Genetic diversity among Lassa virus strains

Abstract

The arenavirus Lassa virus causes Lassa fever, a viral hemorrhagic fever that is endemic in the countries of Nigeria, Sierra Leone, Liberia, and Guinea and perhaps elsewhere in West Africa. To determine the degree of genetic diversity among Lassa virus strains, partial nucleoprotein (NP) gene sequences were obtained from 54 strains and analyzed. Phylogenetic analyses showed that Lassa viruses comprise four lineages, three of which are found in Nigeria and the fourth in Guinea, Liberia, and Sierra Leone. Overall strain variation in the partial NP gene sequence was found to be as high as 27% at the nucleotide level and 15% at the amino acid level. Genetic distance among Lassa strains was found to correlate with geographic distance rather than time, and no evidence of a "molecular clock" was found. A method for amplifying and cloning full-length arenavirus S RNAs was developed and used to obtain the complete NP and glycoprotein gene (GP1 and GP2) sequences for two representative Nigerian strains of Lassa virus. Comparison of full-length gene sequences for four Lassa virus strains representing the four lineages showed that the NP gene (up to 23.8% nucleotide difference and 12.0% amino acid difference) is more variable than the glycoprotein genes. Although the evolutionary order of descent within Lassa virus strains was not completely resolved, the phylogenetic analyses of full-length NP, GP1, and GP2 gene sequences suggested that Nigerian strains of Lassa virus were ancestral to strains from Guinea, Liberia, and Sierra Leone. Compared to the New World arenaviruses, Lassa and the other Old World arenaviruses have either undergone a shorter period of diverisification or are evolving at a slower rate. This study represents the first large-scale examination of Lassa virus genetic variation.

FIG. 1

Maps of Guinea, Liberia, and Sierra Leone (A) and Nigeria (B) showing origins of Lassa virus strains.

FIG. 2

Phylogenetic tree inferred using MP criteria that shows relationships among Lassa virus strains. Lassa virus partial NP gene sequences were aligned with the corresponding regions of the Josiah (GenBank accession number J04324 [5]), GA391 (X52400 [17]), and LP (U80004 [11]) strains of Lassa virus as well as the corresponding regions of OLV virus (U34248 [9]), PIC virus (K02734 [4, 6]), Junin virus strain XJ (U80009 [11]), LCM virus strains Armstrong (M20869 [55]) and WE (M22138 [53]), Ippy virus (U80003 [11]), Mopeia virus strains AN20410 (U80005 [11]) and AN21366 (M33879 [63]), and Mobala virus strain 3076 (AF012530 [11]). Phylogenetic analysis of the partial NP gene data set was carried out using PAUP* 4.0, with characters optimized by accelerated transformation and gaps treated as missing data. The data set contained 667 characters, of which 423 were parsimony informative. The large number of taxa necessitated the use of a heuristic search that was performed using stepwise addition, with random addition (10 replicates) and tree bisection-reconnection branch swapping. The unweighted analysis yielded a single most parsimonious tree of 2,744 steps. The numbers adjacent to each node are percentage bootstrap support (500 replicates)/BSIs. Horizontal branch lengths are indicated by the scale bar. The four Lassa virus lineages are labeled I through IV. Representative strains from each of the four lineages used for further genetic analysis are shown in outline.

FIG. 3

Alignment of the S RNA intergenic region of Lassa virus strains 803213, LP, Josiah, and GA391. The dots indicate sequence identity with the 803213 strain. The dashes indicate alignment gaps.

FIG. 4

Unrooted phylogenetic trees inferred using ML criteria and NP, GP1, and GP2 data sets showing relationships among arenaviruses. Full-length Lassa virus (LAS) gene sequences were aligned with the corresponding gene sequences of the Josiah (J04324 [5]) and GA391 (X52400 [17]) strains of Lassa virus, OLV virus (U34248 [9]), PIC virus strains An3739 (K02734 [6]) and PV2P2 (U77601 [64]), Junin (JUN) virus strains MC2 (D10072 [27]) and XJ (U70799 [1]), LCM virus strains Armstrong (M20869 [55]), WE (M22138 [53]), and MX (Y16308), Mopeia (MOP) virus (M33879 [63]), Tacaribe (TCR) virus (M20304 [25]), Sabia (SAB) virus (U41071 [29]), and Machupo (MAC) virus (X62616 [34]). Phylogenetic estimates were obtained by heuristic search with the HKY85 (35) substitution model and ML estimates of nucleotide frequencies, proportion of invariant sites, and transition/transversion ratios. The log likelihoods for each tree are as follows: NP, −19,134.175; GP1, −9,446.293; and GP2, −6,918.156. Branch lengths are indicated by the scale bars. The letters label nodes in reference to the bootstrap and QP reliability values shown in Table 4.