1999 Workshop Abstracts | Virus Evolution Home Page | Plant Biology Home Page

Rapid and dynamic evolution of quasispecies clouds during equine infectious anemia virus infection

Praisith Baccam^1,2, Micheal Belshan², James Cornette¹, and Susan Carpenter^2
1Department of Mathematics and ²Department of Veterinary Microbiology and Preventive Medicine,
Iowa State University, Ames, Iowa 50011

Equine infectious anemia virus (EIAV) belongs to the lentivirus subfamily of retroviruses and produces a rapid and variable disease course in infected horses. Regions of genetic hypervariability have been observed in the viral genome including the gene encoding the regulatory protein Rev. Rev acts to regulate nucleocytoplasmic transport of incompletely spliced mRNAs which encode the viral structural genes and serve as the progeny RNA. Defects in Rev result in the accumulation of multiply spliced mRNAs, down regulation of viral replication, and decreased expression of viral structural genes. In vivo, down regulation of virus gene expression may be one strategy of virus persistence. Thus, we undertook a longitudinal study to examine the genetic variation in Rev quasispecies at different stages of clinical disease in a pony experimentally infected with the highly pathogenic Wyoming strain of EIAV. The infected pony experienced 8 recurring fever periods through the first 135 days post infection (DPI), after which he was generally afebrile except for two late fevers at 565 and 799 DPI. The serum used to infect the pony as well as eleven sequential sera samples were used to examine variation in rev at different stages of disease, including the acute febrile, recurrent febrile, afebrile, and late febrile stages. Viral RNA isolated from sera was used to amplify rev sequences by RT-PCR, and amplicons were cloned and individually sequenced. This study included 61 clones from the inoculum and approximately 24 clones from each sequential sera samples. Of the 322 clones sequenced, we identified 147 nucleotide and 101 Rev amino acid variants.

The translated Rev variants were used to construct a phylogenetic tree by the neighbor-joining distance method using the program MEGA. The branching pattern of the tree identified 2 major clades, designated A and B, and a minor clade, designated C. All three clades contained variants from the inoculum as well as from the recurrent febrile, afebrile, and late febrile stages of disease. In addition, all clades contained Rev variants that were detected at multiple stages of disease. Despite these similarities, there were striking differences among the three clades. For example, Rev variants isolated from the recurrent or late febrile stages accounted for 65% of all variants in Clade A, but accounted for less than 20% of variants in Clade B. In addition, variants in Clade B were genetically similar, with only 3 amino acids separating the most distinct variants. In contrast, variants in Clade A were separated by up to 8 amino acid changes. The topology of the major clades suggested significant differences in their pattern of evolution. Clade A was characterized by a temporal pattern of evolution, where variants evolved from pre-existing variants. In contrast, a radial pattern of evolution was observed in Clade B. Specifically, 19 variants isolated from all stages of disease appeared to have arisen from one variant, designated R1, which was the predominant variant in the inoculum and in most of the afebrile stage. Taken together, the phylogenetic analysis suggested that there may be two distinct groups of Rev variants that co-exist during the course of infection. Moreover, these two populations appeared to have evolved in different manners.

A cluster analysis was used to further investigate whether rev variants co-exist as two distinct populations. This analysis took advantage of the temporal information by analyzing clusters of rev variants present at each time point. The number of nucleotide differences between the variants was used to define the cluster radius. A computer program was written in which each sequence was considered as the center of a cluster. The optimal clusters were defined as those which maximized the number of variants within the cluster, while minimizing the radius of the cluster. The variant at the center of the clusters was the most representative of the sequences within that cluster. Results of the analysis indicated that each time point could be described by one to three clusters, each with a radius of 3 nucleotides or less. Early during infection, the clusters intersected slightly, but as the disease progressed they grew distinctly further apart. All clusters had a similar radius value; however, the proportion of variants within each cluster was variable. Thus, we calculated a relative cluster size to indicate the proportion of variants within the cluster.

The analysis identified three different clusters, designated Clouds 1, 2, and 3, which generally corresponded to Clades A, B, and C, respectively, from the phylogenetic tree. The center genotype of Cloud 1 continually changed during the course of disease, with 7 different variants observed at the 12 time points. Cloud 2, on the other hand, was more stable with R1 acting as the central genotype in 8 of the 10 times points when Cloud 2 was detected. Cloud 3 was detected only in the afebrile and late febrile stages as a minor cloud. Thus, the cluster analysis supported the phylogenetic analysis with respect to the grouping of the sequences and the pattern of evolution over time. The relative size of the quasispecies clouds was dynamic over the course of infection. For example, Cloud 1 variants were minor during the initial acute febrile stage, grew in frequency to predominate during the recurrent febrile stage, decreased in frequency in the afebrile stage, and re-emerged to predominate in the late febrile stage. Interestingly, Cloud 2 variants were predominant at stages of disease when Cloud 1 variants were minor, and were minor at stages when Cloud 1 variants were predominant. This suggested that sequences in Cloud 1 and Cloud 2 differed in selective advantage. To test this, the variants from the center of the clusters were biologically characterized using a chloramphenicol acetyl transferase reporter assay to quantify Rev activity. Rev variants representative of Clade B and Cloud 2 had significantly lower activity than the Rev variants representative of Clade A and Cloud 1.

The expansion and contraction of the viral sub-populations appears to be correlated with changes in clinical disease. Viral quasispecies are generally considered as a single population of closely related genotypes. However, our analyses suggest that Rev quasispecies co-exist as two distinct sub-populations that differ in selective advantage. Further biological characterization of genotypes representative of each population may offer insight into the selective factors and molecular mechanisms important in lentivirus pathogenesis and persistence.

Abstracts from the Virus Evolution Workshop
Ardmore, OK
October 21 - 24th, 1999

Virus Evolution Workshop - Main Page
Poster Presentations

To contact the organizers:
e-mail: mroossinck@noble.org

Dr. Marilyn Roossinck
Plant Biology Division
The Noble Foundation
P.O. Box 2180
Ardmore, OK 73402

phone: 580 224-6630