A genetic screening strategy identifies novel regulators of the proteostasis network

Abstract

A hallmark of diseases of protein conformation and aging is the appearance of protein aggregates associated with cellular toxicity. We posit that the functional properties of the proteostasis network (PN) protect the proteome from misfolding and combat the proteotoxic events leading to cellular pathology. In this study, we have identified new components of the proteostasis network that can suppress aggregation and proteotoxicity, by performing RNA interference (RNAi) genetic screens for multiple unrelated conformationally challenged cytoplasmic proteins expressed in Caenorhabditis elegans. We identified 88 suppressors of polyglutamine (polyQ) aggregation, of which 63 modifiers also suppressed aggregation of mutant SOD1(G93A). Of these, only 23 gene-modifiers suppressed aggregation and restored animal motility, revealing that aggregation and toxicity can be genetically uncoupled. Nine of these modifiers were shown to be effective in restoring the folding and function of multiple endogenous temperature-sensitive (TS) mutant proteins, of which five improved folding in a HSF-1-dependent manner, by inducing cytoplasmic chaperones. This triage screening strategy also identified a novel set of PN regulatory components that, by altering metabolic and RNA processing functions, establish alternate cellular environments not generally dependent on stress response activation and that are broadly protective against misfolded and aggregation-prone proteins.

Figure 1. Genome-wide RNAi screen for suppression of Q35 aggregation.

(A) Q35 animals (6 days old) show suppression of aggregation relative to EV control. Representative modifier genes: cdk-1 (cyclin-dependent kinase); let-607 (CREB/ATF transcription factor); cyn-11 (cyclophilin); klp-15 (kinesin-like protein); sri-57 (serpentine receptor) and F59C6.5 (NADH-ubiquinone oxidoreductase). yfp-RNAi is the control for RNAi efficiency. Panels IX–XXIV are higher magnification images of the boxed areas in I–VIII. Scale bars: 0.1 mm (I–XVI), 0.05 mm (XVII–XXIV)]. (B) Animals expressing soluble Q24 were used as a control for RNAi phenotypes dissociated from aggregation (scale bar 0.1 mm). (C) Q35 aggregate count (% foci relative to EV control) for a representative group of modifiers (±SD, n>3). Student t-test ***p<0.001 relative to control.

Figure 2. Suppressors of aggregation maintain polyQ in a diffuse state without affecting expression levels.

(A) SDS-PAGE and western blotting analysis of protein samples from Q35 RNAi-treated animals (6 days old), immunoblotted with anti-YFP (32 KDa) and anti-α-tubulin (55 KDa) antibodies. Control refers to EV RNAi (I–IV). YFP/tubulin ratios were calculated from protein band intensities and are shown as an average % of the control (±SD, from ≥3 biological replicates, Student t-test p>0.05). yfp-RNAi in panel II is the positive control for reduced protein levels. (B,C) FRAP analysis confirms suppression of polyQ aggregation to a diffuse state. (B) Q35 protein was subjected to photobleaching in animals treated with control RNAi (left) or hmg-3 RNAi (right) and fluorescence recovery was measured at the indicated time points. (C) Quantitative FRAP analysis indicates the relative fluorescence intensity (RFI) at each time point, and it represents an average of ≥12 independent measurements for each RNAi (5 for the controls). The soluble Q24 control is shown in black solid line, and the Q35 foci control in black dashed line. (D) Native PAGE analysis of whole protein extracts from 6 day old Q35 animals treated with RNAi. Q35 aggregated protein retained at the top of the gel was reduced by each of the modifiers tested.

Figure 3. Common RNAi suppressors of Q35 and Q37 aggregation.

(A) Counter-screen in 5 day old Q37 animals to identify the strongest suppressors of polyQ aggregation. Panels IX–XVI show a higher magnification image of the boxed areas on I–VIII. Scale bar is 0.1 mm. (B) Q37 aggregate count (% foci relative to EV control) for a representative group of modifiers (±SD, n>3). Student t-test ***p<0.001 relative to control. (C) Screen strategy: genome-wide screen with the threshold Q-length for aggregation Q35, and counter-screens with the soluble Q24 and the higher Q-length Q37 strains. Class A refers to Q35 and Q37 common strong modifiers, and Class B to modifiers only affecting Q35.

Figure 4. PolyQ aggregation modifiers tested in the mutant SOD1^G93A model.

(A) Representative RNAi suppressors of SOD1^G93A aggregation: ucr-2.3 (ubiquinol cytochrome c reductase); F43G9.1 (isocitrate dehydrogenase); ZK430.7 (sof1-like rRNA processing protein); gei-11 (GEX-3-interacting protein, Myb transcription factor); let-607 (CREB/ATF transcription factor); R05D7.2 (unknown) and yfp RNAi control. Higher magnification images (IX–XVI) of the boxed areas (I–VIII) show reduced number of SOD1^G93A foci in animals treated with RNAi, relative to the EV control. Scale bar is 0.1 mm. (B) SDS-PAGE and western blot analysis of protein samples from SOD1^G93A RNAi-treated animals (5 days old), immunoblotted with anti-YFP (32 KDa) and anti-α-tubulin (55 KDa) antibodies. Control refers to EV RNAi. YFP/tubulin ratios (bottom) were calculated from protein band intensities and are shown as an average % of the control (I, II) ±SD, from ≥3 biological replicates (Student t-test p>0.05). yfp-RNAi (II) is the positive control for reduced protein levels.

Figure 5. Suppression of polyQ aggregation and toxicity are genetically uncoupled.

(A) Motility measurements of Q35 (black) and wt (grey) animals treated with aggregation suppressor RNAi. Motility is measured in body-length per second relative to wt motility in EV control (% BLPS±SEM). Shown here are the modifiers that did not affect wt motility, grouped in three classes that: enhance Q35 motility defect (Student t-test **p<0.001); cause no change (Student t-test p≥0.05); or suppress Q35 motility defect (Student t-test **p<0.01) relative to Q35 control. Statistical significance between classes calculated by 1-way ANOVA ***p<0.0001; and t-test ***p<0.001. (B) Aggregation modifiers that caused a deleterious effect on wt motility (% BLPS±SEM, relative to EV control). Student t-test ***p<0.001 relative to control. (C) Screening triage for modifier suppressors of protein aggregation and toxicity.

Figure 6. Aggregation modifiers that rescue the folding of endogenous TS mutant proteins.

Modifiers of aggregation and toxicity were tested on endogenous muscle TS mutant proteins. 15°C is the permissive temperature, 25°C is the restrictive temperature and 23°C is the temperature used for RNAi. Misfolding of TS mutant proteins was assessed by measuring the % of animals displaying the associated muscle dysfunction phenotype: (A) unc-15(e1402) (paramyosin), uncoordinated/slow movement; (B) unc-45(e286) (myosin assembly protein), egg laying and paralysis defect; (C) unc-54(e1157) (myosin), slow movement/paralysis; (D) unc-52(e669su250) (perlecan), stiff paralysis (±SD, n>3, Student t-test relative to 23°C control **p<0.01, ***p<0.001). Statistical comparison using 1-way ANOVA ***p<0.001.

Figure 7. Core PN modifiers and activation of the heat shock response.

(A) Screening strategy to identify genetic modifiers that enhance the folding environment and are effective in multiple misfolding models. (B) Suppression of polyQ aggregation by the final nine modifiers dependence on HSF-1. Aggregate quantification on RNAi-treated Q37;hsf-1(sy441) hypomorphic mutant animals relative to control (±SD). Student t-test *p<0.05; **p<0.01; ***p<0.001 (ns, non-significant). (C) Real-time qPCR analysis of the levels of hsp (C12C8.1, F44E5.4, hsp-16.1) genes in RNAi-treated wt animals. Data are relative to the levels of each gene in wt;EV control (±SD) from 3 biological replicates.