Redescription of Paradiplozoon hemiculteri (Monogenea, Diplozoidae) from the type host Hemiculter leucisculus, with neotype designation

Paradiplozoon hemiculteri (Ling, 1973), a member of the Diplozoidae, parasitizes the gills of Asian fish. Not only is the type material unavailable for this species, the original description was poor and somewhat conflicting, and adequate molecular data were not available. What is more, the available morphological and molecular data are inconsistent and fluctuate significantly. Here, we present a redescription of P. hemiculteri based on morphological and molecular data from new isolates collected from the type host, the sharpbelly Hemiculter leucisculus (Basilewsky, 1855), captured at the neotype locality (Shaoguan, Guangdong Province, southern China); a neotype for P. hemiculteri was designated from this collection. The length and width of the body, buccal suckers, pharynx, attachment clamps, sickle and the central hook handle were all measured and the shape of the anterior and posterior part of the median plate and anterior and posterior joining sclerites accurately documented. Phylogenetic analyses based on the sequences of the second rDNA internal transcribed spacer (ITS2) indicated that all new samples clustered together and differed clearly from sequences attributed to P. hemiculteri, which are deposited in GenBank. Our results confirm that P. hemiculteri is the only diplozoid that has demonstrably been found on the gills of H. leucisculus to date.


Introduction
Members of the class Monogenoidea (parasites of salt and freshwater fish) can be classified into two subclasses: Polyonchoinea and Heteronchoinea [4]. Within the Heteronchoinea (infrasubclass Oligonchoinea), the Diplozoidae Tripathi, 1959 represents a specific monogenean group parasitising the gills of freshwater fishes. Diplozoids are obligatory blood-feeding ectoparasites with an unusual life cycle wherein two immature individuals (diporpa) meet on the gills of a fish and subsequently permanently fuse together into an X-shaped adult 'individual' [18,35]. Each adult worm (i.e. each 'side' of a fused individual) has a forebody and an ovarium and testis in each hindbody. The adult has two attachment apparatuses, each having four pairs of clamps and a pair of small central hooks situated on the ventral side of the respective opisthaptors.
According to Khotenovsky (1985) [18], the Diplozoidae is represented by two subfamilies: Diplozoinae Palombi, 1949 (five genera) and Neodiplozoinae Khotenovsky, 1980 (two genera). Morphological features (e.g. size of body, size of clamps) vary widely within diplozoid species, depending on size of the host fish and the developmental stage of the diplozoon [23,24], making determination to species level difficult. In general, the central hooks and clamp sclerites are considered the structures of most taxonomic relevance [5,11,18,22]. In more recent years, genetic analysis of molecular markers has been used to counter morphological similarities when differentiating species [21][22][23]34]. It is widely assumed that the majority of diplozoids are highly specific to their hosts; indeed, most follow their specific fish host throughout their distribution range and they are very likely to have co-evolved together [17,38]. As such, host-species determination has often been used for general identification of the parasite [16]. One exception to this general 'rule' is the generalist parasite Paradiplozoon homoion (Bychowsky & Nagibina, 1959), which has been recorded from more than fifteen cyprinid fish species [10,18,21].
In China, 33 species have been reported from five diplozoine genera, 23 from Paradiplozoon, six from Sindiplozoon, two from Inustiatus, one from Diplozoon, and one from Eudiplozoon [3,9,36,39]. Several authors [9,39] have published information on Chinese diplozoid species, including the original description of Paradiplozoon hemiculteri by Ling (1973); however, the morphological descriptions in these studies have largely been imperfect and/or inaccurate e.g. Chen (1973) [6] or the studies were based on molecular data only [9]. Despite all the published work concerning Chinese diplozoons, the molecular data are not accurate and none of the published studies adequately combined and reviewed morphological and molecular data [22,7,1,[31][32][33]. Moreover, according to the curator of the collection, Mr. Weijian Yao, and specialists from the Institute of Hydrobiology, Chinese Academy of Sciences, where the type material of P. hemiculteri should be located, the types of this species are missing and likely lost. Therefore, we collected new material from Shaoguan, Guangdong Province neotype locality and type host, from which a neotype for P. hemiculteri was selected and designated. In this work, we redescribe P. hemiculteri based on the new morphological and molecular data from the neotype and new specimens from the type host in southern China.

Sampling
As the type specimens of P. hemiculteri are now unavailable, the present study is based on morphological and molecular examination of ten adult diplozoid worms collected from the gills of Hemiculter leucisculus (Basilewsky, 1855), the type host fish species, from the neotype locality in Shaoguan, Guangdong Province, southern China (24.810329°N 113.597547°E) in 2015. Two individuals already fused into the X-shaped adult worm were separated, one haptor being fixed in 96% ethanol for molecular investigation. The second haptor was placed on a slide in a drop of water, covered with the cover slide, while excess water was drawn off with filtration paper until rupture of the parasite's body appeared, and the sample was infiltrated/fixed with ammonium picrateglycerine 1:1 (GAP) [8,20] for morphometric analysis.

Morphological analysis
After the fixation of the haptor in GAP, the sclerites of diplozoid clamps laid nearly in one plane, enabling accurate measurement.
An Olympus BX51 light microscope (Olympus, Japan) equipped with differential interference contrast and Stream Motion digital image analysis software v. 1.5 was used to measure the length and width of the body, buccal suckers, pharynx, attachment clamps and the central hook sickle and handle [18]. All measurements were performed 8-16 times. An Olympus U-DA drawing attachment was used to create outline drawings of the central hook clamps and sickle. The haptoral terminology used herein follows Pečínková et al. (2007) [26].

Molecular analysis
The DNA from ten samples was isolated according to the protocol published by Zavodna et al., (2008) [40]. Sequencing of the second rDNA internal transcribed spacer (ITS2) was used for molecular identification, using PCR primers adopted from Bachellerie & Qu (1993) [2]. The PCR reaction (50 ml) for amplification of the ITS2 region consisted of 5 ml 2x High Fidelity PCR Buffer (Thermo Scientific), 5 ml 25 mM MgCl 2 , 2ml 10mM dNTP, 1 ml (5 U/ml) of Taq DNA polymerase (Thermo Scientific), 32 ml PCR H 2 O, 2 ml 10 mM specific forward primer (DITS2F 5 0 -GGCTYRYGGNGTCGATGAAGA-ACGCAG-3 0 ) and reverse primer (DITS2R 5 0 -GCCGGA-TCCGAATCCTGGTTAGTTTC TTTTCCT-3 0 ), and 1 ml (50 ng) of DNA template. Amplification took place in a MyCycler thermal cycler system (Bio-Rad, USA) using the following protocol: 94°C for 2 min; 30 times 94°C for 30 s, 58°C for 30 s, 72°C for 60 s and a final 10 min extension at 72°C. The PCR products obtained were separated on agarose gel (1%), stained with GoldView (Dongsheng Biotech), purified using a High Pure PCR Product Purification Kit (Roche) and cloned onto Escherichia coli TOP10 using the CloneJET PCR Cloning Kit (Thermo Scientific), according to the manufacturer's protocol. Colonies were subsequently checked for fragment presence using PCR, with positive plasmids isolated using the High Pure Plasmid Isolation Kit (Roche) and sequenced in both directions using DITS2F and DITS2R PCR primers. The individual sequences obtained were compared with the NCBI database using the Basic Local Alignment Search Tool (BLAST).

Phylogenetic analysis
All sequences were aligned and manually edited using Geneious software v 6.1.8 [17]. The final dataset consisted of our data and all available Paradiplozoon spp. sequences from GenBank (Table 1), the whole dataset being aligned using the MAFFT software package [15]. The most suitable nucleotide substitution model was chosen according to Akaike Information Criteria using jModeltest software v 2.1.4 [27]. Phylogenetic analysis was performed on the model with the best likelihood score (GTR + G) on MrBayes v 3.2.2 Bayesian inference software [13,30], using 20 million Markov chain Monte Carlo generations with four chains and four independent runs. Convergence of runs was checked using AWTY software (Are We There Yet [25]) in order to estimate burn-in. Maximumlikelihood phylogeny was performed under the same model using PHYML software v 3.0 [12]. The reliability of branching patterns within trees was tested by the bootstrap method with 1,000 re-samplings. The final trees were visualised in FigTree software v 1.4.2 [28]. The sequence of closely related Inustiatus inustiatus (DQ098893) was used as the outgroup for both phylogenetic analyses (see Table 1).
The opisthaptors in the adult worm each comprise four pairs of clamps organised into two rows with two central hooks between them (Fig. 1b, 2a). The first (anterior-most) and smallest clamp is 56 mm

Molecular analysis
DNA sequences isolated and amplified from the ITS2 fragment (762 bp) of ten adult worms were all similar, with less than 1% variability. The ten sequences have now been deposited in GenBank under accession numbers KY124645 À KY124654.

Phylogenetic analysis
As the results of Bayesian inference and maximum likelihood analysis were almost identical, we pooled the results into one figure. The final tree is based on Bayesian  inference phylogenetic tree topology with branch supports stated for both analysis types (bootstrap and posterior probability; Fig. 3).
All our samples clustered together and created a wellsupported and separated group from those of P. hemiculteri sequences deposited in GenBank. Genetic distances between our sequences ranged between 0.1 and 0.3% (Fig. 3)

Discussion
In reviewing the available literature and corresponding sequence data deposited in public databases, we observed that taxonomic classification of the Asian diplozoon P. hemiculteri, a parasite of the gills of H. leucisculus, remains somewhat controversial at this time. Morphological descriptions of diplozoons isolated from H. leucisculus tend to be either inaccurate or inconsistent, with published illustrations of the clamps and central hook often ignoring key characteristic markers [9,18,39].
Furthermore, some of the sequences available in GenBank (i.e. DQ098884, DQ098887, DQ098888, DQ098892, KY290757-61) are probably misnamed as P. hemiculteri. In order to address this, we performed parallel morphological and molecular analyses on ten samples collected from the gills of the original type host H. leucisculus caught at the neotype locality, Shaoguan, Guangdong Province, southern China.
Based on our re-evaluation of the major morphological features (length and width of the body, buccal suckers, pharynx, attachment clamps, sickle and central hook handle), we determined that our specimens represent P. hemiculteri naturally parasitising the type host H. leucisculus. We evaluated our redescription based on published results of other authors [6,18,39]. While previous drawings of the main morphological characteristics (e.g. the two rows of clamps) were neither consistent nor always accurate, we were able to compare them to our own results and thereby distinguish our samples from other diplozoons. Unfortunately, the relevant type material of P. hemiculteri is now unavailable, hence further morphological comparison was impossible.
Due to the inconsistencies in previously published morphological data on P. hemiculteri, we decided to undertake additional molecular analysis based on amplification of the 762 bp ITS2 fragment from DNA templates of ten adult worms (see above). Our sequence data did not Figure 3. Concatenated phylogenetic tree based on ITS2 sequences for selected diplozoids. Constructed using MrBayes software, the tree includes results for Bayesian inference and maximum-likelihood with PP/bootstrap branch support. Genetic distance between our sequence data is listed below the branch support. Sequences of Paradiplozoon hemiculteri previously stored in GenBank are marked with an asterisk. Data obtained for this study are highlighted by grey gradient. match any sequence saved in GenBank, including some sequences already assigned as P. hemiculteri (GenBank accession numbers DQ098884, DQ098886, DQ098887, DQ098892, KY290757-61 [9]). The genetic distance between ITS2 sequences in our dataset was almost negligible, ranging from 0.1-0.3% (Fig. 3), supporting the assumption that all sequences are related to the same diplozoid species. Recently, Gao (2007) [9] also performed a study on diplozoons originating from H. leucisculus; however, they used molecular methods and sequence data only, without supplementary morphological analysis. The absence of such morphological parameters could have led to incorrect assignment to species. Furthermore, phylogenetic analysis of the majority of P. hemiculteri nucleotide sequences available in public databases (before our own input; marked with an asterisk in Table 1, Fig. 3) indicates that all sequences clustered into one clade together with other species. Moreover, according to the phylogenetic analyses, we might assume that Paradiplozoon is paraphyletic as was also shown in the work of Gao (2007) [9]. However, these results would need revision because the relevant evaluation of potential paraphyly in the Paradiplozoon group cannot be based on the single Sindiplozoon ctenopharyngodoni sequence included in the analysis.
Based on the high morphological similarity and degree of homology in their genomes, it is possible that P. hemiculteri displays high species complexity. Similar high species complexity has been shown in marine capsalid monogeneans [37] and freshwater gyrodactylid monogeneans [14]. Both these studies concluded species complexity based on detailed knowledge of the species' morphological features, combined with high marker similarity during molecular analysis. However, our molecular data could also suggest the existence of cryptic species within Paradiplozoon, as has recently been discovered in other monogenean genera such as Gyrodactylus [19,29]. In these studies, while the main morphological characters of the Gyrodactylus species/lineages (anchor hooks, marginal hooks, ventral bars, marginal hook sickles) were difficult to separate, molecular data (complete cytochrome oxidase I and ITS2 genes) clearly indicated different genotypes of cryptic species. The lack of published morphological descriptions for Chinese diplozoons is a major hindrance for this study as it means these hypotheses cannot be reliably confirmed or refuted at this time. This highlights the importance of combining accurate morphological analysis with molecular analysis of markers as the main means of species identification.
In conclusion, this study provides an accurate redescription of P. hemiculteri, thereby addressing the absence of type material and inaccuracies in the original illustrations and sequence clasifications that have been the norm up until now. Our results confirm P. hemiculteri (Diplozoidae) as the only diplozoid demonstrably parasitising the gills of its Asian fish host, H. leucisculus.