Shaun D. Berry, Vincent N. Fondong, Christine Rey, Donald Rogan, Claude M. Fauquet, Judith K. Brown
Annals of the Entomological Society of America 97 (4), 852-859, (1 July 2004) https://doi.org/10.1603/0013-8746(2004)097[0852:MEFFDB]2.0.CO;2
KEYWORDS: begomovirus, cassava, cytochrome oxidase I, haplotype, whitefly
The Bemisia tabaci (Gennadius) complex contains the only known whitefly vector of plant-infecting begomoviruses, which are the causal agents of mosaic diseases of cassava in Africa and India. Widespread phenotypic variability, together with the absence of definitive morphological taxonomic characters for this whitefly complex, has confounded both the systematics and the study of its virus vector biology. Substantial genetic variability and phylogeographical relationships have been shown for phenotypic, but morphologically identical, variants of B. tabaci based on the mitochondrial (mt) cytochrome oxidase I (COI) sequence, leading to the suggestion that they represent a species complex. Here, phylogenetic relationships were explored, using the mtCOI sequence (780 bp) as a molecular marker, for B. tabaci collected from cassava plants in southern and western Africa, including Cameroon, Zambia, Mozambique, Zimbabwe, Swaziland, and South Africa. Maximum likelihood analyses of mtCOI sequences revealed that most B. tabaci examined were placed into one of three subgroups within the major sub-Saharan African clade, which also contains previously reported populations indigenous to Malawi and Uganda, and collectively shared on overall nucleotide (nt) identity at 88.9–99.7%. Two other reference populations, the monophagous Benin haplotype from Asystasia spp. and a B. tabaci from cassava in the Ivory Coast (IC), were the most divergent outliers of the sub-Saharan clade, each representing the only member of their respective clade (I and V), at the present time. Members of the sub-Saharan clade associated with cassava had as their closest relatives haplotypes I and II of the Mediterranean-Northern Africa clade, with which they shared a collective 84.2–92.9% nt identity (not including the IC cassava reference haplotype). In contrast, the sub-Saharan African clade diverged from the Americas and Southeast Asia/Far East clades at 79.7–85.1 and 77.5–84.9%, respectively. Within the sub-Saharan clade, subclade II contained B. tabaci from Zambia, Mozambique, South Africa, and Swaziland at 95–99% identity. The sub-Saharan subcluster III contained haplotypes from southern and western Africa. Counter to the otherwise phylogeographical relationships observed for cassava-associated B. tabaci from southern Africa, one and two populations from Cameroon (okra) and Zimbabwe (cassava), respectively, grouped with the major Mediterranean-North Africa clade, together with their closest relative associated with okra from IC, are included here as a reference sequence for the first time, with which they collectively formed a new, third subclade. Thus, phylogenetic analysis of B. tabaci mtCOI haplotypes examined thus far from the African continent has revealed five major cassava-associated haplotypes, which grouped primarily based on extant geography, with the exception of one and two collections from Cameroon and Zimbabwe, respectively. Hypotheses explaining the potential distributions of haplotypes are discussed.