A molecular phylogenetic appraisal of the acanthostomines Acanthostomum and Timoniella and their position within Cryptogonimidae (Trematoda: Opisthorchioidea)

Collection of hosts and trematode parasites

As part of our ongoing study in the Celestun lagoon (Sosa-Medina, Vidal-Martínez & Aguirre-Macedo, 2015), we collected specimens of cryptogonimid metacercariae presumed to be of the subfamily Acanthostominae: Acanthostomum americanum (=Atrophecaecum astorquii) Pérez-Vigueras, 1956, and Timoniella (=Pelaezia) loossi Pérez-Vigueras, 1956, from the Ria Celestun Biosphere Reserve, Yucatan Peninsula, Mexico (based on Moravec, 2001; Vidal-Martínez et al., 2001; Brooks, 2004; Miller & Cribb, 2008a). These metacercariae were collected from the euryhaline fish Cichlasoma urophthalmus (Günter, 1862) (Perciformes: Cichlidae) from the Yaxaá water spring (20°53′12.57″N; 90°20′58.86″W), located in the Celestun tropical lagoon (Fig. 1). We also collected cercariae presumed to be of the Cryptogonimidae from the aquatic gastropod Pyrgophorus coronatus (Pfeiffer, 1840) (Hybrobiidae) (see Scholz et al., 2000), at the same location, to test for possible life-cycle links between the cercariae and metacercariae with molecular data. In March 2016 we collected 223 snails of P. coronatus from two localities: Baldiocera Spring (20°54′6.29″N; 90°20′26.46″W) (156 snails) and Yaxaá Spring (67 snails) (the two springs are approximately 1,400 m apart). Snails were collected using strainers, placed separately into glass tubes and maintained in artificial light in the laboratory to stimulate the emergence of cercariae. After 2–3 days, portions of the snails were removed from their shells by dissection under a stereomicroscope. The only representatives of Cyptogonimidae (three cercariae) were collected from a single P. coronatus from Yaxaá Spring. For representatives of other families, of the 156 P. coronatus examined from Baldiocera Spring, we observed two cercaria of Ascocotyle (Phagicola) nana Ransom, 1920 (Heterophyidae) in each of two individual snails; and one metacercaria of Crassicutis cichlasomae Manter, 1936 (Apocreadiidae) from one snail. Both larvae have been previously recorded from P. coronatus (Scholz et al., 2000). Of the 67 P. coronatus examined from Yaxaá Spring, the only cercariae observed belonged to the aforementioned cryptogonimids. We also sampled specimens of other adult cryptogonimids, e.g., Oligogonotylus mayae (Razo-Mendivil, Rosas-Valdez & Pérez-Ponce de León, 2008), from the cichlid fish C. urophthalmus. The protocols for host dissection, examination, collection and preservation, and the morphological study of parasitic specimens followed Vidal-Martínez et al. (2001). We also collected adult specimens of the apocreadiid species Crassicutis cichlasomae, from the same fish host. Crassicutis cichlasomae was used as an outgroup taxon for the phylogenetic analyses in this study, based on its previously established sister group relationship of Ophisthorchioidae (Bray et al., 2009; Fraija-Fernández et al., 2015). Trematodes were identified based on morphological criteria suggested by Vidal-Martínez et al. (2001), Miller & Cribb (2008a), Razo-Mendivil, Rosas-Valdez & Pérez-Ponce de León (2008) and Razo-Mendivil et al. (2010). Reliable identification to genus level is possible for both Timoniella and Acanthosthomum based on metacercariae morphology. Microphotographs of both taxa can be found in Fig. S1. However, identification to species level may be questionable, therefore we hereafter refer to the species as T. cf. loossi and A. cf. americanum. Several metacercariae and adult specimens collected for morphological analysis were deposited as voucher specimens [T. cf. loossi (No. 525), A. cf. americanum (No. 526), C. cichlasomae (No. 527) and O. mayae (No. 528)] in the Colección Helmintológica del CINVESTAV (CHCM), Departamento de Recursos del Mar, Centro de Investigación y de Estudios Avanzados del Instituto Politécnico Nacional, Unidad Mérida, Yucatán, México. Acanthostomine cercariae were not deposited because each specimen was required for the molecular study. Comisión Nacional de Acuacultura y Pesca (PPF/DGOPA-070/16) issued the collecting permits.

DNA extraction, PCR amplification and sequencing

DNA was extracted from individual cercariae, metacercariae and adult trematodes. DNA extraction was performed using the DNAeasy blood and tissue extraction kit (Qiagen, Valencia, CA, USA) following the manufacturer’s instructions. For the four trematode taxa, the partial 28S ribosomal gene region was amplified by Polymerase Chain Reaction (PCR) (Saiki et al., 1988), using 28sl forward (5′-AAC AGT GCG TGA AAC CGC TC-3′) (Palumbi, 1996) and LO reverse (5′-GCT ATC CTG AG(AG) GAA ACT TCG- 3′) (Tkach, Pawlowski & Mariaux, 2000). The primers BD1 forward (5′-GTC GTA ACA AGG TTT CCG TA-3′) and BD2 reverse (5′-TAT GCT TAA ATT CAG CGG GT-3′) (Bowles, Blair & McManus, 1995) were used for ITS1–5.8S–ITS2 fragment. The reactions were prepared using the Green GoTaq Master Mix (Promega, Madison, WI, USA). This procedure was carried out using an Axygen Maxygen thermocycler. PCR cycling conditions by both molecular markers were as follows: an initial denaturing step of 5 min at 94 °C, followed by 35 cycles of 92 °C for 30 s, 55 °C for 45 s, and 72 °C for 90 s, and a final extension step at 72 °C for 10 min. The PCR products were analysed by electrophoresis in 1% agarose gel using TAE 1× buffer and observed under UV light using the QIAxcel® Advanced System. The purification and sequencing of the PCR products were carried out by Genewiz (South Plainfield, NJ, USA; https://www.genewiz.com/).

Molecular data and phylogenetic reconstruction

To obtain the consensus sequences of the larvae and adults of A. cf. americanum, T. cf. loossi, O. mayae and C. cichlasomae, we assembled and edited the chromatograms of forward and reverse sequences using the Geneious Pro v5.1.7 platform (Drummond et al., 2010). To investigate the monophyly of the taxa included in Cryptogonimidae at the subfamily level, the 28S, ITS1, 5.8S and ITS2 sequences that were generated during this study were aligned with sequences of other cryptogonimids, and their sister groups, heterophyid and opisthorchiid taxa (based on Thaenkham et al., 2011; Thaenkham et al., 2012), obtained from GenBank (see GenBank accession numbers in Table S1), using an interface available with MAFFT v.7.263 (Katoh & Standley, 2016), an “auto” strategy and a gap-opening penalty of 1.53 with Geneious Pro, and a final edition by eye in the same platform. The best partitioning scheme and substitution model for each molecular marker was selected by using the “greedy” search strategy in Partition Finder v.1.1.1 (Lanfear et al., 2012; Lanfear et al., 2014) and applying the Bayesian Information Criterion (BIC) (Schwarz, 1978). The nucleotide substitution model that best fit the 28S data was TVM + I + G (Posada, 2003); for ITS1 and ITS2 it was TVMef + G (Posada, 2003); and for 5.8S, it was JC + G (Jukes & Cantor, 1969). Hypervariable regions of 28S, ITS1 and ITS2 alignments were excluded using the Gblocks Web Server (Castresana, 2000; Talavera & Castresana, 2007).

The datasets were analysed by Bayesian inference (BI) and Maximum likelihood analyses (ML) using the CIPRES Science Gateway v. 3.3 (Miller, Pfeiffer & Schwartz, 2010). ML analyses were conducted in RaxML v. 8 (Stamatakis, 2014) using the GTRCAT approximation as a model of nucleotide substitution (Yang, 1994; Yang, 1996; Stamatakis, 2006). BI analyses were carried out with MrBayes v. 3.2.1 (Ronquist et al., 2012). The Bayesian phylogenetic trees were reconstructed for each gene separately using two parallel analyses of Metropolis-Coupled Markov Chain Monte Carlo (MCMC) for 20 × 10⁶ generations each. Topologies were sampled every 1,000 generations and the average standard deviation of split frequencies was observed until it reached <0.01, as suggested by Ronquist et al. (2012). A majority consensus tree with branch lengths was reconstructed for the two runs after discarding the first 5,000 sampled trees. For both ML and BI analyses, model parameters were independently optimized for each partition. Node support was evaluated by non-parametric bootstrapping (Felsenstein, 1985) with 1,000 replicates performed with RAxML (ML) and BI by Posterior probabilities (PP), where bootstrap values ≥75% and PP ≥ 0.95, were considered strongly supported.

DNA sequences and dataset analyses

In total, 36 bi-directional partial 28S (domains 1 and 2) and ITS1-5.8S-ITS2 sequences were obtained from three individual cercariae and three individual metacercariae from A. cf. americanum, as well as three individual metacercariae from T. cf. loossi, O. mayae (one adult specimen), and C. cichlasomae (one adult specimen, outgroup) (Table 1). The partial 28S rDNA sequence fragment consisted of 881 base-pairs (bp) for the cercariae and metacercariae of A. cf. americanum; 880 bp in T. cf. loossi, 871 bp in O. mayae, and 870 bp in C. cichlasomae. The 28S sequences of cercariae and metacercariae of A. cf. americanum from P. coronatus were identical, while the sequences of T. cf. loossi showed a divergence of 0.03%. Nucleotide sequence variation in the 28S alignment from cryptogonimids (excluding the outgroup taxon) from 28S included 722 conserved sites, 537 variable sites, 403 parsimony-informative sites, and 134 singleton sites. The sequence fragments for the ITS1 nuclear marker were between 709 and 781 bp in length for A. cf. americanum; and were 805 bp in T. cf. loossi, 613 bp in O. mayae, and 424 bp in C. cichlasomae. The 5.8S nuclear marker was composed of 160 bp in A. cf. americanum, T. cf. loossi, O. mayae and C. cichlasomae. The length of the ITS2 nuclear marker ranged from 259 bp to 277 bp in A. cf. americanum and from 268 bp to 277 bp in T. cf. loossi; 260 bp in O. mayae, and 295 bp in C. cichlasomae. The ITS1 and ITS2 sequences of A. cf. americanum displayed 4% and 0.7% divergence, respectively, and those from T. cf. loossi displayed 0.9% divergence and 100% pairwise identity; the 5.8S sequences were identical. Nucleotide sequence variation (excluding the outgroup taxa) for ITS1, 5.8S and ITS2 were 62∕69∕50 conserved, 406∕92∕212 variable, 341∕36∕184 parsimony-informative, and 65∕56∕28 singleton sites, respectively.

Phylogenetic reconstructions

We inferred the phylogenetic relationships of Cryptogonimidae, based on the BI and ML analyses, from the following two datasets. The partial 28S gene dataset contained 92 terminals belonging to 81 species, and the combined dataset (28S + ITS1 + 5.8S + ITS2) contained 294 sequences belonging to 81 taxa concatenated (all sequences available from GenBank, see Table S1). The phylogenetic trees constructed from the 28S and the concatenated datasets (28S + ITS1 + 5.8S + ITS2), based on BI and ML analyses, were broadly congruent. For example, all clades with high nodal support values (PP ≥ 0.95 and bootstrap ≥ 75%) and analysed with the concatenated and 28S datasets were recovered with both BI and ML (Fig. 2; Figs. S2–S4). Only three clades were recovered with high nodal support values (PP ≥ 0.95) using BI but not ML (i.e., (Gynichthys diadikidnus, Neoparacryptogonimus ovatus); (Metagonimus takahashii, M. yokogawai); and (Haplorchis yokogawai (Haplorchis popelkae, Haplorchis pumilio))), while only one clade received high nodal support value (bootstrap ≥ 75%) with ML and not BI (i.e., (Haplorchoides sp. (Stictodora sp. isolate St1, Stictodora sp. isolate St2))) (Fig. 2). Conversely, only one difference was observed between the topology of the phylogenetic trees obtained from the 28S and concatenated datasets with BI and ML. Namely, the phylogenetic tree obtained from the ML analysis of the 28S sequence dataset contained a polyphyletic group (without nodal support value), i.e., Siphodera vinaledwardsii, Gynichthys diakidnus, Chelediadema marjoriae, Caecincola parvulus, and Tabascotrema verai (Fig. S3). In all trees, acanthostomines were paraphyletic, with high nodal support values (PP ≥ 0.95 and bootstrap ≥ 75%). Based on all trees, the family Cryptogonimidae appears to have arisen from a paraphyletic Heterophyidae/Opisthorchiidae group. As well, all trees clearly showed that the generated sequences in this study of T. cf. loossi and A. cf. americanum form a monophyletic group with high nodal support values (PP ≥ 0.95 and bootstrap ≥ 75%), respectively. These acanthostomine genera are sister to the remaining cryptogonomids. Furthermore, the genus Acanthostomum is a monophyletic group with high nodal support values (PP ≥ 0.95 and bootstrap ≥ 75%). Lastly, the 28S and ITS1-5.8S-ITS2 fragment sequences of acanthostomine metacercaria from C. urophthalmus were identical to those of cercarie from P. coronatus, and therefore both trematode stages correspond to the same taxa, A. cf. americanum.

The phylogenetic relationships among Cryptogonimidae at the generic level had high support (PP ≥ 0.95) and the genera Siphoderina, Belusca, Varialvus, Caulanus and Latuterus form a monophyletic group (Clade I) (Fig. 2), distributed in the Indo-Pacific region (I-P). Of these, the genera Belusca and Varialvus, Caulanus, and Latuterus have been found parasitizing the marine fish families Haemulidae and Lutjanidae. Furthermore, Retrovarium spp. was found parasitizing Lutjanidae and Haemulidae from the Indo-West Pacific (IW-P) (Fig. 2).