In vivo and in vitro identification of structural and sequence elements ofthe human parechovirus 5 ' untranslated region required for internal initiation

Sequence analysis of the picornavirus echovirus 22 led to its classificatio n as the first member of a new genus, Parechovirus, and renaming as human p arechovirus type 1 (HPeV1). Although distinct from other genera in most of the genome, the 5' untranslated region (5'UTR) shows similarities to that o f cardio/aphthoviruses in some of its structural domains (A to L). The 5'UT R plays an important role in picornavirus translation initiation and in RNA synthesis. To investigate translation in HPeV1, we engineered an extensive range of mutations (including precise deletions and point mutations) into the 5'UTR. Their effects were studied both by in vitro transcription-transl ation using a bicistronic construct and by in vivo studies using an infecti ous, full-length HPeV1 cDNA. These approaches allowed the HPeV1 internal ri bosome entry site (IRES) to be mapped. Deletions within the first 298 nucle otides had little impact in the in vitro system, while deletions of nucleot ides 298 to 538 had a significant effect. Precise removal of domains H and L (nucleotides 287 to 316 and 664 to 682, respectively) did not significant ly reduce translation efficiency in vitro, while domains I, J, and K (nucle otides 327 to 545, 551 to 661, and 614 to 645, respectively) appeared to ha ve much more important roles. Mutation of a phylogenetically conserved GNRA moth (positions 421 to 424) within domain I severely reduced translation. We also confirmed the identity of the AUG (positions 710 to 712) which init iates the open reading frame, the positive identification of which has not been possible previously, as the N terminus of the polyprotein is blocked a nd not amenable to sequence analysis. This is therefore important in unders tanding parechovirus genome organization. Mutation of the AUG or an upstrea m polypyrimidine tract leads to aberrant translation, suggesting they both form part of the parechovirus Yn-Xm-AUG moth. In vivo experiments confirmed the importance of domains I, J, and K, the conserved GNRA moth, polypyrimi dine sequences, and AUG, as mutations here were lethal. These features are also important in the IRES elements of cardio/aphthoviruses, but other feat ures reported to be part of the IRES of some members of these genera, notab ly domains H and L, do not appear to be critical in HPeV1. This adds weight to the idea that there may be functional differences between the IRES elem ents of different picornaviruses, even when they share significant structur al similarity.