The enzootic life-cycle of Borrelia burgdorferi (sensu lato) and tick-borne rickettsiae: an epidemiological study on wild-living small mammals and their ticks from Saxony, Germany

Borrelia burgdorferi (sensu lato) and rickettsiae of the spotted fever group are zoonotic tick-borne pathogens. While small mammals are confirmed reservoirs for certain Borrelia spp., little is known about the reservoirs for tick-borne rickettsiae. Between 2012 and 2014, ticks were collected from the vegetation and small mammals which were trapped in Saxony, Germany. DNA extracted from ticks and the small mammals’ skin was analyzed for the presence of Rickettsia spp. and B. burgdorferi (s.l.) by qPCR targeting the gltA and p41 genes, respectively. Partial sequencing of the rickettsial ompB gene and an MLST of B. burgdorferi (s.l.) were conducted for species determination. In total, 673 small mammals belonging to eight species (Apodemus agrarius, n = 7; A. flavicollis, n = 214; Microtus arvalis, n = 8; Microtus agrestis, n = 1; Mustela nivalis, n = 2; Myodes glareolus, n = 435; Sorex araneus, n = 5; and Talpa europaea, n = 1) were collected and examined. In total, 916 questing ticks belonging to three species (Ixodes ricinus, n = 741; Dermacentor reticulatus, n = 174; and I. trianguliceps, n = 1) were collected. Of these, 474 ticks were further investigated. The prevalence for Rickettsia spp. and B. burgdorferi (s.l.) in the investigated small mammals was 25.3 and 31.2%, respectively. The chance of encountering Rickettsia spp. in M. glareolus was seven times higher for specimens infested with D. reticulatus than for those which were free of D. reticulatus (OR: 7.0; 95% CI: 3.3–14.7; P < 0.001). In total, 11.4% of questing I. ricinus and 70.5% of D. reticulatus were positive for Rickettsia spp. DNA of B. burgdorferi (s.l.) was detected only in I. ricinus (5.5%). Sequence analysis revealed 9 R. helvetica, 5 R. raoultii, and 1 R. felis obtained from 15 small mammal samples. Small mammals may serve as reservoirs for Rickettsia spp. and B. burgdorferi (s.l.). While the prevalence for Rickettsia spp. in M. glareolus is most likely depending on the abundance of attached D. reticulatus, the prevalence for B. burgdorferi (s.l.) in small mammals is independent of tick abundance. Dermacentor reticulatus may be the main vector of certain Rickettsia spp. but not for Borrelia spp.


Background
Tick-borne diseases require invertebrate vectors (ticks) and vertebrate hosts for the completion of their lifecycle [1,2]. Two of the most common tick species in Europe -and at the same time the most important vectors -are the castor bean tick Ixodes ricinus and the meadow tick Dermacentor reticulatus. Their immature life stages (larvae and nymphs) parasitize mostly on small-sized birds and on small mammals. This is why small mammals are essential for the maintenance and distribution of ticks and thus tick-borne diseases [3][4][5][6][7].
Prevalence rates for Borrelia spp. and Rickettsia spp. in I. ricinus ticks in Germany differ and can reach levels of 34 and 61%, respectively [25][26][27][28][29][30][31]. In Germany, the investigations of Rickettsia spp. in wild-living small mammals are scarce and were conducted mostly on Myodes glareolus, Apodemus flavicollis and Erinaceus europaeus [32][33][34]. Earlier, Borrelia spp. was detected in small animals such as Glis glis, E. europaeus, A. flavicollis and Mus musculus in Germany [35][36][37]. However, all studies previously published on Borrelia spp. in small mammals from Germany were focused on the detection of a single locus (ospA gene). In the present study, multi-locus sequence typing (MLST) of eight housekeeping genes was conducted in order to detect different sequence types of B. burgdorferi (s.l.) in small mammals.
The aims of this study were (i) detection of tick-borne rickettsiae and B. burgdorferi (s.l.) by qPCR in captured small mammals and in the questing ticks from selected suburban areas in Saxony, Germany; (ii) species identification of these pathogens by conventional PCR and MLST; and (iii) comparison of prevalence rates of B. burgdorferi (s.l.) and of tick-borne rickettsiae between the respective small mammals and tick species.

Study sites
From 2012 to 2014, small mammals as well as questing ticks were collected at six different study sites in and near the city of Leipzig in Saxony, Germany. Previously, these study sites were described in detail and consecutively named from "E" to "I" (E: 51°16'27.6"N, 12°1 9'18.8"E; F: 51°17'13.0"N, 12°20'40.2"E; G: 51°16'20.3"N, 12°23'12.7"E, H1: 51°18'14.6"N, 12°24'41.4"E; H2: 51°1 7'35.5"N, 12°24'07.5"E, I: 51°18'01.2"N, 12°22'09.5"E) by our group [38]. Three of those six study locations (sites E, F and G) surround a lake which was artificially created from a former brown coal mining area and which is now often frequented by visitors for recreational activities. Site "H" is subdivided in two small areas located in a recreational city park which was created from a former waste disposal area. Site "I" is a part of one of the largest riparian forests in Middle Europe and is located near the city centre of Leipzig. Sites "I" as well as "G" were only investigated in 2012 due to financial restrictions (see complete sequence batches in Additional files 1 and 2).

Small mammals and their attached ticks
Small mammals were captured from March to October in 2012, from January to November in 2013, and from January to October in 2014. Each month, twenty Sherman© live animal traps (H. B. Sherman Traps, Inc., Tallahassee, Fla., USA) were baited with apple slices and placed at each study site for two consecutive nights. Captured small mammals were immediately anesthetized with CO 2 and subsequently euthanized by cervical dislocation (local permit numbers: 36.11-36.45.12/4/10-026-MH, 364.60/2009-102-2). By the use of taxonomic keys, captured animals were morphologically identified [39]. For the present study, the ectoparasites (ticks in particular) were additionally collected from their bodies. Skin samples as well as ticks, which were morphologically identified [40] in advance, were stored at -80°C until further processing.

Collection of questing ticks
Simultaneously to each rodent trapping action, questing ticks were collected monthly by the use of the flagging method at each study site. The ticks were morphologically identified and stored individually at -80°C until further processing [40].

Tissue preparation and DNA extraction
Skin samples were taken individually and then 0.6 g of sterile steel beads (sized 2.8 mm, Peqlab Biotechnologie, Erlangen, Germany) as well as 600 μl phosphate buffered saline were added to each sample. Moreover, 0.6 g ceramic beads (sized 1.4 mm, Peqlab Biotechnologie) and 200 μl PBS were added to each engorged or questing tick. All samples were homogenized at 5700× rpm for 20 s in the Precellys®24 tissue homogenizer (Bertin Technologies). Subsequently, DNA was extracted from all samples with the QIAamp DNA Mini Kit (Qiagen, Hilden, Germany) according to the manufacturer's recommendations for tissue DNA extraction. The quality and the quantity of the DNA samples were measured with a spectrophotometer (NanoDrop® 2000c, Peqlab Biotechnologie).

PCR methods
Initially, small mammal and tick DNA samples were screened for the presence of Rickettsia spp. and Borrelia burgdorferi (s.l.) by qPCR. Real-time PCR analysis targeting the citrate synthase gene (gltA, 70 bp) was performed for Rickettsia spp. as previously described [41]. The initial screening for Borrelia burgdorferi (s.l.) which is targeting the p41 flagellin gene (96 bp) was carried out following a previously published protocol [42].
All Rickettsia-positive samples yielding a cycle threshold value (CT) below 35 were further analysed by a conventional PCR targeting 811 bp of the outer membrane protein B gene (ompB) of SFG rickettsiae [43]. A 1.5% agarose gel was stained with Midori Green (NIPPON Genetics, Düren, Germany) and PCR products were analysed under UV illumination. Five randomly selected samples which were positive for B. burgdorferi (s.l.) by real-time PCR and yielded a CT value below 33 were further analysed by multi-locus sequence typing (MLST) targeting the following housekeeping genes: nifS, pyrG, clpX, pepX, uvrA, rplB, cplA and recG [44]. For all genes a semi-nested or a nested approach was performed as described, however with slight modifications. The first amplification step for the genes clpX, rplB, pepX as well as the second amplification step for the genes rplB, clpA and clpX were performed with a touchdown protocol with 11 cycles with annealing temperatures ranging down from 56 to 46°C, and further 34 cycles with an annealing temperature of 46°C. The first amplification step of the nifS gene was likewise a touchdown protocol with nine cycles with annealing temperatures ranging down from 51 to 43°C, and further 36 cycles with an annealing temperature of 46°C. The annealing temperature of the nifS gene in the second amplification step was 51°C as for the uvrA gene in both amplification steps. The annealing temperature for the first amplification step of the recG gene and for the second amplification step of the pepX gene was 55°C. The annealing temperature of the first amplification step of the pyrG gene and the clpA gene was 47°C. The annealing temperature in the second amplification step was 49°C for the pyrG gene and 50°C for the recG gene.
Sequencing was performed commercially (Interdisziplinäres Zentrum für Klinische Forschung, Leipzig, Germany) for both, Rickettsia spp. and Borrelia spp. MLST, with forward and reverse primers of each gene used for PCR amplification. Results were analysed with the Bionumerics Software (Version 7.6.1. Applied Maths, Inc., Austin, TX USA). Sequences were aligned to available data in GenBank with BLASTn (http://blast.ncbi.nlm.nih.gov/Blast.cgi) Obtained MLST sequences were aligned and compared to sequences from the MLST database (http://pubmlst.org/borrelia).

Statistical analysis
Confidence intervals (95% CI) were determined for prevalences of Rickettsia spp. and B. burgdorferi (s.l.) in small mammals and in the questing ticks by the Clopper and Pearson method with the use of the Graph Pad Software (Graph Pad Software Inc., San Diego, Ca., USA). Pearson's Chi-squared test was used with a type I error α of 0.05 to test the independence of compared prevalences. Fisher's exact test was used for small sample sizes (n < 30) (Graph Pad Software). The odds ratio was calculated testing the association between the D. reticulatus ticks burden on Myodes glareolus and the prevalence of Rickettsia spp. in M. glareolus.

Collection of questing ticks
Altogether 916 questing ticks were collected: 741 I. ricinus (79 females, 105 males, 504 nymphs and 53 larvae), 174 D. reticulatus (72 females and 102 males) and one I. trianguliceps (female). The breakdown of ticks by year and life-cycle stage is shown in Table 2.
PCR analysis for Rickettsia spp. and Borrelia burgdorferi (s.l.) in small mammals  R. raoultii showed 100% identity to a sequence in Gen-Bank (KU961542) which was earlier obtained from a D. marginatus tick from Russia (Katarshov et al. unpublished). The single R. felis sequence showed 100% identity to a sequence in GenBank (GU324467) which was also obtained from A. flavicollis in Germany [33]. The prevalence and the distribution of Borrelia spp. as well as Rickettsia spp. for all small mammal species are shown in Table 3.

Discussion
This study was focussed on the detection of Borrelia burgdorferi (s.l.) and rickettsiae of the spotted fever group in wild-living small mammals and questing ticks from Germany. Borrelia burgdorferi (s.l.) is the causative agent of Lyme disease (LD) which is the most prevalent tick-borne disease in Europe and North America [8,9]. LD may cause severe symptoms with manifestations in the skin, joints, nervous system and heart tissue in humans as well as in companion animals, especially in dogs [45][46][47][48]. Ixodes ricinus is known to be the main vector in Europe, whereas I. scapularis is the main vector in North America, and I. persulcatus in Eurasia  Other: Mustela nivalis (n = 2); Sorex araneus (n = 5); Talpa europaea (n = 1); Microtus agrestis (n = 1) [49][50][51]. The prevalences of B. burgdorferi (s.l.) in I. ricinus in Europe differ regionally. Studies from Europe, e.g. France [52,53], the Netherlands [54], Slovakia [55] and Austria [56], show infection levels in I. ricinus ticks ranging from 3.3 to 22.5%. Earlier studies from Germany also showed high prevalence ranging from 11 to 36.2% in different regions of the country [57][58][59]. The present study confirms I. ricinus as the main vector for B. burgdorferi (s.l.), as the prevalence from this study was in line with previous studies from Europe [52][53][54][55][56]; however being lower than in previous studies from Germany (5.5%) [57][58][59]. The absence of Borrelia burgdorferi (s.l.) in questing I. ricinus larvae suggests a non-existent or insufficient transovarial transmission path [60]. However, transstadial transmission in ticks is verified [61]. Previous studies reported significantly higher prevalence for B. burgdorferi in adult I. ricinus ticks than in nymphs [52,56,59]. Our results are in contrast to these findings as I. ricinus nymphs were significantly more frequently infected than I. ricinus adults. Although in the past, spirochetes were detected in 11% of adult D. reticulatus ticks by immunofluorescence microscopy employing an antibody against B. burgdorferi [62], this non-specific method may likewise detect similar spirochetes such as B. miyamotoi [63]. Moreover, other study confirmed that D. reticulatus is not a suitable vector for B. burgdorferi (s.l.) [64,65]. In our study, none of the D. reticulatus ticks examined tested positive for B. burgdorferi (s.l.); this supports the view that D. reticulatus is of minor importance in the natural life-cycle of this pathogen complex. More than 40 vertebrate species, in particular birds and small mammals like rodents, are considered as reservoir hosts for B. burgdorferi (s.l.) in Europe [12,13]. Previous studies from France, Ireland and Austria showed prevalence of B. burgdorferi (s.l.) in small mammal species ranging from 2.3 to 24% [66][67][68]. The infection level in small mammals in the current study was slightly higher than these obtained in earlier European studies (31.3%). In present research, each species belonging to the order Rodentia was positive and with high prevalence of B. burgdorferi (s.l.) (25.4-62.5%), whereas the insectivores (1 Talpa europaea and 5 Sorex araneus) and the carnivores (1 Mustela nivalis) were all negative. These findings are in line with a study from Austria where all rodent species were positive for B. burgdorferi (s.l.) and also with high prevalence (13.3-77.0%) [68]. The prevalence of spirochetes in rodents from this study was high and independent from their tick burden, and moreover significantly higher than in questing I. ricinus. These results therefore support the hypothesis that the rodent species studied are potential reservoirs for B. burgdorferi (s.l.). They are known to harbour B. japonica, B. afzelii, B. bissettii and the NT29 ribotype as well as the OspA serotype A of B. garinii [69].
Borrelia afzelii was found in all five small mammal samples. Studies from other European countries confirm that B. afzelii is a genospecies which is associated with rodents [70,71]. In Europe, MLST was performed for the identification and genotyping of Borrelia spp. in rodents from central Slovenia [72], questing I. ricinus ticks from Norway [73] and the UK [74], and ticks and rodents from France [75,76]. In Germany, the MLST method has thus far been used in research on phylogenetic relationships and global evolution of the B. burgdorferi (s.l.) species complex [77], and on the population structure and pathogenicity of B. afzelii and B. burgdorferi (s.s.) [78]. To our knowledge, this is the first study using MLST for the detection of allelic combinations of B. burgdorferi in small mammals from Germany. The analysis of the eight housekeeping genes, i.e. nifS, uvrA, clpA, clpX, rplB, recG, pyrG and pepX, revealed ST 165 and 559, both sequence types belonging to B. afzelii. These sequence types were described earlier in I. ricinus ticks from Latvia, Slovenia and France according to Borrelia spp. MLST database (http://pubmlst.org/bigsdb ?db=pubmlst_borrelia_isolates&page=profiles).
Rickettsiae of the spotted fever group may cause a variety of clinical symptoms such as lymphadenopathia, fever and headache in humans [79]. In Europe, there are several different species of varying pathogenic potential (R. aeschlimannii, R. conorii, R. helvetica, R. massiliae, R. monacensis, R. raoultii, R. sibirica and R. slovaca) [15]. In the present study, Rickettsia spp. were detected in all collected tick species (I. ricinus, I. trianguliceps and D. reticulatus). Results from France, the Netherlands, Austria and Poland showed infection levels in I. ricinus ticks ranging from 1.4 to 41% [80][81][82][83]. The prevalence obtained in the present study is in line with these findings. High infection rates (11-50%) for Rickettsia spp. in D. reticulatus were detected in previous investigations from the UK, Slovakia and Croatia [84][85][86]. The infection level in the present research is higher (70.5%), though not as high as in a previous study by our group (85.6%) which was conducted in the same study sites [27]. Transovarial and transstadial transmission of Rickettsia spp. have been described in ticks. Moreover, horizontal transmission during feeding on a bacteriaemic host and co-feeding of Rickettsia-positive arthropods were also demonstrated [87,88]. Dermacentor reticulatus is known to be the main vector of R. raoultii. As the prevalence in adult D. reticulatus ticks was very high but much lower in small mammals, it is probable that transovarial transmission is the main transmission path in D. reticulatus and that rodents are not of primary importance for the maintenance the natural circulation of R. raoultii.
The prevalence of Rickettsia spp. was significantly higher in D. reticulatus than in I. ricinus and in small mammals, pointing out that D. reticulatus-related rickettsiae are maintained independently from a vertebrate reservoir in nature, in contrast to I. ricinus-related rickettsiae. In Europe, there are very few studies about the maintaining and distribution of Rickettsia spp. in wild small mammals [32,33,81]. In Germany, two studies revealed the occurrence of R. helvetica in A. agrarius, A. flavicollis and M. glareolus [27,32,33]. In the present study, Rickettsia spp. was also found in these three rodent species. The study sites of the current research were earlier investigated for Rickettsia spp. by our group. These preliminary studies revealed high prevalences in D. reticulatus (56.7-85.6%), I. ricinus (13.4-17.5%), and small mammals (28.6%) [27,38]. The prevalence rates for Rickettsia spp. in the present study are in line with earlier findings for D. reticulatus (70.5%), however slightly lower for I. ricinus (11.4%) and small mammals (25.3%). In previous investigations on small mammals from Germany, R. felis, R. helvetica, R. monacensis and R. raoultii were detected [27,33]. Our results confirmed the occurrence of all mentioned Rickettsia spp. except for R. monacensis. All R. raoultii-positive rodents were infested with D. reticulatus, the main vector for R. raoultii. Interestingly, the D. reticulatus. This leads to the assumption that small mammals infected with D. reticulatus-related rickettsiae are rather incidental than potential reservoir hosts.

Conclusions
The prevalence for B. burgdorferi (s.l.) in small mammals was high (> 30%) and independent of tick abundance, suggesting small mammals as reservoirs. To our knowledge, this is the first detection of Borrelia spp. sequence types in small mammals from Germany, revealing ST 165 and ST 559 which belong to Borrelia genospecies B. afzelii. Small mammals may also serve as reservoirs for I. ricinus-transmitted Rickettsia spp. Bank voles (Myodes glareolus) had a seven times higher chance of encountering Rickettsia spp. infection while being infested with D. reticulatus in comparison to M. glareolus without D. reticulatus. As the prevalence in questing adult D. reticulatus was very high (> 70%) but much lower in rodents (c.25%), a potential reservoir function of bank voles is unlikely. The prevalence of R. raoultii in M. glareolus can be a result of infestation with infected D. reticulatus. We suggest that transstadial (and likely transovarial) transmission in D. reticulatus is the main mode of maintenance of R. raoultii natural life-cycle.

Acknowledgments
We thank Dana Rüster, Dietlinde Woll, Tessa Foerster, Dr. Anneliese Balling, Dr. André F. Streck and Rayan Ababneh for their excellent practical help during fieldwork and technical assistance. The work of MP and AO was done within the framework of COST action TD1303 EURNEGVEC. This paper has been sponsored by Bayer Animal Heath in the framework of the 12 th CVBD world forum symposium.

Funding
The Federal Environment Agency of Germany (FKZ 371148) funded part of this project.

Availability of data and materials
The data supporting the conclusions of this article are included within the article. The raw data used and/or analyzed during the current study are available from the corresponding author on reasonable request.