Features of Consistent Powassan Virus Lineage II Focus in Southern Maine, United States

ABSTRACT.

Powassan virus lineage II or deer tick virus (DTV) is a rare but increasingly reported human infection in the United States transmitted by Ixodes scapularis ticks. The virus is thought to be maintained in environmental foci that are optimal for tick and vertebrate reservoirs, but details on DTV ecology are poorly understood. We investigated DTV tick infection rates and reservoir host abundance in a focus of consistent DTV activity in Maine, USA. Host and tick abundance, vegetation, and microclimate conditions were measured in three forest sites representing increasing invasive understory infestation. Sites were selected representing native understory, mixed vegetation with some invasive Japanese barberry (Berberis thunbergii), and a highly invasive site dominated by Japanese barberry. Japanese barberry in the mixed vegetation site averaged 1 m in height with space between plants, whereas the highly invasive site had impenetrable Japanese barberry over 1.5 m. The DTV infection rate was greater in the highly invasive site. Density of I. scapularis ticks were significantly lower in the native forest site, and no DTV was found. Another feature of the DTV focus was more stable humid microclimate throughout the year compared with the other sites and a nearby continuous section of forest, consistent with reports from Connecticut, USA. We conclude that invasive Japanese barberry stands provide favorable and consistent microclimate conditions to maintain high DTV infection rates annually among questing I. scapularis ticks. Understanding environmental and landscape features that support high infection rates could lead to the identification of high-risk habitats for contracting DTV.

INTRODUCTION

Powassan virus (POWV) is a member of the tick-borne encephalitis complex of flaviviruses causing severe illness and death in up to 10% to 15% of human cases in the United States.¹ Long-term sequelae including cognitive deficit, muscle weakness, and ventilator dependence may occur in up to 50% of POWV survivors.¹ Two distinct lineages of POWV are recognized: lineage I (prototype) and lineage II, known as deer tick virus (DTV).^2,3 Although POWV lineage I is associated with Ixodes cookei and their common hosts including woodchucks (Marmota monax), limited human exposure to this tick has been documented.^4,5 DTV poses a greater threat to human health because it can be transmitted by Ixodes scapularis, a widespread, generalist feeder that frequently attaches to and feeds on humans in the northeast region.^3,6,7

Important DTV reservoirs have not been determined, and more research is required to identify amplifying hosts in nature.⁸ DTV may circulate among populations of white-footed mice (Peromyscus leucopus), which comprise a large portion of the host population for subadult I. scapularis.⁹ Researchers collecting serum from white-footed mice collected in Rhode Island; Wisconsin; and Nantucket Island, Massachusetts found that 3% to 4% showed prior exposure to DTV²; however, this serological result should be considered with caution due to issues of cross reactivity with other flaviviruses.^10,11 Although the presence of DTV in questing I. scapularis ticks implicates white-footed mice as a potential reservoir host, virus isolation has not been demonstrated from wild caught white-footed mice.^3,12 Recent blood meal analysis of DTV-infected questing ticks suggests shrews (Family Soricidae) as a potential reservoir.¹³

Ixodes scapularis is the vector responsible for transmission of Borrelia burgdorferi s.l., the causative agent for Lyme disease, along with numerous other human pathogens.¹⁴ Unlike B. burgdorferi, which is prevalent throughout the range of the vector, DTV remains focal.¹⁵ This suggests that the transmission patterns for DTV may differ from the transmission patterns for other pathogens transmitted by I. scapularis. Focal and persistently DTV-positive locations may be found during routine pathogen surveillance, but the factors that lead to this persistence are not understood. In 2016 and 2017, a study was conducted to determine the prevalence of DTV in I. scapularis collected from different counties in Maine.¹⁶ Those results showed DTV infection rates in I. scapularis nymphs and adults between 0% and 5%, consistent with rates in neighboring states.^2,6,17–19 However, many of the DTV-positive ticks evaluated in that study were collected from a single site in the Wells National Estuarine Research Reserve (WNERR).

To investigate factors leading to DTV focality further, we conducted a study to determine environmental conditions that sustain virus persistence over time. Specifically, we tested the hypothesis that DTV is maintained in small microhabitats (foci) where environmental factors such as invasive shrub understory and abundant tick hosts sustain virus persistence over time. We systematically collected ticks and measured microclimate and vegetation features from three forested habitats representing a gradient of invasive plant species in the understory. Camera traps were deployed, and small mammals were trapped, their tick burden measured, and tick samples were tested for DTV. We detected the highest infection rates (5.67%) among questing I. scapularis adult ticks from the highly invasive forest habitat. This location contained abundant Japanese barberry (Berberis thunbergia) in the understory with stable humidity and other microclimate conditions that may enable maintenance of DTV. We conclude that invasive Japanese barberry stands provide ideal and consistent microclimate conditions to maintain high DTV infection rates annually among questing I. scapularis ticks.

MATERIALS AND METHODS

Study site.

The study was conducted at WNERR (43.339349°N, 70.551008°W), which is located on the Maine coast with an estuary that runs through the reserve. Average summer temperatures reached 27°C in July and August, with evening lows averaging 15.5°C.²⁰ Average precipitation at the preserve was 10 cm per month (National Weather Service averages for York County). Dominant mature forest tree species at the site included red oak (Quercus rubra L.), red maple, (Acer rubrum L.), yellow birch (Betula lutea Michx.), and white pine (Pinus strobus L.).²¹

Locations with native understory shrub species included high bush blueberry (Vaccinium corymbosum), bayberry (Myrica spp.), and huckleberry (Gaylussacia baccata). Typically, where the forest shrub layer was native, shrub and ground vegetation density were sparse with a layer of deciduous or deciduous/coniferous leaf litter. However, in some portions of WNERR, invasive shrub species, such as Japanese barberry (B. thunbergii), Eurasian honeysuckle (Lonicera spp.), and Asiatic bittersweet (Celastrus orbiculatus), thrived in a dense layer beneath the tree canopy. Prior research at Laudholm Farm (Wells, ME) demonstrated that forest stands with Japanese barberry had a higher density of I. scapularis ticks^21,22 and higher prevalence of DTV than forest stands with native understory shrub species (R. M. Robich, personal communication).

Sampling grids.

In 2020, one grid was established in each of three forest stand habitats: highly invasive understory, invasive understory, and native understory (Figure 1). Location of the grids was based on transect sampling in 2018–2019, which yielded more DTV-positive ticks from transects that ran through invasive understory than native understory.¹⁶ Each grid was marked with flags at 10-m intervals along each transect. Flags were designated according to a letter/number system (e.g., A1, G7). The native grid served as a reference grid. The highly invasive and native habitats were 70 m × 70 m grids, each containing 49 plots (10 m × 10 m). Due to space constraints, the invasive grid was 70 m × 60 m, containing 42 plots (10 m × 10 m).

Figure 1.

Tick sampling plots representing a gradient of native to highly invasive understory. (A) Native grid: understory consists of lowbush blueberries and saplings. (B) Invasive grid: understory consists of knee-high Japanese barberry and ferns, medium sized fallen branches for reference. (C) Highly invasive grid: understory consists of tall impenetrable Japanese barberry, human of average height in the center for height reference. (D) Tick sampling grid locations used in this study within the Wells National Estuarine Research Reserve (WNERR) in Wells, ME. Each habitat is labeled. Blue dots represent data logger locations.

Plant survey and site description.

A survey was conducted on August 29, 2020 to identify and characterize tree and shrub species in each plot of all three grids. Observations were made from each flag looking inward towards the northwest corner to the corresponding 10 m × 10 m plot. The average height of each shrub species was measured in each grid and checked for dominance (i.e., >70% of vegetation present in the space). Height of Japanese barberry was considered in one of four categories: <0.5, 0.5 to 1.0, 1.0 to 1.5, and >1.5. Leaf litter at each flag were recorded as deciduous, coniferous, mixed, or bare. Duff layer depth and soil moisture 2 to 3 cm below the surface was measured (Lincoln Soil Moisture Meter, A.M. Leonard, Piqua, OH). Average percent canopy cover was measured from tree canopy images taken from a height of 109 cm for each cardinal direction using Fiji ImageJ software (Rasband, W.S., ImageJ, U.S. National Institutes of Health, Bethesda, MD).²³

Microclimate conditions.

Hourly temperature and humidity data were collected on Hobo data loggers (Onset Pro v2 #U23-001, Cape Cod, MA) placed in four locations within each grid at 0 and 1 m. Loggers were mounted horizontally on trees or on wooden stakes if no tree was available. A data logger was placed at the center of each grid (flag D4) and at three other locations based on the most representative vegetation (Figure 1A). The highly invasive grid loggers were placed at D4, B7, C5 and between plots C6 and C7. The invasive grid loggers were at D4, D1, A4, and F3 and the native grid loggers were at D4, B2, C6, and F4. Loggers recorded data from June 20, 2020 through January 7, 2021. Saturation deficit was calculated from temperature and relative humidity.²⁴

Questing tick survey.

Each grid was sampled once per week, weather permitting, and all transects were sampled between 0800 and 1500 hours. Tick collection was conducted around marker flags within the grid rather than along transects due to the impenetrability of the dominant Japanese barberry. Host-seeking I. scapularis ticks were collected using a 1-m² white corduroy flag that was checked at 30-second intervals.²² Ticks were stored alive at 4°C and identified according to published keys.^25–27 All ticks were frozen at –80°C individually in 2-mL sterile micro centrifuge tubes.

Small mammal abundance and feeding tick survey.

All animals were captured and handled according to animal use protocols approved by Maine Medical Center’s Institutional Animal Care and Use Committee (#1904). Rodents smaller than chipmunks (Tamias striatus) were not handled with anesthesia; larger mammals were released without handling. Small mammals were trapped in both invasive forest grids using Sherman traps (3310A H.B. Sherman Traps, Inc. Tallahassee, FL) baited with peanuts and apples. Traps were not set in the native forest grid due to the lack of protective understory and potential for heat stress. All traps were placed within 1 m of each flag, set between 1400 and 1600 hours in the evening and collected between 0700 and 0900 hours the next morning. Trapped animals were checked for ticks (pregnant animals were handled first) and then returned to their exact collection location immediately after processing. Live shrews were identified and then returned to their exact locations without being handled due to their fragility in Sherman traps and bite danger to humans. Trapping was completed on 2 consecutive trap nights every other week for a total of 784 trap nights in the highly invasive grid and 672 trap nights in the invasive grid. Traps were not set on days where the high forecasted temperature was >30°C. All mammals were processed for body length, weight, age, and sex. Ticks were removed with forceps by grasping the mouthparts and pulling straight out. Animals were tagged (#100-3; National Band and Tag, Newport, KY) in the ear and released within 5 m of where they were captured. Animals recaptured within 48 hours were eliminated from analysis.^28,29 The ratio of infested animals was calculated as the number of a species with one or more ticks divided by the total number of that species in each study site. Tick burden was calculated as the number of individual mammals divided by the number of ticks from each life stage removed.³⁰

Medium and large mammal survey.

Camera traps were used to determine the diversity of medium and large mammals in the grids.^31,32 Motion triggered camera traps (Moultrie M-880, #MCG-12691, Calera, AL) were placed 1 m above the ground in four locations of each grid facing a clearing or the margin of the salt marsh to capture animals moving into and out of the forest. Camera traps were set for motion trigger with 30-second detection delay and operated for 30 nights during the summer. Animals captured within 30 minutes of the prior image of that animal within the same grid were considered the same individual.

RNA isolation, polymerase chain reaction, and sequencing.

Individual adults, nymphs, and pools of up to 50 larval ticks were crushed using a sterile paperclip (5 cm size, ACCO Brands, Booneville, MS). Samples were then homogenized in 200 µL minimum essential medium (MEM) reagent by pulse vortexing with three stainless steel BBs. Aliquots (14 mL) of each individual homogenate were combined in pools of up to ten questing ticks or ticks from ten animals (140 mL). RNA was extracted using a QIAmp Viral RNA Mini Kit (Qiagen, Germantown MD) according to manufacturer’s protocol. Reverse transcription polymerase chain reaction was used to detect POWV complementary DNA (from either Powassan virus prototype or DTV).¹⁶ Positive samples were then submitted for full genome sequencing for confirmation and lineage typing.³³ Briefly, RNA was treated with heat-labile dsDNase (ArcticZymes, Tromsø, Norway) and converted to cDNA using Superscript III^TM (Invitrogen, Carlsbad, CA) with random primers. Libraries were tagmented, amplified using Nextera^® XT (Illumina, San Diego, CA), and sequenced on an Illumina platform with 150 bp paired-end reads.

STATISTICAL ANALYSES

Data analyses were conducted in R³⁴ using the following packages: ggplot2,³⁵ lubridate,³⁶ dplyr³⁷ for data management, emmeans³⁸ for post hoc comparisons, and brglm2.³⁹ The primary sampling unit was the plot. Ticks collected per hour was summarized across the entire study period (June 13–November 5) by species (I. scapularis, D. variabilis), life stage (nymph/adult), and plot within habitat type (grid). Questing I. scapularis nymphal and adult abundance was calculated for 2-week intervals for each plot for their respective peak season (nymphs June–July, and adults October–November). A Poisson regression model was used to make pairwise comparisons in questing nymphal and adult I. scapularis abundance by habitat (grid); the model included a categorical fixed effect of 2-week interval and used an offset of the log of the minutes of sampling time to adjust for slight variation in sampling effort.

A logistic model was used to compare the infection rate of adult ticks during the fall adult peak (October–November) between the highly invasive grid and invasive grid. Because none of the 19 adult ticks collected from the native forest habitat grid were positive for DTV, we used bias-reduced logistic regression via the brglm2 package. Small mammal abundance was determined as minimum number alive per 100 trap nights by grid. To determine the diversity and variation of mammal populations collected from each grid, the overall proportion of mammal species caught by grid were compared using a χ² test to the ratios of chipmunks, white-footed mice, and red-backed voles (Myodes gapperi).

Analysis of variance was used to compare the mean effect of grid for several vegetation and abiotic factors including height of invasive barberry, percent canopy cover, variation in saturation deficit, and dominance of invasive species. Significance was determined at the α = 0.05 level. A linear model was used to compare the percent canopy cover at the grid level with each individual flag as a replicate measurement. Post hoc pairwise tests with a Tukey correction were used to compare differences by grid.

Saturation deficit was calculated from temperature and relative humidity as a single measure of the drying power of the atmosphere.²⁴ The data were divided into nymphal peak (June 17, 2020 to August 15, 2020) and adult peak (October 1, 2020 to November 15, 2020) time periods. A linear model was used to compare the mean saturation deficit at the ground level in the center of each grid with date as a covariate. Post hoc pairwise tests with a Tukey correction were used to compare differences by grid. Flags in the salt marsh with no canopy cover were not included in the analysis.

RESULTS

Questing ticks.

Overall, we collected 474 I. scapularis and 16 D. variabilis ticks from June 13 to November 5, 2020 by flagging technique (Figure 2). This included 20 I. scapularis larvae, 77 nymphs, 174 females, and 203 males, as well as two D. variabilis nymphs, eight females and six males. The abundance of questing I. scapularis nymphs and adults in the native grid was significantly lower than in either the highly invasive or invasive grids, nymphs (Z = 4.137, P = 0.0001, and Z = 4.437, P <0.0001 respectively) adults (Z = 8.517, P <0.0001, and Z = 7.487, P <0.0001, respectively). The abundance of questing I. scapularis nymphs and adults was not different between the highly invasive and the invasive grids. Questing adult ticks in the highly invasive grid had higher rates of DTV (5.67%, n = 13) than ticks from the invasive grid (1.01%, n = 1), (Z = 2.035, P = 0.0419) (Figure 3). Data for all positive ticks can be found in Table 1.

Figure 2.

Questing I. scapularis ticks per hour collected by flagging from 13 June to 5 November 2020. (n = 174 female, 203 male, 77 nymph, 20 larvae).

Figure 3.

Percent of adult questing I. scapularis ticks positive with DTV RNA collected in plots with different levels of invasive understory at WNERR. Highly invasive (5.67%, n = 213) - Invasive (1.01%, n = 111) Z = 2.035, P = 0.0419.

Table 1.

Location and sex of infected adult I. scapularis ticks in two habitats at the Wells National Research Reserve, Wells, Maine

Tick ID	Grid	Flag	Sex
299P1	Highly invasive	D6	F
299P2	Highly invasive	D6	F
299P3	Highly invasive	D6	M
299S1	Highly invasive	E6	F
338L1	Highly invasive	C3	F
338L2	Highly invasive	C3	F
339M1	Highly invasive	C5	F
338S2	Highly invasive	E4	F
338BB2	Highly invasive	G6	F
422F1	Highly invasive	B2	F
422F3	Highly invasive	B2	F
422I1	Highly invasive	B5	F
422Q2	Highly invasive	E3	M
337X5	Invasive	G5	M

F = female; M = male.

Small mammal abundance, and tick infestation rate and burden.

Small mammal host abundance differed between the highly invasive grid and invasive grids. Significantly more chipmunks were captured in the invasive grid (χ² = 21.9, df = 1, P <0.0001). No differences were found for vole and white-footed mouse captures between grids (χ² = 0.15639, df = 1, P = 0.69 and χ² = 2.4686, df = 1, P = 0.1161, respectively). Deer tick virus was not detected in any of the 711 larvae collected from 185 individual mammals or the 73 nymphs collected from 50 individual animals. Attached larval and nymphal infestation rates as well as attached tick burden by grid and season are provided in Tables 2 and 3.

Table 2.

Species of mammals captured between June 10 and September 16, 2020 in the highly invasive trapping grid at Wells National Estuarine Research Reserve

Species	No. Captured	Ixodes scapularis Larva Burden	Ixodes scapularis Nymph Burden	Percent Infested*	No. DTV Positive Pools
Short-tailed shrew, Blarina brevicauda	7	0	0	0%	0
Southern flying squirrel, Glaucomys volans	1	—	—	—	0
Meadow vole, Mictrotus pennsylvanicus	1	10.00	0	100%	0
Short-tailed weasel, Mustela erminea	1	2.00	0	100%	0
Long-tailed weasel, Mustela frenata	1	—	—	—	0
Red backed vole, Myodes gapperi	40	0.20	0.18	22.5%	0
White-footed mouse, Peromyscus leucopus	125	1.808	0.09	53.6%	0
Norway rat, Rattus norvegicus	1	0	0	0%	0
Masked shrew, Sorex cinereus	3	0	0	0%	0
Red squirrel, Tamiasciurus hudsonicus	1	3	1	100%	0
Chipmunk, Tamias striatus	2	9	1.50	100%	0

DTV = deer tick virus.

^*Percent of animals found to have ≥1 tick feeding on them at the time of capture.

Table 3.

Species of mammals captured between June 10 and September 16, 2020 in the invasive trapping grid at Wells National Estuarine Research Reserve

Species	No. Captured	Ixodes scapularis Larvae Burden	Ixodes scapularis Nymph Burden	Percent Infested*	No. DTV Positive Pools
Short-tailed shrew, Blarina brevicauda	4	0.50	0	25%	0
Southern flying squirrel, Glaucomys volans	3	0	0.67	33%	0
Meadow vole, Mictrotus pennsylvanicus	0	—	—	—	0
Short-tailed weasel, Mustela erminea	1	1	1	100%	0
Long-tailed weasel, Mustela frenata	0	—	—	—	0
Red backed vole, Myodes gapperi	40	0.28	0.03	25%	0
White-footed mouse, Peromyscus leucopus	140	3.11	0.16	75.8%	0
Norway rat, Rattus norvegicus	0	—	—	—	0
Masked shrew, Sorex cinereus	1	0	0	0%	0
Red squirrel, Tamiasciurus hudsonicus	1	0	0	0%	0
Chipmunk, Tamias striatus	32	0.88	0.84	62.5%	0

DTV = deer tick virus.

^*Percent of animals found to have ≥1 tick feeding on them at the time of capture.

Camera traps.

White-tailed deer (Odocoileus virginianus) were captured on camera in all three grids and moving in and out of the highly invasive grid via the salt marsh. Deer were captured once in the native grid, three times in the invasive grid, and 34 times in the highly invasive grid. These numbers were small and not adequate to compare in our study. Coyotes (Canis latrans) were captured once in the native grid and once in the highly invasive grid. For this reason, deer and coyotes were not included in the final analysis.

Microclimate and forest structure.

Mean saturation deficit (Figure 4) varied between the grids during the nymphal peak with the native grid being the highest (model predicted $\overline{x}$ = 2.1, SE = 0.056) followed by the invasive grid (model predicted $\overline{x}$ = 1.5, SE = 0.056) and the highly invasive grid had the lowest (model predicted $\overline{x}$ = 0.8, SE = 0.057). All grids were significantly different from each other (P <0.0001 for all) during the adult peak. The mean saturation deficit of the native grid (model predicted $\overline{x}$ = 0.8, SE = 0.043) was significantly higher than the highly invasive grid ($\overline{x}$ = 0.6, SE = 0.043; t₈₈ = –3.316, P = 0.0038). The invasive grid did not vary from the other two ($\overline{x}$ = 0.7, SE = 0.043). Multiple linear regression was used to test if canopy cover and duff layer significantly predicted soil moisture. The overall regression was statistically significant: R² = 0.197, F(137 df) = 17.99, P = 1.153 e-7. It was found that canopy cover was inversely related to soil moisture (β = –8.13, P = 0.00022); and duff layer depth significantly predicted soil moisture (β = 0.53, P = 0.0056). Percent canopy cover differed among grids (Figure 5) (P <0.001 for all) with the native grid having the highest, most consistent cover (model predicted $\overline{x}$ = 72.7%, SE = 1.49), the invasive grid in the middle (model predicted $\overline{x}$ = 64.1%, SE = 1.61), and the highly invasive grid with the lowest percent canopy cover (model predicted $\overline{x}$ = 51.6%, SE = 1.49).

Figure 4.

Mean saturation deficit (a measure of environmental moisture) at the center flag 0m (D4) ground level of all three sampling grids. (1) Daily average shown over time throughout the nymphal peak (6/17/2020 – 8/15/2020). (2) Mean daily saturation deficit during the nymphal peak with the native grid having the highest (model predicted $\overline{x}$ = 2.1, SE = 0.056), followed by the invasive grid (model predicted $\overline{x}$ = 1.5, SE = 0.056), and the highly invasive grid having the lowest (model predicted $\overline{x}$ = 0.8, SE = 0.057). All grids were different from each other (P <0.0001 for all comparisons). (3) Daily average shown over time throughout the adult peak (10/01/2020 – (11/15/2020). (4) Mean daily saturation deficit during the adult peak, native grid (model predicted $\overline{x}$ = 0.8, SE = 0.043) was higher than the highly invasive grid ($\overline{x}$ = 0.6, SE = 0.043; t₈₈ = −3.316, P = 0.0038). The invasive grid did not significantly differ from the other two grids ($\overline{x}$ = 0.7, SE = 0.043).

Figure 5.

Percent canopy cover by grid. All comparisons P <0.001. Highly invasive grid (model predicted $\overline{x}$ = 51.6%, SE = 1.49). Invasive grid (model predicted $\overline{x}$ = 64.1%, SE = 1.61). Native grid (model predicted $\overline{x}$ = 72.7%, SE = 1.49).

DISCUSSION

Our results are consistent with the theory of focality of vector-borne infection in which zoonotic agents are maintained through an ideal assemblage of pathogen, hosts, and habitat.⁴⁰ We specifically tested the hypothesis that DTV is maintained in small microhabitats (foci) where environmental factors such as invasive shrub understory and abundant tick hosts sustain virus persistence over time. To test our hypothesis, we collected data on tick and host species abundance, vegetation characteristics, and microclimate. We found higher DTV infection rates in the forest grid highly invaded by Japanese barberry compared with forest grids with less dominant Japanese barberry or with native understory vegetation alone. Higher density of I. scapularis was associated with both Japanese barberry invaded forest sites. Our investigation revealed a lack of variance in the population density of small mammal hosts, apart from chipmunks, which exhibited a lower abundance within the densely invaded shrub grid. Additionally, no DTV RNA was detected in the replete larvae recovered from those animals. Notably, the highly invasive area exhibited a diminished canopy density and a reduced saturation deficit.

DTV RNA was not detected in replete I. scapularis larvae collected from trapped animals. Thus, we concluded that none of the trapped animals were likely infectious with DTV at the time of capture, but we cannot infer past infection history. Testing ticks attached to hosts is not a sensitive method of detecting animal infection because DTV viremia is limited (3–10 days) and shorter than our sampling intervals of every 2 weeks. White-footed mice have been implicated as a potential reservoir for DTV because of their association with I. scapularis; however, infection in wild white-footed mice has never been detected.⁴¹ In a laboratory setting, white-footed mice do not show signs of illness and may restrict POWV replication.^3,42

Japanese barberry was the most dominant species in both the moderately and highly invasive shrub grids. We found significantly lower tick abundance in the native grid than either of the invasive grids, and the invasive grids did not differ significantly from each other for either nymphal or adult peak seasons. Our results were consistent with prior work showing that Japanese barberry supports a greater abundance of questing I. scapularis ticks.^21,22 In coastal Maine, nymphs were twice as abundant in Japanese barberry and adults were 2.8 to 16.3 times more abundant than in other native forest types.²² Management of Japanese barberry has been shown to reduce the B. burgdorferi infection prevalence and the abundance of I. scapularis.^29,43 For example, Linkse et al. found that although the numbers of B. burgdorferi infected and uninfected white-footed mice were not affected by removal of Japanese barberry, the level of infestation by juvenile I. scapularis was higher on mice in areas where Japanese barberry was unmanaged.⁴⁴ There are many studies that investigate the effect of presence and absence of Japanese barberry on various aspects of tick-borne disease cycles; however, few have investigated multiple levels of infestation. In contrast to our results showing no differences in abundance of I. scapularis ticks or white-footed mice from invaded sites, D’Antonio⁴⁵ found more white-footed mice in a fully invaded site compared with a partially or noninvaded site and a higher prevalence of I. scapularis ticks between invaded sites.

We detected significant differences in saturation deficit over time between invasive and noninvasive grids. The DTV focality grid had higher moisture levels than the other two grids as evidenced by saturation deficit readings, especially during the nymphal and adult tick peaks. Japanese barberry can retain more constant temperature and humidity, potentially contributing to greater vector competence of I. scapularis. Temperatures higher than 27°C have been shown to negatively impact I. scapularis competence to transmit Borrelia^46,47 in a manner similar to mosquito virus transmission dynamics.^48–50 Humidity above 82% is considered optimal for I. scapularis survival and questing.^51–53 Future studies are warranted to investigate the modulating impact of invasive shrub microhabitats during weather extremes on I. scapularis viability and DTV infection rates.

In addition to the direct benefits of invasive shrub habitats on I. scapularis vitality and vector potential, dominant barberry may create an ideal habitat for amplifying hosts of I. scapularis–borne pathogens.⁵⁴ Barberry can produce an impenetrable barrier for larger host species, creating a situation in which nymphal and adult ticks are more likely to co-feed on small to medium-sized hosts.⁵⁵ In addition, an increase in humidity has been shown to increase the ratio of larvae to nymphs feeding on animals in arenas.²⁴

Recent blood meal analysis of questing I. scapularis nymphs in DTV endemic sites in Rhode Island and Massachusetts coastal islands revealed a 65% of DTV infected ticks fed on shrews.¹³ Unfortunately, our trapping method was not ideal for capturing shrews in the family Soricidae. Those we did trap were considered bycatch and were released quickly for the safety of the field team and the shrews. Alternative trapping methods, such as pit fall traps, may have yielded larger shrew sample sizes.^56,57 Small carnivores have a strong distaste for shrews, leaving birds of prey as their primary predators^58,59 which may not be able enter barberry. The potential significance of shrews as a reservoir requires further investigation.

For this study, we investigate one DTV focality during Summer 2020. In that year, the region experienced an abnormally hot, dry summer with low nymphal abundance.²⁰ No DTV-positive questing ticks were found in the native grid, and tick abundance was too low in this location to make conclusions about the potential infection level. Another limitation of our study was that we did not capture medium-sized mammals, such as striped skunks (Mephitis mephitis) and groundhogs (Marmota monax), although they may represent important tick hosts.^{4,18,28,60,61} The WNERR is an important refuge for the vulnerable New England cottontail (Sylvilagus transitionalis), so efforts to trap and process larger mammals would need to be undertaken with an abundance of caution.⁶² Future efforts to study medium-sized mammals could rely on baited camera trapping or burrow disruption to avoid unwanted impacts to the rehabilitation of cottontails.

Our research highlights key habitat features, abiotic conditions, and animal associations of a consistent DTV focus in Maine. We detected a strong association between invasive Japanese barberry, microclimate conditions, and high DTV infection rates among questing I. scapularis ticks. Our results suggest that DTV foci may be maintained by optimal microclimate conditions supported by invasive plant species. Understanding environmental and landscape features that support high infection rates could lead to the identification of high-risk habitats for contracting this emerging tick-borne virus.