Condensation of Ded1p Promotes a Translational Switch from Housekeeping to Stress Protein Production

Abstract

Cells sense elevated temperatures and mount an adaptive heat shock response that involves changes in gene expression, but the underlying mechanisms, particularly on the level of translation, remain unknown. Here we report that, in budding yeast, the essential translation initiation factor Ded1p undergoes heat-induced phase separation into gel-like condensates. Using ribosome profiling and an in vitro translation assay, we reveal that condensate formation inactivates Ded1p and represses translation of housekeeping mRNAs while promoting translation of stress mRNAs. Testing a variant of Ded1p with altered phase behavior as well as Ded1p homologs from diverse species, we demonstrate that Ded1p condensation is adaptive and fine-tuned to the maximum growth temperature of the respective organism. We conclude that Ded1p condensation is an integral part of an extended heat shock response that selectively represses translation of housekeeping mRNAs to promote survival under conditions of severe heat stress.

Keywords: 5′; Ded1p; UTR; chaperone; condensate; cytosolic pH; evolutionary adaptation; heat shock response; heat stress; phase separation; ribosomal scanning.

Graphical abstract

Figure S1

Extended Data and Quality Control for Ribosome Profiling and RNAseq Experiments, Related to Figure 1 (A) Principal component analysis (PCA) of the normalized count data from duplicate ribosome profiling reads (left) and RNaseq samples (right) taken from yeast treated for 10 min at 30°C, 40°C and 42°C. Plot shows the first two principal components accounting for 59% and 26% of the total data variance, respectively. (B) Clustering distance heatmap of the projection of the Euclidean distance between samples, which approximately corresponds to the biological variation coefficient. (C) Volcano plots displaying differentially translated genes using translation efficiencies. (D) Fraction of soluble Ded1-mCherry in ribosome profiling yeast cell lysate samples after treatment at 30°C, 40°C and 42°C as assessed by immunoblotting. (E) Distribution of the minimum free energy of the 5′UTR of all yeast mRNAs compared to significantly repressed mRNAs at 40°C (in comparison to 30°C) or 42°C (in comparison to 40°C) in wild-type (WT) yeast cells. Significance was confirmed using a two-sided Wilcoxon test.

Figure 1

Heat Shock Promotes a Switch in Gene Expression that Mimics Ded1p Inactivation (A) Genome-wide distribution of log₂-fold-changes of ribosome profiling and RNA abundance between 30°C and 40°C (top left) and 40°C and 42°C (top right). Differentially expressed genes (adjusted p value [p adjust] < 0.05) at the level of transcription (orange) and translation (blue), concordant (transcription and translation in the same direction, green), or in opposite directions (red) are highlighted and identified at the bottom right. Stacked bar charts (bottom left) indicate the absolute numbers of transcriptionally and translationally differentially expressed genes. (B) Distribution of the MFE of the 5′ UTR of all yeast mRNAs compared with the translational efficiencies of mRNAs significantly induced or repressed at 42°C versus 40°C in wild-type (WT) yeast cells. (C) Analysis of the data from Sen et al. (2015), comparing the TE of yeast expressing WT Ded1 or Ded1^ts at 37°C. The distribution of the MFE of all yeast mRNAs was compared with that induced or repressed with a log₂-fold-change of at least 0.5 at the restrictive temperature. Statistical significance was calculated with a two-sided Wilcoxon test. See also Figure S1 and Table S1.

Figure 2

The RNA Helicase Ded1p Assembles in Response to Elevated Temperature (A) Fluorescence images of S. cerevisiae expressing Ded1-GFP and the SG marker protein Pab1-mCherry. Left: yeast before exposure to heat stress. Right: yeast after 10 min at 46°C (right). Cells were imaged at 46°C to observe colocalization with Pab1-mCherry, which only assembles above 44°C (yellow arrows). Scale bar, 5 μm. (B) Quantification of Ded1-GFP foci per cell following a temperature ramp of 10°C/min from a starting temperature of 35°C using live-cell microscopy. Representative images of cells after 5 min are shown on the right. Scale bar, 5 μm. (C). Fraction of soluble Ded1p in cell lysates after 10-min incubation at the indicated temperatures (mean, SD, n = 3). (D) Distribution of the growth rate (h⁻¹) of W303 yeast cells at different temperatures. The trend line (gray) is shown as a guide. See also Figure S2.

Figure S2

The RNA Helicase Ded1p Assembles in Response to Elevated Temperature, Starvation Conditions, and Lowered pH, Related to Figure 2 (A) S. cerevisiae expressing endogenous levels of Ded1-GFP and Pab1-mCherry during exponential growth (left) and after exposure to 60 min of glucose starvation (right). Yellow arrows highlight Ded1p and Pab1p assemblies. Scale bar 5 μm. (B) S. cerevisiae expressing Ded1-GFP incubated in phosphate buffer pH 7.0 (top right) or pH 5.8 (bottom right) for 60 min in the presence of DNP. Yeast prior to DNP treatment shown left. Yellow arrows highlight Ded1p assemblies. Scale bar 5 μm. (C) Representative immunoblot for the sedimentation analysis of Ded1-mCherry after 10 min of heat stress in yeast cells (see Figure 2C for quantification). PGK as a loading control. (D) Representative immunoblot (top) used for the quantification (bottom) of the soluble fraction of Ded1-mCherry after 30 min of heat stress at indicated temperatures (Mean, SD, n = 3).

Figure 3

Ded1p Has an Intrinsic Property to Form Condensates upon Elevated Temperature (A) Fluorescence image of 2 μM purified Ded1-GFP in piperazine-N,N′-bis(2-ethanesulfonic acid (PIPES)/KOH buffer (pH 7) and 100 mM KCl buffer at 25°C (left image) and after 10 min at 42°C (right image). Scale bar, 3 μm. (B) Phase diagram of Ded1p in PIPES/KOH and 100 mM KCl buffer with the control parameters pH and temperature. Analysis was performed after 30 min of incubation. Green dots indicate condensates, and “x” indicates no condensates. (C) Images from a time-lapse video monitoring Ded1p assembly at pH 6.8. Scale bar, 15 μm. (D) Hydrodynamic radius of 0.5, 1, and 2 μM Ded1-GFP in PIPES/KOH (pH 6.8) and 100 mM KCl buffer as a function of temperature. (E) Kinetic analysis of the apparent Ded1-GFP and Pab1-GFP assembly reaction, measured by DLS in PIPES/KOH (pH 6.8) and 100 mM KCl buffer. Plotted is the natural logarithm of the apparent rate constant (ln(k)) as a function of the reciprocal temperature. (F) Analysis of the transition temperature midpoints of Ded1p condensation as determined by light scattering (T_M scattering) and changes in tertiary structure as determined by nano-differential scanning fluorimetry (DSF) (T_M tertiary structure) at varying Ded1p concentrations. See also Figure S3.

Figure S3

Ded1p Has an Intrinsic Property to Form Condensates upon Elevated Temperature and Lowered pH, Related to Figure 3 (A) Brightfield images of 2 μM purified Ded1p protein without fluorescent tag imaged at 25°C in PIPES/KOH pH 7.0, 100mM KCl buffer (left image), after exposure to 42°C for 10 min (top right) or in PIPES/KOH pH 6.0, 100 mM KCl buffer (bottom right). Scale bar 3 μm. (B) Representative nano-DSF measurement monitoring changes in the intrinsic tryptophan/tyrosine fluorescence ratio (F350/330 nm). Dashed line depicts T_M. (C) Light scattering plot of the same sample shown in B. Dashed line depicts T_M. (D) Left: CD spectra of Ded1p at different temperatures starting from 21°C (green) to 45°C (pink). Right: Normalized CD spectra. (E) Left: CD spectra of Ded1p in PIPES/KOH pH 6 and pH 7, 100 mM KCl. Right: Normalized CD spectra.

Figure 4

mRNA Is Sequestered into Ded1p Heat-Induced Condensates and Affects Condensate Physical Properties (A) Images of 2.5 μM purified Ded1-GFP mixed with water, tRNA, rRNA, or mRNA (capped and poly(A)-tailed coding for nano-luciferase with the 5′ UTR of PAB1) at a final RNA concentration of 45 ng/μL in PIPES/KOH (pH 6.8) and 200 mM KCl buffer and incubated at 42°C for 15 min. Scale bar, 5 μm. (B) Mean hydrodynamic radius of Ded1p in the presence or absence of mRNA as a function of temperature in PIPES/KOH (pH 6.8) and 200 mM KCl buffer. (C) Analysis of the soluble RNA fraction after co-incubation with Ded1-GFP for 15 min at 42°C in PIPES/KOH (pH 6.8) and 200 mM KCl buffer. 260-nm values were normalized to the fraction of soluble RNA at 25°C (mean, SD, n = 3). (D) Mean and standard deviation of fluorescence intensities of n = 40 condensates of Ded1 with and without mRNA formed for 10 min at 42°C in PIPES/KOH (pH 6.8) and 200 mM KCl buffer and labeled with Sybr Green II RNA dye. (E) Condensates formed by 5-min incubation at 42°C in PIPES/KOH (pH 6.8) and 200 mM KCl buffer were made to fuse with optical tweezers at room temperature. Top: schematics of controlled condensate fusion probed with optical tweezers. Center: bright-field images of a fusion time course. Scale bars, 2 μm. Bottom: force curve (laser signal [ arbitrary units, AU]) as a function of time. The relaxation time was derived from a single exponential fit (pink line) of the fusion trace. (F) Ded1p condensates with and without mRNA formed after 5 min at 42°C in PIPES/KOH (pH 6.8) and 200 mM KCl buffer (left) and upon increasing the NaCl concentration to 1 M (right). Scale bar, 3 μm. See also Figure S4 and Video S1.

Figure S4

Ded1p Hyper-Dependent mRNAs Are Preferentially Sequestered into Ded1p Heat-Induced Condensates, Related to Figure 4 (A) Controls for Figure 4C. The fraction of soluble mRNA after heat treatment in the presence of GFP or PIPES/KOH pH 6.8, 200mM KCl buffer was measured by nanodrop using an absorbance of 260 nm and normalized to the fraction of soluble mRNA prior to treatment at 25°C (Mean, SD, n = 3). (B) Schematic representation of mRNAs used in microscopy experiment shown in (C). To circumvent effects of mRNA length on Ded1p condensation, mRNAs of similar length were synthesized: full length for the RPL41A mRNA and the first 350 nucleotides of the 5′UTR and coding sequence of PMA1 and SFT2 mRNAs. (C) Representative images of 1 μM Ded1-GFP and 18 ng/μL Cy3 labeled mRNAs after 10 min at 42°C in PIPES/KOH pH 6.8, 150mM KCl buffer in the presence of absence of 18 ng/μL tRNA (n = 3). (D and E) Quantification of the condensed Ded1p fraction and partitioning (I_in/I_out) of mRNA into Ded1p condensates from multiple fields of view are shown in D and E, respectively.

Figure 5

The Phase Behavior of Ded1p Is Modulated by Its IDRs (A) Top: domain structure of Ded1p. Bottom: disorder plot of Ded1p (VSL2 function of the online tool Predictor of Natural Disordered Regions (PONDR); Peng et al., 2005). (B) Mean hydrodynamic radii of 3 μM GFP-tagged Ded1p (Ded1p), Ded1ΔN, and Ded1ΔC as a function of temperature in phosphate (pH 7.4) and 200 mM KCl buffer. (C) Phase diagram of Ded1p and Ded1ΔN in PIPES/KOH (pH 7.5) and 200 mM KCl buffer (left) and Ded1p and Ded1ΔC in PIPES/KOH (pH 6.8) and 200 mM KCl buffer (right) over a temperature and concentration range as determined by the scattering function of nano-DSF. The data of three technical replicates were plotted. Trendlines are shown as a guide. (D) Mean hydrodynamic radii of 4 μM GFP-tagged Ded1p (wild-type) and Ded1-IDR_m (IDR_m) as a function of temperature in phosphate (pH 7.4) and 100 mM KCl buffer. (E) Images of S. cerevisiae expressing Ded1-mCherry and Ded1-IDR_m-mCherry after 30 min at the indicated temperatures. Yellow arrows point to the temperature at which assemblies were first observed. Scale bar, 5 μm. (F) Spot titer assay with 5-fold serial dilutions of strains expressing Ded1-mCherry or Ded1-IDR_m-mCherry grown at 22°C, 30°C, 39°C, and 41°C. Shown are representative images of 3 independent repeats. See also Figure S5.

Figure S5

The Condensation of Ded1p in Response to Elevated Temperature and Lowered pH is Regulated by Its IDRs, Related to Figure 5 (A) Representative images of 2 μM Ded1p, Ded1ΔN and Ded1ΔC in PIPES/KOH buffers ranging from pH 6.0 to 7.5 in 100 mM KCl. Scale bar, 5 μm. (B) Light scattering of GFP-tagged 3 μM Ded1p (blue), 3 μM Ded1ΔN (green), 3 μM (magenta) and 10 μM Ded1ΔC (light magenta) measured using nano-DSF as a function of temperature in PIPES/KOH pH 7.5, 200 mM KCl buffer. Data were normalized to the minimum light scattering value of the corresponding protein. (C) Alignment of the N-terminal region of Ded1p and Ded1-IDR_m with the ClustalW algorithm (Chenna et al., 2003). (^∗): identical amino acids, (:): highly similar amino acid properties, (.): similar amino acids. (D) Quantification of the condensed fraction of Ded1p (wild-type) and Ded1-IDR_m (IDR_m) as a function of temperature as determined by microscopy data shown in (E). (Mean, SD). (E) Representative images of 4 μM GFP-labeled Ded1p (wild-type) and Ded1-IDR_m (IDR_m) in PIPES/KOH pH 6.8, 200 mM KCl buffer at indicated temperatures. Scale bar, 10 μm.

Figure S6

Extended Data and Quality Control for Ribosome Footprinting and RNAseq Experiments of IDRm and Wildtype Ded1p, Related to Figure 6 (A) Clustering distance heatmap of the projection of the Euclidean distance between ribosome profiling samples (left) and RNaseq samples (right). (B) Principal component analysis (PCA) of the normalized count data from the ribosome profiling reads (left) and RNaseq reads (right) of the duplicate samples. (C) Volcano plots showing differentially translated genes. (D) Fraction of soluble Ded1-mCherry and Ded1-IDR_m-mCherry from yeast incubated at 30°C, 40°C and 42°C for 10 min as assessed by immunoblotting (Mean, SD, n = 3). (E) Analysis of the soluble (supernatant) and insoluble (pellet) fractions of yeast expressing Ded1p and Ded1-IDR_m incubated at 30°C, 40°C and 42°C from RNaseq read analysis. Boxplot displays the distribution of the 5′UTR minimum free energies of transcripts differentially enriched in the pellet over the supernatant (left) and supernatant over pellet (right) compared to the 5′UTR of all genes. Significance was confirmed using a two-sided Wilcoxon test. (F) Microscopy images of 0.8 μM Ded1-GFP condensates after 30 min at 40°C or 42°C in pH 6.8 PIPES/KOH, 100mM KCl buffer before (left) and after (right) addition to the translation assays at 25°C. Images were taken right after the addition of the condensates (0 min) and at the end of the assay (90 min). Scale bar, 5 μm.

Figure 6

Heat-Induced Condensation of Ded1p Facilitates a Switch in Translation to Selectively Repress Housekeeping mRNAs (A) Length versus MFE for all yeast 5′ UTR sequences as determined by Kertesz et al. (2010). Heat shock factors and housekeeping proteins are color labeled. (B) Boxplots showing the distribution of the MFE of all yeast mRNAs and those induced or repressed in yeast expressing Ded1-IDR_m compared with WT cells at 42°C. Significance was confirmed with a two-sided Wilcoxon test. Distance clustering heatmaps and systematic comparison by principal-component analysis (PCA) are shown in Figures S6A and S6B. (C) Log₂-fold-change of the TEs of each gene plotted against the MFE. Heat shock factors and housekeeping proteins are color labeled. (D) Schematics of reporter transcripts. 5′ UTRs from mRNAs with structurally complex 5′ UTRs (housekeeping transcripts) or 5′ UTRs with little structure (stress transcripts) were fused to nano-luciferase, a short 3′ UTR and a poly(A) tail. (E) In vitro translation of stress reporter transcripts and housekeeping transcripts as measured by luminescence in the presence of different Ded1p concentrations. Luminescence measurements were normalized to the reaction containing no added Ded1p (mean, SD, n = 3). (F) In vitro translation of different reporter transcripts in the presence of diffuse 0.8 μM Ded1p (25°C) or Ded1p condensates formed at 40°C or 42°C for 30 min in PIPES/KOH (pH 6.8) and 100 mM KCl buffer. Luminescence measurements were normalized to the reaction with the unstructured GIS2 5′ UTR reporter (mean, SD, n = 3). See also Figure S6 and Table S2.

Figure 7

Ded1p Homologs from Different Fungi Exhibit Distinct Temperature Sensitivity (A) Schematic representation of the growth temperature ranges for cold-adapted S. kudriavzevii, mesophilic S. cerevisiae, and thermophilic T. terrestris. The temperatures at which the organisms experience heat stress are marked in magenta. (B) Representative images of S. cerevisiae in which endogenous Ded1 was allele replaced with GFP-tagged DED1 from S. kudriavzevii (S.k.), T. terrestris (T.t.), or mock replaced (S.c.) at the indicated temperatures. Yellow arrows highlight the temperature at which Ded1p assemblies were first observed. Scale bar, 5 μm. (C) Mean hydrodynamic radius of Ded1-GFP homologs as a function of temperature in PIPES/KOH (pH 6.8) and 100 mM KCl buffer at 2 μM protein concentration. (D) Microscopy images of 2 μM Ded1-GFP homologs incubated in PIPES/KOH and 100 mM KCl buffer for 10 min at the indicated temperatures. Scale bar, 5 μm. (E) Model. Under growing conditions, Ded1p facilitates scanning of housekeeping mRNAs that harbor a complex secondary structure in their 5′ UTR, whereas it is not needed for stress mRNAs that harbor simple 5′ UTRs. Other factors may work together with Ded1p to regulate translation of structurally complex housekeeping mRNAs. In addition, stress mRNAs could be specifically regulated by factors that have yet to be identified (factor X). Upon heat stress, Ded1p and potentially other factors condense, and, consequently, Ded1p-dependent housekeeping mRNAs are silenced, whereas stress mRNAs are preferentially translated. See also Figure S7 and Table S3.

Figure S7

Condensation of Ded1p in Response to Elevated Temperature and Lowered pH Is Conserved in Three Fungi, Related to Figure 7 (A) Coomassie stained SDS-PAGE of purified Ded1-GFP from S. kudriavzevii (S.k.), S. cerevisiae (S.c.) and T. terrestris (T.t.). (B) Light scattering of purified Ded1-GFP homologs in PIPES/KOH pH 6.8, 100 mM KCl buffer. Temperature ramp 1°C/min. (C) Representative microscopy images of Ded1-GFP homologs in PIPES/KOH pH 6.8 or PIPES/KOH pH 5.7, 100 mM KCl buffer. Scale bar, 3 μm.