An Anomalous Phylogenetic Position for Deraiotrema platacis Machida, 1982 (Lepocreadiidae) from Platax pinnatus on the Great Barrier Reef

: The monotypic genus Deraiotrema Machida, 1982 has only been reported once, from the orbicular batﬁsh Platax orbicularis (Forsskål) in the waters around Palau in Micronesia (Machida, 1982). It has a body-shape similar to other lepocreadiids from batﬁshes, such as species of Bianium Stunkard, 1930 and Diploproctodaeum La Rue, 1926, but di ﬀ ers in having multiple testes in ventral and dorsal layers. Here we report Deraiotrema platacis Machida, 1982 for just the second time, infecting the dusky batﬁsh Platax pinnatus (Linnaeus) from the waters o ﬀ Lizard Island on the northern Great Barrier Reef. We present a molecular phylogenetic analysis of the position of this genus inferred from 28S rDNA sequences. Surprisingly, we ﬁnd the species most closely related to Echeneidocoelium indicum despite the infection of completely unrelated hosts and the presence of two characters (lateral fold in the forebody and multiple testes) that are found elsewhere in the Lepocreadiidae. We conclude that homoplasy within the Lepocreadiidae is extensive and that morphology-based prediction of relationships has little prospect of success.


Introduction
Recent molecular analyses have revolutionised the classification of the Lepocreadioidea. The process started with a molecular exploration of the superfamily [1] which identified five major clades and that the generally accepted concept of the Lepocreadiidae was polyphyletic. These issues were partly resolved by Bray & Cribb [2], by the recognition of Aephnidiogenidae Yamaguti, 1934 andLepidapedidae Yamaguti, 1958 at the family level. At that stage, nine lepocreadioid genera were considered unplaceable (genera incertae sedis). Most recently, in Bray et al. [3], one of these genera, Gibsonivermis Bray, Cribb & Barker, 1997, was shown to justify recognition of a new monotypic family with the Lepocreadioidea, the Gibsonivermidae. Although we suspect that the overall structure of the Lepocreadioidea is maturing, there remains considerable uncertainty about the relative position of many of the genera. From this it follows that we have limited understanding of character evolution in any of the families. Here we redescribe an enigmatic genus and species, Deraiotrema platacis Machida, 1982, and produce an expanded phylogenetic analysis using 28S DNA, allowing consideration of the distribution of its distinctive defining characters.

Specimen Collection and Morphological Analysis
Fish were obtained by spear-fishing from the waters around Lizard Island, Queensland, Australia (14 • 40 S, 145 • 28 E). Trematodes were collected as described by Cribb & Bray [4]. Specimens were Diversity 2019, 11, 104 2 of 11 pipetted into almost boiling saline for fixation and preserved in formalin or 70% ethanol. Whole-mounts were stained with Mayer's paracarmine or Mayer's haematoxylin, dehydrated in a graded ethanol series, cleared in beechwood creosote or methyl salicylate and mounted in Canada balsam. Measurements were made through a drawing tube on an Olympus BH-2 microscope, using a Digicad Plus digitising tablet and Carl Zeiss KS100 software adapted by Imaging Associates, and are quoted in micrometers, with the range and the mean in parentheses. The following abbreviations are used: NHMUK, the Natural History Museum, London, UK; QM, Queensland Museum Collection, Brisbane, Australia.

Molecular Results
Alignment of the new sequence data with that of related taxa (Table 1) yielded 1302 characters (including indels). Deleted ambiguously aligned regions amounted to 12 bases (<1% of the alignment), resulting in a final dataset of 1290 characters for phylogenetic analysis. Bayesian inference and maximum likelihood analyses resulted in phylograms with identical topologies (Figure 2), in which the species of the Lepocreadiidae formed a well-supported clade. Within the Lepocreadiidae, D. platacis formed a well-supported clade with Echeneidocoelium indicum Simha & Pershad, 1964, which together formed a well-supported clade with three species of Hypocreadium Ozaki, 1936. Notably, D. platacis was phylogenetically distinct from members of the Pelopscreadium/Diploproctodaeum/Lobatocreadium/Diplocreadium/Bianium clade, which are most similar to D. platacis morphologically.

Discussion
These worms appear to be indistinguishable from the only previously known specimens described by Machida [25] from the orbicular batfish Platax orbicularis from the waters off Palau, Micronesia. This is thus the second record of this species. Any slight differences between the original

Discussion
These worms appear to be indistinguishable from the only previously known specimens described by Machida [25] from the orbicular batfish Platax orbicularis from the waters off Palau, Micronesia. This is thus the second record of this species. Any slight differences between the original description and our specimens probably relate to differences in fixation, with the specimens of Machida flattened in contrast to our new specimens.
Deraiotrema is distinguished by the combination of multiple testes, body flaps in the anterior part of the body, and caeca which abuts the body wall, giving the impression that ani may be present. Each of these characters is seen in other lepocreadiids, but the combination is unique.
Species of several lepocreadiid genera have anterior flaps or scoop-like developments of the anterior part of the worm, namely Bianium, Diploproctodaeoides Reimer, 1981, Diploproctodaeum and Diplocreadium Park, 1939 [26]. Strikingly, several species of these genera also infect ephippids, leading us, mistakenly, to expect that D. platacis would be related to them. Notably, none of these have multiple testes although the first three genera have caeca abutting the body wall, possibly forming ani. Other genera have a similar arrangement of the caeca, but no body-flaps, i.e., Pelopscreadium Dronen, Blend, Khalifa, Mohamadain & Karar, 2016 and Mobahincia Bray, Cribb & Cutmore, 2018 [3,27], neither of which have multiple testes. Diploproctia Mamaev, 1970 was illustrated with a scoop by Bray [26], but Machida [28] redescribed the type-and only species D. drepanei Mamaev, 1970 showing that Bray's interpretation of the illustration in Mamaev [29] was in error and the species lacks a scoop. Mamaev [29] stated that separate ani are present in D. drepanei but Machida [28] found no evidence for this and noted that the caeca terminate 'blindly near posterior end of body'. Several lepocreadiid genera have multiple testes, namely Neomultitestis Machida, 1982, Transversocreadium Hafeezullah, 1970, Multitestis Manter, 1931, Rhagorchis Manter, 1931and Multitestoides Yamaguti, 1971. Perhaps the most similar genus to Deraiotrema is Rhagorchis, for which just two species are known, R. odhneri Manter, 1931 (syn. R. varians (Linton, 1940)) and R. manteri Ramadan, 1982 [30-32]. Yamaguti [33] re-drew the holotype of R. odhneri and showed that the ventral sucker is 'often retracted in anterior half of body'. The illustration by Yamaguti shows this is not the same feature as the lateral flaps we found on our specimens, although Yamaguti's illustration of this feature does resemble that in the drawing of D. platacis in Machida [25], where the lateral flaps are said to join posterior to the ventral sucker. We could not detect if the flaps join in our specimens and postulate, again, that the difference is due to flattening. The drawings of Manter [30] and Yamaguti [33] show the caeca in R. odhneri reaching fairly close to the posterior body wall but they do not appear to abut it and it appears that there are 11 testes in a single layer. Rhagorchis odhneri is known only from the monacanthids the orange filefish Aluterus schoepfii (Walbaum), the scrawled filefish A. scriptus (Osbeck) and the fringed filefish Monacanthus ciliatus (Mitchill) in the north-western Atlantic and the Caribbean Sea/Gulf of Mexico area [30,31,[34][35][36][37][38]. The similarities between the species of Deraiotrema and Rhagorchis may be superficial, such that we would be tempted to think that R. odhneri is actually a schistorchiine apocreadiid were it not for the cirrus-sac in the re-drawing of the holotype by Yamaguti [33]; several species of that family with multiple testes are known from monacanthids. Rhagorchis manteri, from the scarid the longnose parrotfish Hipposcarus harid (Forsskål) in the Red Sea [32], is also likely to be a schistorchiine apocreadiid, possibly Plesioschistorchis haridis (Nagaty, 1957), which is also from scarids in the Red Sea [39,40]. Saoud et al. [41] reported Rhagorchis sp. in the bluechin parrotfish Scarus ghobban Forsskål, also from the Red Sea.
In phylogenetic analyses, Deraiotrema platacis formed a well-supported clade sister to Echeneidocoelium indicum, a parasite of remora Remora remora (Linnaeus) and, more usually, the sharksucker Echeneis naucrates Linnaeus. Morphologically and in terms of host-specificity, there is little to connect D. platacis and E. indicum. Echeneidocoelium indicum is a narrow, elongate worm, with two testes and caeca which impinge on the excretory vesicle, but apparently do not form a uroproct or ani [42,43]. One feature shared is the follicular ovary, a feature which is unusual in elongate lepocreadiids, but not in more squat forms. The intimate relationship with E. indicum is all the more surprising for the short branch lengths that separate them. The distance between them is less than that between any of the combinations of the three Hypocreadium species in the analysis. It is apparent, however, that features such as the anterior scoop and the appearance (or presence) of ani, and certainly the acquisition of multiple testes, are likely to be convergent features. As can be seen from the new phylogram ( Figure 2) the anterior scoop appears to be synapomorphic for the Pelopscreadium/Diploproctodaeum/Lobatocreadium/Diplocreadium/Bianium clade (PDLDB clade) but lost in Pelopscreadium and Lobatocreadium; we take this as further evidence that seemingly significant characters are highly labile. Diploproctodaeum as presently recognised by its complete anterior scoop (compared to lateral flaps only in Bianium) is not monophyletic and clearly the relationships of these genera need further study and wider sampling. It seems likely that the 'sponge-like pelops (shoulder pads)' in Pelopscreadium are homologues of the scoop or flaps [17,44]. Lobatocreadium Madhavi, 1972, represented in the tree by L. exiguum (Manter, 1963), has no obvious flap or scoop vestige [45][46][47]. Given the topology of the new phylogram, it seems unlikely that the lateral flaps in Deraiotrema are homologous with those of the PDLDB clade.
The appearance of ani in most members of the PDLDB clade is highly characteristic, with the caeca abutting the body-wall at the point of a distinct concavity in the wall. This feature is not, however, present in all members of the clade, with the caeca in species of Lobatocreadium and Diplocreadium clearly ending blindly well-separated from the body wall. The posterior terminations of the caeca in Mobahincia and Deriaotrema are very similar to those in the PDLDB clade but the condition must be homoplasious.
Multiple testes in two layers is a pattern which has been not described from any other lepocreadiid as far as we are aware. Apart from the multi-testiculate lepocreadiids mentioned above, multiple testes crop up scattered across the digenean tree. Many members of the superfamily Schistosomatoidea Stiles & Hassall, 1898 have multiple testes [48], as do members of the families Typhlocoelidae Harrah, 1922 [49], Syncoeliidae Looss, 1899 [50] and Orchipedidae Skrjabin, 1913 [51]. Sporadic [57]. Despite this difference in testes arrangement S. gurukun (nine testes) and S. raritas (two testes) were shown to form a strongly-supported clade in the cryptogonimid phylogeny [57].
In summary, Deraiotrema has a number of unusual lepocreadiid features; none are unique, although the combination of these features is. The evidence we have, therefore, is that the anterior scoop, caeca abutting the body wall, the follicular ovary and the multiple testes are all features readily homoplasiously derived from the more usual arrangements. We conclude that it should not be assumed that the many remaining unsequenced lepocreadiids genera will not surprise us in their phylogenetic positions. We note that eight of the nine lepocreadioid genera designated as genera incertae sedis by Bray & Cribb [2] remain unsequenced.