Complete mitochondrial genomes and nuclear ribosomal RNA operons of two species of Diplostomum (Platyhelminthes: Trematoda): a molecular resource for taxonomy and molecular epidemiology of important fish pathogens

Background

The genus Diplostomum (Platyhelminthes: Trematoda: Diplostomidae) is a diverse group of freshwater parasites with complex life-cycles and global distribution. The larval stages are important pathogens causing eye fluke disease implicated in substantial impacts on natural fish populations and losses in aquaculture. However, the problematic species delimitation and difficulties in the identification of larval stages hamper the assessment of the distributional and host ranges of Diplostomum spp. and their transmission ecology.

Methods

Total genomic DNA was isolated from adult worms and shotgun sequenced using Illumina MiSeq technology. Mitochondrial (mt) genomes and nuclear ribosomal RNA (rRNA) operons were assembled using established bioinformatic tools and fully annotated. Mt protein-coding genes and nuclear rRNA genes were subjected to phylogenetic analysis by maximum likelihood and the resulting topologies compared.

Results

We characterised novel complete mt genomes and nuclear rRNA operons of two closely related species, Diplostomum spathaceum and D. pseudospathaceum. Comparative mt genome assessment revealed that the cox1 gene and its ‘barcode’ region used for molecular identification are the most conserved regions; instead, nad4 and nad5 genes were identified as most promising molecular diagnostic markers. Using the novel data, we provide the first genome wide estimation of the phylogenetic relationships of the order Diplostomida, one of the two fundamental lineages of the Digenea. Analyses of the mitogenomic data invariably recovered the Diplostomidae as a sister lineage of the order Plagiorchiida rather than as a basal lineage of the Diplostomida as inferred in rDNA phylogenies; this was concordant with the mt gene order of Diplostomum spp. exhibiting closer match to the conserved gene order of the Plagiorchiida.

Conclusions

Complete sequences of the mt genome and rRNA operon of two species of Diplostomum provide a valuable resource for novel genetic markers for species delineation and large-scale molecular epidemiology and disease ecology studies based on the most accessible life-cycle stages of eye flukes.

Trematodes trematodes of the genus Diplostomum von Nordmann, 1832, represent a diverse group of parasitic flatworms (Neodermata: Platyhelminthes: Trematoda) that has attracted the attention of parasitologists and evolutionary ecologists for a long time, but has always stood in the shade of medically important taxa, notably the closely related human-infecting schistosomes (blood flukes). Species of Diplostomum utilise three different hosts to complete their life-cycles (Additional file 1: Figure S1). First intermediate hosts are lymnaeid snails where the dispersive infective stages (free-swimming cercariae) are asexually produced; these identical genetic clones emerge from the snail hosts in vast quantities in the aquatic environment, actively seek and infect second intermediate hosts, a wide range of freshwater fish. Cercariae migrate to the eyes or brain of fish where they develop into a long-lived infective stage (metacercaria). The life-cycle is completed when infected fishes are consumed by the definitive hosts (fish-eating birds) where adults develop and sexual reproduction occurs; eggs are shed, hatch and develop into short-lived free-swimming stages (miracidia), which infect the first intermediate hosts [1].

Due to their wide distributional and host ranges, species of Diplostomum represent attractive model systems for population genetics of monoecious parasites with complex life-cycles alternating asexual and sexual reproduction (e.g. [2, 3]) and for studies of host-parasite coevolution in model vertebrate organisms (e.g. [4]). Furthermore, the metacercariae of Diplostomum spp. in fish eyes are important pathogens causing diplostomiasis (eyefluke disease of fishes) manifested as various degrees of blindness that affects fish feeding, growth and survival, implicated in substantial impacts on natural populations and losses in aquaculture; this has led to intensive field and experimental studies [1, 5]. However, the model systems used in these studies have been referred to as “Diplostomum spathaceum” a collective name for the species found in the lens [5], innominate species of Diplostomum [6] or to a composite group of Diplostomum spp. [7]. The lack of accurate species identification represents a major impediment in the assessment of transmission dynamics, infectivity and virulence that varies among species and strains of Diplostomum [1, 5] and of the effects of these parasites in natural and aquacultured fish populations, as well as in addressing broader questions related to geographical distribution and host ranges of Diplostomum spp.

The taxonomy of the genus Diplostomum is still in a controversial state due to the lack of unequivocal morphological criteria for species discrimination (see [8] for details) and this has resulted in substantially underestimated species richness in large-scale inventories of natural snail, fish and bird populations. Identification of the larval stages of Diplostomum spp. is particularly difficult since linking life-cycle stages requires experimental completion of the life-cycle. The application of DNA-based approaches provides a promising independent method for assessment of species boundaries within the genus and for molecular identification of developmental stages in these parasites with complex life-cycles. The pioneer studies have focused on sequencing of the internal transcribed spacers (ITS1 and ITS1-5.8S-ITS2) of the ribosomal RNA (rRNA) gene cluster [9–11]. However, recent studies have shown that these regions do not provide sufficient resolution for species discrimination within Diplostomum and have provided evidence that the barcode region of the mitochondrial (mt) cytochrome c oxidase subunit 1 (cox1) gene may serve as a more efficient marker in elucidating life-cycles and recognition of cryptic species diversity within Diplostomum. Using the diplostomid-specific primers flanking the cox1 ‘barcode’ region developed by [12], the first molecular prospecting studies predominantly focused on metacercariae in natural fish populations have revealed much higher species diversity than previously estimated from morphology alone in both the Palaearctic and Nearctic [8, 13, 14]. However, the recent expansion of the barcode library for Diplostomum spp. revealed low divergence levels within complexes of cryptic species thus hampering unequivocal species identification using the ‘barcode’ cox1 region alone [8, 14].

To overcome the current limitations, there is a need to supplement the molecular markers available. Characterisation of complete mt genomes represents such an approach whose applicability has already been demonstrated within flatworms of biomedical and veterinary importance (e.g. [15–17]). To greatly alleviate the task of de novo characterisation of mt genomes, recent developments in next generation sequencing and downstream bioinformatics have provided a time- and cost-effective strategy to characterise rich amounts of data, out of which mt sequences can be readily identified and mt genomes reconstructed [18]. Here, we demonstrate and discuss the yields of such a strategy, using representatives of two widely distributed species of Diplostomum as an example of flatworm parasites of increasing relevance. We characterise the first complete mt genomes and nuclear rRNA operons of two closely related species of the basal digenean order Diplostomida, Diplostomum spathaceum and D. pseudospathaceum, assess the existing mt markers and identify new regions of the mt genomes that are promising for simultaneous species delineation and large-scale molecular epidemiology assessments. As a by-product of our mt genome sequencing effort, we have characterised the complete transcribed region of the nuclear rRNA operon and provide the first genome wide assessment of the phylogenetic relationships of the order Diplostomida, one of the two fundamental lineages of the platyhelminth subclass Digenea.

DNA sources, sequencing and assembly

Adults of Diplostomum spathaceum and D. pseudospathaceum were sampled from the intestine of two freshly collected black-headed gulls (Larus ridibundus) from Chropyně and Tovačov (Czech Republic). Individual specimens of Diplostomum spp. were identified on the basis of their morphology [8], stored separately in absolute ethanol before total genomic DNA was extracted from single adult individuals per species with the use of QIAamp DNA Mini kit (Qiagen, Hilden, Germany) according to the manufacturer’s instructions with the final DNA elution done in two steps, each time using 50 μl of AE buffer (Qiagen). The two eluates were combined and the amount of total DNA isolated was measured with Qubit® 2.0 Fluorometer (Life Technologies, Paisley, UK) yielding 2.1 and 1.3 ng/μl total DNA for D. spathaceum and D. pseudospathaceum, respectively. Two samples for next generation sequencing (NGS) were prepared and run at the DNA Sequencing Facility of the Natural History Museum (NHM), London, UK. Genomic DNA was indexed and libraries prepared using TruSeq Nano DNA Sample Preparation Kit (Illumina, Inc., San Diego, USA), and run simultaneously with 2 other samples on a MiSeq Illumina sequencer yielding 250 bp long paired-end reads. Three additional individuals of D. pseudospathaceum were partially characterised through PCR amplification and Sanger sequencing.

The new mt genomes were directly assembled using the mt baiting and iterative mapping (MITObim) approach of [18]. Partial cox1 sequences of [8] for D. spathaceum (JX986887) and D. pseudospathaceum (JX986905) were used as an initial bait to extract the corresponding cox1 reads from the entire Illumina genomic readpool from which a new reference assembly was subsequently derived. The new reference sequence was then automatically subjected to an iterative set of baiting and mapping steps until the total number of mapped reads became stationary. Resulting MITObim assemblies were then imported in Geneious version 7 [19] where the raw paired-end reads were mapped onto them in a single step using custom sensitivity settings to estimate the full mt genome coverage (do not trim; min. overlap = 25; max. mismatches/read = 4 %). Geneious was then used to trim the overlapping regions to create a circular mt molecule and to inspect for any potential mapping/assembly errors in problematic regions (e.g. the repetitive regions). Two such regions (where the two newly characterised mt genomes markedly differ) were detected and the local assembly fit was manually checked in Geneious through a stepwise iterative mapping of Illumina reads extending from conserved flanking regions.

Contiguous sequences of the nuclear rRNA operon were assembled solely within Geneious through iterative mapping of paired-end reads on the previously published ITS1-5.8S-ITS2 sequences of D. spathaceum (JX986844) and D. pseudospathaceum (JX986854) from Tovačov (Czech Republic) [8]. Mapping was done through multiple steps using custom sensitivity settings (do not trim; min. overlap = 50; max. mismatches/read = 10 %) and checking the fit of mapped reads by eye.

Sequence annotation and phylogenetic analyses

Following the assembly, identity and position of individual mt protein- and tRNA-coding regions were determined using a suite of bioinformatics tools. MITOS [20] was used to reveal the position of mt protein-coding and rRNA genes, whereas tRNAscan-SE web server [21] together with ARWEN [22] were employed to localise tRNA genes and reconstruct their secondary structures. Exact boundaries of the mt protein-coding genes were then confirmed by alignment of inferred amino acid sequences with those of mt genomes of Trichobilharzia regenti (NC_009680), Fasciola hepatica (NC_002546) and Schistosoma japonicum (NC_002544), using E-INS-i aligning algorithm of MAFFT [23] implemented in Geneious. Identical aligning strategy was also used to create pairwise alignment of the whole-length of mt genomes of D. spathaceum and D. pseudospathaceum (Fig. 1).

Schematic alignment of linearised mt genomes of Diplostomum spathaceum and D. pseudospathaceum. Outline arrows indicate the direction of transcription of protein-coding and rRNA genes, hairline arrows indicate the position of tRNA genes (see Additional file 4: Table S1 for the single letter abbreviations). Solid horizontal lines in the grey central area depict average nucleotide identities of individual protein- and rRNA-coding regions (the ‘barcode’ region of cox1 defined by Folmer primers indicated by a dotted line); NC non-coding region

Full size image

Exact coding positions of individual nuclear rRNA genes (rDNA), as well as positions of the affiliated transcribed spacers, were determined progressively through the following series of steps: the entire assemblies were BLAST-searched against sequences of Schistosoma japonicum and Trichobilharzia regenti in GenBank and coding rDNA and ITS regions aligned with E-INS-i algorithm of MAFFT; exact boundaries of the ITS1 and ITS2 were identified with the program ITSx 1.0.10 [24]; putative endpoints of lsrDNA and 3′ external transcribed spacer (ETS), as well as the rRNA operon transcription start matching the 5′ end of 5′ ETS, were localised according to [25] and [26]; and finally, the complete annotation was compared with the fully-annotated complete human rRNA repeating unit (Accession No. HSU13369; [27]).

Phylogenetic position of Diplostomum spp. was estimated on the basis of concatenated amino acid and nucleotide sequence data for the two novel and 17 available mt genomes of representatives of both basal (order Diplostomida; seven species) and derived (order Plagiorchiida; ten species) clades of the platyhelminth subclass Digenea, and from an analogous dataset of rDNA sequences (ssrDNA and lsrDNA) made as complete as possible. These included Schistosoma haematobium, S. japonicum, S. mansoni, S. mekongi, S. spindale, S. turkestanicum and Trichobilharzia regenti, all representatives of a single family (Schistosomatidae) of the order Diplostomida, and representatives of six families of the order Plagiorchiida: Clonorchis sinensis, Opisthorchis felineus and O. viverrini (Opisthorchiidae); Dicrocoelium chinensis and D. dendriticum (Dicrocoeliidae); Fasciola gigantica and F. hepatica (Fasciolidae); Haplorchis taichui (Heterophyidae); Paragonimus westermani (Paragonimidae); and Paramphistomum cervi (Paramphistomidae). Sequences representing three species of the platyhelminth class Cestoda (Didymobothrium rudolphii, Diphyllobothrium sp. and Spirometra erinacei; GenBank accession numbers for all taxa provided in Fig. 2) were also included. Since entire mt genomes could not be aligned, all 12 protein-coding gene regions were extracted, nucleotide sequences for individual genes each translated into amino acids using the echinoderm and flatworm mt translation code (NCBI Table 9, [28]), aligned with L-INS-i algorithm of MAFFT, manually curated to exclude ambiguously aligned regions, concatenated, and analysed under the MtZoa and MtREV substitution models [29, 30]. The rDNA dataset was aligned with E-INS-i algorithm of MAFFT, manually curated to exclude ambiguously aligned regions, ssrDNA and lsrDNA partitions concatenated and analysed as a single partition under the GTR + I + Γ substitution model. Phylogenetic analyses of the two genomic loci (conducted at both amino acid and nucleotide levels in the case of the mt data) were carried out individually, under the maximum likelihood (ML) criteria in the program RAxML version 8.1.7 [31]. All model parameters and bootstrap nodal support values were estimated using RAxML; the number of bootstrap repetitions was estimated with the extended majority rule (MRE) bootstopping method [32].

Phylogenetic position of the genus Diplostomum estimated under the maximum likelihood criterion based on analyses of mt protein-coding amino acid data (left) and nuclear rRNA genes (right). Node labels depict bootstrap support calculated from 400 repetitions in RAxML. Rooted phylogram; outgroup: Diphyllobothrium sp., Spirometra erinacei and Didymobothrium rudolphii. Please note the alternative position of Paramphistomum cervi based on the mt amino acid data analysis under the MtREV substitution matrix. Arrows on the rDNA data-based phylogram indicate nodes in conflict with the mt data-based topology

Full size image

Characteristics of the mt genomes

Genomic DNA sequencing of D. spathaceum and D. pseudospathaceum using 1/4 of a MiSeq run each yielded 3.43 and 4.73 million indexed pair-end 250 bp reads, respectively. Of these, two complete mt genomes with overlapping flanking regions were assembled and trimmed to form a consensual circular sequence. The length of the recovered mt genomes of D. spathaceum and D. pseudospathaceum (GenBank accessions KR269763 and KR269764) was 14,784 and 14,099 bp, respectively. The total number of reads mapped on the trimmed mt genomes was 7762 and 5326, accounting for 0.23 and 0.11 % of the genomic readpool obtained, with minimum coverage of 61 and 53 reads, respectively (see Table 1 for further details).

The mt genomes of D. spathaceum and D. pseudospathaceum were fully annotated (Fig. 1), both featuring an identical, complete suite of 12 intron-less protein-coding genes typical for the Platyhelminthes (all lacking the atp8 gene): ATP synthase subunit 6 (atp6), cytochrome b (cob), three subunits of cytochrome c oxidase (cox1–3), and seven subunits of NADH dehydrogenase (nad1–6, nad4L). Further, the two assemblies contained small and large subunits of the mt rRNA (rrnS, rrnL), and a total of 22 transfer RNA genes (including two tRNA copies for amino acids leucine and serine; see inferred secondary structures in Additional file 2: Figure S2 and Additional file 3: Figure S3). All of these genes were encoded unidirectionally, on the plus strand (Fig. 1), in widespread agreement with the protein- and rRNA-coding backbone gene order observed in all other parasitic flatworms, with the exception of the derived African schistosomes ([17]; see positions and sequence lengths of individual genes in Table 2); nad4L overlapped the first 40 nucleotides of nad4 (Table 2). We found no sign of protein-coding gene duplications. The observed adenine + thymine (AT) content of the mt genomes was high, accounting for 69.3 % in D. spathaceum and 70.4 % in D. pseudospathaceum. The most AT-rich protein-coding genes were nad3 (73.7 %) in D. spathaceum and nad4L (75.0 %) in D. pseudospathaceum. In contrast, the lowest AT content was found in cox2 (67.2 and 64.9 %, respectively).

Codon usage of individual mt protein-coding genes is provided in Additional file 4: Table S1. Most mt protein-coding genes of both Diplostomum spp. had ATG as the translation initiation codon whereas nad1, nad2, nad3 and nad5 genes use alternative codons (mostly GTG; Table 2). Surprisingly, nad2 of D. pseudospathaceum starts with a putative TTG, 9 bp upstream of the position of nad2 start codon GTG in D. spathaceum. It seems relatively unlikely that the codon GCA (which corresponds to the position of GTG in D. spathaceum) represents the genuine start codon of nad2 in D. pseudospathaceum. We found no evidence of a sequencing error or a single nucleotide polymorphism in this region based on comparisons with three additional partially sequenced adults of D. pseudospathaceum. However, direct support from mRNA sequences is required before concluding the presence of an alternative initiation codon in Diplostomum spp. Contrary to the start codons, translation termination codons in the two species did not match entirely and the stop codons TAG, TAA and putative truncated T were all found frequently (Table 2).

Genetic divergence of mt genomes and nuclear rRNA operons

A schematic pairwise alignment of the mt genomes of the two species of Diplostomum, depicting the order of the 36 genes and the non-coding regions, is provided in Fig. 1. The entire gene order matched exactly but the two congeneric species differed notably in the total length of the mt circle. The difference stems mainly from the presence of extensive insertions at two loci: a 781 bp long non-coding insertion between rrnS and cox2 genes in D. spathaceum and a 124 bp long insertion in D. pseudospathaceum found immediately upstream of the common non-coding region situated between trnE and trnG genes. The insertion in D. spathaceum mt genome comprises two conspicuously imperfect repeats (pairwise sequence identity of 79.8 %) not present in the D. pseudospathaceum mt genome, and the insert in D. pseudospathaceum represents a partly repeated unit found in the 188 bp long non-coding region common to both species. Although surprising, extensive manual checking of assemblies demonstrated that these inserts are genuine. Analysis of pairwise nucleotide sequence identities of the protein-coding and rDNA regions of the mt genomes revealed a range of sequence conservation from levels as low as 81.3 % in nad4 and 81.7 % in nad5 to as high as 88.9 % in the most conserved gene cox1. The barcode region of cox1 was even more conserved with 90.6 % of nucleotides of the 722 bp long fragment [33] being identical (Fig. 1). The rrnL and rrnS gene nucleotide identities reached 90.1 and 92.6 %, respectively.

The transcribed region of the nuclear rRNA operon, although representing only about half the length of the mt genome, mapped 2.4 and 5.2× more MiSeq reads in D. spathaceum and D. pseudospathaceum, respectively (Table 3). Using the methodological approach of [25] and [26] to supplement a series of multiple sequence alignments and other rDNA annotation strategies (see Methods above), we identified a putative nuclear rRNA operon transcription start site, followed by a 726 bp sequence of 5′ ETS, 1979 bp of ssrDNA, 607–608 bp of ITS1, 157 bp of 5.8S rDNA, 294–295 bp of ITS2, 4210 bp of lsrDNA, and 17 bp long 3′ ETS, forming together a 7991 and 7993 bp long transcribed rDNA region of D. spathaceum and D. pseudospathaceum, respectively (see Table 3, GenBank accessions KR269765 and KR269766). The total length of the entire operon then likely extends over 9000 bp; however, given the limited length (250 bp) of the raw MiSeq reads and the observed presence of multiple repeat motifs in the ribosomal intergenic spacer (IGS) region separating individual rRNA operons, we refrained from assembling the entire rDNA tandem repeat unit and limited our comparisons to the transcribed portion of the region.

Pairwise alignment of rRNA operons revealed remarkably high levels of nucleotide conservation with coding regions reaching from 99.8 % (ssrDNA) and 99.9 % (lsrDNA) to 100 % (5.8S rDNA) of identical nucleotides and high nucleotide identities for the spacers (97.1 % in 5′ ETS, 98.8 % in ITS1 and ITS2, and 100 % in 3′ ETS). Mapping of the raw pair-end MiSeq reads on the rRNA operon under relaxed settings did not reveal any polymorphic sites within the rDNA except for position 3597 in ITS2 of D. spathaceum.

Phylogenetic analyses

Phylogenetic estimates of species of the Digenea based on mitochondrial and rDNA data for 19 species revealed contradictory topologies concerning, most importantly, the phylogenetic placement of the genus Diplostomum (Fig. 2). The analysis based on mt protein-coding genes supported its position as a sister lineage to the order Plagiorchiida (and thus not part of the order Diplostomida), whereas the analysis of nearly complete sequences of rDNA recovered it as a basal lineage of the Diplostomida supporting this branching pattern with maximum bootstrap values. Considering the analyses of mt data, the only topological difference between estimates employing MtZoa and MtREV substitution matrices was the relatively more derived position of Paramphistomum cervi under the MtREV model. Analysis of the same mt data at the level of nucleotides resolved the same placement of Diplostomum as a sister lineage of the Plagiorchiida and yet another alternative position of Paramphistomum cervi as the most basal lineage of the Plagiorchiida (results not shown). In fact, the phylogenetic position of Paramphistomum cervi represents the only other difference between the mt- and rDNA-based trees receiving significant statistical support (it forms a sister lineage to Paragonimus westermani in the latter) and will remain problematic given the low statistical support in either of the analyses.

Mapping the mt gene arrangement across the phylogenetic tree inferred from mt data revealed that both species of Diplostomum exhibit a closer match to the conserved gene order of the plagiorchiid digeneans with only one or two immediately neighbouring tRNAs positions switched (Fig. 3), rather than to the more extensively rearranged order observed in the mt genomes of the representatives of the Diplostomida (basal schistosomes display at least four single tRNA relocations, derived schistosomes a further two shifts of larger genome chunks involving several protein-coding and tRNA genes). The mt gene order in Diplostomum spp. resembles most closely that of representatives of the flatworm class Cestoda, a sister lineage of the class Trematoda (e.g. [34]) used as an outgroup in our analyses (Fig. 3).

Schematic diagram of the mitochondrial gene order of the Trematoda mapped onto the mtDNA amino acid phylogenetic estimate from Fig. 1. Diplostomum spp. gene order highlighted by grey background, altered gene position relative to the Diplostomum spp. indicated in black. Outgroup: Diphyllobothrium spp. and Spirometra erinacei

Full size image

To the best of our knowledge this is the first application of the NGS approach to completely characterise mt genomes and nuclear rRNA operons of a trematode in a single step. Given the fact that within parasitic flatworms (Neodermata) there have been just two monogenean species of Gyrodactylus sequenced using NGS platform (Illumina HiSeq 2000^TM) to develop an automated in silico mt genome assembly approach [18], this is the first use of a NGS platform with limited output to easily characterise multiple-copy genomic loci within parasitic flatworms. We combined a total of four indexed TruSeq Nano library samples (only two samples published here) prepared from untreated DNA extractions on a single Illumina MiSeq run to assess the capability of a simple and straightforward approach to reliably characterise the genomic loci of interest. Of the two loci, nuclear rRNA operons received about five- to nine-fold greater coverage than the mt genomes, whose minimal values in turn were 61 and 53 reads in D. spathaceum and D. pseudospathaceum, respectively. We consider this a sufficiently high coverage that allows for combining two- or three-fold more flatworm specimens on a single MiSeq run using the same library preparation strategy as employed here in order to ensure that characterisation of complete novel mt genomes and nuclear rRNA operons is as cost-effective as possible while receiving sufficient coverage in future studies.

Our study made available reference mt genomes of closely related Diplostomum spp. thus providing a rich resource for future approaches to species delimitation that, in turn, will enable the exploration of molecular epidemiology within a large group of widespread fish pathogens. Recent application of cox1 ‘barcoding’ proved essential in discovering vast previously unrecognised species diversity within the genus [8, 13, 14]. The currently used primers for amplification of the cox1 ‘barcode’ region of [12] work reasonably well and seem to allow for differentiation and identification of a number of Diplostomum spp. However, problematic low divergence has already been demonstrated within three cryptic species complexes, “D. baeri”, “D. huronense” and “D. mergi” (see [8, 14]). We predict that expansion of the exploration of genetic diversity across host populations will inevitably reveal cryptic species thus making species delineation crucial for defining the host ranges and geographical distribution of Diplostomum spp. and the successful development of population genetic studies. Our study profited from the availability of DNA from the adult stages of Diplostomum spp. and taxonomic expertise. Given that most of the recently molecularly delimited species-level lineages of Diplostomum cannot be formally described (awaiting the collection of the adult stages), we see an increasing need to employ the NGS approach presented here to characterise any morphologically determinable Diplostomum specimens available to allow for building a solid baseline for molecular barcoding on which ecological and epidemiological studies could be based in future. Moreover, it is already exciting to visualise an idea of using the outputs of NGS, most easily the mt genome sequences, to test their power in addressing the phylogenetic interrelationships of Diplostomum spp. when analysed at the amino acid level.

Although the currently utilised cox1 ‘barcode’ fragment seems useful in assigning individual isolates of Diplostomum to both described species and novel molecularly defined lineages, its use for species delineation has a number of disadvantages. First, single molecular markers for robust molecular systematic estimates are insufficient, and multi-gene approaches are preferred; the recently applied coalescent-based approaches [14] and the molecular resources provided here hold significant promise for species delimitation within Diplostomum. Secondly, phylogenetic estimates based on the cox1 ‘barcode’ fragment [8, 14] generally lack nodal support at the internal nodes that define the interrelationships of individual lineages and therefore the phylogenetic utility of this region is limited. Finally, and most importantly, we have shown that the cox1 gene in fact represents the most conserved protein-coding region of the mt genome of Diplostomum spp. (Fig. 1), as already shown for other parasitic flatworms (e.g. [35, 36]) and that the ‘barcode’ region is even more conserved. The comparative mt genome assessment allowed us to identify genes with the greatest interspecific variation that are promising for the development of molecular diagnostic markers for species recognition [35]. Certainly the best candidates are the nad4 and nad5 genes that are both relatively long and the least conserved, even when comparing such closely related taxa as D. spathaceum and D. pseudospathaceum, and thus have the potential to add phylogenetic signal and possibly resolve current taxonomic problems. The use of these new markers in large-scale population studies on Diplostomum spp. will offer the advantage of simultaneous species delimitation and assessment of intraspecific genetic variation and thus boost molecular epidemiology studies based on the most accessible life-cycle stages, i.e. the larval forms in the snail and fish populations.

The accelerated development of methods for next-generation biodiversity assessment such as environmental DNA (eDNA) and metabarcoding [37] offer significant promise for large-scale spatial studies related to disease ecology. The first study using water samples to assess the presence of the pathogenic parasite of frogs Ribeiroia ondatrae in wetland habitats indicates the high potential for detecting free-living larval stages of macroparasite infectious agents in the aquatic environment [38]. Our study provides a resource for the design of primer pairs targeting very short DNA sequences that would allow the use of NGS tools to identify unambiguously Diplostomum spp. in both bulk samples (natural assemblages of metacercariae in fish eyes) and freshwater samples containing DNA of the dispersive free-living larval stages (miracidia, cercariae).

The novel data on mt genomes and nuclear rRNA operons of Diplostomum spp. allowed the first genome wide estimation of the phylogenetic relationships of the order Diplostomida, one of the two fundamental lineages from which extant digeneans have diversified [39]. Surprisingly, the currently available mitogenomic data for the Digenea analysed at both the amino acid and nucleotide levels invariably recovered, albeit with much weaker statistical support, the Diplostomidae as a sister lineage of the order Plagiorchiida rather than as a basal lineage of the Diplostomida, the latter inferred by both the benchmark phylogeny of the Digenea of [39] and the current rDNA-based phylogenetic analysis for the set of taxa used in mt genome-based analysis (Fig. 2). This is perhaps the most striking finding given the significance and depth of the basal dichotomy in the Digenea between the Plagiorchiida and Diplostomida on both the molecular [39] and life history levels [40]. If the mt-based phylogeny reflects actual organismal phylogeny, with Diplostomidae as an early diverging lineage within the Plagiorchiida, the seemingly synapomorphic life history characteristics of Diplostomida would need to be reconsidered. For example, the active penetration of a vertebrate host by cercaria in the Diplostomida (with the exception of Brachylaimoidea), or the utilization of tetrapod vertebrates as definitive hosts (with the exception of the Aporocotylidae), would be viewed as plesiomorphies for the Digenea. However, addressing such questions would currently be premature since the mt genomes of Diplostomum spp. could be compared solely with those of taxa representing the most derived lineages of both the Diplostomida and the Plagiorchiida. Additional mt genome data, especially for the earlier diverging lineages of the Diplostomida (Brachylaimoidea, Clinostomidae, Aporocotylidae and Spirorchidae) and Plagiorchiida (e.g. Bivesiculoidea, Transversotrematoidea) would allow testing of this hypothesis.

In light of the conflict between nuclear ribosomal and mitochondrial protein estimates of phylogeny, it may be less surprising that mt genome organisation of Diplostomum spp. also closely matches that of the nearly perfectly conserved mt genomes of representatives of the Plagiorchiida, rather than that of any representative of the Schistosomatidae, the only members of the order Diplostomida with mt genomes characterised prior to our study. Compared with the mt gene order in all species of the Plagiorchiida, there was a single minor alteration in the mt genomes of Diplostomum spp., i.e. a reciprocally switched position of immediately neighbouring genes trnP and trnN (Fig. 3). This tRNA gene position switch likely represents an autapomorphy for Diplostomum, because trnP is always found immediately downstream to trnN throughout remaining trematodes as well as cestodes, the sister lineage of the trematodes (e.g. [34]). Species of the plagiorchiidan genera Dicrocoelium, Paragonimus and Paramphistomum then display a further common switch of neighbouring trnE and trnG gene positions; however, this minor reorganisation does not seem to reflect actual cladogenetic events and might have arisen several times during digenean evolution. Contrary to the members of the Plagiorchiida, species of the Schistosomatidae display a significantly altered gene order, including a few tRNA transpositions (as seen in Trichobilharzia regenti and two Asian schistosomes) and two larger mt genome region rearrangement events within the more derived African schistosomes (see [17, 41] for details). Further mt genome data for a wider range of taxa, especially for representatives of the remaining families of the Diplostomida (see above), would provide important insights into the patterns and possible mechanisms of mt genome evolution in early divergent digeneans.

The genus Diplostomum represents a taxonomically complex group, with unsatisfactorily resolved species diversity, host-associations and distribution patterns. Application of molecular tools offers a breakthrough to overcome previous problems and obstacles caused by the existence of cryptic diversity and the morphological uniformity of larval stages, especially metacercariae in fish. Our results represent a significant step towards a considerably better understanding of the convoluted systematics of the genus and a valuable resource for marker design that will enhance the development of large-scale biodiversity and molecular epidemiology assessments of these important pathogens in the freshwater environment.

This project was supported by the project Postdok_BIOGLOBE (CZ.1.07/2.3.00/30.0032) co-financed by the European Social Fund and the state budget of the Czech Republic (Institute of Parasitology, RVO: 60077344) and the Czech Science Foundation (15-14198S). We thank Kevin Hopkins (Natural History Museum, London, UK) for running the Illumina MiSeq sequencer. Access to computing and storage facilities owned by parties and projects contributing to the Czech National Grid Infrastructure MetaCentrum, provided under the programme “Projects of Large Infrastructure for Research, Development and Innovations” (LM2010005), is also greatly appreciated.

Affiliations

1. Institute of Parasitology, Biology Centre of the Czech Academy of Sciences and Faculty of Science, University of South Bohemia, Branišovská 31, 370 05, České Budějovice, Czech Republic

   • Jan Brabec
   • , Aneta Kostadinova
   •  & Tomáš Scholz

2. Department of Life Sciences, Natural History Museum, Cromwell Road, London, SW7 5BD, UK

   • D. Timothy J. Littlewood

Corresponding author

Correspondence to Jan Brabec.

Competing interests

The authors declare that they have no competing interests.

Authors’ contributions

J.B., A.K. and D.T.J.L. conceived the research; J.B. performed the majority of the laboratory work, analysed the data and drafted the manuscript; D.T.J.L. supported the laboratory work and contributed to data analysis; A.K. identified parasites and revised the first drafts of the manuscript; T.S. supervised the project; all authors commented on the manuscript and revised the article. All authors read and approved the final manuscript.

Reprints and Permissions