Replication of MERS and SARS coronaviruses in bat cells offers insights to their ancestral origins.

Previous findings of Middle East Respiratory Syndrome coronavirus (MERS-CoV)-related viruses in bats, and the ability of Tylonycteris-BatCoV HKU4 spike protein to utilize MERS-CoV receptor, human dipeptidyl peptidase 4 hDPP4, suggest a bat ancestral origin of MERS-CoV. We developed 12 primary bat cell lines from seven bat species, including Tylonycteris pachypus, Pipistrellus abramus and Rhinolophus sinicus (hosts of Tylonycteris-BatCoV HKU4, Pipistrellus-BatCoV HKU5, and SARS-related-CoV respectively), and tested their susceptibilities to MERS-CoVs, SARS-CoV, and human coronavirus 229E (HCoV-229E). Five cell lines, including P. abramus and R. sinicus but not T. pachypus cells, were susceptible to human MERS-CoV EMC/2012. However, three tested camel MERS-CoV strains showed different infectivities, with only two strains capable of infecting three and one cell lines respectively. SARS-CoV can only replicate in R. sinicus cells, while HCoV-229E cannot replicate in any bat cells. Bat dipeptidyl peptidase 4 (DPP4) sequences were closely related to those of human and non-human primates but distinct from dromedary DPP4 sequence. Critical residues for binding to MERS-CoV spike protein were mostly conserved in bat DPP4. DPP4 was expressed in the five bat cells susceptible to MERS-CoV, with significantly higher mRNA expression levels than those in non-susceptible cells (P = 0.0174), supporting that DPP4 expression is critical for MERS-CoV infection in bats. However, overexpression of T. pachypus DPP4 failed to confer MERS-CoV susceptibility in T. pachypus cells, suggesting other cellular factors in determining viral replication. The broad cellular tropism of MERS-CoV should prompt further exploration of host diversity of related viruses to identify its ancestral origin.

Although dromedary camels are now known to be the immediate animal source of the recent MERS epidemic, the evolutionary origin of MERS-CoV remains obscure 31,32 . When the virus was first discovered, it was found to be closely related to Tylonycteris bat CoV HKU4 (Ty-BatCoV HKU4) and Pipistrellus bat CoV HKU5 (Pi-BatCoV HKU5) previously discovered in lesser bamboo bat (Tylonycteris pachypus) and Japanese pipistrelle (Pipistrellus abramus) respectively in Hong Kong 3,11,33,34 . Five other lineage C betacoronaviruses closely related to MERS-CoV were subsequently detected in bats, including BtVs-BetaCoV/SC2013, Hypsugo BatCoV HKU25 from China bats, and coronavirus Neoromicia/PML-PHE1/RSA/2011 (NeoCoV), BtCoVNeo5038/KZN/RSA/2015, and BatCoV PREDICT/PDF-2180 from African bats [35][36][37][38][39] . Besides bats, a lineage C betacoronavirus, Erinaceus CoV VMC/DEU, subsequently defined as a novel species, Hedgehog coronavirus 1, has also been discovered in European hedgehogs, a group of animals being phylogenetically closely related to bats 40 . While none of these animal viruses represent the immediate ancestor of MERS-CoV, the spike protein of Ty-BatCoV HKU4 was most closely related to that of MERS-CoV, and was shown to utilize the MERS-CoV receptor, human dipeptidyl peptidase 4 (hDPP4) or CD26, for cell entry 41,42 . Previous studies have also shown that MERS-CoV was able to infect bat cell lines and Jamaican fruit bats 43,44 . These findings suggest that bats may be the primary host of the ancestor of MERS-CoV.
Although MERS-CoV has been shown to replicate in various animal cell lines including bat cells [43][44][45][46][47][48] , the broad tissue tropism was mainly demonstrated using type strain EMC/2012, and no comparison was made between different MERS-CoV strains. Moreover, cells from the bat hosts of MERS-CoV-related viruses, such as T. pachypus and P. abramus which harbor Ty-BatCoV HKU4 and Pi-BatCoV HKU5, respectively, were not included in previous studies, which may be due to the geographical limitation of these bat species. To better understand the replicative ability of MERS-CoV in bat cells, which may provide clues on the origin of MERS-CoV, we developed diverse primary bat cell lines from different bat species, including Rhinolophus sinicus (the host of SARSr-Bat-CoV) and T. pachypus (the host of Ty-BatCoV HKU4), and tested their susceptibilities to infection by different strains of MERS-CoV, SARS-CoV, and HCoV-229E. The DPP4 mRNA sequences of six bat species and their expression in bat cells were determined to correlate with viral replication results. Our findings showed differential cell tropism between different strains of MERS-CoV, SARS-CoV, and HCoV-229E, which offers insights to the origin of MERS-CoV.

Results
Five of 12 tested bat cell lines are susceptible to MERS-CoV EMC/2012 infection Since lineage C betacoronaviruses closely related to MERS-CoV were detected in bats, we developed 12 diverse primary bat cell lines from seven different bat species, including Hipposideros pomona, Miniopterus pusillus, Myotis ricketti, Pipistrellus abramus (the host of Pi-BatCoV HKU5), Rhinolophus sinicus (the host of SARSr-BatCoV and Rs-BatCoV HKU2), Tylonycteris pachypus (the host of Ty-BatCoV HKU4), Rousettus leschenaultii (the host of many viruses including Ro-BatCoV HKU9), which were subjected to infection with MERS-CoV at multiplicity of infection (MOI) of 1. Viral titers were determined by RT-qPCR on day 5 p.i. Five of the 12 cell lines (M. ricketti lung, P. abramus kidney, R. sinicus kidney and lung, and R. leschenaulti kidney cells) and Vero cells propagated MERS-CoV with at least one log 10 increase in viral load. The highest increase in viral load was observed in R. sinicus kidney and lung cells, which was comparable with that observed in Vero cells (Fig. 1). Cytopathic effects (CPEs) were observed in infected M. ricketti lung and R. sinicus lung cells with rounding of cells (Fig. 2). The infectivities of the viruses from culture supernatants were confirmed by passage in Vero cells with CPE. H. pomona kidney, M. pusillus kidney, M. ricketti kidney, P. abramus lung, R. leschenaulti lung, and T. pachypus kidney and lung cells did not support MERS-CoV infection.

SARS-CoV can replicate in R. sinicus cells
Since Chinese horseshoe bats are the major reservoir of SARS-related-CoVs, we also tested the 12 bat cell lines, including R. sinicus cells, for susceptibility to a clinical strain of SARS-CoV. SARS-CoV strain HKU-39849 can replicate in R. sinicus kidney but not lung cells, with at least one log 10 increase in viral load (Fig. 3a). The other bat cells did not support SARS-CoV infection.

HCoV-229E cannot replicate in bat cells
The 12 bat cells were also tested for susceptible to HCoV-229E infection. HCoV-229E strain ATCC VR-740, previously isolated from a man with upper respiratory illness, cannot replicate in any of the tested bat cells (Fig. 3b).

mRNA transcript sequence analysis and expression of DPP4 in bat cells
Partial DPP4 mRNA transcript sequences (corresponding to nt 688-1040 of hDPP4 which includes residue 229-346, where the critical residues for MERS-CoV spike protein binding are present) were determined for six of the seven bat species from which the 12 tested bat cells were developed. The sequence of M. pusillus was not determined, as RT-PCR for the DPP4 mRNA from bat cells was negative. Phylogenetic analysis showed that the bat DPP4 mRNA sequences formed a distinct cluster being closely related to sequences from human and nonhuman primates; while the dromedary camel DPP4 sequence was most closely related to that of wild boar (Fig. 4a). Previous studies have identified 14 critical residues in hDPP4 for binding of MERS-CoV spike protein 53,54 . Upon multiple sequence alignment of the corresponding regions that contain these critical residues, most of the residues are conserved in the six bat DPP4 sequences in this study (Fig. 3b). Notably, T. pachypus DPP4 (Tp-DPP4) contains an I → K substitution at position 295 compared with hDPP4. On the other hand, residue R336 of hDPP4 was only conserved in T. pachypus and H. pomona, both not susceptible to MERS-CoV infection.
RT-qPCR of bat DPP4 mRNA in the corresponding bat cells was performed to determine the mRNA expression abramus kidney (susceptible to MERS-CoV) was significantly higher than that in its lung (non-susceptible) cells (P = 0.0185 by Student's t test). Similarly, the mRNA expression level in R. leschenaulti kidney (susceptible to MERS-CoV) was significantly higher than that in its lung Compared with bat cells that are non-susceptible, bat cells that are susceptible to MERS-CoV showed a significantly higher mRNA expression level (P = 0.0174 by Student's t test) (Fig. 5).
Overexpression of Tp-DPP4 does not confer MERS-CoV EMC/2012 susceptibility to T. pachypus cells Since DPP4 was not expressed in T. pachypus cells while this bat species hosts Ty-BatCoV HKU4, a close relative of MERS-CoV which can utilize hDPP4 for cell entry 41,42 , we overexpressed Tp-DPP4 in T. pachypus lung and kidney cells for infection by MERS-CoV. While a slight increase (less than one log 10 ) in viral replication was noted in both mock and Tp-DPP4-overexpressed cells on day 5 p.i., no significant difference was noted between mock and Tp-DPP4-overexpressed cells (Fig. 6), suggesting that overexpression of Tp-DPP4 does not confer MERS-CoV susceptibility to T. pachypus cells.

Discussion
The present study provides further support to the bat origin of MERS-CoV. Previous studies have shown that MERS-CoV can infect cell lines from different animals, including cells from various bat species [43][44][45][46][47] . However, bat cells from host species of MERS-CoV-related viruses, such as T. pachypus and P. abramus which harbor Ty-BatCoV HKU4 and Pi-BatCoV HKU5, respectively, were not included in these studies. In this study, MERS-CoV EMC/2012 was able to replicate in cells from four different bat genera/species belonging to three different bat families including Pteropodidae, Rhinolophidae, and Vespertilionidae. None of these bat cells were previously reported to support MERS-CoV replication. While lineage C betacoronaviruses have not been detected in bats outside the Vespertilionidae family, such diverse cellular tropism should prompt further studies to explore the host diversity of this group of CoVs and hence the possible evolutionary origin of MERS-CoV. In particular, the ability of MERS-CoV to replicate in cells from P. abramus, the host of Pi-BatCoV HKU5, may suggest that the ancestor of MERS-CoV could have originated from Pipistrellus-related bats. Interestingly, MERS-CoV showed the highest replicative ability in cells from R. sinicus. This suggests that MERS-CoV may potentially infect this bat species which is also the natural reservoir of SARSr-CoVs and animal origin of SARS-CoV 21 . Nevertheless, three other MERS-CoV strains belonging to either clade A2 or B3 showed differential replicative abilities in the five bat cell lines that were susceptible to EMC/ 2012 strain (clade A1), with strain D998/15 (clade A2) unable to replicate in all five tested cell lines. Interestingly, it has been shown in a recent study that some clade C MERS-CoV strains from Nigeria and Burkina Faso replicated at a lower titre in Calu-3 cells 52 . This is in line with the present study that our clade A2 strain D998/15 (with its spike gene belonging to clade C), conferred less broad cellular tropism. Our findings suggest that different clades or strains of MERS-CoV may possess different cellular tropism and host range, while all MERS-CoVs can replicate well in the immune-deficient Vero cells.
The broader cellular tropism of MERS-CoV than SARS-CoV and HCoV-229E in bat cells may reflect the host diversity of lineage C betacoronaviruses in bats. Bats are now known to be the recent origin of at least two human CoVs, including SARS-CoV and HCoV-229E. SARS-CoV is likely a recombinant virus arising from horseshoe bats before it jumped to civet as the intermediate host and then human 16,24 . SARSr-CoVs in bats were shown to utilize the SARS-CoV receptor, human angiotensinconverting enzyme 2 (hACE2), for cell entry 55 . Similar to SARS-CoV, HCoV-229E is likely to have originated from bats. CoVs closely related to HCoV-229E have been detected in bats of the genus Hipposideros in Africa 56 . More recently, CoVs even closer to HCoV-229E were identified in dromedary camels in the Middle East, which were able to utilize the receptor of HCoV-229E, human aminopeptidase N (hAPN) for cell entry 57 . This suggests Recently, a CoV, closely related to HCoV-NL63 in most genome regions except the spike protein, was detected from a bat of the genus Triaenops in Kenya 58 . However, further studies are required to identify even closer relatives of HCoV-NL63 to ascertain its possible bat origin. In contrast to SARSr-CoVs and HCoV-229E which were mainly found in horseshoe and roundleaf bats, respectively, a more diverse host range was observed in lineage C betacoronaviruses. Yet, the different bat species harboring lineage C betacoronaviruses all belonged to the family Vespertilionidae. Specifically, Ty-BatCoV HKU4, Pi-BatCoV HKU5, Hp-BatCoV HKU25, and BtVs-Beta-CoV/SC2013 were detected in bats belonging to the genera, Tylonycteris, Pipistrellus, Hypsugo, and Verspetilio, respectively, in China, whereas NeoCoV and BatCoV PREDICT/PDF-2180 were detected in bats belonging to Neoromicia and Pipistrellus in Africa 36,37 . In this study, SARS-CoV could only replicate in R. sinicus kidney cells but not other bat cells, which reflects its evolutionary origin from horseshoe bats. On the other hand, HCoV-229E was unable to replicate in any tested bat cells including cells from Hipposideros. This may suggest a relatively narrow host range of SARS-CoV and HCoV-229E in bats compared with lineage C betacoronaviruses including MERS-CoV.
The close phylogenetic relationship between bat and primate DPP4 sequences may reflect the replicative ability of MERS-CoV in bat cells. Moreover, cells that supported MERS-CoV replication showed significantly higher DPP4 mRNA expression than those that were non-susceptible to MERS-CoV, suggesting that cellular DPP4 expression is critical for viral infection. Since Ty-BatCoV HKU4 from T. pachypus was previously shown to be able to utilize hDPP4 for cell entry 41,42 , we expected MERS-CoV to be able to utilize Tp-DPP4 for receptor binding and infect T. pachypus cells. While the inability of MERS-CoV to replicate in T. pachypus cells may be partly explained by the lack of DPP4 expression, overexpression of Tp-DPP4 was unable to confer viral susceptibility. Since mRNA expression may not always correlate with cell surface expression of the receptor, further studies are required to study the receptorbinding interphase between MERS-CoV and Tp-DPP4 and whether other cellular factors may play a role in determining viral replicative abilities in T. pachypus cells, which may offer further insights into the evolutionary origin and mechanisms of interspecies transmission of MERS-CoV. Interestingly, one of the critical residues for binding to MERS-CoV spike protein, R336, was only conserved in T. pachypus and H. pomona DPP4 (cells from these two species did not support MERS-CoV replication), but not in the other four sequenced bat DPP4 (cells from these four species supported MERS-CoV replication). This suggests that this residue may not be critical for receptor binding. Binding and mutagenesis studies may help better understand the role of receptor-binding-interphase during viral evolution and interspecies jumping.

Ethics statement
The collection of bat samples for developing bat cell lines was approved by the Committee on the Use of Live

Cell lines
The bat cells, Vero (African green monkey kidney) cells, and HFL (human embryonic lung fibroblast) cells used in this study are described in Table 1. For development of primary bat cell lines, bats were captured in Hong Kong and euthanized before dissection for recovery of cells from organs aseptically. Briefly, the kidney and lung tissue were rinsed with cold PBS and cut into small pieces. Cold 0.25% trypsin-EDTA was added to the tissues and incubated at 4°C overnight. Tissues were then incubated at 37°C on shaking platform for 30 min. Supernatants were filtered through cell strainers to remove large pieces of tissues 59 . Bat cells were harvested by spinning down the supernatant at 1200 rpm for 8 min and were grown in DMEM/ F12 supplemented with 15% FBS. Vero and HFL cells were grown in MEM supplemented with 10% FBS. All cells were incubated at 37°C with 5% CO 2 .

Infection of bat cell lines
Viral titers were determined as median tissue culture infective dose (TCID 50 ) per ml in confluent cells in 96well microtiter plates. Cells were seeded onto 24-well tissue culture plates, at 2 × 10 5 cells per well with the respective medium and incubated at 37°C and 5% CO 2 for 24 h prior to experiment. Cells were washed once with PBS and inoculated with 1 MOI of MERS-CoV or SARS-CoV, or 0.01 MOI of 229E for 1 h. After 1 h of viral adsorption, the supernatant was removed and cells were washed twice with PBS. The cells were maintained in MEM supplemented with 1% FBS for Vero and HFL cells and DMEM/F12 supplemented with 1% FBS for bat cells, before further incubation for 5 days.
Amplification and sequencing of partial bat DPP4 mRNA transcripts Total RNA was extracted from bat cell lysates using RNeasy Mini Spin Column (QIAgen). cDNA was PCR amplified with primers, 5′-GTCACCAGAGGGTCAT AAA-3′ and 5′-CCACTTCCTCTGCCATCAAA-3′. The PCR mixture (25 μl) contained cDNA, PCR buffer, 200 μM (each) dNTPs, and 1.0 U Iproof Polymerase (Bio-Rad, Hercules, CA, USA). The mixtures were amplified for 40 cycles of 98°C for 10 s, 55°C for 30 s, and 72°C for 72 s and a final extension at 72°C for 10 min in an automated thermal cycler (Applied Biosystems, Foster City, CA, USA). RT-PCR products were gel purified using QIAquick gel extraction kit (Qiagen), and sequenced with an ABI Prism 3700 DNA Analyzer (Applied Biosystems). The sequences obtained were compared with sequences of DPP4 genes in GenBank database. Phylogenetic tree construction was performed based on an amino acid alignment of partial DPP4 sequences (corresponding to residue 229-346 of hDPP4) using the Neighbor-Joining method with JTT model by MEGA 6.0, with bootstrap values calculated from 1000 trees.

DPP4 expression analysis
To study DPP4 expression profiles in different bat cell lines, cell lysates were collected for total RNA extraction using RNeasy Mini Spin Column (QIAgen). RNA was eluted in 50 μl of RNase-free water and was used as a template for one-step RT-qPCR with SuperScript III platinum One-step qRT-PCR system (Invitrogen, Carlsbad, CA, USA). RT-qPCR assays were performed using conserved primers designed by multiple alignment of available bat DPP4 gene sequences (Table 2), using β-actin for normalization. cDNA was amplified in a LightCycler 480 (Roche, Basel, Switzerland) with 25 μl reaction mixtures containing 2 × reaction mix, 5 μl RNA, 25 μM ROX reference dye, 50 μM primers, and 10 μM probe at 50°C for 30 min, then 95°C for 2 min followed by 50 cycles of denaturation, annealing and extension. Experiments were performed in triplicates, and results  were expressed as the mean expression level of DPP4/ β-actin. The relative expression between different bat cells was then calculated by the ΔΔCt method.
tpDPP4 overexpression in T. pachypus cells T. pachypus DPP4 (tpDPP4) sequence was cloned into pLenti7.3/V5-TOPO vector (Invitrogen). The construct was transfected into 293FT cells together with ViraPower Packaging Mix (Invitrogen) using Lipofectamine 2000 (Life Technologies, Carlsbad, CA, USA). Lentivirus was harvested from the supernatant and concentrated using PEG-it (System Biosciences, Palo Alto, USA). T. pachypus cells resistant to MERS-CoV were transduced using the concentrated lentivirus for tpDPP4 overexpression and were subsequently subjected to inoculation with 1 MOI of MERS-CoV.

Nucleotide sequence accession numbers
The nucleotide sequences of bat DPP4 obtained from this study have been deposited in the GenBank sequence database under accession numbers MH345671-MH345676.