Ecoepidemiology and complete genome comparison of different strains of severe acute respiratory syndrome-related Rhinolophus bat coronavirus in China reveal bats as a reservoir for acute, self-limiting infection that allows recombination events

Abstract

Despite the identification of severe acute respiratory syndrome-related coronavirus (SARSr-CoV) in Rhinolophus Chinese horseshoe bats (SARSr-Rh-BatCoV) in China, the evolutionary and possible recombination origin of SARSr-CoV remains undetermined. We carried out the first study to investigate the migration pattern and SARSr-Rh-BatCoV genome epidemiology in Chinese horseshoe bats during a 4-year period. Of 1,401 Chinese horseshoe bats from Hong Kong and Guangdong, China, that were sampled, SARSr-Rh-BatCoV was detected in alimentary specimens from 130 (9.3%) bats, with peak activity during spring. A tagging exercise of 511 bats showed migration distances from 1.86 to 17 km. Bats carrying SARSr-Rh-BatCoV appeared healthy, with viral clearance occurring between 2 weeks and 4 months. However, lower body weights were observed in bats positive for SARSr-Rh-BatCoV, but not Rh-BatCoV HKU2. Complete genome sequencing of 10 SARSr-Rh-BatCoV strains showed frequent recombination between different strains. Moreover, recombination was detected between SARSr-Rh-BatCoV Rp3 from Guangxi, China, and Rf1 from Hubei, China, in the possible generation of civet SARSr-CoV SZ3, with a breakpoint at the nsp16/spike region. Molecular clock analysis showed that SARSr-CoVs were newly emerged viruses with the time of the most recent common ancestor (tMRCA) at 1972, which diverged between civet and bat strains in 1995. The present data suggest that SARSr-Rh-BatCoV causes acute, self-limiting infection in horseshoe bats, which serve as a reservoir for recombination between strains from different geographical locations within reachable foraging range. Civet SARSr-CoV is likely a recombinant virus arising from SARSr-CoV strains closely related to SARSr-Rh-BatCoV Rp3 and Rf1. Such frequent recombination, coupled with rapid evolution especially in ORF7b/ORF8 region, in these animals may have accounted for the cross-species transmission and emergence of SARS.

FIG. 1.

Map showing the locations of bat sampling and tagging in Hong Kong. Squares represent the locations where bats were positive for SARSr-Rh-BatCoV, dark circles represent locations where bats were positive for Rh-BatCoV HKU2, and triangles represent locations where the bats were positive for both Rh-BatCoV HKU2 and SARSr-Rh-BatCoV. The percentages indicate the proportion of bats positive for SARSr-Rh-BatCoV, Rh-BatCoV HKU2, or SARSr-Rh-BatCoV/Rh-BatCoV HKU2 at each location. Blank circles represent locations negative for SARSr-Rh-BatCoV and Rh-BatCoV HKU2. The red circle represents the location of Shenzhen Dongman market (SZDM) in China where civet SARSr-CoV was first identified. The arrows indicate the direction of migration of Chinese horseshoe bats as demonstrated in the tagging exercise.

FIG. 2.

Schematic diagram showing the number of Rhinolophus sinicus (Chinese horseshoe) bats tagged and recaptured, and the presence of coronaviruses among these tagged bats. The numbers in boldface type indicate the number of bats successfully recaptured. The numbers in roman type (not boldface type) following dashed lines are the numbers of bats not recaptured in subsequent visits. The numbers in parentheses are the number of recaptured bats positive for coronaviruses. CoV +ve, coronavirus positive; CoV -ve, coronavirus negative.

FIG. 3.

Bootscan analysis using the genome sequence of civet SARSr-CoV strain SZ3 as the query sequence. Bootscanning was conducted with Simplot version 3.5.1 (F84 model; window size, 1,500 bp; step size, 300 bp). SARSr-Rh-BatCoV strain Rf1, SARSr-Rh-BatCoV strain Rp3, and SARSr-Rh-BatCoV strain Rm1 were examined by bootscan analysis.

FIG. 4.

(A) Bootscan analysis using the genome sequence of SARSr-Rh-BatCoV strain Rf1 as the query sequence (A) and phylogenetic analysis of its partial sequences to the corresponding regions in other SARSr-CoVs (indicated by the letters B to D above the graph). Bootscanning was conducted with Simplot version 3.5.1 (F84 model; window size, 1,500 bp; step size, 300 bp) on a gapless nucleotide alignment, generated with ClustalX. SARSr-Rh-BatCoV strain 279/04 (279), civet SARSr-CoV strain SZ3, and SARSr-Rh-BatCoV strain HKU3-1 were examined by bootscan analysis. (B to D) Phylogenetic trees were constructed for the regions corresponding to positions 16400 to 20700 (B), 20700 to 25000 (C), and 25000 to 3′ end (D) by the neighbor-joining method using Kimura's two-parameter correction, and bootstrap values were calculated from 1,000 trees. Shaded strains represent strains included in bootscan analysis. Hel, helicase. Bars, 0.01 nucleotide substitution (B and C) or 0.005 nucleotide substitution (D).

FIG. 5.

(A) Bootscan analysis using the genome sequence of SARSr-Rh-BatCoV strain Rf1 as the query sequence (A) and phylogenetic analysis of its partial sequences to the corresponding regions in other SARSr-CoVs (indicated by the letters B to D above the graph). Bootscanning was conducted with Simplot version 3.5.1 (F84 model; window size, 1500 bp; step size, 300 bp) on a gapless nucleotide alignment, generated with ClustalX. SARSr-Rh-BatCoV strain Rm1, SARSr-Rh-BatCoV strain 273/04 (273), and SARSr-Rh-BatCoV strain Rp3 were examined by bootscan analysis. (B to D) Phylogenetic trees were constructed for the regions before position 18300 (B), positions 18300 to 19900 (C), and after position 19900 (D) by the neighbor-joining method using Kimura's two-parameter correction, and bootstrap values were calculated from 1,000 trees. Shaded strains represent strains included in bootscan analysis. Bar, 0.01 nucleotide substitution.

FIG. 6.

Estimation of the time of interspecies transmission of SARSr-CoV. Squares denote the MRCA of all SARSr-CoV (1972), the MRCA of human/civet SARSr-CoV and the closest SARSr-Rh-BatCoV (1995), and the MRCA of human and civet SARSr-CoV (2001), respectively.