Isolation of ACE2-dependent and -independent sarbecoviruses from Chinese horseshoe bats

Abstract

While the spike proteins from severe acute respiratory syndrome coronaviruses-1 and 2 (SARS-CoV and SARS-CoV-2) bind to host angiotensin-converting enzyme 2 (ACE2) to infect cells, the majority of bat sarbecoviruses cannot use ACE2 from any species. Despite their discovery almost 20 years ago, ACE2-independent sarbecoviruses have never been isolated from field samples, leading to the assumption these viruses pose little risk to humans. We have previously shown how spike proteins from a small group of ACE2-independent bat sarbecoviruses may possess the ability to infect human cells in the presence of exogenous trypsin. Here, we adapted our earlier findings into a virus isolation protocol and recovered two new ACE2-dependent viruses, RsYN2012 and RsYN2016A, as well as an ACE2-independent virus, RsHuB2019A. Although our stocks of RsHuB2019A rapidly acquired a tissue-culture adaption that rendered the spike protein resistant to trypsin, trypsin was still required for viral entry, suggesting limitations on the exogenous entry factors that support bat sarbecoviruses. Electron microscopy revealed that ACE2-independent sarbecoviruses have a prominent spike corona and share similar morphology to other coronaviruses. Our findings demonstrate a broader zoonotic threat posed by sarbecoviruses and shed light on the intricacies of coronavirus isolation and propagation in vitro. IMPORTANCE Several coronaviruses have been transmitted from animals to people, and 20 years of virus discovery studies have uncovered thousands of new coronavirus sequences in nature. Most of the animal-derived sarbecoviruses have never been isolated in culture due to cell incompatibilities and a poor understanding of the in vitro requirements for their propagation. Here, we built on our growing body of work characterizing viral entry mechanisms of bat sarbecoviruses in human cells and have developed a virus isolation protocol that allows for the exploration of these understudied viruses. Our protocol is robust and practical, leading to successful isolation of more sarbecoviruses than previous approaches and from field samples that had been collected over a 10-year longitudinal study.

Keywords: bat; coronavirus; cross-species transmission; sarbecovirus; zoonosis.

Fig 1

Isolation of clade 1 and clade 2 RBD sarbecoviruses on human cell lines. (A) Field samples were used to inoculate Huh-7 cells in the presence of trypsin. Viral titers were quantified in supernatants by qRT-PCR. (B) Whole-genome nucleotide sequences were compared to RsWIV1 with a sequence similarity plot. Open reading frame (ORF) positions are indicated under the x-axis. (C) Cladogram analysis of RBD amino acid sequences (corresponding to SARS-COV spike aa323–510) for sarbecoviruses. RBD indels and receptor preferences are indicated for each functional phylogenetic clade. Viruses isolated in this study are in bold font. (D) Viral isolates were inoculated on indicated cell cultures, and viral replication was monitored by qRT-PCR. n/a, not available; qRT-PCR, quantitative RT-PCR.

Fig 2

Clade 2 RBD sarbecoviruses do not use any known coronavirus receptors for cell entry. (A) BHK-21 cells were transfected with human orthologs of known coronavirus receptors and then infected with viral isolates. Replication was quantified by qRT-PCR. (B) VSV-based pseudotyped virus particles bearing chimeric SARS-CoV spikes with the indicated virus RBDs were generated in HEK 293T cells and concentrated in OptiPrep. Spike was detected in cell lysates and pseudotyped virus particles by probing for FLAG. (C) Huh-7 cells or cells transduced to express human ACE2 were infected with pseudotyped virus particles, and luciferase was measured as a readout for cell entry. (D) BHK-21 cells were transfected with the indicated bat ACE2 alleles and infected with viral isolates. Replication was monitored by qRT-PCR.

Fig 3

Clade 2 RBD virus adaptation to the cell culture. (A) V976L mutation emerged in RsHuB2019A virus stocks. (B) Pseudotyped virus particles were produced with full-length spike wild type (WT) or the V967L mutant. Spike was detected in producer cells and pseudotyped virus particles by western blot for FLAG. (C) Indicated cells were infected with pseudotyped virus particles in the presence or absence of trypsin. (D) Schematic overview of the dual-reporter fusion assay developed for this study. T7 polymerase drives the expression of GFP and luciferase separated by self-cleaving 2A peptide from porcine teschovirus-1 (P2A). (E) HEK 293T cells expressing receptor or (F) empty vector and T7 polymerase were combined with cells expressing spike and the T7-driven reporter. Luciferase was measured as a readout for cell fusion. Dotted lines indicate data from 1:4 ratio of receptor: spike cells. (G) Overview of RsHuB2019A spike with in silico predicted trypsin digest sites indicated. Location of V976L is indicated in green. (H) Concentrated pseudotyped virus particles were combined with a wide range of trypsin dilutions or (I) a fine range of trypsin dilutions and incubated at 37°C. Spike digestion was assessed by western blot for the FLAG epitope.

Fig 4

Electron microscopy of purified viral isolates. Viral stocks for (A) RsYN2012, (B) RsYN2016A, or (C) RsHuB2019A were visualized by transmission electron microscopy. The bottom images were taken at a higher magnification to show detail.