|Home | About | Journals | Submit | Contact Us | Français|
A genetically distinct hantavirus, designated Oxbow virus (OXBV), was detected in tissues of an American shrew mole (Neurotrichus gibbsii), captured in Gresham, Oregon, in September 2003. Pairwise analysis of full-length S- and M- and partial L-segment nucleotide and amino acid sequences of OXBV indicated low sequence similarity with rodent-borne hantaviruses. Phylogenetic analyses using maximum-likelihood and Bayesian methods, and host-parasite evolutionary comparisons, showed that OXBV and Asama virus, a hantavirus recently identified from the Japanese shrew mole (Urotrichus talpoides), were related to soricine shrew-borne hantaviruses from North America and Eurasia, respectively, suggesting parallel evolution associated with cross-species transmission.
Members of the genus Hantavirus (Family Bunyaviridae) are harbored by specific rodent species with which they share long-standing virus-host relationships (Monroe et al., 1999; Morzunov et al., 1998). Recently, the previously well-accepted notion of co-divergence has been opposed in favor of preferential host switching and local host-specific adaptation (Ramsden et al., 2009). Nevertheless, growing evidence suggests an additional congruent pattern of co-evolution of hantaviruses harbored by shrews. Analysis of the full genome of Thottapalayam virus (TPMV), a hantavirus isolated from the Asian house shrew (Suncus murinus) (Carey et al. 1971), demonstrates a separate phylogenetic clade consistent with an early evolutionary divergence from rodent-borne hantaviruses (Song et al., 2007a; Yadav et al., 2007). Moreover, long-ignored reports of serologic and antigenic evidence of hantavirus infection in shrews (Gavrilovskaya et al., 1983; Tkachenko et al., 1983) have been validated by the identification of genetically distinct hantaviruses in the Eurasian common shrew (Sorex araneus) from Switzerland (Song et al., 2007b), Chinese mole shrew (Anourosorex squamipes) from Vietnam (Song et al., 2007c), and northern short-tailed shrew (Blarina brevicauda), masked shrew (Sorex cinereus) and dusky shrew (Sorex monticolus) from the United States (Arai et al., 2007, 2008a). Hantavirus RNAs have also been detected in the Therese’s shrew (Crocidura theresae) from Guinea (Klempa et al., 2007), Ussuri white-toothed shrew (Crocidura lasiura) from Korea, vagrant shrew (Sorex vagrans), Trowbridge’s shrew (Sorex trowbridgii) and American water shrew (Sorex palustris) from the United States, and flat-skulled shrew (Sorex roboratus) and Laxmann’s shrew (Sorex caecutiens) from Russia (H.J. Kang, J.-W. Song and R. Yanagihara, unpublished observations).
Although their evolutionary origins remain obscure, the discovery of hantaviruses in shrews (Order Soricomorpha, Family Soricidae) from widely separated geographic regions, spanning four continents, challenges the long-held view that rodents are the principal and primordial reservoir hosts. In addition, these findings raise the possibility that other soricomorphs, particularly moles (Family Talpidae), may harbor hantaviruses. Recent identification of a novel hantavirus, designated Asama virus (ASAV), in the Japanese shrew mole (Urotrichus talpoides) (Arai et al., 2008b) prompted us to intensively analyze tissues from the American shrew mole (Neurotrichus gibbsii). We now report on a phylogenetically distinct hantavirus, called Oxbow virus (OXBV), which provides additional support for the emerging concept that ancestral soricomorphs, rather than rodents, may have served as the original mammalian hosts of hantaviruses. Moreover, cross-species transmission events among soricomorphs on separate continents probably influenced the early evolution of hantaviruses.
Novel hantavirus RNAs were detected by RT-PCR in lung, kidney, heart, salivary gland, adrenal gland, stomach, large intestine and terminal colon of an American shrew mole, captured in Oxbow Regional Park, in Gresham, Oregon, in September 2003 (Fig. 1). Amplified products from the above-listed tissues showed no sequence variation in a 370- and 450-nucleotide region of the OXBV S and L segments, respectively. The widespread tissue distribution of OXBV RNA was not unlike that found in hantavirus-infected rodent reservoir hosts. Also, the detection of OXBV RNA in salivary gland, terminal colon and kidney suggests modes of virus transmission similar to that observed in rodents (Lee et al., 1981; Yanagihara et al., 1985). Future longitudinal investigations, using a capture-recapture strategy, are warranted to clarify the intraspecies transmission dynamics of OXBV (Clay et al., 2009), including the level of hantavirus shedding in secretions and excretions, as well as the frequency, duration and heterogeneity of contacts among American shrew moles.
Earlier attempts to amplify hantavirus RNAs in the same tissues had been repeatedly unsuccessful, using oligonucleotide primers based on rodent-borne hantaviruses. Newly designed primers, based on the most highly conserved regions of recently acquired genomes of shrew-borne hantaviruses, were ultimately effective and allowed amplification of the full-length S- and M-genomic segments and partial L segment of OXBV.
Pair-wise alignment and comparison of the OXBV genome with representative rodent-borne hantaviruses showed low nucleotide sequence similarity in the S segment, ranging from 59.7% to 62.6% (Table 1). OXBV sequences were even more divergent from TPMV and Imjin virus (MJNV), two crocidurine shrew-borne hantaviruses: approximately 50% in the S and M segments and 60% in the L segment. In contrast, much higher nucleotide sequence similarity of approximately 70% to 73% was found in the S, M and L segments of OXBV and Jemez Springs virus (JMSV) harbored by the dusky shrew (Arai et al., 2008a), whereas ASAV was more closely related to Seewis virus (SWSV), a hantavirus recently identified in the Eurasian common shrew (Song et al., 2007b). OXBV and ASAV genomic sequences were distinct, differing by approximately 30% at the nucleotide level (Table 1).
The full-length 1,705-nucleotide S-genomic segment of OXBV contained a single open reading frame (ORF), encoding a predicted nucleocapsid (N) protein of 428 amino acids (nucleotide positions 39 to 1,325), and a 380-nucleotide 3′-noncoding region (NCR), which exhibited significant variability from previously identified hantaviruses. The hypothetical NSs ORF was absent, as in murine rodent- and soricine and crocidurine shrew-borne hantaviruses. As determined by various prediction software available in the NPS@ structure server (Combet et al., 2000), the overall secondary structure of the OXBV N protein resembled that of other hantaviruses. Specifically, the predicted secondary structure of the OXBV N protein was composed of 53.5% a-helices and 6.5% β-sheets, with the characteristic coiled-coil domain in the N-terminal region (residues 1 to 31 and 48–68) (Fig. 2A). β-pleated sheets were present at the presumed RNA-binding domain region (residues 175–217) (Xu et al., 2002).
The full-length 3,645-nucleotide M-genomic segment of OXBV encoded a predicted glycoprotein of 1,141 amino acids (nucleotide positions 42 to 3,467). In scanning the OXBV glycoprotein, the highly conserved WAASA amino acid motif cleavage site was found at positions 650 to 654. Also, the OXBV glycoprotein precursor had seven potential N-linked glycosylation sites (six in Gn at amino acid positions 138, 353, 405, 524, 617 and 623; and one in Gc at position 934), of which five are found among all hantavirus glycoproteins (Fig. 2B). Both the Gn and Gc of OXBV showed predicted transmembrane helices (Fig. 2C) serving as hydrophobic anchor domains at the C-terminal region, as determined by TMHMM, a program for predicting transmembrane helices based on a hidden Markov model (Krogh et al., 2001; Möller et al., 2001).
The partial 4,396-nucleotide L-genomic segment of OXBV encoded an incomplete RNA-dependent RNA polymerase of 1,452 amino acids (nucleotide positions 39 to 4,396). In the amino acid sequences of the RNA-dependent RNA polymerases of all rodent-borne hantaviruses, major conserved regions have been reported for the polymerase function, which are designated as motifs A, B, C, D, E and premotif A or F. Motifs A, B and D have conserved aspartate, glycine and lysine, respectively. In motif C are two conserved aspartic acid residues. The XDD motif is essential for catalytic activity and motif E contains the E(F/Y)XS site. Premotif A has a conserved lysine and two arginine residues (Kukkonen et al., 2005). The five major motifs for viral RNA polymerases were also present in OXBV.
The coding regions of the full-length S- and M- and partial L-segment sequences of OXBV were analyzed extensively for recombination using multiple recombination-detection methods, including GENECONV, Bootscan, Chimaera, 3SEQ, RDP, SiScan, MaxChi and HyPhy Single Recombinant Breakpoint (Pond et al., 2005). The vast majority of these programs failed to disclose any evidence of recombination. Although separate regions of potential recombination were found in a few instances, there was no consistency or concordance between the detection methods, calling into question the validity of the identified sequences or the biological significance of recombination versus general heterogeneity in sequence evolutionary rates.
Phylogenetic analysis of the OXBV genome, using the maximum-likelihood (ML) and Markov Chain Monte Carlo tree-sampling methods, with the GTR+I+Γ model of evolution, indicated high bootstrap support for close phylogenetic relationships between OXBV and ASAV with hantaviruses harbored bysoricine shrews in the New and Old World, respectively (Fig. 3), consistent with cross-species transmission, occurring independently in different geographic regions during the distant past and with subsequent host adaptation (Arai et al., 2008b). However, while host switching has been operative, such events alone do not adequately account for the co-existence and widespread geographic distribution of genetically distinct hantaviruses among host species in two divergent taxonomic Orders of small mammals. Importantly, the well-supported phylogenies of hantaviruses and their rodent and soricid reservoir hosts spanning across four continents suggest that hantaviruses have likely co-diverged with some of their hosts during part of their long evolutionary history (Fig. 3).
Despite their physical resemblance, American and Japanese shrew moles are evolutionarily distinct, diverging prior to the diversification of the Talpini and represent separate lineages (Sanchez-Villagra et al., 2006; Shinohara et al., 2004). Based on analysis of mtDNA and nuclear gene sequences, the split between Old World and New World Sorex has been estimated at 13.9 MYA (95% CI: 10.2–17.5 MYA) (Dubey et al., 2007). As supported by the fossil record, three independent lineages of Soricinae shrews colonized North America during the middle Miocene. Accordingly, OXBV would have emerged during this period, long after ASAV. That is, since ASAV is in a clade that includes hantaviruses identified in both Sorex (M-segment) and Anourosorex (S-segment) (Fig. 3), one must hypothesize that it is older (95% CI: 14.0–20.5 MYA), as it would predate the Old World-New World dichotomy in Sorex.
Particularly during irruptions in rodent reservoir populations, hantavirus spillover into sympatric and syntopic hosts are more likely to occur (Gavrilovskaya et al., 1983; Mills et al., 1998). Such repeated or recurring cross-species transmission events could then result in long-term virus-host adaptation. A frequently cited example of host switching during hantavirus evolution is Topografov virus, which is genetically very similar to Khabarovsk virus but is found in the Siberian lemming(Lemmus sibirica), instead of the reed vole (Microtus fortis) (Vapalahti et al., 1999) or Maximowicz’s vole (Microtus maximowiczii). However, while belonging to different genera, these rodent species are members of the same Family (Cricetidae) and Subfamily (Arvicolinae). Similarly, all other known examples of so-called host switching of hantaviruses have occurred across rodent species within the same Family or Subfamily. By contrast, in the case of OXBV, reported here, and ASAV reported previously (Arai et al., 2008b), their phylogenetic positions are consistent with cross-species virus transmission involving mammalian hosts of two separate and distinct families (Talpidae and Soricidae) within the Order Soricomorpha. These findings would seem to strengthen the conjecture that soricomorphs may have been the primordial mammalian hosts of ancestral hantaviruses. In this regard, to what extent greater flexibility or non-specificity in host selection was a primitive characteristic that was later reduced or lost in hantaviral evolution is unknown.
Phylogenetic trees were reconstructed for co-phylogeny mapping from a virus tree into a host tree in TreeMap 2.0β (Charleston and Page, 1998) in order to ascertain the co-evolutionary history of hantaviruses and their hosts. Similar topologies with high bootstrap support for each genomic segment (at both the nucleotide and amino acid levels) were found for the segregation of hantaviruses, according to the Subfamily of their rodent and shrew reservoir hosts (Fig. 3). Specifically, hantaviruses and their rodent and soricid hosts showed congruent topologies in their phylogenetic relationships—with the exception of OXBV and ASAV. Overall, these data support the contention that hantaviruses have generally co-diverged with and adapted to their reservoir hosts over many millions of years (Monroe et al., 1999; Morzunov et al., 1998). That is, for part of their evolutionary history, hantaviruses have co-diverged with their hosts, namely in the divergence of shrews and their viruses from a still-unknown shared common ancestor, and to some extent during the subsequent diversification of shrews. For OXBV and ASAV, cross-species transmission of hantaviruses from hosts of one Family to another (Soricidae to Talpidae), with subsequent adaptation, appears to have occurred at separate times on different continents. Again, this observation suggests that the primordial or ancestral hantaviruses may have been comparatively more broad in their host selection than the more well-known host specificities of present-day rodent-borne hantaviruses. Intensive attempts to identify hantaviruses in other talpid species and sympatric soricid species, now underway, will help to clarify the evolutionary history of hantaviruses.
The largest incongruity between hantavirus and host co-divergence remains the divergent position of two hantaviruses identified in crocidurine shrews, namely TPMV in the Asian house shrew (Suncus murinus) (Carey et al., 1971) and MJNV in the Ussuri white-toothed shrew (Crocidura lasiura) (J.-W. Song and R. Yanagihara, unpublished observations). These viruses remain deeply divergent with respect to other hantaviruses, including the more recently identified hantaviruses reported here and elsewhere. Better resolution of the phylogeny of the soricid hosts is necessary, using multiple genetic loci. However, in unrooted trees based on each hantavirus genomic segment, the branch lengths leading to TPMV and MJNV were far longer than the other branches (data not shown), indicating greater evolutionary change and higher likelihood of older age. Thus, the deep divergences among crocidurine shrew-borne hantaviruses also suggest that soricomorphs were the original mammalian hosts. In this scenario, rodents would have acquired hantaviruses more recently from shrews (and moles).
Confirmation of the identity of the American shrew mole, in which OXBV was detected, was achieved by phylogenetic analysis of the 1,140-nucleotide mtDNA cytochrome b gene. Phylogenetic trees showed that rodents, shrews and moles segregated into three separate clades along two distinct lineages, one consisting of soricids and talpids, and the other of rodents (Fig. 3). Shrew-moles occupy similar ecological niches as shrews (Campbell and Hochachka, 2000) and share physical features of both soricids (small body size, posteriorly directed pelage, paired ampullary glands and long and pointed noses) and talpids (heavy insectivorous dentition, large head and enlarged forefeet). Importantly, the geographical ranges of the American and Japanese shrew moles do not overlap. The Japanese shrew mole is limited strictly to Japan, while the distribution of the American shrew mole (Subfamily Talpinae, Tribe Neurotrichini), the only member of the genus Neurotrichus and the smallest and most primitive mole in North America, is restricted to an area approximately the size of Japan in the western regions of North America (Fig. 1), extending from southwestern British Columbia in Canada to central California in the United States, where sympatric and syntopic shrew and rodent hosts of hantaviruses (such as the vagrant shrew and Trowbridge’s shrew and the deer mouse) are also found.
Shrew-moles are active at all hours throughout the day and night, with intermittent periods of rest and/or sleep (Dalquest and Orcutt, 1942). They construct networks of runways or underground burrows, generally at two levels, one being shallow and the other deeper. Being almost completely blind, they rely on their prehensile nose to locate prey, which consists largely of worms and insects. Unlike most shrew species that are solitary and reclusive, shrew-moles tend to be gregarious, often traveling in groups of more than 10 individuals. How these and other life history characteristics of the American shrew mole contribute to the transmission dynamics of OXBV remains open to future investigation.
In our opportunistic search for non-rodent-borne hantaviruses, we were fortunate in gaining access to frozen tissues from shrew moles trapped coincidentally, as part of an epizootiological study of Sin Nombre virus infection in deer mice (Peromyscus maniculatus) in Oregon (L. Dizney and L. Ruedas, unpublished observations). Since sera from shrew moles were not collected and because appropriate immunological reagents for detecting talpid-borne hantaviruses are unavailable, we relied on RT-PCR. Total RNA was extracted, using the PureLink Micro-to-Midi total RNA purification kit (Invitrogen, San Diego, CA, USA), from tissues obtained from 10 American shrew mole (Neurotrichus gibbsii), captured between July and October 2003 in Oregon: five from Oxbow Regional Park (45.4879°N, 122.2970°W), a 1,200-acre natural area park located within the Sandy River Gorge in Gresham (Fig. 1); one from Tryon Creek State Park (45.4337°N, 122.6690°W) in Multnomah County; and four from Tualatin River National Wildlife Refuge (45.3957°N, 122.8305°W) in Washington County. cDNA was prepared at 65°C for 3 min, 42°C for 20min, 50°C for 50 min and 70°C for 3 min, using the SuperScript III First-Strand Synthesis System (Invitrogen) and primer (OSM55:5′-TAGTAGTAGACTCC-3′) designed from the conserved 5′-end of the S, M and L segments of hantaviruses.
PCR was performed as described previously, with each 20-μL reaction containing 250 μM dNTP, 2 mM MgCl2, 1 U of AmpliTaq polymerase (Roche, Basel, Switzerland) and 0.25 μM of oligonucleotide primers, which were designed from highly conserved regions of the hantavirus genome. Considerable trial-and-error testing was required to devise suitable primers and refine cycling conditions. Eventually, the following nested PCR primers sets provided the initial sequences of the newfound hantavirus: S segment (OSM55, 5′-TAGTAGTAGACTCC-3′ and HTN-S6, 5′-AGCTCNGGATCCATNTCATC-3′; and Han-S604F, 5′-GCHGADGARHTN ACACCNGG-3′ and Han-S974R, 5′-TCNGGNGCHCHNGCAAANAHCCA-3′); M segment (OSM55, 5′-TAGTAGTAGACTCC-3′ and TM-2957R, 5′-GAACCCCADGCCC CNTCYAT -3′ and HTN-M1190F, 5′-GGNCCNGGDGCWNVHTGTGA-3′; and HTN-M2020R, 5′-CCATGDGCAKTRTCANTCCA-3′); and L segment (Han-L1880F, 5′-CARAAR ATGAARNTNTGTGC-3′ and Han-L3470R, 5′-TTRAACATNSNYTTCCACATHTC-3′; and Han-L2520F, 5′-ATNWGHYTDAARGGNATGTCNGG-3′ and Han-L2970R, 5′-CCNGGNGA CCAYTTNGTDGCATC-3′). Initial denaturation at 94°C for 5 min was followed by two cycles each of denaturation at 94°C for 40 sec, two-degree step-down annealing from 48°C to 38°C for 40 sec, and elongation at 72°C for 1 min, then 32 cycles of denaturation at 94°C for 40 sec, annealing at 42°C for 40 sec, and elongation at 72°C for 1 min, in a GeneAmp PCR 9700 thermal cycler (Perkin-Elmer, Waltham, MA). Amplicons were separated by electrophoresis on 1.5% agarose gels and purified using the QIAQuick Gel Extraction Kit (Qiagen, Hilden, Germany). DNA was sequenced directly using an ABI Prism 377XL Genetic Analyzer (Applied Biosystems, Foster City, CA).
To predict secondary structures of nucleocapsid protein and glycoprotein, the whole amino acid sequences were submitted to NPS@ structure server (Combet et al., 2000). Glycosylation and transmembrane sites were predicted at the NetNlyc 1.0 and Predictprotein (Gavel and von Heijne, 1990) and TMHMM version 2.0 (Krogh et al., 2001) respectively. The program COILS (Lupas et al., 1991) was used to scan of N protein for expect of coiled-colis region.
Newly acquired, full-length S- and M- and partial L-segment sequences of OXBV, amplified from an American shrew mole, were aligned and compared with publically available hantavirus sequences, using the ClustalW method, implemented in Lasergene program version 5 (DNASTAR, Inc., Madison, WI) (Thompson et al., 1994) and transAlign (for coding sequences) (Bininda-Emonds, 2005). Phylogenetic trees were estimated using the ML method implemented in PAUP* (Phylogenetic Analysis Using Parsimony, 4.0b10), RAxML Blackbox web-server (Stamatakis et al., 2008) and MrBayes 3.1 (Ronquist and Huelsenbeck, 2003). The optimal evolutionary model was estimated under the GTR+I+Γ model of evolution, as selected by using ModelTest v.3.7 (Posada and Crandall, 1998). Parameters were re-estimated during successive rounds of ML heuristic searches using the TBR and SPR algorithms implemented in PAUP*. Respective individual parameter estimates for S-, M- and L-segment sequence alignments were as follows for base frequencies: A (0.3406, 0.3446, 0.3729); C (0.1940, 0.1680, 0.1388); G (0.2269, 0.1877, 0.1861); T (0.2385, 0.2997, 0.3023); substitution rate matrices: A–C (3.3442, 4.0319, 9.0177); A–G (5.1156, 5.8287, 10.6997); A–T (2.4196, 1.8545, 1.9089); C–G (1.5578, 2.9970, 5.4793); C–T (7.7701, 9.6963, 29.0711); G–T (1.0000, 1.0000, 1.0000); proportions of invariable sites (0.2050, 0.1440, 0.2210); among-site rate heterogeneity gamma distribution shape parameters (0.9840, 0.7660, 0.5560).
Topologies were evaluated by bootstrap analysis of 1,000 iterations, using neighbor-joining trees in PAUP*, posterior node probabilities based on 2 million generations and estimated sample sizes well over 100 (implemented in MrBayes) and 100 ML bootstrap replicates implemented in RAxML. With a robust phylogeny of shrew- and rodent-borne hantaviruses (as defined by bootstrap support of >70% or 0.70), we readdressed the co-evolutionary relationship between hantaviruses and their hosts that forms the basis of our predictive paradigm for hantavirus discovery. We employed host-parasite phylogenetic comparisons to detect co-divergence or host switch in TreeMap 2.0β (Charleston and Page, 1998).
To verify the identity of the hantavirus-infected mole and to study its phylogenetic relationship to other reservoir hosts, genomic DNA was extracted from frozen kidney tissue using the QIAamp DNA Mini Kit (Qiagen) according to the manufacturer’s instructions. The entire 1,140-nucleotide region of the cytochrome b gene of mtDNA was amplified by PCR, using well-tested primers (forward: 5′-CGAAGCTTGATATGAAAAACCATCGTTG-3′; and reverse: 5′-CTGGTTTACAAGACCAGAGTAAT-3′). PCR was performed in 50-μL reaction mixtures, containing 200 μM dNTP and 1 U of AmpliTaq polymerase (Roche, Basel, Switzerland). Cycling conditions consisted of 8 cycles at 94°C for 40 sec, 55°C for 40 sec and 72°C for 1 min, followed by 30 cycles at 94°C for 40 sec, 50°C for 40 sec and 72°C for 1 min, and 1 cycle at 72°C for 7 min. Amplified DNA was purified and then submitted for automated fluorescent sequencing. Host phylogenies based on mtDNA cytochrome b sequences were generated, using ML and Bayesian methods.
We thank Metro for providing access to Oxbow Regional Park. Permits for mammal trapping were obtained from the Oregon Department of Fish and Wildlife. This research was supported in part by U.S. Public Health Service grants R01AI075057 from the National Institute of Allergy and Infectious Diseases, and P20RR018727 (Centers of Biomedical Research Excellence) and G12RR003061 (Research Centers in Minority Institutions) from the National Center for Research Resources, National Institutes of Health, as well as grants from the American Society of Mammalogists, Sigma Delta Epsilon-Graduate Women in Science, Northwest Health Foundation and Forbes-Lea Fund at Portland State University.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.