Ribosomal frameshifting and transcriptional slippage: From genetic steganography and cryptography to adventitious use

Atkins, John F.; Loughran, Gary; Bhatt, Pramod R.; Firth, Andrew E.; Baranov, Pavel V.

Excluding the special case of antizyme 3, there is frequent occurrence of either a glycine or proline, as in human antizyme 1, encoded 7 codons 5 of the shift site and flanked on one or other side by a proline or glycine ( Table 2 ). The location of the proline in human antizyme 1 in a complex of antizyme and some of its interacting partners is now known (276, 291) . Because antizyme has other interacting partners, caution is needed about function in intact protein except to note that proline would be very different from glycine. However, though context dependent (see below) tandem proline or proline glycine (see below) can be slow-to-decode. Proline is an exceptionally poor donor and acceptor for peptide bond formation for steric rather than electronic reasons (467) . Polyproline containing peptidyl-tRNA is prone to destabilization leading to drop-off (467) , and the ribosomes are trapped in a pre-translocational state with a free E-site (468) . In eukaryotes this is recognized by a protein with a tRNA-like shape, elongation factors eEF5 (formerly eIF5A), that aids polymerization, Reviews (469, 470) . Distinct from the type of association between polyamines and antizyme, in eEF5 a key lysine residue is uniquely post-translationally modified with polyaminederived hypusine. Hypusine, and different modifications on the bacterial counterpart of eEF5, EF-P, likely evolved for the stabilization of the CCA-end of P-site tRNA (470, 471) [Importance of CCA characteristics are shown by tRNA with CCA mutated to GCA or ACA causing frameshifting at a specific sequence and this being enhanced by mutants of hrpA (440, 472) that encodes an RNA helicase (473) . EF-P is also relevant to frame maintenance (474) .] eEF5's hypusine moiety interacts with the backbone of the CCA-end of the P-site tRNA and entropically steers the substrate into a more favorable position in the peptidyl-transfer center (467, 468, 475) . Reduced hypusination of eEF5 due to limiting polyamines, would likely lead to a greater stall presumably with increased drop-off (whether the third codon base in the key codons preferentially has non C or G in the third codon base to facilitate such hypothetical regulatory dropoff has not been addressed). Irrespective, perhaps reduced hypusination can influence the nascent peptide already in the peptide tunnel or some other feature of the ribosome that has an effect on frameshifting 6 codons later. Studies in bacteria have shown that the extent of stalling at proline codon stall sites is strongly influenced by the identity of the amino acids specified 5 up to a distance of 3 to 5 codons, though the effect decreases with distance. Though the pattern is complex depending on the identity of the stall site, codons for H, K, Q, R or W enhance stalling in contrast to those for C, G, L, S or T (476,477) (Note in the StopGo mechanism, below, the flanking amino acid on the N-terminal side is N, and UGG encoding W is strongly conserved as the second codon 5 of the antizyme shift site). Features of the nascent peptide just inside the peptide exit tunnel are likely relevant. The identity of the codon directly 3 and the extent of spacing between ribosomes can also be relevant (476, 478) . [A progressive effect over several codons of the gene 60 bypassing nascent peptide signal is described 2 paragaraphs down.] Sequence 5 of the mammalian antizyme 1 shift site was shown to be important for the sensing of polyamine levels that influence frameshifting efficie