Hantavirus surveillance and genetic diversity targeting small mammals at Camp Humphreys, a US military installation and new expansion site, Republic of Korea

Small mammal surveillance was conducted (2008–2010, 2012) at Camp (Cp) Humphreys, a US Army installation and new expansion site, Republic of Korea (ROK), to identify hemorrhagic fever with renal syndrome health threats to US military/civilian populations during its ongoing expansion phase. Small mammals were collected using Sherman live capture traps and transported to Korea University where they were euthanized, tissues removed, and assayed to determine hantavirus IgG antibody-positive and hantavirus-positive rates by RT-PCR. A total of 2,364 small mammals were captured over 11,300 trap nights (capture rate = 20.92%). Apodemus agrarius was the most commonly collected (76.65%), with capture rates of 9.62% and 21.70% for Cp Humphreys and the expansion site, respectively. Overall, Hantaan virus (HTNV) IgG antibody-positive (Ab+) rate for A. agrarius was 2.15% (39/1,812). A total of 5.43% (10/184) Crocidura lasiura, 0.79% (2/254) Microtus fortis and 2.44% (1/41) Micromys minutus were serologically IgG Ab+ for hantaviruses. HTNV-specific RT-PCR demonstrated that 28.2% (11/39) HTNV Ab+ A. agrarius harbored the 328-nt sequence of the GC glycoprotein-encoding M segment of HTNV. Among them, the whole genome sequences of 3 HTNV strains were obtained by conventional RT-PCR and Rapid Amplification cDNA Ends PCR. Phylogenetic analyses of the HTNV strains from Cp Humphreys and the expansion site, Pyeongtaek, show a greater diversity of rodent-borne hantaviruses compared to HTNV previously identified in Gyeonggi province of the ROK. Thus, this study provides significant insights for raising HFRS threat awareness, analysis, and risk reduction strategies in southern Gyeonggi province.

Introduction Cp Humphreys is bounded by a small town (Anjeong-ri, Paengseong-eup), farmland, and the expansion site. It consists of an airstrip, asphalt roads, and structures for military operations and housing, streams, man-made water impoundments, and drainage ditches. Most areas bordering structures and roadside ditches were well maintained, while unmanaged grasses/herbaceous vegetation and shrubs border ponds, streams, and some drainage ditches, providing limited space, cover, and food for small mammals.
The expansion site under development is bounded by Cp Humphreys and the Anseongcheon River. During 2008, areas immediately adjacent to Cp Humphreys were under development for the construction of structures for military operations and housing, roads, drainage systems, and recreation areas. A large area not under development consisted of low lying hills of grasses/herbaceous vegetation, small groves of trees, major irrigation/drainage systems to reduce flooding, rice paddies, and limited dry-land farmland that lay fallow with unmanaged grasses/herbaceous vegetation. The major irrigation/drainage systems were bordered by moderate/tall grasses along the banks extending to roadways that harbored relatively high small mammal populations. Grading and filling of the unmanaged fallow lands continued and by 2010 much of the expansion site had been graded and cleared of grasses/herbaceous vegetation in preparation for land fill and construction. Habitats of small mammals (e.g., rodents and soricomorphs) and predators (e.g., weasels, raccoon dogs, and feral cats) were destroyed as lands were sequentially graded and filled, leaving these areas devoid of vegetation except for limited areas of grasses/herbaceous vegetation bordering ditches, major drainage systems for flood control, and temporary roads that provided habitat for small mammals.

Small mammal trapping
Trapping was conducted monthly from January 2008-December 2009, quarterly during 2010, and semiannually in 2012. Areas surveyed included: limited tall grasses/herbaceous vegetation that bordered streams, retaining ponds, fence lines, park perimeters, and open fields at Cp Humphreys and expansive tall grasses/herbaceous vegetation bordering dirt and paved roads, streams, ditches, water impoundments, flood control drainage systems, spillways, and the Anseongcheon River at the expansion site, which did not interfere with military or construction activities. Collapsible live-capture Sherman1 traps (7.7x9x23 cm; H.B. Sherman, Tallahassee, FL) were set in grasses/herbaceous vegetation (providing shade) at 4-5 m intervals (25-50 traps/trap line) in the late afternoon over a 2-3 day period and picked up early the following morning as previously described [4,[6][7][8][9][10]. Traps positive for small mammals were sequentially numbered, placed in a secure container, and transported to Korea University where they were euthanized, identified to species using morphological techniques, sexed, weighed and then, tissues (lung, liver, kidney, and spleen) collected and stored at -80˚C until used [14].

Real-time quantitative PCR (RT-qPCR)
cDNA was synthesized using a High Capacity RNA-to-cDNA kit (Applied Biosystems) by adding 1 μg of total RNA from lung tissues of HTNV-positive A. agrarius. RT-qPCR was conducted using SYBR Green PCR Master mix (Applied Biosystems) on a StepOne Real-Time PCR System (Applied Biosystems). The primer sequences included a forward primer; 5'-TT ATTGTGCTCTTCATGGTTGC-3' and a reverse primer; 5'-CATCCCCTAAGTGGAAGTTG TC-3' for HTNV S segment [22]. The cycling program was a cycle of 95˚C for 10 min, followed by 40 cycles of 15 s at 95˚C and 1 min at 60˚C.

Whole genome sequencing of HTNV
Total RNA was isolated from lung tissues of HTNV-positive A. agrarius using a Hybrid R Kit (GeneAll Biotechnology, Seoul, ROK). cDNA was synthesized using High Capacity RNA-to-cDNA kit (Applied Biosystems) with random hexamer or 5 0 -TAGTAGTAGACTCC-3 0 . Using Ex Taq DNA polymerase (TaKaRa BIO Inc., Shiga, Japan), first and second RT-PCR were performed at 95˚C for 4 min, followed by 6 cycles of denaturation at 94˚C for 30 sec, annealing at 37˚C for 30 sec, elongation at 72˚C for 1 min 30 sec, and then 32 cycles of denaturation at 94˚C for 30 sec, annealing at 42˚C for 30 sec, and elongation at 72˚C for 1 min 30 sec, and 72˚C for 5 min (ProFlex PCR System, Life Technology, CA, USA). To complete whole genome sequencing of HTNV, rapid amplification of cDNA ends (RACE) for 3' and 5' termini were performed using 3' and 5'-Full RACE Core Sets (Takara, Shiga, Japan) according to manufacturer specifications.

Genetic and phylogenetic analyses
Alignments of whole genome sequences of HTNV L, M, and S segments were facilitated using the Clustal W method (Lasergene program version 5, DNASTAR Inc. Madison, WI). The phylogenetic tree was generated by neighbor joining (NJ) and maximum likelihood (ML) methods (Molecular Evolutionary Genetics Analysis, 6.0). Genetic distances were computed, and topologies evaluated by bootstrap analysis of 1,000 iterations [23].

Capture rates for years and months
Annual capture rates for A. agrarius for both Cp Humphreys and the expansion site for years 2008-2009 were more than 4-fold and 3-fold higher, respectively, than for years 2010 and 2012, when a large portion of unmanaged vegetation was cleared at the expansion site ( Table 2,
Males accounted for 69.94% and 88.37% of those weighing 30-40 g and >40 g (old rodents), while females accounted for 66.30% and 54.21% of those weighing 10-20 g and 20-30 g (young rodents), respectively. On the average, male A. agrarius weighed 0.4-0.8 g more than females for each weight category, except for those weighing 20-30 g where females was heavier by 0.6 g more than the males.
Seasonal variations in population weights coincided with reproductive activity, e.g., increased proportions of lower weight individuals following high gravid rates ( Table 5).
The proportion of A. agrarius weighing 20 g declined from a high of 82.18% in January to a low of 10.37% by July before increasing to a high of 71.37% by December (Fig 3). In contrast, the proportion of A. agrarius weighing >30 g increased from a low of 1.98% in January to a high of 51.76% by September as a result of a maturing populations, gravid females, and abundant food supply, but rapidly declined to 1.76% by December that followed high numbers of gravid females during August (41.67%) and September (45.16%). Serological prevalence of HTNV at Cp Humphreys and the expansion site, Pyeongtaek Among 1,812 A. agrarius captured at Cp Humphreys and the expansion site, IFAT showed that 39 (2.15%) rodents were positive for anti-HTNV IgG. HTNV Ab+ rates among A. agrarius were significantly higher for Cp Humphreys (2.41%, range 0.0-7.89%) than for the expansion site (2.11%, range 0.0-3.51%) (χ 2 = 17.279, df = 1, P<0.001) (Fig 4). However, based on trap rates (Cp Humphreys, 9.62; expansion site, 21.70), the number of HTNV Ab+ A. agrarius captured/100 traps was 2-fold greater at the expansion site (0.46) compared to Cp Humphreys (0.23). The quarterly seasonal proportions of HTNV Ab+ A. agrarius for all weight categories at Cp Humphreys and the expansion site varied from 0.95% (Oct-Dec) to 2.86% (Jan-Mar) ( Table 6).
Overall, monthly HTNV Ab+ rates for males (2.25%) and females (2.07%) were similar, but varied monthly from 0.0-4.08% for males and 0.0-5.95% for females (Fig 5). The proportion of A. agrarius males and females that were serologically positive for HTNV for weight categories 10-20 g (1.52%), 20-30 g (2.33%), and 30-40 g (2.31%) were similar (Fig 6). However, the serological positivity (9.30%) of HTNV Ab for A. agrarius weighting > 40 g (oldest rodents) was only observed in males and was significantly higher than the other weighted groups (Oneway ANOVA test, p<0.0001).  Table 7). The partial sequences of HTNV M segment (328nt length) were trimmed and used for analysis. All HTNV strain sequences were submitted to GenBank (Accession numbers; KX119152-119162). The partial M segment sequences (coordinates 1,994 to 2,321) of 11 HTNV strains from Cp Humphreys and the expansion site were phylogenetically compared to HTNV strains previously identified in military training sites, northern Gyeonggi province (Fig 7). The nucleotide and amino acid homologies of the 11 HTNV strains from Cp Humphreys and the expansion site varied between 0-3.1% and 0-2.8%, respectively.

Quantitative RT-PCR, whole-genome sequencing, and phylogenetic analyses
To obtain whole genome sequences of HTNV in serological and molecular positive A. agrarius from Cp Humphreys and the extension region, Pyeongtaek, HTNV RNA copies were   The genetic diversity and phylogenetic relationship of HTNV in Cp Humphreys and the extension region were determined in comparison to strains obtained from lung tissue of seropositive rodents previously captured at a variety of HFRS-endemic areas, e.g. Twin Bridge Training Area (TBTA) North, TBTA South, and Dagmar North in Paju, Yeoncheon, and Pocheon (Fig 9). The L segment of the HTNV formed an independent outgroup of all of HTNV in Gyeonggi province. The phylogenetic analysis of the M segment showed a well-  supported genetic lineage with HTNV 76-118. The S segment formed a geographic-specific group within HTNV strains, including HTNV 76-118, in Gyeonggi province.

Discussion
Cp Humphreys was designated a major US military hub with an estimated final US military and civilian population of >20,000 personnel. To accommodate for the increased population and expansion for military operations and outdoor recreational areas, adjacent lands that consisted mostly of low-lying rice paddies were purchased. The resulting environmental modifications of purchased properties included sequential grading of terrain and filling of low-lying fallow rice paddies for the construction of roads, ditches, major drainage system/recreational areas, military housing, schools, hospital and medical clinics, and other structures designed for military operations Apodemus agrarius is associated with unmanaged lands characterized by abundant grasses/ herbaceous vegetation in rural areas, including military training sites [24][25][26][27]. Similar to this survey and other annual and multi-year surveys, A. agrarius was the most commonly collected small mammal at US and ROK operated military training sites and installation field environments [4,[6][7][8][9][10]. Compared to the expansion site of unmanaged grasses during 2008-2009, rodent populations were much higher than for Cp Humphreys where habitat was often limited to narrow strips of unmanaged vegetation along drainage systems/holding ponds. While capture rates of A. agrarius associated with the expansion site were high and movement of large trucks that created dusts on dirt, gravel, and hardened roads, the transmission risks of HTNV were reduced by very low HTNV Ab+ rates. Although no cases were reported among US military and civilian personnel, there may have been cases among local contractors and truck drivers that we were not aware of since these cases were not reported through the military medical system.
The Korea Centers for Control and Prevention (KCDC) [28] reports approximately 400-500 cases of HFRS annually, which are mainly caused by HTNV and SEOV. HTNV is the most common causative agent of HFRS in rural areas of the ROK and is characterized by severe medical manifestations and high mortality rate (9.46%) among US military personnel in Korea with good quality medical care from 1986-2014 [2,[4][5][6][7][8][9][10][11][12][13][14][15]. In Korea, human infections of HTNV among military members are usually associated with high populations of A. agrarius in field environments or mice-infested vacant buildings in combination with "dust-creating" activities (e.g., back-blast from artillery, convoy operations, and track and wheeled vehicle maneuvers/operations in field environments), while SEOV infections are usually associated with urban environments activities (e.g., dry sweeping or vacuuming rodent infested buildings) where R. norvegicus predominates [3,5]. While HFRS caused by HTNV infections poses a serious health threat in Korea, it is classified by the US National Medical Intelligence Center (NCMI) as a rare disease, frequently occurring is small clusters. The most recent cluster among US military personnel deployed to the ROK was observed in 2005, when three US soldiers acquired HFRS at TBTA associated with exposure of contaminated dusts in wheeled vehicle cabs (cavalry unit) [3]. During the same year, another HFRS case was acquired at Firing Point 60, Yeoncheon, associated with the back-blast of artillery. More recently (Nov., 2014) a single case was reported for a US soldier conducting convoy and driver's training at Dagmar North when HTNV seropositive rates in A. agrarius were 19.3% (considered a HFRS high-risk area), 27-30 days post-exposure that preceded infection [29]. The epidemiology of these cases was only accomplished through comparative analysis of the HTNV RNA from HFRS patients and associated rodents where the soldiers had trained, as the HTNV varies geographically [3].
Since 1986, only one case of SEOV has been reported from a US Airman that was vacuuming a rat-infested building and who had a relatively mild case of HFRS [5]. R. norvegicus, the primary reservoir for SEOV, is routinely captured by the Department of Public Works near housing and other facilities at Cp Humphreys. These resources would provide risk analyses for SEOV risks among US populations residing or working in buildings infested with rats. Similar to rodent-borne disease surveillance conducted at training sites near the DMZ and other US military installations, A. agrarius was the most frequently collected small mammal [4,[6][7][8][9][10]. Low to moderate A. agrarius capture rates were reported for limited tall grass habitats at Cp Humphreys and were similar to capture rates observed at other installations, e.g., Osan, Gunsan, and Gwangju Air Bases (unpublished data). During 2008-2009, high capture rates were observed for expansive tall grasses/herbaceous vegetation habitats at the expansion site and were similar to capture rates observed for expansive tall grass habitats at US and ROK operated training sites near the DMZ [4,[6][7][8][9][10]. However, trap rates were significantly lower during 2010 and 2012 following grading and removal of much of the vegetation from the landscape that provided food and harborage for small mammals. Additionally, in part, the decline may have been due to over predation as the predator populations (e.g., raccoon dogs, feral cats, weasels, and predatory birds) were pushed into space-limited habitats surrounded by farming activities and urban environments. Over time, predator populations will likely stabilize based on available food sources and small mammal populations may rebound to near previous levels for undisturbed areas.
HFRS risks are associated with a combination of factors, including: environmental, reservoir host bionomics, and types of human exposure. Overall, HTNV Ab+ rates for Cp Humphreys/expansion site and Osan Air Base (50 km south of Seoul and 20 km north of Pyeongtaek), are very low, usually 6%, when compared to US and ROK operated training areas near the DMZ where seasonal HTNV Ab+ rates varied up to 60% during monthly surveys and overall annual rates varied from 15% to 25% [4,7,9,10]. Limited surveys at Gunsan and Gwangju Air Bases, near the southern tip of the Korean Peninsula, were indicative of low small mammal populations, as well as none of the A. agrarius were HTNV Ab+ (TA Klein, personal communication). The reason for high HTNV Ab+ rates near the DMZ that decrease over distance to the tip of the peninsula is not understood, but may be related to reproductive behaviors. For training areas near the DMZ, there were observed low reproductive periods during the winter (0-0.3%), followed by a late spring increase in reproduction (4.2-24.6%) (Apr-May), low reproduction during the summer (0-1.3%), and very high reproduction in the late summer/early fall (27.3-70.0%) (Aug-Sep). A large influx of HTNV naïve mice observed at training areas near the DMZ during the fall/early winter periods when temperatures become cooler and habitat is shrinking as vegetation dies likely results in increased territorial disputes, wounding, and higher rates of HTNV transmission [30]. At Cp Humphreys and the expansion site, peak numbers of gravid females were observed earlier (June, 40.0%) and similarly in August and September (42.5 and 25.2%, respectively), while moderate numbers of gravid females were observed during the early spring (April/May, 16.0-16.2%), July (15.6%), and early winter (October/November, 9.4-11.8%) [4,[6][7][8][9][10]. In the southern area, young naïve A. agrarius broods throughout the summer may reduce territorial disputes in the fall due to relatively sufficient habitat and food. This proposed decreased movement and competition of naïve young rodents at Cp Humphreys and the expansion site may impact negatively on rodent-to-rodent HTNV transmission and result in lower HTNV Ab+ rates than those observed at the military training areas located near the DMZ, northern Gyeonggi province [31]. Additionally, the greatly reduced numbers of A. agrarius during 2010 likely reduced the potential for acute infections and corresponding viral shedding during the late fall/early winter when the majority of HFRS cases are reported. Although gravid females were observed throughout the early spring and summer, similar to training areas near the DMZ, the overall age (based on weight) of the population increased through September before rapidly declining as a result of the influx of young naïve rodents during the late fall reproductive cycle. By January, much of the population (based on weight) was replaced by young mice born during the late fall, indicating that the life span of A. agrarius live is approximately one year [24].
The overall HTNV seropositive rates of Cp Humphreys were higher than observed for the extensive tall grass habitats for undisturbed fallow rice paddies of the expansion area. However, the numbers of HTNV seropositive mice/100 traps were nearly 2-fold greater for the expansion site compared to Cp Humphreys, thereby increasing HFRS risks associated with less disturbed and unmanaged lands. The movement of potentially contaminated soil and vegetation and soil covered concrete roads created the potential for contaminated dusts and HTNV infections, especially for truck drivers and construction site monitors and workers. HTNV risks, while present at Cp Humphreys, are very low as a result of hard surface roads and recreation sites with short-cut grasses in the center, greatly reducing HTNV reservoir host habitats.
In this study, 39 (2.15%) of 1,812 A. agrarius were HTNV seropositive. The partial genome sequence of HTNV M segment was identified from 11 (28.21%) rodent lung tissues of HTNV Ab+ samples. RT-qPCR results showed varied viral loads in both sero-and molecular positive samples. The whole genome sequences of HTNV tripartite RNA were obtained from Aa08-1111, Aa09-189, and Aa09-198 that contained higher number of HTNV RNA copies. The termini of 3' and 5' sequences were determined by RACE PCR. Both end sequences of HTNV L, M, and S segments contained a mismatch at 9 th and the noncanonical U-G pair at 10 th nucleotides, suggesting the incomplete complementarity as previously described [32]. The total length of HTNV L segment for Cp Humphreys and the expansion site, Pyeongtaek, was three nucleotides shorter (6,530nt) than that of HTNV 76-118, demonstrating the deletion of 5'-AUC-3' at the 5' end of the L segment. U at the 12 th nucleotide on the M segment was defined compared to that on the HTNV 76-118 M segment.
The phylogenetic analyses of HTNV strains from Cp Humphreys and the extension site demonstrated a greater diversity of the rodent-borne hantavirus; the L segment showed distinct outgroup from entire HTNV strains, previously described in Gyeonggi province. The M segment formed a genetic cluster with HTNV 76-118, while the S segment was a geographic lineage within HTNV strains in Gyeonggi province. The natural reassortment and recombination of HTNV tripartite RNA genomes were observed near DMZ areas, northern Gyeonggi province [29,33]. Thus, the phylogenetic position and characterization of HTNV in Pyeongtaek will be clarified when additional genomic sequences of HTNV are acquired in southern areas of Korean peninsula.
A total of 10 (5.43%) C. lasiura were positive for MJNV, which was identified from shrews distributed in ROK and China, and the sera do not cross react with other rodent-borne hantaviruses [17]. Recently, there was a report that African shrew-borne hantaviruses were likely to infect humans [34]. Whether MJNV in C. lasiura poses a human health threat remains to be investigated. A total of 2/254 (0.79%) M. fortis and 1/41 (2.44%) M. minutus were serologically positive for hantaviruses, which was likely the result of interspecies transmission of HTNV since tissues were negative for hantaviruses by RT-PCR.
In summary, the characterization of US military installations undergoing expansion, in combination with small mammal surveillance, provides epidemiological information for the relative abundance of reservoir populations, hantavirus Ab+ rates, and other bionomic and environmental factors that are necessary to identify potential HTNV transmission risks. These transmission risks combined with human activities and exposure, which can be applied for disease risk analyses, are essential to the process of developing strategies for disease prevention. Comprehensive and long term rodent-borne disease surveillance should be the goal of US military preventive medicine to not only identify changes in HFRS disease risks due to modification of feral lands, but subsequently to better understand HFRS disease risks to soldiers, civilians, and family members residing and/or working on the installation. The whole genome sequences of HTNV at Cp Humphreys and the extension site show a greater diversity of rodent-borne hantaviruses in the ROK. Taken together, these data provide the robust impact to increase our knowledge of military activities, environmental conditions, and the genetic diversity of HTNV that can be applied to strategies to improve land management, disease risk mitigation, and the understanding of hantavirology.