Death by translation: ribosome-assisted degradation of mRNA by endonuclease toxins

Like other forms of life, bacteria utilize a variety of mechanisms to survive stress. Some of these mechanisms mitigate the physiological impact of the stress, whereas others enable cells to simply wait out the stressful condition. Among the cellular mediators of the latter response are toxin-antitoxin (TA) systems [1]. When activated in response to stress, TA systems cause bacterial cells to enter a dormant state in which growth ceases and their metabolic needs are greatly diminished. Subsequently, when more favorable conditions return, the cells re-awaken and resume growth. Mechanisms of this kind have also been implicated in antibiotic tolerance, in which a small fraction of an otherwise susceptible bacterial population manages to survive exposure to an antibiotic by hibernation, repopulating the host when the antibiotic treatment ends [2].

TA systems are widespread throughout the bacterial realm, where almost all species encode several of them in discrete genetic modules from which a toxin and its cognate antitoxin are both expressed [3,4]. Among the most common are those that comprise a ribonuclease (toxin) and a protein inhibitor (antitoxin) that sequesters the ribonuclease in a catalytically inactive complex [5]. Upon exposure to stress, selective degradation of the antitoxin unleashes the toxin, whose catalytic action forces the cell into a dormant state. Some toxins of this kind (e.g., MazF, ChpB, MqsR, and HicA in Escherichia coli) are sequence-specific endonucleases that choose their RNA targets rather indiscriminately [6–9]. Others (e.g., RelE, YoeB, HigB, YafO, and YafQ in E. coli) are activated by ribosome binding, which causes them to preferentially target ribosome-associated transcripts [10–12]. By selectively degrading easily replenishable mRNAs and avoiding rRNAs and tRNAs, toxins of the latter variety cause less damage, which in principle should make it easier for cell growth to resume once the stress ends.

On the basis of bioinformatic analysis, Lupas and co-workers proposed several years ago that the E. coli prlF-yhaV operon encodes another such TA system conserved in many proteobacterial species [13]. They went on to show that YhaV is indeed a ribonuclease that ordinarily is inhibited by complex formation with PrlF but can arrest cell growth when produced in excess of its antitoxin [14]. Unexpectedly, despite statistically significant sequence similarity to RelE, purified YhaV appeared capable of degrading RNA in the absence of ribosomes.

In this issue of FEBS Letters, Choi et al. revisit the question of whether the activity of YhaV is ribosome-independent and present multiple lines of contrary evidence indicating that its endonuclease activity is in fact potentiated by ribosome binding [15]. In their hands, purified E. coli YhaV cleaves RNA only in the presence of ribosomes. Its activation appears to be specific for bacterial ribosomes, as it is able to abort mRNA translation in vitro when added to a cell-free system derived from E. coli but not one derived from rabbit reticulocytes. Immunoblot analysis of polysomes extracted from E. coli and fractionated on sucrose gradients reveals that YhaV co-sediments with 70S ribosomes and 50S ribosomal subunits, consistent with an affinity for E. coli ribosomes. Furthermore, when selectively overproduced in E. coli, YhaV rapidly degrades mRNA, including transcripts that are normally long-lived, but leaves ribosomal RNA unscathed. It cleaves these mRNAs within the protein-coding region, usually between codons and with little apparent sequence specificity, which is further evidence that its activity is dependent on translation.

What then to make of the previous report that YhaV does not require ribosomes for its endonuclease activity [14]? Significantly, that investigation did not test whether the activity observed for purified YhaV could be stimulated by ribosomes; nor was the target specificity of the toxin examined in vivo. Furthermore, both that study and the current inquiry by Choi et al. were complicated by the need to isolate the toxin as a complex with its neutralizing antitoxin, free it from the complex by denaturation, and then refold it in the absence of the antitoxin in order to examine the endonuclease activity of the purified toxin in vitro. The multifaceted approach employed by Choi et al. to address the ribosome dependence of YhaV and the compatibility of their observations in vitro and in vivo engender confidence in their conclusion that YhaV is catalytically activated by ribosome binding.

X-ray crystallographic analysis of ribosome-associated RelE, YoeB, and HigB has provided crucial information about their mechanism of action and the structural basis for their specificity [16–18]. Each binds to the A site of the small ribosomal subunit and cuts within the mRNA codon there by using a distinct set of active site residues to promote nucleophilic attack on the scissile phosphate by the adjacent 2′ hydroxyl group. Concurrent association of the ribosome with the toxin and its mRNA target appears to stimulate endonucleolytic cleavage by a combination of effects that include optimizing the conformation of both the enzyme and substrate and helping to mediate their interaction. Interestingly, those three endonucleases each cut after the second nucleotide of the targeted codon [10,16,17,19], whereas Choi et al. report that YhaV preferentially cleaves after the third. Insights into the mechanism and specificity of YhaV, and how its activity is potentiated by ribosome binding, await the determination of its structure in a complex with ribosome-bound RNA.