Introduction

Water mites (Hydrachnidia) are very diverse and species rich group of macroinvertebrates. They occupy almost all freshwater environments. An updated version of Limnofauna Europaea (www.watermite.org) shows the improvement of the knowledge on European water mite biodiversity. From the 1062 species listed in 1978 year, 28% have been synonymized or excluded because of their uncertain status (species incertae), while at the same time over 200 species were added1. There is still a clear gap in alpha-taxonomy and knowledge upon phenotypic polymorphism and cryptic diversity within Hydrachnidia. Recent publications2,3,4,5,6,7,8,9,10 indicate the presence of many unrecognized species, especially in the southern part of Europe, which can be distinguished by molecular methods.

The publication of the three parts of an identification key11,12,13 initiated a new trend in researchers on European water mites, facilitating or enabling ecological and biological research. The release of these keys was preceded by numerous revisions of individual genera14,15,16 based on morphological data. However, the development of molecular barcoding17 quickly suggested that alpha-taxonomy is still weakly recognised and there are many species-complex containing cryptic or pseudo-cryptic species. An important work on integrative taxonomy and phylogeny of water mites was published by Dabert et al.18. Their analyses, based on nuclear ribosomal genes such as 18S, 28S and fragments of the mitochondrial gene of the cytochrome c oxidase subunit 1 (cox1), provided evidence about the relationships within the group. Recent taxonomical studies2,3,4,5,6,7,8 indicated a lack of knowledge about phenotypic polymorphism and cryptic or pseudo-cryptic diversity within Hydrachnidia. Therefore, more genetic data are needed. One of the ways to achieve this is to sequence complete mitochondrial genomes. There is an important, rapidly growing literature dedicated to the sequencing of complete mitogenomes, but despite this, up to now only 12 mitogenomes of Hydrachnidia have been made available and published19,20,21,22,23.

To participate in this effort, we have undertaken the sequencing of the complete mitogenome of Hygrobates turcicus Pešić, Esen & Dabert, 2017, a species belonging to genera extensively studied by DNA-barcoding2,3,4,5,6,9,10. We compared the characteristics of this mitogenome with those of Hygrobates longiporus Thor, 1898 and Hygrobates taniguchii Imamura, 1954, and used the data to perform a multigene phylogeny.

Results

Mitogenome of H. turcicus

The mitogenome of H. turcicus (GenBank accession number OM336267) is 15,006 bp long (Fig. 1) (Table 1). It contains 13 conserved protein-coding genes, two rRNAs, and 22 tRNAs. It is colinear with the mitogenomes of H. longiporus and H. taniguchii, but its size is ca. 1300 bp longer (Table 1). The extra-length originates mostly from a non-coding region, which is located between protein-coding genes ND3 and ND4L. A comparison of the length, start and stop codon of the protein-coding genes of Hygrobates spp. is presented in Table 2. Two genes of H. turcicus present a premature termination, namely cox2 and cox3. In the case of cox3, this is common to the three species of Hygrobates spp. The distribution of the start codons is 7 ATG, 3 ATT, 2 ATA and 1 TTG. This TTG start codon was found in the ND5 gene, which discriminates H. turcicus from H. longiporus and H. taniguchii. A canonical start codon could not be identified in this case.

Figure 1
figure 1

Map of the mitochondrial genome of Hygrobates turcicus. Genes belonging to different functional groups are color coded differently and the GC, AT content of the genome are plotted on the inner circle as dark and light gray, respectively.

Full size image
Table 1 List of the mitogenomes used during this study with their accession number and sizes.
Full size table
Table 2 Comparison of the mitochondrial protein-coding genes of the three species of Hygrobates.
Full size table

Comparison with other populations of H. turcicus

The percentages of identities between the available cox1 genes of H. turcicus are listed in Table 3. These percentages range between 99.52 and 99.84%, with the highest percentage (99.84%) being found 10 times out of 16.

Table 3 Percentages of identity as calculated by Clustal Omega between the cox1 gene obtained during this study and the same gene as available on GenBank.
Full size table

Multigene phylogeny

The phylogenetic tree inferred from concatenated mitochondrial protein-coding genes displays high bootstrap values, ranging from 98 to 100%. The snail parasites Riccardoella spp. appear as an outgroup. Phylogenetic analyses clustered the analysed Parasitengona species into 6 maximally supported clades, three of them (1. Sperchon plumeifer and Mideopsis roztoczensis; 2. Unionicola parkeri and U. foili; 3. H. longiporus, H. taniguchii and H. turcicus) corresponding to the water mites. H. turcicus was recovered with high statistical support (98%) as a sister branch to H. longiporus and H. taniguchii (Fig. 2).

Figure 2
figure 2

Cladogram illustrating the phylogenetic relationships for Hygrobates turcicus based on complete mitochondrial genome sequences. Mitochondrial genome rearrangement events are mapped on the branches of the best scoring maximum likelihood tree generated with RAxML-NG. Each node has 100% bootstrap support value.

Full size image

Discussion

H. turcicus is a species distinguished recently on the basis of DNA-barcoding from the group of H. fluviatilis-complex species. It was described from Turkey and next mentioned from Bulgaria3,5. It is closely related with H. ulii Pesic, Saboori, Zawal & Dabert, 2019, and together with H. balcanicus Pesic, 2020, it is a separate clade in relation to the other species of H. fluviatilis-complex5,6. It is worth noting that based on cox1 comparisons (Table 3), the specimen sequenced in the current study didn’t exhibit a higher conservation with the other specimen from Bulgaria (GenBank: MN520308), more precisely from the Strymon river, than with 9 others specimens originating from Turkey. This might prove to be a limitation for accurate biogeographical studies that need to be considered when using single gene barcoding.

Molecular barcoding has proven useful to unveil the genetic diversity between closely related species, if not cryptic or semi-cryptic species of water mites, especially when obtained through the sequencing of the cox1 gene. It is usually sufficient to perform molecular phylogenies within these species. For more distantly related species, e.g. belonging to different families, more conserved nuclear genes such as the small subunit of the ribosomal RNA gene are sometimes preferred10,18,24. What we would like to emphasize in our work is that complete mitogenomes can prove to be also useful. The amount of data retrieved from the concatenation of all protein-coding genes led to the obtention of a phylogenetic tree with optimal support at the nodes. The results obtained in this study combine geographically distant but taxonomically related species (H. longiporus, H. taniguchii and H. turcicus) into one clade, thereby establishing a sister group for the clade comprising the genus Unionicola, which belongs to the same superfamily (Hygrobatoidea) and contrasting the rest of the water mite species (Sperchon clupeifer and Mideospis roztoczensis). At the same time, all species of water mites constitute one group, a sister group of species belonging to Trombidia (genus Leptotrombium). This illustrates the great usefulness of complete mitogenomes for the recognition of relationship between geographically and taxonomically distant taxa.

Based on our results and subsequent comparisons with the works of other authors22, we could also notice some differences among the genus Hygrobates for what concerns the start and stop codons of their mitochondrial genes. In the near future, it would be interesting to sequence mitogenomes of other species closely related to H. turcicus, to see how much these features are conserved.

Finally, studies such as ours will be helpful in the near future for members of the community who work on biomonitoring based on metabarcoding or environmental DNA. We might cite the recent article from Blattner et al.25, which includes Hygrobates norvegicus among other bioindicator species. This study was based on the amplification of several mitochondrial genes. Sequencing complete mitogenomes of duly identified specimen of water mites will help documenting the databases for later uses in similar studies.

Materials and Methods

Biological material

The H. turcicus samples were collected from stones on the bottom of Bulgaria, Vidima River near Debnewo (42°56′41.4′′N, 24°48′44.6′′E, depth 0.4 m, water flow 0.71 m/s, temperature 10 oC, pH 8.53, oxygen 110%, conductivity 279 µS/cm, hardness 121 CaO mg/l), including 11 males, 27 females, 2 deutonymphs. Collected in 12.x.2019 by Zawal, Michoński & Bańkowska. One male and one female were dissected and slide-mounted for morphological identification.

DNA extraction, sequencing, assembly and annotation

Water mites were collected by hand netting. Specimen were sorted out, initially identified and preserved in 96% ethanol, which is a method generally used in genetic research material5. Up to 50 specimens identified as H. turcicus were pooled together, and their DNA was extracted using the DNeasy Blood and Tissue Kit (Qiagen GmbH, Hilden, Germany) as described previously23. Exoskeletons were retrieved after DNA extraction and mounted in Hoyer’s medium. Sequencing was performed at the Beijing Genomics Institute in Shenzhen, China, on a DNBSEQ platform in accordance with the company's procedure. A total of ca. 40 million clean 150 bp paired-end reads were obtained and assembled using SPAdes 3.14.026 using a k-mer of 125. The contig corresponding to the mitogenome was extracted, and the Consed27 package was used to verify its extremities. Annotations were done with the help of MITOS28 and manually curated.

Comparison with other populations of H. turcicus

The cox1 gene of H. turcicus was aligned with other 16 other sequences downloaded from GenBank and trimmed to a final size of 624 bp. The trimmed sequences were aligned on Clustal Omega online (ebi.ac.uk/Tools/msa/clustalo) to calculate the percentages of identities. We also computed the overall mean of genetic distances based on the Kimura 2-parameter model using MEGA 7.0. Standard error estimate(s) were obtained using bootstrap (1000 replicates).

Multigene phylogeny

We aligned the 13 complete mitochondrial genomes with MAFFT version 7.51029, using Riccardoella tokyoensis and Riccardoella reaumuri as outgroup terminals. We conducted maximum likelihood (ML) analyses using RAxML-NG30 under three different strategies. (1) One of the IR regions was removed from all mitochondrial genomes to reduce overrepresentation of duplicated sequences before we ran RAxML-NG on the unpartitioned alignment under GTR + I + G substitution model as a single partition; (2) The same data was partitioned by gene, exon, intron and intergenic spacer regions, allowing separate base frequencies, α-shape parameters, and evolutionary rates to be estimated for each; (3) we inferred the best-fitting partitioning strategy with PartitionFinder231 for the alignment. The best fitting nucleotide substitution models were inferred with jModelTest232. Phylogenetic trees were visualized and edited with FigTree 1.4.433. Support for the ML tree branches was calculated using the nonparametric bootstrap method with 1000 replicates.