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.

Shortly after the discovery of splicing in 1977, it was widely suspected that it was involved in generating an mRNA from which GagPol was synthesized by standard decoding. However, suggestive evidence was quickly obtained that for the gammaretrovirus, Moloney murine leukemia virus, whose pol is in the same frame as its gag, stop codon readthrough was involved and this was confirmed in 1985 (92, 93) . By 1984 two groups were testing whether frameshifting was involved in retroviruses where pol is in the −1 frame with respect to gag. A dilemma faced by both groups at the time was how to distinguish between a low level of ribosomal frameshifting and polymerase slippage (related to acknowledgements in (76) ). The initial 'proof' for retroviral gag-frameshifting being ribosomal rather than due to standard decoding of altered RNA arising from polymerase slippage, was the reported absence of fusionderived product in E. coli cell-free protein synthesis programmed with SP6 transcripts that yielded fusion product in reticulocyte lysates (76) . However, cassettes with the gag-pol shift region do exhibit frameshifting in such sys-tems with ∼40% of the efficiency shown in mammalian systems (94) (95) (96) (97) (98) . Ironically for one of us, current in vitro work is revealing a modest level of HIV reverse transcriptase realignment (slippage) at the same 7 nt shift site at which the ribosomal frameshifting that yields GagPol occurs, to yield product with an extra base or lacking a base. This polymerase slippage is influenced by the ratio of two of the 4 dNTPs (C. Penno, P.V.B. and J.F.A. unpublished) (and possibly by an RNA structural feature different from that recently described (99) ). dNTP depletion is part of the hosts response to viral infection and lentiviruses, such as HIV-2, have developed a vpx gene to counteract this effect (100, 101) . The problem with the control in the initial published work on retroviral frameshifting (76) does not detract from its great importance and the elegance of both it and the subsequent work by the same group including the discovery of heptanucleotide −1 frameshift sites to which 2 tRNAs could pair in alternative frames (102) . A similar comment also pertains to the relatively recent finding that a proportion of the HIV frameshifting products are due to −2 rather than −1 frameshifting (98) .