Ribosomal progression through the open reading frames within mRNAs is frequently considered as uneventful when compared with the highly regulated initiation step. However, both RNA and nascent peptide can interact with the ribosome to influence how translation proceeds and can modify gene expression in several ways. 2A peptides are a class of sequences that, as nascent chains, pause ribosomes and drive a translation-termination reaction on a sense (proline) codon, followed by continued downstream translation. In the present paper, what is known about the 2A reaction is discussed, and 2A is compared with other sequences that, as nascent peptides, pause or stall translation.
Once translation has initiated on an mRNA, ribosomes proceed through the elongation phase until they reach a stop codon, where the nascent chain is released by translation-terminating RFs (release factors). The course of translation is, however, not always as smooth as this description suggests. mRNA structure and the availability of cognate tRNAs for the codons present can influence elongation rate. Similarly, nascent peptides can, through interactions with the exit tunnel of the large ribosomal subunit, cause ribosomes to pause or even stall. Such alterations to elongation rate can affect downstream events such as folding of the nascent chain, accessibility of regulatory regions such as the Shine–Dalgarno sequence of downstream cistrons (in polycistronic bacterial mRNAs) and scanning of small ribosomal subunits to the main open reading frames of eukaryotic mRNAs. Pauses to translational elongation may also be linked to translational recoding, events such as programmed stop codon readthrough and frameshifting, wherein translation diverges from standard decoding rules. The present review discusses several peptides, of both prokaryotic and eukaryotic origin, that, as nascent chains, affect ribosomal progression, with a particular focus on a class of peptides, termed 2A, that drive a recoding event, a co-translational ‘skip’ in peptide bond formation.
2A peptides, also termed CHYSELs (cis-acting hydrolase elements), are short ~19-amino-acid sequences, found primarily in a variety of viruses, but also in repeat [non-LTR (long terminal repeat) retrotransposon] sequences in the genomes of trypanosomes (reviewed in [1–3]). These peptides direct ribosomes into an extremely unusual series of events: termination of translation on a sense codon followed by (re-)initiation on the same codon (Figure 1) .
Model of the 2A reaction
The first 2A peptide identified, and the most studied, is that from the picornavirus FMDV (foot-and-mouth disease virus) . It corresponds to the 2A region of the viral polyprotein (which lends its name to the whole class of peptides) plus the first residue, proline, of 2B. It is on the codon specifying this proline residue that the reaction takes place: a glycine residue encoded by the preceding codon becomes the final amino acid of the first (upstream) translation product, and the proline residue is the first of the downstream product [4,6]. Both of these amino acids are critical for activity of the FMDV sequence . Since replacement of the proline codon with one encoding a different amino acid results in continued translation, tRNAs must have access to the A site when ribosomes reach the 3′ end of a sequence coding for a 2A peptide. Somehow, however, proline is unable to incorporate into the growing peptide chain.
Significant insight into the 2A reaction mechanism has come through experiments using fungal systems (Saccharomyces cerevisiae and Neurospora crassa), evolutionarily far removed from the higher eukaryotic hosts of FMDV [4,7]. This highlights a particularly useful feature of 2A peptides: they function in all eukaryotic systems in which they have been tested, providing a useful tool for co-expression of more than one protein from a single ORF (open reading frame) . This does not, however, extend to prokaryotes, where 2A does not function . Key observations from these experiments include mapping a ribosomal pause with the final glycine and proline codons of 2A in the P and A sites respectively, and finding that genetic depletion of translation terminating RF leads to more ribosomes incorporating 2A into a full-length ORF product. This indicates that RF catalyses release of the upstream reaction product despite the presence of a proline codon in the ribosomal A site. Intriguingly, overexpression of proteins containing 2A leads to both increased readthrough of stop codons in trans and impaired growth of yeast strains with limited RF activity. These data are consistent with the suggestion that ribosomes translating 2A somehow titrate RF activity. This could be through an atypically long dwell time on ribosomes at 2A or that after the unusual termination reaction RF does not recycle efficiently. As yet, little is known about the second step in the reaction: the restart to translation that initiates synthesis of the downstream product.
The final glycine and proline codons of 2A are part of a longer motif in its C-terminal portion. This motif, D(V/I)EXNPGP, is almost invariant among 2A peptides. In contrast, the N-terminal portion of 2A is highly variable. The minimal FMDV sequence with any activity is 13 amino acids long, and includes five of these non-conserved amino acids, indicating that they provide some important function to the peptide [5,9]. Furthermore, although 2A is defined as being 19 amino acids long, this being the length of the 2A segment of the FMDV polyprotein, inclusion of upstream sequences from the virus, up to approx. 30 codons in total, into reporters enhances 2A activity . Thus amino acids upstream of the conserved motif of 2A somehow provide a platform that promotes its activity.
Barring the final few amino acids, several picornaviral 2A peptides can be modelled into an amphipathic α-helix [8,10]. Molecular dynamics simulations confirm that this is indeed a favourable conformation for the FMDV peptide, and suggested that the helix might extend to asparagine at position 16 . The final amino acids were then postulated to form a tight (type VI reverse) turn that might move the peptidyl-tRNAGly sufficiently from the usual positioning within the ribosomal peptidyltransferase centre to preclude formation of a new peptide bond between the peptide chain and the secondary amine of the incoming (A site) prolyl-tRNA.
This model leads to important questions, not least whether data can be gathered that supports (or not) the possibility that 2A is structured in the ribosomal exit tunnel. The exit tunnel can accommodate nascent chains in a helical conformation , and, indeed, several peptides have been shown to adopt compact structures in the exit tunnel (e.g. [13–15]). Many more 2A peptides are now known than when the initial suggestion was made that they may form an α-helix. Among this expanded set, a fraction do not have significant α-helical propensity (J.D. Brown, unpublished work). Furthermore, mutagenesis of FMDV 2A has revealed that changes to the peptide across most of its length, not just the conserved region, are detrimental to activity (F. Yan, P. Sharma and J.D. Brown, unpublished work). Thus, whether or not it adopts an α-helical conformation, the identity of most of the amino acids within the peptide is critical for function. The ‘variable’ N-terminal portion of FMDV 2A is then perhaps best considered as a single solution to the problem of correctly positioning the conserved C-terminal region in the peptidyltransferase centre, with other 2A peptides having arrived at other solutions.
SecM (secretion monitor)
SecM lies upstream of and regulates SecA, encoding the ATPase that drives secretion of many proteins from bacteria through the SecYEG PCC (protein-conducting channel). During translation, the nascent SecM protein induces arrest of translational elongation . The stall is released when SecM, which contains an N-terminal signal sequence, becomes engaged with the PCC. As ribosomes stalled on SecM disrupt secondary structure in the mRNA that would otherwise occlude the Shine–Dalgarno sequence of the SecA ORF, increased duration of pausing on SecM correlates with greater synthesis of SecA. The sequence of SecM that directs stalling is the conserved motif from amino acids 150 to 166 (F150XXXXWIXXXXGIRAGP166) of the 170-amino-acid protein. The arrest occurs with the penultimate Gly165 of the motif in the P site and the final Pro166 codon in the ribosomal A site . Prolyl-tRNA binds to the A site, but proline is not incorporated into the peptide [18,19]. Clearly, the organization/function of the peptidyltransferase centre of the ribosome is disturbed by the action of the nascent SecM arrest motif.
Mutation to ribosomal components (both RNA and protein) that form the narrowest point in the exit tunnel reduced stalling on SecM, indicating that interactions between the peptide and this ribosomal region are key for its activity . Insertion of fluorescent probes into nascent SecM peptides allowed FRET (fluorescence resonance energy transfer) studies to be carried out on the stalled SecM–ribosome complex, and these revealed that the peptide was in a compact (probably α-helical) conformation in the exit tunnel . Cryo-EM (electron microscopy) and single-particle reconstruction revealed dramatic alterations to ribosome structure driven by the SecM peptide . This is a key finding as it demonstrates that the nascent chain can have profound effects on ribosomal conformation.
Residues within SecM necessary for stalling are distributed along its length, consistent with multiple interactions with the exit tunnel driving stalling, and include several predicted to lie close to the exit tunnel constriction during the stall to translation. However, a recent study revealed that, with the exception of the final Pro166 codon and Arg163, the importance of other ‘key’ amino acids, including those proximal to the exit tunnel constriction, is context-dependent, rather than absolute . Indeed, altered arrest motifs with as few as three of the nine key residues of the native Escherichia coli SecM are completely functional. This strongly suggests that residues in the motif work together as a module, presumably affecting the structure of the nascent chain to position the critical Arg163 correctly with respect to the Pro166 codon in the A site.
The tryptophanase operon of E. coli, responsible for catabolism of tryptophan, encodes a leader peptide, TnaC . This 24-amino-acid sequence regulates expression of downstream genes, based on the cellular concentration of tryptophan. Stalling on TnaC is tryptophan-dependent , and makes a Rho-dependent transcription termination site (on the spacer between the sequence of TnaC and TnaA) inaccessible. When the tryptophan concentration is low, ribosomes release the peptide and termination by Rho reduces transcription of TnaA and TnaB.
Ribosomes stall on TnaC with its stop codon in the A site, and thus mechanistically RFs must be prevented from acting on the TnaC–tRNAPro ester bond. A recent cryo-EM study revealed that, in the arrested TnaC–ribosome complex, two conserved nucleotides in the ribosomal PCC adopt conformations that preclude RF binding . This study also showed that the peptide is in extended conformation (unlike nascent SecM), making contacts with the exit tunnel at several points. Key residues of TnaC are tryptophan at position 12 and aspartic acid 16 in addition to the final proline 24 . Similar to SecM, mutations that alter the exit tunnel constriction affect the ability of TnaC to impose arrest, with some affecting both TnaC- and SecM-induced stalling .
Cytomegalovirus UL4 uORF2
Expression of the human cytomegalovirus UL4 gene is inhibited by the 22-amino-acid peptide encoded by its second upstream ORF (uORF2: MQPLVLSAKKLSSLLTCKYIPP) . Similar to TnaC, the nascent uORF2 peptide arrests ribosomes at its own termination codon, resulting in a persistent covalent linkage to the last tRNA, which, also similar to TnaC, is proline [28,29]. Reduced translation of the downstream UL4 ORF is then due to stalled ribosomes preventing others from scanning beyond uORF2 to access UL4-coding sequences. Intriguingly, ribosomes stalled at the uORF2 recruit higher levels of eRF1 (eukaryotic RF1), but not eRF3, over ribosomes synthesizing a non-functional mutant peptide, suggesting a very different mechanism to that used by the TnaC nascent chain to inhibit termination . Furthermore, addition of eRF1 carrying mutations to key residues involved in the catalysis of termination (Gly183 and Gly184 of the conserved GGQ motif) to translation reactions, reduced ribosomal stalling on uORF2, suggesting that these amino acids are critical for generating the stalled complex. The final two proline residues of uORF2 are required for stalling, and a model proposed to explain these observations envisages the proline residues and the GGQ motif of eRF1 interacting to stabilize an intermediate in termination .
The AAP (arginine attenuator peptide)
The mRNA encoding the small subunit of carbamoyl phosphate synthase (a component of the arginine biosynthetic pathway) of a variety of fungal species, including S. cerevisiae and N. crassa, contain a uORF, translation of which restricts synthesis of the main ORF-encoded product in response to high arginine levels . Again, as with nascent TnaC and cytomegalovirus UL4 uORF2 peptides, ribosomes stall with the stop codon of the uORF in the A site . However, the final amino acid encoded by the uORF is not proline, and this is not a conserved position in the peptide. The N-terminus of the uORF-encoded peptide, or AAP, comprises non-conserved amino acids, which are not required for its activity, whereas the sequence becomes conserved in the middle, with several non-conserved amino acids making up its C-terminus . Mutagenesis indicates that the conserved amino acids are important for stalling. An intriguing feature of the AAP is that its ability to stall ribosomes does not require a stop codon, i.e. it functions as part of a longer ORF [31,33]. In this case, ribosomes stall at several positions just downstream of the AAP. Mechanistically then, the AAP would appear to be very distinct from other peptides discussed above.
Pause for thought
The above few examples illustrate that the ribosome can be stalled by a variety of different peptide sequences through alternative mechanisms. A recent high-resolution structure of a yeast ribosome–nascent chain complex included density for the nascent chain suggesting a particular orientation was preferred, even though the nascent chain present was not one that affects translation activity . Perhaps, then, interaction of the nascent chain with the exit tunnel is a frequent occurrence (rather than the exit tunnel being a slippery tunnel). This may provide cues for initial folding events , as well as the possibility for feeding regulatory information back to the peptidyltransferase centre and other parts of the ribosome. In this regard, it is notable that transmembrane sequences can drive alterations to the organization of the ribosome–translocon complex while still located within the exit tunnel .
Although the peptides discussed above have rather different sequences, common features do emerge. First, proline appears in either the A site or the P site in several stalling peptides. Proline is relatively poor as both a participant in termination and as a nucleophile for peptide bond formation [36–39]. Additionally, its secondary amino group has both conformational and steric constraints, and reactions involving it are perhaps more likely to be sensitive to slight perturbations in ribosomal geometry than those of other amino acids. These properties of proline make it a prime candidate for inclusion in the active site of the ribosome when stalling is promoted. Secondly, it is clear that, whereas there are always key residues necessary for any given pausing (or in the case of 2A recoding) event, the constraints on these are perhaps not as strict as first envisaged, since they work together with others in the peptide sequence to provide critical positioning of groups within the peptidyltransferase centre and to modify ribosomal geometry. The recent studies on SecM highlight this very clearly . In the context of the 2A reaction, identification of new candidate 2A sequences in sequence databases has relied on the conserved motif at its C-terminus. Perhaps mutagenesis studies will find different combinations of C-terminal amino acids that can function in the reaction, and lead to new classes of 2A peptides being uncovered.
Differences between peptides and their activities are also remarkable. The compact structure of SecM and extended nature of TnaC in the exit tunnel indicate that very different modes of peptide–exit tunnel interaction can affect ribosomal activity. Structural studies also indicate that, whereas SecM alters ribosomal conformation globally , TnaC does not, promoting instead subtle alterations to the peptidyltransferase centre . This indicates a relaying of information from the exit tunnel to the peptidyltransferase centre, probably through the nascent chain itself. Some peptides act constitutively, whereas others are inducible regulatory elements, acting under particular circumstances, frequently in the presence of specific cofactors.
2A directs ribosomes to pause, recruit RFs in the absence of a stop codon in the A site to terminate translation and then promotes a restart to translation. How is this all achieved? Generation of substrates that stabilize intermediates in the 2A reaction pathway and study of these biochemically and structurally will be key for furthering understanding of the 2A reaction. A number of situations are known where eukaryotic translation initiation proceeds without the need for a cap and scanning of small ribosomal subunits. Internal ribosome entry sites recruit ribosomes directly to initiation sites, and several examples of coupled termination–re-initiation events are known. These are driven by initial recruitment of the small ribosomal subunit, or, in the case of re-initiation, by its retention on the mRNA after termination. 2A, acting through the exit tunnel in the large subunit must function differently, preventing the action of recycling factors  that would break apart the 80S ribosome, and facilitating translocation of the ribosome to place the final proline codon of 2A in the P site.
Post-Transcriptional Control: mRNA Translation, Localization and Turnover: A Biochemical Society Focused Meeting held at University of Edinburgh, U.K., 8–10 June 2010. Organized and Edited by Matthew Brook (Edinburgh, U.K.), Mark Coldwell (Southampton, U.K.), Simon Morley (Sussex, U.K.) and Nicola Gray (Edinburgh, U.K.).
This work was supported by a Medical Research Council Senior Non-Clinical Research Fellowship and by the Biotechnology and Biological Sciences Research Council [grant numbers C20418 and BB/E009093] to J.D.B. and by Universities UK and Newcastle University to P.S..
Present address: Faculty of Life Sciences, University of Manchester, Stopford Building, Oxford Road, Manchester M13 9PT, U.K.