After obtaining approval from the University of Hawaii Institutional Animal Care and Use Committee, we retrieved lung and liver tissues of 30 northern short-tailed shrews captured from several regions within the United States during 1994–1999 (Table 1) from deep-freeze storage at the University of New Mexico Museum of Southwestern Biology. Total RNA was extracted from shrew tissues by using the PureLink Micro-to-Midi Total RNA Purification Kit (Invitrogen, San Diego, CA, USA). cDNA was then prepared by using the SuperScript III First-Strand Synthesis System (Invitrogen) for reverse transcription–PCR (RT-PCR) with oligonucleotide primers designed from TPMV and other hantaviruses: medium (M) (outer: 5′-GGACCAGGTGCADCTTGTGAAGC-3′, 5′-GAACCCCADGCCCCNTCYAT-3′; inner: 5′-TAAVTTCAMCAACATGTCT-3′, 5′-CATGAYATCTCCAGGGTCHCC-3′) and large (L) (outer: 5′-CAGTCWACARTTGGTGCAAGTGG-3′, 5′-TCCATKATWGACATBGMRCCA-3′; inner: 5′-YTMATGTATGTTAGTGCAGATGC-3′, 5′-GRTTAAACATACTCTTCCACATCTC-3′). For confirmation, RNA extraction and RT-PCR were performed independently in a laboratory in which hantaviruses had never been handled. Amplicons were sequenced directly by using an ABI Prism 377XL Genetic Analyzer (Applied Biosystems, Foster City, CA, USA).

Of the 30 northern short-tailed shrews tested, hantavirus M-segment sequences were amplified from lung tissues of 3 of 12 animals captured in Camp Ripley (46.185°N, 94.4337°W), a 53,000-acre, state-owned military and civilian training center near Little Falls, in Morrison County, Minnesota, USA, in August 1998 (Table 1). Pairwise alignment and comparison of the 1,390-nt region (463 aa) spanning the Gn and Gc glycoprotein–encoding M segment indicated differences of 33.6%–41.9% and 32.7%–47.4% at the nucleotide and amino acid levels, respectively, from representative hantaviruses harbored by Murinae, Arvicolinae, Neotominae, and Sigmodontinae rodents (Table 2). No insertions or deletions were found in the regions sequenced compared with sequences of other hantaviruses. The new hantavirus, designated Camp Ripley virus (RPLV), showed sequence similarity of 98.1%–98.5% among the 3 strains.

Analysis of a 490-nt (163-aa) region of the L genomic segment amplified from 2 of the 3 shrews indicated that RPLV differed from rodentborne hantaviruses by 26.8%–34.1% at the nucleotide level and 20.7%–34.5% at the amino acid level (Table 2). RPLV differed from Tanganya virus (TGNV), a hantavirus detected recently in the Therese shrew (Crocidura theresae) in Guinea (7), by 27.3%–28.8% and 24.0%–25.0%, respectively. The higher degree of sequence similarity in the L segment between RPLV and other hantaviruses probably signifies the limits of functional preservation of the RNA-dependent RNA polymerase.

Repeated and exhaustive phylogenetic analyses based on nucleotide and deduced amino acid sequences of the M and L genomic segments generated by the maximum-likelihood method indicated that RPLV was distinct from rodentborne hantaviruses (with high bootstrap support based on 100 maximum likelihood replicates) (Figure). Similar topologies were consistently derived by using various algorithms and different taxa (including La Crosse virus) and combinations of taxa, which suggested an ancient evolutionary origin. However, definitive conclusions about the molecular phylogeny of RPLV and its relationship to TGNV and other soricidborne hantaviruses must await complete-genome sequence analyses.

As we had previously encountered in sequencing the entire genome of TPMV (J.-W. Song, R. Yanagihara, unpub. data), the divergent genome of RPLV presented challenges in designing suitable primers for RT-PCR. We were also constrained by the limited availability of tissues from the 3 infected shrews and the need to retain small portions of tissues for future virus isolation attempts. Consequently, we have been hitherto unable to obtain the full-length sequence of RPLV.

The northern short-tailed shrew (family Soricidae, subfamily Soricinae), 1 of 2 poisonous mammals in North America (8), inhabits forests and grasslands within the central and eastern half of the United States, extending north to Canada, west to Montana, and south to Tennessee and Georgia. Cytochrome b mitochondrial DNA and 16S rRNA sequence analyses support a monophyletic origin for the genus Blarina, with phylogeographic structuring of northern short-tailed shrews into well-defined groups to the east and west of the Mississippi River basin (9). Current studies will examine whether RPLV is harbored by the eastern haplogroup of northern short-tailed shrews and by the southern short-tailed shrew (Blarina carolinensis), a closely related species, which inhabits the southeastern United States, extending as far north as southern Illinois and south-central Virginia and as far south as central Florida.

Given the sympatric and synchronistic coexistence of northern short-tailed shrews with Neotominae and Arvicolinae rodents (such as Peromyscus leucopus and Microtus pennsylvanicus) and their ferocious territorial behavior, hantavirus spillover may be possible. Viruses closely related antigenically to Hantaan virus have been isolated from the Asian house shrew (Suncus murinus), greater white-toothed shrew (Crocidura russula), and Chinese mole shrew (Anourosorex squamipes) (10–12), which suggests that shrews are capable of serving as incidental hosts for hantaviruses typically harbored by rodents.

Shrews that harbor genetically distinct hantaviruses pose a compelling conceptual framework that challenges the long-accepted dogma that rodents are the sole reservoirs of hantaviruses. Viewed within the context of the recent detection of TGNV in the Therese shrew in Guinea, the identification of RPLV in the northern short-tailed shrew in the United States indicates that renewed efforts, facilitated by the rapidly expanding sequence database of shrew-borne hantaviruses, will lead to the discovery of additional hantaviruses in soricids throughout Eurasia, Africa, and the Americas. Our preliminary studies indicate 3 other novel soricid-borne hantaviruses in the Republic of Korea, Vietnam, and Switzerland. To establish if >1 of these newly identified hantaviruses is pathogenic for humans will require development of robust serologic assays (13) and application of other sensitive technologies, such as microarray analysis (14,15), for rapid detection of shrewborne hantavirus RNA in human tissues and bodily fluids.