Signal sequences on messenger RNA that initiate protein synthesis are not thought to be interchangeable between life's domains. The finding that a signal from an arthropod virus can function in bacteria questions this idea. See Letter p.110
All domains of life, from prokaryotes (archaea and bacteria) to eukaryotes (organisms that include plants, animals and fungi) use the ribosome apparatus to synthesize proteins by translating genetic code carried by messenger RNAs. Although the general steps of protein synthesis are evolutionarily conserved, the way in which ribosomes are recruited to an mRNA molecule differs depending on the specific phylogenetic domain1. In this issue, Colussi et al.2 (page 110) reveal the surprising finding that a eukaryotic ribosome-recruiting signal is functional in prokaryotic bacteria, thereby challenging the prevailing dogma that prokaryotic and eukaryotic ribosome recruitment are mutually exclusive.
In prokaryotes, ribosomes are generally recruited by a specific signal on the mRNA called the ribosome-binding site (RBS, also called the Shine–Dalgarno sequence). The RBS positions the ribosome over the AUG start codon, a sequence that initiates protein synthesis1. By contrast, eukaryotic mRNAs do not contain an RBS, but instead contain covalent modifications — the 5′ cap at the 5′ end, and a poly(A) tail at the 3′ end — that promote the recruitment of ribosomes and protect the mRNA from degradation. Eukaryotes also use at least 12 protein initiation factors (compared to three in prokaryotes) that help to recruit the small subunit of the ribosome to the 5′ cap and facilitate scanning of the mRNA by the ribosome to find the AUG codon.
Some eukaryotic viral RNAs contain alternative signals called internal ribosome entry sites3 (IRESs) to recruit the ribosome. When host-protein synthesis is inactivated during virus infection, an IRES can bypass normal signals, such as the 5′ cap and some initiation factors, to enlist the ribosome for viral protein synthesis. The simplest known IRES lies within regions found between the genes of dicistroviruses — a family of RNA viruses that infect arthropods4.
Dicistrovirus IRESs are about 200 nucleotides long and fold into a unique RNA structure. What makes them so remarkable is their ability to bind directly to the ribosome without the need for any initiation factors and to initiate protein synthesis at a non-standard start codon5 (not AUG). Structural6 and biochemical studies7 have shown that part of the IRES structure binds to the conserved core of the ribosome by mimicking a transfer RNA, which normally delivers an amino acid to the growing protein chain. The dicistrovirus IRES has therefore evolved to mimic a component of the normal translational machinery, permitting the viral RNA to hijack the ribosome.
Colussi and colleagues tested whether the dicistrovirus IRES can function in bacteria. To do this, they constructed mRNA that encodes reporter proteins — in this case, luminescent proteins. They found that inclusion of the IRES into the mRNA promotes expression of the reporter proteins in the bacterium Escherichia coli. Impressively, the researchers generated a comprehensive set of mutations within the IRES to work out how it functions in bacteria. They observed that some mutations that disrupt the IRES RNA structure reduced expression of the reporter proteins, demonstrating that the integrity of this structure is crucial for function.
To gain further insight into how IRESs that evolved to hijack ribosomes in eukaryotes can work in prokaryotic bacteria, Colussi et al. used X-ray crystallography to acquire a high-resolution image of the dicistrovirus IRES bound to the bacterial ribosome. Remarkably, the IRES binds to the ribosomal core, which is present across all domains of life, thus explaining how it can interact with both eukaryotic and bacterial ribosomes. The authors also show that an IRES structure from an unrelated virus — classical swine fever virus — does not support reporter-protein expression in bacteria, indicating that only the dicistrovirus IRES enables ribosome recruitment in E. coli.
But that is not the whole story. The investigators found that maximal reporter expression in bacteria depends not only on the IRES, but also on an RBS hidden in the reporter-system sequence and on an AUG start codon of the reporter mRNA. Moreover, their mutational analysis indicated that protein synthesis does not start at the IRES, but at the AUG codon. Colussi and co-workers therefore propose that, after being recruited to the IRES, the bacterial ribosome moves downstream to interact with the RBS and to start protein synthesis at the AUG codon (Fig. 1). Alternatively, another ribosome might be recruited to the RBS and initiate protein synthesis, a model that needs to be examined further.
It therefore seems that the authors' reporter system uses a hybrid of eukaryotic and prokaryotic signals to promote protein synthesis in bacteria. In prokaryotic translation, RNA structures are dynamic, and some can control protein expression by either burying the RBS within their structure or altering conformation to expose the RBS for ribosome recruitment8. Although leaderless RNAs — which are translated in the absence of the signals that usually support ribosome binding and translation efficiency — can also function across domains9, Colussi et al. have shown for the first time that a bona fide signal in an RNA structure promotes protein synthesis in two domains of life.
This exciting study raises new questions. If the bacterial ribosome repositions itself from the IRES to the RBS, as the authors suggest, then how does it do so? Is the movement akin to ribosomal scanning of mRNAs in eukaryotes? And does the IRES function in bacteria without initiation factors, as it does in eukaryotic cells? Colussi and colleagues report that the dicistrovirus IRES binds to bacterial ribosomes slightly differently and more transiently than it does to eukaryotic ribosomes. Importantly, the IRES interacts with eukaryotic ribosomal protein rpS25, which is not present in bacterial ribosomes10. Further studies are needed to address how the IRES manipulates bacterial ribosomes.
Although the IRES functions in bacteria, does this phenomenon have a physiological role, and do other bacterial RNA structures function similarly? Another question is whether the IRES could work in archaea, or interact with the prokaryotic-like ribosome of mitochondria (specialized organelles that act as cellular powerhouses). And could the IRES be a remnant of a molecular fossil of the ancient 'RNA world' that is widely presumed to have preceded the evolution of DNA and proteins? Finally, this study opens up the possibility that other RNA structures function as signals across domains, and that eukaryotes and bacteria have more in common than was previously thought.