolved by denaturing PAGE. The gel was stained with SYBR gold nucleic acid stain and imaged. E, densiometric quantification of panel D. All signals were first subtracted by the background. Then signal from the top third of the gel was divided by the total signal per lane. Data are presented as mean ± SD, n = 5 for (NLuc)-6(G₄C₂)₇₀. Two-tailed paired t test, ∗p < 0.05. DHX36, DEAH-box helicase 36; KCl, potassium chloride; NLuc, nanoluciferase; NS, not significant.

Figure 2

The effect of DHX36 KD on C9-RNA and C9RAN reporter expression.A, schematic of AUG-FF, AUG-NLuc Ctrl, C9-RAN, and CGG-RAN luciferase reporters. B, experimental timeline for doxycycline treatment and reporter transfection. C, immunoblots detecting DHX36 in parental, Ctrl, and DHX36 KD HeLa cells with and without doxycycline treatment. D, relative expression of AUG and C9-RAN translation in GA (+0), GP (+1), and GR (+2) frames with 70 repeats between Ctrl and DHX36 KD HeLa cells. NLuc signals were normalized to AUG-FFluc translation. E, immunoblot of RAN translation products from 70 repeats of G₄C₂ in GA frame and 100 repeats of CGG in +1 reading frame in Ctrl and DHX36 KD HeLa cells. GFP was blotted as transfection control, and GAPDH was blotted as loading Ctrl. F, expression of +1CGG₁₀₀ RAN translation reporters measured by luciferase assay. NLuc signals were normalized to AUG-FFluc signals to compare between Ctrl and DHX36 KD HeLa cells. G, abundance of NLuc mRNA from AUG and GA₇₀ in DHX36 Ctrl and KD HeLa cells. NLuc mRNAs were normalized to FF mRNA and compared with DHX36 Ctrl. Data in (D) and (F) are represented as mean ± SD, n = 9 to 12. Data in (G) are mean ± SD, n = 3. Two-tailed Student's t test with Bonferroni and Welch's correction, ∗p < 0.05; ∗∗p < 0.01; ∗∗∗p < 0.001; and ∗∗∗∗p < 0.0001. AUG-FF, AUG-initiated firefly luciferase; Ctrl, control; DHX36, DEAH-box helicase 36; KD, knockdown; NLuc, nanoluciferase; NS, not significant; RAN, repeat-associated non-AUG.

Figure 3

C9RAN reporter expression in DHX36 KO Jurkat cell lines and in vitro cell lysates.A, immunoblots to DHX36 from WT and DHX36 KO Jurkat cells. B, relative expression of AUG and C9-RAN translation from GA (+0), GP (+1), and GR (+2) reading frames in WT and DHX36 KO Jurkat cells. NLuc signal were normalized to AUG-FF and compared between WT and DHX36 KO Jurkat cells. Data are represented as mean ± SD, n = 9. C, in vitro translation using lysates derived from DHX36 WT and KO Jurkat cells for AUG-NLuc RNA and C9-RAN in GA frame RNA. NLuc signals were normalized to signal from DHX36 WT. Data are represented as mean ± SD, n = 24. Two-tailed Student's t test with Bonferroni and Welch's correction, ∗p < 0.05; ∗∗∗p < 0.001; and ∗∗∗∗p < 0.0001. AUG-FF, AUG-initiated firefly luciferase; DHX36, DEAH-box helicase 36; NLuc, nanoluciferase; RAN, repeat-associated non-AUG.

Figure 4

The effect of decreased DHX36 on C9RAN reporter expression is G₄C₂repeat length dependent.A, schematic of previously published luciferase reporters of C9-RAN in GA frame harboring different repeat sizes. B and C, relative expression of AUG and C9-RAN translation from GA frames with 3, 35, and 70 repeats in Ctrl and DHX36 KD HeLa cells (B) and DHX36 WT and KO Jurkat cells. NLuc signals were normalized to AUG-FF. Data are represented as mean ± SD, n = 9 to 12. One-way ANOVAs were performed to compare the statistical differences between repeat length in DHX36 KD or KO cell lines. Two-tailed Student's t test with Bonferroni and Welch's correction were then performed to confirm the differences between multiple comparison, ∗p < 0.05; ∗∗p < 0.01; and ∗∗∗∗p < 0.0001. AUG-FF, AUG-initiated firefly luciferase; Ctrl, control; DHX36, DEAH-box helicase 36; G₄C_2, GGGGCC; KD, knockdown; NLuc, nanoluciferase; RAN, repeat-associated non-AUG.

Figure 5

DHX36 overexpression enhances C9RAN reporter expression from G₄C₂repeat RNA.A, relative expression of AUG and C9-RAN translation when cotransfecting reporter RNA and overexpression of empty vector, WT, or E335A DHX36 DNA plasmids in HeLa cells. Data are represented as mean ± SD, n = 9. Two-tailed Student's t test with Bonferroni and Welch's correction, ∗p < 0.05; ∗∗∗p < 0.001; and ∗∗∗∗p < 0.0001. B, Western blot analysis of cotransfected AUG and C9-RAN luciferase reporters in RNA and empty vector, DHX36 WT, or DHX36 E335A DNA plasmids in HeLa cells. GAPDH was blotted as internal control. DHX36, DEAH-box helicase 36; G₄C_2, GGGGCC; RAN, repeat-associated non-AUG.

Figure 6

KD of DHX36 prevents stress-dependent upregulation of C9RAN reporter expression.A, relative expression of RAN translation in G₄C₂ and CGG repeat treated with 2 μM thapsigargin (Tg) or DMSO in Ctrl and DHX36 KD HeLa cells. NLuc (left) and FF (right) signals were represented as ratio of Tg-treated cells to DMSO-treated cells and compared between Ctrl and DHX36 KD HeLa cells. B, immunoblots of G₄C₂ and CGG RAN luciferase reporters in Ctrl and DHX36 KD HeLa cells treated with 2 μM Tg. GFP was blotted as a transfection Ctrl, and GAPDH was blotted as internal Ctrl. For panel A, data are represented as mean ± SD, n = 6. Two-way ANOVA was performed to discern effect of Tg treatment across cell types. Two-tailed Student's t test with Bonferroni and Welch's correction was performed to assess differences between individual groups. ∗p < 0.05; ∗∗p < 0.01; and ∗∗∗∗p < 0.0001. Ctrl, control; DHX36, DEAH-box helicase 36; DMSO, dimethyl sulfoxide; G₄C₂, GGGGCC; KD, knockdown; NLuc, nanoluciferase; RAN, repeat-associated non-AUG.

Figure 7

Model of DHX36 modulation of C9RAN translation. DHX36 binds to G-quadruplex DNA and RNA structures. DHX36 aids transcription through large stretches of G₄C₂-repeat RNA in vitro, but its effects in human cells with large repeats are unclear. Depletion of DHX36 decreases RAN translation from both CGG- and C9-repeat reporters—both of which are capable of forming G-quadruplex structures–whereas increased DHX36 expression enhances C9RAN translation. These findings suggest a direct role for DHX36 in RAN translation of GC-rich repeats. DHX36, DEAH-box helicase 36; G₄C_2, GGGGCC; RAN, repeat-associated non-AUG.