Indopolystoma n. gen. (Monogenea, Polystomatidae) with the description of three new species and reassignment of eight known Polystoma species from Asian frogs (Anura, Rhacophoridae)

The polystomes (Monogenea, Polystomatidae) radiated across semi-aquatic tetrapods including all three amphibian orders, freshwater turtles and the hippopotamus. Prior to this study, phylogenetic analyses revealed that the most diverse and widespread genus, Polystoma, was not monophyletic; a lineage comprising four undescribed species from the bladder of Zhangixalus spp. (Rhacophoridae) in Asia occupied a deep phylogenetic position. Regarding vicariance biogeography and molecular dating, the origin of this lineage is correlated with the breakup of Gondwanaland in the Mesozoic period. Based on a Bayesian analysis of four concatenated genes (18S, 28S, COI and 12S) and morphological evidence, one new genus, Indopolystoma n. gen., and three new species, sampled in Japan and China, are described here: Indopolystoma viridi n. sp. from Z. viridis of Japan, Indopolystoma elongatum n. sp. from Z. arboreus of Japan, and Indopolystoma parvum n. sp. from Z. omeimontis of China. Indopolystoma is unique amongst polystome genera infecting anurans by possessing a small haptor relative to the body size, posteriormost marginal hooklet C1 much bigger than hooklets C2–C8 with conspicuous broad blade and guard and a pair of hamuli lacking a deep notch. Eight species of Asian Polystoma, all from rhacophorids, are transferred as Indopolystoma carvirostris (Fan, Li & He, 2008) n. comb., I. hakgalense (Crusz & Ching, 1975) n. comb., I. indicum (Diengdoh & Tandon, 1991) n. comb., I. leucomystax (Zhang & Long, 1987) n. comb., I. mutus (Meng, Song & Ding, 2010) n. comb., I. pingbianensis (Fan, Wang & Li, 2004) n. comb., I. rhacophori (Yamaguti, 1936) n. comb., and I. zuoi (Shen, Wang & Fan, 2013) n. comb.


Introduction
In contrast to the digeneans that can be found in all groups of vertebrates, monogeneans are mostly parasites of marine and freshwater fishes [52]. With the exception of a few monogeneans that were assigned to the Gyrodactylidae Cobbold, 1864, Iagotrematidae Mañé-Garzón & Gil 1962, and Lagarocotylidae Kritsky, Hoberg & Aubry, 1993, only a single family, the Polystomatidae Gamble, 1896, radiated across semi-aquatic tetrapods including all three amphibian orders (anurans, salamanders and caecilians), freshwater turtles and the common hippopotamus. The Polystomatidae, in modern classification, belong to the order Polyopisthocotylea Odhner, 1912. Nowadays, polystomatids are globally in excess of 180 described species in 26 genera, most of which are endoparasitic in the bladder of amphibian adults (18 genera) and in the pharyngeal cavity, bladder or conjunctival sacs of freshwater turtles (five genera). Whereas fish usually harbor a high diversity of monogeneans on their gills [28,41], no more than two species of polystomes have thus far been recorded per species of anuran host [7,15]. Finally, a high degree of hostspecificity was assumed for polystomatids of especially anuran hosts (see [49] for a review on the diversity of polystomatids).
The Polystomatidae thus provided the opportunity to trace host-parasite co-evolution over an exceptionally long period of time, namely from the ecological transition from marine to terrestrial life at about 425 million years ago (Mya) [50]. Whereas flatworm groups often display diverse body plans, monogeneans and in particular polystomatids show limited interspecies variation [45]. Although hardly any information is known about ancestral forms, the molecular phylogenies published in Bentz et al. [4,5], Verneau et al. [50,51], Badets et al. [2] and Héritier et al. [22] gave an invaluable timescale to date evolutionary events and to infer origins of major monophyletic groups within the family. The extant anuran polystomatids in Asia are less than 20 species that belong to five genera, Diplorchis Ozaki, 1931, Eupolystoma Kaw, 1950, Neoriojatrema Imkongwapang & Tandon, 2010, Polystoma Zeder, 1800 and Sundapolystoma Lim & Du Preez, 2001. Of these, the ubiquitous Polystoma, which is the most speciose-polystome genus known from anurans of the suborder Neobatrachia Reig, 1958, encompasses 14 parasite species (six from China, four from Japan and one each from Sri Lanka, India, Iran and Turkey). Earlier studies based on the phylogeny and historical biogeography of polystomes infecting species of the Neobatrachia [2] revealed that Polystoma was not a monophyletic taxon and that the deep-branched lineage including Polystoma species sampled from rhacophorids of India, Japan and China was strongly correlated with the breakup of Gondwanaland in the Mesozoic period. Badets et al. [2] suggested from cophylogenetic and vicariance analyses supplemented by molecular dating that this lineage probably arose on the Indian subcontinent about 177 Ma when western and eastern Gondwanan components were fully separated, and later colonized southeast Asia following host dispersal after India collided with Asia close to 86 Mya.
India is one of the largest landmass countries in Asia and also well-recognized as a rich biogeographic area in terms of species diversity and endemic species, with its boundaries falling in Himalaya, Western Ghats, Indo-Burma and Sundaland biodiversity "hot spots" [34,37]. For instance, India has a striking anuran diversity with 395 known species [21] of which 286 (73%) are endemic [1]. As shown by many authors, the geological history of India played a crucial role in shaping the current diversity, endemicity, and distribution patterns of amphibian lineages. Before joining Laurasia, India was part of Gondwanaland and gradually became detached from other landmasses during its northward journey across the Tethys Sea [29]. It broke off from Africa about 130 Mya [29] and subsequently from Madagascar about 88 Mya [42]. Its collision with southern Asia occurred during the Paleocene or Early Eocene at 66-56 Mya [3] and gave rise to biotic exchange [6].
According to plate tectonics, rifting and drifting of continents following the breakup of Gondwana provided ample time for animal differentiation. Therefore, the long period of isolation of the Polystoma lineage in the Indian subcontinent [2] should have been sufficient to restrict specific morphological marks for this higher taxon. In the present study, we focused on several specimens of the three undescribed Asian Polystoma species reported in Verneau et al. [51], Badets et al. [2] and Héritier et al. [22] to provide formal descriptions of this new taxon and species. Based on genetic and morphological characters, we bring some evidence that this lineage, which includes polystomes of Asian rhacophorids, is a new genus within the Polystomatidae, and we also reassign eight polystomes previously described as Polystoma to this genus.

Polystome sampling and morphology
Polystomes were recovered from the bladder of three Asian rhacophorids belonging to Zhangixalus Jiang et al. 2019 [21,27], namely Z. viridis (Hallowell) and Z. arboreus (Okada & Kawano) that were both collected in Japan by Hideo Hasegawa on 8 February 1986, and 27 June 2003, respectively, and Z. omeimontis (Stejneger) that was collected in China by Annemarie Ohler on 11 May 2004. A single parasite specimen from each host species was fixed in alcohol for molecular analyses and processed in Badets et al. [2] and Héritier et al. [22]. Whereas some of the material collected in Japan was stained and mounted in Canada balsam, all specimens collected in China were preserved in alcohol. We therefore stained all of them but one with acetocarmine and mounted them permanently in Canada balsam. Specimens were examined using a Nikon NiE compound microscope (Nikon, Netherlands) fitted with a Nikon DS-Ri1 digital camera and drawn using Adobe Illustrator software. Measurements were taken, in micrometers, using a Nikon NIS elements D software program and expressed as the mean, followed by the range in parentheses.

Sequence collection
In order to establish the phylogenetic relationships of the polystomes assumed to belong to a new genus, namely polystomes recovered from Zhangixalus frogs, we selected the four Asian Polystoma species reported in Badets et al. [2] and Héritier et al. [22] (Ozaki, 1948). All these species with their respective accession numbers for two nuclear (18S and 28S) and two mitochondrial (12S and COI) genes are reported in Table 1. Prior to running phylogenetic analyses, we noticed that COI sequences reported in Héritier et al. [22] for polystomes infecting Rhacophorus maximus Günther (known today as Zhangixalus smaragdinus (Blyth)) (JF699303) and Z. viridis (KR856171) were almost identical, differing by only two substitutions, while pairwise comparisons of 12S, 18S and 28S sequences showed higher molecular divergences. This suggested inversion of DNA samples during the amplification process. Using primers L-CO1p and H-Cox1R and following the PCR procedure described in Héritier et al. [22], we therefore re-amplified the COI fragment from both polystome DNA samples recovered by these authors and selected the new sequences for phylogenetic and genetic analyses.

Sequence analyses
18S and 28S sequences were aligned according to the procedure described in Badets et al. [2] and Héritier et al. [22] who took into account the rRNA secondary structure (stems and loops) of both genes. Partial COI and 12S gene sequences were aligned independently using Clustal W under default parameters [44] implemented in MEGA7 software [30]. Because it was too difficult to assess homologous characters within a highly variable region in the 12S, that specific region was deleted prior to running phylogenetic analyses. Using ModelTest implemented in PAUP* version 4.0b9 [43], evolutionary models were estimated independently for the four partitions from the Akaike Information Criterion [38]. All partitions with their own evolutionary model (18S: nst = 6; rates = invgamma; ngammacat = 4; 28S: nst = 6; rates = invgamma; ngammacat = 4; COI: nst = 2; rates = invgamma; ngammacat = 4; 12S: nst = 6; rates = gamma; ngammacat = 4) were subsequently concatenated and a Bayesian analysis was conducted using MrBayes 3.04b [24], with four chains running for one million generations and sampled every 100 cycles. Convergence was assessed with the program Tracer v1.7.1 (http://beast.community/tracer) [39]. A consensus tree was then reconstructed after removing the first 1000 trees (10%) as the burn-in phase. Finally, COI and 28S genetic divergences (p-distances) as well as total differences were also computed for species delimitations following thresholds designed in Du Preez et al. [17]. When all positions containing missing data and/or gaps were eliminated, there were a total of 342 and 1300 positions in the final COI and 28S datasets, respectively.  (Table 2) and 3.7% in the 28S (Table 3), we can indeed consider it is a new genus according to its morphological characteristics (see below).
Taking into account that uncorrected p-distances estimates within Asian polystomes are well beyond 10% in the COI ( Table 2) and 0.2% in the 28S (Table 3), we can assume that there are four separate species according to the 1.2% and 0.07% genetic divergences that were considered as the species-level threshold within polystomes of amphibians from COI and 28S sequences, respectively [17]. Furthermore, though several substitutions were found between Asian polystomes, at least one unique change (autapomorphy) was observed in each of the four undescribed species, regardless of the gene of interest, COI or 28S. These results reinforced our hypothesis of four distinct polystome species.
Finally, a Bayesian tree inferred from the analysis of a dataset comprising 14 full-length 18S sequences (Table 1)  and Héritier et al. [22]. B. w. refers to Blommersia wittei and Z. s. to Zhangixalus smaragdinus. See also Table 1 for other host species. Scale bar represents 0.1 substitution/site. (EU734835), P. zuoi Shen, Wang & Fan, 2013 (KF850147) and two other undescribed Polystoma spp. infecting Rana chaochiaoensis Liu (U734834) and Hyla annectans (Jerdon) (EU979386) of China, showed that the last two species were more closely related to species of Polystoma than they were to species of Indopolystoma (results not shown).

Generic diagnosis
Body large and oblong. Intestical caeca bifurcate, diverticulated, confluent posteriorly with posterior diverticulum barely entering haptor. Intestinal anastomoses usually absent but at most a single anastomosis may be present. Vas deferens extends antero-medially, opens into seminal vesicle that opens into genital bulb, armed with 8-9 genital spines. Ovary comma shaped and prominent, sinistral, in anterior 20% of body. Oviduct arises from posterior region of ovary, connected by genito-intestinal canal to sinistral caecum, receives common vitelline duct, ascends giving rise to short tubular uterus that often holds a single egg but as many as 40. Vitellaria distributed throughout body proper except in region around ovary and reproductive ducts, extending marginally into haptor; left and right vitelline ducts join to form common vitelline reservoir near ovary, with duct to oviduct. Two prominent vaginae, antero-lateral to ovary; left and right vaginal ducts connected to respective vitelline ducts. Egg operculate, oval and lacking a filament. Haptor short relative to body size (haptor/total body length ratio < 0.15 for most species) with three pairs of suckers, one pair of hamuli and 16 marginal hooklets. Hamuli curved, unbranched in base (handle and guard not well separated) and with short recurved hook. Prominent big posteriormost  marginal hooklet C1 (see numbering, [36]) with prominent broad blade and guard, in contrast with smaller hooklets C2-C8. Indopolystoma spp. are parasites of the bladder of rhacophorid frogs from Asia. Etymology: The prefix indo refers to India, which was assumed to be the center of origin for this new genus [2].

Differential diagnosis
Indopolystoma viridi is similar to I. elongatum and I. parvum in terms of body shape, haptor/total body length ratio and shape of haptoral sclerites. However, it differs from the same two species by the general morphology of intestinal caeca and its body size (8550 lm vs. 14,791 lm for I. elongatum and 4714 lm for I. parvum). It differs from all other species of Indopolystoma in having intestinal diverticula without anastomoses.  Note. To the exception of the newly described species, body measurements for all other species were extracted or estimated from Crusz and Ching [12], Diengdoh and Tandon [13], Fan et al. [19,20], Meng et al. [33], Shen et al. [40], Yamaguti [53] and Zhang and Long [54].    (Fig. 5).

Differential diagnosis
Indopolystoma elongatum is well characterized by its body size and shape. This species is much bigger and more elongated (body length 14,791 lm) than any other species of Indopolystoma, though there is an overlap of size values with I. indicum. Indopolystoma elongatum can be easily distinguished from the later by the number of intrauterine eggs. None of the specimens of I. elongatum have more than a single egg in utero while I. indicum has as many as 40.
Remarks: Zhangixalus arboreus hosts two polystomes, namely I. elongatum and I. rhacophori (see below), which is uncommon within anuran polystomes. However, Z. arboreus and Z. schlegelii occur sympatrically in Japan [1]. The possibility of a misidentification can thus not be excluded especially since molecular evidence on host identity is currently not available. We consider for now that both I. elongatum and I. rhacophori are separate species primarily on the basis of body length and haptor/total body length ratio (0.06 for I. elongatum vs. 0.18 for I. rhacophori).

Differential diagnosis
Indopolystoma parvum can be easily distinguished from I. viridi and I. elongatum by its body size, haptor shape and general morphology of intestine. This species is much smaller than I. elongatum (4714 lm vs. 14,791 lm) while it is only half the size of I. viridi (4714 lm vs. 8550 lm). It shows haptor sub-spherical (vs. sub-rectangular) and intestinal caeca with haptoral anastomosis. It differs from all other congeners, apart from I. pingbianensis, in lacking medial anastomoses. Indopolystoma parvum is smaller than I. pingbianensis (4714 lm vs. 9428 lm).   [21,26].
Remarks: Although the authors of the original description did not draw the marginal hooklets [20], the general morphology of this species, including haptor/total body length ratio (0.13) and hamuli shape (unbranched), is consistent with the diagnosis of Indopolystoma. Furthermore, a phylogeny based on partial 18S sequences only (unpublished results) showed that this species fell within the clade of Indopolystoma species.
Indopolystoma carvirostris was originally recorded in China from P. cavirostris. However, P. cavirostris only occurs in Sri Lanka [21]. According to Inger et al. [26], Chinese records of P. cavirostris likely apply to R. bisacculus or R. verrucosus.
Site: Bladder. Type-locality: Hakgala Strict Natural Reserve, Sri Lanka. Remarks: Although the authors of the original description did not draw the marginal hooklets [12], the general morphology of this species, including haptor/total body length ratio (0.11) and hamuli shape (unbranched), is consistent with the diagnosis of Indopolystoma.
Remarks: Although the authors of the original description did not draw the marginal hooklets [13], the general morphology of this species, including haptor/total body length ratio (0.10) and hamuli shape (unbranched), is consistent with the diagnosis of Indopolystoma.
Site: Bladder. Type-locality: Yunnan province, China. Remarks: Although the authors of the original description did not draw the marginal hooklets [20], the general morphology of this species, including haptor/total body length ratio (0.09) and hamuli shape (unbranched), is consistent with the diagnosis of Indopolystoma.
Site: Bladder. Type-locality: Kurama, near Kyoto, Japan. Remarks: Although the author of the original description did not draw the marginal hooklets [53], the general morphology of this species, including hamuli shape (unbranched), is consistent with the diagnosis of Indopolystoma. The haptor/ total body length ratio of about 0.18 is bigger than that of any other Indopolystoma spp. with the exception of I. zuoi (Shen, Wang & Fan, 2013) n. comb. As discussed earlier for I. elongatum, which infests the same host, we consider for now that both I. elongatum and I. rhacophori are two separate species primarily on the basis of body length and haptor/ total body length ratio.  [21].
Remarks: Although the authors of the original description did not draw the marginal hooklets [40], the general morphology of this species, including hamuli shape (unbranched), is consistent with the diagnosis of Indopolystoma. The haptor/ total body length ratio of about 0.26 is so much bigger than that of any other Indopolystoma spp. Nevertheless, a phylogeny based on partial 18S sequences only (unpublished results) showed that this species fell within the clade of Indopolystoma species.
Site: Bladder. Type-locality: India. Remarks: Indopolystoma sp. was tentatively assigned to P. indicum from Z. smaragdinus by Verneau et al. [51], Badets et al. [2] and Héritier et al. [22]. However, because we did not have any information on morphological characteristics of this species, which is nested in a clade with I. viridi, I. elongatum and I. parvum ( Fig. 1; see also [2,22,51]), we must for now consider it as an undescribed species of Indopolystoma.

Discussion
In this paper, one genus and three new species are being described, and eight previously described species of Polystoma as well as an undescribed species from Z. smaragdinus are being transferred to the new genus. Whereas species of Polystoma in Asia infect mostly ranids and hylids, species of Indopolystoma are only reported from species assigned to rhacophorid genera, such as Feihyla, Kurixalus, Polypedates, Rhacophorus, Taruga and Zhangixalus. These results clearly illustrate that rhacophorids are frequent hosts for Indopolystoma in Asia in which polystome speciation and diversification would have occurred during the long isolation of India from Madagascar and Africa. The Rhacophoridae is currently represented by 422 valid species arranged in 20 genera [1,21]. As such, they account for roughly 6% of the living anurans of the world. These neobatrachians occur almost exclusively in India as well as in southeast Asia, with only one genus, Chiromantis, having species known from Africa [1,21]. Therefore, we may expect a greater diversity of polystomes within Indopolystoma which should be restricted to Asia, where rhacophorids have undergone spectacular radiation "out of India" [32].
The interspecific morphological variation of polystomes is generally limited [45]. Herein, the haptor along with sclerotized structures (or sclerites) makes Indopolystoma a unique genus amongst all polystome genera infecting anuran hosts. Despite their morphological plasticity, the haptoral sclerites which are the "hallmark of monogeneans" [9] remain a significant character for morphological identification. Within amphibian polystomes, the haptoral sclerites are typically represented by 16 marginal hooklets and one pair of hamuli, although a few exceptions are known [14,25,31]. These characters have been largely investigated because of their usefulness in polystome delimitation [8,10,11,14,16,25,35,36,[46][47][48]. The species of Indopolystoma are characterized by a posteriormost marginal hooklet C1, with prominent broad blade and guard, much larger than those of hooklets C2-C8, unlike that of Polystoma and Diplorchis spp. where the hooklets are all morphologically similar, although posteriormost marginal hooklet C1 is also larger than hooklets C2-C8. On the other hand, all marginal hooklets are equal in length and similar in shape within species of Eupolystoma, Neoriojatrema and Sundapolystoma. Whereas the presence of hamuli within Indopolystoma allows the differentiation of that genus from Eupolystoma and Neoriojatrema in which hamuli are lacking, their particular structure with a handle not separated from the guard, i.e. they lack a deep notch in base, is not unique as it is similar with some species of Polystoma. Finally, the haptor/total body length ratio is also of value for delimitating Indopolystoma. For all species of Indopolystoma, with the exception of I. rhacophori and I. zuoi, this value is less than 0.15, while it is greater for most other anuran polystomes. Chiromantis rufescens (Günther) is currently the only rhacophorid frog in Africa known to host a polystome, namely Polystoma chiromantis Dupouy & Knoepffler, 1978. Although marginal hooklets were not described in the original description [18], this parasite shares the elongated body and small haptor of Indopolystoma. According to Imasuen (unpublished thesis), marginal hooklet C1 of P. chiromantis has the typical shape as seen in Polystoma species. Therefore, in the absence of molecular evidence, we herein consider this species as belonging to Polystoma, which could have arisen from host-switching in Africa.
In conclusion, even though three main characters, i.e. the shape of the posteriormost marginal hooklet C1, the haptor/total body length ratio, and host species belonging to Rhacophoridae, constitute key characters for the morphological delimitation of Indopolystoma, it is important that genotyping of several polystome worms be conducted prior to the description process, as recommended by Héritier et al. [23].