Survey of Borreliae in ticks, canines, and white-tailed deer from Arkansas, U.S.A.

Background In the Eastern and Upper Midwestern regions of North America, Ixodes scapularis (L.) is the most abundant tick species encountered by humans and the primary vector of B. burgdorferi, whereas in the southeastern region Amblyomma americanum (Say) is the most abundant tick species encountered by humans but cannot transmit B. burgdorferi. Surveys of Borreliae in ticks have been conducted in the southeastern United States and often these surveys identify B. lonestari as the primary Borrelia species, surveys have not included Arkansas ticks, canines, or white-tailed deer and B. lonestari is not considered pathogenic. The objective of this study was to identify Borrelia species within Arkansas by screening ticks (n = 2123), canines (n = 173), and white-tailed deer (n = 228) to determine the identity and locations of Borreliae endemic to Arkansas using PCR amplification of the flagellin (flaB) gene. Methods Field collected ticks from canines and from hunter-killed white-tailed were identified to species and life stage. After which, ticks and their hosts were screened for the presence of Borrelia using PCR to amplify the flaB gene. A subset of the positive samples was confirmed with bidirectional sequencing. Results In total 53 (21.2%) white-tailed deer, ten (6%) canines, and 583 (27.5%) Ixodid ticks (252 Ixodes scapularis, 161 A. americanum, 88 Rhipicephalus sanguineus, 50 Amblyomma maculatum, 19 Dermacentor variabilis, and 13 unidentified Amblyomma species) produced a Borrelia flaB amplicon. Of the positive ticks, 324 (22.7%) were collected from canines (151 A. americanum, 78 R. sanguineus, 43 I. scapularis, 26 A. maculatum, 18 D. variabilis, and 8 Amblyomma species) and 259 (37.2%) were collected from white-tailed deer (209 I. scapularis, 24 A. maculatum, 10 A. americanum, 10 R. sanguineus, 1 D. variabilis, and 5 Amblyomma species). None of the larvae were PCR positive. A majority of the flaB amplicons were homologous with B. lonestari sequences: 281 of the 296 sequenced ticks, 3 canines, and 27 deer. Only 22 deer, 7 canines, and 15 tick flaB amplicons (12 I. scapularis, 2 A. maculatum, and 1 Amblyomma species) were homologous with B. burgdorferi sequences. Conclusions Data from this study identified multiple Borreliae genotypes in Arkansas ticks, canines and deer including B. burgdorferi and B. lonestari; however, B. lonestari was significantly more prevalent in the tick population than B. burgdorferi. Results from this study suggest that the majority of tick-borne diseases in Arkansas are not B. burgdorferi.


Background
In the Eastern and Upper Midwestern regions of North America, Ixodes scapularis (L.) is the most abundant tick species encountered by humans and the primary vector of Borrelia burgdorferi (causative agent of Lyme disease), whereas in the southeastern region Amblyomma americanum (Say) is the most abundant tick species encountered by humans but it cannot transmit B. burgdorferi [1,2]. Since its first description in the 1970s, Lyme disease is the most frequently reported vector-borne disease in the northern hemisphere [3]. However, in some areas of the world Lyme borreliosis may be caused by Borrelia genotypes other than B. burgdorferi and includes a range of symptoms and pathologies [3,4]. In the southern United States B. lonestari is associated with Southern Tick Associated Rash Illness (STARI) or Masters disease [5,6]. This bacteria is common throughout the southeast and has been identified primarily in A. americanum; however, the etiology of B. lonestari remains undetermined [7][8][9][10][11][12][13][14][15] and recent reports indicate B. lonestari may not be pathogenic [16].
Health professionals often question cases of Lyme disease from the southeastern United States because symptoms may be confused with other tick-borne illnesses and not all patients produce the erythema migrans or bull's eye rash used for diagnosis [15,17]. Additionally, these cases are rarely fatal, but can cause cardiac, neurological and joint problems [17]. However, the potential vectors (I. scapularis), pathogens (B. burgdorferi) and hosts (Peromyscus species) are all present in the south [17]. In Georgia, B. burgdorferi was isolated and characterized from field collected I. scapularis and cotton mice (Peromyscus gossypinus); and these field isolates were transmitted via I. scapularis to hamsters (Cricetidae) and mice (Muridae) [1]. Field collected I. scapularis from Alabama could acquire, maintain, and transmit B. burgdorferi [18] and field collected ticks and rodents from the southern U.S. were positive for B. burgdorferi [1,[19][20][21][22]. Previously in Arkansas, ticks and reservoir species were screened for Borrelia in nine northeast Arkansas counties using indirect fluorescence antibody (IFA), both I. scapularis and A. americanum were identified as potential vectors, and deer mice (Peromyscus spp.) and marsh rice rats (Oryzomys palustris Harlan) as potential reservoir hosts [23]. Although the environment is suitable for Borrelia transmission [17], many researchers and physicians do not believe the southern U.S. is at risk for Borrelia related diseases [2,4,15]. The exact cause for the reduced incidence for Lyme disease in the southern United States is unknown, but hypotheses include the abundance of other tick species in the area, the habitat, host dynamics, and tick genetics [2,24].
Previous research indicates that B. burgdorferi IFA positive canines geographically associated with I. scapularis infested deer, increase the likelihood of B. burgdorferi transmission to humans because the nymphs that feed on the deer may also feed on canines and their owners [25,26]. Recent Arkansas tick studies identified five tick species infesting canines and white-tailed deer with A. americanum infesting canines and I. scapularis primarily infesting white-tailed deer [27]. A population genetic study of I. scapularis based on the 16 S mtDNA identified both the American and Southern lineages (formerly known as I. scapularis and I. dammini) on canines and white-tailed deer [28], which is a concern because the American I. scapularis lineage is more likely than the Southern I. scapularis lineage to transmit Lyme disease [29]. The objective of this study was to identify Borrelia species within Arkansas ticks, canines, and white-tailed deer to determine the identity and locations of Borreliae endemic to Arkansas using PCR amplification of the flagellin (flaB) gene. The CDC defines an endemic county as "endemic for Lyme disease if there are at least two confirmed human cases that were acquired in that county or there are established populations of I. scapularis infected with B. burgdorferi" [30].

Tick, canine, and white-tailed deer collections
This project used the collections reported in Trout and Steelman [27]. Briefly, ticks were collected from canines (March 2006 to November 2007) and white-tailed deer (Oct. 2007 -Jan. 2008) and stored in 80% ethanol until visual identification of species and life stage [31][32][33]. Additionally, a small volume of canine or white-tailed deer blood (0.5-1 cc) was obtained from each host and stored on a FTA card (FTA W Indicating Micro Card, Whatman International Ltd, Maidstone, England) in an envelope labeled with collection information. FTA cards were stored at room temperature until DNA extraction. All FTA cards, all ticks from canines, and at least one specimen of each tick species from each white-tailed deer were subjected to further analyses.

Tick DNA extraction and detection for Borrelia species
To minimize DNA contamination, DNA extractions and PCR were conducted in different laboratories and reagents and equipment were dedicated to each procedure. Tick identification and extractions were conducted in the Veterinary Entomology Laboratory (VELUA) and PCR reactions were conducted in the Insect Genetics Laboratory (IGLUA) at the University of Arkansas. During each PCR, at least one blank reagent to detect contamination (negative control) and an appropriate positive control ensured PCR reagents and conditions were used. PCR and reaction product analyses were performed on the ticks according to the protocols of Trout et al. [28]. After tick identification to species, each specimen was dried on a paper towel to allow the residual ethanol to evaporate and was then cut longitudinally and half of the tick was subjected to the Qiagen Dneasy Insect Protocol (Qiagen Inc., Valencia, CA). Extracted DNA from half of the tick was stored at −20°C until further analyses. The remaining tick half was stored at −20°C for additional analysis or morphological confirmation as required. To evaluate the tick DNA, each sample was first assessed by PCR with Ixodidae specific mitochondrial primers (16 S + 2/16 S-1) using previously described cycling parameters [34]. If the mitochondrial marker amplified, then the tick was subjected to a genus specific flaB PCR to identify the presence of Borrelia DNA [6,35].

Host DNA extraction and detection for Borrelia species
Each canine and white-tailed deer FTA card was screened in duplicate to determine the presence of Borrelia DNA. The FTA card was cut into halves, where half of the card remained at VELUA for an initial screening and the other half was sent to the University of North Texas Health Science Center in Ft. Worth (UNTHSC) for verification screening. Each FTA card half was removed from its envelope and a 1.2 mm disc was punched from the card using a sterilized Harris Micro Punch and the paper disc was washed for DNA extraction according to the Whatman protocol (Harris MicroPunch W , Whatman International Ltd, Maidstone, England). After the punch had dried at room temperature, it was subjected to PCR analyses. To ensure DNA detection from FTA punches were not inhibited in any manner, FTA punches were assessed for PCR amplification of host cytochrome b genes [36]. As with the ticks, canines and white-tailed deer specimens were screened for Borrelia species by PCR in a genus specific manner [6,35].

Statistical analyses
Summary statistics and relative abundance of each tick species were calculated to determine overall trends within the population using JMP (α = 0.05) [37]. Fisher's exact tests and two-tailed T-tests were performed in Excel 2007 to determine if the species of tick or tick host had any effect on the probability that the tick would be PCR positive for Borrelia [38]. Since spatial and temporal sampling of ticks and hosts were different (e.g., ticks from white-tailed deer were collected only in the fall) and this sampling structure may affect findings, comparisons across and among cohorts were not conducted.
Sequence identification of Borrelia species PCR products, from flaB positive samples, were sent to UNTHSC for DNA sequence analysis. The PCR reaction products were hydrolyzed with ExoSAP-IT (USB Corporation, Cleveland, OH) and sequence determination was performed using a BigDYE Terminator v.3.1 Cycle Sequencing kit (Applied Biosystems, Inc., Foster City, CA) followed by capillary electrophoresis on a ABI PRISM 310 Genetic Analyzer (Applied Biosystems Inc., Foster City, CA) using the method of Williamson et al. [39]. Sequences were edited, aligned, and analyzed with Sequencher 4.7 (Gene Codes, Corporation, Ann Arbor MI) and compared with sequences in GenBank (National Center for Biotechnology Information, Bethesda, MD).

Spatial identification of Borrelia species
Canine and deer collection data along with their corresponding tick collections and Borrelia DNA presence/ absence data were georeferenced into ArcMap 9.0 (ESRI Redlands, CA USA) and projected to the NAD 1983 UTM Zone 15 N of the GCS North American 1983 Geographic coordinate system. Boolean operations and symbology were used to identify locations with PCR positive ticks and hosts. The resulting Boolean map was overlaid on an existing county map of Arkansas [40].

Identification of Borrelia DNA in the tick population
A total of 2123 ticks were included in this study and represented five tick species; I. scapularis (33%), A. americanum (31%), R. sanguineus (16 %), D. variabilis (9%), and A. maculatum (8%). An additional 3% were Amblyomma ticks that could not be identified to species because the tick specimens were damaged. The tick population was comprised of each life stage, but most were adults (83%) and more were collected from domesticated canines (67%) than from hunter-killed whitetailed deer (33%). Overall, 583 of the 2123 (27.5%) ticks tested by PCR produced an amplicon for the Borrelia flaB gene ( Table 1)

Identification of Borrelia DNA in canines and their ticks
Of the 173 canines and their 1427 ticks subjected to the flaB PCR, 5.8% (n = 10) and 22.7% (n = 324) were positive respectively. All ten positive canine flaB amplicons were sequenced and six samples were homologous to B. burgdorferi IP2 (AY345236), two were homolgous to B. lonestari NC/MD (AF273670), one was homologous to B. lonestari fragments MO2002-V1 (AY850063), and one was homologous to strain B. burgdorferi OK HS-2 (FJ871032) ( Table 2). Contingency tests revealed that the number of flaB positive and negative ticks were significantly different based on tick species (× 2 = 31.9; df = 5; P < 0.0001) such that I. scapularis (35.5%) had the highest flaB positive rate. The most frequently encountered tick on canines was A. americanum and 23.8% (151/634) were flaB PCR positive. Of the 173 canines, 87 canines did not have a positive tick, 27 canines had one positive tick, and 59 canines had more than one positive tick. Of the 59 canines with more than one positive tick, 18 canines had multiple specimens of the same species that were positive, 13 canines had specimens of two separate tick species positive, and two canines had specimens of three separate tick species PCR positive.
In canine collected ticks, the B. lonestari strain NC/ MD (AF273670) was also the most common amplicon and amplified in nymphs, males, and females. Positive ticks were identified primarily in northwest Arkansas where many of the canines were collected (Figure 2A). Ticks from B. burgdorferi positive canines did not generate B. burgdorferi sequences. Instead these ticks were either negative for flaB or produced flaB positive amplicons that were homologous with B. lonestari.     white-tailed deer did not have a positive tick (n = 91), but 70 white-tailed deer had more than one PCR positive tick of which 24 deer had more than one flaB positive tick species and one deer had three different tick species flaB positive. Fragments amplified from all five deer-collected tick species were most similar to the B. lonestari sequence AY850063. Positive ticks were identified throughout Arkansas ( Figure 2B).

Discussion
Borreliae were identified in field collected ticks, canines, and white-tailed deer throughout the state suggesting the bacteria (as a family) are endemic in Arkansas. Of interest was the diversity of Borrelia genotypes identified from the variety of different field collections ( Figure 1, Table 2). Five different B. lonestari genotypes and three different B. burgdorferi genotypes were identified in Arkansas. None of the genotypes appeared to be more common in one tick species or host; however, B. burgdorferi genotypes were only identified in ticks collected from deer -the significance of this finding remains to be determined. Specifically, B. burgdorferi genotypes were amplified only from twelve I. scapularis, two A. maculatum, and an Amblyomma tick that was too damaged for the species to be determined.  [39] and with B. lonestari in Arkansas white-tailed deer serum [26]. We documented Borrelia in five tick species infesting domesticated canines and white-tailed deer; consequently, Borreliae are likely endemic in Arkansas and in the southeastern U.S. by a more diverse tick repertoire than previously thought. Identifying Borrelia in ticks indicates that field collected ticks have fed on infected animals at some point in their life cycle and identifying Borrelia in canines and white-tailed deer indicates that infected ticks have transmitted Borrelia to these hosts. We observed a trend for ticks collected from canines to be positive with B. lonestari whereas ticks from deer could be positive with either B. lonestari or B. burgdorferi. This may be related to the species of tick collected on each host and the sampling periods, as canines were primarily infested with A. americanum and collected year round whereas white-tailed deer were primarily infested with I. scapularis and collected in the fall. Also, ticks feeding on white-tailed deer may actually be feeding on hosts that are incompetent reservoirs of Lyme disease [42]. Additional evidence is required to assert the potential for each role in transmission. Past research has indicated that several tick species are not active vectors of B. burgdorferi (e.g., transmission through salivary glands) [18,20], but additional research should isolate Borrelias from the study area and confirm these tick species are not passive vectors (e.g., transmission studies in a suitable animal model). Additionally, often ticks are misidentified and physicians do not think "beyond Lyme disease" to diagnose the patient properly [2]. Other tick-borne diseases transmitted in Arkansas include rickettsia spotted fevers [32,43] and Ehrlichiosis [13,39], perhaps the cases of Lyme disease reported in 2007 were actually misdiagnosed.
Monitoring tick populations on canines could provide reliable early detection of tick-borne disease outbreaks [44]. Monitoring field populations of ticks for B. burgdorferi and other tick-borne diseases is essential to minimize transmission and maximize management. Recent landscape changes in Arkansas are expanding urban environments into deer habitats [45,46]. Previously, this pattern of urban development into previous deer habitats in Lyme Connecticut led to the increased diagnoses of Lyme disease in the northeastern United States [47]; similar reports have indicated B. burgdorferi emerged independently in the Midwest and eastern United States [48][49][50].
Additional research into the involvement of Borrelia spp. in the southeastern zoonotic cycle (e.g., abundance of tick/host infection, reservoir maintenance, vector competency) should be conducted in order to determine the life cycle and etiology of the various Borreliae detected in this survey and what roles, if any, they may play relative to subsequent borreliosis observed in humans. Future work should also evaluate the interaction of A. americanum carrying B. lonestari, A. maculatum carrying B. burgdorferi, and I. scapularis carrying B. burgdorferi and what effect, if any, this might have on the pathogenicity of the Borreliae, as both tick species have been identified simultaneously infesting the same host [27,51]. It is possible that there are multiple interactions occurring between tick species (e.g., species competition for host), the variety of hosts (e.g., Ixodid ticks feed on three different hosts), and the different Borreliae (infesting both the tick and the host), which warrants additional research.

Conclusions
Data from this study indicate multiple Borrelia genotypes are endemic to Arkansas because flaB was amplified from ticks, canines, and white-tailed deer, but the exact role each genotype plays in transmission and epidemiology remains undetermined. There were significantly more amplicons of B. lonestari than B. burgdorferi present suggesting that the majority of tick-borne diseases in Arkansas are not B. burgdorferi.

Competing interests
The authors declare we have no competing interests with the work presented in this manuscript.
Authors' contributions RTF, CDS, and ALS conceived the study and designed the experiments. PMB and PCW conducted the work in Texas, and RTF conducted the work in Arkansas. RTF, KLK, PMB, and PCW analyzed the results. RTF and PCW wrote and edited the manuscript. All authors read and approved the final version of the manuscript.