Molecular Phylogeny of Hantaviruses Harbored by Insectivorous Bats in Côte d’Ivoire and Vietnam

The recent discovery of genetically distinct hantaviruses in multiple species of shrews and moles prompted a further exploration of their host diversification by analyzing frozen, ethanol-fixed and RNAlater®-preserved archival tissues and fecal samples from 533 bats (representing seven families, 28 genera and 53 species in the order Chiroptera), captured in Asia, Africa and the Americas in 1981–2012, using RT-PCR. Hantavirus RNA was detected in Pomona roundleaf bats (Hipposideros pomona) (family Hipposideridae), captured in Vietnam in 1997 and 1999, and in banana pipistrelles (Neoromicia nanus) (family Vespertilionidae), captured in Côte d’Ivoire in 2011. Phylogenetic analysis, based on the full-length S- and partial M- and L-segment sequences using maximum likelihood and Bayesian methods, demonstrated that the newfound hantaviruses formed highly divergent lineages, comprising other recently recognized bat-borne hantaviruses in Sierra Leone and China. The detection of bat-associated hantaviruses opens a new era in hantavirology and provides insights into their evolutionary origins.

Despite their phylogenetic relatedness to the European mole (Talpa europaea) within the Laurasiatheria [20,21], as well as their rich genetic diversity, vast geographic range and ability to host many disease-causing viruses [22][23][24], bats (order Chiroptera) have not been extensively studied as potential reservoirs of hantaviruses. Although serological evidence of hantavirus infection was reported in the common serotine (Eptesicus serotinus) and greater horseshoe bat (Rhinolophus ferrumequinum) captured in Korea [25], genetic analysis of hantavirus isolates from these bat species suggested laboratory contamination [26].
The genetic diversity of newfound hantaviruses recently detected in insectivorous bats preclude any possibility of contamination: Mouyassué virus (MOYV) in the banana pipistrelle (Neoromicia nanus) from Côte d'Ivoire [27]; Magboi virus (MGBV) in the hairy slit-faced bat (Nycteris hispida) from Sierra Leone [28]; Xuan Son virus (XSV) in the Pomona roundleaf bat (Hipposideros pomona) from Vietnam [29]; Huangpi virus (HUPV) in the Japanese house bat (Pipistrellus abramus) and Longquan virus (LQUV) in the Chinese horseshoe bat (Rhinolophus sinicus), Formosan lesser horseshoe bat (Rhinolophus monoceros) and intermediate horseshoe bat (Rhinolophus affinis) from China [30]. The primary goal of this multi-national collaborative study was to extend the search for hantaviruses in bats and to obtain more of the MOYV and XSV genomes. Our data indicate that bat-borne hantaviruses and Nova virus, a hantavirus hosted by the European mole, comprise a highly divergent phylogenetic lineage, suggesting that ancestral bats and/or soricomorphs, rather than rodents, may have served as the early reservoir hosts of primordial hantaviruses.

Hantavirus Detection and Sequence Analysis
Exhaustive attempts to detect hantaviruses were unsuccessful in nearly all of the 454 bat tissue samples (Table 1 and Figure 1), despite employing oligonucleotide primers and PCR cycling conditions used to find MOYV [27] and XSV [29]. In addition, hantavirus RNA was not detected in any of the 79 rectal swab and fecal samples. Because LQUV was previously found in four species of horseshoe bats in China [30], we expected to find the same or a similar hantavirus in the greater horseshoe bat, captured on Jeju Island in Korea. However, this was not the case, in spite of using LQUVspecific primers. Nevertheless, we did manage to obtain more of the MOYV and XSV genomes. That is, the original report of MOYV in the banana pipistrelle (Figure 2A,B) was based on a 423-nucleotide region of the L segment [27]. Through repeated trial-and-error efforts, suitable primers were designed to obtain an additional 1268 nucleotides of the L segment ( Table 2).
In addition, Arai and colleagues previously reported a novel hantavirus, designated XSV, in one of five Pomona roundleaf bats, captured during July 2012 in Xuan Son National Park in Phu Tho province in northern Vietnam [29]. In analyzing archival kidney tissues from 44 Pomona roundleaf bats trapped in Tuyê n Quang and Quang Nam provinces, hantavirus L-segment sequences were detected in five animals ( Figure 2C,D). Although a 15.7%-19.2% difference was found at the nucleotide level with prototype XSV, the high amino acid sequence similarity was consistent with these sequences representing genetic variants of XSV. Pair-wise alignment and comparison of the full-length S segment of XSV, amplified and sequenced from four bats ( Table 2), indicated sequence similarity of 58.9%-60.3% at the amino acid level with LQUV, the only other bat-borne hantavirus for which the entire S segment has been sequenced. And sequence analysis of a 663-nucleotide (221 amino acid) region of the Gc envelope glycoprotein-encoding M segment showed that XSV differed by >45% from representative hantaviruses harbored by rodents and most soricomorphs. Collectively, the high level of sequence divergence in the N protein and Gc glycoprotein between XSV and other hantaviruses suggests that it might represent a new hantavirus species, using the guidelines proposed by Maes and co-workers [31]. However, the definitive taxonomic classification of XSV and other bat-borne hantaviruses must await their isolation in cell culture.    Nombre virus (SNV) and Andes virus (ANDV), as predicted using methods available on the NPS@ structure server [32]. Alpha helices are represented by blue bars, beta strands by red bars, and random coils and unclassified structures by magenta and gray bars, respectively.

Nucleocapsid Secondary Structure
In employing software available on the @NPS structure server [32], the overall predicted secondary structures of the N proteins were similar. That is, despite the relatively low amino acid sequence similarity among the rodent-, shrew-, mole-and bat-borne hantaviruses, the N protein comprised two major α-helical domains packed against a central β-pleated sheet ( Figure 2E). However, the central β-pleated sheet motif of XSV, including the RNA-binding region (amino acid positions 175 to 217), was unlike that of other hantaviruses, even that of LQUV, which more closely resembled murid rodent-borne hantaviruses, such as Hantaan virus (HTNV 76-118), Dobrava virus (DOBV Greece) and Seoul virus (SEOV 80-39) ( Figure 2E). The distinctive α-helix motif between two β-strands of the RNA-binding region, observed in the prototype mole-borne hantavirus, Nova virus (NVAV MSB95703), as well as HTNV and SEOV, but not in LQUV, may have a significant effect on binding specificity.

Phylogenetic Analysis
Phylogenetic analyses, based on S-, M-and L-genomic sequences, indicated that XSV and MOYV shared a common ancestry with other bat-borne hantaviruses (Figure 3). In all analyses, NVAV from the European mole segregated with the bat-associated hantaviruses, which was reminiscent of trees based on the complete mitochondrial genomes of the European mole and bats [20,21]. The basal position of chiropteran-borne hantaviruses and selected soricomorph-borne hantaviruses, such as Nova virus in the European mole, Thottapalayam virus in the Asian house shrew and Imjin virus in the Ussuri white-toothed shrew, in phylogenetic trees based on the S-and L-genomic sequences suggests that soricomorphs and/or chiropterans, rather than rodents, may have been the primordial mammalian hosts of ancestral hantaviruses (Figure 3). Geographic-specific clustering was evidenced by the close phylogenetic relationship between prototype XSV VN1982 from Phu Tho province and XSV F42640 and XSV F42682 from neighboring Tuyê n Quang province in northern Vietnam. On the other hand, XSV F44583, XSV 44601 and XSV 44580 from Quang Nam province in central Vietnam clustered together. Although limited differences were present in phylogenetic trees based on each segment, tree topologies were generally congruent and supported by significant bootstrap values (>70%) and posterior node probabilities (>0.70).

Bats as Hosts of Hantaviruses
The phylogeny of bats is not fully resolved [21]. The order Chiroptera was traditionally divided in two suborders, Megachiroptera and Microchiroptera. However, due to the paraphyly of the Microchiroptera, a new taxonomic nomenclature, comprising the suborder Yinpterochiroptera (megabats or fruit bats in the Pteropodidae family in Megachiroptera and a few Microchiroptera families) and Yangochiroptera (the remaining Microchiroptera families), has been proposed [33]. In the former classification, bat species hosting hantaviruses belong only to the Microchiroptera suborder, but in the Yinpterochiroptera-Yangochiroptera classification, they belong to both suborders, suggesting that primordial hantaviruses may have emerged in an early common ancestor of bats.
Within the Microchiroptera, hantaviruses are found in bats belonging to four phylogenetically distant families, namely Hipposideridae (Old World leaf-nosed bats) and Rhinolophidae (horseshoe bats) in the suborder Yinpterochiroptera, and Nycteridae (hollow-faced bats) and Vespertilionidae (vesper bats) in the suborder Yangochiroptera. The families Hipposideridae and Vespertilionidae are among the most speciose insectivorous bats, with member species distributed across Africa, Europe, Asia, the Americas and Australia. Their vast geographic distribution provides unlimited opportunities to search for related bat-associated hantaviruses.
Compared to the multitude of hantaviruses reported from approximately 50% of soricomorph species tested [34,35], the cumulative number of newly recognized bat-borne hantaviruses is exceedingly low, if one considers the 533 bat samples tested in the present study, along with the nearly 1200 bat specimens analyzed in four other studies [27][28][29][30]. The modest proportion of hantavirus RNA detection in bat tissues may be attributed to the highly divergent nature of their genomes, as well as the very focal or localized nature of hantavirus infection, small sample sizes of bat species, primer mismatches, suboptimal PCR cycling conditions, and variable tissue preservation with degraded RNA [27,29]. Alternatively, bats may be less susceptible to hantavirus infection or may have developed immune mechanisms to curtail viral replication and/or persistence. For answers to such questions, and myriad others, reagents need to be developed and multidisciplinary collaborative studies must be designed to collect optimal specimens to isolate and characterize these newfound bat-borne hantaviruses. Only then will a better understanding be gained about their evolutionary origins and phylogeography, co-evolution history, transmission dynamics and pathogenic potential.

Samples
Archival frozen, ethanol-fixed and RNAlater ® -preserved tissues from bats, captured during 1981-2012 in Brazil, China, Cote d'Ivoire, Guinea, Korea, Republic of Georgia, Vietnam and the United States ( Figure 1 and Table 1), were tested for hantavirus RNA by RT-PCR, using newly designed and previously employed oligonucleotide primers [12,18,27,29]. Of the 533 samples tested, the majority consisted of lung (310) and kidney (51) tissues (Table 1). RNA extracted from rectal swabs and feces (79) were also tested. Bats were from seven families (Hipposideridae, Molossidae, Nycteridae, Pteropodidae, Phyllostomidae, Rhinolophidae and Vespertilionidae), 28 genera and 53 species (Figure 1). The University of Hawaii Institutional Animal Care and Use Committee approved the use of archival tissues as being exempt from protocol review.

Phylogenetic Analysis
Maximum likelihood and Bayesian methods, implemented in RAxML Blackbox webserver [36] and MrBayes 3.1 [37], under the best-fit GTR+I+Γ model of evolution [38] and jModelTest version 0.1 [39], were used to generate phylogenetic trees. Two replicate Bayesian Metropolis-Hastings Markov Chain Monte Carlo runs, each consisting of six chains of 10 million generations sampled every 100 generations with a burn-in of 25,000 (25%), resulted in 150,000 trees overall. The S, M and L segments were treated separately in phylogenetic analyses. Topologies were evaluated by bootstrap analysis of 1000 iterations, and posterior node probabilities were based on 2 million generations and estimated sample sizes over 100 (implemented in MrBayes) [18].

Conclusions
Mammalian reservoirs of zoonotic viruses typically do not display host restrictions within a given taxonomic order. Also, infection is usually chronic, persistent and subclinical. For example, rodents of multiple genera and species, belonging to four subfamilies in the order Rodentia, serve as reservoirs of hantaviruses in Eurasia, Africa and the Americas and do not exhibit clinical disease or survival disadvantage. In addition, recently, hantaviruses exhibiting far greater genetic diversity have been detected in healthy-appearing shrews and moles representing many genera in six subfamilies within the order Soricomorpha in Eurasia, Africa and North America. Similarly, as mentioned earlier, bat species belonging to both suborders of Chiroptera host hantaviruses without evidence of apparent disease. However, some might contend that the low prevalence of hantavirus RNA in a few bat species, and the absence of hantavirus infection in the majority of bat species analyzed to date, would argue against a long-standing hantavirus-reservoir host relationship, and instead support spillover or host switching. That is, the gleaning feeding behavior of some bats, such as Nycteris, presents the possibility of acquired infection from excreta of well-established terrestrial reservoirs of hantaviruses. However, this seems highly improbable because bat-borne hantaviruses are among the most genetically diverse described to date.
With the discovery of divergent hantavirus lineages in three taxonomic orders of placental mammals, there is renewed interest in investigating their genetic diversity, geographic distributions, and evolutionary dynamics [34,35]. Newfound knowledge that insectivorous bats harbor a distinctly divergent lineage of hantaviruses emphasizes the truly complex evolutionary origins and phylogeography of a group of viruses once thought to be restricted to rodents. At this point, it would not be surprising if hantaviruses are found in small mammals belonging to other taxonomic orders, such as Erinaceomorpha (hedgehogs) and even Afrosoricida (tenrecs). Such anticipated discoveries may provide additional insights into the dynamics of hantavirus transmission, potential reassortment of genomes, and molecular determinants of hantavirus pathogenicity. As importantly, a sizable expansion of the hantavirus sequence database would provide valuable tools for refining diagnostic tests and enhancing preparedness for future outbreaks caused by emerging hantaviruses.