Regulated Formation of an Amyloid-like Translational Repressor Governs Gametogenesis

Abstract

Message-specific translational control is required for gametogenesis. In yeast, the RNA-binding protein Rim4 mediates translational repression of numerous mRNAs, including the B-type cyclin CLB3, which is essential for establishing the meiotic chromosome segregation pattern. Here, we show that Rim4 forms amyloid-like aggregates and that it is the amyloid-like form of Rim4 that is the active, translationally repressive form of the protein. Our data further show that Rim4 aggregation is a developmentally regulated process. Starvation induces the conversion of monomeric Rim4 into amyloid-like aggregates, thereby activating the protein to bring about repression of translation. At the onset of meiosis II, Rim4 aggregates are abruptly degraded allowing translation to commence. Although amyloids are best known for their role in the etiology of diseases such as Alzheimer's, Parkinson's, and diabetes by forming toxic protein aggregates, our findings show that cells can utilize amyloid-like protein aggregates to function as central regulators of gametogenesis.

Figure 1

B-isox binds Rim4 in meiosis I (see Figure S1 for detailed workflow) (A) Diagram of Rim4 (B) RIM4-3V5, pGAL-NDT80, GAL4.ER (A30868) cells were induced to sporulate at 30°C. After 6 hours when cells had arrested in G 2 due to the lack of Ndt80 lysates were prepared. Lysates were incubated with 100 μM b-isox (or DMSO control) and Rim4 abundance was examined by Western blot analysis. Pgk1 did not bind b-isox. (C) Cells were grown as in (B) except cells were released from the G2 block by the addition of 1 μM β-estradiol. Lysates were prepared from G2-arrested cells (0 hour) and at 1, 2, and 3 hours post-release and analyzed as in (B). (D) Meiotic staging data for (C). The percentage of metaphase I, anaphase I, metaphase II, and anaphase II cells was determined. (E) pGAL-NDT80 (A15055) strains carrying the GAL4.ER were grown and lysates prepared as in (C). Precipitated proteins were ITRAQ-labeled and analyzed by mass spectrometry. Shown are the ratios of b-isox enrichment/DMSO for each identified protein converted to z-scores. (F) Meiotic staging data from (E).

Figure 2

Rim4 forms amyloid-like aggregates in vitro (A, B) Transmission electron microscopy (TEM) of fibers formed by recombinant Rim4 and Rim4Δ294C (aa 1-420) incubated at 4°C overnight. The top three panels are micrographs depicting negatively stained samples (prepared at 40 mg/ml and 1 mg/ml). The bottom left panels depict Rim4 samples analyzed by cryo EM at concentrations of 40 mg/ml (left) and 4 mg/ml (right). Scale bars are shown in white. (C) Recombinant Rim4 and Rim4Δ294C (various concentrations diluted to 1 mg/ml just before assay) were incubated with 16 ng/ml Thioflavin T, and the fluorescence intensity was measured as in (Nilsson, 2004). Histone H1 at 4 and 40 mg/ml was used as a control. (D) 10 μg of recombinant protein was slot-blotted onto nitrocellulose and assayed for reactivity with α-amyloid fibril OC. Rim4, Rim4Δ294C, and Histone H1 concentrated to 40 mg/ml were assayed. Amyloid beta (Aβ, Anaspec) concentrated to 0.25 mg/ml was used as a positive control.

Figure 3

Evidence that Rim4 forms aggregates in cells (A, B) Rim4 forms static aggregates in cells. Strains heterozygous for HTB1-mCherry and homozygous for either RIM4-EGFP (A36245) or rim4Δ204C-EGFP (A36706) were induced to sporulate at 30°C. The mob ility of Rim4 aggregates was determined by FRAP of live cells. Cells were imaged starting at 5 hours after transfer into sporulation medium. A haploid mitotic culture expressing Pgk1-GFP (A34422) was used as a soluble control. Bleached areas were normalized to an unbleached control area. Time point zero was the first observed time point after bleaching. Representative cells are shown in (A) with Rim4 in green and the nucleus in red (scale bar = 5 μm). Signal recovery over time is shown in (B). Plotted is the mean and SEM of 10 cells per strain. (C-E) pGAL-NDT80, GAL4.ER cells harboring RIM4-mCherry (A35632) or rim4Δ204C-mCherry (A37237) were induced to sporulate at 30°C. After 6 hours when cells had arrested in G2 due to the lack of Ndt80, the cells were fixed and imaged. (C) Single-plane images of representative cells arrested in G2 are shown. Scale bar = 5 μm (D, E) Rim4-mCherry signal (vacuole excluded) was quantified in the central z-plane. The mean Rim4 signal is not significantly different between the two strains however the standard deviation (aggregation proxy) is significantly different. n = 50 cells per strain. (F) Cells either harboring a RIM4-EGFP (A36775) or rim4Δ204C-EGFP (A36773) fusion were induced to sporulate at 30°C. After 1 h our of growth in batch culture, the cells were loaded onto a microfluidics chip and imaged every 20 minutes. See also Figures S3A-C and Movies S3-S6. A quantification of Rim4-EGFP signal over time is shown. Plots were aligned according to the first meiotic division. Cells harboring RIM4-EGFP and rim4Δ204C-EGFP are shown in black and red, respectively. Plotted is the mean and SEM of 10 cells per strain. Scale bar = 5 μm

Figure 4

Rim4 forms amyloid-like aggregates in vivo to repress translation (A) Rim4 amyloid-like aggregates form during premeiotic S phase which are cleared during meiosis II. pGAL-NDT80 cells (A30868) were induced to sporulate and Rim4’s ability to form SDS-resistant aggregates was analyzed as in (Alberti et al., 2009). Samples were incubated in 2% SDS for 10 min and resolved on 1.7% agarose (with 0.1% SDS) and Rim4 was detected by immunoblot analysis. A strain harboring Rim4 lacking its C-terminal LC sequences (A33848) is shown as a control. (B, C) Western blot analysis of time course shown in 4A. Total levels of Rim4-3V5, Clb3-3HA, and Pgk1 (loading control) are shown in (B). Quantification of blots is shown in (C). Rim4 SDS-resistant aggregates were quantified from SDD-AGE in Figure 4A and Clb3 from SDS-PAGE in Figure 4B. Total levels of Rim4 and Clb3 were normalized to Pgk1. (D-F) CLB3 translational repression begins between 2 and 3 hours after induction of sporulation (see Figure S4A for detailed workflow). A culture harboring RIM-3V5, GAL4.ER, and pGAL-5´UTR_CLB3-CLB3 (A35430) was split and induced to sporulate at 30°C. pGAL-5´UTR_CLB3-CLB3 was induced by β-estradiol at the indicated time. Rim4-3V5, Clb3-3HA, and Pgk1 (loading control) protein and CLB3 mRNA and rRNA (loading control) levels were determined at the indicated times following β-estradiol addition (D). SDS-resistant Rim4 aggregates were analyzed in (E). A quantification of Rim4 aggregate formation (open circles) and Clb3 translation (Clb3 protein/CLB3 mRNA after 60 min induction: closed circles) is shown in (F).

Figure 5

The LC sequences, but not RNA-binding are required for Rim4 aggregation (A) TEM of fibers formed by Rim4-F139L (RRM mutant) concentrated to 27 mg/ml and incubated at 4°C overnight. (B) RNA binding is not required for amyloid-like aggregation. Rim4 was analyzed by SDD-AGE six hours after transfer into sporulation-inducing conditions of cells harboring rim4Δ271C-3V5 (A35408), a no-tag control (A15055), RIM4-3V5 (A35430), and rim4-F139L-3V5 (rrm mutant A35324). (C) Diagram of Rim4 C-terminal truncations (D) Rim4 can be pelleted from detergent-containing undenatured lysate. Strains harboring RIM4-3V5 (A35430), rim4Δ138C-3V5 (A35402), and rim4Δ271C-3V5 (A35408) were arrested in G2 as above. Native lysates were prepared and spun for 60 min at 250,000g. Rim4 levels were determined in total lysate, pellet and supernatant. (E, F) rim4 mutants that fail to form amyloid-like aggregates are defective for translational control. Strains harboring RIM4-3V5 (A35430), heterozygous RIM4-3V5/rim4? (A35841) and constructs shown in (C) (A35399, A35402, A35405, A35408) as well as pGAL-5´UTR_CLB3-CLB3, ndt80?, ZIP1-GFP, SPC42-mCherry, and GAL4.ER were arrested in G2 (6 hours) and pGAL-5´UTR_CLB3-CLB3 was induced by β-estradiol addition. Rim4-3V5, Clb3-3HA, and Pgk1 (loading control) protein and CLB3 mRNA and rRNA (loading control) were determined (E). SDS-resistant Rim4 aggregates (at 6 hours) were analyzed in (F). The Sup35 aggregate-forming strain ([PSI+]; A25208) was used as a positive control.

Figure 6

Effects of the LC sequences on Rim4 RNA binding and CLB3 mRNA mobility in sucrose gradients (A, B) An aggregation-defective rim4 mutant has decreased RNA-binding activity. Extracts from RIM4-3V5 (A35430) and rim4Δ138C-3V5 (A35402) cells arrested in G2 (6 hours) were incubated with in vitro transcribed 3´ biotinylated RNAs conjugated to streptavidin magnetic beads. Beads were recovered and boiled in SDS loading buffer to release bound factors. Shown is the amount of Rim4 recovered using the indicated RNA baits (A). Quantifications (B) were input-normalized and are shown as fold enrichment over the no-RNA control. (C, D) Rim4 and its bound mRNA targets exist in heavy mRNPs. Strains harboring RIM4-3V5 (A35430) and rim4Δ138C-3V5 (A35402) were arrested in G2 as above except CLB3 was expressed by the addition of β-estradiol for 60 minutes prior to collection. Samples were fixed in 1% formaldehyde on ice for 15 mins (stopped by the addition of 0.1 M glycine) to preserve the RNPs (Valazek et al., 2007). Native lysates were fractionated on 10-50% sucrose gradients. Sedimentaton of Rim4, Pgk1 as well as CLB3 mRNA and 18S rRNA was determined by immunoblot and qPCR, respectively. RNA quantification was normalized to luciferase RNA which was spiked into each fraction in equal amounts to normalize for RNA extraction efficiency.

Figure 7

Formation of Rim4 amyloid-like aggregates is regulated by starvation (A, B) Starvation triggers Rim4 aggregation and translational repression (for detailed workflow see Figure S6A). Haploid GAL4.ER, pGAL-5´UTR_CLB3-CLB3 strains with or without pCUP1-RIM4-3V5 (A34980, A28492) were diluted to OD₆₀₀ = 1.8 in either YEPD or SPO medium. pCUP1-RIM4 was induced by the addition of 25 μm CuSO₄. pGAL-5´UTR_CLB3-CLB3 was induced by 2 μM β-estradiol prior to (0 hr) and after (6 hr) CuSO₄ induction for 60 minutes. Rim4-3V5, Clb3-3HA, and Pgk1 (loading control) protein and CLB3 mRNA and rRNA (loading control) levels were determined at the indicated times in (A). SDS-resistant Rim4 aggregates (prior to and after CuSO₄ induction) were analyzed by SDD-AGE in (B). (C) Monomeric Rim4 is converted into amyloid-like aggregates upon starvation (for detailed workflow see Figure S6B). A haploid pGAL-RIM4-3V5 strain was grown in BYTA medium in the presence of 1 μM β-estradiol for 3 hours to induce Rim4 expression. Cells were then shifted to SPO medium. Rim4 SDS-resistant aggregates were assayed. Rim4 and Pgk1 protein levels from TCA extracts were analyzed in (Figure S6C). (D, E) DAZL forms SDS-resistant aggregates in mouse gametogenesis. Brain, intestine, liver, and testes were harvested from a 2 month old C57BL/6J male mouse. Samples were Dounce homogenized in SDD-AGE buffer and analyzed by SDD-AGE (D) and SDS-PAGE (E) DAZL was detected by immunoblot analysis. (F) Model for Rim4-mediated translational control. See text for details.