Current molecular epidemiology of Lassa virus in Nigeria

Abstract

Recent Lassa virus strains from Nigeria were completely or partially sequenced. Phylogenetic analysis revealed the predominance of lineage II and III strains, the existence of a previously undescribed (sub)lineage in Nigeria, and the directional spread of virus in the southern part of the country. The Bayesian analysis also provided estimates for divergence times within the Lassa virus clade.

Fig. 1.

Phylogenetic analysis of Lassa virus. (Left panel) Phylogenies were inferred using complete nucleotide sequences of GP (1,473 nucleotides), NP (1,707 nucleotides), and L genes (6,654 nucleotides) of all Old World arenaviruses. For clarity of presentation, only the Lassa virus clade is shown; the complete trees are shown in Fig. S1 in the supplemental material. Analysis was performed with BEAST software with the following parameters: general time-reversible model with gamma-distributed sites (GTR+gamma), relaxed lognormal clock with mean substitution rate fixed at 1, 10⁷ steps with sampling every 10⁵ steps, and results of two independent runs combined (effective sampling size, >200 for all parameters). The data sets were also analyzed using PhyML software with the following parameters: GTR+gamma and consensus of 500 bootstrap trees (not shown). The topology of Bayesian and PhyML trees was congruent for a given gene. BEAST and PhyML support values are indicated on the branches (posterior/bootstrap). (Right panel) Phylogenies were inferred using partial nucleotide sequences of GP (237 nucleotides), NP (631 nucleotides), and L genes (342 nucleotides) of all Lassa virus strains. For clarity of presentation, only Nigerian Lassa virus strains are shown; the complete trees are shown in Fig. S2 in the supplemental material. Analysis was performed with BEAST software with the following parameters: GTR+gamma, strict clock, Bayesian skyline demographic model, 10⁷ steps with sampling every 10⁵ steps, and results of two independent runs combined (effective sampling size, >200 for all parameters). Trees obtained with an exponential growth model were essentially identical (not shown). Posterior values are indicated on the branches. Nodes representing the most recent common ancestors of the Lassa virus clade and lineages II, III, and IV are marked with letters A to D, respectively. GenBank accession number and year and place of isolation are shown with the strains. The origin of Lassa virus strains is indicated by a prefix: SL, Sierra Leone; LIB, Liberia; GUI, Guinea; IC, Ivory Coast; NIG, Nigeria. Lassa virus lineages are indicated with each tree.

Fig. 2.

Map of Nigeria showing sites of Lassa virus circulation. The possible spread of virus is indicated by dashed arrows. The Lassa virus lineages are indicated with the corresponding areas in which the virus is endemic. The potential lineage/sublineage represented by strain Nig05-A08 is indicated with a question mark. (Modified from a map by Daniel Dalet that is freely available at http://d-maps.com.)