Genomic and phylogenetic characterization of Brazilian yellow fever virus strains

Abstract

Globally, yellow fever virus infects nearly 200,000 people, leading to 30,000 deaths annually. Although the virus is endemic to Latin America, only a single genome from this region has been sequenced. Here, we report 12 Brazilian yellow fever virus complete genomes, their genetic traits, phylogenetic characterization, and phylogeographic dynamics. Variable 3' noncoding region (3'NCR) patterns and specific mutations throughout the open reading frame altered predicted secondary structures. Our findings suggest that whereas the introduction of yellow fever virus in Brazil led to genotype I-predominant dispersal throughout South and Central Americas, genotype II remained confined to Bolivia, Peru, and the western Brazilian Amazon.

Fig 1

(a) Brazilian YFV strains according to the genotype and cyclization patterns; (b) schematic representation of the five distinct patterns found for 3′NCR of 12 YFV Brazilian isolates. (II) 3′NCR/deleted YFVSACM1 and YFVSACM2; (III) long 3′NCR Uniq motif-plus (YFVSAUM)/deleted YFVSADM2; (IV) long 3′NCR/deleted YFVSADM2; (V) long 3′NCR/YFVSADM1; (VI) extralong 3′NCR (YFVSACM1/YFVSACM2). Pattern I corresponds to African strains exhibiting the RYF1, RYF2, RYF3, CS2, and 3′CYC motifs (I-a; West African strains) or lack of RYF2 repetition sequence (I-b; East and Central African strains). (c) Predicted secondary structures generated by the mfold program for the five distinct patterns (II to VI) described for YFV strains isolated in Brazil. Conserved structures are indicated by arrows, brackets, colors, or boxes. Abbreviations: CS, conserved sequences; RCS, repeated conserved sequence; SmLs, small loop; LSH, long stable hairpin; 3′CYC, cycling sequence within CS1 of the 3′NCR; 5′CYC, cycling sequence within the capsid (Cap) gene; HS, hairpin sequence; UAR, upstream AUG region; NCR, noncoding region.

Fig 1

(a) Brazilian YFV strains according to the genotype and cyclization patterns; (b) schematic representation of the five distinct patterns found for 3′NCR of 12 YFV Brazilian isolates. (II) 3′NCR/deleted YFVSACM1 and YFVSACM2; (III) long 3′NCR Uniq motif-plus (YFVSAUM)/deleted YFVSADM2; (IV) long 3′NCR/deleted YFVSADM2; (V) long 3′NCR/YFVSADM1; (VI) extralong 3′NCR (YFVSACM1/YFVSACM2). Pattern I corresponds to African strains exhibiting the RYF1, RYF2, RYF3, CS2, and 3′CYC motifs (I-a; West African strains) or lack of RYF2 repetition sequence (I-b; East and Central African strains). (c) Predicted secondary structures generated by the mfold program for the five distinct patterns (II to VI) described for YFV strains isolated in Brazil. Conserved structures are indicated by arrows, brackets, colors, or boxes. Abbreviations: CS, conserved sequences; RCS, repeated conserved sequence; SmLs, small loop; LSH, long stable hairpin; 3′CYC, cycling sequence within CS1 of the 3′NCR; 5′CYC, cycling sequence within the capsid (Cap) gene; HS, hairpin sequence; UAR, upstream AUG region; NCR, noncoding region.

Fig 2

(A) Predicted 3D structure of the YFV E protein. The protein domains I, II, and III and fusion-peptide subdomain are shown in circles. Neutralizing epitopes are represented as red spheres along the E protein domains. (B) The E protein DIII is highlighted; comparison between ASIBI (blue) and YFV BeH 413820 (green) 3D structures. (B1) Conformational changes in YFV BeH 413820 strain due to amino acid substitution (K331R). (B2a) YFV BeH 413820 strain; N359 and E360 amino acids related to rescue of virulence in attenuated strains. (B2b) Amino acid changes in the DIII virulence-related epitope E360D. Note the hydrogen bond (dotted line) between amino acids N359 and D360 in the YFV strain BeH 413820 that are related to virulence. The number 4.81 corresponds to the distance in angstroms revealing the presence of a hydrogen bond, which is associated with viral persistence. (B3) TGD motif revealing conformational changes for ASIBI (blue) and the wild-type YFV BeH 413820 strain (green).

Fig 3

(a) Bayesian maximum clade credibility tree demonstrating the phylogenetic relationships among the YFV complete genomes. Major groups are indicated (I and II). Subgroups according to geographic location (West Africa, East Africa, Trinidad, and Brazil) are colored. Numbers under the bar represent the timing scaled for the most recent common ancestor. Node ages and substitution rates (substitutions/site/year) are placed over each main node in the tree. Numbers within parentheses and brackets represent the 95% HPD Bayesian posterior probabilities and dispersion rates expressed in percentages, respectively. Spatiotemporal diffusion of YFV throughout Africa to the Americas (b). Temporally framed snapshots of the dispersal patterns and respective 95% Bayesian credibility regions (BCR) are indicated (light-green areas) for the years 1940 (c), 1970 (d), and 2000 (e) (Maps in panels b to e reproduced from Google Earth [www.earth.google.com/].)