Portal de Periódicos da CAPES

Transmission cycles of Borrelia burgdorferi and B. bissettii in relation to habitat type in northwestern California

2009; Wiley; Volume: 34; Issue: 1 Linguagem: Inglês

10.1111/j.1948-7134.2009.00010.x

1948-7134

Lars Eisen, Rebecca J. Eisen, Jeomhee Mun, Daniel J. Salkeld, Robert S. Lane,

Yersinia bacterium, plague, ectoparasites research

This study was undertaken to determine which rodent species serve as primary reservoirs for the Lyme disease spirochete Borrelia burgdorferi in commonly occurring woodland types in inland areas of northwestern California, and to examine whether chaparral or grassland serve as source habitats for dispersal of B. burgdorferi- or B. bissettii-infected rodents into adjacent woodlands. The western gray squirrel (Sciurus griseus) was commonly infected with B. burgdorferi in oak woodlands, whereas examination of 30 dusky-footed woodrats (Neotoma fuscipes) and 280 Peromyscus spp. mice from 13 widely-spaced Mendocino County woodlands during 2002 and 2003 yielded only one infected woodrat and one infected deer mouse (P. maniculatus). These data suggest that western gray squirrels account for the majority of production by rodents of fed Ixodes pacificus larvae infected with B. burgdorferi in the woodlands sampled. Infections with B. burgdorferi also were rare in woodrats (0/47, 0/3) and mice (3/66, 1/6) captured in chaparral and grassland, respectively, and therefore these habitats are unlikely sources for dispersal of this spirochete into adjacent woodlands. On the other hand, B. bissettii was commonly detected in both woodrats (22/47) and mice (15/66) in chaparral. We conclude that the data from this and previous studies in northwestern California are suggestive of a pattern where inland oak-woodland habitats harbor a B. burgdorferi transmission cycle driven primarily by I. pacificus and western gray squirrels, whereas chaparral habitats contain a B. bissettii transmission cycle perpetuated largely by I. spinipalpis, woodrats, and Peromyscus mice. The dominant role of western gray squirrels as reservoirs of B. burgdorferi in certain woodlands offers intriguing opportunities for preventing Lyme disease by targeting these animals by means of either host-targeted acaricides or oral vaccination against B. burgdorferi. North-coastal California is a primary focus of Lyme disease in the western United States (Fritz and Vugia 2001). Knowledge of the enzootic transmission cycle of the causative agent Borrelia burgdorferi sensu stricto (hereinafter referred to as B. burgdorferi), and its mode of transmission to humans, are critical for implementing locally appropriate measures to prevent Lyme disease. Early studies on tick vectors and mammalian reservoirs of B. burgdorferi in northern California and southern Oregon suggested an enzootic transmission cycle involving two species of ticks, Ixodes pacificus and I. spinipalpis, and several species of rodents including the dusky-footed woodrat (Neotoma fuscipes), the California kangaroo rat (Dipodomys californicus), the brush mouse (Peromyscus boylii), the deer mouse (P. maniculatus), and the piñon mouse (P. truei) (e.g., Burgdorfer et al. 1985, Brown and Lane 1992, Peavey and Lane 1995, Burkot et al. 1999). Of the above-mentioned ticks, only I. pacificus commonly bites humans and thus serves as a bridging vector of B. burgdorferi (Merten and Durden 2000). Whereas many of the ecological studies of B. burgdorferi in northern California before the mid-1990s targeted rodents and their ticks in chaparral habitats, Clover and Lane (1995) discovered that dense woodlands are the principal habitats where humans are at elevated risk for exposure to I. pacificus nymphs. To clarify this risk, most subsequent ecological and risk-assessment studies were conducted in various subtypes of dense woodland (e.g., Eisen et al. 2003a, 2006; Lane et al. 2004). To make matters even more interesting, the only previous study specifically focused on rodents in woodlands in northern California (Wright et al. 2000) demonstrated that dusky-footed woodrats and Peromyscus spp. mice were not the primary reservoirs for B. burgdorferi in the Sierra Nevada foothills; only one of 83 dusky-footed woodrats and none of 48 Peromyscus spp. mice were found to be infected with B. burgdorferi. Furthermore, the initial impression of enzootic maintenance of B. burgdorferi in California was transformed in 1998 by the discovery that Californian B. burgdorferi comprises a species complex of B. burgdorferi sensu lato (s.l.) group spirochetes. These include the human pathogen B. burgdorferi, B. bissettii which is not known to infect people in the United States, and several uncharacterized Borrelia spp. (Postic et al. 1998, Dennis and Hayes 2002) that were described recently (Postic et al. 2007). Re-examination of Californian isolates from naturally infected rodents of the previously mentioned species, and data from studies on other species of Neotoma woodrats, revealed that these rodents in some habitats, particularly chaparral or grassland, most commonly carry B. bissettii or B. burgdorferi s.l. spirochetes rather than B. burgdorferi (Postic et al. 1998, 2007; Lane et al. 1999; Vredevoe et al. 2004; Brown et al. 2006). The potential complexity of, and overlap among, the enzootic transmission cycles of B. burgdorferi and B. bissettii was underscored by experimental studies demonstrating that both I. pacificus and I. spinipalpis can transmit either of these spirochetes and that Peromyscus mice and Neotoma woodrats are effective reservoirs for both B. burgdorferi and B. bissettii (Lane et al. 1994, Peavey and Lane 1995, Dolan et al. 1997, Eisen et al. 2003b). In view of this new knowledge, it became necessary to conduct further studies in northwestern California aimed at determining enzootic transmission cycles of B. burgdorferi and B. bissettii in common habitats such as grassland, chaparral and, especially, the woodland environment where humans are at greatest risk of exposure to I. pacificus nymphs. One part of the data generated by these efforts revealed that an arboreal rodent, the western gray squirrel Sciurus griseus, is an important reservoir for B. burgdorferi in oak woodlands in Mendocino County (Lane et al. 2005). The specific aims of the present study were to: 1) determine which rodent species besides squirrels may serve as primary reservoirs for B. burgdorferi in commonly occurring woodland types in northwestern California and 2) examine whether chaparral or grassland may serve as source habitats for dispersal of B. burgdorferi- or B. bissettii-infected rodents into adjacent woodlands. Rodent studies in 2002 and subsequent drag sampling for host-seeking I. pacificus nymphs in the spring of 2003 encompassed 13 woodland sites with a leaf and/or fir-needle litter ground cover (hereinafter called woodland) located throughout Mendocino County (Figure 1). These sites represented three woodland types occurring commonly in the county: 1) southeastern oak woodlands dominated by Quercus spp. oaks and Pacific madrone (Arbutus menziesii), 2) western redwood/tanoak woodlands dominated by coast redwood (Sequoia sempervirens), tanoak (Lithocarpus densiflorus), and Douglas fir (Pseudotsuga menziesii), and 3) mixed oak/Douglas fir woodlands in the central and northwestern parts of the county. Climate conditions differ dramatically between these woodland types during the warm part of the year, with a gradient ranging from cool and moist in western redwood/tanoak woodlands to dry and hot in oak woodlands to the southeast (Eisen et al. 2003a). In 2002, oak woodland was represented by four sites at the University of California Hopland Research and Extension Center (HREC) and one site at a private property in Potter Valley (total of five sites). Redwood/tanoak woodland included single sites at the Montgomery Woods State Reserve, Hendy Woods State Park, and the Angelo Coast Range Reserve (Angelo-Redwood) (total of three sites). Oak/Douglas fir woodland was represented by three sites at a private ranch community (CHR) and single sites at a private property in Willits and the Angelo Coast Range Reserve (Angelo-Creek) (total of five sites). Locations of study sites in Mendocino County. Habitat types include oak woodland (triangles) bordering on grassland (gray triangles) or chaparral (black triangles), oak/Douglas fir woodland (squares), and redwood/tanoak woodland (stars). Rodent studies in 2003 and subsequent drag sampling for host-seeking I. pacificus nymphs in the spring of 2004 were carried out solely at the HREC. Chaparral and grassland commonly abut woodlands in the study area (Figure 2). These studies included four sites comprised of woodland and adjacent chaparral (Beasley, Hunt Club, James II, Maude), and four sites including woodland and adjoining grassland (Cell tower, Parson, Pepperwood, Tank) (Figure 1). Illustration of habitat heterogeneity at the Hopland Research and Extension Center. Rodent trapping and processing procedures and tick identification methodology were described previously by Eisen et al. (2004a, b). Briefly, rodents were trapped using Sherman traps (8 × 8-10 × 23-30 cm; H.B. Sherman Traps Inc., Tallahassee, FL) and Tomahawk traps (15 × 15 × 50 cm; Tomahawk Live Trap, Tomahawk, WI). Captured animals were anesthetized, identified to species, sex, reproductive status, and age (juvenile vs adult), measured, ear-tagged (National Band and Tag Co., Newport, KY), examined for presence of ticks (tick infestation data from 2002 and 2003 were published earlier by Eisen et al. 2004a, b), ear-punch biopsied, and bled retroorbitally. Records of Ixodes spinipalpis include I. neotomae, which was relegated to a junior synonym of I. spinipalpis (Norris et al. 1997). Rodent sampling during the county-wide 2002 study focused exclusively on woodlands and was conducted from 23 April - 28 June. A four-day trapping session was carried out once per site from 23 April - 4 May at HREC, 7-10 May at Potter Valley, 14-25 May at CHR, 28-31 May at Willits, 4-7 June at Angelo and 11-14 June at Montgomery Woods State Reserve and Hendy Woods. The sites at HREC were re-sampled during 25-28 June. This sampling schedule was designed to coincide, as closely as possible, with the peak activity period of I. pacificus immatures in the above-mentioned areas (Eisen et al. 2003a). Rodent sampling at HREC in 2003 was conducted twice per site over a four-day period, once during 5-30 May and again during 9 June - 4 July. The sampling sites extended from woodland to the woodland/grassland or woodland/chaparral ecotone and into the adjoining grassland or chaparral habitat (Eisen et al. 2004a). In 2003, sampling efforts targeting mammals in woodlands at the HREC also included shooting of western gray squirrels (Lane et al. 2005) and extensive but unsuccessful efforts to capture shrews and moles (Eisen et al. 2004a). The decision was made to shoot squirrels in 2003 because this rodent rarely entered live-traps in 2002 despite extensive trapping efforts. Host-seeking I. pacificus nymphs, which presumably resulted from larvae fed during spring or summer of the previous year (Padgett and Lane 2001), were collected by drag sampling in woodlands in spring 2003 as a follow-up to the county-wide 2002 rodent sampling, and in spring 2004 as a follow-up to the 2003 HREC rodent sampling. For an in-depth description of the drag-sampling methodology used, see Tälleklint-Eisen and Lane (2000). Sites were sampled on a single occasion from 12 May - 5 June 2003 in the county-wide study and twice from 28 April - 17 May 2004 at HREC. We did not attempt to collect host-seeking I. pacificus nymphs in either grassland or chaparral at HREC in 2004 because vegetative features of these habitats render their collection ineffective (Clover and Lane 1995). Animal capture and handling procedures followed protocols on file with the Animal Care and Use Committee at the University of California at Berkeley and the California Department of Fish and Game. Ear-punch biopsies (EPBs) and tick specimens were stored in 95% ethanol at room temperature prior to processing. EPBs are widely used for detection of B. burgdorferi s.l. in rodents (Sinsky and Piesman 1989). DNA extraction from EPBs or host-seeking I. pacificus nymphs, detection of B. burgdorferi s.l. DNA by polymerase chain reaction (PCR), and sequencing of DNA in positive samples followed methods described in detail by Lane et al. (2004, 2005). B. burgdorferi s.l. DNA was detected with two sets of primers in a nested PCR format that targets an approximately 200-bp fragment of the 5S-23S rRNA intergenic spacer region (ISR) of the B. burgdorferi s.l. genospecies complex. PCR amplicons were sequenced using both directions of the second set of PCR primers. DNA sequences were aligned using Sequencher 4.5 (Codon Code, Ann Arbor, MI). Only data for B. burgdorferi and B. bissettii are presented here. In 2002, PCR-based detection of B. burgdorferi s.l. in rodents was complemented by culture of EPBs (one 2-mm punch per pinna) in BSK-H medium with rifampicin (Sigma, St Louis, MO; 25-50 μg rifampicin per ml). Cultures were held at 34-35°C and examined for spirochetes twice, at weeks 1 and 3, by dark-field microscopy at 400 X magnification. Statistical analyses were carried out using the JMP® statistical package (Sall et al. 2005) and the results were considered significant when P < 0.05. Specific tests used are indicated in the text. Our initial intent was to calculate reservoir potential among rodent species in woodlands based on their abundance, larval loads, and specific infectivity to feeding larvae following Mather et al. (1989). This approach was abandoned when we later realized that the western gray squirrel, a trap-shy rodent for which we lacked reliable abundance data in 2002-2003, is a primary reservoir for B. burgdorferi in woodlands. Instead, we present a crude calculation based on minimal numbers of infected animals present and their tick loads to indicate the importance of western gray squirrels relative to other rodents as reservoirs for B. burgdorferi in HREC woodlands in 2003. Thirteen Mendocino County sites representing three different woodland types (oak/Pacific madrone, mixed oak/Douglas fir, redwood/tanoak) were examined in 2002 to determine which rodent species serve as important reservoirs for B. burgdorferi in this key habitat type for human exposure to I. pacificus nymphs. Peromyscus spp. mice were the rodents most commonly captured in woodlands in 2002 (accounting for 87.8% of all captured individuals), followed by dusky-footed woodrats (8.3%), California meadow voles (Microtus californicus; 1.8%), and Tamias spp. chipmunks (1.3%) (Table 1). We also captured one northern flying squirrel (Glaucomys sabrinus) and one western gray squirrel. Among the 229 rodents captured, we recorded a single infection with B. burgdorferi (Table 1); a deer mouse collected from the HREC-Pepperwood site was positive both by PCR and by culture of ear-punch biopsies. No rodent was infected with B. bissettii. The fact that host-seeking I. pacificus nymphs collected from the same sites in 2003 (and presumably having fed as larvae in spring of 2002) had infection rates with B. burgdorferi that reached 4-5% (Table 2) indicated that our trapping efforts in 2002 failed to include the primary spirochetal reservoir(s). Prevalences of infection with B. burgdorferi among host-seeking I. pacificus nymphs were similar for oak woodlands and oak/Douglas fir woodlands (Table 2; 5.0% and 4.2%, respectively; Fisher's exact two-tailed test, P= 0.83) but higher for these woodland types combined than for redwood/tanoak woodlands (4.6% versus 0%; P= 0.04). Further, nymphs were more commonly infected with B. burgdorferi than with B. bissettii in oak woodlands (Table 2; 5.0% versus 0.4%; P < 0.005), whereas the difference was not significant in either oak/Douglas fir woodlands (4.2% versus 1.7%; P= 0.17) or redwood/tanoak woodlands (0% in both cases; P= 1.0). Studies at the HREC in 2003 included efforts to take western gray squirrels by shooting (Lane et al. 2005) and yielded more animals infected with B. burgdorferi (12 out of 268) than the 2002 county-wide woodland study (one out of 229) (P= 0.004). Infected animals from 2003 included seven western gray squirrels and one dusky-footed woodrat from woodlands, one brush mouse from each of the grassland and chaparral habitats, and two piñon mice from chaparral (Table 3). Although lack of data for abundance of western gray squirrels precludes a more elaborate quantification of its relative importance as a reservoir for B. burgdorferi in HREC woodlands, we note that this species accounted for 87.5% (7/8) of the B. burgdorferi-infected rodents collected in woodlands (Table 3) and that the average number of I. pacificus larvae infesting western gray squirrels in this habitat from late May to late June (13.8) was higher than it was for either dusky-footed woodrats (8.5) or Peromyscus mice (1.6) (data on tick infestation from Eisen et al. 2004a). These data suggest that western gray squirrels contributed >90% of I. pacificus larvae that became infected with B. burgdorferi after feeding on rodents at HREC in 2003. The clearest association with B. burgdorferi occurred for western gray squirrels in woodlands (Table 3). There was no similar pattern for other rodents. Dusky-footed woodrats were equally unlikely to be infected with B. burgdorferi in woodlands (1/11), grassland (0/3), or chaparral (0/47) (Likelihood Ratio Test; χ2 = 3.50, P= 0.17). Peromyscus mice tended to be infected with B. burgdorferi more commonly in chaparral (3/66) or grassland (1/12) than in woodland (0/79), but the differences were not significant (χ2 = 5.96, P= 0.051). We also explored whether infection prevalences with B. burgdorferi or B. bissettii differed for woodrats or mice between chaparral-associated habitats and grassland-associated ones. There were no differences in the prevalence of infection with B. burgdorferi between chaparral-associated habitats and grassland-associated habitats for either dusky-footed woodrats (0/59 vs 1/10 animals infected; P= 0.15) or Peromyscus mice (3/107 vs 1/74 animals infected; P= 0.65). On the other hand, there appears to be a strong association between B. bissettii and chaparral at HREC (Table 3). Dusky-footed woodrats captured in chaparral-associated habitats were far more likely to be infected with B. bissettii (45.8%, n = 59 animals examined) than woodrats from grassland-associated habitats (0%, n = 10) (P= 0.005). Similarly, Peromyscus mice from chaparral-associated habitats were more commonly infected with B. bissettii (17.8%, n = 107) than mice from grassland-associated habitats (0%, n = 74) (P < 0.001). The prevalence of infection with B. burgdorferi among host-seeking I. pacificus nymphs in 2004 was similar for woodlands bordering grassland and woodlands bordering chaparral (Table 4; 6.5% and 5.3%, respectively; P= 0.55). The same held true for B. bissettii (0% in nymphs from woodlands bordering grassland vs 0.7% in nymphs from woodlands bordering chaparral; P= 0.25). Host-seeking I. pacificus nymphs were more commonly infected with B. burgdorferi than with B. bissettii in both woodlands bordering grassland (P < 0.001) and woodlands bordering chaparral (P < 0.001). Because B. bissettii-infected woodrats and Peromyscus mice were far more common in chaparral than in adjoining woodlands (Table 3), we examined how frequently these rodents moved between chaparral and woodland areas. In the four chaparral-associated sites, we trapped 25 dusky-footed woodrats and 46 Peromyscus mice on multiple occasions in 2003. Although most woodrats (72%) were captured exclusively within either chaparral or woodland, six animals utilized the woodland-chaparral ecotone or a combination of the ecotone and either chaparral or woodland. Further, usage of both chaparral and woodland was recorded for one juvenile female woodrat. The results for Peromyscus mice were similar. Most mice (78%) were captured only within chaparral or woodland but nine animals utilized the ecotone or ecotone and either chaparral or woodland, and a single animal (an adult male brush mouse) was trapped in both chaparral and woodland. Although none of the two animals trapped in both chaparral and woodland harbored detectable infections, five (19%) of 26 rodents trapped in the ecotonal area (including one deer mouse and a woodrat trapped in both ecotone and woodland) were infected with B. bissettii. Rodents examined for pathogens in both May and June 2003 (with a 5 week period between samples for individual animals) included 24 brush mice, 13 dusky-footed woodrats, four piñon mice, and two deer mice. All ten animals infected with B. bissettii or uncharacterized B. burgdorferi s.l. in May (six dusky-footed woodrats and four Peromyscus mice) also were infected in June. Of 26 Peromyscus mice and seven woodrats that were free of infection in May, only one brush mouse and one woodrat had acquired infections by June (B. bissettii in both cases). Two dusky-footed woodrats and three brush mice were examined for pathogens in May-June of both 2002 and 2003. In spring of 2002, all five animals were free of B. burgdorferi s.l. infection. By spring of 2003 one of the brush mice had acquired B. burgdorferi and one of the woodrats was found to be infected with B. bissettii. We found that in Mendocino County woodlands, Neotoma woodrats and Peromyscus mice are only infrequently infected with B. burgdorferi (Tables 1, 3) whereas western gray squirrels commonly harbor this spirochete (Table 3, Lane et al. 2005). Indeed, our 2003 findings suggest that western gray squirrels could have accounted for >90% of the production by rodents of fed I. pacificus larvae infected with B. burgdorferi in HREC woodlands. The dominant role of western gray squirrels as reservoirs of B. burgdorferi in certain woodlands offers intriguing opportunities for preventing Lyme disease by targeting these animals by means of either host-targeted acaricides or oral vaccination against B. burgdorferi. Infections with B. burgdorferi were rare in woodrats (0/47) and mice (3/66) captured in chaparral, and this habitat therefore appears unlikely to serve as a source for dispersal of this spirochete into adjacent woodlands. On the other hand, B. bissettii was commonly detected in both woodrats (22/47) and mice (15/66) in chaparral. The results from the present study and previous ones reporting on tick infestation of rodents at the HREC (Lane and Loye 1991, Brown and Lane 1992, Leprince and Lane 1996, Casher et al. 2002, Eisen et al. 2004a) are suggestive of a pattern where the oak woodland habitat harbors a B. burgdorferi transmission cycle driven primarily by I. pacificus and western gray squirrels, whereas chaparral habitats contain a B. bissettii transmission cycle perpetuated in large part by woodrats and Peromyscus mice together with I. spinipalpis. A similar B. bissettii transmission cycle has been reported from brushlands in Colorado (Burkot et al. 2001). In the case of B. burgdorferi, the transmission cycle occurring in HREC woodlands most closely resembles the situation observed in the British Isles where tree squirrels can be important reservoirs of B. burgdorferi s.l., whereas mice and voles appear to contribute little to the force of transmission (e.g., Craine et al. 1997, Gray et al. 2000). This is in stark contrast to the northeastern United States where the white-footed mouse (P. leucopus) is considered an important reservoir for B. burgdorferi (e.g., Mather et al. 1989, LoGiudice et al. 2003). One plausible explanation for this difference in importance of mice as B. burgdorferi reservoirs is low intensity of spirochete transmission among Peromyscus mice in California or Apodemus mice in the British Isles due to minimal loads of I. pacificus or I. ricinus nymphs. At HREC in 2003, for example, western gray squirrels carried an average of 6.2 I. pacificus nymphs, whereas no nymphs were recorded from Peromyscus mice (Eisen et al. 2004a). There appears to be a potent association between oak/Pacific madrone woodlands and B. burgdorferi, and another between chaparral and B. bissettii, in northwestern California. Nevertheless, woodrats, mice and host-seeking I. pacificus nymphs collected in woodlands occasionally are infected with B. bissettii (Tables 2, 4; Eisen et al. 2004c) and woodrats and mice from chaparral sometimes harbor B. burgdorferi (Table 3; Brown et al. 2006). Limited movements of rodents between chaparral and woodland, with woodrats and mice typically nesting and replete ticks detaching in the more sheltered chaparral habitat, may explain why the prevalence of infection with B. bissettii in I. pacificus nymphs in 2004 was so low (0.7%) in woodlands abutting chaparral despite the capture of five B. bissettii-infected rodents in the same woodland areas in 2003 (Table 4). Likewise, restricted, if not rare, movements of B. burgdorferi-infected western gray squirrels from woodlands to contiguous chaparral and nesting by squirrels exclusively in woodlands offer a plausible explanation for the infrequent detection of B. burgdorferi-infected woodrats or Peromyscus mice in chaparral (Table 3). Although considerable progress has been made towards untangling the web of B. burgdorferi s.l. transmission in California, we still are far from resolving the full complexity of enzootic transmission. This area contains a wondrous variety of habitat types, vertebrates ranging in spirochete reservoir roles from zooprophylactic (certain lizards) to unknown (birds, shrews) and to highly competent (California kangaroo rats, dusky-footed woodrats, western gray squirrels), and a tremendous genetic variability within the B. burgdorferi s.l. complex. The latter is underscored by the recent description of Borrelia californiensis and Borrelia genomospecies 1 and 2 (Postic et al. 2007). Those descriptions bring to five the number of B. burgdorferi s.l. genomospecies known to occur in California, and molecular evidence indicates that there yet is a wealth of uncharacterized B. burgdorferi s.l. spirochetes in the state (Brown et al. 2006). Perhaps more than for any other area where Lyme disease is a public health problem, California presents us with a situation where peculiar local B. burgdorferi s.l. transmission cycles are the norm rather than the exception and Lyme disease control must be locally adapted to stand a chance of succeeding. Numerous approaches to control Lyme disease by targeting tick vectors in vegetation or on reservoir or non-reservoir hosts of B. burgdorferi have been explored over the last two decades (reviewed by Piesman and Eisen 2008). Approaches that target rodent reservoirs of B. burgdorferi are predicated on first determining which reservoir species contribute most heavily to production of B. burgdorferi-infected vector ticks and thereafter to design and implement methodologies to prevent this production. The latter can conceptually be done in at least three different ways: 1) reduction of tick loads on rodent reservoirs to break the cycle of enzootic transmission of B. burgdorferi (e.g., Mather et al. 1987, Lane et al. 1998), 2) use of reservoir-targeted vaccines against B. burgdorferi (Tsao et al. 2004), or 3) treatment of reservoirs with antibiotics (Zeidner et al. 2004). Early efforts in California to apply pesticides to B. burgdorferi reservoirs were, based on the hypothesis that the dusky-footed woodrat is a primary reservoir for this spirochete in key risk habitats, focused primarily on woodrats (e.g., Leprince and Lane 1996, Lane et al. 1998). The recent realization that woodrats in Californian brushlands most commonly carry B. bissettii or currently uncharacterized B. burgdorferi s.l. spirochetes of unknown pathogencity to humans (Postic et al. 1998, Vredevoe et al. 2004, Brown et al. 2006) prompted a renewed search for key reservoirs of B. burgdorferi. These efforts have focused primarily on woodlands because this habitat presents the greatest risk of human exposure to B. burgdorferi-infected I. pacificus nymphs (e.g., Clover and Lane 1995; Eisen et al. 2003a, 2004c). Although dusky-footed woodrats are potential reservoir hosts of B. burgdorferi in certain oak-dominated, and possibly redwood-forest, habitats in northwestern California (Brown et al. 2006), the results of this and another study (Wright et al. 2000) demonstrate that they, along with Peromyscus mice, play minimal roles as B. burgdorferi reservoirs in many wooded habitats of northern California and thus are poor targets for control efforts in those habitats. Sciurus spp. squirrels are considered important reservoirs of Lyme disease spirochetes in parts of Europe (Craine et al. 1997, Humair and Gern 1998), and current evidence now implicates the western gray squirrel as a primary reservoir of B. burgdorferi in northern Californian woodlands (Lane et al. 2005, Salkeld et al. 2008). As noted earlier by Lane et al. (2005), this introduces intriguing new possibilities for prevention of Lyme disease in the far-western United States by targeting these squirrels by means of host-targeted acaricides (e.g., delivered through bait-boxes) or oral vaccination against B. burgdorferi. As an interesting side-note, a recent study from the eastern United States (Dolan et al. 2004) reported problems with keeping eastern gray squirrels from destroying bait-boxes aimed at attracting smaller rodents and exposing them to acaricide (squirrel vandalism). The western gray squirrel similarly may be attracted to bait-boxes and, if so, treated with acaricides in order to reduce tick loads and disrupt enzootic transmission of B. burgdorferi in high-risk areas. Potential non-rodent reservoirs of B. burgdorferi in Californian woodlands include insectivores, particularly shrews, and birds. Shrews have emerged as key reservoirs of Lyme disease spirochetes both in Sweden (Tälleklint and Jaenson 1994) and the northeastern United States (Brisson et al. 2008). Although shrews have rarely been collected in studies on tick hosts and vertebrate reservoirs of B. burgdorferi s.l. in California (e.g., Casher et al. 2002, Eisen et al. 2004a, b), we cannot yet rule out the possibility that this is because they behave in a manner that allows them to elude capture by standard rodent and pitfall traps. The role of birds as potential reservoirs of B. burgdorferi in Californian woodlands remains unclear. Some birds can be heavily infested with I. pacificus immatures in California but evidence that they commonly are infected with B. burgdorferi is still lacking (e.g., Manweiler et al. 1990, Wright et al. 2000, Slowik and Lane 2001, Eisen et al. 2004a, Lane et al. 2006). Additional work on the roles of both birds and Sciurus spp. squirrels in the dynamics of B. burgdorferi transmission in northern California is under way. This will further clarify the potential for using reservoir-targeted control to prevent Lyme disease in California. We thank the involved private landowners, the California Department of Parks and Recreation, and the University of California Angelo Coast Range Reserve for allowing us to collect rodents and ticks. The University of California Hopland Research and Extension Center granted us permission to collect rodents and ticks and kindly provided logistical support. Field or laboratory assistance was provided by D. Bonilla, I. Jones, J.E. Kleinjan, and E. Omi-Olsen. The research project was supported by Grant Number RO1AI022501 from the National Institutes of Allergy and Infectious Diseases. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institute of Allergy and Infectious Diseases or the National Institutes of Health.