Nearctic Species in the Palearctic: Trans-Beringian Range, Phylogeny and Phylogeography of Pterostichus (Cryobius) mandibularoides (Coleoptera, Carabidae)

Subgenus Cryobius is one of the most numerous among the megafauna of tundra soils, but studies on its species distribution, taxonomy, and ecology are lacking. Phylogeny and phylogeography reconstructions of insects with taxonomic complexity have become possible using an integrative approach. Here, we report that specimens of Pterostichus (Cryobius) mandibularoides, described from North America, were detected in Eurasia. Thus, this species has a trans-Beringian range with high distributions in North America, as well as a disjunctive part of the range on the northeastern edge of Asia within Chukotka and Wrangel Island. Eight COI haplotypes with closed relationships (1–2 mutation steps) were detected within the whole range, and one 28S rRNA haplotype was detected for Eurasia. Bayesian phylogeny revealed that P. mandibularoides had the most recent common ancestor with sister species P. brevicornis and P. nivalis. Mean genetic distances of both markers were similar and higher between P. mandibularoides and both P. brevicornis and P. nivalis (>5% ± 1.0%) than between the latter species (<4%± 1.0%). The obtained results change the previous view about brevicornis group stock differentiation within Cryobius in the Arctic and require a revision of the phylogeny and phylogeography of brevicornis group species and Cryobius altogether.


Introduction
The Bering Strait was the last bridge connecting the tundra fauna of Eurasia and North America during the last Ice Age [1][2][3][4]. A recent study based on time-calibrated phylogenies of a broad range of specific taxon groups (plants, fungi, arthropods, amphibians, reptiles, and mammals) showed that the biotic exchange was asymmetric with a higher degree of dispersal from Eurasia to North America [4]. However, a number of species were restricted to only one side of Western or Eastern Beringia [2]. Range reconstructions and the phylogeny of some arctic species with taxonomic complexity have become possible via the integrative approach. A disjunctive range on Novaya Zemlya and Wrangel Island with close affinities was detected for Bombus glacialis (Friese, 1902) [5]. It was shown that the widespread Arctic tiger moth Arctia tundrana (Tshistjakov, 1990) has a low genetic divergence among distanced populations from Kolguev Island (north Europe) to Chukotka [6]. Widely distributed specimens of North American moth Udea itysalis (Walker, 1859) were registered on Kolguev Island (north Europe) [7]. Moreover, specimens of Gnorimoschema vastificum (Braun, 1929) to date known only from North America were detected on Yugorsky Peninsula (north Europe) [8]. Some insect taxa with Nearctic distribution were found on Wrangel Island, which were not registered in other regions of Eurasia. They are represented in various taxonomic groups: Aphididae [9], Erebidae [10], and the dipteran from families Empididae [11], Syrphidae [12], Agromyzidae [13], and Muscidae [14,15].
Among insects of the soil megafauna in the Arctic and Subarctic, the species of subgenus Cryobius demonstrate high frequency and distribution [16]. Nevertheless, there are only a few studies on the phylogeny and ecology of Cryobious in the North Holarctic [17][18][19][20][21][22]). The most studied subgenus group is brevicornis, with the last revision in 1998 by Erjiomin [22]. Five species were detected in the group: P. brevicornis (Kirby, 1837) with circumpolar distribution, P. empetrikola (Dejean, 1828) and P. nivalis F. (Sahlberg, 1844) with trans-Beringian range, P. kolymensis (Erjiomin, 1998) with a restricted range in East Siberia, and P. mandibularoides (Ball, 1966) widely distributed in North America [19,22]. Three species, P. brevicornis, P. kolymensis, and P. mandibularoides, have a close morphology with similar body size, pronotum shape, and male genitalia. Furthermore, for species P. brevicornis, three subspecies were detected: P. brevicornis brevicornis (Kirby, 1837), Pterostichus brevicornis delicatus (Casey, 1918), and Pterostichus brevicornis yasudai (Morita, 2002). Only P. b. brevicornis has a Holarctic range, while the other two subspecies are distributed on restricted territories of the islands of the Bering Sea and Chukotka for Pterostichus b. delicatus, and Hokkaido Island for Pterostichus b. yasudai [19,[23][24][25]. In our samples of soil arthropods from the northeast edge of Russia, several Cryobius species were found and examined using the integrative approach with genetic tools. The specimens of P. mandibularoides were detected in samples from Chukotka and Wrangel Island. Thus, the aims of this study were to (1) clarify the range distribution of P. mandibularoides, and (2) focus on its phylogeography and phylogeny within the brevicornis group using the integrative approach.

Data Collection
Specimens of P. mandibularoides were collected on Chukotka (Alkatvaam and Meynipilgino settlements, subzone of southern tundra) and Wrangel Island (Somnitelnye Mountain and Somnitelnaya Bay, subzone of Arctic tundra) ( Figure 1, Table 1). The specimens were sampled using pitfall traps and hand collection (with an exhaustor).   Table S1.
Ground beetles were preserved in 96% ethanol or dried after collection at room temperature (21-25 °С) for subsequent DNA and morphological analyses. The P. mandibularoides specimens were studied morphologically using a stereomicroscope (Leica M165C, Germany) in accordance with the approaches of Ball [19] and Lindroth [20,21].  Table S1. Ground beetles were preserved in 96% ethanol or dried after collection at room temperature (21-25 • C) for subsequent DNA and morphological analyses. The P. mandibularoides specimens were studied morphologically using a stereomicroscope (Leica M165C, Germany) in accordance with the approaches of Ball [19] and Lindroth [20,21]. Furthermore, we explored the morphological structures of the taxonomically closest species to P. mandibularoides from the brevicornis group: P. nivalis and P. brevicornis (subspecies P. brevicornis brevicornis) ( Table S1). Images of morphological details were taken using a stereomicro-scope Axio Zoom V16 (Carl Zeiss, Berlin, Germany) and Canon EOS 80D Digital Camera (Canon Inc., Japan, Tokyo) ( Figure 2). The full morphological descriptions with dichotomic keys for these species are given in the checklists of Ball and Erjiomin [19,22]. According to imago habitus, P. mandibularoides is most like P. brevicornis, and both species can overlap in all characteristics [19]. Hence, the main distinctions between these species are in the form of the male genitalia [19,22]. The median lobe of P. mandibularoides in the lateral aspect without curving of the shaft features an apex with a slight ventral lip ( Figure 2F) [19]. For the male of P. brevicornis, the median lobe has a truncate or narrowly rounded apex, and sides of the shaft in ventral aspect were almost parallel ( Figure 2D) [19]. According to the combination of morphology similarities, Ball concluded that P. mandibularoides and P. brevicornis are more closely related to one another than they are to P. nivalis [19].
Furthermore, we explored the morphological structures of the taxonomically closest species to P. mandibularoides from the brevicornis group: P. nivalis and P. brevicornis (subspecies P. brevicornis brevicornis) ( Table S1). Images of morphological details were taken using a stereomicroscope Axio Zoom V16 (Carl Zeiss, Germany, Berlin) and Canon EOS 80D Digital Camera (Canon Inc., Japan, Tokyo) ( Figure 2). The full morphological descriptions with dichotomic keys for these species are given in the checklists of Ball and Erjiomin [19,22]. According to imago habitus, P. mandibularoides is most like P. brevicornis, and both species can overlap in all characteristics [19]. Hence, the main distinctions between these species are in the form of the male genitalia [19,22]. The median lobe of P. mandibularoides in the lateral aspect without curving of the shaft features an apex with a slight ventral lip ( Figure 2F) [19]. For the male of P. brevicornis, the median lobe has a truncate or narrowly rounded apex, and sides of the shaft in ventral aspect were almost parallel ( Figure 2D) [19]. According to the combination of morphology similarities, Ball concluded that P. mandibularoides and P. brevicornis are more closely related to one another than they are to P. nivalis [19].  Overall, 65 specimens of P. mandibularoides were revealed in our samples from Chukotka and Wrangel Island. Cryobius samples from Chukotka specimens of P. mandibularoides were not found in the suburbs of Pevek settlement (closest mainland locality to Wrangel Island) and in the suburbs of Hatyrka settlement (about 100 km from Meynipilgino settlement near the coast of Bering Sea). We also examined the collected species of brevicornis group in samples from West and Central Siberia, as well as North Europe, and we did not detect P. mandibularoides. Unfortunately, we had no samples of species of the brevicornis group from localities adjacent to the Chukotka territory of East Siberia. The specimens were deposited in the Zoological Institute of the Russian Academy of Sciences (Saint Petersburg, Russia) from East Siberia, belonging to P. brevicornis or P. kolymensis (without ID numbers). One paratypus of P. mandibularoides is deposited in the coleoptera collection of the aforementioned Zoological Institute ( Figure S1). A female specimen was explored by Ball

DNA Extraction, PCR, and Sequencing
For molecular analysis, 11 specimens of P. mandibularoides were used (Table S1). In the present study, we obtained the cytochrome c oxidase subunit I (COI) gene and the large subunit ribosomal RNA 28S (28S rRNA) gene sequences. The sequence dataset involved 22 COI and eight 28S rRNA sequences combined with our materials (11 COI and eight 28S rRNA sequences) and published GenBank (NCBI and (BOLD) IDS) data (11 COI sequences) of P. mandibularoides (Table S1). Moreover, we used datasets of sequences from the brevicornis group, P. nivalis and P. brevicornis, which are taxonomically closely related species to P. mandibularoides. For the P. nivalis dataset, we used 79 COI and four 28S rRNA sequences (GenBank: 70 COI sequences; our materials: nine COI and four 28S rRNA sequences). For the P. b. brevicornis dataset, we used 170 COI and 33 28S rRNA sequences (GenBank data: 93 COI sequences; our materials: 77 COI and 33 28S rRNA sequences). All datasets of the aforementioned three species from GenBank belong to North American (Alaska, Canada) specimens, and our materials were sampled in North Eurasia (Table S1).
DNA was extracted from dissected abdominal muscle tissue and legs using a standard phenol/chloroform procedure [26]. The mitochondrial gene COI was amplified using two primer pairs: C1-J-1718 with C1-N-2329 [27], and LCO 1490 [28] with LepR [29]. Amplified COI barcode fragments were 612 bp and 660 bp in length, respectively. The primers D23F [30] and D2 [31] (Table S1). All sequences were analyzed using BioEdit v. 7.2.5 [32]. The alignment of the COI and 28S rRNA sequences was performed using the Muscle and the ClustalW algorithms in MEGA X [33]. Poorly aligned positions and divergent regions from the alignment of the 28S rRNA gene were eliminated using the online GBlocks server v0.9b [34,35].

Phylogenetic and Phylogeographic Analyses
Both genes were analyzed, combined and separately (Table S2). The resulting aligned sequences of the two gene fragments (COI, 28S rRNA) were joined to obtain a combined nucleotide sequence alignment (total length of 1130 bp). This combined dataset was collapsed into a set of unique haplotypes (including haplotypes of outgroup taxa Pterostichus (Cryobius) ventricosus (Eschscholtz, 1823) and Pterostichus (Phonias) strenuus (Panzer, 1797)) using an online FASTA sequence toolbox [36]. The best evolution models were selected for each gene separately on the basis of the corrected Akaike information criterion (AICC) of jModelTest [37] in MEGA X [33]. The GTR + G model of sequence evolution was best for both studied genes: COI gene (G = 0.15) and 28S rRNA (G = 0.16). Phylogenetic relationships were reconstructed on the basis of Bayesian inference performed in MrBayes v.3.2.6 [38] through the CIPRES Science Gateway [39]. The following parameters were used for analyses: nchains = 4, nruns = 4, samplefreq = 1000, temp = 0.2. The first 15% of trees were discarded as burn-in (pre-convergence part), and the majority rule consensus tree was calculated from the remaining trees. Convergence of the MCMC chains to a stationary distribution was checked using Tracer v. 1.7 [40]. The effective sample size (ESS) value for each parameter sampled from the MCMC analysis was always recorded as >600. Trees were viewed using FigTree v. 1.4.4 [41]. Genetic differences (p-distances) between analyzed haplotypes were calculated using MEGA X [32].
The phylogeographic analyses were performed on the basis of a median-joining network approach using Network v. 5.0.0.1 software with default settings [42]. Additionally, 264 COI sequences from 28 localities of Northern Europe and North America were used (Table S1). The sequence length leaving the dataset was 571 bp long due to the removal of missing sites and cuts of different lengths of available sequences.

Results
The sequenced specimens of P. mandibularoides consisted of eight unique COI and one 28S rRNA haplotypes. An equal number of unique COI haplotypes (four haplotypes) were detected for North America and northeast Russia. The mtDNA haplotypes were closely related to each other with 1-2 mutation steps but did not share a common haplotype within samples of subdivided continents and Wrangel Island (Figure 3). The mean uncorrected pairwise COI p-distance (±SEM) between North American and Far Eastern samples of P. mandibularoides was 0.49% ± 0.17%. According to both median-joining network analysis and pairwise p-distance, P. mandibularoides had equidistant relations with P. brevicornis and P. nivalis. The pairwise COI p-distance for complete haplotype dataset and pairwise 28S rRNA p-distance for Eurasian samples did not exceed 5% ± 1% between P. mandibularoides and the other two species ( Table 2). Haplotypes of P. brevicornis and P. nivalis were more closely related to each other and did not exceed 4% ± 1% mean p-distance for both genetic markers ( Table 2).
Bayesian phylogeny revealed the same patterns as network analysis, with P. mandibularoides representing the most distant lineage among three studied species of brevicornis group with a fully supported clade ( Figure 4). P. b. brevicornis appeared to be a sister species to P. nivalis, and their clades had quite high support (Figure 4). The subclade of Canadian sequences of P. mandibularoides with a high probability value had more distance with samples from America and Chukotka. Table 2. Genetic divergences (mean uncorrected p-distances ± standard error estimations, %) between P. mandibularoides and the two other related species of the brevicornis group on the mitochondrial COI gene for the full dataset (below diagonal) and the nuclear 28S rDNA gene for Eurasian samples (above diagonal) sequences.   Table S1).

Species
Bayesian phylogeny revealed the same patterns as network analysis, with P. mandibularoides representing the most distant lineage among three studied species of brevicornis group with a fully supported clade ( Figure 4). P. b. brevicornis appeared to be a sister species to P. nivalis, and their clades had quite high support (Figure 4). The subclade of Canadian sequences of P. mandibularoides with a high probability value had more distance with samples from America and Chukotka.  Table S1).  . The majority-rule consensus phylogenetic tree of P. mandibularoides, P. b. brevicornis, and P. nivalis recovered from Bayesian inference analysis, based on the combined mitochondrial and nuclear sequence dataset (COI, 28S rRNA). Sequences of P. mandibularoides from Chukotka and Wrangel Island are highlighted in green with accession numbers to NCBI. New haplotype codes are as indicated in Table S1. Numbers near branches indicate the Bayesian posterior probability (BPP). The scale bar indicates the branch length.

Discussion
This study showed that the previous Nearctic range of P. mandibularoides must be extended into the Palearctic along the northeastern edge of Asia (Chukotka and Wrangel Island). The observed disjunctive range with a wide marine barrier between the mainland of Chukotka and Wrangel Island (140 km) makes it possible to assume a wider dis- Figure 4. The majority-rule consensus phylogenetic tree of P. mandibularoides, P. b. brevicornis, and P. nivalis recovered from Bayesian inference analysis, based on the combined mitochondrial and nuclear sequence dataset (COI, 28S rRNA). Sequences of P. mandibularoides from Chukotka and Wrangel Island are highlighted in green with accession numbers to NCBI. New haplotype codes are as indicated in Table S1. Numbers near branches indicate the Bayesian posterior probability (BPP). The scale bar indicates the branch length.

Discussion
This study showed that the previous Nearctic range of P. mandibularoides must be extended into the Palearctic along the northeastern edge of Asia (Chukotka and Wrangel Island). The observed disjunctive range with a wide marine barrier between the mainland of Chukotka and Wrangel Island (140 km) makes it possible to assume a wider distribution of P. mandibularoides at the Last Glacial Age in Eurasia before sea transgression, at least on trans-Beringian territory. A similar result was detected using time-calibrated phylogeny for some species of bumblebees from the lapponicus group [5]. Our result is another species with a range wider than the Nearctic region, which continues the list of recent detections of Nearctic species in the Holarctic for the taxa of different insects [7][8][9][10][11][12][13][14][15].
P. mandibularoides is widely distributed in North America but has low occurrences within ground beetle compositions of the northeastern part of the range. In the mountainous part of Wrangel Island and in both localities in Chukotka, it was found together with P. brevicornis. At the same time, in a much more significant number of Chukotka localities, where P. b. brevicornis was collected (Table S1), P. mandibularoides was absent. This is especially interesting, since the species was collected in the northern and southern parts of the tundra zone, in biotopes with different ecological characteristics, which suggests that it is quite polytopic.
Phylogeny by both genetic markers did not correspond with the previous conclusion on the two species (P. mandibularoides and P. brevicornis), which did not share common ancestry within the brevicornis group in the late Pleistocene [19]. The result indicates that P. mandibularoides had the most recent common ancestor with sister species P. brevicornis and P. nivalis. Bayesian phylogeny was also confirmed by the genetic distances of both mtDNA and 28 rRNA, with lower p-distance measures between P. b. brevicornis and P. nivalis, and higher p-distance measures between both species and P. mandibularoides. The number of unique COI haplotypes for P. mandibularoides samples (eight haplotypes) was comparable with another species from the brevicornis group-P. nivalis (nine unique COI haplotypes); however, the number of species samples was three times higher for P. nivalis. The ranges of both species overlap in Alaska and Chukotka, but P. nivalis is found principally in coastal areas [19]. In comparison, P. b. brevicornis has a circumpolar range with 23 COI haplotypes in northern territories, overlapping with the ranges of the other two species. P. b. brevicornis is distributed in different subzones of tundra, north taiga, and mountain [16,19,21,43].
According to the latest taxonomy revision of the brevicornis group in the Palearctic, a new species, P. kolymensis (Erjiomin, 1998) was described in specimens from west Chukotka and the boundary territory with Yakutia [22]. P. kolymensis has similar male genitalia to P. mandibularoides and morphologically differs from P. mandibularoides in terms of width of body and shape of pronotum angles. Furthermore, there is no reliable information about the biology, ecology, and range of this species [22]. Considering the complexity of species identification, with high variability of morphological structures between closely related species of subgenus Cryobius, the specimens of P. kolymensis have to be distinguished using the integrative approach with genetic analysis. In particular, this methodology was required to confirm new data on P. mandibularoides distribution in the Palearctic.
Three morphologically close species with overlapping ranges from the brevicornis group (P. brevicornis, P. mandibularoides, and P. kolymensis) have been registered for the northeastern territories in the Palearctic. Thus, paleontological data are hardly useful for time-calibrated phylogeny of this group. Some body structures such as elytra or pronotum are not strong morphological features for Cryobius species taxonomy, and male genitalia are rarely found among fossil ground beetles. Fossil insects are being used as ecology indicators for Arctic paleoenvironmental reconstructions in the Pleistocene and the Holocene [3,[44][45][46][47][48]. Carabidae and particularly Cryobius frequently predominate in samples from different stratigraphic layers of Arctic territories, and they are associated with mesic to moist tundra habitats or meadows and bogs in boreal forest [47,48]. For instance, in the taiga and tundra of North Europe, specimens of P. b. brevicornis have been sampled in broader environmental conditions [49][50][51][52][53]. Specimens of P. nivalis are usually found on loamy soil in rather dry tundra [19]. North American specimens of P. mandibularoides were sampled on bare, damp, sandy soil under the dead leaves of Salix and Alnus [19]. However, in the extreme northeast of Asia, on Wrangel Island and in Chukotka, this species was collected in both wet and dry biotopes. It is interesting to note that specimens with different haplotypes were associated with these different types of biotopes on Wrangel Island.

Conclusions
Specimens of P. mandibularoides were detected in the trans-Beringian range with high distributions in North America, and the disjunctive part of the range was on the northeastern edge of Asia, within Chukotka and on Wrangel Island, where it was found in areas with different environmental conditions. Bayesian phylogeny for P. mandibularoides revealed a recent common ancestor with sister species P. b. brevicornis and P. nivalis. We also conclude that the integrative approach in general would be required for further studies of Cryobius species and the brevicornis group. Overall, the lack of studies on Cryobius taxonomy, distribution, and ecology makes the species the weak link in paleoenvironmental reconstructions of the Arctic and the Subarctic in the Pleistocene and Holocene, which could lead to biased conclusions.
Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/d14060415/s1, Figure S1: Paratypus of P. mandibularoides (female speciment) is deposited in the coleoptera collection of the Zoological Institute of the Russian Academy of Sciences (Saint Petersburg, Russia); Table S1: List of sequenced specimens of Pterostichus mandibularoides, Pterostichus brevicornis and Pterostichus nivalis, including the location wiht BOLD and NCBI's GenBank accession numbers; Table S2: The 28S rRNA gene polymorphism in three species of the brevicornis group.