Widespread regulation of translation by elongation pausing in heat shock

Abstract

Global repression of protein synthesis is a hallmark of the cellular stress response and has been attributed primarily to inhibition of translation initiation, although this mechanism may not always explain the full extent of repression. Here, using ribosome footprinting, we show that 2 hr of severe heat stress triggers global pausing of translation elongation at around codon 65 on most mRNAs in both mouse and human cells. The genome-wide nature of the phenomenon, its location, and features of protein N termini suggested the involvement of ribosome-associated chaperones. After severe heat shock, Hsp70's interactions with the translational machinery were markedly altered and its association with ribosomes was reduced. Pretreatment with mild heat stress or overexpression of Hsp70 protected cells from heat shock-induced elongation pausing, while inhibition of Hsp70 activity triggered elongation pausing without heat stress. Our findings suggest that regulation of translation elongation in general, and by chaperones in particular, represents a major component of cellular stress responses.

Figure 1. Heat shock induces a global increase in ribosome occupancy around 5′ ends of open reading frames

(A) Polysome profiles of mouse fibroblasts (3T3 cells) in normal growth conditions (blue) or following mild heat shock (42° for 8 hours, green) or severe but not lethal heat shock (44° for 2 hours, red). The polysome region was defined as 5somes and higher. P:M (Polysome to Monosome) ratio is 0.79 in Control, 0.18 in HS8M and 0.08 in HS2S. Profile is representative of three replicate experiments, where relative P:M ratios in HS8M were 22.5% ± 0.9% of Control values and relative P:M ratios in HS2S were 7.2% ± 2.7% of Control. (B) Normalized footprint density along mRNAs (see Experimental procedures). (C) Raw footprint count per position (smoothed) across the Serpine1 transcript (ENSMUST00000041388, 5′ UTR and first 500 nt of the transcript are shown). (D) Distribution of r5′LR values in HS8M (green) and HS2S (red). Population means are indicated. (E) Changes in 5′LR values following heat shock. See Figure S1 and Table S1 for more details.

Figure 2. Ribosomes are paused after translating about 65 amino acids in severe heat shock

(A) Polysome profiles of harringtonine treated Control (left graph, Control) and severe heat shock (right graph, HS2S) treated cells. Heat shocked or control cells were harvested after treatment with 0.1mM harringtonine for 1 minute (dark blue and purple lines for Control and HS2S respectively) or 3 minutes (cyan and pink lines for Control and HS2S respectively) or without any treatment (blue and red lines). P:M ratios are: 0.8 in Control, 0.33 in Control+Har 1min, 0.07 in Control+Har 3min, 0.10 in HS2S, 0.09 in HS2S+Har 1 min, and 0.05 in HS2S+Har 3min. A replicate experiment showed a similar polysome collapse (Figure S2A). (B) Illustration of the KS location and p-value calculation for Vimentin (ENSMUST00000028062). Cumulative plots of reads along the ORF (bottom panel) are shown for Control (blue) and HS2S (red) (see Experimental Procedures). (C) Distribution of the KS locations for all unique translated genes which were found to be significantly paused (at an FDR of 5%) in HS8M relative to Control (green) and in HS2S relative to Control (red). The number of significant genes is indicated, out of 5125 genes. (D) Raw footprint counts (smoothed) along the Atf4 transcript (ENSMUST00000109605) show an escape from the global pause. uORF positions are marked. See Figure S2 and Table S2 for more details.

Figure 3. Association of HSC70 with polysomes is altered following heat shock

(A) Western blot of HSC70/HSP70 (using pan-HSP70 antibody, see Experimental procedures for details) protein levels (middle panel) across a polysome gradient in Control and severe heat shock cells (HS2S, top panel). RPL10A was used to assess ribosome abundance in each fraction (middle panel). Levels of comigrating HSC70/HSP70 protein and RPL10A protein were measured using ImageJ software. Ribosomal association of HSC70/HSP70 was quantified by normalizing the amount of HSC70/HSP70 protein by the amount of RPL10A protein in each fraction (HSC70/HSP70 / RPL10A) for both control (blue bars) and heat shock (red bars) conditions. P:M ratio is 1.9 in Control and 0.18 in HS2S. Profile is representative of six experiments, where relative P:M ratios of HS2S were 8.5% ± 2.9% of Control. (B) Quantification and significance of the reduction in HSC70/HSP70 ribosome association following heat shock. Mean and standard error of the ratio (log₂) of HSC70/HSP70 association (HSC70/HSP70 signal/RPL10A signal), in HS2S to control conditions, in each fraction, over 5 biological replicate experiments. Fractions in which HSC70/HSP70 and/or RPL10A signal was undetectable were excluded from analysis (see Methods for details). Significance of reduction in HSC70 ribosomal association is given by t-test. * indicates p < 0.05; ** indicates p < 0.01. (C) Hydrophobicity scores (see Supplemental Experimental procedures) of the first 70 amino acids encoded by all highly translated mRNAs (in black), of significantly paused mRNAs (in magenta) and of non-significantly paused mRNAs (in orange). The background representing the overall average Kyte-Doolittle hydrophobicity of all highly translated mRNAs (grey dashed line) is shown for reference. T-test p-value for the difference in the mean hydrophobicity values of the first 25 amino acids between all significantly paused mRNAs and non-significant mRNAs was 3.8 × 10⁻⁷. (D) Hsp70 binding site scores (Supplemental Experimental Procedures) were plotted as in (C). T-test p-value for the difference in the mean Hsp70 scores of the first 25 amino acids between all significantly paused mRNAs and non-significant mRNAs was 5.3 × 10⁻⁶. See also Figure S3.

Figure 4. Modulation of chaperone activity regulates translation elongation pausing

(A) Polysome profiles of a thermotolerance experiment. 3T3 cells were treated with mild (green), severe (red), or mild prior to severe (magenta) heat shock, or left untreated (blue). P:M ratios are 1.4 in Control, 0.33 in HS8M, 0.1 in HS2S, and 0.2 in HS8M2S. Profile is representative of two replicate experiments, where relative P:M ratios were 22.4% ± 1.3% of Control in HS8M, 5.8% ± 1.7% of Control in HS2S, and 12.2% ± 2.7% of Control in HS8M2S. (B) Normalized footprint density along mRNAs for a biological replicate experiment and for cells pre-treated with mild heat shock (8 h at 42° C) prior to severe heat shock (magenta). (C) Polysome profiles of cells treated with the Hsp70 inhibitor VER-155008 (Massey et al., 2010) at a 20 μM concentration (Hsp70 inhibitor, cyan line) or DMSO (Control, blue line) for 3hrs. P:M ratios are 1.52 in Control and 0.7 in Hsp70 inhibitor. Profile is representative of three replicate experiments, where relative P:M ratios for Hsp70 inhibitor were 46.5% ± 3.8% of Control. (D) Normalized footprint density along mRNAs for control cells after 3 hours of VER-155008 treatment (cyan). Control and severe heat shock plots (dotted lines) are identical to those in B. (E) Normalized footprint density along mRNAs for 293T cells overexpressing GFP or Hspa1a, with or without 2 hours of severe heat shock. See also Figure S4.

Figure 5. The Hsp70 chaperone – translation machinery interactome

(A) HSC70-translational machinery interactome in a network representation, was drown using Cytoscape. Proteins are represented by circles, and the size of the circle corresponds to the HSC70 interaction score (log2([HSC70, luminescence]/[FLAG ELISA]). Different groups of interest are indicated. (B) Hsp70 interactome changes upon 2 hours of severe heat shock was tested using co-IP. Blots were quantified using ImageJ and shown are log2 of Hsp70 (interaction) normalized to FLAG (expression), of proteins that changed more than 1.5 fold. Striped bars represent proteins with increased normalized interaction and significant destabilization, defined as a significant reduction in FLAG levels – at least 3 fold in one replicate and at least 2 fold in the other. Dark grey bars represent increased interactions with Hsp70 with stable protein levels. Light grey bars represent decreased interactions without significant destabilization. Mean and standard deviation of two experiments is shown. (C) co-IP WB data for several proteins. See Supplemental Experimental procedures, Figure S5, Table S3 and Table S4 for more details.

Figure 6. Heat shock induced translation elongation pausing is modulated by Hsp70 chaperones

During severe heat shock, ribosomes often pause at 5′ ends of mRNAs after translating about 65 amino acids (Figs. 1, 2), resulting in accumulation of ribosomes upstream and depletion of ribosomes downstream of this point; Following severe heat shock, HSC70/HSP70 association with the ribosome is reduced (Fig. 3A, 3B). As Hsp70 proteins interact with nascent peptide chains as they emerge from the exit tunnel (Beckmann et al., 1990; Nelson et al., 1992), nascent chains might be exposed. More significantly paused mRNAs tend to encode N-termini with stronger Hsp70 requiring characteristic (Fig. 3C and 3D). Modulation of Hsp70 levels, by thermotolerance or overexpression, rescue heat shock induced elongation pausing, while inhibition of Hsp70 activity leads to pausing in the absence of stress (Fig. 4). Hsp70 chaperones interact with various translation factors and ribosomal proteins, and this interaction is regulated following heat shock (Fig. 5). HSC70/HSP70 may affect ribosome dynamics via its interaction elongation factors, regulation of ribosomal proteins in the exit tunnel, or through reduced interaction with nascent chains.