A complete mitochondrial genome of the Roman snail Helix pomatia Linnaeus, 1758 has been sequenced. The length and gene order correspond to that of other available helicid mitogenomes. We used the mitogenome sequence to reappraise the relationships among the four presumed principal groups of the helicid subfamily Helicinae. The results support the idea that the subfamily is divided between two western Palaearctic diversification centres: Iberian Peninsula and western Maghreb in the west, and Anatolia, the Aegean and Caucasus in the east. One group, the tribe Helicini, diversified in the east and the remaining three currently recognised tribes in the west. However, the exact relationships among lineages of the non-Helicini tribes could not be resolved.

Allognathini , biogeography, Helicini , land snail, Mediterranean, mitochondrial genome, Otalini , phylogeny, Thebini

Phylogenetic research on land snails has been thus far dominated by studies based on mitochondrial markers, mostly partial sequences of the cox1 (COI) and rrnL (16S rRNA) genes, not exceeding 2000 bp in total. These genes dominate not only in terms of the amount of sequences generated, but usually also by the variability of the sequences used if combined with sequences of nuclear loci. Although modern techniques targeting many loci simultaneously begin to advance also into the phylogenetics of pulmonate land snails (Teasdale et al. 2016), the mitochondrial markers will likely continue to be used not only due to relatively low costs and accessibility of the methods, but especially thanks to the wealth of previously published data. Mitochondrial genome sequences may then become handy for primer design, to facilitate assembly of further mitochondrial genomes from outputs of high-throughput sequencing, or directly as the data for phylogenetic analyses. In fact, one of the best resolved backbone phylogenies of Stylommatophora to date was reconstructed from mitochondrial genome sequences (Groenenberg et al. 2017).

We sequenced a complete mitogenome of Helix pomatia Linnaeus, 1758, a large and common European edible land snail species, to provide a reference for further work focused on the family Helicidae Rafinesque, 1815. We then used other helicid mitogenomic sequences to evaluate the support for the basal branching order in the subfamily Helicinae, which was reported by Neiber and Hausdorf (2015) from analyses based on partial sequences of rrnL and cox1 genes along with part of the nuclear rRNA gene cluster (5.8S rRNA, internal transcribed spacer 2, 28S rRNA).

The subfamily Helicinae includes several charismatic species, which are large, common, and edible. As the most prominent, we may mention Cepaea nemoralis (Linnaeus, 1758), famous for evolutionary studies of its colour polymorphism (e.g. Cain and Sheppard 1950, Jones et al. 1977, Silvertown et al. 2011), Cornu aspersum (Müller, 1774), the most commonly reared and consumed helicid species which became a pest in several regions around the globe (e.g. Rees 1955, Ansart et al. 2009), and Helix pomatia, a characteristic snail of Central Europe and the type species of the family’s type genus. Each of these three species represents a different lineage of the three presumed clades constituting the Helicinae, which were proposed and assigned a tribe status in the first comprehensive molecular phylogenetic study of Helicoidea by Razkin et al. (2015). Cepaea represents Allognathini Westerlund, 1902, HelixHelicini Rafinesque, 1815, and CornuOtalini Pfeffer, 1930. Razkin et al. (2015) also recognized a fourth tribe, the monotypic Thebini Wenz, 1923 represented by Theba Risso, 1826. The phylogenetic position of Theba has been contentious (Perrot 1939, Gittenberger and Ripken 1987, Bouchet and Rocroi 2005, Schileyko 2013, Razkin et al. 2015), but Helix, Cornu and Cepaea were generally considered closer to each other than to Theba, for which a separate tribe or even subfamily has been erected.

Neiber and Hausdorf (2015) used a more informative dataset than Razkin et al. (2015) and found Cepaea, Cornu and Theba to be more closely related to each other than to Helix and some other eastern Mediterranean genera. Their results suggest a principal biogeographic division within Helicinae, running approximately along the Apennine Peninsula and separating two equally old clades with centre of diversification in the western (Cornu, Cepaea and Theba, with their respective relatives) and the eastern (Helix and relatives; Korábek et al. 2015) Mediterranean basin, respectively. Considering the Paleogene/Neogene palaeogeographic evolution of the western Mediterranean (Giusti and Manganelli 1984, Rosenbaum et al. 2002, Advokaat et al. 2014, van Hinsbergen et al. 2014), it may be expected that the western taxa are indeed more closely related to each other.

Alternatively, the eastern group may be an offshoot of the western diversity, if the western group is paraphyletic. The systematic concept and monophyly of Otalini remains an open issue, as they (in the sense of Neiber and Hausdorf 2015) may be paraphyletic to both Thebini and Allognathini, and the east-west divide has not been unambiguously supported. The absence of an east-west divide in the phylogeny would lend support to a Western European origin of the whole subfamily Helicinae and to the north-western Maghreb being a refugium of its oldest and perhaps phylogenetically most distinctive lineages. Apart from biogeographic implications, the branching order at the base of the Helicinae has for us also a practical significance for the selection of appropriate outgroup taxa for analyses of relationships between Helix and related genera.

There are already published complete mitochondrial genome sequences of Ce. nemoralis (Terrett et al. 1996; NCBI accession number NC_001816), Co. apersum (Gaitán-Espitia et al. 2013; NCBI accession number NC_021747) and Theba pisana (Müller, 1774) (Wang et al. 2018b; Genbank accession number MH362760), as well as of a convenient outgroup species Cylindrus obtusus (Draparnaud, 1805) (Groenenberg et al. 2012; NC_017872). Combined with the newly sequenced mitogenome of Helix pomatia they offered an opportunity to verify the relationships between the four Helicinae tribes.

We have sequenced a complete mitogenome of H. pomatia, from a specimen representative of a common central-European lineage (Korábek et al. 2018) of this broadly distributed snail species. The length and gene content and order (Table 2) correspond with those of mitogenomes of Ce. nemoralis, T. pisana and Co. aspersum, the other Helicinae species with available mitogenome sequences.

The phylogenetic analyses in all cases recovered the expected split between the eastern Helix and the western group of Theba, Cepaea, Cornu, assuming Cylindrus as an outgroup (Figure 1). This grouping received equivocal support in the analysis based on the nucleotide data, but it increased with amino acid data and further when using site-heterogeneous model. The support thus increased as the analyses got less sensitive to substitution saturation and among site heterogeneity. In the western group, the results united Theba and Cepaea to the exclusion of Cornu, but this branch received positive support only from bootstrap with the nucleotide data. Especially analyses of the amino-acid alignment with the site-heterogeneous model based on alternative guide trees resulted in very low support for the branch uniting Cepaea and Theba. The Bayesian analysis of nucleotide data with P4 recovered the same topology as the maximum likelihood analysis but with full support, regardless whether homo- or heterogeneous model was used.

Figure 1. 

The inferred phylogenetic relationships between the five helicid mitogenomes available. Branch supports are given for maximum likelihood analyses based on nucleotide data, amino acid data under homogeneous model, and amino acid data under site-heterogeneous model (in this order). The latter was run under alternative settings, see details in the text. Shimodaira–Hasegawa-like approximate likelihood ratio test (SH-aLRT) and standard bootstrap percentages (BP) are reported. Values of SH-aLRT >90 % and BP >75 % are considered positive support. Cylindrus is included as an outgroup.

Because Theba and Cepaea had the longest branches in the tree, which had short internal branches, and there is no independent indication of the relationship between Theba and Cepaea, we suspected the result to be caused by the LBA (Bergsten 2005). Theba is annual to biannual and semelparous (Heller 1982, Cowie 1984), unlike the other helicids, which are more long-lived (Pollard 1975, Bisenberger et al. 1999; Ansart et al. 2009). The genus may generally cause problems in phylogenetic inference if the difference in life history results in differece in substitution rate (Thomas et al. 2010; cf. Saclier et al. 2018). In addition, analysis of the nucleotide composition of the five taxa revealed substantial differences, where Cornu (GC 30.5%) and Cepaea (GC 41.6%) differed the most, which could obscure potential relationship between the two species. The low support with presumably more robust methods suggests that the relationship between Cepaea and Theba could be an artefact.

Our results are consistent with the hypothesis that Helicinae are principally divided into a western (mainly Iberia, western Maghreb, Macaronesia) and an eastern (Caucasus, Anatolia, Greece) group (Neiber and Hausdorf 2015). The split between these two lineages probably occurred no later than during the Late Eocene‒Early Oligocene (Neiber and Hausdorf 2015), but the east-west pattern in the distribution of the two lineages persists despite 30 million years having elapsed since the split. Only two Helix species naturally represent the eastern group west of ca 9°E (one in Europe, one in Africa; Neubert 2014, Korábek et al. 2018), and two species of Cepaea (Europe; Welter-Schultes 2012) and few species of Eremina Pfeiffer, 1855 (northern Africa; Ali et al. 2016) represent the western group east of 18°E.

Despite analysing a substantially higher number of genes than Razkin et al. (2015) and Neiber and Hausdorf (2015), we also could not resolve the relationships between Cornu, Cepaea and Theba. We assume that additional lineages should be analysed in order to resolve all the major phylogenetic and biogeographic problems within the western branch of Helicinae. These include Macularia Albers, 1850, whose phylogenetic position is equally problematic as that of Theba (Nordsieck 1987; Schileyko 2006; Neiber and Hausdorf 2015), the closest relative of Theba (perhaps Eremina, but see Holyoak et al. 2018), and the type genera of the tribes Otalini and Allognathini. Nevertheless, it is likely that only analyses of multiple nuclear loci will yield robust estimates of the basal relationships of Helicinae due to high and lineage-specific evolutionary rates of the mtDNA and the saturation of nucleotide substitutions.

We thank Tereza Kosová for help in the lab, Lucie Juřičková for financial support, and Marco T. Neiber for review and language polishing. Access to computing and storage facilities owned by parties and projects contributing to the National Grid Infrastructure MetaCentrum provided under the programme “Projects of Large Research, Development, and Innovations Infrastructures” (CESNET LM2015042) is greatly appreciated. MR was supported by Charles University Research Centre program (204069).