Leaving the tropics: The successful colonization of cold temperate regions by Dolicheremaeus dorni (Acari, Oribatida)

Abstract Species diversity is generally higher in the tropics compared to the temperate zones. The phenomenon that one species of an almost exclusively tropical living genus was able to adapt successfully to the cold northern regions is rather rare. However, the oribatid mite Dolicheremaeus dorni represents such a species and is in the focus of this study. While 180 Dolicheremaeus species are confined to the tropics and subtropics, only five species are known to occur in temperate climates and D. dorni represents the only species with a wider distribution in this climatic region. This species is distributed in Central and Southern Europe and was now recorded for the first time in Austria. A morphological and molecular genetic investigation of specimens from Austria, Poland and Croatia confirmed this distribution pattern and revealed specific geographic clades and haplotypes for each population and hence indicate low gene flow between populations. A further molecular genetic analysis of the 18S rRNA gene sequence of D. dorni confirmed its phylogenetic position within Carabodoidea. Based on record information, this species is associated with trees or tree habitats and seems to be rather a generalist than a specialist for a specific substrate (e.g., tree species) or food source.

With more than 180 species, Dolicheremaeus is the most diverse genus of the family Otocepheidae, which includes 39 described genera (Norton & Behan-Pelletier, 2009). The preferred habitats of these taxa are wet decaying, spongy woods in tropical regions. As a possible evolutionary adaptation to the high moisture and heavy rainfall present in the tropics, adults of this genus have a special feature allowing to breathe under flooded conditions, namely respiratory taenidia of a type more commonly found in (semi-)aquatic oribatid mites (Norton & Behan-Pelletier, 2009;Trav e, 1986). Given the large variety of Dolicheremaeus species and their more or less exclusive occurrence in tropic areas of the world, cold temperate European biota seem to be unfavorable for these organisms. However, there is one species, namely Dolicheremaeus dorni (Balogh, 1937), which has been found sporadically in some, mainly more southern, European countries, for example, Greece, Montenegro, Southern Romania, and Southern France (Balogh, 1937;Mahunka, 1982;Tarman, 1977;Trav e, 1986). Additionally, Bulanova-Zachvatkina (1967) described a species, Dolicheremaeus georgii, from the Trans-Carpathians which is morphologically clearly distinct from D. dorni, but since its original description, there were no more findings of this species. Accordingly, D. dorni represents the only species of this genus showing a wider distribution in European regions.
Despite there is a huge number of otocepheid species described (more than 400, following the classification of Norton & Behan-Pelletier, 2009), no barcoding sequences of the mitochondrial cytochrome c oxidase subunit I gene (COI) are available. However, there are seven sequences of different nuclear markers representing four species of Otocepheidae recorded in GenBank.
In this study, we investigate the geographic distribution and the population structure of D. dorni and discuss the unusual occurrence of this species in non-tropical areas. As there is a limit of genetic data in the Otocepheidae, we additionally provide the first COI sequences of D. dorni and a second otocepheid species Spinotocepheus sp., and present their phylogenetic placement within the Oribatida by the use of the standard nuclear small subunit rRNA (18S rRNA) gene. Furthermore, we integrate a short redescription including leg drawings to the manuscript.

| Sampling
In total, 49 individuals of D. dorni were assayed in this study, whereof all were firstly used for a (at least rough) morphological investigation including body size measurements. Afterward, 14 individuals were used for genetic analyses. All of them were analyzed for a fragment of the COI gene for intraspecific studies. Furthermore, one single individual of an undetermined Spinotocepheus species was analyzed with the same methods as mentioned before for D. dorni individuals. To study the phylogenetic placement, we sequenced part of the 18S rRNA gene (18S) of one D. dorni and the Spinotocepheus individual too. These were then aligned together with 54 oribatid mite 18S sequences from GenBank, including all available sequences of possible sister taxa/groups according to the classification scheme of Norton and Behan-Pelletier (2009). Species of the supercohort Palaeosomatides are generally considered as the most primitive Oribatida group; therefore, we decided to use them as outgroup taxa.
Detailed information on herein investigated individuals is shown in Table 1. Individual data obtained from GenBank are given in the Table A1.
For morphological comparisons, we used the specimens described by Weigmann (2014) Total genomic DNA of single individuals was extracted by means of the rapid Chelex 100 resin protocol described in Richlen and Barber (2005). Body remnants (cuticle structures) of all investigated specimens were kept and frozen for a later preparation of permanent slides serving as vouchers. All voucher specimens are deposited in the mite collection at the Institute of Zoology, University of Graz (voucher IDs are same as sample IDs; see Table 1).
Also, PCR amplification of the 18S sequences was performed in two overlapping fragments of approximately 950 and 1500 bp length each, using the same protocol and primer pairs (18Sfwd/ rev960 and fw1230/rev18S) as described in Dabert, Witalinski, Kazmierski, Olszanowski, and Dabert (2010).
Purification of all PCR products and DNA sequencing followed standard protocols as described in Sch€ affer et al. (2008) using same primers as for PCR amplification. In case of 18S sequences, two additional internal sequencing primers were used (fw390 and fw770; Dabert et al., 2010). DNA fragments were purified with Sephadex TM G-50 (Amersham Biosciences) following the manufacturer's instruction and visualized on an ABI PRISM 3130xl automated sequencer (Applied Biosystems). All sequences are available from GenBank with the accession numbers MG719346 to MG719360 for COI and MG719344 and MG719345 for 18S (see also Table 1 & Table A1).

| Data analysis
All COI sequences were verified by comparisons with known oribatid mite sequences from GenBank and aligned by eye in MEGA version 6 (Tamura, Stecher, Peterson, Filipski, & Kumar, 2013). To infer and visualize the genealogical relationships among the D. dorni individuals, the COI data were used for a TCS network reconstruction (Clement, Snell, Walker, Posada, & Crandall, 2002) using the program PopART (http://popart.otago.ac.nz).
For 18S sequence alignment, the R-Coffee web server (Moretti, Wilm, Higgins, Xenarios, & Notredame, 2008; available at http:// www.tcoffee.org) which takes into account the predicted secondary structures, was used. To eliminate poorly aligned positions/regions of the resulted RNA alignment, the program Gblocks v0.91b (Castresana, 2000) was applied under default parameters, except "Minimum Length of A Block" was set to a smaller value (5 instead of 10) as recommended by the authors for rDNA-like alignments.
The final 18S alignment had a total length of 1375 bp. All alignments are provided as Supporting Information.
For BI inference, number of substitution types was set to six (GTR model) for each data partition and among-site rate variation was drawn from a gamma distribution. Posterior probabilities were obtained from a Metropolis-coupled Markov chain Monte Carlo simulation (two independent runs, eight chains with 15 million generations each, chain temperature 0.2, and trees sampled every 1000 generations). After checking parameter values of the sampled chains in Tracer v1.6 (Rambaut & Drummond, 2007; available at http://tree.bio.ed.ac. uk/software/tracer/), the first 10% of the sampled trees were excluded as burn-in. From the remaining trees, a majority rule consensus tree was calculated.

| Morphological analysis
In general, mite specimens were mounted in Berlese medium (a mixture of arabic gum, aqua dest., glucose, chloral hydrate, and glacial ethanoic acid) as permanent slides.
T A B L E 1 Sampling locality, coordinates, sample (=voucher) ID, and sequence GenBank accession numbers for all Dolicheremaeus dorni (Dd) and Spinotocepheus sp. (Spin_sp) specimens analyzed in this study  Poland (orange colored), one from Carinthia/Lavam€ und (red colored), and one from Croatia (green colored). As the uncorrected pairwise differences between the two studied othocepheid species ranged from 22.3% to 23.2%, we avoided it to include the Spinotocepheus sp. haplotype in the network reconstruction.

| 18S sequences
The results of both methods, BI and ML, yielded highly similar topologies (Figure 2 & Figure S1). Differences were either due to unresolved parts in the ML tree compared to the BI analysis (there, however, nodes were poorly supported) or in lower node supports of some taxa. In general, Parhyposomatides and Enarthronotides formed one clade at the basis of the phylogeny with Desmonomatides as sister group which is congruent with previously published data (Dabert et al., 2010;Pachl et al., 2012Pachl et al., , 2017. Also within Desmonomatides, the topology went quite along with the morphologybased system after Norton and Behan-Pelletier (2009). Nearly all included superfamilies were resolved as monophyletic entities excepting Ceratozetoidea and Crotonioidea-in latter case, however, only weakly supported by both analyses (Figure 2 & Figure S1).  Figure 1a Epieremulus granulatus (Balogh & Mahunka, 1979)] were the sister group of Carabodoidea and both together the sister clade of Oppioidea. Prodorsum (Figures 3a and 4a). Cerotegument finely granular, except for area between costulae showing large granules. All prodorsal setae robust and slightly barbed; rostral setae (ro) long (approx. Gastronotic region (Figures 3a and 4a). Cerotegument granular, granules loosely distributed. Lateral condyles of notogaster (co.nl.) triangular in shape and tips slightly covered by prodorsal lateral condyles (co.pl.) in dorsal view. Ten pairs of robust, long (length 55-75 lm), and slightly barbed notogastral setae, c, la, lm, lp, h 1-3 , p 1-3 .

| Genetics
All herein investigated D. dorni specimens, originating from six European countries, represent one and the same species. This is in contrast to other studies on mites, insects, or other invertebrates, which have shown that presumed widespread taxa often represent com- Furthermore, our results revealed similar topologies as already published phylogenies (e.g., Iseki & Karasawa, 2014;Pachl et al., 2012Pachl et al., , 2017, aside from the different taxa and taxonomic classification used. According to the system provided by Norton and Behan-Pelletier (2009), Oppiodea and Gustavioidea might be closely related to the Carabodoidea, which in fact is supported by our phylogenetic data.
However, the resulted sister grouping of Carabodoidea and Cepheoidea is questionable, with the reason that the accommodation of the family Anderemaeidae, represented by E. granulatus in this study, to the Cepheoidea is still under discussion and therefore might be wrong (Norton & Behan-Pelletier, 2009;Woas, 2002

| Ecology
Basically all tropical Dolicheremaeus species can be found in soil and litter of wet decaying, spongy wood (Norton & Behan-Pelletier, 2009;Aoki, 1965Aoki, , 1967etc.). The temperate D. dorni was originally described from decaying leaves in the area of Baile Herculane (Meridional Carpathians) (Balogh, 1937) and Mahunka (1982)  In this context, another record of this species seems to be quite interesting, namely those from a study of Pernek, Wirth, Blomquist, Avtzis, and Moser (2012) who detected specimens in samples of the fir bark beetle Pityokteines curvidens Germ. caught in pheromone traps in Croatia. However, as this is the only known case of such an association, it would be highly speculative to suggest phoretic behavior for D. dorni. Pernek et al. (2012) also stated that the finding of this taxon is more likely the result of accidental dispersal than an active phoretic behavior of the mite (see also Norton, 1980). Moreover, phoresy increases the dispersal ability of individuals leading to positive effects on population demography, evolution, and community success of a species (Clobert, Danchin, Dhondt, & Nichols, 2001). Given the rare and accidental records in Europe (Figure 1a) but also the clear signal of four geographically distinct clades in our network reconstruction (Figure 1b), phoresy seems not to be a common phenomenon in the studied species.
However, as D. dorni was found in litter, on bark, and tree-associated mushrooms, this species clearly seems to be associated with tree habitats but at the same time seems to be a generalist within these habitats. Maybe this generalistic nature is one of the reasons why D. dorni could colonize a larger area within cold temperate regions.

| Diversity and distributions
Presently, there are 185 species and nine subspecies of the genus distinctus (Aoki, 1982), Dolicheremaeus elongatus (Aoki, 1967), Dolicheremaeus infrequens (Aoki, 1967), and Dolicheremaeus orientalis (Aoki, 1955)]. Accordingly, there is a clear latitudinal gradient with the lowest species number in temperate regions and the highest number in tropical areas which may be explained by the tropical conservatism hypothesis (Wiens & Donoghue, 2004). This general concept suggests that (i) species richness is higher in tropical biomes because most taxa originated in the tropics, (ii) tropic taxa had more time and area available for speciation, and (iii) species are specialized for tropical climates and only few were able to disperse out of the tropics and adapt to the cold (often freezing) temperatures of middle-and high-latitude regions. These three points are also met by the hypothesis of Pachl et al. (2017) stating that the desmonomatan radiation started on the super continent Pangaea where mites were mainly exposed to tropical climatic conditions.

| Morphology
The present specimens are well in accordance with the original description of D. dorni given by Balogh (1937) large-and small-sized animals of each population even show the same haplotype. Interestingly, Weigmann (2014) already demonstrated large intraspecific size differences in the German individuals (more than 100 lm between smallest and largest, equaling a fifth of overall body size), and this unusual variation is also present in Austrian and Croatian populations (Table S2). Oribatid mites are known to show a size-dependent sexual dimorphism with females being basically larger than males (e.g., Behan-Pelletier, 2015) and the same kind of dimorphism can be found in D. dorni. However, males are only by trend smaller and body sizes of both sexes do largely overlap so that the found large variation cannot be explained by such a sexual dimorphism. Jacot described Dolicheremaeus rubripedes (Jacot, 1938) and stated "size quite variable" (Jacot, 1938;p.51), which indicates that large intraspecific variation can also be found in other Dolicheremaeus species and hence variable body size may simply be a generic trait.
Other However, a comparison with the other two non-tropical Dolicheremaeus species, namely D. georgii from Trans-Carpathians and D. montanus from Eastern-Kirghizia (Ghilarov, 1975), shows that they are quite similar in terms of morphology, and they mainly differ in the shape of notogastral setae and the length of prodorsal lamellae from each other.
F I G U R E 6 Map showing the worldwide distribution of the genus Dolicheremaeus. Colors refer to climate zones; size of circle symbol is relative to species numbers in the respective area

| CONCLUSIONS
Morphological and molecular genetic analyses clearly demonstrate that D. dorni shows a wider distribution in Central Europe. Nearly all investigated populations show specific haplotypes indicating that there is actually no or low gene flow between the populations.
Based on all the records of the temperate D. dorni, we suggest that this species is basically associated with tree habitats, whereas preferences for specific tree species or specific microhabitats on the trees could not be detected.
Presently, D. dorni represents the only species of this genus that was able to colonize a wider region within the cold temperate climate zone.