Many members of the trematode family Heterophyidae Odhner, 1914, use fishes as intermediate hosts and humans as definitive hosts [1, 2]. Six species in particular, Centrocestus formosanus, Haplorchis pumilio, H. taichui, H. yokogawai, Procerovum varium and Stellantchasmus falcatus [2, 3] are among the most clearly recognized human pathogens and mostly occur in eastern Asia including China, the Philippines, Korea, Taiwan, Thailand, Laos, Cambodia and Vietnam [3–11]. Heterophyidiasis caused by these and related species has now become a major public health threat, not only in Asia but in parts of Africa and the Americas [3, 5, 10, 12, 13]. Humans acquire heterophyid infection by consumption of undercooked or raw freshwater fishes containing infective metacercariae [3, 14]. Infection with multiple species is frequently reported in Vietnam and elsewhere [3, 5, 7, 9, 14].

DNA sequences are commonly used for molecular diagnosis and systematic/phylogenetic studies. Although markers are often chosen from the mitochondrial genome, sequences from the nuclear ribosomal transcription unit (rTU) (including 18S, ITS1, ITS2 and 28S) are particularly useful and reliable for this purpose [10, 15–22]. A single rTU consists of three coding regions (the 18S, 5.8S and 28S rRNA genes) separated by two internal transcribed spacers (ITS1 and ITS2) [17]. Short external transcribed spacers (ETS) are found 5' of the 18S gene and 3' of the 28S gene. Adjacent rTUs in the ribosomal array are separated by a long non-transcribed intergenic spacer (IGS) region [17, 23, 24]. Sequences from various portions of the rTU (18S, ITS1, ITS2 and 28S) have been widely used for inference of phylogenetic relationships and taxonomic clarification within and between many trematode families (e.g. [15, 18, 22, 25–31]). Sequences of complete or near-complete rTUs are only available for a few species of trematode [12, 16, 20, 32, 33]. Clearly, however, such data will be valuable for many kinds of comparative analysis, including systematics/phylogenetics and studies on intra- and interspecific or even intra- and interindividual variation in trematodes [15, 18, 20, 34, 35]. In particular, these data are needed for the large family Heterophyidae, which comprises more than 30 genera, many containing species infecting humans [1, 2, 12, 15, 34]. Heterophyid species in Vietnam have well been described epidemiologically and morphologically, but molecular data useful for diagnosis and identification, as well as taxonomy, are still limited [5–7, 9, 14].

The aim of this paper is to present the sequence of near-complete ribosomal transcription units from Haplorchis pumilio and H. taichui, commonly found in humans. Portions of the 28S rRNA gene from other heterophyids infecting humans in Vietnam are also presented, i.e. Centrocestus formosanus, Haplorchis yokogawai, Procerovum varium and Stellantchasmus falcatus. The data will be used to explore the phylogenetic positions of these genera in the family Heterophyidae and in the class Trematoda.

In this study, we have presented sequences of the near-complete ribosomal transcription units (rTUs) for two common species of the family Heterophyidae, Haplorchis pumilio and H. taichui, which infect humans in Vietnam. The obtained sequences encompass virtually the complete 18S gene (typical length range 1.7–2.9 kb) and almost half of the 28S gene (typical length range 3.3–5.5 kb) [16, 17]. Also obtained were the complete ITS1, 5.8S gene and ITS2 sequences for these species.

We have found repetitive sequences tandemly arranged in the ITS1 of H. pumilio and in the ITS2 of H. taichui. ITS sequences of both species have been reported from Israel [12]. Israeli H. pumilio possessed only two short tandem repeats (30 bp) in their ITS1, in strong contrast to the Vietnamese sequences, in which the ITS1 contained five complete repeats and one incomplete copy. The ITS1 sequences differed substantially in length between Vietnamese and Israeli individuals of the same species, 1,106 vs 640 bp in H. pumilio; and 797 vs 582 bp in H. taichui, due to differences in numbers of tandem repeats. These indicate intraspecific polymorphism as reported commonly in trematodes [8, 12, 33]. Likewise, ITS2 showed repetitive sequence differences between individuals from different locations. The presence of repeats in the internal transcribed spacers of trematodes has been reported for several taxa, including those in Schistosomatidae, Opisthorchiidae, Heterophyidae, Paramphistomatidae and others [8, 32, 33, 40]. The presence of repeats, variation in length and sequence variation, within and between species, all contribute to difficulties when trying to align ITS regions. This is particularly so when phylogenetically divergent species are being compared and suggest that this region is not suitable for deep-level phylogenies [17]. At the level of genus and species, however, alignments of ITS sequences have proved valuable for phylogenetic studies and molecular taxonomy [17, 41, 42].

The 18S and 28S rDNA sequences, however, are of considerable value for species identification and phylogenetic analysis [12, 15, 16, 18, 19, 25, 26, 30, 43, 44]. Alignment of these genes is generally straightforward, even among distantly related species, and long repeats do not occur.

The topology of the phylogenetic tree inferred from 40 trematode sequences in this study (Fig. 2) generally agreed well with previous findings. Most genera represented by multiple sequences formed well-supported monophyletic clusters. One striking exception was the sequence of Procerovum varium, which rendered Haplorchis paraphyletic. This relationship has also been noticed by others (e.g. [10]). Clearly, the definitions of these two genera will need to be revisited. The three families constituting the Opisthorchioidea, the Heterophyidae, Cryptogonimidae and Opisthorchiidae, are very poorly resolved in the tree. The Heterophyidae is not a monophyletic taxon. Indeed, two genera of nominal heterophyids, Euryhelmis and Cryptocotyle, appear to have closer affinities with the Opisthorchiidae than with the Heterophyidae. This relationship was also found by Thaenkham et al. [34] using 18S rDNA sequences, and by Thaenkham et al. [18] using concatenated 18S and 28S sequences. Paraphyly of the Heterophyidae with respect to the Opisthorchiidae was also demonstrated by [15] using 18S and 28S sequences. An additional heterophyid genus, Centrocestus, had an affinity with members of the Cryptogonimidae, or appeared as basal within the Opisthorchioidea (Fig. 2). Such a placement was not supported by analysis of concatenated 18S and 28S sequences by [18]. It is clear that the entire superfamily Opisthorchioidea presents broad systematic and taxonomic challenges to be met in the future using combined morphological and molecular approaches.

In conclusion, the present study determined and annotated the near-complete ribosomal transcription unit (rTU), consisting of 18S, ITS1, 5,8S, ITS2 and 28S rRNA genes and spacers, from H. pumilio and H. taichui from human hosts in Vietnam. The ITS1 in H. pumilio and ITS2 in H. taichui contained tandem repeats. The 28S rDNA sequences are conserved among individuals within a species but variable between species and genera. Based on 28S rDNA sequence analysis of 40 sequences representing 19 species in the superfamily Opisthorchioidea, six common human pathogenic heterophyids, Centrocestus formosanus, Haplorchis pumilio, H. taichui, H. yokogawai, Procerovum varium and Stellantchasmus falcatus were clearly resolved. In addition, the phylogenetic tree inferred from these sequences again confirmed anomalies in molecular placement of some members of the family Heterophyidae and demonstrates the need for reappraisal of the entire superfamily Opisthorchioidea. The new sequences provided here supplement those already available in public databases and add to the array of molecular tools that can be used for the diagnosis of heterophyid species in human and animal infections.