New insights into flavivirus evolution, taxonomy and biogeographic history, extended by analysis of canonical and alternative coding sequences

Abstract

To generate the most diverse phylogenetic dataset for the flaviviruses to date, we determined the genomic sequences and phylogenetic relationships of 14 flaviviruses, of which 10 are primarily associated with Culex spp. mosquitoes. We analyze these data, in conjunction with a comprehensive collection of flavivirus genomes, to characterize flavivirus evolutionary and biogeographic history in unprecedented detail and breadth. Based on the presumed introduction of yellow fever virus into the Americas via the transatlantic slave trade, we extrapolated a timescale for a relevant subset of flaviviruses whose evolutionary history, shows that different Culex-spp. associated flaviviruses have been introduced from the Old World to the New World on at least five separate occasions, with 2 different sets of factors likely to have contributed to the dispersal of the different viruses. We also discuss the significance of programmed ribosomal frameshifting in a central region of the polyprotein open reading frame in some mosquito-associated flaviviruses.

Fig 1. Bayesian phylogeny of the ORF ‘global genus’ amino acid dataset.

Only posterior probabilities of 0.9 are included. The tree is midpoint rooted. Bar, 0.3 substitutions per site.

Fig 2. Maximum clade credibility tree summarized from the Bayesian molecular clock analysis.

All nodes attained maximal probability support except for those indicated by a * (posterior probability < 0.95). The node age uncertainty is indicated using 95% highest posterior density (HPD) interval bars in blue. Old World and New World ancestral states are indicated by black and red branches/nodes respectively. Nodes of interest are listed A to N as referred in the text.

Fig 3. Frameshift stimulatory elements in viruses of the JEV serogroup.

(a) Previously identified frameshift site (Y_CCU_UUU; Y = C or U; orange) and 3'-adjacent stable pseudoknot structure responsible for stimulating-1 PRF in the NS2A-encoding region of JEV and related flaviviruses. Stems 1 and 2 of the pseudoknot are indicated in blue and red respectively. (b) The shift site and pseudoknot are preserved in the newly sequenced KOUV but not in YAOV or CPCV. Substitutions that preserve the base-pairings in stem 1 (blue) or stem 2 (red) of the pseudoknot are indicated in pale blue and orange respectively. In YAOV, a simple stem-loop (brown) was predicted at an appropriate spacing from the shift site to act as a stimulator of-1 PRF. CPCV maintains the shift site but multiple possible 3'-proximal structures (not shown) could be predicted. SLEV sequences lack a suitable shift site at this genomic location. (c) Predicted frameshift stimulatory elements (shift site and 3'-adjacent stem-loop) in YAOV.

Fig 4. Synonymous site conservation analysis for selected flavivirus clades.

Alignments of 249 JEV serogroup, 49 NTAV/TMUV clade, 89 DENV, 6 DGV/LAMV clade and 29 CxFV/QBV clade polyprotein ORF sequences were analyzed for synonymous site variability as decribed previously (Firth et al., 2011 PMID 21525127). The accession numbers of all sequences used in the analysis are available on request. Red lines indicate the probability (p-value) of obtaining not more than the observed number of synonymous substitutions, in a 25-codon sliding window, under a null model of neutral evolution at synonymous sites. Dashed grey lines indicate an approximate 5% false positive threshold after correcting for multiple tests (i.e. ~136 x 25-codon windows in the ~3400-codon polyprotein ORF). Statistically significant peaks in synonymous site conservation are indicative of overlapping functional elements, either coding or non-coding. Genome maps are shown for each clade. UTR lengths may be uncertain for less well-studied clades. Known and predicted overlapping ORFs accessed via-1 PRF are shown in pink. The predicted overlapping ORF in the DGV/LAMV clade is much shorter in DGV than in other members of the clade; the long form of the ORF is indicated. Note that p-values can not be directly compared between different clades because the statistical significance (i.e. p-value) of observed reductions in synonymous site variabilty depends on the diversity of the specific sequence alignment being analyzed.

Fig 5. Synonymous site conservation analysis for selected flavivirus clades.

Alignments of 6 UGSV/BOUV clade, 6 WESSV/SEPV clade, 56 YFV, and 144 POWV/TBEV clade polyprotein ORF sequences were analyzed for synonymous site variability as decribed previously (Firth et al., 2011 PMID 21525127). The accession numbers of all sequences used in the analysis are available on request. Red lines indicate the probability (p-value) of obtaining not more than the observed number of synonymous substitutions, in a 25-codon sliding window, under a null model of neutral evolution at synonymous sites. Dashed grey lines indicate an approximate 5% false positive threshold after correcting for multiple tests (i.e. ~136 x 25-codon windows in the ~3400-codon polyprotein ORF). Statistically significant peaks in synonymous site conservation are indicative of overlapping functional elements, either coding or non-coding. Genome maps are shown for each clade. UTR lengths may be uncertain for less well-studied clades and are omitted for the UGSV/BOUV clade due to lack of sequence data. Known and predicted overlapping ORFs accessed via-1 PRF are shown in pink. Note that p-values can not be directly compared between different clades because the statistical significance (i.e. p-value) of observed reductions in synonymous site variabilty depends on the diversity of the specific sequence alignment being analyzed.

Fig 6. Predicted frameshift stimulatory elements in WESSV and SEPV.

Frameshifting is predicted to occur on a conserved G_GUU_UUU heptanucleotide (orange) in the NS2A-encoding region, stimulated by a 3'-adjacent stem-loop structure. The-1 frame stop codon is indicated in red. The WESSV RefSeq NC_012735 is shown; nucleotide differences in the SEPV RefSeq NC_008719 are indicated in blue.