RNA buffers the phase separation behavior of prion-like RNA binding proteins

Abstract

Prion-like RNA binding proteins (RBPs) such as TDP43 and FUS are largely soluble in the nucleus but form solid pathological aggregates when mislocalized to the cytoplasm. What keeps these proteins soluble in the nucleus and promotes aggregation in the cytoplasm is still unknown. We report here that RNA critically regulates the phase behavior of prion-like RBPs. Low RNA/protein ratios promote phase separation into liquid droplets, whereas high ratios prevent droplet formation in vitro. Reduction of nuclear RNA levels or genetic ablation of RNA binding causes excessive phase separation and the formation of cytotoxic solid-like assemblies in cells. We propose that the nucleus is a buffered system in which high RNA concentrations keep RBPs soluble. Changes in RNA levels or RNA binding abilities of RBPs cause aberrant phase transitions.

Fig. 1. Prion-like RBPs phase-separate at their physiological concentrations.

(A) Domain structure. PLD, prion-like domain; RRM, RNA recognition motif; RGG, arginine- and glycine-rich region; ZF, zinc finger; NLS, nuclear localization sequence. (B) Representative images of immunostained HeLa cells. Dashed lines indicate the nuclear boundary. Scale bar, 5 μm. (C) Quantification of the nuclear enrichment of RBPs. Error bars represent SD. (D) Calculated cellular and nuclear (Nuc) concentrations of RBPs in HeLa cells. (E) Left, live HeLa cell nucleus expressing GFP-tagged FUS from a bacterial artificial chromosome (BAC). Arrows point to paraspeckles. Right, FUS-GFP phase-separated in vitro at 7.5 μM. Scale bars, 2 μm. (F) Quantification of the fractions of FUS present in condensed and soluble states in vivo and in vitro. Error bars represent SD. (G) Top, HeLa cell nuclei expressing GFP-tagged RBPs from BACs. White arrows indicate condensates. Bottom, purified RBPs phase-separate at their respective nuclear concentrations. Scale bars, 2 μm.

Fig. 2. RNA regulates the phase behavior of prion-like RBPs.

(A) Representative images of purified FUS-GFP (5 μM) in vitro in the presence of total RNA. (B) Quantification of the fraction of condensed FUS-GFP. Cin, fraction of total protein in the droplets; Cout, fraction of total protein in the soluble phase outside the droplets. The value of FUS enrichment in the droplet phase in the absence of RNA was normalized to 1. Error bars represent SD. (C) In vitro phase separation assay with EWSR1, TAF15, hnRNPA1, or TDP43 in the presence of total RNA. (D) Addition of RNase A to a sample of FUS-GFP (5 μM) solubilized with 300 ng/ml of total RNA. (E) Left, FUS-GFP (5 μM) solubilized with 800 ng/ml tRNA in vitro. Right, FUS phase separation triggered by the addition of 100 ng/μl Neat1 RNA in the presence of tRNA (800 ng/μl). Scale bars in (A), (C), (D), and (E), 2 μm.

Fig. 3. RNA keeps prion-like RBPs in a soluble state in the nucleus.

(A) Montage of a HeLa cell expressing FUS-GFP after microinjection with RNase A. Scale bar, 2 μm. (B) HeLa cells expressing FUS-GFP. White lines indicate cell outlines, orange lines indicate nuclear outlines, and boxes indicate regions of FCS measurements. Scale bar, 5 μm. (C) Autocorrelation curves obtained from FCS of FUS-GFP. τ is the autocorrelation time and Gn (τ) is the autocorrelation function, which is normalized to the amplitude of 1 at τ = 10 ms. (D) Quantification of the amount of slow FUS (methods are described in supplementary materials). Error bars represent SD. **P < 0.01. (E) Autocorrelation curves obtained from FCS of FUS-GFP variants in the nucleus. wt, wild type; mutRGG, mutations in the first RGG; mutRRM/ZnF, mutations in the RRM and zinc finger (ZnF); PLD, lacks all the RNA binding domains. (F) Quantification of slow FUS in the nucleus, obtained from two-component fits of the curves in (E) (methods are described in supplementary materials). Error bars represent SD. tD2, decay time for slow FUS. (G) HeLa cells showing variable FUS-GFP expression. Scale bar, 5 μm. (H) HeLa cells expressing different FUS-GFP variants with mutations in RNA binding domains (fig. S16). Scale bar, 5 μm. (I) Number of nuclear FUS-GFP assemblies per cell (n > 30) as a function of mean protein intensity (AI). Shading represents the confidence interval of the fitted linear regression model, which is plotted as a solid line. (J) Number of cells with more than 100 nuclear assemblies. n > 100 cells. Error bars represent SD. In (F) and (J), *P < 0.05 and ***P < 0.01 in comparison with the wild type.

Fig. 4. RNA regulates aberrant liquid-to-solid phase transitions of prion-like RBPs.

(A) In vitro phase-separated Gly156→Glu (G156E) variant FUS–GFP in the absence or presence of total RNA after 24 hours. Scale bar, 2 μm. (B) FUS-GFP–expressing HeLa cell nucleus after RNase A microinjection. Scale bar, 1 μm. (C) Montage of FUS-GFP droplets formed after RNase A microinjection. The droplets fuse in the first 5 min (blue box) but dissociate after 30 min, resulting in “sticky droplets” (red box). (D) Montage of FUS-GFP droplets formed in vitro (7 μM). The fusion of freshly formed droplets is compared with 3-hour-old droplets. Scale bar, 5 μm. (E) Fluorescence recovery after photobleaching (FRAP) of nuclear FUS-GFP assemblies less than 5 min (blue box) or more than 30 min (red box) after RNase A microinjection. (F) FRAP of nuclear FUS-GFP assemblies in HeLa cells after RNA degradation as shown in (B) and (E) (n > 10 cells). (G to I) FRAP of nuclear assemblies in HeLa cells expressing full-length FUS [(G) and (I)] or FUS-PLD (H). The cell in (I) was also treated with F22. Scale bars, 1 mm. (J) Mobile fraction of photobleached assemblies in (G) to (I) (n > 15 cells). Error bars represent SD. (K) Three-dimensional (3D) rendering of FUS-PLD nuclear assemblies.The insets show aberrant “sticky droplets.” Scale bar, 1 mm. (L) Time series to track the lifetime of FUS-GFP HeLa cells. H2B-mCherry was used to detect cell death. Scale bar, 5 mm. (M) Quantification of the fraction of cells undergoing cell death. Error bars represent SD. *P < 0.05, **P < 0.01, and ***P < 0.001 in comparison with the wild type.