Figure 1

Figure 1. . . Location of bat sampling sites in Ghana and Europe. The 7 sites in Ghana (A) and the 5 areas in Europe (B) are marked with dots and numbered from...

Fecal specimens were collected from 10 bat species in Ghana and 4 Pipistrellus species in Europe (Table 1). Bats were caught during 2009–2011 with mist nets, as described (9), in 7 locations across Ghana and 5 areas in Germany, the Netherlands, Romania, and Ukraine (Figure 1). The species, age, sex, reproductive status, and morphologic measurements of the bats were recorded. Fecal pellets were collected and suspended in RNAlater Stabilization Reagent (QIAGEN, Hilden, Germany). RNA was purified as described (11). CoV was detected by using nested reverse transcription PCR (RT-PCR) targeting the RNA-dependent RNA polymerase (RdRp) gene (12) (see Table 1 for assay oligonucleotides).

A novel CoV was detected in insectivorous Nycteris cf. gambiensis specimens (Technical Appendix); GenBank accession nos. JX899382–JX899384). A real-time RT-PCR was designed to permit sensitive and quantitative detection of this CoV (Table 1). Only Nycteris bats were positive for CoV (46 [24.9%] of 185 specimens) (Table 1). Demographic factors predictive of CoV in captured Nycteris bats were assessed. Juvenile bats and lactating females were significantly more likely to be CoV-infected than were adult and nonlactating female bats, respectively (Table 2). Virus concentrations in feces from Nycteris bats were high (median 412,951 RNA copies/g range 323–150,000,000 copies/g).

The 398-bp CoV RdRp screening fragment was extended to 816 bp, as described (5), to enable more reliable taxonomic classification. We previously established RdRp-grouping units (RGU) as a taxonomic surrogate to enable prediction of CoV species on the basis of this 816-bp fragment when no full genome sequences could be obtained. According to our classification, the amino acid sequences in the translated 816-bp fragment of the tentative betacoronavirus species (RGU) differed from each other by at least 6.3% (5). The new Nycteris bat CoV differed from the 2c-prototype viruses HKU4 and HKU5 by 8.8%–9.6% and from EMC/2012 by 7.5% and thus constituted a novel RGU. A partial RdRp sequence fragment of a P. pipistrellus bat CoV from the Netherlands, termed VM314 (described by us in [10]), was completed toward the 816-bp fragment to refine the RGU classification of EMC/2012. EMC/2012 differed from VM314 by only 1.8%.

Because of the genetic similarity between EMC/2012 and VM314, we specifically investigated Pipistrellus bats from 4 European countries for 2c betacoronaviruses. We detected betacoronaviruses in 40 (14.7%) of 272 P. pipistrellus, P. nathusii, and P. pygmaeus bats from the Netherlands, Romania, and Ukraine (Table 1; GenBank accession nos. KC243390-KC243392) that were closely related to VM314. The VM314-associated Pipistrellus bat betacoronaviruses differed from EMC/2012 by 1.8%. The difference between EMC/2012 and HKU5 was 5.5%–5.9%. In summary, HKU5, EMC/2012, and the VM314-associated clade form 1 RGU according to our classification system, and the VM314-Pipistrellus bat clade contains the closest relatives of EMC/2012. HKU4 and the Nycteris CoV define 2 separate tentative species in close equidistant relationship.

Figure 2

Figure 2. . . RNA-dependent RNA polymerase (RdRp) gene and Spike genephylogenies including the novel betacoronaviruses from bats in Ghana and Europe. A) Bayesian phylogeny of an 816-nt RdRp gene sequence fragment corresponding...

We conducted a Bayesian phylogenetic analysis. In this analysis, the Nycteris bat CoV clustered as a phylogenetically basal sister clade with HKU4, HKU5, and EMC/2012 and the associated European Pipistrellus viruses (Figure 2).

To confirm the RdRp-based classification, we amplified the complete glycoprotein-encoding Spike gene and sequenced it for the novel Nycteris bat virus. The phylogenetically basal position of the novel Nycteris bat virus within the 2c clade resembled that in the CoV RdRp gene (Figure 2). Partial sequences that could be obtained from the 3′-end of the Spike gene of three 2c Pipistrellus bat betacoronaviruses confirmed their relatedness to EMC/2012 (Figure 2).

We detected novel clade 2c betacoronaviruses in Nycteris bats in Ghana and Pipistrellus bats in Europe that are phylogenetically related to the novel hCoV EMC/2012. All previously known 2c bat CoVs originated from vespertilionid bats: VM314 originated from a P. pipistrellus bat from the Netherlands and HKU4 and HKU5 originated from Tylonycteris pachypus and P. abramus bats, respectively, from the People’s Republic of China. The Nycteris bat virus in Africa extends this bat CoV clade over 2 different host families, Nycteridae and Vespertilionidae (Technical Appendix). Detection of genetically related betacoronaviruses in bats from Africa and Eurasia parallels detection of SARS-CoV in rhinolophid bats from Eurasia and related betacoronaviruses in hipposiderid bats from Africa (9).

The relatedness of EMC/2012 to CoVs hosted by Pipistrellus bats at high prevalence across different European countries and the occurrence of HKU5 in bats of this genus from China highlight the possibility that Pipistrellus bats might indeed host close relatives of EMC/2012. This suspicion is supported by observations that tentative bat CoV species (RGUs) are commonly detected within 1 host genus (5). Within the Arabian Peninsula, the International Union for Conservation of Nature (www.iucn.org) lists 50 bat species, including P. arabicus, P. ariel, P. kuhlii, P. pipistrellus, P. rueppellii, and P. savii bats. Because of the epidemiologic link of EMC/2012 with the Arabian Peninsula (6,7), bats from this area should be specifically screened.

The genomic data suggest that EMC/2012, like hCoV-229E and SARS-CoV, might be another human CoV for which an animal reservoir of closely related viruses could exist in Old World insectivorous bats (4,9). Whether cross-order (e.g., chiropteran, carnivore, primate) host switches, such as suspected for SARS-CoV, have occurred for 2c clade bat CoVs remains unknown. However, we showed previously that CoVs are massively amplified in bat maternity colonies in temperate climates (13). This amplification also might apply to the Nycteris bat CoV because, as shown previously for vespertilionid bats from temperate climates (14), detection rates of CoV are significantly higher among juvenile and lactating Nycteris bats. In light of the observed high virus concentrations, the use of water from bat caves and bat guano as fertilizer for farming and the hunting of bats as wild game throughout Africa (15) may facilitate host switching events. To our knowledge, no CoV has been isolated directly from bats. Further studies should still include isolation attempts to obtain full virus genomes and to identify virulence factors that may contribute to the high pathogenicity of EMC/2012 (7).