Figure 3.

Potential downstream RNA secondary structures in all sequences analysed. (Continued in Figure 4.) As of 20 April 2008, there were 357 alphavirus sequences in GenBank with coverage of the U UUU UUA motif in the 6K cistron. The 100 nt region starting from the U UUU UUA motif, and including the first 93 nt of 3'-adjacent sequence, was extracted from all 357 sequences (although in 26 sequences a shorter region had to be used due to incomplete sequence data). Shown here are the 108 unique ≤100-nt sequences, plus an additional seven duplicate sequences also included since they have different species/strain annotations. The total number of duplicate sequences represented by each sequence shown is given in column 1, while column 2 gives an example GenBank accession number for the sequence, and column 3 gives the virus name abbreviation. Potential RNA secondary structures were identified using a combination of RNAfold and alidot [36], pknots [37], and manual inspection. Bases within potential stems are indicated either in colour or with underlines (if overlapping other potential stems) and potential base-pairings are indicated with brackets – '()', '[]' or '<>'. '<>' signify more dubious base-pairings, including stems that were experimentally shown not to affect frameshifting efficiency (dual luciferase assays with inserts comprising the U UUU UUA motif and 3'-adjacent sequence; B Chung et al, in preparation). Base variations that maintain base-pairings are marked in bold. Note that not all sequences in GenBank represent functional (infectious) viruses and it is possible that certain sequences whose shift site and/or predicted RNA structure do not conform with the majority of isolates for the same species may represent defective viruses – for example the non-standard slippery heptanucleotide in the SPDV sequence AJ012631 is due to a 36-codon deletion in 6K relative to other SPDV sequences.