Extinction and rapid emergence of strains of dengue 3 virus during an interepidemic period.

Strains of dengue 3 (DEN-3) virus circulating in Thailand prior to 1992 appear to have disappeared from that location and to have been replaced by two new lineages which have evolved locally, rather than being introduced. Similar DEN-3 virus extinctions may have occurred previously in Thailand in 1962 and 1973. Although no causal relationship could be shown, this strain replacement event was accompanied by DEN-3 replacing DEN-2 as the serotype recovered most frequently from patients in Thailand. Although this implies a change in selection pressure, we found no evidence for positive natural selection at the level of either the E protein or the E protein gene. Further, the extinction of the pre-1992 strains and the appearance of the new lineages occurred during an interepidemic period, suggesting that a genetic bottleneck, rather than selection, might have been important in the emergence of these two new strains of virus. The pre-1992 DEN-3 virus lineage could still be found in 1998, to the west, in Myanmar. The ratio of nonsynonymous-to-synonymous nucleotide changes within a DEN-3 virus population from a single patient was less than the ratio among the consensus sequences of DEN-3 viruses from different patients, suggesting that many of the nonsynonymous nucleotide changes which occurred naturally in the E protein were deleterious and removed by purifying selection.

B cells are the principal circulating mononuclear cells infected by dengue virus.

Although dengue virus infects a variety of cells in vitro, little is known about cell types infected in vivo. Since blood is a readily accessible tissue, we chose to determine which circulating blood cells are infected by dengue viruses. We collected blood mononuclear cells from acutely ill dengue patients and separated the cells by flow cytometry into subsets for virus isolation. Cells were sorted into groups corresponding to the cluster designations CD3, CD14, CD16 and CD20. Virus was isolated from sorted groups by inoculation into Toxorhynchites splendens mosquitos. The majority of the virus was recovered from the CD20 or B cell positive subset. Little virus was isolated from monocytes, NK cells or T cells. Virus was isolated from B cells regardless of the age or sex of the patient, virus serotype isolated, or the patient's history of dengue virus infection. The location of cell associated virus was determined by proteolytic digestion of surface virus. There was an equal distribution of virus between the intracellular compartment and the surface of B cells. The intracellular localization of virus was confirmed by immunocytochemistry. Since this study focused on circulating cells, no inferences were made regarding infection of cells in solid tissues.