Spotted Fever Group Rickettsiae in Ticks and Small Mammals from Grassland and Forest Habitats in Central Germany

Rickettsiae of the spotted fever group (SFG) are zoonotic tick-borne pathogens. Small mammals are important hosts for the immature life stages of two of the most common tick species in Europe, Ixodes ricinus and Dermacentor reticulatus. These hosts and vectors can be found in diverse habitats with different vegetation types like grasslands and forests. To investigate the influence of environmental and individual factors on Rickettsia prevalence, this study aimed to analyse the prevalence of SFG rickettsiae in ticks and small mammals in different small-scale habitats in central Germany for the first time. Small mammals of ten species and ticks of two species were collected from grasslands and forests in the Hainich-Dün region, central Germany. After species identification, DNA samples from 1098 ticks and ear snips of 1167 small mammals were screened for Rickettsia DNA by qPCR targeting the gltA gene. Positive samples were retested by conventional PCR targeting the ompB gene and sequencing. Rickettsia DNA was detected in eight out of ten small mammal species. Small mammal hosts from forests (14.0%) were significantly more often infected than those from grasslands (4.4%) (p < 0.001). The highest prevalence was found in the mostly forest-inhabiting genus Apodemus (14.8%) and the lowest in Microtus (6.6%), which inhabits grasslands. The prevalence was higher in D. reticulatus (46.3%) than in the I. ricinus complex (8.6%). Adult ticks were more often infected than nymphs (p = 0.0199). All sequenced rickettsiae in I. ricinus complex ticks were R. helvetica, and the ones in D. reticulatus were R. raoultii. Unlike adults, questing nymphs have had only one blood meal, which explains the higher prevalence in I. ricinus adults. Interestingly, habitat type did influence infection probability in small mammals, but did not in ticks. A possible explanation may be the high prevalence in Apodemus flavicollis and A. sylvaticus which were more abundant in the forest.


Introduction
Hard ticks are haematophagous arthropods that can be found in diverse habitats with different kinds of vegetation. While feeding, they may serve as vectors for multiple tick-borne pathogens (TBPs) like protozoan parasites, bacteria, and viruses. The most common tick species in central Europe is the castor bean tick Ixodes ricinus [1]. It can be found in various environments, from forests over grasslands to urban areas [2]. As three-host non-nidicolous ectoparasites, immature I. ricinus ticks (larvae and nymphs) feed

Study Sites
In total, 21 study sites surrounding the Hainich-Dün National Park in Thuringia, central Germany were examined. The area is one of the largest continuous deciduous forests in Germany, with European beech (Fagus sylvatica) being the dominant tree species. The region is characterised by forest and agricultural areas which are cultivated to different extents. While some forests and grasslands are extensively managed, the forests of the National Park itself are protected (https://www.biodiversity-exploratories.de/en/regions/ hainich-duen/, accessed on 25 April 2023). A more detailed description of the study area was given in Król et al. [32].

Tick Collection
A total of 1115 questing ticks were collected for a previous study by flagging 100 m 2 simultaneously with small mammal trapping at 17 of the 21 sites, once per season (spring, summer, and autumn) in 2018 and 2019 [32] ( Figure A1). The study sites for ticks were composed of one plot in the forest and one in the bordering grassland-forest ecotone, as described before [32]. Collected ticks were specified for sex, life stage, and species under a light microscope (Motic ® SMZ-171, Moticeurope, S.L.U., Barcelona, Spain) according to taxonomic keys [33,34]. DNA was extracted using a QIAmp DNA Mini Kit (Qiagen, Hilden, Germany) according to the manufacturer's instructions for DNA isolation. Extracted DNA of 1094 ticks was available from the study of Król et al., where further details of tick handling, taxonomic identification, and DNA extraction procedure are described [32]. In addition to the given dataset, four ticks (three nymphs and one female of I. ricinus) that were all collected in spring 2019 were processed after the same protocol. Seventeen individuals (four I. ricinus nymphs and thirteen D. reticulatus adults) had to be excluded from further molecular biological analyses due to insufficient material conservation. Data on I. ricinus and I. inopinatus (presumably 16 individuals) were merged under the terminus "I. ricinus complex" [35].
In total, 1115 ticks belonging to 2 species were collected (Table A1). The most prevalent species was the I. ricinus complex (91.7%), followed by D. reticulatus (8.3%). The most frequently found life stages were the nymphs of the I. ricinus complex (74.9%). Most ticks were flagged in spring (74.1%) in comparison to summer (18.3%) and autumn (7.6%). In ecotones less ticks (35.2%) were flagged than in forests (64.8%). After excluding the abovementioned ticks, 1018 ticks of the I. ricinus complex and 80 individuals of D. reticulatus were further processed (Table 1).

Small Mammal Collection
Extracted DNA derived from small mammal skin samples was available from a former study [32]. The samples were taken from about 0.5 cm × 0.5 cm ear tissue. DNA extraction was performed using the same methodology as for the ticks. Small mammals were trapped at 21 study sites, each in a paired system of a forest plot and an adjacent grassland plot per site [32] ( Figure A1).

Real-Time PCR, Conventional PCR and Sequencing
Quantitative real-time PCR (qPCR) was used to screen all tick and small mammal samples for a 70 base pair (bp)-sized region of the citrate synthase gene (gltA) of Rickettsia spp. The mix was prepared as described previously [38] using the LightCycler ® FastStart DNA Master HybProbe (Roche Diagnostics GmbH, Mannheim, Germany). The cycling protocol for the Thermocycler (Stratagene Mx3000P, Agilent, Santa Clara, CA, USA) included 95 • C for 10 min followed by 45 cycles of denaturation at 95 • C for 25 s, annealing at 55 • C for 30 s, and extension at 72 • C for 30 s.
Tick samples tested with a cycle threshold (ct) < 36 and small mammal samples tested with a ct < 35 were subsequently examined by conventional PCR targeting a fragment of the gene encoding for the outer membrane protein B (ompB, 811bp) of the SFG rickettsiae. A previously described protocol [39] was followed with the primers "120-2788" and "120-3599" with one adjustment of the initial denaturation temperature of 94 • C.
For visualisation of the PCR products, gel electrophoresis was performed-8 µL of the samples were mixed with 2 µL of loading dye (TriTrack DNA Loading Dye (6×), Thermo Scientific™, Waltham, MA, USA) and separated on a 1.5% agarose gel.
PCR products were prepared for sequencing using the NucleoSpin ® Gel and PCR Clean-up kit (Macherey-Nagel GmbH & Co. KG, Düren, Germany) according to the protocol recommended by the manufacturer. Sequencing was performed commercially by Eurofins Genomics Germany GmbH (Ebersberg, Germany). After editing and aligning the sequences with Bionumerics (Applied Maths NV, Sint-Martens-Latem, Belgium), a comparison was conducted to sequences present in GenBank on the Basic Local Alignment Tool (BLAST; https://blast.ncbi.nlm.nih.gov/Blast.cgi, accessed on 9 March 2023). Two obtained sequences from small mammals and 49 obtained sequences from ticks from 2019 were submitted to GenBank under the accession numbers OQ694692-OQ694742.

Statistical Analysis
For Rickettsia spp. prevalence of ticks and small mammals, a confidence interval (95% CI) was formulated using the Clopper and Pearson method with Graph Pad software (GraphPad Software, San Diego, CA, USA). For small mammals, a generalised linear mixed model (GLMM) with binomial error distribution was computed with R software (version 4.1.2. for Windows; RStudio, Boston, MA, USA) and the lme4 package. It was used to investigate the dependence of individual Rickettsia infection status (dependent binary variable; Rickettsia spp. positive = 1; Rickettsia spp. negative = 0) in relation to season (independent binary variable: summer vs. spring), habitat (independent binary variable: forest vs. grassland), and small mammal species (independent categorical variable) [40]. Because of low numbers of trapped individuals, the following five species were excluded from GLMM analysis: M. agrestis, A. amphibius, S. araneus, S. minutus, and C. russula.
In addition, GLMMs were generated for A. flavicollis, Cl. glareolus, and M. arvalis separately. Independent variables were sex of the individual (independent binary variable: female vs. male), habitat (see above), and season (see above).
For ticks, a slightly different model was calculated. The dependent variable was the binominal "presence of Rickettsia". The independent variables were (i) developmental stage (binary: adult vs. nymph), (ii) habitat (binary: forest vs. ecotone), (iii) season (three stages: autumn, spring, and summer), and (iv) tick species (binary: I. ricinus complex vs. D. reticulatus). The first three independent variables were used for a model with exclusively I. ricinus complex ticks. Due to the low number of individuals collected, the model was not carried out for D. reticulatus ticks.
The interaction term for the GLMM included three and four independent variables with at least two levels each for small mammals and I. ricinus complex ticks, and ticks, respectively. To state potential differences of all variables separately, marginal means accessing the emmeans package within R and a post hoc test were computed. The significance threshold for p was set at ≤0.05.

Co-Infections of Borrelia spp. and Rickettsia spp. in Ticks and Small Mammals
Results of detection of Borrelia spp. DNA in small mammals (1167 individuals) and tick samples (1094 individuals) that were also investigated in this study have been published in an earlier study [32]. In the current study, co-infections of Borrelia spp. (previous study) with Rickettsia spp. (current study) were analysed. To investigate the degree of cooccurrence, the approach proposed by Malini et al. was applied [41]. Here, the R package CooccurrenceAffinity was applied to calculate a maximum likely estimator to evaluate whether or not the pairing of Rickettsia spp. and Borrelia spp. was more or less likely than the expected prevalence [42]. Seasonal analyses were made for ticks and small mammals separately, and all trapped individuals regardless of species were taken into account.
The GLMM with the factors of season and habitat (Table A3) revealed statistical significance for the lower probability of infection of an individual of Cl. glareolus compared to A. agrarius (p = 0.034). No statistical significance was observed when only the two species were compared and no other factors were taken into account (p = 0.206). Clethrionomys glareolus was less often infected with Rickettsia spp. than both A. flavicollis (p = 0.007) and A. sylvaticus (p < 0.001).
In the GLMMs computed individually for the three most abundant species, sex and season did not influence the prevalence in M. arvalis, Cl. glareolus and A. flavicollis (Table A3). Out of these three species, habitat mattered only for M. arvalis. The prevalence was higher in forests, but 92.6% of the animals of that species were trapped in grasslands (p < 0.001) ( Table 2).
Although excluded from the analysis with GLMM, this is the first detection of Rickettsia spp. DNA in S. minutus in Germany to our knowledge.
Of the 115 individuals that were tested positive in qPCR, 24 yielded a ct-value <35. Only two of them (both A. flavicollis) yielded a result in the ompB gene. Both amplicons could be sequenced and were 100% identical to R. helvetica strain AS819 (deposited in GenBank: MF163037), which was isolated from an I. ricinus tick [43] (Table 3).

Discussion
Rickettsiae of the SFG are pathogens of public health concern. Studies, including analyses of Rickettsia prevalence in both ticks and small mammal hosts regarding various aspects, like particular habitats and the possible influence on each other, are largely overlooked. This study presents, for the first time, data on Rickettsia species and their prevalence in ticks and small mammals in central Germany with different habitat features.
In the current study, the most prevalent tick species was Ixodes ricinus, which is the most abundant tick species in Germany [45]. Ixodes ricinus ticks in central Europe have a bimodal activity pattern peaking in warm and humid months, with a larger peak in spring and a smaller peak in autumn. In the Mediterranean area, a unimodal activity pattern with one big peak in spring is common, which is now also more often observed in central Europe, as shifts in temperature and precipitation result in a decline in the impact of microclimatic conditions like sufficient humidity [46,47]. In our study, we also did not find an activity peak in autumn, but observed a unimodal activity pattern with a peak in spring for nymphs and adults. The ratio of I. ricinus complex adult ticks collected in ecotones and forests was similar, fitting their described primary habitat of forests and shrubbery [48]. In all three considered seasons, I. ricinus complex nymphs were more often collected in the forest compared to the ecotone. One previous study from southern Germany showed a positive effect of humidity on the occurrence of nymphs [46]. In the forest areas investigated, a base layer with higher humidity serves as protection from desiccation during hotter months, explaining why more ticks were collected in the forest sites compared to the ecotone [2].
Less than a tenth of the collected ticks belonged to the species Dermacentor reticulatus [45]. Larvae and nymphs of D. reticulatus commonly show nidicolous behaviour and live in burrows and nests of small mammals [33]. This is why only adult individuals were collected in the current study. Dermacentor reticulatus ticks show a quiescence phase over summer and two activity peaks, one in spring and one in autumn [24,49], which we observed in our study. Most D. reticulatus ticks (93.6%) were flagged in the ecotone which is in concordance with the described natural habitat of bushy pastures, meadows, and open forests [48,49].
As expected, most Clethrionomys glareolus (97.1%) and Apodemus flavicollis (91.7%) were trapped in the forest. Opposite to that, Microtus arvalis was mostly trapped in grasslands (92.6%). This distribution matches the preference of Cl. glareolus and A. flavicollis for forests and M. arvalis for grasslands [50,51]. Apodemus sylvaticus was trapped with a proportion of 74.1% in forests, which reflects their well-known common habitat shift between forests and grasslands [7,52]. Proportions of trapped A. agrarius, Sorex araneus, and M. agrestis were balanced in forest and grassland. These three species do not show such a strong link to either one of the two considered habitats [7,51,53].
Small mammals serve as reservoirs for many zoonotic pathogens including rickettsiae of the SFG group of diverse pathogenicity. In Germany, most known severe human cases were, however, not autochthonous [54]. The two most common species in Germany, Rickettsia helvetica and R. raoultii, are nowadays known to cause clinically unspecific symptoms such as eschars in humans [24,55].
In central Europe, DNA of R. helvetica and R. raoultii has been detected in arthropods such as fleas and various tick species like I. ricinus, D. reticulatus, and D. marginatus. Animals in which rickettsial DNA has been detected include rodents, racoons (Procyon lotor), roe deer, wild boars, and lizards [16,20,27,56,57].
To our knowledge, this is the first detection of Rickettsia spp. DNA in S. minutus from Germany. Rickettsia spp. DNA could not be recorded in a former study from Germany, including 72 individuals of S. minutus, of which 16 likewise originated from the state of Thuringia [58]. A study from Norway has shown a lower infestation rate of S. minutus with ticks compared to investigated individuals from the genus Apodemus [59], which has also been displayed for shrews in a study from France [52]. This, in turn, might lead to a lower risk of infection with TBPs. As in our study, Rickettsia spp. DNA has been detected in S. araneus in other studies before [58,60].
Interestingly, individuals in forests were more often infected than those in grasslands. Among the investigated genera, Apodemus had the highest prevalence. As described above, the two most abundant species of the genus Apodemus, namely, A. flavicollis and A. sylvaticus, were mostly trapped in the forest. Apodemus spp. showing a higher prevalence than other small mammals has been described before. The prevalence in A. sylvaticus (18.5%), A. flavicollis (15.4%), and A. agrarius (8.9%) falls in line with previous findings from Germany with prevalence ranges of 0-16.2% [58,61], 13.0-23.4% [58,61,62], and 0-9% [58,62], respectively. A few studies on A. flavicollis from several European countries showed prevalence rates of 0% in whole blood, 1.7% in organ samples, 5.7% in whole blood, and 29.4% in spleens in Poland [63], Croatia [64], Slovakia [65], and Lithuania [60], respectively. One study from Italy found a prevalence of 6% in ear pinna samples without distinguishing between Apodemus species [66].
Difficulties in sequencing of Rickettsia spp. DNA in material of small mammal origin due to poor sensitivity of conventional PCR has been described before [58]. In our investigations, only two out of one hundred and fifteen small mammal samples positive in qPCR (24 of which had a ct < 35) could be sequenced. Both were from A. flavicollis samples and showed the highest similarity to R. helvetica strain AS819, which has been isolated from I. ricinus ticks before. Rickettsia helvetica has been found in A. flavicollis in several studies from Germany before [20,58]. One study from the Netherlands found R. helvetica in small rodents [16]. Rickettsia raoultii, on the other hand, seems to be rather rare in small mammal samples [58]. One study found R. raoultii in Cl. glareolus which were all infested with D. reticulatus [62]. A larger number of sequencing results in our study could have added more certainty as to whether small mammals, in particular Apodemus spp., take a less essential part in the life cycle of R. raoultii. One study from southern Germany investigated Dermacentor ticks and mainly voles without detecting any Rickettsia spp. DNA in rodent samples [67]. Two studies from China examined other mammals as potential reservoirs of R. raoultii and detected R. raoultii in horses (Equus ferus) and red foxes (Vulpes vulpes) [68,69]. Rickettsia helvetica in roe deer and wild boar and R. slovaca in wild boar were found in studies from the Netherlands [16] and Algeria [70], respectively.
Rickettsia spp. prevalence in tick species differed from 8.6% in the I. ricinus complex and 46.3% in D. reticulatus. Dermacentor reticulatus often shows a higher infection rate with Rickettsia than I. ricinus [71]. For example, a study from northeast Germany noted a prevalence of 64.0% in D. reticulatus [49]. The determined prevalence in the I. ricinus complex in our study conforms to prevalences described in other studies on I. ricinus from Germany [21,72,73]. However, the methods and life stages of ticks differed among the published studies; therefore, direct comparison of prevalence has to be regarded with caution. Adult ticks showed a statistically higher infection risk than nymphs, which have had only one blood meal as larvae. Nevertheless, it has to be noted that we only analysed nymphs from the I. ricinus complex and no immature stages from D. reticulatus. In a study also analysing feeding immatures, it has been proposed that transovarial transmission of Rickettsia spp. in ticks plays a more important role in D. reticulatus than in I. ricinus [20]. Success of transovarial and transstadial transmission of R. raoultii in D. reticulatus have been presented to be 90.0% and 98.0%, respectively [74]. This supports the observation in our study that D. reticulatus has a higher prevalence than I. ricinus. Also, their preference for voles as hosts, which had a comparably low prevalence in our study, might reinforce this approach.
Regarding only I. ricinus ticks, the infection risk was higher in adults than in nymphs. This has been observed in most but not all studies examining exclusively I. ricinus [31,72,75]. In spring and summer, ticks were significantly more often infected than in autumn. In the statistical analysis regarding solely I. ricinus ticks, no seasonal influence could be noted. In other European studies, in which only I. ricinus ticks were considered, the seasonal influence differed. Overall, no pattern can be noted as in some studies from Germany ticks collected in summer and autumn had higher infection rate than in spring [19,76], and in another study ticks collected in summer had a higher infection rate than in autumn and spring [72]. In opposite to that stands the finding from a study from Denmark in which ticks flagged in spring were significantly more often infected than ticks from summer and autumn [77]. In one study analysing only D. reticulatus from northeast Germany, no seasonal influence on prevalence was recognised [49].
Habitat type did not influence the prevalence of Rickettsia in ticks in our study. Other studies in Europe also did not find any statistical effect of different landscapes on Rickettsia prevalence in I. ricinus ticks [29][30][31]. It seems that microclimatic factors may play a more important role for Rickettsia abundance in ticks than the habitat itself. A study design of field and experimental studies taking the microclimate into account could be useful to identify additional driving parameters for tick infection.
Numbers of co-infections with both Borrelia spp. and Rickettsia spp. in small mammal species were low, except in S. araneus (10%). The co-infection rates in ticks with these two bacteria were low in our study but have been shown to correlate positively before [78]. Another study with a higher prevalence of both bacteria than in our study did not show this positive association [79]. In our study, the only observed effect was a slight tendency of a negative association of Rickettsia spp. infection with Borrelia spp. occurrence in ticks in spring. Studies investigating co-infections in small mammals are scarce. As S. araneus was the only small mammal species with a high co-infection rate in our study and was rarely regarded in previous studies, further research focusing on this correlation should be considered.
In our study, 68.8% of the tick samples that were positive in qPCR could be amplified by conventional PCR, sequenced and identified as either R. helvetica or R. raoultii. Rickettsia helvetica is the most abundant Rickettsia species in Germany, as in our study. Interestingly, it was the only I. ricinus associated Rickettsia species that was determined. Even though R. monacensis detection is not that rare in Germany, it seems to occur more often in southern Germany [31,72,73,76]. The R. helvetica sequences from our study were identical to those previously found in I. persulcatus from Russia and in I. ricinus from Germany. Rickettsia raoultii sequences found here had been detected in China earlier. Compared to other tick species, R. raoultii has a strong link to D. reticulatus [9] and was also the only Rickettsia species we could detect in D. reticulatus in our study. As I. ricinus and D. reticulatus are the two most common tick species in central Europe, their monitoring is one important tool to gain knowledge about risk factors for human and animal infections.

Conclusions
The prevalence of Rickettsia spp. in small mammals and ticks determined in this study falls in line with previous studies from Germany. Adult ticks had a statistically higher infection risk than nymphs. Dermacentor reticulatus showed a significantly higher prevalence than I. ricinus ticks. In our study, we found R. raoultii and R. helvetica, which are both associated with cases of human illness. Interestingly, this study showed no influence of habitat type on the prevalence of Rickettsia in ticks but in small mammals, which were significantly more often infected in forests than in grasslands. A possible explanation may be the high prevalence in small mammals of the genus Apodemus, which are more abundant in forests. Nevertheless, habitat type should always be considered in a one health perspective, as it has a massive impact on the abundance of potential pathogen reservoirs.       Table A5. Co-infections of Borrelia spp. and Rickettsia spp. in small mammals. Data of Borrelia spp. investigation were available from Król et al. [32]. No.: Number. Table A6. Results from the co-occurrence seasonal analysis for all trapped ticks and mammals. Alpha maximum likelihood estimator (MLE) can be interpreted as an odds ratio for co-occurrence. The accompanying confidence interval (CI) was calculated according to Blaker [80].