Deciphering translation - a genome-wide analysis of translation in Escherichia coli

Protein synthesis is one of the central elements in every living cell. For this process, mRNAs coding for genes are simultaneously competing for the same translation machinery (ribosomes and amino acids). But the ultimate determinants of cellular functions are the proteins, a good understanding of this process is therefore of crucial importance. For this purpose, we developed a genome-wide, mechanistic, mathematical model of translation, taking into account individual codon kinetics. It is used to quantify the rate limiting steps of translation and gain insight into the mechanism of regulation of expression between different proteins. We observe for each gene in a first low-occupancy phase a nearly linear increase in protein synthesis rate with amount of bound ribosomes to the mRNA, followed by a plateau of maximum synthesis rate. A further increase in the ribosomal coverage of the mRNA leads to a decrease in protein production rate, and might act as sinks for system resources like ribosomes. Native E. coli gene sequences are also compared to randomly obtained synonymous sequences, coding the same protein, and it is found some genes of higher significance for the cell have been better optimized for protein synthesis rate than other.

Deciphering Translation - a Genome-Wide Analysis of Translation Limitations and Competitions in Saccharomyces cerevisiae

Vassily Hatzimanikatis, Julien Racle

Protein synthesis is one of the central elements in every living cell. For this process, the mRNAs coding for genes are simultaneously competing for the same translation machinery, i.e. for the same ribosomes and the same protein building blocks, the amino acids. But the ultimate determinants of cellular functions are the proteins, it is therefore of crucial importance to understand well this translation process. Nowadays, it is possible to measure the changes in the levels of all mRNAs and proteins in the cell, under different environmental and genetic perturbations. The large amount of data from transcriptomic and proteomic studies requires development of efficient systems-wide mathematical and computational frameworks to analyze and interpret the data. For this purpose, we developed a genome-wide, mechanistic, mathematical model for translation in S. cerevisiae. This model includes information on the translation state of each mRNA (i.e. polysome size), and mRNA length, and can be used to estimate genome-scale kinetics of the protein synthesis machinery. It is then used to quantify the rate limiting steps of the translation process and gain insights into the mechanism of regulation of expression between different proteins. This framework allows us therefore to quantify the effect of a change in the total number of ribosomes, or also the effect of varying the level of a single mRNA on the various proteins. This information is then used to determine functional couplings in the yeast genome. We find that while 95% of the mRNA species are initiation and elongation limited, approximately 5% are elongation limited and have higher specific protein synthesis rates. We also observed that the mRNAs with ribosomal coverage greater than 0.8 display a decrease in the specific protein production rate with increasing coverage, and can be acting as sinks for system resources like ribosomes. It was also possible to determine which mRNA species have the most / least influence on the other, and we found that the most influential species are linked to proteins which form structural constituents of ribosomes, proteins related to translation elongation factor activity or key enzymes of the central carbon metabolism. Thus a perturbation to these key mRNA species has the maximum impact on protein synthesis from other mRNA species. Our results thus have implications in the understanding of the translation machinery response to perturbations and of the global systems response.

2010

A Novel Pulse-Chase SILAC Strategy Measures Changes in Protein Decay and Synthesis Rates Induced by Perturbation of Proteostasis with an Hsp90 Inhibitor

Vassily Hatzimanikatis, Céline Hernández, Julien Racle

Standard proteomics methods allow the relative quantitation of levels of thousands of proteins in two or more samples. While such methods are invaluable for defining the variations in protein concentrations which follow the perturbation of a biological system, they do not offer information on the mechanisms underlying such changes. Expanding on previous work [1], we developed a pulse-chase (pc) variant of SILAC (stable isotope labeling by amino acids in cell culture). pcSILAC can quantitate in one experiment and for two conditions the relative levels of proteins newly synthesized in a given time as well as the relative levels of remaining preexisting proteins. We validated the method studying the drug-mediated inhibition of the Hsp90 molecular chaperone, which is known to lead to increased synthesis of stress response proteins as well as the increased decay of Hsp90 “clients”. We showed that pcSILAC can give information on changes in global cellular proteostasis induced by treatment with the inhibitor, which are normally not captured by standard relative quantitation techniques. Furthermore, we have developed a mathematical model and computational framework that uses pcSILAC data to determine degradation constants kd and synthesis rates Vs for proteins in both control and drug-treated cells. The results show that Hsp90 inhibition induced a generalized slowdown of protein synthesis and an increase in protein decay. Treatment with the inhibitor also resulted in widespread protein-specific changes in relative synthesis rates, together with variations in protein decay rates. The latter were more restricted to individual proteins or protein families than the variations in synthesis. Our results establish pcSILAC as a viable workflow for the mechanistic dissection of changes in the proteome which follow perturbations. Data are available via ProteomeXchange with identifier PXD000538.

Public Library of Science2013

Circadian clock- and feeding-dependent regulation of translation initiation and elongation in the liver

Cédric Gobet

Circadian rhythms allow organisms to anticipate and adapt to periodic environmental changes related to the day-night cycle. These rhythms originate from a genetic network ticking in almost every cell of our body and the coordination between different organs, thus allowing to generate physiological outputs at the right time of the day. In the liver, which is one of the major metabolic organ of the body, gene expression has to be highly regulated by the circadian clock and feeding rhythms to generate such a timely behavior. Indeed, at every step from transcription to post-translational modifications, the circadian clock and metabolism interplay. In this context, we quantified transcription, mRNA accumulation and translation in liver from wild-type and clock-deficient mice fed ad libitum or only during the night period. RNA-Sequencing and ribosome profiling combined with mathematical modeling were used to assess rhythmic regulation along the gene expression process. Transcription was found to be the main driver of temporal mRNA accumulation and its subsequent translation. We observed a diurnal translation efficiency in genes with 5'-Terminal Oligo Pyrimidine tract (5'-TOP) sequences or with Translation Initiator of Short 5'-UTR (TISU) motifs. Remarkably, the rhythmic translation efficiency was mainly driven by feeding cues although clock-deficiency slightly affected the amplitude and phase of the oscillations. In a subsequent work, we emphasized the above analysis of translation regulation with a focus on the ribosome elongation process. A statistical model and bioinformatics pipeline were developed to infer codon dwell times from ribosome profiling data. The method has the particularity of integrating pairs of codons between the sites of the ribosome and to fit every sequencing read on an individual basis. We validated our approach on a yeast dataset and then studied extensively the mouse liver translation elongation landscape. A large dynamic range between the fastest and longest codons was uncovered in the ribosome A and P sites with strong synergic interaction of pairs of codons clustering by amino acids. In order to challenge our system, we performed drug free ribosome profiling, modified to reduce library biases, in liver of mice fed ad libitum or starved. Unexpectedly, codon dwell times were conserved between the two conditions despite a gene expression landscape reflecting low energy state of the mice and subsequent amino acid shortage. Finally, to understand how the ribosome elongation velocity is regulated by (aminoacyl)-tRNA levels, we adapted and modified a published DNA hybridization-based tRNA profiling method in liver of mice fed ad libitum or starved. While the quantified tRNAs abundance showed a large dynamic range and specific isoacceptors patterns, no differences in (aminoacyl)-tRNAs were found between the two feeding conditions. Nevertheless, tRNAs for isoleucine, asparagine, aspartate and arginine exhibited a relative low loading state compared to other tRNA in both conditions. Altogether, correlation between codon usage, codon dwell times and tRNA abundances provided insights of how translation elongation is regulated in mammals. Together, this thesis aimed to dissect the mechanisms regulating gene expression by the interplay of the circadian clock and metabolism, as well as understand how translation elongation is regulated in mammals.

EPFL2018