Global Epidemiology of Bat Coronaviruses

Bats are a unique group of mammals of the order Chiroptera. They are highly diversified and are the group of mammals with the second largest number of species. Such highly diversified cell types and receptors facilitate them to be potential hosts of a large variety of viruses. Bats are the only group of mammals capable of sustained flight, which enables them to disseminate the viruses they harbor and enhance the chance of interspecies transmission. This article aims at reviewing the various aspects of the global epidemiology of bat coronaviruses (CoVs). Before the SARS epidemic, bats were not known to be hosts for CoVs. In the last 15 years, bats have been found to be hosts of >30 CoVs with complete genomes sequenced, and many more if those without genome sequences are included. Among the four CoV genera, only alphaCoVs and betaCoVs have been found in bats. As a whole, both alphaCoVs and betaCoVs have been detected from bats in Asia, Europe, Africa, North and South America and Australasia; but alphaCoVs seem to be more widespread than betaCoVs, and their detection rate is also higher. For betaCoVs, only those from subgenera Sarbecovirus, Merbecovirus, Nobecovirus and Hibecovirus have been detected in bats. Most notably, horseshoe bats are the reservoir of SARS-CoV, and several betaCoVs from subgenus Merbecovirus are closely related to MERS-CoV. In addition to the interactions among various bat species themselves, bat–animal and bat–human interactions, such as the presence of live bats in wildlife wet markets and restaurants in Southern China, are important for interspecies transmission of CoVs and may lead to devastating global outbreaks.


Introduction
Coronaviruses (CoVs) infect humans and a wide variety of animals, causing diseases in the respiratory, enteric, hepatic, and neurological systems with varying severity. CoVs are classified into four genera, Alphacoronavirus, Betacoronavirus, Gammacoronavirus and Deltacoronavirus. Within Betacoronavirus, they can be further subclassified into lineages A, B, C and D [1]. In 2018, these four lineages were reclassified as subgenera of Betacoronavirus, and renamed as Embecovirus (previous lineage A), Sarbecovirus (previous lineage B), Merbecovirus (previous lineage C) and Nobecovirus (previous lineage D) [2]. In addition, a fifth subgenus, Hibecovirus, was also included (Figure 1) [2]. As a result, of the unique mechanism of viral replication, CoVs have a high frequency of recombination [3][4][5][6][7][8]. CoVs may achieve rapid adaptation to new host and ecological niches, as a result of their tendency for recombination and the inherently high mutation rates, which are typical of RNA viruses [9].
Viruses 2019, 11 FOR PEER REVIEW result of their tendency for recombination and the inherently high mutation rates, which are typical of RNA viruses [9]. Bats are a unique group of mammals of the order Chiroptera. Traditionally, bats have been classified into two suborders: the megabats, which are mostly frugivorous, and the microbats, which are mostly insectivorous. However, recent molecular studies have supported a revision of the classification into the suborders Yinpterochiroptera and Yangochiroptera, of which Yinpterochiroptera consists of the megabats and several microbat species [10]. Bats are globally distributed, although individual bat species have their own geographical niche [11]. Bats are highly diversified and are the group of mammals with the second largest number of species [12]. Such high diversification of bat species provides various cell types and receptors which facilitate them being potential hosts to a large  Bats are a unique group of mammals of the order Chiroptera. Traditionally, bats have been classified into two suborders: the megabats, which are mostly frugivorous, and the microbats, which are mostly insectivorous. However, recent molecular studies have supported a revision of the classification into the suborders Yinpterochiroptera and Yangochiroptera, of which Yinpterochiroptera consists of the megabats and several microbat species [10]. Bats are globally distributed, although individual bat species have their own geographical niche [11]. Bats are highly diversified and are the group of mammals with the second largest number of species [12]. Such high diversification of bat species provides various cell types and receptors which facilitate them being potential hosts to a large variety of viruses. Bats are the only group of mammals capable of sustained flight, which enables them to disseminate the viruses they harbor and enhance the chance of interspecies transmission. Traditionally, bats are recognized to be hosts of several highly pathogenic viruses, such as rabies virus and other lyssaviruses, Hendra virus, Nipah virus and Ebola virus [13][14][15][16][17][18][19][20][21].
Before the Severe Acute Respiratory Syndrome (SARS) epidemic, bats were not known to be hosts for CoVs. After the SARS epidemic, there was a boost in interest regarding searching for novel CoVs in various mammals. In the last 14 years, bats have been found to be hosts of at least 30 CoVs with complete genome sequences available [7,[22][23][24][25][26][27][28][29][30][31][32], and many more if those without genome sequences are included [33][34][35][36][37][38][39][40][41][42]. Most notably, horseshoe bats were found to be the reservoir of SARS-like CoVs, while palm civet cats are considered to be the intermediate host for SARS-CoVs [43][44][45]. In addition, several betaCoVs from subgenus Merbecovirus were also found to be closely related to the Middle East Respiratory Syndrome (MERS)-CoV in dromedary camels and humans [23,29,46,47]. In this article, we review the various aspects of the global epidemiology of bat CoVs. The detailed molecular evolution, phylogenetic analysis and recombination studies, which have been reviewed by others [48][49][50], will not be included in this review.

Bat-Animal and Bat-Human Interactions: Interspecies Jumping of Bat CoVs
In addition to the interactions among the various bat species themselves, interactions between bats and other animals, as well as interactions between bats and human are important for interspecies transmission of viruses. Scientists have proposed several possible activities or events that have led to successful interspecies jumping of CoVs in the last decade. For instance, bats are used as food in Southern China and other countries in Southeast Asia. Live bats are found in wild life wet markets and restaurants in Southern China, which have greatly facilitated bat-human and bat-animal interactions. One of the first cases of SARS occurred in a chef from Heyuan. He worked in a restaurant in Shenzhen and had regular contact with wild food animals, including bats [54]. Shortly afterwards, SARS-CoVs were isolated from caged Himalayan palm civets from wild live markets in Guangdong [55][56][57]. There are many different predators of bats, and what could consume them depends on their locations. Some flying animals, such as owls and hawks, are natural predators of bats. Owls can be active at night when the bats are out. Owls are able to capture bats without any warning while they are in flight [58]. Weasels and raccoons have also been identified as bat predators in some locations and they often lurk around places where the bats reside; and interestingly, SARS-CoV was also detected in a raccoon dog in a Chinese wet market [57,59]. All these interactions between bats and other animals and between bats and human may contribute to CoV interspecies jumping.
Theoretically, in order to achieve successful interspecies jumping, several conditions need to be satisfied. Firstly, a host reservoir of CoVs should be established. Numerous surveillance studies conducted by different groups of scientists have proven that bats serve as a major reservoir of alphaCoVs and betaCoVs [7,20,22,25,42,51,52,[60][61][62][63][64][65]. The wide diversity of bat species and their capability of sustained flight allow them to occupy a broad worldwide habitat.
Secondly, direct or indirect transmission routes between donor and recipient hosts should be identified. However, there are few reports on direct contact between bats and human or other animals, except for consumption by bat predators or human [58]. Yet, the possibility of spillover events should not be overlooked. It is believed that the spillover of bat CoVs mainly occurs by host viral shedding, an indirect route whereby interspecies transmission could be achieved.
Thirdly, when viral spillover events are considered, tissue tropism of CoVs is a major determining factor [66]. Unlike Hendra virus and Nipah virus, both alphaCoVs and betaCoVs have much higher detection rates in intestinal and fecal samples than in throat or urine samples, although the donor host may not manifest apparent pathologies or diseases [7,43,[67][68][69]. Therefore, bat excretions are the major environmental source shedding CoVs in spillover events. The close contact between the recipient host's susceptible tissue and bat excretions or contaminated fomites becomes the essential link for transmission.
Fourthly, survival fitness of CoVs in the environment determines the chance of successful spillover events. CoVs are enveloped, positive-sense single-stranded RNA viruses with surface spike protein projection. A high genomic mutation rate allows CoVs to evolve, and leads to high diversity with potential for host receptor adaptation [4,5,7,70]. However, the presence of viral lipid envelopes renders CoVs sensitive to environmental conditions such as desiccation, heat, extreme pH, UV light and presence of detergents [71]. Prolonged exposure to unfavorable natural environments leads to rapid viral decay and loss of infectivity [71][72][73]. Therefore, the particular location of viral shedding events and relevant environmental condition should be taken into account when predicting the potential of interspecies transmission.
This notwithstanding, available evidence provides insight that recipient hosts which share similar receptor identity to human are potential candidates as intermediate hosts in interspecies jumping events [96][97][98][99]. A common phenomenon observed in both SARS-CoV and SADS-CoV outbreaks is that the outbreaks involved caged or farmed animals which were restricted in defined areas with bats residing around [55,56,63,[100][101][102]. In the case of SARS-CoV, civets in the wild were found to be free from the infection [43]. In other words, human activities facilitated the viral spillover events by bringing susceptible recipient hosts to the vicinity of viral sources. Surveillance on interspecies transmission should be placed around wet markets, farms and abattoirs to safeguard human from novel zoonotic diseases.

Geographical Distribution of Bat CoVs
Generally, both alphaCoVs and betaCoVs have been detected in bats in Asia, Europe, Africa, North and South America and Australasia [7,22,30,37,42,65,68,[103][104][105][106][107][108][109][110]. In general, alphaCoVs seem to be more widespread than betaCoVs (Figure 2), and their detection rate is also higher. In our experience, the prevalence of alphaCoVs was around twice that of betaCoVs for bats in Hong Kong ( Figure 3) (unpublished data). Regional patterns of bat CoV outbreaks at species level can be deduced from the population distribution of their respective bat hosts. There were three major CoV outbreaks in the past 15 years, including SARS-CoV from Sarbecovirus, MERS-CoV from Merbecovirus and SADS-CoV from Rhinacovirus [54,63,100,101,111,112]. Apart from MERS-CoV, whose most recent common ancestor still remains unknown, bats have been confirmed to be the origin of SARS-CoV and SADS-CoV [7,43,100,101,113]. Interestingly, both SARS-CoV and SADS-CoV were discovered in horseshoe bats, mainly Rhinolophus sinicus and Rhinolophus affinis [7,100]. The outbreaks were located in southeast China, where a large diversity of horseshoe bats resides, especially the two species (Rhinolophus sinicus and Rhinolophus affinis), mentioned above ( Figure 4) [7,100]. This suggests that the next CoV outbreak can be geographically predicted by the specific bat species' distribution.
Viruses 2019, 11 FOR PEER REVIEW 5 bats, mainly Rhinolophus sinicus and Rhinolophus affinis [7,100]. The outbreaks were located in southeast China, where a large diversity of horseshoe bats resides, especially the two species (Rhinolophus sinicus and Rhinolophus affinis), mentioned above ( Figure 4) [7,100]. This suggests that the next CoV outbreak can be geographically predicted by the specific bat species' distribution.     Several hypotheses have been proposed based on available evidence regarding the contributions of bat species diversity together with their respective habits towards CoV epidemiology. Firstly, since both SARS-like CoV and SADS-like CoV are found in both Indomalaya and Palearctic biogeographic realms and a few horseshoe bat species like Rhinolophus ferrumequinum reside across both realms, these horseshoe bats may act as a bridge to carry the CoVs from one realm to another (Figure 4) [44,62,122,[126][127][128][129][130][131]. Secondly, two CoV species from Nobecovirus, Ro-BatCoV HKU9 and Ro-BatCoV GCCDC1, are found in Rousettus leschenaultia, a species of fruit bats located in southern Asia [27,28,60]. Later studies also discovered Ro-BatCoV HKU9 from other species of Rousettus bats and Ro-BatCoV GCCDC1 from Eonycteris spelaea [132]. Most of the Rousettus and Eonycteris bats reside within the Indomalaya realm, with a few species of Rousettus bats found in the Afrotropic realm ( Figure 5). Therefore, these ecozones may become hotspots for potential interspecies jumping events of Nobecovirus. Thirdly, Merbecovirus diversity is positively correlated with the diversity of bats in the family Vespertilionidae. It is observed that CoVs from Merbecovirus are able to infect different bat genera in the large Vespertilionidae family with distinct and diverse habitats, providing a clue that Merbecovirus is more widespread geographically than Sarbecovirus and Norbecovirus ( Figure 5) [23,24,28,29,32,61]. Several hypotheses have been proposed based on available evidence regarding the contributions of bat species diversity together with their respective habits towards CoV epidemiology. Firstly, since both SARS-like CoV and SADS-like CoV are found in both Indomalaya and Palearctic biogeographic realms and a few horseshoe bat species like Rhinolophus ferrumequinum reside across both realms, these horseshoe bats may act as a bridge to carry the CoVs from one realm to another ( Figure 4) [44,62,122,[126][127][128][129][130][131]. Secondly, two CoV species from Nobecovirus, Ro-BatCoV HKU9 and Ro-BatCoV GCCDC1, are found in Rousettus leschenaultia, a species of fruit bats located in southern Asia [27,28,60]. Later studies also discovered Ro-BatCoV HKU9 from other species of Rousettus bats and Ro-BatCoV GCCDC1 from Eonycteris spelaea [132]. Most of the Rousettus and Eonycteris bats reside within the Indomalaya realm, with a few species of Rousettus bats found in the Afrotropic realm ( Figure 5). Therefore, these ecozones may become hotspots for potential interspecies jumping events of Nobecovirus. Thirdly, Merbecovirus diversity is positively correlated with the diversity of bats in the family Vespertilionidae. It is observed that CoVs from Merbecovirus are able to infect different bat genera in the large Vespertilionidae family with distinct and diverse habitats, providing a clue that Merbecovirus is more widespread geographically than Sarbecovirus and Norbecovirus ( Figure 5) [23,24,28,29,32,61].

Sarbecovirus (betaCoVs): SARS-like CoV in Horseshoe Bats
Since the first discovery of SARS-like CoV in Chinese horseshoe bats in Hong Kong in 2005, several molecular epidemiology studies have been carried out to look for this highly fatal CoV globally [7,43,44,64,126,127,129,130,133]. Several conclusions can be drawn from these studies. Firstly, most bat SARS-like CoVs were detected in various species of horseshoe bats (genus Rhinolophus of suborder Yinpterochiroptera), although a few strains have also been found in Aselliscus stoliczkanus, Chaerephon plicatus and Hipposideros larvatus [64,109,134]. So far, the largest population of completely sequenced bat SARS-like CoVs were found in Chinese horseshoe bats (Rhinolophus sinicus) [7,64]. However, there are not sufficient data to establish the prevalence of SARS-like CoVs in different bat host species, especially the species under the genus Rhinolophus. Interestingly, geographical factor does contribute to the diversity of SARS-like CoVs. Available genome sequences showed that the majority of SARS-like CoVs found in Yunnan province shared higher nucleotide identity with human and civet SARS-CoVs in the range of 90-95%, while those found in southeast China, Korea and Europe shared only 77-90% genome nucleotide identity (unpublished data) [64,129,131,135]. However, despite having significantly different genomes, especially in the receptor binding domain where over 20% nucleotide differences were observed, both the SARS-like CoVs found in Yunnan province (southwest China) and those from southeast China were able to infect Chinese horseshoe bats (unpublished data). Further knowledge of the interaction between SARS-like CoVs and bat host receptors will shed light upon this issue. Moreover, bat SARS-like CoVs are only found in Asia, Europe and Africa, but not America or Australasia ( Figure 5). This is probably because of a relatively small population of horseshoe bats in America and Australasia. Moreover, Yunnan Province in China is an ecological niche where many different species of horseshoe bats reside (Figure 4). It is also the geographical region where the highest diversity of bat SARS-like CoV was observed, with many recombination events among the various strains of bat SARS-like CoV [64,135,136]. In fact, Yunnan is one of the provinces with the highest animal diversity in China, therefore harboring a large number of viruses and facilitating their recombination and interspecies jumping events [137]. Intriguingly, there are no observable pathogenic features reported in horseshoe bats that are infected with SARS- Figure 5. Geographical distribution of bat CoVs from the genus Betacoronavirus. Each colored region represents the country which reported the discovery of bat CoV from different sub-genera. Navy-blue regions represent the countries which discovered bat CoVs from Sarbecovirus. Yellow regions represent the countries which discovered bat CoVs from Merbecovirus. Purple regions represent the countries which discovered bat CoVs from Nobecovirus.

Sarbecovirus (betaCoVs): SARS-like CoV in Horseshoe Bats
Since the first discovery of SARS-like CoV in Chinese horseshoe bats in Hong Kong in 2005, several molecular epidemiology studies have been carried out to look for this highly fatal CoV globally [7,43,44,64,126,127,129,130,133]. Several conclusions can be drawn from these studies. Firstly, most bat SARS-like CoVs were detected in various species of horseshoe bats (genus Rhinolophus of suborder Yinpterochiroptera), although a few strains have also been found in Aselliscus stoliczkanus, Chaerephon plicatus and Hipposideros larvatus [64,109,134]. So far, the largest population of completely sequenced bat SARS-like CoVs were found in Chinese horseshoe bats (Rhinolophus sinicus) [7,64]. However, there are not sufficient data to establish the prevalence of SARS-like CoVs in different bat host species, especially the species under the genus Rhinolophus. Interestingly, geographical factor does contribute to the diversity of SARS-like CoVs. Available genome sequences showed that the majority of SARS-like CoVs found in Yunnan province shared higher nucleotide identity with human and civet SARS-CoVs in the range of 90-95%, while those found in southeast China, Korea and Europe shared only 77-90% genome nucleotide identity (unpublished data) [64,129,131,135]. However, despite having significantly different genomes, especially in the receptor binding domain where over 20% nucleotide differences were observed, both the SARS-like CoVs found in Yunnan province (southwest China) and those from southeast China were able to infect Chinese horseshoe bats (unpublished data). Further knowledge of the interaction between SARS-like CoVs and bat host receptors will shed light upon this issue. Moreover, bat SARS-like CoVs are only found in Asia, Europe and Africa, but not America or Australasia ( Figure 5). This is probably because of a relatively small population of horseshoe bats in America and Australasia. Moreover, Yunnan Province in China is an ecological niche where many different species of horseshoe bats reside (Figure 4). It is also the geographical region where the highest diversity of bat SARS-like CoV was observed, with many recombination events among the various strains of bat SARS-like CoV [64,135,136]. In fact, Yunnan is one of the provinces with the highest animal diversity in China, therefore harboring a large number of viruses and facilitating their recombination and interspecies jumping events [137]. Intriguingly, there are no observable pathogenic features reported in horseshoe bats that are infected with SARS-like CoVs [7,138]. In fact, there are few reports on bats being symptomatic due to either alphaCoVs or betaCoVs infection. This special feature makes bats the ideal reservoir hosts for CoVs.

Merbecovirus (betaCoVs)
When MERS-CoV was first isolated in the Middle East in 2012 and its genome sequenced; it was found that MERS-CoV was most closely related to Ty-BatCoV HKU4 in Tylonycteris pachypus and Pi-BatCoV HKU5 in Pipistrellus abramus, which were the only known members of Merbecovirus at that time [28,46]. Since the emergence of MERS-CoV, more than 10 additional betaCoVs of the subgenus Merbecovirus have been discovered from various bats of suborder Yangochiroptera in Asia, Europe, Africa and North and South America, most commonly from bats of family Vespertilionidae [23,24,29,31,32,35,42,61,105,110]. So far, no betaCoVs from the sub-genus Merbecovirus have been detected in bats of suborder Yinpterochiroptera. However, unlike SARS-CoV, of which most of the genome sequences of bat SARS-CoVs are >90% identical to those of civet and human SARS-CoVs, the genome sequences of MERS-CoV in human and dromedaries possess only around 65-80% nucleotide identities to those of the other members of subgenus Merbecovirus from different bats [23,24,29,31,32,46,61]. Compared to Sarbecovirus and Nobecovirus, the bat hosts that harbor viruses from the Merbecovirus sub-genus are more diverse. This is in line with the presence of more CoV species belonging to Merbecovirus.

Nobecovirus (betaCoVs)
Ro-BatCoV HKU9 and Ro-BatCoV GCCDC1 are the two CoV species from the sub-genus Nobecovirus. Compared with other bat CoVs from Sarbecovirus and Merbecovirus, Nobecovirus involves fewer bat host species. This might be due to their host specificity towards fruit bats like Rousettus bats and Eonycteris bats, which belong to the family Pteropodidae [28,60,132]. The majority of bat species from the Pteropodidae family are found in tropical and subtropical areas, which is consistent with the locations at which Ro-BatCoV HKU9 and Ro-BatCoV GCCDC1 are reported [27,60,65,132]. Several other deadly zoonotic viruses are also found in several bats belonging to the Pteropodidae family by molecular detection or viral isolation, including Ebola virus from Epomops franqueti, Hypsignathus monstrosus and Myonycteris torquata, Marburg virus from Rousettus aegyptiacus, Hendra virus from Pteropus alecto, Pteropus conscpicillatus, Pteropus poliocephalus and Pteropus scapulatus, and Nipah virus from Pteropus lylei and Pteropus hypomelanus [14,[139][140][141][142][143]. Pteropine orthoreovirus (PRV) is a zoonotic virus from the family Rheoviridae discovered from humans and Pteropus bats [144]. Interestingly, Ro-BatCoV GCCDC1 has been reported to harbor a p10 fusogenic gene, which was obtained from orthoreovirus [27]. This evidence suggests that CoVs are capable of undergoing recombination with viruses from other families, especially in the case of Nobecovirus, which share similar host tropism with the deadly viruses like Ebola virus, Marburg virus, Hendra virus and Nipah viruses. The zoonotic potential of Nobecovirus should not be underestimated, and surveillance of the Pteropodidae family in tropical and subtropical regions like Southeast Asia for emerging CoVs is necessary for future outbreak precaution.

Concluding Remarks
In the last 15 years, we have witnessed a large number of novel CoVs being discovered. Surprisingly, bats are the group of mammals that harbor the largest number of CoVs. All these discoveries and the genomes sequenced have given us unprecedented opportunities to understand the evolution of CoVs as well as the paths of interspecies transmission, which sometimes have led to devastating outbreaks, such as SARS and MERS. In the next decade, using the genome information and with the help of sophisticated technologies such as construction of wild-type and mutant infectious clones, we would be able to design experiments to further dissect the molecular mechanism of interspecies jumping. Moreover, more intensive surveillance should be performed in geographical areas which are relatively under-studied, so that a more comprehensive and detailed picture of the global epidemiology can be grasped.
Author Contributions: A.C.P.W., S.K.P.L. and P.C.Y.W. reviewed and analyzed the literature and data; A.C.P.W., S.K.P.L. and P.C.Y.W. drafted the manuscript. A.C.P.W., X.L., S.K.P.L. and P.C.Y.W. revised and edited the manuscript.