The internal ribosomal entry site (IRES) from encephalomyocarditis virus (EMCV) is a popular RNA element used widely in experimental and pharmaceutical applications to express proteins in eukaryotic cells or cell-free extracts. Inclusion of the wild-type element in monocistronic or bicistronic messenger RNAs (mRNAs) confers a high level of cap-independent translation activity to appropriately configured cistrons. The history of this element and the experimental consequences of sequence derivations inherent to commercial IRES vectors are less well known. Compared head-to-head with dual-luciferase reporter constructs, a native EMCV IRES in a bicistronic configuration directed 8- to 10-fold more protein than a simi larly configured pIRES vector. It also produced nearly twice as much protein as p#CITE®-1, an early monocistronic iteration, harboring a suboptimal A7 sequence in a crucial structural motif. The results indicate that investigators should be aware of and carefully report the sequence of their IRES in any comparative study. The preferred IRES (viral bases 273–845) and the minimum IRES (viral bases 400–836) for optimum activity are illustrated.
The internal ribosome entry site (IRES) from encephalomyocarditis virus (EMCV) is a noncoding RNA fragment noed for its ability to initiate high levels of cap-independent protein synthesis in mammalian cells and cell-free extracts. In 1986, shortly after EMCV-R was sequenced (1), the region responsible for this activity was localized to a genome fragment about 430 bases long, immediately 5′ to the AUG, which begins the viral polyprotein open reading frame (ORF) (Figure 1A) (2). When the enhancer element was excised and linked to other portions of the virus ORF, the resulting T7 transcripts were highly active messenger RNAs (mRNAs) in the absence of 5′ capping reactions.
This useful discovery was commercialized in 1990 by Novagen (Madison, WI, USA) in the form of p#CITE®-1, a vector that allowed easy linkage of exogenous cistrons onto the cap-independent translation enhancer for transcription of hybrid mRNAs and facile protein expression in cell-free extracts. p#CITE-1 enjoyed a wide distribution because of the high level of enhancer-conferred protein synthesis (3). In 1992, that popularity led to the release of p#CITE-2, a derivative vector, differing in several important features. First, an expanded multiple cloning site (MCS) was engineered 3′ of the natural virus AUG, to allow ligation of cistrons from multiple reading frames. The original plasmid, based on viral pE5LVP0 (2) had utilized only native BalI (MscI) and NcoI sequences (Figure 1B). Second, an ACC triplet was inserted upstream of the AUG, to create a new NcoI site and to bring this codon into better accord with canonical (at that time) eukaryotic initiation sequences. The third difference corrected an isolate-specific discrepancy within the enhancer element, whereby the published EMCV-R sequence, 5′-GGTTAAAAAACGTC-3′ (the A6 sequence), had by chance, been recloned by Novagen as 5′-GGTTAAAAAAACGTC-3′ (the A7 sequence).
The questionable oligo(A) was in a highly conserved cardiovirus and aphthovirus bifurcation loop at the junction of stems J and K (Figure 1A), and p#CITE-2 was adjusted relative to p#CITE-1 to conform with the native sequence (4). The consequence of this change was not immediately appreciated. It was only later reported that A6 versus A7 segments actually required somewhat different cohorts of translation factors for optimal activity (5) and therefore had the potential to respond differently in a range of cell types. It is now recognized that this particular loop is crucial to transla tional function, because it interacts with translation initiation factors (6) and, in its genome context, with the EMCV Leader protein (7).
In 1988, the availability of vectors to enhance translation, in any format, sparked lively interest in new eukaryotic expression techniques. One particularly innovative experimental series linked the (pE5LVP0) segment downstream of a cap-dependent reporter gene in a bicistronic configuration. Translation of the EMCV-driven cistron proved independent of the upstream gene, an activity that could not be attributed to leaky scanning or ribosome read-through mechanisms (8). Dubbed an internal ribosome entry site, or IRES, the moniker has stuck with this element and has since been applied to many other virus or cellular RNA fragments that also direct eukaryotic translation in a cap-independent manner. To be sure, other viral noncoding segments, most notably from polio–virus (9), were tested at about this time in analogous bicis tronic constructions and also shown to be IRESs. The superior activity of the EMCV IRES in multiple cell and cell-free systems, however, popularized its use. In 1995, Clontech capitalized on the bicistronic potential and released a new vector (pIRES) that could be delivered directly into mammalian cells as cDNA. Transfection with pIRES induced nuclear transcription of capped bicistronic mRNAs, driven by a cytomegalovirus (CMV) promoter. Translation of any upstream gene “A” inserted into the 5′ -most MCS-A was cap-dependent. Linked in tandem was a modified EMCV IRES, a second MCS (B), and a poly(A) signal sequence. Protein synthesis for gene “B” inserted at MCS-B was IRES-dependent. Since the native EMCV IRES usually overexpressed protein relative to an equivalent capped mRNA, Clontech chose to partially disable the bicistronic EMCV segment to attenuate B cistron translation down to the more moderate level of the A cistron (10). Accordingly, pIRES vectors have A7 rather than A6 bifurcation sequences, and the MCS-B does not include a native EMCV AUG (Figure 1B).