Cut in translation: ribosome‐dependent mRNA decay

Abstract

Transcription and translation are two complex mechanisms that are tightly coupled in prokaryotic cells. Even before the completion of transcription, ribosomes attach to the nascent mRNA and initiate protein synthesis. Remarkably, recent publications have indicated an association between translation and decay of certain mRNAs. In this issue of The EMBO Journal, Leroy et al (2017) depicts a fascinating mechanism of mRNA degradation, which involves the ribosome‐associated ribonuclease Rae1 in Bacillus subtilis. In a translation‐dependent manner, Rae1 binds the ribosomal aminoacylation (A)‐site and cleaves between specific codons of the targeted mRNA.

During their lifetime, all RNA molecules undergo numerous cleavages leading to their maturation and, ultimately, degradation. There is a large variety of ribonucleases (RNases) responsible for these modifications, which depend on the nature of the RNA substrates (e.g. endonuclease/exonuclease, single‐ or double‐stranded specific RNase). No less than 21 RNases are present in the Gram‐negative model organism Escherichia coli (Mackie, 2013) and 19 in the Gram‐positive bacterium Bacillus subtilis (Durand et al, 2015). RNases are both scavengers of useless ribonucleic acids and key players in essential processing events and main regulatory systems.

In this issue of The EMBO Journal, Leroy et al (2017) investigate the role of an uncharacterized ribonuclease, YacP, renamed Rae1 for ribosome‐associated endonuclease 1 in B. subtilis. Rae1 is a ribosomal aminoacylation (A)‐site RNase that specifically cleaves the yrzI polycistronic transcript in the coding sequence of the S1025 ORF, encoding a 17‐aa‐long peptide. Intriguingly, Rae1 cleaves the S1025 mRNA in the course of active translation in a ribosome‐dependent manner. This ribosome‐dependent cleavage happens through a hydrolytic mechanism specifically upstream of Lys14 (AAG) codon (Fig 1A).

Figure 1. Illustration of ribosome‐associated RNases.

(A) During translation of certain mRNAs, the endonuclease Rae1 enters in the aminoacylation (A)‐site of B. subtilis ribosome. Here, Rae1 induces a cleavage upstream of a Lys(AAG) or Lys(AAA) codon. However, inducing signals remain unknown. (B) RelE is a type II toxin, component of the RelE/RelB toxin–antitoxin system found in numerous bacteria including E. coli. Under normal conditions, RelB antitoxin sequesters RelE and prevents toxicity. Upon nutritional stress, RelB is degraded and RelE reaches the ribosomal A‐site to induce a cleavage between the second and third nucleotides (generally a purine) of the targeted codon. For example, the CAG codon is strongly recognized by RelE. Here, the 16S rRNA plays a crucial role for RelE activity by reorienting correctly the mRNA for the subsequent cleavage reaction. (C) Bacteria mainly use the trans‐translation system to prevent toxic effect of non‐stop translation complexes. This system is composed of the transfer‐messenger RNA (tmRNA) and the SmpB protein, forming a ribonucleoprotein complex. This complex enables ribosome recycling, proteolysis of the nascent polypeptide and degradation of the aberrant mRNA through the recruitment of the 3′‐5′ exoribonuclease RNase R. Even if the mechanism remains elusive, ribosome‐associated RNase R presumably promotes mRNA degradation immediately after its release from the rescued ribosome.

Additional ribosome‐dependent RNases were previously described. They principally belong to the type II toxins of the RelE family (e.g., RelE, YoeB, YafQ, and HigB), scattered throughout the prokaryotic kingdom (Goeders et al, 2013). Toxin–antitoxin (TA) systems are categorized as six distinct groups (types I to VI) depending on the mechanism used to neutralize the toxin (Page & Peti, 2016). In type II TA systems, the antitoxin is a labile protein that tightly binds and inactivates the stable toxin to prevent any lethal effect.

In E. coli, at least four type II TA systems contain a ribosome‐dependent RNase. A well‐characterized example is the RelE/RelB toxin–antitoxin system. In this case, RelB protein sequesters the RelE toxin, thwarting its activity (Fig 1B). Upon nutritional stress, RelE is released from this complex due to the degradation of RelB by the Lon protease. Then, RelE‐dependent cleavage of targeted mRNAs rapidly blocks bacterial growth and induces cell persistence. Type II toxins of the RelE family generally cleave targeted mRNAs between the second and third nucleotides of the codon localized in the ribosomal A‐site (Goeders et al, 2013).

Components of type II TA systems are usually encoded in the same operon. Because rae1 is part of a large operon encoding translation‐related proteins, Rae1 is unlikely to have an antitoxin partner. However, according to Prof. Kenn Gerdes, “In principle, a partner antitoxin could be encoded anywhere on a chromosome” (Gerdes, 2000). Thus, it cannot be excluded that Rae1 could be part of a toxin–antitoxin system.

While all examples described above require ribosomes for their specific activity, there are several differences among ribosome‐dependent RNases: weak sequence similarity, distinct interaction sites on the ribosome (e.g., rRNA and/or ribosomal proteins), and canonical or non‐canonical cleavage mechanism. What is more, RNases recognize a variety of codon sequence within the ribosomal A‐site. Although RelE and YoeB preferentially cleave upstream of purines, HigB and YafQ generally target AAA codons (Goeders et al, 2013), which is reminiscent of Rae1's cleavage site. Indeed, Rae1 cleaves optimally at AAA and AAG lysine codons.

However, the Rae1‐specific cleavage was determined using a single substrate, the yrzI mRNA. Leroy et al (2017) could not extend this model to all 46 mRNAs identified as potentially targeted by Rae1, casting doubt on the existence of a specific consensus sequence. This is no surprise, as the substrate specificity of many RNases is still poorly understood. Determination of substrates and recognition sequences are rarely evident. Even the well‐characterized prokaryotic RNase E lacks a clear sequence motif and preferentially targets unstructured A/U‐rich regions (Mackie, 2013). Recent innovative experimental approaches such as RNase co‐immunoprecipitation (Waters et al, 2017) or global profiling of cleavage products (Chao et al, 2017) could settle this challenge.

More RNases, and not only TA‐related RNAses, are known to associate with the ribosome. Indeed, the exoribonuclease RNase R specifically binds to stalled ribosomes (Domingues et al, 2015). Between 2 and 4% of translating ribosomes stall and need to be rescued (Keiler, 2015). Generally, stalling occurs when a ribosome remains trapped at the end of an mRNA resulting in the formation of a non‐stop translation complex. To rescue stalled ribosomes, bacteria mainly use a translation quality control system called trans‐translation system. This system is composed of the hybrid transfer‐messenger RNA (tmRNA) and the small protein B (SmpB), which enable the recycling of ribosomes. Then, both the truncated polypeptide and mRNA are subject to degradation. In this example, the decay of defective mRNAs is notably induced by RNase R, which is recruited by the trans‐translation system (Fig 1C). It is tempting to speculate that tmRNA and trans‐translation could be involved in rescuing mRNAs cleaved by Rae1.

Interestingly, most of the putative Rae1 mRNA substrates code for proteins involved in stress responses (e.g., iron starvation, nutritional stress), suggesting a role in stress management for Rae1. A direct role for type II TA systems has also been established in cell persistence upon stressful conditions (Page & Peti, 2016). Hence, they could act as “adaptation enzymes”, each responding to specific environmental stimuli (Neubauer et al, 2009). Revealing the whole regulon of ribosome‐dependent mRNA decay will certainly help deepen the physiological purpose of this captivating mechanism.

Altogether, this report sheds more light on the relationship between translation and mRNA decay in bacteria. Hence, from its synthesis to its degradation, each event occurring in the lifetime of an mRNA could be coupled, suggesting a concealed synergy in gene expression mechanisms.

See also: M Leroy et al (May 2017)