Characterization of the Candiru antigenic complex (Bunyaviridae: Phlebovirus), a highly diverse and reassorting group of viruses affecting humans in tropical America

Abstract

The genus Phlebovirus of the family Bunyaviridae consists of approximately 70 named viruses, currently assigned to nine serocomplexes (species) based on antigenic similarities. Sixteen other named viruses that show little serologic relationship to the nine recognized groups are also classified as tentative species in the genus. In an effort to develop a more precise classification system for phleboviruses, we are attempting to sequence most of the named viruses in the genus with the goal of clarifying their phylogenetic relationships. In this report, we describe the serologic and phylogenetic relationships of 13 viruses that were found to be members of the Candiru serocomplex; 6 of them cause disease in humans. Analysis of full genome sequences revealed branching inconsistencies that suggest five reassortment events, all involving the M segment, and thus appear to be natural reassortants. This high rate of reassortment illustrates the inaccuracy of a classification system based solely on antigenic relationships.

Fig. 1.

Geographic distribution of Candiru viruses in Central America and North America. Each number indicates a virus in the Candiru complex and the approximate geographic location of its isolation.

Fig. 2.

Phylogenetic analysis of the available sequences of phlebovirus L ORF. A set of 66 L phlebovirus sequences comprising all partial or complete sequences from GenBank available on 1 October 2010 were aligned using the CLUSTAL algorithm (as implemented in the MEGA package version 3) at the amino acid level with additional manual editing to ensure the highest possible quality of alignment. NJ analysis at the amino acid level was performed due to the observed high variability of the underlying nucleotide sequences. The statistical significance of tree topology was evaluated by bootstrap resampling of the sequences 1,000 times. Phylogenetic analyses were performed using MEGA software (33). Sequences marked with black dots corresponded to partial sequences.

Fig. 3.

Phylogenetic analysis of the available sequences of phlebovirus M ORF. A set of 118 M phlebovirus sequences comprising all partial or complete sequences from GenBank available on 1 October 2010 were analyzed under conditions identical to those for Fig. 2. Sequences marked with black dots correspond to partial sequences.

Fig. 4.

Phylogenetic analysis of the available sequences of phlebovirus S ORF. (Left panel) N ORF. A set of 127 N phlebovirus sequences comprising all partial or complete sequences from GenBank available on October 1, 2010 were analyzed under identical conditions to Fig. 2. (Right panel) NS ORF. A set of 94 NS phlebovirus sequences comprising all partial or complete sequences from GenBank available on 1 October 2010 were analyzed under conditions identical to those for Fig. 2.

Fig. 5.

Phylogenetic analysis of the complete sequences of the Candiru complex. All Candiru sequences were aligned using the CLUSTAL algorithm at the amino acid level with additional manual editing to ensure the highest possible quality of alignment. NJ analysis at the nucleotide level was performed. The statistical significance of tree topology was evaluated by bootstrap resampling of the sequences 1,000 times. Phylogenetics analyses were performed using MEGA software (33). Three clades were identified: clade A (green shading), clade B (blue), and clade C (red).

Fig. 6.

Pairwise sequence analysis. To establish a potential cutoff for phlebovirus classification, we pursued pairwise sequence comparison of all published phlebovirus sequences. Calculations were performed using MEGA software (33) to calculate the p-distance of each segment at the nucleotide level. p-distance values were grouped in three groups representing: blue, distances among different strains of the same virus (intraspecies); red, distances between members of different species (intergroup); and green, distances among members of the Candiru complex.

Fig. 7.

Serological distance among members of the Candiru complex as determined by CF. CF values were transformed to rAHM, as previously described for influenza virus (40, 54). rAHM values were grouped in two groups representing: distances among strains belonging to the same clade, and distances between strains of different clades. In the median log scores box plot, the whiskers represent minimum and maximum ranges. Lines in box represent the first interquartile range, followed by the median, and third interquartile range. The rAHM index among members of the same clade (n = 34) and the rAHM index between members of different clades (n = 44) are shown. The data were log transformed to better approximate a normal distribution.