RNA-binding ability of FUS regulates neurodegeneration, cytoplasmic mislocalization and incorporation into stress granules associated with FUS carrying ALS-linked mutations

Abstract

Amyotrophic lateral sclerosis (ALS) is an uncommon neurodegenerative disease caused by degeneration of upper and lower motor neurons. Several genes, including SOD1, TDP-43, FUS, Ubiquilin 2, C9orf72 and Profilin 1, have been linked with the sporadic and familiar forms of ALS. FUS is a DNA/RNA-binding protein (RBP) that forms cytoplasmic inclusions in ALS and frontotemporal lobular degeneration (FTLD) patients' brains and spinal cords. However, it is unknown whether the RNA-binding ability of FUS is required for causing ALS pathogenesis. Here, we exploited a Drosophila model of ALS and neuronal cell lines to elucidate the role of the RNA-binding ability of FUS in regulating FUS-mediated toxicity, cytoplasmic mislocalization and incorporation into stress granules (SGs). To determine the role of the RNA-binding ability of FUS in ALS, we mutated FUS RNA-binding sites (F305L, F341L, F359L, F368L) and generated RNA-binding-incompetent FUS mutants with and without ALS-causing mutations (R518K or R521C). We found that mutating the aforementioned four phenylalanine (F) amino acids to leucines (L) (4F-L) eliminates FUS RNA binding. We observed that these RNA-binding mutations block neurodegenerative phenotypes seen in the fly brains, eyes and motor neurons compared with the expression of RNA-binding-competent FUS carrying ALS-causing mutations. Interestingly, RNA-binding-deficient FUS strongly localized to the nucleus of Drosophila motor neurons and mammalian neuronal cells, whereas FUS carrying ALS-linked mutations was distributed to the nucleus and cytoplasm. Importantly, we determined that incorporation of mutant FUS into the SG compartment is dependent on the RNA-binding ability of FUS. In summary, we demonstrate that the RNA-binding ability of FUS is essential for the neurodegenerative phenotype in vivo of mutant FUS (either through direct contact with RNA or through interactions with other RBPs).

Figure 1.

Mutating conserved RNA-binding sites 4F-L (F305L, F341L, F359L, F368L) of FUS blocks its RNA-binding ability and does not cause any obvious toxicity in yeast and Drosophila. (A) Pull-down of HA-tagged FUS constructs UV-crosslinked to RNA from neuroblastoma (N2a) cell lines stably expressing FUS WT, FUS 4F-L, FUS R518K or FUS 4F-L R518K (top panel). Subsequent to the pull-down, the RNA was radiolabeled and the complex was displayed by SDS–PAGE. An RNA smear indicating RNA binding is seen in the FUS-WT and FUS R518K lanes but is highly diminished in the RNA-binding-incompetent (FUS 4F-L and FUS 4F-L R518K) samples. Controls (first three lanes) consisted of the control of non-transfected cells (lane 1), cells transfected with FUS WT but not UV-crosslinked prior to the pull-down (lane 2) and a pull-down of UV-crosslinked FUS WT with a non-specific antibody (lane 3), all show no binding to RNA. (Middle panel) Autoradiograph probed with anti-HA antibody shows that HA-FUS protein was pulled down in all expected lanes, demonstrating that FUS 4F-L protein levels are still high despite the elimination of bound RNA. (Bottom panel) Whole-cell lysate of input shows that HA-FUS protein is present in all samples except for the non-transfected cells (first lane). Loading of whole-cell lysate input samples was done in the same order as the top two panels, but without empty wells between the samples. (B) Over-expression of RNA-binding-incompetent FUS alone (FUS 4F-L) or with an ALS-causing mutation (FUS 4F-L R518K) results in significantly less toxicity in yeast compared with FUS WT and FUS R518K. All proteins were expressed using the galactose-inducible GAL1 promoter. (C) Western blotting indicates that FUS WT, FUS R518K and RNA-binding-incompetent FUS accumulate to similar levels in yeast following galactose-induced over-expression. (D) Ectopic expression of RNA-binding-incompetent FUS (4F-L constructs) does not cause any external eye degeneration in Drosophila (lower panel) when compared with FUS WT, FUS R518K and FUS R521C (upper panel). (E) Western blot of two independent transgenic fly lines expressing FUS WT, FUS 4F-L, FUS R521C, FUS 4F-L R521C, FUS R518K and FUS 4F-L R518K showed equal FUS protein levels in the fly heads. (F) We performed quantification of eye phenotypes using previously published criteria and found that normally toxic FUS mutations (R518K and R521C) were alleviated of their toxicity when those FUS proteins were rendered RNA-binding-incompetent (4F-L). The asterisks represent significant difference between different groups. ***P, 0.0001.

Figure 2.

Ectopic expression of RNA-binding-incompetent FUS does not cause any behavioral abnormalities in Drosophila. (A) Eclosion assay: Motor neuron expression of either UAS-FUS R518K or UAS-FUS R521C, but not UAS-FUS WT, led to pupal lethality in Drosophila. The animals expressing RNA-binding-incompetent FUS (single or double mutants) in the motor neurons eclosed normally when compared with driver-alone control and WT FUS (error bars represent standard error). (B) Larval crawling assay: Motor neuron expression of mutant FUS R521C and R518K, but not WT, in the fly motor neurons led to larval paralysis when compared with driver alone. RNA-binding-incompetent FUS (4F-L)-expressing animals showed no defect in their crawling ability when compared with controls. (C) Body-wall contraction: Expression of mutant FUS led to reduced body peristaltic contractions when compared with FUS WT and driver-alone control. However, RNA-binding-incompetent FUS-expressing animals showed body peristaltic contractions similar to FUS WT and driver-alone controls. (D) Larval righting assay: Motor neuron expression of mutant FUS compromised the turning ability of third instar larvae when compared with FUS WT and driver-alone control. Note that FUS R521C larvae lost all their ability to right themselves. Interestingly, RNA-binding-incompetent FUS-expressing animals demonstrated normal larval turning ability similar to FUS WT and driver-alone controls. The asterisks represent significant difference between groups *P < 0.05, **P < 0.001, ***P < 0.0001 and NS (not significant).

Figure 3.

Ectopic expression of mutant FUS, but not RNA-binding-incompetent FUS, leads to brain atrophy with ubiquitin-positive pathology in vivo. (A) Ectopic expression of FUS R518K and FUS R521C in fly motor neurons leads to brain atrophy and reduced brain size. However, the expression of RNA-binding-incompetent FUS alone or with ALS-causing mutations does not cause any gross defect in the brain size. (B) Quantification of larval brain size.

Figure 4.

FUS carrying the ALS-causing mutation R518K alone distributes itself to the nucleus and cytoplasm in mammalian neuronal cells but RNA-binding-incompetent FUS (4F-L alone and 4F-L R518K) localizes in the nucleus similar to WT FUS. (A) FUS WT (anti-HA: green) predominantly localizes to the nucleus (DRAQ5: blue), (B) whereas FUS R518K is distributed to the nucleus and cytoplasm (D). Interestingly, RNA-binding-incompetent FUS 4F-L (C) itself or with an ALS-associated mutation FUS 4F-L R518K (E) localized exclusively to the nucleus in the neuronal (N2a) cells. (F) Quantification of cytoplasmic and nuclear distribution of FUS. The asterisks represent significant difference between different groups. ***P, 0.0001.

Figure 5.

Cytoplasmic mislocalization of FUS carrying ALS-causing mutations (R518K and R521C) in Drosophila motor neurons is reversed by making FUS RNA-binding-incompetent. (A) FUS WT (anti-HA: green) predominantly localized in the nucleus (anti-Lamin: red) of the Drosophila motor neurons (B), whereas FUS R518K or FUS R521C distributed itself to both the nucleus and cytoplasm (D and F). Interestingly, RNA-binding-incompetent FUS itself (C) or with ALS-associated mutations were predominantly nuclear (E and G).

Figure 6.

Incorporation of mutant FUS into SGs is dependent on its RNA-binding ability. Each HA-FUS stable N2a cell line was treated with 0.5 mm sodium arsenite for 1 h, stained with the nuclear dye DRAQ5 (blue), anti-HA (green) and anti-TIAR (red). (A) SGs were induced in untransfected N2a cells. (B) Anti-HA staining reveals HA-FUS is primarily nuclear and does not incorporate into SGs. (C) RNA-binding-incompetent mutations do not cause HA-FUS incorporation into SGs. (D) HA-FUS R518K colocalizes with the SG marker TIAR. (E) RNA-binding-incompetent FUS carrying the ALS-causing mutation FUS R518K was excluded from TIAR SGs.