Emergence and pandemic spread of small ruminant lentiviruses

Abstract

Small ruminant lentiviruses (SRLVs) cause chronic, persistent infections in populations of domestic sheep (Ovis aries) and goats (Capra hircus) worldwide. The vast majority of SRLV infections involve two genotypes (A and B) that spread in association with the emergence of global livestock trade. However, SRLVs have likely been present in Eurasian ruminant populations since at least the early Neolithic period. Here, we use phylogenetic and phylogeographic approaches to reconstruct the origin of pandemic SRLV strains and infer their historical pattern of global spread. We constructed an open computational resource ('Lentivirus-GLUE') via which an up-to-date database of published SRLV sequences, multiple sequence alignments (MSAs), and sequence-associated metadata can be maintained. We used data collated in Lentivirus-GLUE to perform a comprehensive phylogenetic investigation of global SRLV diversity. Phylogenies reconstructed from genome-length alignments reveal that the deep divisions in the SRLV phylogeny are consistent with an ancient split into Eastern (A-like) and Western (B-like) lineages as agricultural systems disseminated out of domestication centres during the Neolithic period. These findings are also consistent with historical and phylogeographic evidence linking the early 20^th century emergence of SRLV-A to the international export of Central Asian Karakul sheep. Investigating the global diversity of SRLVs can help reveal how anthropogenic factors have impacted the ecology and evolution of livestock diseases. The open resources generated in our study can expedite these studies and can also serve more broadly to facilitate the use of genomic data in SRLV diagnostics and research.

Keywords: emergence; evolution; iatrogenesis; lentivirus; retrovirus.

Figure 1.

Dissemination of Karakul and early SRLV outbreaks. Arrows on the map indicate documented international exportations of Karakul during the period 1900–70. During this period, a cluster of similar disease syndromes emerged in sheep, all of which are either known or considered likely to have been caused by SRLV-A infection (see inset key). Their geographic locations are approximately indicated on the map following the key. ‘Graaff-Reinet disease’ was first reported in South Africa in 1915 (De Kock 1929), and the same year, ‘Montana sheep disease’ was reported for the first time in the USA (Marsh 1923). In Europe, ‘maedi-visna’ (Sigurdsson 1954) and ‘La Bouhite’ (France; Lucam 1942) both emerged during the 1930s and 40s. Further outbreaks of maedi-visna were reported in Canada in 1956 and in East Africa (Kenya) in 1966 (Table S4).

Figure 2.

The MSA hierarchy in Lentivirus-GLUE. The GLUE software framework allows hierarchical linking of MSAs to reflect taxonomic relationships (Singer et al. 2018). The figure shows a schematic representation of the hierarchical ‘MSA tree’ data structure in Lentivirus-GLUE, which recapitulates the relationships among lentivirus sequences. Numbers shown adjacent to nodes/tips correspond to rows in Table 1. Abbreviations: SIV = simian immunodeficiency virus; JDV = Jembrana disease virus; FIV = feline immunodeficiency virus; EIAV = equine infectious anaemia virus.

Figure 3.

Rooted SRLV phylogenies support an ancient East–West split. (A) A ML phylogeny showing the evolutionary relationships between lentiviruses. The phylogeny is based on an alignment of Gag–Pol polyproteins spanning 1,397 amino acid residues in the Gag–Pol open reading frame and the rtREV substitution model. (B) A ML phylogeny showing the evolutionary relationships between SRLV genotypes and rooted on the BLVs. The phylogeny is based on an alignment of Gag–Pol polyproteins spanning 1,598 amino acid residues in Gag–Pol and the rtREV substitution model. (C) A ML phylogeny showing the evolutionary relationships between eighty-five genome-length SRLV sequences (8,491 nucleotide residues), constructed using the general time reversible (GTR) model of nucleotide substitution, and rooted on SRLV-E. Two of the major clades contained within this tree are shown enlarged, with taxa labels, in panels (D) A-like viruses and (E) B-like viruses. Scale bars indicate evolutionary distance in substitutions per site. Asterisks indicate nodes with ML bootstrap support >85 per cent, based on 1,000 replicates. Abbreviations: BLV = bovine lentivirus; SIV = simian immunodeficiency virus; JDV = Jembrana disease virus; FIV = feline immunodeficiency virus; EIAV = equine infectious anaemia virus; rtREV = retrovirus-specific model.

Figure 4.

SRLV diversity in the Fertile Crescent region. The map in panel (A) indicates the ‘Fertile Crescent’ region (shaded). The overlapping square indicates the location of the area shown in panel (B). Circles indicate sampling sites within the Levant region: A = Northern Lebanon (Qornet el Sawda, Arez); B = Bekaa Valley (three sites: North (Aammiq), West (Swairi, Manara, Rashaya), and Eastern (Nahle, Maqne, Knaisse)); D = Northern Jordan and Jordan Valley. (C) ML phylogeny based on an alignment of SRLV capsid(CA) gene nucleotide sequences and showing the evolutionary relationships of Lebanese SRLV isolates to globally sampled SRLVs. The tree was reconstructed using the general time reversible (GTR) model of nucleotide selection as selected using the likelihood ratio test and is midpoint rooted for display purposes. Taxa labels show GenBank accession numbers. Isolate names are shown for complete genome isolates (denoted by coloured circles adjacent taxa labels). Asterisks indicate nodes with ML bootstrap support >85 per cent, based on 1,000 bootstrap replicates. Brackets to the right indicate genotypes and subgroups.

Figure 5.

A proposed model of SRLV-A origin and early spread. A schematic representation of our proposed model of SRLV-A evolution. The broken scale bar shows time on three different scales, from left to right as follows: (1) thousands of years (Kya); (2) centuries; and (3) decades. Selected periods of ancient and contemporary human history are indicated. The underlying tree (lighter - see key), represents the historical dissemination of sheep breeds out of the Fertile Crescent (F.C.) region and into other geographic areas as indicated. The overlaid tree (darker ) shows the hypothesised pattern of SRLV-A dissemination among global sheep populations. Historical outbreaks of SRLV-associated disease are indicated by numbers in square markers, as follows: (1) Graaff-Reinet disease; (2) Montana sheep disease; (3) Maedi-Visna; and (4) La Bouhite. Squares adjacent arrows containing roman numerals indicate the international export of Karakul stocks between continental regions as follows: (i) Central Asia to Europe; (ii) Central Asia to North America; (iii) Europe to Southwest Africa. (iv) Mainland Europe to The arrowed circle containing the ‘?’ symbol indicates a possible introduction of SRLV-A in Africa in thin-tailed sheep during the 1960s-70s. Please note that, other than this, the figure does not represent the extensive international trade that occurred from ∼1960 onwards, only the earlier export events pertinent to our model.