Novel Bat Alphacoronaviruses in Southern China Support Chinese Horseshoe Bats as an Important Reservoir for Potential Novel Coronaviruses

While bats are increasingly recognized as a source of coronavirus epidemics, the diversity and emergence potential of bat coronaviruses remains to be fully understood. Among 1779 bat samples collected in China, diverse coronaviruses were detected in 32 samples from five different bat species by RT-PCR. Two novel alphacoronaviruses, Rhinolophus sinicus bat coronavirus HKU32 (Rs-BatCoV HKU32) and Tylonycteris robustula bat coronavirus HKU33 (Tr-BatCoV HKU33), were discovered from Chinese horseshoe bats in Hong Kong and greater bamboo bats in Guizhou Province, respectively. Genome analyses showed that Rs-BatCoV HKU32 is closely related to BatCoV HKU10 and related viruses from diverse bat families, whereas Tr-BatCoV HKU33 is closely related to BtNv-AlphaCoV and similar viruses exclusively from bats of Vespertilionidae family. The close relatedness of Rs-BatCoV HKU32 to BatCoV HKU10 which was also detected in Pomona roundleaf bats from the same country park suggests that these viruses may have the tendency of infecting genetically distant bat populations of close geographical proximity with subsequent genetic divergence. Moreover, the presence of SARSr-CoV ORF7a-like protein in Rs-BatCoV HKU32 suggests a common evolutionary origin of this accessory protein with SARS-CoV, also from Chinese horseshoe bats, an apparent reservoir for coronavirus epidemics. The emergence potential of Rs-BatCoV HKU32 should be explored.


Introduction
The Severe Acute Respiratory Syndrome (SARS) and more recently the Middle East Respiratory Syndrome (MERS) have proven the emergence potential of animal coronaviruses (CoVs) and aroused immense interest in the discovery of novel CoVs in animals and humans. While SARS coronavirus (SARS-CoV) was originated from horseshoe bats in China as its animal reservoir and transmitted to humans after amplification in palm civets from wildlife markets [1,2], dromedary camels in the Middle East are the immediate animal source of the MERS epidemic caused by MERS coronavirus (MERS-CoV) [3][4][5][6]. Bats also harbor MERS-CoV-related viruses, which may suggest a possible bat origin [7][8][9][10][11][12][13][14][15], although the evolutionary origin of MERS-CoV remains to be ascertained.
Besides SARS-CoV and MERS-CoV, bat CoVs closely related to other human CoVs, including human CoV 229E and human CoV NL63, were also recently discovered [26][27][28], suggesting that bats are the important animal source of CoVs that may emerge in humans. On the other hand, bat CoVs may also evolve to infect other animals. For example, porcine epidemic diarrhea virus (PEDV) is phylogenetically closely related to Scotophilus bat coronavirus 512 (Sc-BatCoV 512), suggesting cross-species transmission events between bats and pigs [29]. In 2016-2017, outbreaks of severe watery diarrhea were reported in suckling piglets from farms in Guangdong Province, China, which were found to be caused by swine acute diarrhea syndrome coronavirus (SADS-CoV) [30][31][32]. SADS-CoV is very close to and likely to have emerged from Rhinolophus bat CoV HKU2 (Rh-BatCoV HKU2), first discovered in Hong Kong and detected in a wide range of horseshoe bats including Rhinolophus sinicus, Rhinolophus affinis and Rhinolophus ferrumequinum [30,33]. In particular, the spike protein of SADS-CoV shared 93-98% amino acid identity to that of Rh-BatCoV HKU2 from Rhinolophus affinis, supporting recent interspecies jumping from bats to pigs [30].
To further explore the diversity of CoVs in bats and understand the genetic evolution of CoVs, we collected bat samples from Hong Kong and mainland China. Diverse CoVs belonging to alphaCoVs and betaCoVs were detected, including two novel alphaCoVs, as confirmed by complete genome sequencing and characterization, supporting bats as an important reservoir for CoVs. The evolutionary relationship of the two novel alphaCoVs to other known CoVs is also discussed.

Collection of Bat Samples
Oral and alimentary samples were collected from bats captured from HKSAR, and Guizhou and Guangdong Provinces, mainland China, during 2013−2015 using procedures described previously [1]. To prevent cross contamination, sterile disposable swabs with protective gloves were used during sample collection and changed between samples. All samples were immediately placed in viral transport medium (Earle's balanced salt solution, 20% glucose, 4.4% NaHCO 3 , 5% bovine albumin, vancomycin 50,000 µg/mL, amikacin 50,000 µg/mL, nystatin 10,000 units/mL) before transportation to the laboratory. For prolonged storage, all samples were stored at −80 • C before further studies.

Detection of Bat CoVs by RNA Extraction, RT-PCR and DNA Sequencing
Viral RNA was extracted form oral and alimentary samples using QIAamp viral RNA minikit (Qiagen, Hilden, Germany).
Eluted RNA was used as the template for reverse transcription-PCR (RT-PCR). Detection of CoV was performed by amplifying a 440-bp fragment of the RNA-dependent RNA polymerase (RdRp) gene of CoVs using conserved primers (5'-GGTTGGGACTATCCTAAGTGTGA-3' and 5'-ACCATCATCNGANARDATCATNA-3') [18]. Reverse transcription was performed using a SuperScript III kit (Invitrogen, San Diego, CA, USA). PCR mixture (25 µL) was prepared and PCR conditions were set as described previously in an automated thermal cycler (Applied Biosystems) [34]. Amplified PCR products were gel-purified using the QIAquick gel extraction kit (QIAgen). Both strands of PCR products were sequenced with an ABI Prism 3130× genetic Analyzer (Applied Biosystems, Foster City, CA, USA), using the above primers. Comparison of the PCR products' sequences with other known CoVs' RdRp genes from GenBank sequence database was performed.

Viral Culture
Attempts to isolate Rs-BatCoV HKU32 and Tr-BatCoV HKU33 were performed by inoculating samples with RT-PCR positive results to different cells. Viral replication was detected by cytopathic effect observation and viral detection of culture supernatant collected from passages by RT-PCR.

Complete Genome Sequencing of Rs-BatCoV HKU32 and Tr-BatCoV HKU33
Viral genomes of two Rs-BatCoV HKU32 (TLC26A and TLC28A) and one Tr-BatCoV HKU33 (GZ151867) were amplified and sequenced using RNA directly extracted from their alimentary samples respectively as templates. Both viral RNAs were reverse transcribed to cDNA by a combined random-priming and oligo(dT)-priming strategy. For GZ151867, the amplified cDNA sample was barcoded and sequenced using the Ion Torrent sequencing platform. The average sequencing throughput of these samples was 12.90 Mbp and the average read length was 150.1 bp. The single-end reads were de novo assembled using SPAdes Genome Assembler version 3.10.0 using default parameters [35]. Coronavirus-matching contigs were searched using BLASTN version 2.5.0+ against NCBI nucleotide database (nt) version downloaded on October 4th 2016 with e-value cutoff at 1 × 10 −5 [36].
Degenerated primers were designed according to multiple alignments of the genomes of other alphaCoVs with complete genomes available, using strategies as described previously [18,20,37]. Additional primers were designed based on the results of the first and subsequent rounds of sequencing or the NGS sequencing for the sample GZ151867. SMARTer 5'/3' RACE kit (Clontech, Mountain View, CA, USA) was used to perform rapid amplification of cDNA ends and confirm the 5' ends of the genomes. For Rs-BatCoV HKU32, a total of 53 sets of primers (available on request) were used for PCR. For Tr-BatCoV HKU33, a total of 44 sets of primers (available on request) were used for PCR. Sequences were assembled and edited manually to produce complete sequences of the three viral genomes.

Phylogenetic and Genome Analysis of Rs-BatCoV HKU32 and Tr-BatCoV HKU33
The genomes of Rs-BatCoV HKU32 and Tr-BatCoV HKU33 were aligned and analyzed with other alphaCoVs with complete genome sequences available from Genbank using online sequence alignment server MAFFT version 7 [38]. The nucleotide sequences of the genomes and the deduced amino acid sequences of the open reading frames were analyzed and compared with other alphaCoVs using ORFfinder (https://www.ncbi.nlm.nih.gov/orffinder/). Maximum-likelihood phylogenetic tress with 1000 bootstrap replicates of ORF1ab and S genes were constructed using PhyMLv3.0 (The French Institute of Bioinformatics & France Genomique, Montpellier, France) [39]. Smart Model Selection from PhyML was used to calculate the best-fit substitution model for ML analyses.

Expression of ORF10 Accessory Gene and Determination of Leader-Body Junction Sequence
The leader-body junction site and flanking sequences of the ORF10 subgenomic mRNA in Rs-BatCoV HKU32 strain TLC28A was sequenced and determined using RT-PCR method as described previously. cDNA obtained from RT was used as the template for PCR amplification with a forward primer (5'-GCGTCTCATCCCCTCAA-3') located in the leader sequence and a reverse primer (5'-GAACCAGCGATACAATCAATG-3') located in the body of the ORF10 subgenomic mRNA. PCR mixture was prepared as described previously. The mixtures were amplified for 60 cycles of 94 • C for 1 min, 55 • C for 1 min, and 72 • C for 1 min and a final extension at 72 • C for 10 min. Amplified PCR products were subjected to gel purification and sequencing as described previously.

Bat Coronaviruses Surveillance and Identification of Two Novel Alphacoronaviruses
The other positive bat samples contained known bat alphaCoVs and betaCoVs. Hi-BatCoV HKU10 was detected in 2 samples from Hipposideros pomona captured in Hong Kong, with 99% nucleotide identity to the corresponding partial RdRp sequence of Hi-BatCoV HKU10 isolate TLC1310A (GenBank accession no. JQ989268.1) (

Genome Features of the Two Novel Alphacoronaviruses, Rs-BatCoV HKU32 and Tr-BatCoV HKU33
The complete genomes of two strains of Rs-BatCoV HKU32, TLC26A and TLC28A, and Tr-BatCoV HKU33 strain GZ151867 were sequenced and determined to characterize their genome features.

Novel alphaCoV Species: Rs-BatCoV HKU32
Both genomes of Rs-BatCoV HKU32 TLC26A and TLC28A possessed genome sizes of 29201 nucleotides, with 40.3% G + C content. They shared 99% overall nucleotide identity to each another. Rs-BatCoV HKU32 strain TLC28A was selected as the reference strain for the following genomic analyses.
Similar to other alphaCoV genomes, Rs-BatCoV HKU32 consisted of 10 putative open reading frames (ORFs) including the essential ORF1ab, S, E, M and N ( Table 2). 3 accessory genes were located between the S and N genes while 2 accessory genes were found downstream of N gene ( Figure 2). Rs-BatCoV HKU32 ORF3 accessory protein shared low amino acid identity (31%-51%) to the respective accessory proteins in other alphaCoVs while Rs-BatCoV HKU32 ORF10 accessory protein shared 29% amino acid identity to SARSr-CoV ORF7a accessory protein. A putative transcription regulatory sequence (TRS) motif, 5'-CUAAAC-3', was identified at the 3' end of the leader sequence and preceded most ORFs except the S, ORF3, ORF5a and ORF9 (Table 2). An alternative TRS motif for S and ORF5 genes was found to be 5'-CUAAAU-3', while that for ORF3 and ORF9 was 5'-CUAAAU-3' and 5'-CUGAAC-3', respectively ( Table 2). The characteristics of putative nonstructural protein and predicted putative cleavage sites of Rs-BatCoV HKU32 are shown in Tables 3 and 4.
Comparative genomic analyses showed that Rs-BatCoV HKU32 shared 58.9% overall nucleotide identity with BtRf-AlphaCoV and 67.2% with Hi-BatCoV HKU10. To determine whether Rs-BatCoV HKU32 was a novel alphaCoV species, 7 conserved replicase domains of Rs-BatCoV HKU32 were selected for analyses according to the CoV species demarcation criteria by the ICTV [23]. Five known alphaCoVs with complete genome sequences available and close phylogenetic relationship to Rs-BatCoV HKU32 were chosen for comparison. The 7 concatenated domains of Rs-BatCoV HKU32 shared 83.2%, 83.3%, 78.6% 70.3 and 69.5% amino acid identity with those of BtRf-AlphaCoV, BtMs-AlphaCoV, Ro-BatCoV HKU10, PEDV and Tr-BatCoV HKU33, respectively, which were below the threshold of 90% amino acid identity ( Table 5). The results supported that Rs-BatCoV HKU32 represents a novel CoV species in the AlphaCoV genus.

Genome Features of the Two Novel Alphacoronaviruses, Rs-BatCoV HKU32 and Tr-BatCoV HKU33
The complete genomes of two strains of Rs-BatCoV HKU32, TLC26A and TLC28A, and Tr-BatCoV HKU33 strain GZ151867 were sequenced and determined to characterize their genome features.

Novel alphaCoV Species: Rs-BatCoV HKU32
Both genomes of Rs-BatCoV HKU32 TLC26A and TLC28A possessed genome sizes of 29201 nucleotides, with 40.3% G + C content. They shared 99% overall nucleotide identity to each another. Rs-BatCoV HKU32 strain TLC28A was selected as the reference strain for the following genomic analyses.
Similar to other alphaCoV genomes, Rs-BatCoV HKU32 consisted of 10 putative open reading frames (ORFs) including the essential ORF1ab, S, E, M and N ( Table 2). 3 accessory genes were located between the S and N genes while 2 accessory genes were found downstream of N gene ( Figure 2). Rs-BatCoV HKU32 ORF3 accessory protein shared low amino acid identity (31%-51%) to the respective accessory proteins in other alphaCoVs while Rs-BatCoV HKU32 ORF10 accessory protein shared 29% amino acid identity to SARSr-CoV ORF7a accessory protein. A putative transcription regulatory sequence (TRS) motif, 5'-CUAAAC-3', was identified at the 3' end of the leader sequence and preceded most ORFs except the S, ORF3, ORF5a and ORF9 (Table 2). An alternative TRS motif for S and ORF5 genes was found to be 5'-CUAAAU-3', while that for ORF3 and ORF9 was 5'-CUAAAU-3' and 5'-CUGAAC-3', respectively ( Table 2). The characteristics of putative nonstructural protein and predicted putative cleavage sites of Rs-BatCoV HKU32 are shown in Tables 3 and 4.
Comparative genomic analyses showed that Rs-BatCoV HKU32 shared 58.9% overall nucleotide identity with BtRf-AlphaCoV and 67.2% with Hi-BatCoV HKU10. To determine whether Rs-BatCoV HKU32 was a novel alphaCoV species, 7 conserved replicase domains of Rs-BatCoV HKU32 were selected for analyses according to the CoV species demarcation criteria by the ICTV [23]. Five known alphaCoVs with complete genome sequences available and close phylogenetic relationship to Rs-BatCoV HKU32 were chosen for comparison. The 7 concatenated domains of Rs-BatCoV HKU32 shared 83.2%, 83.3%, 78.6% 70.3 and 69.5% amino acid identity with those of BtRf-AlphaCoV, BtMs-AlphaCoV, Ro-BatCoV HKU10, PEDV and Tr-BatCoV HKU33, respectively, which were below the threshold of 90% amino acid identity ( Table 5). The results supported that Rs-BatCoV HKU32 represents a novel CoV species in the AlphaCoV genus.

Homologous SARSr-CoV ORF7a-Like Accessory Protein in Rs-BatCoV HKU32
A homologous SARSr-CoV ORF7a-like accessory protein, ORF10, was found in Rs-BatCoV HKU32, located at nucleotide position 28593 to 28955 with 120 amino acids. The ORF7a accessory gene (also known as X4) found in SARSr-CoVs from both human and animal sources is a type I transmembrane protein [40]. InterProScan analysis showed that the Rs-BatCoV HKU32 ORF10 protein also possessed four domains, an N-terminal signal peptide, followed by a luminal domain, transmembrane segment and cytoplasmic tail. Rs-BatCoV HKU32 ORF10 protein possessed 29% amino acid identity to SARSr-CoV ORF7a protein.
While the function of SARSr-CoV ORF7a protein remained to be elucidated, studies have shown that this accessory protein was expressed during SARS-CoV replication. ORF7a protein was shown to exit the endoplasmic reticulum through COPII transport machinery and is targeted to the Golgi apparatus [41]. To determine if ORF10 accessory gene is expressed in Rs-BatCoV HKU32, the leaderbody junction sites and flanking sequences of ORF10 subgenomic mRNA were determined ( Figure  5). The subgenomic mRNAs was successfully amplified and sequenced directly from both samples TLC26A and TLC28A. The sequences were aligned to the leader sequence, which confirmed 5'-CUAAAC-3' as the core sequence of the TRS motif. The leader TRS and subgenomic mRNA of ORF10 accessory gene exactly matched each other.

Homologous SARSr-CoV ORF7a-Like Accessory Protein in Rs-BatCoV HKU32
A homologous SARSr-CoV ORF7a-like accessory protein, ORF10, was found in Rs-BatCoV HKU32, located at nucleotide position 28593 to 28955 with 120 amino acids. The ORF7a accessory gene (also known as X4) found in SARSr-CoVs from both human and animal sources is a type I transmembrane protein [40]. InterProScan analysis showed that the Rs-BatCoV HKU32 ORF10 protein also possessed four domains, an N-terminal signal peptide, followed by a luminal domain, transmembrane segment and cytoplasmic tail. Rs-BatCoV HKU32 ORF10 protein possessed 29% amino acid identity to SARSr-CoV ORF7a protein.
While the function of SARSr-CoV ORF7a protein remained to be elucidated, studies have shown that this accessory protein was expressed during SARS-CoV replication. ORF7a protein was shown to exit the endoplasmic reticulum through COPII transport machinery and is targeted to the Golgi apparatus [41]. To determine if ORF10 accessory gene is expressed in Rs-BatCoV HKU32, the leader-body junction sites and flanking sequences of ORF10 subgenomic mRNA were determined ( Figure 5). The subgenomic mRNAs was successfully amplified and sequenced directly from both samples TLC26A and TLC28A. The sequences were aligned to the leader sequence, which confirmed 5'-CUAAAC-3' as the core sequence of the TRS motif. The leader TRS and subgenomic mRNA of ORF10 accessory gene exactly matched each other.

Discussion
In this study, two novel alphaCoVs, Rs-BatCoV HKU32 from Rhinolophus sinicus in Hong Kong and Tr-BatCoV HKU33 from Tylonycteris robustula in Guizhou Province, were discovered. Their classification as novel species within the genus AlphaCoV is supported by the results from pairwise comparison of their 7 concatenated domains to those of known alphaCoVs based on the CoV species demarcation criteria by the ICTV. Phylogenetically, Rs-BatCoV HKU32 is most closely related to BtRf-AlphaCoV, BtMs-AlphaCoV, Ro-BatCoV HKU10 and Hi-BatCoV HKU10. On the other hand, Tr-BatCoV HKU33 is most closely related to BtNv-AlphaCoV, AlphaCoV BatCoV/P.kuhlii/Italy/206645-41/2011 and AlphaCoV BatCoV/P.kuhlii/Italy/206679-3/2010. Nevertheless, different phylogenetic positioning with closely related alphaCoVs was observed between ORF1ab and the S gene, suggesting that the S gene may have evolved in a path different from other genome regions. The S protein of CoVs is responsible for receptor recognition and contains epitopes for neutralizing antibodies, and hence is often subjected to selective pressure. Therefore, its evolutionary path may be different from that of other genome regions. For example, we have previously found that another bat alphaCoV, Rs-BatCoV HKU2, contains an evolutionarily distinct spike protein which is only distantly related to other alphaCoVs [33].
The close phylogenetic relationship between Rs-BatCoV HKU32 and BatCoV HKU10 may be explained by the geographical proximity of their hosts. Besides Rs-BatCoV HKU32 and Tr-BatCoV HKU33, diverse alphaCoVs and betaCoVs were also detected in our bat samples. Hi-BatCoV HKU10 and Coronavirus PREDICT CoV-37, belonging to alphaCoVs, were detected in bats in Hong Kong, while SARSr-BatCoVs and Ty-BatCoV HKU4, belonging to lineage B (Sarbecovirus) and C (Merbecovirus) betaCoVs, respectively, were detected in bats in Guangdong and Guizhou provinces. It is of note that Rs-BatCoV HKU32 and BatCoV HKU10 were detected in bats of different families but captured from the same sampling location at a country park in Hong Kong. This suggests that these viruses may have evolved among geographically close but phylogenetically distant bat populations. We have also previously described interspecies transmission of BatCoV HKU10 between Rousettus leschenaultii and Hipposideros pomona bats which belonged to two different bat families [42]. Given that the genetic cluster formed by Rs-BatCoV HKU32, BatCoV HKU10, BtRs-AlphaCoV and BtMs-AlphaCoV were detected from diverse bat families, this suggests that these alphaCoVs may have a tendency for cross-species transmission. This is in contrast to Tr-BatCoV HKU33 and related viruses which were limited to the family Vespertilionidae. Greater bamboo bats, the host of Tr-BatCoV HKU33, and other Vesper bats are geographically widespread and able to occupy distinct and diverse habitats [43]. Interestingly, MERS-CoV-related bat lineage C betaCoVs (Merbecovirus) have also been exclusively detected in the family Vespertilionidae [10,11,[13][14][15]44,45]. Further studies are needed to understand the mechanism for host tropism, especially the stringency of host receptor usage by different alphaCoVs, the interspecies transmissibility of coronaviruses among different bats and the contribution of bat ecology to this phenomenon.
The presence of a homologous SARSr-CoV ORF7a-like protein in Rs-BatCoV HKU32 suggests a common evolutionary origin of this accessory protein from viruses in Chinese horseshoe bats Figure 5. Rs-BatCoV HKU32 strain TLC28A mRNA leader-body junction and flanking sequences. The subgenomic ORF10 mRNA sequence are shown in alignment with the leader and genomic sequences. Identical nucleotides between the leader sequence and subgenomic mRNA sequence are labeled in green. Identical nucleotides between genome and subgenomic mRNA sequence are labeled in blue. The putative TRS is labeled in bold type, in purple and is underlined. Start codon AUG is labeled in red.

Discussion
In this study, two novel alphaCoVs, Rs-BatCoV HKU32 from Rhinolophus sinicus in Hong Kong and Tr-BatCoV HKU33 from Tylonycteris robustula in Guizhou Province, were discovered. Their classification as novel species within the genus AlphaCoV is supported by the results from pairwise comparison of their 7 concatenated domains to those of known alphaCoVs based on the CoV species demarcation criteria by the ICTV. Phylogenetically, Rs-BatCoV HKU32 is most closely related to BtRf-AlphaCoV, BtMs-AlphaCoV, Ro-BatCoV HKU10 and Hi-BatCoV HKU10. On the other hand, Tr-BatCoV HKU33 is most closely related to BtNv-AlphaCoV, AlphaCoV BatCoV/P.kuhlii/Italy/206645-41/2011 and AlphaCoV BatCoV/P.kuhlii/Italy/206679-3/2010. Nevertheless, different phylogenetic positioning with closely related alphaCoVs was observed between ORF1ab and the S gene, suggesting that the S gene may have evolved in a path different from other genome regions. The S protein of CoVs is responsible for receptor recognition and contains epitopes for neutralizing antibodies, and hence is often subjected to selective pressure. Therefore, its evolutionary path may be different from that of other genome regions. For example, we have previously found that another bat alphaCoV, Rs-BatCoV HKU2, contains an evolutionarily distinct spike protein which is only distantly related to other alphaCoVs [33].
The close phylogenetic relationship between Rs-BatCoV HKU32 and BatCoV HKU10 may be explained by the geographical proximity of their hosts. Besides Rs-BatCoV HKU32 and Tr-BatCoV HKU33, diverse alphaCoVs and betaCoVs were also detected in our bat samples. Hi-BatCoV HKU10 and Coronavirus PREDICT CoV-37, belonging to alphaCoVs, were detected in bats in Hong Kong, while SARSr-BatCoVs and Ty-BatCoV HKU4, belonging to lineage B (Sarbecovirus) and C (Merbecovirus) betaCoVs, respectively, were detected in bats in Guangdong and Guizhou provinces. It is of note that Rs-BatCoV HKU32 and BatCoV HKU10 were detected in bats of different families but captured from the same sampling location at a country park in Hong Kong. This suggests that these viruses may have evolved among geographically close but phylogenetically distant bat populations. We have also previously described interspecies transmission of BatCoV HKU10 between Rousettus leschenaultii and Hipposideros pomona bats which belonged to two different bat families [42]. Given that the genetic cluster formed by Rs-BatCoV HKU32, BatCoV HKU10, BtRs-AlphaCoV and BtMs-AlphaCoV were detected from diverse bat families, this suggests that these alphaCoVs may have a tendency for cross-species transmission. This is in contrast to Tr-BatCoV HKU33 and related viruses which were limited to the family Vespertilionidae. Greater bamboo bats, the host of Tr-BatCoV HKU33, and other Vesper bats are geographically widespread and able to occupy distinct and diverse habitats [43]. Interestingly, MERS-CoV-related bat lineage C betaCoVs (Merbecovirus) have also been exclusively detected in the family Vespertilionidae [10,11,[13][14][15]44,45]. Further studies are needed to understand the mechanism for host tropism, especially the stringency of host receptor usage by different alphaCoVs, the interspecies transmissibility of coronaviruses among different bats and the contribution of bat ecology to this phenomenon.
The presence of a homologous SARSr-CoV ORF7a-like protein in Rs-BatCoV HKU32 suggests a common evolutionary origin of this accessory protein from viruses in Chinese horseshoe bats (Rhinolophus sinicus). While they share the same host specificity for Chinese horseshoe bats, Rs-BatCoV HKU32 is an alphaCoV while SARSr-CoV is a lineage B betaCoV (Sarbecovirus). While the two proteins in SARSr-CoVs and Rs-BatCoV HKU32 share only 29% amino acid identity, they both possess typical conserved domains of a type I transmembrane protein. Moreover, this ORF10 (SARSr-CoV ORF7a-like) gene is shown to be expressed in Rs-BatCoV HKU32, suggesting it may play a role in viral replication. Besides SARSr-CoVs and Rs-BatCoV HKU32, another alphaCoV, Rs-BatCoV HKU2 has also been detected in this horseshoe bat species which is the origin of the SARS epidemic. Interestingly, Rs-BatCoV HKU2 was also recently found to have evolved and emerged in swine population causing epidemics in China, supporting Chinese horseshoe bats as an important animal source for CoV epidemics in both humans and animals [30][31][32]. This horseshoe bat species, which ranges from northern India to southern China, often resides in caves or any man-made cave-like structures such as abandoned tunnels in Hong Kong [43]. It is therefore important to preserve their natural habitats and avoid contact with these bats. Besides SARS-CoV and MERS-CoV, HCoV 229E, an alphaCoV, is also likely originated from bats approximately 219 to 333 years ago (Figure 4) [28], suggesting that alphaCoVs, like betaCoVs, are also the potential source for new epidemics in humans. Further studies are required to elucidate the function of SARSr-CoV ORF7a protein and the emergence potential of Rs-BatCoV HKU32.