Discovery of Novel Bat Coronaviruses in South China That Use the Same Receptor as Middle East Respiratory Syndrome Coronavirus

Between the years 2012 and 2015, we surveyed 1,059 bats belonging to 13 genera collected from Guangdong, Guangxi, and Sichuan provinces in China. We performed reverse transcription-PCR (RT-PCR) on these specimens and found 89 strains of lineage C BetaCoVs from among 88 positive specimens (Table 1 and Fig. 1A). These lineage C BetaCoVs included (i) 23 (25%) strains closely related to HKU4 from 115 samples from bamboo bats (Tylonycteris pachypus and Tylonycteris spp.); (ii) 15 (10%) strains closely related to HKU5 from 149 samples from pipistrelle bats (Pipistrellus abramus, Pipistrellus pipistrellus, Pipistrellus minus, and Pipistrellus spp.); and (iii) 50 (14%) strains closely related to MERS-CoV from 364 samples, including pipistrelle bats (P. abramus and P. pipistrellus), great evening bats (Ia io), and particolored bats (Vespertilio superans) (Table 1 and Fig. 1B). Lineage C BetaCoVs were not detected from other tested bat species (Table 1). We selected four of these strains that exhibited >90% sequence identity in the 440-bp RdRp region to that of MERS-CoV. After sequencing the complete RdRp and S genes of these viral strains, we selected the two that were the most closely related to MERS-CoV for sequencing the entire full-length genome.

Both of the sequenced strains were derived from I. io samples and were named BtCoV/Ii/GD/2013-845 and BtCoV/Ii/GD/2014-422. Both genomes are 30,113 nucleotides long, with G+C contents of 40.80% and 40.98%, respectively. They exhibit genomic structures and transcription regulatory sequences (TRS) that are identical to those of other lineage C BetaCoVs (Fig. 1C and Table 2). The only exceptions are ORF4b and ORF8b, which do not have TRS preceding them; hence, these two open reading frames (ORFs) may be translated from bicistronic mRNAs (Table 2; subgenome analysis data not shown). The replicase ORF1ab is ∼21,300 nucleotides long and encodes a number of putative proteins, including Nsp3 (papain-like protease; PLpro), Nsp5 (chymotrypsin-like protease; 3CLpro), Nsp12 (RdRp), Nsp13 (helicase), and other proteins of unknown function (Table 3). A separate comparison of the amino acid sequences of seven conserved ORF1ab domains is presented in Table 4, as suggested by the International Committee on Taxonomy of Viruses (ICTV) for formal CoV species delineation. The sequence identities of the seven concatenated domains of ORF1ab between these two novel MERS-related CoVs and MERS-CoV are around 90%, which is the threshold for defining a single viral species as proposed by the ICTV.

We next examined the complete genomic sequences of BtCoV/Ii/GD/2013-845 and BtCoV/Ii/GD/2014-422 (Table 4; see also Fig. 3A). The two genomes share only 88.1% nucleotide identity, suggesting considerable polymorphism between these two strains. The K_a/K_s ratios (where K_s is the number of synonymous substitutions per synonymous site and K_a is the number of nonsynonymous substitutions per nonsynonymous site) are generally low in the two MERS-related CoVs for all genes except ORF8b, suggesting that the remaining genes have evolved under purifying selection (data not shown). Particularly divergent are the 5′-terminal regions of the S gene, as well as of ORF4a, ORF4b, and ORF5. As the 5′-terminal region of the S gene encodes the RBD, it is interesting that two CoVs isolated from the same host species in the same sampling location contain divergent RBDs. Overall, the genomic sequences of both BtCoV/Ii/GD/2013-845 and BtCoV/Ii/GD/2014-422 are more similar to that of MERS-related CoV than to those of other lineage C BetaCoVs.

We further investigated the individual genes and predicted protein sequences of the two novel MERS-related CoVs (Table 4). The membrane (M) protein is the most conserved, with an identity of 95.4% between the two novel strains and MERS-CoV. ORF3 is the least conserved, with 50.0% identity between the two novel strains and MERS-CoV. Phylogenetic analysis of polyprotein 1 and the E, M, and N proteins suggests that the two novel strains are more closely related to MERS-related CoV than to other lineage C BetaCoVs (Fig. 2). The predicted RdRp sequences of the two newly detected strains are closely related to MERS-CoV and other MERS-related CoVs, with 94.4 to 97.0% amino acid identities. In contrast, their spike proteins are highly divergent from those of MERS-CoV and other MERS-related CoVs, with 43.3 to 64.9% amino acid identities (Table 4). Phylogenetic analysis suggests that the RdRp genes form an independent cluster closely related to MERS-CoV. In comparison, the S genes form an independent cluster closely related to HKU4 (Fig. 2).

We further focused on the spike protein, which is the most divergent among all CoV proteins and responsible for receptor binding and virus entry. The S proteins of BtCoV/Ii/GD/2013-845 and BtCoV/Ii/GD/2014-422 are 1,349 and 1,356 residues long, respectively. Their ectodomains include residues 1 to 1293 and 1 to 1298, respectively, while their transmembrane anchors include residues 1294 to 1316 and 1299 to 1321, respectively, and their RBDs include residues 371 to 592 and 372 to 593, respectively. Based on our analysis, at least two recombination events have occurred among the S genes of these two viral strains and other lineage C BetaCoVs (Fig. 3A). The recombination event involving the RBD (BtCoV/Ii/GD/2013-845 breakpoints at nucleotide [nt] 22694/23467; BtCoV/Ii/GD/2014-422 breakpoints at nt 22732/23306) includes a bootscan value of <10⁻³⁸, indicating significant evidence for this event (Fig. 3B). Unlike the full-length S protein, their spike RBDs are more similar to the HKU4 RBD than to the MERS-CoV or HKU5 RBDs. Therefore, recombination may have occurred between the ancestor of the two novel MERS-related CoVs and HKU4. Moreover, the bootscan results also showed that the S1 subunits of the two viruses were derived from the ancestor of HKU4, while their S2 subunits were derived from the ancestor of HKU5 (Fig. 3). The complicated recombination history between these lineage C BetaCoVs suggests frequent gene transfers, especially of the S gene, among different CoVs, which may be responsible for the cross-species transmission of these CoVs.

Thirteen residues in the MERS-CoV S-RBD were shown to directly interact with human DPP4 as part of the RBD-human DPP4 complex. The S-RBDs of the novel MERS-related CoVs are identical to that of MERS-CoV at six residues, making them more similar to MERS-CoV than any other known lineage C BetaCoV (Fig. 4A).

To understand the receptor usage of the newly detected novel MERS-related CoVs, we selected the BtCoV/Ii/GD/2014-422 spike protein for detailed functional studies, as its spike RBD is slightly more identical to MERS-CoV than BtCoV/Ii/GD/2013-845 (64.9% versus 64.7%). Using four alternative approaches, we investigated the binding interactions between the BtCoV/Ii/GD/2014-422 spike protein and DPP4. First, an AlphaScreen protein-protein binding assay showed that the BtCoV/Ii/GD/2014-422 RBD binds both human and bat DPP4 (Fig. 4B). Both MERS-CoV and HKU4 RBDs, used as controls, bound to human and bat DPP4, but the HKU5 RBD did not. Second, dot blot hybridization assays confirmed the above-described results (Fig. 4C). Third, we established a BtCoV/Ii/GD/2014-422 spike-mediated pseudovirus entry assay. To this end, retroviruses pseudotyped with BtCoV/Ii/GD/2014-422 spike were incubated with bat cells exogenously expressing or not expressing DPP4. The results showed that bat cells exogenously expressing human or bat DPP4 support entry of BtCoV/Ii/GD/2014-422 pseudoviruses, while bat cells not expressing DPP4 do not (Fig. 5A and B). Lastly, the entry of BtCoV/Ii/GD/2014-422 pseudoviruses into hDPP4-expressing bat cells was inhibited by anti-DPP4 antibodies (Fig. 5B). Taken together, these results reveal that the BtCoV/Ii/GD/2014-422 spike protein recognizes the human and bat DPP4 receptors via the RBD.

We used BtCoV/Ii/GD/2014-422 spike-mediated pseudovirus entry assay to test the cross-neutralization reactivity of MERS-CoV antibodies. Neither antiserum nor monoclonal antibody could neutralize the entry of BtCoV/Ii/GD/2014-422 pseudoviruses (Fig. 6), suggesting no cross-neutralization reaction between BtCoV/Ii/GD/2014-422 and MERS-CoV.

All sampling procedures were performed by veterinarians with approval from the Animal Ethics Committee of the Wuhan Institute of Virology (WIVH5210201). The study was conducted in accordance with the Guide for the Care and Use of Wild Mammals in Research of the People's Republic of China.

The animal surveillance program was conducted from the summer of 2012 to the autumn of 2015 in three southern provinces (Guangdong, Guangxi, and Sichuan) of China. Alimentary specimens were collected using procedures described previously (18). All specimens were collected using disposable swabs and protective gloves that were changed between samples for preventing cross contamination and were immediately placed in viral transport medium. All samples were kept in liquid nitrogen or dry ice before transportation to the laboratory for RNA extraction.

Viral RNA was extracted from the alimentary specimens using a High Pure viral RNA kit (Roche Diagnostics GmbH, Mannheim, Germany). RNA was extracted from 200 μl of sample, used as the template for RT-PCR, eluted in 50 μl of RNase-free water provided by the viral RNA extraction kit, and stored at −80°C.

Two pairs of previously described conserved nested primers (1st round, CoV-FWD3, 5′-GGTTGGGAYTAYCCHAARTGTGA-3′; CoV-RVS3, 5′-CCATCATCASWYRAATCATCATA-3′; 2nd round, CoV-FWD4/Bat, 5′-GAYTAYCCHAARTGTGAYAGAGC-3′; CoV-RVS3, 5′-CCATCATCASWYRAATCATCATA-3′) for amplifying a 440-bp fragment of the CoV RdRp gene were used for initial CoV screening (19). The first round of screening PCR was performed with a Superscript III one-step RT-PCR with Platinum Taq kit (Invitrogen, San Diego, CA). The second round of screening PCR was performed with a Platinum Taq DNA polymerase kit (Invitrogen, San Diego, CA).

Host species were identified first morphologically and then by amplifying and sequencing the NADH dehydrogenase subunit 1 (ND1) gene and/or the cytochrome b (CytB) gene (20, 21). The sequence data, excluding primer sequences, were assembled, and BLASTN (default parameters) was used to identify host species based on the most closely related sequences with the highest query coverage and a minimum identity of 95%.

Two novel MERS-related CoV strains, named BtCoV/Ii/GD/2013-845 and BtCoV/Ii/GD/2014-422, were selected for amplification of their complete genomes through gene walking with primers based on the sequences of known MERS-related CoVs as previously described (primers provided upon request) (22). The 5′ and 3′ ends of the viral genomes were confirmed by rapid amplification of cDNA ends (RACE) using a 5′/3′ RACE kit (Clontech, Heidelberg, Germany). Sequences were assembled and manually edited to produce final sequences of the viral genome with Lasergene SeqMan (v7; DNASTAR, Madison, WI).

Putative open reading frames (ORFs) were predicted using NCBI's ORF Finder (https://www.ncbi.nlm.nih.gov/orffinder/), with a minimal ORF length of 150 nt, and then inspected manually. The nucleotide sequences of the genomes and the deduced amino acid sequences of the ORFs were preliminarily compared to those of other CoV full-length genome sequences available from EMBOSS Needle (http://www.ebi.ac.uk). Sequencing alignment were performed in MEGA (v6.0) with the ClustalW method (23, 24). The similarity plot (window size, 400 bp; step size, 40 bp) was generated with Simplot (v3.5.7) (25). Pairwise comparison of the complete genome and putative proteins was calculated using MEGA 6.0 (24) with a previously described method (26). The TRSs and nonstructural protein (nsp1 to nsp16) prediction was done according to described methods (27). Phylogenetic trees were constructed using the maximum likelihood method set as previously described (5), with 100 bootstrap replications in MEGA (v6.0) (24). Protein family analysis was performed using PFAM (28) and InterProScan (29). The genome recombination bootscan plot (window size, 1,500 bp; step size, 300 bp) was first scanned using the Recombination Detection Program (RDP; v4.94 Beta) (30) and then generated using Simplot (v3.5.7) (25).

The corrected Nei-Gojobori method (Jukes-Cantor) in the KaKs Calculator (v2.0) (31) and the Ka/Ks calculation tool were used to calculate the number of synonymous substitutions per synonymous site (K_s) and the number of nonsynonymous substitutions per nonsynonymous site (K_a) for each coding region between each pair of strains.

The HEK293T (human embryonic kidney) and Tb1-Lu (Tadarida brasiliensis bat lung) cell lines were obtained from the ATCC (www.atcc.org). These cell lines were maintained in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum (FBS), 2 mM l-glutamine, and 1% penicillin-streptomycin (Life Technologies Inc., Grand Island, NY).

The RBDs of the spike proteins of MERS-CoV (residues 367 to 588; GenBank accession no. AFS88936), HKU4 (residues 372 to 593; GenBank accession no. ABN10839), BtCoV/Ii/GD/2014-422 (residues 357 to 592), and HKU5 (residues 375 to 586; GenBank accession no. ABN10875) were expressed and purified as previously described (32, 33). Briefly, CoV RBDs containing an N-terminal honey bee melittin signal peptide and a C-terminal human IgG4 Fc tag were expressed in insect cells using the Bac-to-Bac expression system (Life Technologies Inc.), secreted into cell culture medium, and purified sequentially on HiTrap protein G HP columns and Superdex 200 gel filtration columns (GE Healthcare, Pittsburgh, PA).

The ectodomains of human DPP4 (residues 39 to 766; GenBank accession no. NP_001926) and Pipistrellus bat DPP4 (residues 36 to 760; GenBank accession no. KC249974), containing an N-terminal human CD5 signal peptide and an N-terminal His tag, were expressed and purified using the same procedure as that for CoV RBDs, except they were purified sequentially on HiTrap nickel-chelating HP columns and Superdex 200 gel filtration columns (GE Healthcare) (17).

Interactions between CoV RBDs and human or bat DPP4 were measured using AlphaScreen, as previously described (34, 35). Briefly, each of the Fc-tagged CoV RBDs at a final concentration of 3 nM was mixed with 100 nM His₈-tagged hDPP4 or bDPP4 in half of an AreaPlate (PerkinElmer, Waltham, MA) for 1 h at room temperature as recommended by the manufacturer. AlphaScreen nickel chelate donor beads and AlphaScreen protein A acceptor beads (PerkinElmer) were added to the mixture at a final concentration of 5 μg/ml. The mixtures were incubated at room temperature for 1 h and protected from light. The assay plates were read using an EnSpire plate reader (PerkinElmer).

Dot blot hybridization assays were performed as previously described (36). Briefly, 2 μl of each 50 μg/ml Fc-tagged CoV RBD was dotted onto a nitrocellulose membrane. The membranes were dried completely and blocked with 5% skim milk at 37°C for 1 h. A sample of 20 μg/ml His₈-tagged hDPP4 or bDPP4 was preincubated at 37°C for 1 h, added to the membrane, and incubated at 4°C overnight. The membrane was washed five times with phosphate-buffered saline with 0.1% Tween 20 (PBST), incubated with anti-His₆ mouse monoclonal IgG1 horseradish peroxidase-conjugated antibody (Santa Cruz Biotechnology, Santa Cruz, CA) at 37°C for 2 h, and washed five times with PBST. Finally, the bound proteins were detected using ECL plus (GE Healthcare).

Retroviruses pseudotyped with MERS-CoV spike, BtCoV/Ii/GD/2014-422 spike, or no spike (mock) were used to infect Tb1Lu cells trans-expressing hDPP4, bDPP4, or no DPP4 in 96-well plates. Measurements of infection and luciferase activity were conducted as described previously (17).

Tb1Lu cells expressing hDPP4, bDPP4, or no DPP4 were preincubated with 0 or 10 μg/ml goat anti-hDPP4 polyclonal antibody (R&D Systems) at 37°C for 1 h and then infected by equal amounts of MERS-CoV or BtCoV/Ii/GD/2014-422 spike-pseudotyped retroviruses. Measurements of infection and luciferase activity were conducted as described previously (17).

The mouse serum and monoclonal antibody against MERS-CoV-spike were kindly provided by Lanying Du at New York Blood Center. MERS-CoV or BtCoV/Ii/GD/2014-422 spike-pseudotyped retrovirus was preincubated with mouse serum or monoclonal antibody against MERS-CoV-spike, as previously described (37), for 10 min at 37°C. The mixture was used to infect Tb1Lu cells expressing hDPP4, bDPP4, or no DPP4. Measurements of infection and luciferase activity were conducted as described previously (17).

The nucleotide sequences of the complete genomes of the lineage C BetaCoVs obtained in this study have been submitted to GenBank under accession numbers MG021451 and MG021452. The GenBank accession numbers of the reference complete genomes are the following: MERS-CoV, JX869059, KC164505, and KJ713298; bat MERS-related CoV, KJ473821, KC869678, and NC_034440; HKU4-CoV, NC_009019, EF065506–EF065508, and KJ473822; HKU5-CoV, NC_009020, EF065510–EF065512, and KJ473820; hedgehog coronavirus, KC545383 and KC545386; HCoV-OC43, AY391777; murine hepatitis virus, AY700211; severe acute respiratory syndrome virus, NC_004718.