Lysine acetylation regulates the RNA binding, subcellular localization and inclusion formation of FUS

Abstract

Amyotrophic lateral sclerosis (ALS) is a neurodegenerative disease characterized by the preferential death of motor neurons. Approximately 10% of ALS cases are familial and 90% are sporadic. Fused in sarcoma (FUS) is a ubiquitously expressed RNA-binding protein implicated in familial ALS and frontotemporal dementia (FTD). The physiological function and pathological mechanism of FUS are not well understood, particularly whether post-translational modifications play a role in regulating FUS function. In this study, we discovered that FUS was acetylated at lysine-315/316 (K315/K316) and lysine-510 (K510) residues in two distinct domains. Located in the nuclear localization sequence, K510 acetylation disrupted the interaction between FUS and Transportin-1, resulting in the mislocalization of FUS in the cytoplasm and formation of stress granule-like inclusions. Located in the RNA recognition motif, K315/K316 acetylation reduced RNA binding to FUS and decreased the formation of cytoplasmic inclusions. Treatment with deacetylase inhibitors also significantly reduced the inclusion formation in cells expressing ALS mutation P525L. More interestingly, familial ALS patient fibroblasts showed higher levels of FUS K510 acetylation as compared with healthy controls. Lastly, CREB-binding protein/p300 acetylated FUS, whereas both sirtuins and histone deacetylases families of lysine deacetylases contributed to FUS deacetylation. These findings demonstrate that FUS acetylation regulates the RNA binding, subcellular localization and inclusion formation of FUS, implicating a potential role of acetylation in the pathophysiological process leading to FUS-mediated ALS/FTD.

Figure 1

FUS is acetylated. (A) Endogenous FUS-IP from N2A cells treated with DACi cocktail (nicotinamide (30 mM), sodium butyrate (50 mM) and TSA (3 μM)) for 6 h. Immunoblotting was performed using the indicated antibodies. (B) Mass spectrometric identification of the acetylated FUS peptide RGGRGGGDRGGFGPGK⁵¹⁰MDSRGEHRQDRRERPY. (C) 3× FLAG-tagged WT, K315Q/K316Q, K510Q, K315Q/K316Q/K510Q or FLAG vector control were transfected into HEK293T cells. After 24 h, cells were treated with DACi cocktail for 6 h, followed by FLAG-IP and immunoblotting with the indicated antibodies. (D) Quantification of FUS acetylation from three independent experiments ±standard deviation (SD). Student’s t-test was performed for individual comparisons against WT (*P ≤ 0.05). (E) The domain structure of FUS showing the RRM domain sequence and acetylation sites. (F) NMR solution structure of FUS RRM domain (pink) showing K315 and K316 in the KK-loop bound to stem–loop RNA (protein data bank entry 6GBM). (G) Crystal structure of TNPO1/FUS-NLS (protein data bank entry 4FQ3) illustrating the FUS-NLS domain (pink) K510 adjacent to TNPO1 (purple) D693. Molecular graphics of FUS RRM and NLS domains visualized using UCSF ChimeraX (79).

Figure 2

Acetylation of FUS impairs RNA binding and the FUS-NLS-TNPO1 interaction. (A) 3× FLAG-tagged FUS constructs were transfected into N2A cells. After 48 h, cells were lysed and FLAG-IP followed by reverse transcription and qPCR against FUS pre-mRNA surrounding exon 7 was performed. FUS mRNA was normalized with the FLAG levels in the FLAG-FUS-IPs and with the Rpl13a mRNA levels in the total extracts. Averages of three independent experiments are shown, ±SD. Student’s t-test was performed for individual comparisons against WT. (B) The indicated 3× FLAG-tagged FUS constructs were transfected into N2A cells and treated with DACi cocktail where indicated. FLAG-IP was performed 48 h after transfection with the inclusion of RNase cocktail as indicated, followed by immunoblotting with the indicated antibodies. (C) Quantification of (B) from three independent experiments. Student’s t-test was performed for individual comparisons against WT. (D) In vitro TNPO1 pulldown with Sulfolink-immobilized acetylated or non-acetylated FUS-NLS peptides. Different amounts of GST–TNPO1 were incubated with FUS-NLS peptide immobilized on the beads at 4°C for 3 h. The amount of TNPO1 pulled down with FUS-NLS peptides were evaluated by western blot. (E) Quantification of (D) from three independent experiments, ±SD. Student’s t-test was performed comparing the band intensities of non-acetylated and acetylated peptide pulldowns at the same concentration (*P ≤ 0.05; **P ≤ 0.005; ***P ≤ 0.001; NS, not significant).

Figure 3

The effect of FUS acetylation on cellular localization and stress granule formation. (A) N2A cells were transfected with EGFP-tagged WT, K315Q/K316Q, K510Q and K315Q/K316Q/K510Q FUS. The nuclei were visualized with DAPI. Samples were examined by confocal microscopy. Scale bars, 10 μm. (B) Quantification of nuclear and cytoplasmic EGFP-tagged FUS intensity ± SD (n > 200 cells) using ImageJ scripts. One-way ANOVA was performed to determine statistical significance (*P ≤ 0.05; **P ≤ 0.005; ***P ≤ 0.001). (C) The percentage of EGFP-positive cells with cytoplasmic inclusions (n > 100 cells; *P ≤ 0.001). (D) N2A cells were transfected with EGFP-tagged WT, P525L and P525L/K315Q/K316Q FUS. The nuclei were visualized with DAPI. Samples were examined by confocal microscopy. Scale bars, 10 μm. (E) The percentage of EGFP-positive cells with cytoplasmic inclusions (n > 100 cells; *P ≤ 0.001). (F) N2A cells were transfected with EGFP-tagged WT or P525L FUS. One set of P525L-transfected cells were treated with DACi cocktail (30 mM nicotinamide, 50 mM sodium butyrate and 3 μM TSA). The nuclei were visualized with DAPI. Samples were examined by confocal microscopy; scale bars, 10 μm. (G) The percentage of EGFP-positive cells with cytoplasmic inclusions (n > 100 cells; *P ≤ 0.001).

Figure 4

K510 acetylation is increased in familiar FUS ALS. (A) FUS-K510 acetylation levels in ALS patients with R521G or P525R FUS mutations versus control subjects. Immunoblotting was performed using the indicated antibodies. (B) Quantification of FUS-K510 acetylation normalized against total FUS levels. Error bars represent SD between individuals. One-way ANOVA was performed to determine statistical significance (*P ≤ 0.05).

Figure 5

The regulators of FUS acetylation. (A) 3× FLAG-FUS or 3× FLAG-vector was co-transfected with HA-CBP or HA-vector into N2A cells. FLAG-IP was performed, followed by immunoblotting using a pan-acetylated lysine antibody and other indicated antibodies. (B) N2A cells were treated with different concentrations of the CBP/p300 inhibitor A-485. Immunoblotting was performed using the FUS K510 acetylation antibody and other antibodies as indicated. (C) Quantification of (B) from three independent experiments ±SD. Student’s t-test was performed for individual comparisons against no treatment (*P ≤ 0.05). (D) N2A cells were transfected with 3× FLAG-vector or 3× FLAG-FUS and treated with 8 μM CBP/p300 inhibitor A-485 for 16 h in the presence of DACi cocktail. FLAG-IP was performed followed by immunoblotting with a pan-acetylated lysine antibody and a FLAG antibody. The cell lysate was examined using the FUS K510 acetylation antibody and other indicated antibodies. (E). N2A cells were treated with 30 mM nicotinamide and/or 3 μM TSA for 6 h. FLAG-IP was performed followed by immunoblotting with a pan-acetylated lysine antibody and a FLAG antibody. (F) Quantification of (E) from three independent experiments, ±SD. One-way ANOVA was performed to determine statistical significance (*P ≤ 0.05; **P ≤ 0.005). (G). N2A cells were treated with 30 mM nicotinamide and/or 3 μM TSA for 6 h. Cells were harvested and lysed after treatment and immunoblotting was performed using the FUS K510 acetylation antibody and other indicated antibodies. (H) Quantification of (G) from three independent experiments ±SD. One-way ANOVA was performed to determine statistical significance (*P ≤ 0.05; **P ≤ 0.005; ***P ≤ 0.001). (I) HEK293T cells were transfected with 3× FLAG-empty vector, 3× FLAG-SIRTs 6–7, 3× FLAG-HDAC3 and 3× FLAG-ROA1 (hnRNPA1) used as a positive control. After 48 h, endogenous FUS-IP was performed, followed by immunoblotting using the indicated antibodies.

Figure 6

Proposed model of the role of FUS acetylation in the modulation of FUS subcellular localization and inclusion formation.