Discovery and biological confirmation of a highly divergent Tacaribe virus in metatranscriptomic data from neotropical bats

Abstract

First isolated from neotropical fruit bats in Trinidad in 1956, Tacaribe virus (TCRV) has rarely been detected since. We searched for New World arenavirus reads in roughly 5.7 million sequencing runs available on public databases using Serratus. We recovered a complete genome of a divergent TCRV in metatranscriptomic data derived from heart and eye tissue of an adult male Jamaican fruit-eating bat sampled in the Dominican Republic, 2014. In total, 2,733 reads were mapped resulting in mean coverages of 7.4-fold for the L and 10.2-fold for the S segment. Re-testing original bat specimens showed the highest viral loads in liver tissue (245 copies/mg). Sanger sequencing of PCR amplicons from liver confirmed correctness of and completed the genome recovered from metatranscriptomic data, revealing conserved arenavirus genomic organization, length, intergenic regions, and genome termini. The newly found TCRV strain tentatively named DOM2014 clustered in a basal sister relationship to all other known TCRV strains with which it shared between 83.3%-86.0% genomic and 91.8%-93.7% translated amino acid sequence identity across protein-coding regions. DOM2014 showed a conserved glycine, proline, proline, threonine (GPPT) nucleoprotein motif, which is essential for TCRV interferon β antagonism. Our data confirm the association of TCRV with the bat genus Artibeus put into question by lethal experimental infections and scarce bat-derived TCRV genomic data. Broad genetic diversity and geographic spread require assessments of TCRV strain-associated pathogenicity, particularly for DOM2014 as a highly divergent TCRV strain. Confirmation of genomic database findings by testing original specimens provides robustness to our findings and supports the usefulness of metatranscriptomic studies.

Importance: Clade B New World arenaviruses (NWA) include rodent-borne lethal hemorrhagic fever viruses, whereas Tacaribe virus (TCRV) stands out because of its detection in bats and its presumably low zoonotic potential. However, the bat association of TCRV was put into question by lethal experimental neotropical fruit bat infections and rare TCRV detection in bats. Scarce genomic data include near-identical viruses from Caribbean bats and ticks from the US sampled 50 years later. The prototype TCRV isolate used for experimental risk assessments has an extensive passage history in suckling mouse brains. Exploring the true genetic diversity, geographic distribution, and host range of bat-borne NWA is pivotal to assess their zoonotic potential and transmission cycles. We analyzed metatranscriptomic data for evidence of NWA identifying a highly divergent TCRV in bats and confirmed virus detection in original biological materials, supporting the association of TCRV with neotropical bats and warranting investigation of strain-associated TCRV pathogenicity.

Keywords: Tacaribe; arenaviruses; bats; genomics; virology; virus evolution.

FIG 1

Genome reconstruction and analyses. Genome coverage of TCRV strain DOM2014 compared to the TCRV reference genome for the L (A) and S segments (B). Coding regions and the GPPT motif are indicated. Pairwise genome comparison of DOM2014 with TCRV and TETV (C). L and S segments were concatenated before analyses in SSE version 1.4 (http://www.virus-evolution.org/Downloads/Software/) using a sliding window of 400 bases and an increment of 100 bases. Re-sequenced TCRV strain TRVL-11573 was included, which also represents the almost identical TCRV strain Florida. Organ distribution of TCRV strain DOM2014 (D). Viral copies were determined by real-time RT-PCR. RNA was extracted from approximately 30 mg of each tissue. Bat sampling was approved by the Dominican Ministry of Environment and Natural Resources.

FIG 2

Phylogeny of TCRV strains and genetically related clade B NWA. L segment shown in panel A, S segment shown in panel B. Arenavirus genomes were aligned in MEGA 11 using the MUSCLE algorithm. Maximum likelihood nucleotide sequence-based trees were calculated in MEGA11 using a general time reversible model with uniform substitution rates and a complete deletion option. For each tree, 500 replicates were calculated to provide bootstrap support. Included L and S segment sequences were TCRV strain Florida (MW150032, MW150033), TCRV TRVL-11573 (MT081317, MT081316), TCRV isolate A354 (ON648817, ON648821), TETV (ON648820, ON648824), Ocozocoautla de Espinosa virus (JN897398), Junín virus (NC005080, NC005081), Machupo virus (NC005079, NC005078), Sabiá virus (NC006313, NC006317), Chapare virus (NC010563, NC010562), and Guanarito virus (NC005082, NC005077). Countries are abbreviated according to ISO alpha-3 codes: BRA, Brazil; DOM, Dominican Republic; TTO, Trinidad; USA, United States of America.