We sequenced the complete mitochondrial genome of the flatworm Australopacifica atrata. The species, originally described from New South Wales, Australia, has been found in various locations in the British Isles, New Zealand and in the United States of America; it is thus potentially invasive. The genome is 16513 bp long, encodes for 12 protein coding genes, two ribosomal RNA genes and 20 tRNA genes, and is completely colinear with the other two available Rhynchodeminae. In addition, it shares with them some unusual characters discriminating them from members of the other subfamilies of Geoplanidae, the most noticeable being the extra length of its cox2 gene. The data allow a reliable multigene phylogeny to be derived, and also provide a means of accurate biomonitoring of possible invasiveness by A. atrata.

Biomonitoring, Continenticola, cox2 gene, multigene phylogeny, next generation sequencing, Parakontikia ventrolineata, Tricladida

In the recent years, an increasing number of reports have emerged from Europe and abroad concerning invasive terrestrial flatworms of the family Geoplanidae Stimpson, 1857 (Winsor et al. 2004; Sluys 2016). The most recent classification of the Geoplanidae (Almeida et al. 2021) lists five subfamilies within the Geoplanidae, among which three might be of particular concern in the context of biological invasions. The most infamous species in these subfamilies are probably the ‘hammerhead flatworm’ Bipalium kewense Moseley, 1878 (Bipaliinae Von Graff, 1896) from South East Asia (Winsor 1983; Justine et al. 2018), followed by Obama nungara Carbayo, Álvarez-Presas, Jones & Riutort, 2016 (Geoplaninae Stimpson, 1857) from South America (Carbayo et al. 2016; Lago-Barcia et al. 2019; Justine et al. 2020b; Fourcade 2021) and Platydemus manokwari de Beauchamp, 1963 (Rhynchodeminae Von Graff, 1896) from Papua New Guinea (Justine et al. 2014; Justine et al. 2015; Gastineau et al. 2020; Justine et al. 2021). With their large size and their feeding habits, these predators of soil invertebrates have attracted most attention.

Two species, smaller than the aforementioned large ones and with a mostly scavenging behaviour, are also potentially invasive (Winsor et al. 2004). The first is Parakontikia ventrolineata (Dendy, 1892) Winsor 1991; the species is about 30 mm in length with a dark grey-black body, with longitudinal black stripes on the dorsum, and with the eponymous paired greyish stripes on its ventral surface. Originally from Queensland, Australia, it has been found elsewhere in Australia, and has also been reported in France (Justine et al. 2014; Gastineau et al. 2020), Spain (Álvarez-Presas et al. 2014) and possibly in South Africa (Jones et al. 1998). Also New Zealand, Hawaii, United Kingdom, Madeira, USA, and Mexico (records summarized in Winsor et al. 2004), Costa-Rica, Colombia, Ecuador, Chile, Argentina, and confirmed for South Africa (iNaturalist records, https://www.inaturalist.org/). The second species is Australopacifica atrata (Steel, 1897). Originally described as Geoplana atrata from New South Wales, Australia, it was transferred to Parakontikia on the basis of its external morphology and internal anatomy (Winsor 1991), and subsequently to Australopacifica (Jones, 2019). The formal definition of Australopacifica is “Geoplanidae, but not classifiable into the present taxonomic genera because of insufficient morphological information; geographical distribution largely in Australasia and Indo Pacific islands. A collective group to temporarily assign species inquirendae and nomina dubia” (Ogren and Kawakatsu 1991). It is also found in other Australian states and territories: Victoria (Winsor 1973), Queensland (Winsor 1997), the Australian Capital Territory, and Tasmania (Atlas of Living Australia records, https://www.ala.org.au/). The species has been recorded in New Zealand (Winsor et al. 2004), in Wales and South England (Jones 2019), and in California, USA (iNaturalist records). The species superficially resembles Pa. ventrolineata, except that it has a uniformly black dorsum, with a characteristic median dark stripe on its ventral surface (Figs 1, 2). Both species are gregarious, collectively predating upon slugs and snails and other soil species such as isopods (Smith 1979; Barker 1989; Winsor et al. 2004), and are also necrophagous scavengers (Smith 1979; Jones 2019). They have been recorded in France and the UK on various fallen fruit lying on the soil, and in cavities in growing strawberries; whether the planarians are simply attracted to rotting fruit or actually feeding on the fruit remains has yet to be adequately explained (Jones 2019; Justine et al. 2020b).

Figure 1. 

Living flatworm Australopacifica atrata (Steel, 1897), showing underside of mid body.

In recent years, several complete mitochondrial genomes from different species of terrestrial invasive flatworms have been sequenced (Solà et al. 2015; Gastineau et al. 2019; Gastineau et al. 2020; Gastineau and Justine, 2020; Justine et al. 2020a; Justine et al. 2022). Full mitogenomes provide a complete sequence for the widely used molecular barcode COI or cox1 (the gene of the cytochrome c oxidase subunit 1) and allows us to perform robust multigene phylogenies. Full mitogenomes may provide additional information such as those related to the gene order or the presence of pseudogenes resulting from duplications. It also prevents amplification of nuclear copies of mitochondrial pseudogenes by PCR, a problem that sometimes occurs in molecular barcoding (Song et al. 2008). In the case of terrestrial flatworms, earlier reports have noted that there are mitogenomic features common to several species of the same subfamily that are not present in other subfamilies (Solà et al. 2015; Gastineau et al. 2020; Gastineau and Justine 2020; Justine et al. 2020a). However, the number of available mitogenomes remains still too low to allow for more general conclusions. Prior to this study, GenBank did not contain any sequence corresponding to the genus Australopacifica.

Figure 2. 

Living flatworm Australopacifica atrata (Steel, 1897), dorsal view.

Here we report the complete mitochondrial genome of a morphologically identified specimen of A. atrata and compare it with other species. We provide further evidence concerning the specific features of the mitogenomes of Rhynchodeminae compared to land flatworms in other subfamilies. We present a multigene phylogeny demonstrating its proximity with Pa. ventrolineata. Finally, we discuss the next steps that should take place in the investigation of invasive terrestrial flatworms by means of next generation sequencing.

The mitogenome of A. atrata is shorter than those of the two other Rhynchodeminae Pl. manokwari (19959 bp, MT081580) and Pa. ventrolineata (17210 bp, MT081960), but larger than the largest recorded for a Bipaliinae, Bipalium vagum Jones & Sterrer, 2005 (17149 bp, MZ561468) or the Geoplaninae, O. nungara (14909 bp, KP208777) and Amaga expatria (14962 bp, MT527191). However, the presence of repeated sequences in the mitogenome of Geoplanidae has been strongly suggested (Solà et al. 2015). These repeated portions cannot be fully resolved by short reads sequencing, so the criteria of the length of the mitogenome should be taken with care.

So far, the Rhynchodeminae can be differentiated from Bipaliinae and Geoplaninae by the position of the t RNA-Cys. For A. atrata, Pl. manokwari and Pa. ventrolineata, it is located between the protein coding genes cox3 and atp6, clustering with t RNA-Ile, t RNA-Gln and t RNA-Lys, as the first t RNA of this cluster. For B. kewense as well as O. nungara and A. expatria, it is located between the protein coding gene ND2 and the 12S rRNA gene, clustering with t RNA-Met and t RNA-His, located after these two t RNA. Another difference regarding t RNA is the apparent lack of a tRNA-Thr among all Rhynchodeminae. Among the other species, t RNA-Thr has been found between the 16S rRNA gene and the protein-coding cob gene, clustering with t RNA-Leu and t RNA-Asn. It is worth emphasising here that there is a difference between the Geoplaninae, in which the order of this cluster is 16S, t RNA-Thr, t RNA-Leu, t RNA-Asn, cob and B. kewense, in which t RNA-Leu and t RNA-Thr exchange their positions. However, it might be noted that for the recently described species of Bipaliinae, Humbertium covidum Justine, Gastineau, Gros, Gey, Ruzzier, Charles & Winsor, 2022 (Justine et al. 2022), it was also impossible to find a tRNA-Thr. Therefore, it would be better to take a conservative view of this feature, and not to assign it a too high value for classification as a molecular synapomorphy.

Several studies failed to find a start codon for various mitochondrial genes of terrestrial flatworms such as O. nungara, A. expatria, Diversibipalium multilineatum Makino & Shirasawa, 1983 and B. vagum (Solà et al. 2015; Justine et al. 2020a, 2022). This also seems to be the case for other members of the Continenticola, including Planariidae and Dugesiidae such as Crenobia alpina Dana (KP208776) (Solà et al. 2015), Schmidtea mediterranea Benazzi, Baguñà, Ballester, Puccinelli & Del Papa (JX398125) (Ross et al. 2016), Dugesia ryukyuensis Kawakatsu (AB618488), and Dugesia japonica Ichikawa & Kawakatsu (AB618487) (Sakai and Sakaizumi 2012).

For all three Rhynchodeminae, the ND5 gene has an early termination because of the presence of tRNA-Ser immediately following the last T residue of the gene, for which we suspect that the functional TAA stop codon is obtained by being completed by the addition of 3’ A residues to the mRNA, while a canonical stop codon was found for all other species. Also, it is interesting to note that there is an overlap between the ND4L and the ND4 genes, and that the size of this overlap is always 32 bp. This character, which is common to Pl. manokwari and Pa. ventrolineata, has been also reported in O. nungara and C. alpina. A rapid investigation of the mitogenomes of all available Bipaliinae shows that this overlap is totally lacking. Trying to simulate this overlap by extending the open reading frame at its N terminal ending leads to a fictious polypeptide with no initial methionine but that may comprise several stop codons in its early part.

As was observed in both the other Rhynchodeminae, the cox2 gene has an important extra length. This extra length does not result from a missing stop codon, as it is located in the middle of the gene, and not on the 3’ extremity. The size of the cox2 putative protein of A. atrata is nearly identical to those of Pa. ventrolineata (434 and 433 amino acids respectively) (Table 1). The size is comparable to Pl. manokwari (452 amino acids), and far bigger than those observed among Geoplaninae and Bipaliinae, where this size ranges from 225 to 260 amino acids. It is noteworthy that Girardia spp. also display substantially larger cox2 genes (KP090061, MW972220). However, an alignment performed with all the amino-acid sequences showed that this extra length is not located in the middle of the ORF, as for Rhynchodeminae, but at the C terminal ending. As verified on CDD/SPARCLE (Marchler-Bauer et al. 2017; accessed on 08/24/2021), the last conserved domain of the putative protein of Girardia spp. seems to be approximately at the amino-acid 227, for a total length of the predicted protein of 389 amino acids. CDD/SPARCLE positions this same conserved domain at the amino-acid 348 in A. atrata, for a total length of the predicted protein of 434 amino acids.

Following this work, we see three major paths of investigation. The first one will continue to focus on Rhynchodeminae. It will be interesting to compare all the characters considered here (colinearity, composition in t RNA, type of termination for the ND5 gene, overlap between ND4L and ND4 and of course the extra length of the cox2 gene) with more species. Of the five tribes of Rhynchodeminae, we have now sequenced representatives of two: the Rhynchodemini (i.e. Pl. manokwari) and the Caenoplanini (Pa. ventrolineata and A. atrata). However, living examples of many of the other species assigned to the remaining tribes will be difficult to find considering their origin and repartition, but at least one of them, Anzoplanini Marionfyfea adventor Jones & Sluys, is present in Europe (Jones and Sluys 2016). Also, Anzoplana trilineata Winsor, can be found in south-eastern Australia (Winsor, 2006). Thus, we hope to obtain specimens of both species in the near future. The second would be to extend the search of shared mitogenomic characters among the subfamilies which have not been investigated until now, such as Timyminae and Microplaninae. The third and last path deals with the relationships between the genera Parakontikia and Australopacifica. Our results support the earlier provisional classification of A. atrata within the genus Parakontikia (Winsor 1991). We provide here the first sequence for a species presently assigned to the genus Australopacifica, which contributes towards the resolution of systematic relationships of species currently included in this heterogeneous collective genus.

We are most grateful to Mitch Findlater-Smith for collecting specimens of A. atrata, and to Reiner Richter for kindly allowing us permission to use his photographs of A. atrata supplied as ARR05976.JPG and ARR05978.JPG under the CC-BY 4.0 license. ProSciTech Pty. Ltd., Townsville, Queensland, is thanked for generously donating consumables for specimen collection. We also thank Prof. Hugh D. Jones and Dr Marta Álvarez-Presas for kindly sharing with us their molecular barcoding results.