Information

10.8: Regulation of Translation - Biology

10.8: Regulation of Translation - Biology


We are searching data for your request:

Forums and discussions:
Manuals and reference books:
Data from registers:
Wait the end of the search in all databases.
Upon completion, a link will appear to access the found materials.

Gene expression is primarily regulated at the pre-transcriptional level, but there are a number of mechanisms for regulation of translation as well. One well-studied animal system is the iron-sensitive RNA-binding protein, which regulates the expression of genes involved in regulating intracellular levels of iron ions. Two of these genes, ferritin, which safely sequesters iron ions inside cells, and transferrin, which transports iron from the blood into the cell, both utilize this translational regulation system in a feedback loop to respond to intracellular iron concentration, but they react in opposite ways. The key interaction is between the iron response elements (IRE), which are sequences of mRNA that form short stem-loop structures, and IRE-BP, the protein that recognizes and binds to the IREs. In the case of the ferritin gene, the IRE sequences are situated upstream of the start codon. When there is high iron, the IRE-BP is inactive, and the stem-loop structures are melted and overrun by the ribosome, allowing translation of ferritin, which is an iron-binding protein. As the iron concentration drops, the IRE-BP is activated and binds around the IRE stem-loop structures, stabilizing them and preventing the ribosome from proceeding. This prevents the production of ferritin when there is little iron to bind.

Transferrin also uses iron response elements and IRE-binding proteins, but in a very different mechanism. The IRE sequences of the transferrin gene are located downstream of the stop codon, and play no direct role in allowing or preventing translation.

However, when there is low intracellular iron and there is a need for more transferrin to bring iron into the cell, the IRE-BP is activated as in the previous case, and it binds to the IREs to stabilize the stem-loop structures. In this case; however, it prevents the 3’ poly-A tail degradation that would normally occur over time. Once the poly-A tail is degraded, the rest of the mRNA is destroyed soon thereafter. As mentioned in the transcription chapter, the longer poly-A tails are associated with greater persistence in the cytoplasm, allowing more translation before they are destroyed. The IRE-BP system in this case externally prolongs the lifetime of the mRNA when that gene product is needed in higher amounts.

Since mRNA is a single-stranded nucleic acid and thus able to bind complementary sequence, it is not too surprising to find that one of the ways that a cell can regulate translation is using another piece of RNA. Micro RNAs (miRNAs) were discovered as very short (~20 nucleotides) non-protein-coding genes in the nematode, C. elegans. Since their initial discovery (Lee et al, Cell 75: 843-54, 1993), hundreds have been found in various eukaryotes, including humans. The expression pattern of the miRNA genes is highly specific to tissue and developmental stage. Many are predicted to form stem-loop structures, and appear to hybridize to 3’-untranslated sequences of mRNA thus blocking initiation of translation on those mRNA molecules. They may also work through a mechanism similar to the siRNA discussed below, but there is clear evidence that mRNA levels are not necessarily altered by miRNA-directed translational control.

MicroRNAs are currently under investigation for their roles as either oncogenes or tumor suppressors (reviewed in Garzon et al, Ann. Rev. Med. 60: 167-79, 2009). Approximately half of known human miRNAs are located at fragile sites, breakpoints, and other regions associated with cancers (Calin et al, Proc. Nat. Acad. Sci. (USA) 101: 2999-3004, 2004). For example, miR-21 is not only upregulated in a number of tumors, its overexpression blocks apoptosis - a necessary step to allow abnormal cells to continue to live and divide rather than die out. Conversely, miR-15a is significantly depressed in some tumor cells, and overexpression can slow or stop the cell cycle, even inducing apoptosis.

Another mechanism for translational control that uses small RNA molecules is RNA interference (RNAi). This was first discovered as an experimentally induced repression of translation when short double-stranded RNA molecules, a few hundred nucleotides in length and containing the same sequence as a target mRNA, were introduced into cells. The effect was dramatic: most of the mRNA with the target sequence was quickly destroyed. The current mechanistic model of RNAi repression is that first, the double-stranded molecules are cleaved by an endonuclease called Dicer, which cleaves with over-hanging single-stranded 3’ ends. This allows the short fragments (siRNA, ~20nt long) to form a complex with several proteins (RISC, RNA-induced silencing complex). The RISC splits the double-stranded fragments into single strands, one of which is an exact complement to the mRNA. Because of the complementarity, this is a stable interaction, and the double-stranded region appears to signal an endonuclease to destroy the mRNA/siRNA hybrid.

The final method of controlling levels of gene expression is control after the fact, i.e., by targeted destruction of the gene product protein. While some proteins keep working until they fall apart, others are only meant for short-term use (e.g. to signal a short phase in the cell cycle) and need to be removed for the cell to function properly. Removal, in this sense, would be a euphemism for chopped up and recycled. The ubiquitin-proteasome system is a tag-and-destroy mechanism in which proteins that have outlived their usefulness are polyubiquitinated. Ubiquitin is a small (76 amino acids, ~5.6 kDa), highly conserved (96% between human and yeast sequences) eukaryotic protein (Figure (PageIndex{9})) that can be attached to other proteins through the action of three sequential enzymatic steps, each catalyzed by a different enzyme.

E1 activates the ubiquitin by combining it with ATP to make ubiquitin-adenylate, and then transfers the ubiquitin to itself via a cysteine thioester bond. Through a trans(thio) esterification reaction, the ubiquitin is then transferred to a cysteine in the E2 enzyme, also known as ubiquitin-conjugating enzyme. Finally, E3, or ubiquitin ligase, interacts with both E2-ubiquitin and the protein designated for destruction, transferring the ubiquitin to the target protein. After several rounds, the polyubiquitinated protein is send to the proteasome for destruction.

Mutations in E3 genes can cause a variety of human medical disorders such as the neurodevelopmental disorders Angelman syndrome, Hippel-Lindau syndrome, or the general growth disorder known as 3-M syndrome. Mechanisms linking malfunction in ubiquitination pathways and symptoms of these disorders are not currently known.

Proteasomes are very large protein complexes arranged as a four-layered barrel (the 20S subunit) capped by a regulatory subunit (19S) on each end. The two outer rings are each composed of 7 α subunits that function as entry gates to the central rings, each of which is composed of 7 β subunits, and which contain along the interior surface, 6 proteolytic sites. The 19S regulatory units control the opening and closing of the gates into the 20S catalytic barrel. The entire proteasome is sometimes referred to as a 26S particle.

A polyubiquitinated protein is first bound to the 19S regulatory unit in an ATP-dependent reaction (the 19S contains ATPase activity). 19S unit opens the gates of the 20S unit, possibly involving ATP hydrolysis, and guides the protein into the central proteolytic chamber. The protease activity of proteasomes is unique in that it is a threonine protease, and it cuts most proteins into regular 8-9 residue polypeptides, although this can vary.

As we will see in the cell cycle chapter, proteasomes are a crucial component to precise regulation of protein functions.


Translational regulation

Translational regulation refers to the control of the levels of protein synthesized from its mRNA. This regulation is vastly important to the cellular response to stressors, growth cues, and differentiation. In comparison to transcriptional regulation, it results in much more immediate cellular adjustment through direct regulation of protein concentration. The corresponding mechanisms are primarily targeted on the control of ribosome recruitment on the initiation codon, but can also involve modulation of peptide elongation, termination of protein synthesis, or ribosome biogenesis. While these general concepts are widely conserved, some of the finer details in this sort of regulation have been proven to differ between prokaryotic and eukaryotic organisms.


Ribosomal structure and the mechanism of translation

The crystal structures of ribosomes published in the past few years have revolutionized our understanding of the structural basis of tRNA selection and the peptide-bond-forming activity of the ribosome. The precise mechanisms of the distinct steps of protein synthesis are still unknown, however. This issue was addressed by several speakers, including Venki Ramakrishnan (MRC Laboratory of Molecular Biology, Cambridge, UK), who presented his recent work showing that the ribosome promotes accurate tRNA selection at the ribosomal A site and that recognition of cognate codon-anticodon interaction induces the 30S ribosome subunit to adopt a closed conformation. This movement most probably accelerates the rate of GTP hydrolysis and the following accommodation step, observed by other groups from kinetic analysis. Other presentations focused on structural rearrangements of the ribosome during elongation and translocation and, together, these structural data highlighted the dynamic nature of ribosome structure during the different steps of translation and prompted the audience to ponder which conformational changes are rate-limiting during translation.

Structural analysis of the eukaryotic ribosome when associated with translation factors has also brought new insights. In eukaryotes, initiation of translation is generally dependent on the presence of a 5' cap structure on the messenger RNA. Cap-dependent translation initiation is a complex process, facilitated by a large number of initiation factors (eIFs) that form a complicated network of cooperative interactions with the 40S ribosomal subunit. John McCarthy (Manchester Interdisciplinary Biocentre, UK) reported cryo-electron microscopy (cryo-EM) reconstructions, which indicate that binding of eIF1A to the 40S ribosomal subunit induces significant conformational changes in the subunit. These movements may create a recruitment-competent state of the 40S subunit that mediates the cooperative binding of other eIFs to form the 43S initiation complex. Moreover, the structure of the 43S complex indicates that the 40S to 43S transition involves a large rotation of the head of the small subunit this is thought to reflect the opening of the mRNA channel which, in turn, may facilitate mRNA binding and subsequent scanning.

The cap-independent pathway of translation initiation, utilized by both viral and cellular mRNAs, exploits highly structured translation-initiation regions on mRNAs dubbed internal ribosome entry sites (IRESs). The IRES from the cricket paralysis virus (CrPV) directly assembles elongation-competent ribosomes in the absence of the canonical eIFs and the initiator tRNA, methionyl-tRNAi. Eric Jan, from Peter Sarnow's group (Stanford University, USA) described experiments exploiting cryo-EM to visualize the CrPV-IRES bound to human 40S subunits and the 80S ribosome. The IRES was shown to form specific contacts with the components of the ribosomal A, P and E sites and to induce conformational changes in the ribosome. These changes were similar to those observed when the hepatitis C virus (HCV) IRES binds to the 40S subunit and when the elongation factor eEF2 binds to the 80S ribosome. This suggests that the CrPV IRES functions as an RNA-based translation factor that actively manipulates the ribosome to mediate the virus's unusual mode of translation initiation. Collectively, the structural data on the ribosome and its associated complexes presented during the meeting led the audience to an appreciation of the ribosome as a dynamic machine whose contortions are subject to the considerable influence of both regulatory proteins and RNA structures.


Results and discussion

Global translational control during environmental stress

To study translational changes in stress conditions, we prepared polysome profiles from unstressed cultures of fission yeast cells (control) and from the same cultures exposed for different times to oxidative stress, heat shock, and the DNA-damaging agent methylmethane sulfonate (MMS). The relatively mild doses of stress were chosen based on previous experiments to prevent substantial cell death [3, 7]. Severe stress conditions lead to general translational repression [11–13, 21, 22, 35]. We detected no substantial differences, however, between the overall polysome profiles from cells exposed to any of the stresses and unstressed control cells (data not shown), indicating that translation was not extensively altered on a global scale under the selected, relatively mild conditions [3, 7].

To analyze translational control, we extracted mRNA from four equal fractions throughout the polysome profiles, followed by labeling and hybridization onto DNA microarrays against labeled genomic DNA as reference (Figure 1a). To test whether translational profiles obtained from this medium-resolution approach reflected the data from high-resolution translation profiling using 12 fractions [36], profiles from the mRNAs with the highest and lowest ribosome occupancy were compared. There was good overall agreement between profiles from medium- and high-resolution translational profiling (Figure 1b): the mRNAs with the highest ribosome occupancy, corresponding to efficiently translated mRNAs, peaked in the higher fractions (red profiles in Figure 1b: fractions 3 to 4 in medium-resolution profiling fractions 7 to 12 in high-resolution profiling), while the mRNAs with the lowest ribosome occupancy peaked at the lower fractions (green profiles in Figure 1b). Although somewhat less sensitive than high-resolution profiling, this comparison indicates that four fractions are sufficient to capture the essence of the translational activity, providing more information than simply comparing monosome with polysome fractions.

Experimental layout and data analysis for translation profiling. (a) Four equal mRNA fractions were each competitively hybridized to DNA microarrays against genomic DNA as common reference. Polysome profiles were prepared from unstressed control cells and from cells exposed to stress for 5, 15 and 60 minutes. (b) Comparison of polysome profiles for medium-resolution translation profiling with high-resolution profiling applied before [36]. The graphs show polysome profiles of the 10% mRNAs with the lowest ribosome occupancy (green) and the 10% mRNAs with the highest ribosome occupancy (red), measured by medium-resolution (left graph) and by high-resolution translation profiling (right graph). The profiles represent the average from three independent biological repeats. Ribosome occupancy was determined based on previous high-resolution translation profiling [36]. (c) Two complementary data analysis approaches to uncover translationally regulated mRNAs (Materials and methods). Left: the total difference for a given mRNA between the translation profile under stress and the translation profile in the control was calculated by summing up the differences of each fraction (indicated by arrows). Right: scores for each mRNA in each condition were calculated as described in Materials and methods. A translation ratio was then obtained by dividing the score of a given mRNA in a stress condition by the score of the same mRNA in the control condition (see Additional file 1 for data from translational profiling analysis). The combined lists from both approaches were then visually inspected to generate high-confidence, curated lists of translationally regulated genes (Table 1). (d) Left graph: translation profiles for sod1mRNA before and after exposure to 39°C. Right graph: translation profiles for sks2mRNA before and after exposure to H2O2. Blue lines, control (unstressed) samples orange lines, 15 minutes after stress red lines, 60 minutes after stress. Multiple lines of the same color represent translation profiles from independent biological repeats.

To identify mRNAs with altered translational profiles among the nuclear-encoded protein-coding genes, we initially applied a combination of two complementary automated approaches: 1) using a measure of the overall difference in mRNA profiles between stress and control samples and 2) using a ratio of weighted translation scores between stress and control samples (Figure 1c Additional file 1). These two methods gave largely (approximately 90%) overlapping yet complementary results, with the first one informing about overall differences and shifts in translational profiles and the second one informing about the levels and directions of translational changes. The profiles from these candidate mRNAs uncovered by either approach were then visually inspected to create a high-confidence set of translationally regulated mRNAs. Two typical examples of mRNAs showing translational control in response to stress are shown in Figure 1d: the sod1 mRNA (encoding a superoxide dismutase) is gradually shifted towards higher polysomal fractions in response to heat stress, reflecting translational up-regulation, while the sks2 mRNA (encoding a ribosome-associated molecular chaperone) is strongly shifted from the higher polysomal fractions towards fractions of free mRNA in response to oxidative stress, reflecting translational down-regulation. Table 1 shows the numbers of translationally regulated mRNAs in the different conditions, before and after filtering by visual inspection. Below, we will refer to the high-confidence data set as translationally regulated mRNAs.

Changes in mRNA levels and translation are globally coordinated

We detected 757 mRNAs that showed translational regulation during exposure to oxidative stress (at 60 minutes), and 364 mRNAs that showed such changes in heat stress (at 15 minutes), whereas only 87 mRNAs showed translational regulation after the exposure to DNA damage (Table 1 Additional file 2). Notably, the translationally up-regulated mRNAs for all three stress conditions were significantly enriched in up-regulated CESR genes (P

9 × 10 -55 to 2 × 10 -22 ), while translationally down-regulated mRNAs were significantly enriched in down-regulated CESR genes (P

3 × 10 -173 to 5 × 10 -80 ) [3, 7]. These substantial overlaps between CESR genes and translationally regulated genes suggest that mRNAs regulated at the level of transcription are often also regulated in the same direction at the level of translation.

To directly compare the regulation of mRNA abundance with the regulation of translation, we also measured changes in relative mRNA levels by expression profiling of the same cell samples used for translational profiling. These mRNA expression data were highly similar to those previously described [3]. Table 1 shows the numbers of mRNAs whose levels substantially change in the various stress conditions, compared to the numbers of translationally regulated mRNAs. The response to heat was more rapid than the response to oxidative stress, which is reflected in both translational and mRNA up-regulation peaking at 15 minutes in the former and at 60 minutes in the latter stress. The response to DNA damage was relatively small compared to the other two stresses, both at the translation and mRNA levels. The down-regulation of translation was much more pronounced in oxidative stress than in the other stresses.

We then clustered the profiles of all the mRNAs that showed substantial changes in abundance and/or translation in the different stress conditions (Figure 2 Table 1). This analysis highlights the overall coordination between mRNA and translation profiles, which were typically regulated in the same direction. Accordingly, the average translation and expression ratios for the regulated mRNAs were significantly positively correlated in all stress conditions tested (Figure 3). Taken together, along with results on transcriptional control for these genes [3] (S Marguerat, K Lawler, A Brazma and JB, submitted), these data strongly support the idea that, during environmental stress in fission yeast, most mRNAs are regulated both at the level of transcription and at the level of translation in a concordant manner. Moreover, regulation at the level of mRNA turnover during oxidative stress is also globally coordinated with transcription (S Marguerat, K Lawler, A Brazma and JB, submitted). It is not known how such global coordination at multiple levels of gene expression is achieved by the cells.

Clustering of genes regulated during environmental stress. (a) Hierarchical cluster analysis with columns representing experimental time points and rows representing the 1,355 genes that were regulated after exposure to H2O2, at the level of either mRNA and/or translation as defined in Table 1. Columns 1 and 2: mRNA expression levels after 15 and 60 minutes in stress, respectively, relative to expression in the same cells before stress are color coded as indicated at the bottom. Columns 3 and 4: translation ratios (Figure 1c Additional file 1 Materials and methods) after 15 and 60 minutes in stress, respectively, are color coded as indicated at the bottom. Data from biologically repeated samples are averaged, with missing data in gray. The red bars indicate clusters 1 and 2 described in the text. The graphs at right show average translation profiles for the genes of clusters 1 and 2 before and after exposure to H2O2. Blue lines, control (unstressed) samples orange lines, 15 minutes after stress red lines, 60 minutes after stress. Multiple lines of the same color represent average translation profiles from independent biological repeats. Annotated gene lists of these clusters are provided in Additional file 3. (b) Cluster analysis as in (a) for the 1,071 genes that were regulated after exposure to 39°C, at the level of either mRNA and/or translation (Table 1). The red bar indicates cluster 3 described in the text. The graph at right shows the average translation profiles for cluster 3 genes before as well as 15 and 60 minutes after exposure to 39°C, with details as in (a). An annotated gene list of cluster 3 is provided in Additional file 3. (c) Cluster analysis as in (a) for the 233 genes that were regulated after exposure to DNA damage, at the level of either mRNA and/or translation (Table 1).

Correlation between mRNA and translational regulation. Scatter plots showing linear regressions of the average log2 changes in the oxidative (top), heat (middle), and MMS (bottom) stress time-course experiments. The gene lists were the same as those used for clustering in Figure 2 for the three stress experiments. The Pearson's correlations and probabilities (two-tailed test) are indicated for each experiment.

Genes differentially regulated at mRNA and translational levels

Although transcriptional and translational regulation were generally coordinated, we detected substantial gene groups that opposed this global trend (indicated as clusters 1 to 3 in Figure 2). The 74 genes of cluster 1 were translationally up-regulated in response to oxidative stress, most notably at 60 minutes, while their mRNA levels were down-regulated (Figure 2a Additional file 3). No strong enrichment for functional categories was evident, apart from an overlap with genes that are strongly periodically regulated during the cell cycle (P

4 × 10 -7 ) [37, 38]. Most of the 16 overlapping genes function in mitosis or cell division and could be important for stress recovery to re-start the cell proliferation after H2O2-induced arrest.

The 123 genes of cluster 2 showed the reverse trend to cluster 1 genes: they were translationally down-regulated, while their mRNA levels were up-regulated (Figure 2a Additional file 3). Cluster 2 genes were enriched for the Gene Ontology (GO) [39] terms 'oxidoreductase activity' (P < 1 × 10 -8 ) and 'amino acid biosynthesis' (P

3 × 10 -10 ). Cluster 2 was also enriched for genes highly expressed at transcriptional and translational levels in unstressed cells [36, 40]: they showed higher mRNA levels (P < 1 × 10 -8 ), higher RNA polymerase II occupancy (P

8 × 10 -3 ), longer mRNA half-life and polyA tails (P

7 × 10 -3 and 2 × 10 -4 , respectively) as well as higher ribosome occupancy and density (P < 1 × 10 -8 and 4 × 10 -3 , respectively) compared to all mRNAs in unstressed cells. Shorter mRNAs are more efficiently translated in unstressed cells [13], but cluster 2 genes were not biased with respect to mRNA size. It is possible that the antagonistic translational down-regulation, which is maximal at 60 minutes after stress induction, balances protein production of these highly expressed genes for eventual stress recovery.

The 208 genes of cluster 3 showed a transient boost in translation in response to heat, but their mRNA levels were all down-regulated (Figure 2b Additional file 3). Cluster 3 was strongly enriched for genes encoding ribosomal proteins (P

3 × 10 -147 ). In contrast, in response to oxidative stress, genes for ribosomal proteins showed similar translation profiles to unstressed cells at 15 minutes before becoming strongly down-regulated at the translational level at 60 minutes. Cell growth and proliferation are tightly linked to ribosome biogenesis [41] the translational up-regulation of ribosomal protein genes at 15 minutes in heat stress could therefore reflect a transient boost in growth in response to the shift from 32°C to 39°C, as it takes some time to reach temperature equilibrium and Schizosaccharomyces pombe shows the fastest growth at approximately 35°C. At 60 minutes in heat stress, however, most of the cluster 3 genes became translationally down-regulated (Figure 2b), probably reflecting subsequent stalling of growth at 39°C. Intriguingly, a minority of seven ribosomal protein genes did not become transiently induced at the translational level in heat (rpl301, rpl302, rpl401, rpl402, rpl501, rps401, rps403), which might reflect functional specialization of different ribosomal proteins as suggested for budding yeast [42].

Concordant changes at mRNA and protein levels for induced but not for repressed genes

Given that changes in mRNA levels and in translation were largely coordinated, we expected that changes in protein levels mostly reflect changes in mRNA levels. We applied a proteomics approach to determine the relative changes in protein levels at multiple times after addition of oxidative stress compared to unstressed cells (Figure 4). The same samples were also interrogated with microarrays for mRNA expression profiling. Two independent biological repeats were performed for protein and mRNA profiling. We could obtain spectrum count data for 4,644 S. pombe proteins in at least one sample, and could detect 2,147 proteins in all 12 samples of both repeats (minimum of two identified unique peptides per protein Additional file 4). Of these 2,147 proteins, 234 (11%) showed significant changes in abundance during the stress time course (Materials and methods).

Proteome profiling during oxidative stress. Scheme delineating the experimental procedures applied to measure protein levels at different time points immediately before and after exposure to H2O2. Cells were harvested at the indicated time points, followed by preparation of protein lysates and digestion of proteins into peptides. Peptides were separated by strong ion exchange SCX chromatography into 24 fractions. Each fraction was separated by reversed phase chromatography and directly eluted into a ThermoFisher LTQ Orbitrap mass spectrometer (for details see Materials and methods). Transcript levels were determined from the same samples using DNA microarrays. Proteins and mRNAs were measured in two independent biological repeats of the time-course experiment, and quantified using QTools [60]. MS/MS, tandem mass spectrometry.

Figure 5a shows the expression profiles of all genes whose mRNA abundance was regulated during oxidative stress and whose proteins could be detected in all 12 samples. The data for transcriptome profiling in the two experiments (mRNA1 and mRNA2, performed with different microarray platforms and in different laboratories) were highly similar overall. Moreover, changes in mRNA and translation profiles were largely mirrored by changes in protein profiles, especially for up-regulated transcripts (Figure 5a). The inverse analysis, starting from proteins whose profiles significantly changed during oxidative stress, showed a similar picture of highly concordant up-regulation at multiple regulatory levels (Figure 5b). While the genes up-regulated at the mRNA level produced up-regulation of the corresponding proteins, the down-regulated mRNAs showed much weaker relationships with protein profiles (Figure 5a,b). Accordingly, the linear correlation between maximal mRNA and protein changes was highly significant for up-regulated mRNAs, while there was no correlation for down-regulated mRNAs (Figure 5c). This pattern was also evident from the average mRNA and protein expression profiles (Figure 5d).

mRNA, translation, and protein regulation during oxidative stress. (a) Hierarchical cluster analysis with columns representing experimental time points and rows representing 811 genes whose mRNAs showed significant expression changes after exposure to H2O2 and whose proteins could be detected in all conditions. mRNA1: mRNA expression relative to the unstressed samples is color coded as indicated at the bottom, using same samples as for the translation experiment (Figure 2). Translation: translational efficiency relative to the unstressed samples is color coded as indicated at the bottom. mRNA2: mRNA expression relative to the unstressed samples is color coded as indicated at the bottom, using the same samples as for the proteome experiment. Protein: protein expression relative to the unstressed samples is color coded as indicated at the bottom. Average data of biological repeats are shown, with missing data in gray. (b) Cluster analysis as in (a) for the 232 genes that encode proteins showing significant changes in expression in the proteome experiment and with data in >50% of all conditions used for clustering. (c) Scatter plot showing linear regressions of the maximum average log2 changes in mRNA and protein expression across the time-course experiment mRNA2/Protein shown in (a) for proteins that were detected in all conditions. Yellow dots, 193 genes showing >1.5-fold induction in mRNA expression after exposure to H2O2 in at least 4 of 7 stress time points in experiments mRNA1 and mRNA2 shown in (a) blue dots, 226 genes showing >1.5-fold repression in mRNA expression after exposure to H2O2 in at least 4 of 7 stress time points in experiments mRNA1 and mRNA2. Pearson's correlations and probabilities (two-tailed test) are indicated for induced (yellow) and repressed (blue) mRNAs. (d) Graph showing average mRNA (green) and protein (red) expression profiles (log2 ratios) in the experiment mRNA2/Protein shown in (a). Solid and dashed lines indicate average profiles for 193 and 226 genes showing >1.5-fold induction or repression, respectively, in mRNA expression after exposure to H2O2 in at least 4 of 7 stress timepoints in experiments mRNA1 and mRNA2 shown in (a).

Although we limited our analysis to those proteins that were detectable in all conditions, the poor correlation between down-regulated mRNAs and corresponding proteins could reflect that low abundance proteins are less reliably quantified by mass spectrometry. On the other hand, it is plausible that this poor correlation reflects the biological reality of protein regulation, as proteins with long half lives are expected to maintain stable expression for some time after shutting down production. Notably, similar findings were recently reported by Lee et al. [33] during osmotic stress in budding yeast. These authors applied mathematical modeling to their transcriptome and proteome data sets, suggesting that reduction in transcript abundance may serve to redirect ribosomes to newly produced mRNAs.

The dynamic range for regulation of mRNA abundance was substantially larger than for regulation of protein abundance (Figure 5a,b,d). Similar results were obtained in a recent study in budding yeast applying a different proteomics approach [33]. Moreover, while most up-regulated mRNAs transiently peaked in expression at 60 minutes after stress induction and then decreased again, the corresponding proteins showed a delayed and gradual increase in expression up to 180 minutes (Figure 5a,b,d). The delayed up-regulation of proteins may reflect the time required for translation along with protein half-lives. Similar mRNA and protein expression patterns have recently been observed in budding yeast [33, 43]. Lee et al. [33] have shown that the 'burst' in mRNA expression serves to accelerate protein expression before mRNA levels adjust to maintain a new steady-state.

Overall, the average changes in protein expression were substantially correlated with corresponding changes in mRNA expression and, to a lesser extent, with changes in translation (Figure 6). Factoring in both transcription and translation by simply multiplying relative mRNA and translational changes, however, did not further improve the correlation with protein expression (Figure 6). The global correlation between mRNA and protein expression we observed here is stronger than the more modest correlations reported in previous papers [24–32], but is comparable to higher correlations reported in some smaller-scale studies [44, 45] and recent global studies [33, 34].

Relationships between mRNA, translation, and protein regulation. Scatter plots showing linear regressions of the average log2 changes of proteins and of average changes in mRNA2 experiment (top Figure 5), average changes in translation (middle), and combined changes in mRNA2 and translation (bottom product of average changes). The gene list is the same as used in Figure 5a. The Pearson's correlations and probabilities (two-tailed test) are indicated for each comparison.

To explore any effects of post-transcriptional regulation during oxidative stress, we identified genes that went against the overall trend of concordant regulation at mRNA and protein levels. Only 22 and 19 proteins showed increased and decreased expression, respectively, in the absence of changes in mRNA expression (Additional file 5). Translation of the corresponding mRNAs was not strongly regulated, however. These data raise the possibility of additional regulation at the protein level, by stress-induced protein stabilization or degradation. A higher number of proteins did not show significant expression changes, although the corresponding mRNAs showed increased or decreased expression (63 and 156 mRNAs, respectively Additional file 5). It is likely that the latter reflect the overall limited correlation between down-regulated mRNAs and protein expression discussed above (Figure 5). However, some of these discrepancies could also be explained by compensatory regulation at the level of translation, especially for the mRNAs with increased expression: 16 of the 63 up-regulated mRNAs were translationally down-regulated (cluster 2 genes in Figure 2a), but only 2 of the 156 down-regulated mRNAs were translationally up-regulated (cluster 1 genes in Figure 2a). These patterns suggest that, in some cases, translation is regulated to counter the stress-induced changes in mRNA expression so that the resulting protein expression does not substantially change. It is possible that the proteins encoded by these mRNAs are not immediately required under the given condition but are prepared at the mRNA level to become rapidly available on short notice ('translation on demand') [46]. The majority of cluster 1 and 2 genes showed changes in protein expression consistent with the changes in mRNA expression, indicating that, in most cases, transcriptional regulation dominates translational regulation. This conclusion is supported by the findings that mRNA expression correlated better than translational changes with protein expression, and factoring in translation did not improve the overall correlation (Figure 6). Finally, a few proteins showed abundance changes in the opposite direction to their mRNAs: 20 proteins increased and 12 proteins decreased in expression, while the corresponding mRNAs decreased and increased, respectively (Additional file 5). While these few exceptions could reflect technical noise, it was striking that several ribosomal proteins and translation factors showed such opposite regulation. This finding raises the possibility of some remodeling of the translation machinery via regulation of protein stability and the involvement of specialized ribosome subunits during stress [42].


Biology 10/8/2020

a. DNA encodes information that is transcribed into RNA, and RNA encodes information that is translated into proteins.

b. DNA encodes information that is translated into RNA, and RNA encodes information that is transcribed into proteins.

c. Proteins encode information that is used to produce other proteins of the same amino acid sequence.

b. The mutation occurs only in the germ line cells and is passed on to all of the individual's offspring.

c. The mutation occurs only in the germ line cells and is passed on to half of the individual's offspring.

a. toxic metabolic products produced by cellular metabolism

b. spontaneous changes in nucleotide structure

c. errors in DNA replication

d. transposon insertions within genes

a. A mutation caused by a chemical mutagen that causes a thymine base to pair with another thymine instead of adenine

b. The new base that results when a thymine base is alkylated by a chemical mutagen

c. A site where two adjacent thymine bases become covalently cross-linked to each other

d. A location in a DNA molecule where there are two thymine bases in a row


Watch the video: Translation Regulation (January 2023).