Aim: Freshwater planarians may have a wide geographic range despite their assumed low vagility. Found across four continents, Dugesia may have either an ancient origin on a large paleo landmass, followed by colonisation in different regions before continental fragmentation, or a more recent origin and subsequent transoceanic dispersal. We seek to resolve between these two hypotheses.

Location: Africa, Eurasia, and Australasia

Taxon: Genus Dugesia (Platyhelminthes: Tricladida: Dugesiidae)

Methods: We used data from the sequencing of six gene fragments and comprehensive taxonomic sampling of Dugesia from across its distribution range to reconstruct the phylogeny of this genus using maximum likelihood and bayesian inference methods. We conducted two phylogenetic dating analyses using Platyhelminthes fossils and palaeogeological events. Basing on the time-calibrated molecular phylogenetic framework we evaluated the contribution of vicariance and dispersal to the biogeographic evolution of Dugesia. By reconstructing the ancestral areas and present-day potential distribution using BioGeoBEARS and niche modelling, we elucidated the biogeographic history of the genus.

Results: The present-day distribution of Dugesia is a result of different vicariance and dispersal events. However, we also found evidence of transoceanic dispersal. Consistent with previous hypotheses, Dugesia dates to the Upper Jurassic in the Afro-Malagasy Gondwana region. We unveiled a novel biogeographic scenario for the genus, involving multiple events of colonisation in Eurasia from continental Africa via at least three dispersal routes.

Main conclusions: Dugesia is an ancient genus having reached its present distribution through a complex history of dispersal and vicariant events following its origin in southern Gondwana. Despite the low vagility of Dugesia, we found evidence of their overseas dispersal.

DNA extraction, PCR and sequencing. Total genomic DNA extraction was carried out using either DNAzol® (Molecular Research Center Inc. Cincinnati, OH) or the Wizard® Genomic DNA Purification Kit (Promega, Madison, WI, USA), following the manufacturer's instructions. The following gene fragments were amplified by polymerase chain reaction (PCR): (1) the mitochondrial gene cytochrome c oxidase subunit I (Cox1); from the nuclear ribosomal cluster (2) 18S ribosomal gene (18S), (3) 28S ribosomal gene (28S), and (4) the ribosomal internal transcribed spacer-1 (ITS−1); finally, two other anonymous nuclear makers coined elsewhere as (5) DUNUC3 and (6) DUNUC5 (Leria et al., 2020. https://doi.org/10.1016/j.ympev.2019.05.010). 

Alignment. Nuclear genes were aligned using the online software MAFFT version 7 under the G-INS-i algorithm (Katoh & Standley, 2013. https://doi.org/10.1093/molbev/mst010). The intronic region of DUNUC5 was curated by eye after alignment using Geneious 10.2.3 (https://www.geneious.com,  Kearse et al., 2012). The Cox1 sequences were translated into amino acid sequences (genetic code 9 in NCBI) to verify the absence of stop codons. Thereafter, Cox1 sequences were aligned using the translation align function in Geneious 10.2.3 ('MAFFT alignment' option). All alignments were also inspected by eye in Geneious 10.2.3.

Four datasets were prepared for the different analyses. Dataset I included 18S, 28S, DUNUC3, Cox1 for Dugesia representatives and used Recurva and Schmidtea as outgroups, to root the Dugesia (Sluys et al., 2013. https://doi.org/10.1111/zoj.12077). Dataset II included Cox1, 18S, 28S, ITS-1, DUNUC3, DUNUC5 analysed concatenated or independently (no outgroup included), to obtain the phylogenetic trees and to carry out dating analyses. Dataset III is constituted by 18S gene sequences from Catenulida and Rhabditophora (including Tricladida) representatives, and a chaetognath species as outgroup to run a dating analysis. Dataset IV, to run biogeographical analyses, was obtained by pruning dataset II to include only one representative per biogeographical region but including at least one representative per clade (species of Recurva and Schmidtea were used as outgroup representatives), in this dataset each gene was analysed independently.

Four datasets were prepared for the different analyses:

Dataset I includes 18S, 28S, DUNUC3, Cox1 concatenated for Dugesia representatives and used Recurva and Schmidtea genera as outgroups, to root the Dugesia (Sluys et al., 2013).

Dataset I (4,444 bp):

18S - 1 - 1,737

28S - 1,738 - 3,316

Dunuc3 - 3,317 - 3,688

Cox1 - 3,689 - 4,444

Dataset II includes 18S, 28S, Cox1, DUNUC3, DUNUC5, ITS-1 concatenated (no outgroup included), to obtain the phylogenetic trees and to carry out dating analyses.

Dataset II (5,501 bp):

18S - 1 - 1,607

28S - 1,608 - 3,184

Cox1 - 3,185 - 3,934

Dunuc3 - 3,935 - 4,300

Dunuc5 - 4,301 - 4,622

ITS-1 - 4,623 - 5,501

Dataset III is constituted by 18S gene sequences from Catenulida and Rhabditophora (including Tricladida) representatives, and a chaetognath species as outgroup to run a dating analysis.

Dataset IV, to run biogeographical analyses, was obtained by pruning dataset II to include only one representative per biogeographical region but including at least one representative per clade (species of Recurva and Schmidtea were used as outgroup representatives).

Dunuc2 alignment includes the sequences obtained for these nuclear gene fragment. Since few sequences were obtained they were not analysed in the paper, nonetheless they have been uploaded to GenBank, and the data on exon and intron limits is given in a table on the paper.

See readme file for more details on the datasets