Uncoupled histories in the largest continental arc of North America: goodeid fishes and their gyrodactylid parasites influenced differently by the Trans-Mexican Volcanic Belt
Description
We characterized monopisthocotylans infecting endemic goodeid fishes south of the Trans-Mexican Volcanic Belt (TMVB), a major North American barrier separating the Nearctic and Neotropical biogeographic provinces. Considering TMVB emergence powerfully influenced goodeid diversification and distribution, we assessed whether it also affected Gyrodactylus lineages associated with these nearctic hosts. Previously, six Gyrodactylus spp. were known to infect goodeids (Goodea atripinnis, Giradinichthys multiradiatus) within and north of the TMVB; we recorded six gyrodactylids (two known, four undescribed species) infecting Ilyodon and Allotoca spp. south of the TMVB (S-TMVB). Using phylogenetic (internal transcribed spacer, cytochrome oxidase subunit II) and species delimitation (ASAP, MSC models) analyses, we tested three hypotheses: 1) ancient parasite lineages whose association with goodeids probably predates TMVB emergence occur both sides of this barrier; 2) following TMVB emergence, some nearctic gyrodactylids will only be found S-TMVB; 3) gyrodactylids infecting goodeids S-TMVB switched hosts from other Cyprinodontiformes. Analyses showed that: 1) Gyrodactylus tomahuac occurs both sides of the TMVB; 2) no nearctic parasites were found S-TMVB, confirming TMVB is a biogeographical barrier; 3) four undescribed gyrodactylids infecting goodeids S-TMVB are morphologically and phylogenetically closely-related to parasites infecting profundulid fishes, suggesting host switches from these neotropical hosts whose northern distribution limit is the TMVB. Our data reveal a complex and profound evolutionary history of Gyrodactylus species parasitizing goodeids, uncoupled from that of their hosts. Phylogenetic analyses and Multispecies Coalescent models suggest three main parasite groups are associated with goodeid fishes, and that gyrodactylids have historically switched hosts between Cyprinodontiformes (Goodeidae, Profundulidae, Poeciliidae).
Files
Steps to reproduce
Sequences of ITS1-2 rDNA and cox2 mtDNA from this study were run in a BLAST, and species of Gyrodactylus with the highest percentage of identity were included in the alignment. ITS alignments included 19 sequences from gyrodactylids infecting goodeid fishes characterized by Rubio-Godoy et al. (2016) and García-Vásquez et al. (2018). Additionally, some species of Gyrodactylus that parasitize profundulids and poecilids in Mexico were included. Gyrodactylus pakan Razo-Mendivil, García-Vásquez & Rubio-Godoy, 2016 was used as the outgroup. Alignments and phylogenetic analyses (ITS1-2 rDNA, cox2 mtDNA, and ITS1-2+cox2) were implemented in PhyloSuite v2 (Zhao et al., 2025) with the plugin programs using the methodology described by Jakovlić et al. (2024). Sequences of ITS1-2 and cox2 were aligned using the (--auto strategy) command of MAFFT v7.526 (Katoh and Standley, 2013), implemented in PhyloSuite v2; all the alignments are accessible at Mendeley Data, doi: XXX. ModelFinder v3.0.1 (Kalyaanamoorthy et al., 2017) was used to select the best-fit model using BIC criterion (ITS=TVM+F+I+G4; cox2=TIM3+F+I+G4). The datasets of ITS1-2+cox2 were concatenated in PhyloSuite v2 and the TVM+I+G+F and TIM3+I+G+F models were implemented. Maximum-likelihood (ML) phylogenies of ITS1-2, cox2, and ITS1-2+cox2 were inferred using IQ-TREE v3.1.0 (Minh et al., 2020; Nguyen et al., 2015; Wong et al., 2025) under the TVM+F+I+G4 and TIM3+F+I+G4 models for 5000 ultrafast bootstrap replicates (Hoang et al., 2018; Minh et al., 2013), as well as the Shimodaira-Hasegawa-like approximate likelihood-ratio test (Guindon et al., 2010). Bayesian Inference (BI) phylogenies for individual and concatenated genes were inferred using MrBayes v3.2.7a (Ronquist et al., 2012) under the TVM+I+G+F and TIM3+I+G+F models (four chains, two parallel runs, 10 million generations, sampled every 1000 generations), in which the initial 25% of sampled data were discarded as burn-in and the outputs were examined with Tracer v1.4 (Rambaut et al., 2018). We obtained the posterior probabilities of clades from the 50% majority rule consensus of sampled trees after excluding the initial 25% as burn-in for MrBayes. Phylogeny was visualized using FigTree v1.4.4 (Rambaut et al., 2018).