Contributions to the phylogeny of Ixodes (Pholeoixodes) canisuga, I. (Ph.) kaiseri, I. (Ph.) hexagonus and a simple pictorial key for the identification of their females

Background In Europe, hard ticks of the subgenus Pholeoixodes (Ixodidae: Ixodes) are usually associated with burrow-dwelling mammals and terrestrial birds. Reports of Pholeoixodes spp. from carnivores are frequently contradictory, and their identification is not based on key diagnostic characters. Therefore, the aims of the present study were to identify ticks collected from dogs, foxes and badgers in several European countries, and to reassess their systematic status with molecular analyses using two mitochondrial markers. Results Between 2003 and 2017, 144 Pholeoixodes spp. ticks were collected in nine European countries. From accurate descriptions and comparison with type-materials, a simple illustrated identification key was compiled for adult females, by focusing on the shape of the anterior surface of basis capituli. Based on this key, 71 female ticks were identified as I. canisuga, 21 as I. kaiseri and 21 as I. hexagonus. DNA was extracted from these 113 female ticks, and from further 31 specimens. Fragments of two mitochondrial genes, cox1 (cytochrome c oxidase subunit 1) and 16S rRNA, were amplified and sequenced. Ixodes kaiseri had nine unique cox1 haplotypes, which showed 99.2–100% sequence identity, whereas I. canisuga and I. hexagonus had eleven and five cox1 haplotypes, respectively, with 99.5–100% sequence identity. The distribution of cox1 haplotypes reflected a geographical pattern. Pholeoixodes spp. ticks had fewer 16S rRNA haplotypes, with a lower degree of intraspecific divergence (99.5–100% sequence identity) and no geographical clustering. Phylogenetic analyses were in agreement with morphology: I. kaiseri and I. hexagonus (with the similar shape of the anterior surface of basis capituli) were genetically more closely related to each other than to I. canisuga. Phylogenetic analyses also showed that the subgenus Eschatocephalus (bat ticks) clustered within the subgenus Pholeoixodes. Conclusions A simple, illustrated identification key is provided for female Pholeoixodes ticks of carnivores (including I. hexagonus and I. rugicollis) to prevent future misidentification of these species. It is also shown that I. kaiseri is more widespread in Europe than previously thought. Phylogenetic analyses suggest that the subgenus Pholeoixodes is not monophyletic: either the subgenus Eschatocephalus should be included in Pholeoixodes, or the latter subgenus should be divided, which is a task for future studies.


Background
Hard ticks (Acari: Ixodidae) are regarded as the most important vectors of pathogens [1]. Among them, the genus Ixodes Latreille, 1795 contains the highest number of species, exceeding 244 [2]. Traditionally, this genus was subdivided into subgenera, eight of which have representatives in the western Palaearctic [3]. The subgenus Pholeoixodes was erected [4] based on common morphological and ecological features of its members. For instance, the females of this subgenus have relatively short palps, there are no auriculae on the ventral basis capituli, and the first tarsi show a subapical dorsal hump [5]. Pholeoixodes species are usually associated with burrow-dwelling mammals, as well as terrestrial birds that nest in cavities (tree holes or burrows). Species of this subgenus usually feed on mammals, particularly carnivores (mainly Canidae, Mustelidae) and hedgehogs (Erinaceidae), in the western Palaearctic including I. canisuga Johnston, 1849, I. kaiseri Arthur, 1957, I. crenulatus Koch, 1844, I. hexagonus Leach, 1815 and I. rugicollis Schulze & Schlottke, 1929. Revised extensively by Babos [6], the systematics of the subgenus Pholeoixodes has yet to be fully understood. In particular, problems with identification and nomenclature led to the incorrect definition of geographical ranges. For instance, I. crenulatus was considered to occur in western, central and eastern Europe [7]. However, recent studies on ticks from carnivores confirmed its presence in Romania [8], but not in central and western Europe [9][10][11]. Furthermore, the validity of I. crenulatus was questioned because of its uninformative description that often results in it being misidentified as I. hexagonus [12]. Also, based on the morphological similarities of I. crenulatus and I. kaiseri, their synonymization was proposed [13], and later rejected [14].
Another controversy regarding the systematics and morphology of Pholeoixodes species is related to I. rugicollis. In the original drawing of the gnathosoma of I. rugicollis [23] (Fig. 1a), there was no indication of two frontal bumps on the anterior surface of basis capituli (between the basis of the hypostome and the palps). Later this species was even depicted without frontal bumps [7]. Furthermore, the drawings of the gnathosoma of I. rugicollis in its original description [23] and redescription [24] do not show broad separation of the porose areas, unlike what was reported in an electron microscopic study [25]. Ixodes rugicollis was also synonymized with I. cornutus [26], the porose areas of which are relatively close to each other [14].
In light of the above uncertainties, when contradictions arise with the traditional morphology-based identification of ticks, molecular comparison of representative specimens may provide additional and important clues to solve problems. Recently, North American members of the subgenus Pholeoixodes have been included in such molecular analyses [27]. However, no phylogenetic study has been performed on Pholeoixodes tick species collected from carnivores in Europe. Therefore, the aims of the present study were to morphologically identify female ticks collected from dogs, foxes and badgers in nine European countries, with subsequent molecular analyses using two mitochondrial markers. The study also aimed to clarify the taxonomic status and phylogenetic relationships of some Pholeoixodes species.

Sample collection and tick identification
Ticks were collected between 2003 and 2017. Specimens from Germany were stored frozen, and others were stored in 96% ethanol. Female ticks (except those from Serbia, from which the DNA was extracted previously) were examined morphologically with a VHX-5000 digital microscope (Keyence Co., Osaka, Japan). For species identification of females within Pholeoixodes, the following literature sources and type-materials were used: I. hexagonus [14], I. canisuga (neotypes loaned by the Natural History Museum, London, UK [NHM], accession numbers NHMUK 010305616-8: collected from dogs, UK), I. kaiseri ( [14,28]; paratype deposited in the United States National Ticks Collection [USNTC], accession number USNMENT00859298; and paratype loaned by NHM, accession number 1957.1.28.1: collected from fox, Egypt), I. crenulatus [14], I. rugicollis (syntype deposited in the USNTC, accession number USNMENT00865840: collected from pine marten, Germany). Adult ixodid ticks are usually easier to identify at the species level than immature stages, therefore morphological comparisons focused on adult female specimens. Nymphs were identified molecularly with PCR and sequencing two mitochondrial markers, which were compared to those of female ticks.

Molecular and phylogenetic analyses
DNA was extracted from the ticks (from part of the idiosoma and/or legs) individually with the QIAamp DNA Mini Kit (Qiagen, Hilden, Germany) according to the manufacturer's instructions, including an overnight digestion in tissue lysis buffer and Proteinase-K at 56°C. The cytochrome c oxidase subunit I (cox1) gene was chosen as the primary target for molecular analysis, on account of its suitability as a DNA-barcode sequence for tick species identification [29]. The PCR (modified from [30]) amplifies an approximately 710 bp long fragment of the cox1 gene. The primers HCO2198 (5′-TAA ACT TCA GGG TGA CCA AAA AAT CA-3′) and LCO1490 (5′-GGT CAA CAA ATC ATA AAG ATA TTG G-3′) were used in a reaction volume of 25 μl, containing 1 U (0.2 μl) HotStarTaq Plus DNA polymerase (Qiagen), 2.5 μl 10× CoralLoad Reaction buffer (including 15 mM MgCl 2 ), 0.5 μl PCR nucleotide Mix (0.2 mM each), 0.5 μl (1 μM final concentration) of each primer, 15.8 μl ddH 2 O and 5 μl template DNA. For amplification, an initial denaturation step at 95°C for 5 min was followed by 40 cycles of denaturation at 94°C for 40 s, annealing at 48°C for 1 min and extension at 72°C for 1 min. Final extension was performed at 72°C for 10 min.
To complement the results obtained with the cox1 gene, all samples that showed different cox1 haplotype within a country, were also tested for another mitochondrial marker. This PCR amplifies an approximately 460 bp fragment of the 16S rRNA gene of Ixodidae [31], with the primers 16S + 1 (5′-CTG CTC AAT GAT TTT TTA AAT TGC TGT GG-3′) and 16S-1 (5′-CCG GTC TGA ACT CAG ATC AAG T-3′). Other reaction components and cycling conditions were the same as above, except for annealing at 51°C. PCR products were visualized in a 1.5% agarose gel. Purification and sequencing were done by Biomi Inc. (Gödöllő, Hungary). If identical sequences were found in a country, only one representative sequence was submitted to GenBank (accession numbers for the cox1 gene: KY962011-KY962051; for the 16S rRNA gene: KY962052-KY962077). Positions of nucleotide differences according to haplotypes are provided in Tables 1, 2. For comparison and phylogenetic analyses, the sequences were trimmed to the same length (cox1: 631 bp, 16S rRNA gene: 402 bp). Tick species from other studies were included in the phylogenetic analyses only if their sequence(s) available in GenBank had 99-100% coverage with the sequences in this study. This dataset was resampled 1000 times to generate bootstrap values. Phylogenetic analyses were conducted with the Maximum Likelihood method by using MEGA version 6.0. The MEGA model selection method was applied to choose the appropriate model (GTR and Jukes-Cantor model for cox1 and 16S rRNA genes, respectively). The ratio of haplotypes between different geographical regions was compared by Fisher's exact test (condition of significance: P < 0.05).

Morphological identification of female ticks
Altogether 113 female ticks (all Pholeoixodes spp.) were compared morphologically with adequate descriptions, and, when available, with type material. For their species identification, the following simplified keys were compiled, taking into account distinctive cardinal criteria observed in the present study.
Key to the females of Pholeoixodes spp. of carnivores in Europe Pronounced and pointed bumps on anterior surface of basis capituli between hypostome and palps, palps "stalked" and laterally curved (convex); space between porose areas more than twice as wide as diameter of porose area (Fig. 1)    represented by seven and five individuals, respectively, while the remaining nine were found only once. These haplotypes showed up to three nucleotide differences from each other, corresponding to 99.5-100% sequence identity (628-631/631 bp). The distribution of some haplotypes reflected a clear geographical pattern (Table 1). Haplotype "H" was significantly more frequently identified in samples from western Europe than from central and south-eastern Europe (P < 0.0001). Several other haplotypes were identified only in one country (e.g. "B" in Hungary; "C", "D" and "K" in Romania; "E", "G" and "I" in Croatia; "F" in France; "J" in UK). However, haplotype "A" occurred in all evaluated regions of Europe (Table 1). Ticks identified as I. kaiseri had nine ("L to T") cox1 haplotypes, which showed a higher rate of polymorphism compared to I. canisuga, i.e. up to five nucleotide differences from each other (626-631/631 bp = 99.2-100% sequence identity). The occurrence of these haplotypes was restricted to central and south-eastern Europe (Table 1). Haplotypes "M" and "O" were unique to Germany, whereas the others occurred in Hungary and south-eastern Europe (Table 1). Ticks identified as I. hexagonus had five ("U to Y") cox1 haplotypes, which showed up to three nucleotide differences from each other, i.e. 99.5-100% sequence identity (628-631/631 bp).
Haplotype "X" was only identified in Germany and Austria, whereas all the others were present in southeastern Europe ( Table 1).
The phylogenetic relationships of cox1 and 16S rRNA haplotypes are shown in Figs. 6 and 7, respectively. Morphological identification of the three species was supported by the phylogenetic analyses, because all morphologically a priori identified individuals of the three Pholeoixodes species grouped in the phylogenetic trees (Figs. 6, 7). The topologies of both phylogenetic trees reflect that (based on the investigated sequences) the subgenus Pholeoixodes is not monophyletic. Isolates of I. kaiseri and I. hexagonus formed two sister groups, whereas samples of I. canisuga were more closely related to the bat tick species I. vespertilionis, I. ariadnae and I. simplex (subgenus Eschatocephalus). Thus, the phylogenetic group of Pholeoixodes spp. also contained the clade of Eschatocephalus spp. (Figs. 6, 7).

Discussion
This is the first comprehensive phylogenetic analysis involving three tick species from the subgenus Pholeoixodes collected from carnivores in Europe. In the present study, the morphological identification of I. canisuga and I. hexagonus was confirmed with molecularphylogenetic methods, validating the morphological characters selected and described in the above taxonomic key. This identification key highlights differences among Pholeoixodes spp. in the shape of the anterior surface of basis capituli (between the hypostome and the palps). While the shape of the anterior surface of basis capituli has been reported to be a distinctive character between various Ixodes spp., including I. canisuga and I. hexagonus [32], this character has not been incorporated into the identification keys provided by other authors (e.g. [20]). These data suggest that the shape of the anterior surface of basis capituli should always be observed for the morphological differentiation of female Pholeoixodes ticks of carnivores (i.e. I. canisuga vs I. hexagonus and I. kaiseri).
The data presented here expand the known geographical range of I. kaiseri in Europe. Recently, ticks resembling I. kaiseri were reported in Poland [33]. Although I. kaiseri was previously reported from Romania [34], this was not confirmed by later studies involving or reporting ticks of carnivores [8,[35][36][37]. Here, the evidence is provided for the occurrence of I. kaiseri in Romania, as well as in three other countries, where it had not been reported yet (Germany, Hungary and Serbia). These new records are probably not a consequence of recent emergence of this tick species in new regions, but rather that its specimens have hitherto been misdiagnosed. Therefore, the keys presented here will be useful for future studies, which will help to define the geographical range of I. kaiseri.
Ixodes crenulatus was not found in the present study, not even in Romania, where it has been recently reported to occur [8]. On the other hand, reports of its occurrence in western Europe (Ireland, UK, Germany) are more than half a century old [7], and require verification. The diagnosis of this species is difficult, because the redescription (although adequate) is not easily accessible and is written in Russian [14], and no typespecimen is available [2]. The most important diagnostic feature of I. crenulatus females, the longitudinal ridges on the basis capituli [14] bear a resemblance to the ridge (rounded bumps) of the plateau seen in I. canisuga (Fig. 5), therefore further studies involving both species will be needed to reassess their synonymy, which was proposed by some authors [2].
Similarly, based on the inadequacy (and contradictions) of descriptions, illustrations in some former and recent reports on I. rugicollis, the relevant data should be interpreted with caution, taking into account the type specimens and key of females presented here (in Romania [7]: porose areas different, frontal bumps absent; in Poland [38]: porose areas not shown, frontal bumps are rounded). Also, I. rugicollis has recently been identified in Hungary [39] in part according to the description of I. cornutus [14], which was regarded as a synonym of I. rugicollis [26]. However, based on the type specimens investigated here and the identification criteria presented in the key, this synonymy cannot be maintained.
Phylogenetic analyses performed here suggest that the subgenus Pholeoixodes is not monophyletic. Taking into account that in both the cox1 and 16S rRNA gene phylogenetic analyses I. canisuga specimens (and I./Ph. lividus) formed one clade with bat ticks of the subgenus Eschatocephalus, it is necessary to test the parsimony of the inclusion of the latter subgenus in Pholeoixodes. Alternatively, Pholeoixodes should be divided into two subgenera. These data, on the other hand, confirm the phylogenetic relevance of morphological traits, because these bat tick species lack auriculae and a b Fig. 5 The similarity of Ixodes canisuga females to I. crenulatus. Basis capituli with dorsal ridge (between arrows) in case of two identical haplotypes ("A") of I. canisuga, which thus resemble I. crenulatus. a Female from Bosnia and Herzegovina. b Female from Hungary long/pointed internal spur on coxa I, similarly to most Pholeoixodes spp.
The results presented here should also be considered in a geographical context. Although geographical structuring of several (but not all) cox1 mitochondrial lineages was observed, sequence and phylogenetic analyses of the 16S rRNA gene did not reflect the same pattern. These findings suggest that genetic exchange within each Pholeoixodes species is not limited between different European populations investigated here, i.e. these tick species are not subdivided into geographically distinct populations, but undergo constant gene flow. One underlying reason may be that populations of an important host species, the red fox (Vulpes vulpes) are apparently in contact and mix throughout Europe, allowing gene flow, with little spatial structuring [40]. The emergence of golden jackals (Canis aureus) towards western Europe [41] may have also contributed to the dispersal of tick species investigated here. This could also counterbalance genetic differences that would have otherwise resulted from separation of populations of the main hosts (hedgehog species) of I. hexagonus (i.e. Erinaceus europaeus in Germany vs E. roumanicus in Hungary and south-eastern Europe).

Conclusions
Reference sequences are provided for I. canisuga and I. kaiseri based on female specimens determined according to simple identification keys (including I. hexagonus and I. rugicollis), to prevent future misdiagnoses of these species. These results confirm that the shape and the morphological features of the anterior surface of basis capituli, the details of spurs on coxa I and the relative width of the scutum of females are important characters for species identification among Pholeoixodes ticks of carnivores. It is also demonstrated that I. kaiseri is more widespread in Europe than previously thought. Based on phylogenetic analyses, the subgenus Pholeoixodes is not monophyletic, as the subgenus Eschatocephalus clustered within its clade. This should be further elaborated by future taxonomic studies.