Death Valley National Park is the largest national park in the contiguous 48 United States and is well known for having some of the hottest desert valleys in North America. Nearly 1 million persons visit the park each year; most visitation occurs during late autumn through mid-spring. Scotty’s Castle, located in the northeastern region of the park (37.031°N, 117.340°W, elevation 950 m), is a popular tourist attraction that has ≈100,000 visitors annually (US National Park Service, pers. comm., 2015). The attraction consists of an extensive 2-story historic villa and associated outbuildings, used as offices, residences, a visitor’s center, and storage, designed in Mission and Spanish Colonial Revival architecture. Below the villa lies a complex of service tunnels that are also included in tours. A year-round natural spring provides surface water flow for ≈0.3 km, creating an oasis in the desert.

We sampled rodents on the Scotty’s Castle grounds in March 2010, April 2011, and October 2011. During each sampling event, we placed 100 aluminum Sherman live-traps (HB Sherman Traps Inc., Tallahassee, FL, USA) throughout the compound, including the riparian zone, inside and around occupied and unoccupied buildings, and in the service tunnels below the villa. We baited traps with dry oats and peanuts, set late in the afternoon, and retrieved captured rodents the following morning. We anesthetized the rodents, then collected reproductive and morphometric data, and identified the rodents to species. We collected a minimum of 11 µL of blood from the retrobulbar sinus of each mouse for ELISA testing for SNV IgG (10); all Peromyscus spp. mice were humanely euthanized for SNV molecular testing. Blood samples and carcasses were analyzed at the California Department of Public Health Viral and Rickettsial Disease Laboratory (Richmond, CA, USA).

Figure

Figure. Phylogenetic tree of Gn glycoprotein sequences comparing hantaviruses sampled from 48 Peromyscus eremicus and 1 P. maniculatus (DEVA 10 022) mice collected in Death Valley National Park, California, USA (detail...

A total of 109 mice were captured during the 3 sampling events (300 trap nights): 100 (91.7%) P. eremicus, 5 (4.6%) P. maniculatus, and 4 (3.7%) P. crinitus. For P. eremicus mice, antibodies reactive to SNV antigen, PCR positive tissue, or both were detected in 13 (40.6%) of 32 collected in March 2010, 13 (32.5%) of 40 collected in April 2011, and 20 (71.4%) of 28 collected in October 2011. For P. maniculatus mice, antibodies reactive to SNV antigen, PCR positive tissue, or both were detected in 1 (100%) of 1 collected in March 2010, 0 (0%) of 1 collected in April 2011, and 2 (66.6%) of 3 collected in October 2011 (Table). None of the 4 P. crinitus mice tested positive for SNV antibodies or viral RNA. We compared viral RNA sequence (Gn) results for 48 P. eremicus and 1 P. maniculatus mice collected in each of the 3 collection periods to related hantaviruses and found a close consensus (>98%) to Convict Creek viruses 74 and 107 ([GenBank accession nos. L33474 and L33684) (11) (Figure).

The P. maniculatus deer mouse is recognized as the primary reservoir for SNV in the western United States. Published estimates for SNV seroprevalence are consistently higher in P. maniculatus mice than for any other Peromyscus species. Rodent hantavirus surveillance in California during 2001–2010 (7) detected the serum antibody to SNV among 14% (1,058/7,621) of deer mice statewide; concurrently, SNV seroprevalence for P. eremicus mice was significantly lower at 3.7% (102/2,723) and did not exceed this highest site-specific estimate at any individual surveillance site.

We identified serum antibodies to SNV in 40% of P. eremicus mice sampled. The consistently high seroprevalence over 3 sample periods suggests that SNV is efficiently transmitted and maintained within this population. The sequence characterization of viral RNA from seropositive P. eremicus mice further substantiates that the virus closely resembles type strains of pathogenic SNV associated with HPS.

The factors necessary to sustain a virus–reservoir relationship are both intrinsic and extrinsic. Hantaviruses are believed to have coevolved with their respective rodent hosts (1). P. eremicus mice are most closely phylogenetically related to P. maniculatus and P. leucopus mice, reservoirs for hantaviruses SNV and Monogahela virus, respectively, which are recognized to cause HPS (12). The close genetic similarity among these species may best enable P. eremicus mice among Peromyscus spp. mice in California to serve as a viable alternative host for SNV or to harbor a coevolved hantavirus of similar SNV lineage.

Hantaviruses are transmitted between rodent hosts through direct contact. Thus, a minimum population density is required to sustain transmission within an isolated group. The optimal habitat provided by the oasis setting of our study enables the typically solitary cactus mouse (13) to achieve a greater population density not generally found in most of the sylvan desert habitats where these mice are native (0.21–3.3/hectare) (14). Further studies in similar high density/optimal peridomestic habitats, which are associated with higher SNV infection prevalence (1), are needed to establish whether this high level of infection is reflective of the unique environment or an alternate phenotypic expression or strain, considering the enhanced transmission and maintenance of SNV in cactus mice detected at this location.

These findings underscore the importance of rodent exclusion and management in and around rural and semirural buildings where risk for contact between rodents and humans is high, even in the absence of P. maniculatus mice (9,15). Park leadership and staff were notified of our study results and given training on hantavirus awareness and prevention, and hantavirus pamphlets were made available for visitors to the park.