Introduction

The Spanish toothcarp (Aphanius iberus Valenciennes, 1846) is a small fish species endemic to the Mediterranean Coast of the Iberian Peninsula [1,2,3]. It belongs to the Cyprinodontiformes order [4] within the family Aphaniidae and although it can tolerate varying environmental conditions, including high temperatures and elevated salt concentrations [5], its populations have been in decline due to urbanism and agricultural exploitation which have dramatically transformed the eastern landscape of the Iberian Peninsula. Due to its fragmented distribution and its severe population decline in recent years as a result of habitat destruction, the Spanish toothcarp is currently listed as “Endangered” (IUCN: EN) [3, 6, 7], with about 20 isolated populations along the Iberian Mediterranean slope and it is currently restricted to salt marshes, coastal lagoons and natural springs referred to locally as Ullals [1, 3, 8]. To aid in the recovery of the species, conservation programs including two EU LIFE Projects (LIFE96 NAT/E/003118; LIFE04/NAT/ES/000035) are dedicated to the restoration of the natural habitat of A. iberus and the implementation of captive breeding programs. Currently, there is debate regarding the genus assignment of this species where some authors argue its belonging to the genus Apricaphanius Freyhof & YoĞurtÇuoĞlu [9], while others say it belongs to the genus Aphanius Nardo, 1827 [10]. Until further evidence arises, we have opted for the more conservative option in order to remain consistent with the International Code of Zoological Nomenclature, and therefore the genus, in this study, is referred to as Aphanius. We have also maintained Aphanius farsicus over Esmaeilius persicus which was recently proposed by Freyhof & YoĞurtÇuoĞlu [9].

Genetic studies involving A. iberus have mainly been focused on evaluating the genetic structure of the species under a conservation framework in order to identify its potential conservation units either with allozymes [8], partial gene sequences [11,12,13,14] or microsatellites [15]. However, the phylogenetic relationships within the order Cyprinodontiformes are still controversial and not yet clear especially since very few complete mitochondrial genomes are available causing some inconsistencies in relation to the phylogenetic position of Aphanius [10, 16, 17]. In this study, the entire mitochondrial genome of A. iberus was obtained for the first time in the context of an A. iberus reference genome sequence and a phylogenetic tree was constructed within the Cyprinodontiformes order. This new data will not only be valuable for improving our understanding of the evolutionary history of the species and its phylogenetic relationships, but also for future evolutionary and genomic studies.

Methods

Three specimens of A. iberus were collected in El Palmar (Valencia, Spain) and euthanized following the appropriate protocols. The DNA was then isolated using the MagAttract HMW DNA isolation kit (Qiagen). The final elution was done in a volume of 100 μL. DNA was quantified using the Qubit High Sensitivity dsDNA Assay (Thermo Fisher Scientific).

Sequencing data within the context of a reference genome was obtained from PacBio and Illumina DNA sequencing in order to obtain the adequate assembly and annotation. Prior to PacBio library preparation, the sample was further purified and size-selected in order to maintain the largest fragments. Next, the recommended SMRTbell Express Template Prep Kit 2.0 (PacBio) was used to prepare the library, following the manufacturer’s instructions. The library was sequenced in a Sequel II sequencer (PacBio), using a SMRT Cell 8 M, under the Long-reads mode. A total of 7.3 million long reads (~ 9500) were obtained and quality-checked using the software SequelTools [18]. For posterior Illumina library preparation, the Illumina DNA Prep kit was used strictly following the manufacturer’s instructions. The fragment size distribution and concentration of the library was checked in the Agilent 2100 Bioanalyzer (using the Agilent HS DNA Kit). Then, the library was sequenced in a fraction of a NovaSeq PE150 flow cell, aiming for a total output of 50 gigabases, that yielded close to 435 million short paired-end reads (~ 150). The raw fastq files were quality-checked using the software FastQC v0.11.5 [19].

For the mitochondrial genome assembly NOVOPlasty v4.2 [20] was used as a seed-extend based assembler to reconstruct organelle genomes from whole-genome sequencing data, starting from a related or distant seed sequence. The Cytochrome C Oxidase Subunit 1 (COX1) gene of the species Aphanius vladykovi (NCBI Reference Sequence: MN702439.1) was selected as the seed, with a k-mer length of 33 bp.

The quality of the resulting mitochondrial assembly was then evaluated with the package QUAST 5.0.2 [21]. The number of contigs and the total length (in base pairs, bp) of the mitochondrial genome assembly are represented in Table 1. The mitogenome assembly generated was queried against the NCBI’s (National Center for Biotechnology Information) nr/nt (nucleotide) database using the Basic Local Alignment Search Tool BLAST. The best match found was the cyprinodontiform species Cyprinodon variegatus (Accession Number: KT288182.1) with a nucleotide identity percentage of 82% and the closely related Orestias ascotanensis exhibited a nucleotide identity percentage of 81.5%.

Table 1 Number of contigs and their total length
Full size table

The Mitochondrial genome annotation Server (MITOS2) [22] was used to automatically annotate the mitochondrial genome selecting RefSeq 63 Metazoa as the reference database and the vertebrate mitochondrial genetic code. The protein prediction method from Al Arab et al. [23] was enabled.

Results

The mitogenome of Aphanius iberus is a circular doble-stranded DNA sequence that is 16,708 bp long including 13 protein-coding genes, 22 transfer RNA genes, 2 ribosomal RNA genes, and the putative control region (Table 2, Figs. 1 and 2). The base percentage composition showed smaller G+C content (45.06%) compared to A+T content (54.94%). MITOS2 software annotated several peculiarities and found some OH overlaps in the following genes: (atp8,atp6):10; (nad4l,nad4):7; (nad5,nad6):4; (nad3,trnR):2; (trnI,trnQ):1; (trnQ,trnM):1; (nad2,trnW):1; (atp6,cox3):1; (cox3,trnG):1; (trnT,trnP):1.

Table 2 Organization of the Aphanius iberus mitogenome
Full size table
Fig. 1
figure 1

A graphical map of the mitochondrial genome of A. iberus showing the GC content and annotation results using the CGView server [24] (see Figure)

Full size image
Fig. 2
figure 2

Linear representation of the A. iberus mitochondrial genome

Full size image

The mitochondrial genome was deposited on GenBank with the Accession Number OP884090; (BankIt2647138 Seq1). The Whole Genome Shotgun project has been deposited at DDBJ/ENA/GenBank under the accession JAPXFQ000000000. The version described in this paper is version JAPXFQ010000000. BioProject: PRJNA913687. BioSample: SAMN32303939.

A phylogenetic analysis was performed and included both coding regions (Nad1, Nad2, Cox1, Cox2, Atp8, Atp6, Cox3, Nad3, Nad4L, Nad4, Nad5, Nad6, CytB) and non-coding regions 12s (rrnS) and 16s (rrnL). Eighty-three cyprinodontiform species were selected and the mentioned mitogenome regions were downloaded from GenBank (Table 3) with the aim to evaluate the phylogenetic position of the mitogenome of Aphanius iberus. The following outgroup taxa were selected: Oryzias uwai (Accession Number: MN832874) and Bedotia geayi (Accession Number: AP006770) from the closely related Beloniformes and Atheriniformes orders respectively [25]. Geneious software [26] was used to align sequences using the MUSCLE alignment method including all of the 83 cyprinodontiform species, the two outgroup taxa, and the new mitogenome obtained for Aphanius iberus (Table 3). PartitionFinder2 software [27] was used to search for the best evolutionary model for each gene separately. The results revealed GTR+I+G as the best fit model for each gene, therefore, the entire mitogenome alignment was considered to belong to only one partition. Afterwards, two phylogenetic approximations were conducted based on two different methods. First, the Maximum-Likelihood reconstruction analysis was performed using the option MFP + MERGE in the IQ-TREE software [29]. The support for each node was evaluated with the SH-like approximate likelihood ratio test [30] and 1.000 ultrafast bootstrap (UFBoot2) approximations [31]. Then a Bayesian inference was performed with MrBayes [32]. Two analyses were run for 10,000,000 generations simultaneously, each with two parallel runs and four MCMC chains with a sampling frequency of 1000 generations. We rejected the first 25% of generations as burn-in and obtained the 50% majority rule consensus tree and the posterior probabilities (PP). The convergence of the runs was corroborated using Tracer v1.7.1 [33]. Finally, both phylogenetic trees inferred were imported into the software FigTree [34] and presented together with bootstrap values (over 100) and Bayesian probabilities (over 1) as the branch support (Fig. 3).

Table 3 Species selected from Genbank and classified by order and family with their respective accession numbers, authors and submission dates
Full size table
Fig. 3
figure 3

Phylogenetic tree rendered by Maximum likelihood and Bayesian inference based on the complete mitochondrial genome of 83 cyprinodontiform fishes. The phylogenetic position of Aphanius iberus is highlighted in red. The numbers on the branches indicate bootstrap (top of the branch) and posterior probability (bottom of the branch) values. (Color figure online)

Full size image

Discussion

The phylogenetic analysis suggests the close relationship of Aphanius iberus with Orestias ascotanesis. The evolutionary relationship between A. iberus and O. ascotanensis exhibits high branch support with a bootstrap value of 73 and a posterior probability value of 0.99 (Fig. 3). This result was congruent with the traditional morphometric reconstructed trees [35,36,37,38] based on recent molecular studies [10, 17, 25] and with a geological vicariant hypothesis involving Cyprinodontiformes [16]. All of them revealed Aphaniidae (Aphanius iberus) as a monophyletic family separated from the Cyprinodontidae family (Cyprinodon and Jordanella genera, both included in our phylogenetic tree) [4] with a close proximity to Valenciidae [39]. Recently the complete mitogenome of Aphanius farsicus was published by Teimori & Motamedi, where they show A. farsicus positioned within the same evolutionary clade as the Cyprinodon and Jordanella genera [40]. Despite the genetic proximity of A. farsicus with A. iberus postulated by some studies [9, 10, 17], our analysis corroborated the close and highly supported mitochondrial relationship between A. farsicus and the Cyprinodon-Jordanella clade instead of with A. iberus. This latter issue questions the monophyly of the Aphaniidae family and poses the need for a further revision of this particular fish family within the order Cyprinodontiformes.

Mitochondrial DNA has proved to be a useful marker for deciphering the phylogenetic relationships at both inter- and intraspecific species levels [41]. On the one hand, its relatively fast evolutionary rate and nucleotide polymorphism has enough resolution to identify species and to even differentiate genetic groups within species. On the other hand, DNA recombination and the lack of complex genomic structures found in nuclear DNA (e.g. repeated elements, pseudogenes or introns) make for straightforward analyses. In fact, one of the mitochondrial genes, the cytochrome c oxidase subunit 1, has been designated as a barcode for species identification in many taxonomic studies [42, 43], but not without debate [44]. Some limitations of its use have been reported in the literature, including discrepancies with nuclear phylogenies [45, 46] or difficulty to identify hybridizations when these evolutionary events are present due to the maternal inheritance of the mitochondrial genome [47]. For these reasons, and due to the fact that only a few complete mitogenomes have ever been published for the genera Aphanius (two mitogenomes) and Orestias (one mitogenome), more research should be done in order to clarify the phylogenetic relationships of these genera and their species within the Cyprinodontiformes order in an evolutionary context.

Conclusion

In this study, we have revealed that the genus Aphanius s.l. is not monophyletic. We have additionally proposed the close relationship between the western Mediterranean species Aphanius iberus and Orestias ascotanensis, a species from the Andean Region in South America. Our results are congruent with previous phylogenetic and biogeographical vicariant hypotheses involving Cyprinodontiformes. However, due to limitations of the mitochondrial genome and the fact that only a few were analyzed, further studies are required. Nevertheless, the information gained from this study is valuable for improving our understanding of the evolutionary history of A. iberus and for future genomic studies.