Lyme disease in the United States is caused by the bacterial spirochete Borrelia burgdorferi s.s. (Johnson, Schmid, Hyde, Steigerwalt, and Brenner), which is transmitted by tick vectors Ixodes scapularis (Say) and I. pacificus (Cooley and Kohls). Borrelia lonestari, transmitted by the tick Amblyomma americanum L., may be associated with a related syndrome, southern tick-associated rash illness (STARI). Borrelia lonestari sequences, reported primarily in the southeastern states, have also been detected in ticks in northern states. It has been suggested that migratory birds may have a role in the spread of Lyme disease spirochetes. This study evaluated both migratory waterfowl and nonmigratory wild turkeys (Meleagris gallopavo silvestris, Eastern wild turkey) for B. burgdorferi and B. lonestari DNA sequences. A total of 389 avian blood samples (163 migratory birds representing six species, 125 wild turkeys harvested in habitats shared with migratory birds, 101 wild turkeys residing more distant from migratory flyways) were extracted, amplified, and probed to determine Borrelia presence and species identity. Ninety-one samples were positive for Borrelia spp. Among migratory birds and turkeys collected near migration routes, B. burgdorferi predominated. Among turkeys residing further away from flyways, detection of B. lonestari was more common. All A. americanum ticks collected from these areas were negative for Borrelia DNA; no I. scapularis were found. To our knowledge, this represents the first documentation of B. lonestari among any birds.
Lyme disease is the most common vector-borne disease in the United States and has been reported from nearly every state. From 1993 through 2006, a total of 386 cases of Lyme disease were reported from Tennessee (CDC 2007). The causative agent of Lyme disease, isolated from the black-legged tick, Ixodes scapularis (Say) (Burgdorfer et al. 1982), is the bacterial spirochete, Borrelia burgdorferi s.s. (Johnson, Schmid, Hyde, Steigerwalt, and Brenner), hereafter referred to as B. burgdorferi. Lyme disease is characterized by erythema migrans (EM), a bulls-eye rash that develops around the tick bite. A related syndrome, also exhibiting EM at the site of tick feeding, has been designated as southern tick-associated rash illness (STARI) (Burkot et al. 2001, James et al. 2001). The causative agent of STARI remains to be determined, although Borrelia lonestari (Barbour et al. 1996) has been suggested as the responsible spirochete (Barbour et al. 1996). Genetic sequences characteristic of B. lonestari have been reported from the lone star tick, Amblyomma americanum L. (Burkot et al. 2001, Stromdahl et al. 2003). The majority of reports documenting detection of B. lonestari have been from the southeastern United States (Moore et al. 2003, Stegall-Faulk et al. 2003). Borrelia lonestari has also been reported in northeastern states (Taft et al. 2005), suggesting that this putative agent of STARI is not confined to southern states. It would seem likely that the range of host ticks, I. scapularis or I. pacificus (Cooley and Kohls) for B. burgdorferi and A. americanum for B. lonestari, would be a primary determinant for where each syndrome occurs. However, detection of each Borrelia species has been documented where tick vectors are reportedly sparse.
Ixodes scapularis is established throughout much of the northeastern United States where Lyme disease is reported most frequently; however, I. scapularis is more limited in south central states, including Tennessee (Dennis et al. 1998). A. americanum is widely distributed throughout southern states, from Texas to the eastern seaboard (Childs and Paddock 2003). Limited tick surveys in Tennessee have recovered large numbers of A. americanum but not I. scapularis (Stegall-Faulk et al. 2003). Ludyjan-Ybarra (2004) collected 8,199 ticks in middle Tennessee during 2002–2003. A. americanum accounted for 91% of these ticks, and no I. scapularis were found.
Studies have documented the presence of B. burgdorferi in ixodid ticks that had been removed from birds (Stafford et al. 1995, Nicholls and Callister 1996, Durden et al. 1997). Some birds also have been reported to serve as competent hosts for B. burgdorferi, both in the laboratory (Burgess 1989, Richter et al. 2000) and in the wild (McLean et al. 1993, Ginsberg et al. 2005). It has been suggested that bird migration may have an important role in the distribution of Lyme disease spirochetes (Weisbrod and Johnson 1989, Reed et al. 2003).
This study was undertaken to evaluate both migratory and nonmigratory birds in Tennessee for the presence of B. burgdorferi and B. lonestari sequences. Blood was obtained from wild turkeys (Meleagris gallopavo silvertris, Eastern wild turkey) and migratory waterfowl. Among the wild turkey populations sampled, one group of these nonmigratory birds resided near the Tennessee National Wildlife Refuge (TNWR), a major resting and feeding area for migrating waterfowl. The second group of Eastern wild turkeys was harvested at sites more distant from the migratory refuge. We report detection of both Borrelia, although distinctly different species predominated among the bird groups sampled. No Borrelia spp. sequences were detected among A. americanum ticks collected in habitats of bird residence.
Materials and Methods
Birds and Blood Sample Handling.
Avian blood samples were obtained at Tennessee Wildlife Resources Agency check stations or directly from hunters. Samples were collected over three hunting seasons during November, December, January, February, April (2003–2004), November, December, January, February, April (2004–2005), and November, December, January, and March (2005–2006). Samples were taken from 226 nonmigratory wild turkeys. Turkeys were harvested in 10 different counties. One group of 101 wild turkeys represented a population residing more distantly from the TNWR. This habitat is a mosaic of pastureland and woodland, interspersed with small communities and very limited row crop agriculture. Counties included DeKalb, Jackson, Macon, Smith, White, and Wilson. The other group of 125 wild turkeys resided in counties within or near the TNWR (counties: Dickson, Montgomery, Robertson, and Stewart). The wild turkey habitat near the TNWR is represented by row crop and pastureland agriculture, grasslands, and wood areas, both riparian and upland woods. These areas include bottomland along rivers and areas drained by tributaries of rivers. Table 1 documents the presence of migratory waterfowl at each area of turkey harvest. Table 1 data were obtained from the Christmas Bird Count conducted by the National Audubon Society (website, 2007) during 2003–2004, 2004–2005, and 2005–2006. Only migratory birds evaluated during this study are listed. Far greater numbers of migratory birds were documented in sites near the refuge. Blood samples were also obtained from 163 migrating waterfowl representing six species: American black duck (Anas rubripes), Canada goose (Branta canadensis), mallard (Anas platyrhynchos), northern pintail (Anas acuta), ring-necked duck (Aythya collaris), and wood duck (Aix sponsa). These waterfowl were harvested in six counties: Bedford, Chester, Coffee, Dyer, Giles, and Rutherford (see Fig. 1 for Tennessee county locations). Blood was collected into 15-ml centrifuge tubes by us at check stations or by hunters in the field and transferred into cryovials in the laboratory and held at -20°C until extraction.
Tick Collection and Preparation.
Ticks were collected in 2004 (June, July, September) and 2005 (June, July, August) by dragging or CO2 traps as described (Sonenshine 1993). Collections were made in nine counties that represented areas where birds were harvested (DeKalb, Franklin, Giles, Jackson, Montgomery, Robertson, Rutherford, Stewart, White). Ticks were stored in 70% ethanol. Adult and nymph ticks were identified to species using a dissecting microscope (Keirans and Durden 1998). Only A. americanum were evaluated for Borrelia sequences because others have reported that Dermacentor fail to transmit Borrelia spp. (Sanders and Oliver 1995, Piesman and Happ 1997). Ticks were pooled for extraction and amplification of DNA as two adults (89 pools) or four nymphs (40 pools). Ticks were cut in half, placed in extraction Lysis Buffer (ZR Genomic DNA II Kit; Zymo Research, Orange, CA), and homogenized for 1 min with a hand-held Kontes Pellet Pestle Cordless Motor (Fisher, Pittsburgh, PA).
Extraction and Amplification.
DNA was extracted from tick homogenates and bird blood samples using the ZR Genomic DNA II Kit according to manufacturer’s instructions. Samples underwent two rounds of amplification using primers specific for a conserved region of the Borrelia flagellin gene, which would permit amplification of both B. burgdorferi and B. lonestari. Primer sequences for the first amplification were as follows: forward, 5′CAAAAATTAATACACCAGCAT; reverse, 5′GCAATCATAGCCATTGCAGA (all primers were prepared by Integrated DNA Technologies, Coralville, IA). Nested amplification primer sequences were as follows: forward, 5′CTAATGTTGCAAATCTTTT; reverse, 5′GCATCTTTAATTTGAGCATA (Haynes et al. 2005). Predicted nested fragment sizes were 319 bp for B. burgdorferi and 298 bp for B. lonestari. Taq polymerase, buffer, and dNTPs were obtained from Promega (Madison, WI). Uninfected chicken blood (Carolina Biological Supply, Burlington, NC) or dH2O controls were included as negative controls both for extractions and amplifications. Barrier pipette tips (Molecular BioProducts, San Diego, CA) were used for all manipulations. Sequences characteristic of B. lonestari, isolated from an A. americanum tick in Tennessee (Stegall-Faulk et al. 2003; GenBank accession AF408410) and B. burgdorferi strain B-31 were used as positive controls to verify amplification, gel fragment size, and probe hybridization. Positive controls were not used during sample extraction or amplification. Ten to 12 samples were extracted/amplified at a time. Polymerase chain reaction (PCR) products were resolved on a 1.25% agarose gel (NuSieve; FMC BioProducts, Rockland, ME) and visualized with ethidium bromide.
Dot Blot Hybridization.
Nested PCR products showing fragment sizes characteristic of Borrelia spp. were hybridized with B. burgdorferi-specific and B. lonestari-specific probes. The probe sequence used to identify B. burgdorferi was as follows: 5′ATCTATAAAGAATAGTACT, and for B. lonestari was 5′AGCTCAAGGTGGGATTAGC (Taft et al. 2005) (probes were prepared by Integrated DNA Technologies). PCR products were applied to two nylon membranes (Zeta Probe; Bio-Rad Laboratories, Hercules, CA). One membrane was hybridized with the B. burgdorferi-specific probe and the second with the probe specific for B. lonestari. Probes were labeled with digoxigenin (Roche Applied Science, Indianapolis, IN) and allowed to hybridize overnight at 50°C. Membranes were washed and blocked according to manufacturer instructions for the DIG Nucleic Acid Detection Kit (Roche Applied Science). The presence of digoxigenin was detected using anti-digoxigenin Fab fragments supplied in the kit.
χ2 tests were used to determine differences between Borrelia spp. and the bird groups. Numbers of B. burgdorferi and B. lonestari were compared between nonmigratory birds more distant from the TNWR and nonmigratory birds near the TNWR, nonmigratory birds more distant and migratory birds, and nonmigratory birds within the TNWR and migratory birds.
A total of 387 ticks were collected from counties representing locations of waterfowl and wild turkey harvests. Forty-eight D. variabilis (Say) were recovered: 43 adults and 5 nymphs. A. americanum comprised the remaining 339 ticks: 178 adults and 161 nymphs. No I. scapularis were recovered. None of the A. americanum ticks showed evidence of Borrelia spp. DNA.
Blood samples from 389 birds were tested for evidence of Borrelia spp. sequences. Ninety-one samples (23%) were positive by PCR for either B. burgdorferi or B. lonestari. Three avian populations were represented, and the counties of collection are shown in Fig. 1: (1) nonmigratory wild turkeys residing away from the wildlife refuge (black in Fig. 1), (2) nonmigratory wild turkeys collected in or near the refuge (light gray), and (3) migrating waterfowl (dark gray). The heavy black lines in Fig. 1 represent the Land Between the Lakes region, part of the TNWR, a major resting and feeding area for migrating birds. Christmas bird counts (Table 1, averaged from National Audubon Society data) document the presence of the migratory birds in each area of wild turkey collection during the same seasons that samples were collected in this study. Substantially greater numbers of migrating waterfowl were observed in the refuge area (light gray counties, Fig. 1), whereas fewer waterfowl were reported in areas more distant from the refuge (black counties, Fig. 1).
Detection of Borrelia spp. was accomplished through primary and nested amplification of blood extracts. Nested PCR products were evaluated by gel electrophoresis. Two representative samples from each group of birds (turkeys more distant from the TNWR, turkeys harvested in proximity to the TNWR, and migratory waterfowl) are presented in Fig. 2. Gel lanes 2, 3, and 4 are controls, negative, positive B. burgdorferi, and positive B. lonestari, respectively. Lanes 6, 9, and 12 contain nested PCR fragment sizes predicted for B. burgdorferi (319 bp). PCR products in lanes 7, 10, and 13 are slightly smaller and migrate further on the gel, characteristic of B. lonestari (298 bp). Fragment sizes are confirmed by the standard ladder in lanes 1 and 15 and the positive controls in lanes 3 and 4.
Verification of Borrelia spp. identity was done by species-specific probe hybridization (Fig. 3). All positive nested PCR products (91 samples) were applied to each membrane. The membrane in Fig. 3A was hybridized with the probe that binds to sequences characteristic of B. burgdorferi, whereas 3B was hybridized with the B. lonestari--specific probe. Nested PCR products from wild turkeys away from the TNWR were placed in positions A1 through B10. The first three spots, A1, A2, and A3, contained DNA characteristic of B. burgdorferi, and the remaining 19 spots were B. lonestari. PCR samples from wild turkeys harvested near the TNWR were placed in C1 through E4. B. burgdorferi samples were evident in C1 through D6 (18 samples) and B. lonestari in D7 through E4 (10 samples). Samples from migratory waterfowl were applied to positions E7 through H11. Dark spots at E7 through G11 (29 samples) indicated B. burgdorferi, whereas G12 through H11 (12 samples) were B. lonestari.
Distinct differences between avian groups were evident in relation to which species of Borrelia predominated (Fig. 4). Borrelia spp. identity was significantly different (P < 0.001) for turkeys collected more distant from the TNWR. There was no difference for Borrelia spp. between migratory birds and turkeys harvested within the TNWR. Among wild turkeys residing farther away from the TNWR, 19 of 101 turkey blood samples (18%) had sequences characteristic of B. lonestari, whereas only 3 (2%) had evidence of B. burgdorferi. The higher percentage of B. lonestari was consistent for all counties representing this group of wild turkey collection (Table 2). For wild turkeys harvested near the TNWR, the percentages of each Borrelia spp. were nearly identical to that found among migrating birds. B. burgdorferi sequences were detected in 14% of these turkeys compared with 17% of migratory birds, and B. lonestari was detected at an 8% level for wild turkeys compared with 7% for waterfowl (Fig. 4). With little exception, the higher rate of B. burgdorferi among this group of birds was maintained regardless of the county of sample collection (Table 2).
Among the six species of migratory fowl, mallards were collected most frequently, and Borrelia spp. sequences were detected in 29% of samples (Table 3). Canada geese were the next largest migratory group sampled, and only B. burgdorferi sequences were detected. No Borrelia spp. were detected among northern pintail or American black ducks, although the number sampled was low.
A number of studies have evaluated ticks and their mammalian or avian hosts for Borrelia spp. Evidence of Lyme spirochete presence has been obtained through spirochete culturing, microscopy, immunologic testing, or molecular methods. In studies by Olsen et al. (1995a, 1996), PCR detection of Borrelia spirochetes was more sensitive than culture. This study evaluated ticks and blood samples from birds by DNA amplification and probe hybridization to detect and differentiate Borrelia spp. The detection of Borrelia sequences does not necessarily indicate active infection. Conclusions from this study are somewhat limited in the absence of direct acarological data from birds. A. americanum ticks collected from locations where birds were harvested were negative for Borrelia spp. DNA. A. americanum was the predominant tick recovered, and no I. scapularis were found, in agreement with previous tick surveys conducted in middle Tennessee (Stegall-Faulk et al. 2003, Ludyjan-Ybarra 2004). Tick collections in this study were conducted in summer months and may have missed I. scapularis. To our knowledge, there have been no reports of B. burgdorferi detection in birds or ticks in Tennessee. This is the first study to report the presence of B. lonestari sequences among any birds. Infection of wild turkeys in this study by Borrelia spp. is caused by ticks in their environment, whereas waterfowl encounter different habitats and ticks during migration. Several studies have shown extensive parasitism of wild turkeys by A. americanum (Jacobson and Hurst 1979, Kollars et al. 2000, Mock et al. 2001). Wild turkey in this study may have been exposed to B. lonestari through A. americanum.
This study did not evaluate migratory passerine birds, although these birds could participate in transporting Borrelia spirochetes or ticks into Tennessee. The involvement of passerines and their ticks has been studied in relation to Borrelia spp. distribution (Anderson et al. 1986, Morshed et al. 2005). Migratory passerines have been documented to carry ticks long distances, potentially establishing new foci for Lyme disease (Weisbrod and Johnson 1989, Smith et al. 1996). Because tick attachment may last several days and a migrating bird could travel hundreds of miles daily (Reed et al. 2003), it is possible that migration has an impact on the geographic occurrence of Lyme disease or the proposed STARI syndrome. A. americanum was collected from a passerine in Canada (Scott et al. 2001), well outside the normal range for this tick. The potential contribution of waterfowl to the spread of Lyme spirochetes and ticks is less well understood. To our knowledge, there are few reports of ticks parasitizing waterfowl. In one North American study of the presence of B. burgdorferi in ticks in migrating birds, no ticks were found on mallards, lesser scaup (Aythya affinis), or a Canada goose (Ogden et al. 2008). Hutcheson et al. (2005) showed transmission of West Nile virus to pekin ducklings (Anas domesticus) by the seabird tick Carios capensis (Neumann). B. lonestari has also been isolated from C. capensis (Reeves et al. 2006). There have been reports for evidence of tickborne encephalitis virus in mallards (Ernek et al. 1975, Sonenshine and Mather 1994). This virus is transmitted by Ixodes ricinus L. and I. persulcatus (Schulze), each a known vector of Borrelia spp. To better elucidate the contribution of waterfowl to the spread of Borrelia spp., it would be advantageous to determine what ticks are normally present on these birds. In this study, it is unknown whether Borrelia-infected waterfowl may have served as host for resident ticks, whether already-infected ticks were transported to Tennessee through bird migration, or whether waterfowl have any impact on the spread of Borrelia spp. in the state, serving only as "accidental" hosts.
The role of birds in the distribution and maintenance of Borrelia spp. has been extensively studied in Europe and the United Kingdom. Olsen et al. (1995b) evaluated ticks taken from migratory birds in Scandinavia for B. burgdorferi s.l. Just >26% of I. ricinus ticks, the predominant tick recovered, were infected as determined by PCR amplification. A large study of migratory birds found that passerine birds were responsible for transporting Lyme borreliosis–infected ticks and could serve as a reservoir for Borrelia spirochetes (Comstedt et al. 2006). PCR-based studies conducted in Germany suggested that I. ricinus ticks became infected with B. burgdorferi s.l. after feeding on Borrelia-infected birds (Kipp et al. 2006, Pichon et al. 2006). In Switzerland, Humair et al. (2007) showed by molecular analysis of I. ricinus bloodmeals that birds had served as host for some Borrelia-infected ticks.
In North America, laboratory trials documented American robins (Turdus migratorius) as efficient hosts for Lyme disease spirochetes (Richter et al. 2000, Ginsberg et al. 2005). Spirochetes were recovered from blood 3 wk after subcutaneous infection of Northern bobwhite quail (Colinus virginianus) (Bishop et al. 1994). Burgess (1989) infected mallards orally or intravenously with B. burgdorferi. Spirochetes, detected by immunofluorescence, were present in droppings 22 d after infection and for 29 d in blood. A study in Slovakia recovered B. burgdorferi from a cloacal swab taken from a mallard (Schwarzova et al. 2006). Potential fecal–oral transmission of B. burgdorferi among waterfowl needs to be more fully investigated. In this study, 29% of mallards showed Borrelia spp. sequences in blood (Table 3). Although transmission through fecal contamination remains to be established, it is conceivable that some of these waterfowl were infected orally, either within the state or elsewhere. However, it is unlikely that wild turkeys would have been infected through mallard excretion of Borrelia spirochetes.
Although nonmigratory birds have little role in extensive geographic spread of Borrelia spp., their potential role as reservoir hosts could perpetuate the tick–host cycle. One report suggested that turkeys have little impact on the maintenance of B. burgdorferi in California (Lane et al. 2006). However, at another site in California, 53% of blood samples from ground-dwelling birds were reactive with B. burgdorferi antibody (Wright et al. 2000). Ground-dwelling birds were shown to be infested with significantly higher numbers of ticks than canopy birds (Wright et al. 2006). Without evidence of Borrelia spirochetes in ticks collected directly from wild turkey in this study, it is not possible to draw conclusions regarding what role turkeys in Tennessee may have in relation to Borrelia spp. maintenance. The dissimilar detection rates of Borrelia spp. among the resident wild turkey groups in this study (Table 2; Fig. 4) may be related to a number of complex interactions including habitat variation, the presence of different vectors and mammalian hosts, or the existence of ticks associated with migratory passerines or waterfowl.
This study showed the presence of Borrelia spp. DNA sequences among birds in Tennessee. Future study into waterfowl, their ticks, and the presence of Borrelia spirochetes would provide important information regarding their potential role in the distribution of Lyme disease and STARI.
R. Fox, Assistant Director of the Tennessee Wildlife Resources Agency, is acknowledged for assistance with avian blood sample collection at TWRA check stations. We thank M. Travis who coordinated sample collection among hunters and S. Wright for technical assistance. This work was supported, in part, by National Science Foundation Grants 0216716 and 0227754).