Phylogeography of horseshoe bat sarbecoviruses in Vietnam and neighbouring countries. Implications for the origins of SARS-CoV and SARS-CoV-2

Abstract

Previous studies on horseshoe bats (Rhinolophus spp.) have described many coronaviruses related to SARS-CoV (SARSCoVr) in China and only a few coronaviruses related to SARS-CoV-2 (SARSCoV2r) in Yunnan (southern China), Cambodia, Laos and Thailand. Here, we report the results of several field missions carried out in 2017, 2021 and 2022 across Vietnam during which 1218 horseshoe bats were sampled from 19 locations. Sarbecoviruses were detected in 11% of faecal RNA extracts, with much more positives among Rhinolophus thomasi (46%). We assembled 38 Sarbecovirus genomes, including 32 SARSCoVr, four SARSCoV2r, and two recombinants of SARSCoVr and SARSCoV2r (RecSar), one showing a Spike protein very similar to SARS-CoV-2. We detected a bat co-infected with four coronaviruses, including two sarbecoviruses. Our analyses revealed that Sarbecovirus genomes evolve in Vietnam under strong geographical and host constraints. First, we found evidence for a deep separation between viruses from northern Vietnam and those from central and southern Vietnam. Second, we detected only SARSCoVr in Rhinolophus thomasi, both SARSCoVr and SARSCoV2r in Rhinolophus affinis, and only RecSar in Rhinolophus pusillus captured close to the border with China. Third, the bias in favour of Uracil in synonymous third codon positions of SARSCoVr extracted from R. thomasi showed a negative correlation with latitudes. Our results also provided support for an emergence of SARS-CoV in horseshoe bats from northern Yunnan and emergence of SARS-CoV-2 in horseshoe bats from northern Indochina subtropical forests (southern Yunnan, northern Laos and north-western Vietnam).

Keywords: Chiroptera; Southeast Asia; coronavirus genome; molecular evolution; phylogeny; recombination.