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.

Natural perturbation of aminoacyl-tRNA levels occurs under conditions of amino acid starvation, a common situation in nature especially for a high proportion of bacteria, and is exacerbated in relA mutants when tRNA undermodification is more prevelant (433) . The effect of this on frameshifting has been studied extensively in E. coli (409) . What is unclear is under natural conditions, the extent to which such 'hungry codon' frameshifting leads to out-offrame termination with subsequent proteolysis contributing to recycling of scarce amino acids for the synthesis of proteins that are vital under such conditions. [While this type of frameshifting could contribute to the synthesis of transframe encoded proteins that have novel function specific for starvation conditions, no examples are known.] Almost all tRNAs have standard anticodon loop sizes. Among the few exceptions are a mitochondrial tRNA from yeast S. cerevisiae (434) , tRNAs in a phage (435) that has many relatives, (436) and a mitochondrial tRNA from the glass sponge Iphiteon panacea (252) . The S. cerevisiae tRNA has an extra base in its anticodon loop but neither it nor the other two, are known to mediate frameshifting. However, frameshifting does occur naturally in expression of several mitochondrial genes in Polyrhachis species and I. panacea. One of the I. panacea mitochondrial tRNAs whose anticodon loop very likely has 9 nts is tRNA Gly and it has 4 bases in the anticodon location that matches the UGGA codon at which the shift to the +1 frame occurs. The frameshifting properties of mutant tRNAs with 9 or 10 nt anticodon loops has been studied in E. coli with spontaneously arising mutants (437) or ones sought in code expansion studies (438); though enlarged anticodon loops do not necessarily imply other than triplet pairing inside WT ribosomes (439, 440) , the use of the anticodon loop of a tRNA switching stacking to present an alternative offset triplet anticodon has been proposed (441) . However, even if quadruplet pairing or an anticodon loop 'roll' occurred in a few cases of mitochondrial frameshifting, it would be exceptional and not the general explanation. Nevertheless, while much of the natural frameshifting in mitochondrial expression to circumvent potential deleterious consequences of mutational inserts follows the same principles as just described, some of the features are mitochondrial specific. At most of the insert sites in Polyrhachis ants, the zero-frame codon 3 adjacent to that for the peptidyl-tRNA Gly involved in the switch are AGY codons. Its cognate tRNA has a characteristic restricted to a subset of mitochondrial tRNAs in that it lacks the DHU stem, and has a distinctive structure. Perhaps it is slow-to-decode. The +1 frameshifting in decoding several genes in glass sponges occurs at extremely rare UGG (A), or CGG (A) codons (with all the inserts being T, or rarely C, 5 adjacent to a GGA codon). Though most of their frameshifting has been similarly interpreted to that of Ty3 (252), other possibilities merit consideration (mitochondrial ribosomes also have an E-site, despite an earlier proposal to the contrary, and so dissociation potential of the corresponding codon: anticodon interaction may be relevant).