Diplectanids from Mycteroperca spp. (Epinephelidae) in the Mediterranean Sea: Redescriptions of six species from material collected off Tunisia and Libya, proposal for the 'Pseudorhabdosynochus riouxi group’, and a taxonomic key

Diplectanid monogeneans are gill parasites that can infect fish in huge numbers and thus become harmful, especially in maricultured fish. It is therefore useful to have taxonomic tools, such as keys, to identify species. The following diplectanid species from groupers of the Mediterranean Sea were studied: five species of Pseudorhabdosynochus Yamaguti, 1958, including P. riouxi (Oliver, 1986) Kritsky & Beverley-Burton, 1986 from the dusky grouper Mycteroperca marginata, P. enitsuji Neifar & Euzet, 2007, P. bouaini Neifar & Euzet, 2007, P. dolicocolpos Neifar & Euzet, 2007 and P. sinediscus Neifar & Euzet, 2007 from the goldblotch grouper M. costae, and Echinoplectanum echinophallus (Euzet & Oliver, 1965) Justine & Euzet, 2006 from the dusky grouper. New material was obtained from fish collected from off Tunisia and Libya and compared to the type-material and voucher specimens in museum collections. Identifications of fish were confirmed by barcoding of cytochrome c oxidase subunit I (COI) sequences. The sclerotized vagina was considered the most important structure for systematics. The three species P. riouxi, P. bouaini, and P. enitsuji share a common general structure of the sclerotized vagina with a conspicuous spherical secondary chamber. We thus propose the ‘Pseudorhabdosynochus riouxi group’ to accommodate them. Pseudorhabdosynochus dolicocolpos has an elongate vaginal structure that is completely different from all its congeneric species reported from the Mediterranean Sea, and Pseudorhabdosynochus sinediscus has a sclerotized vagina in which the secondary chamber is not visible, and a haptor without squamodiscs. A taxonomic key to diplectanid species on Mycteroperca spp. in the Mediterranean Sea is proposed; it includes ten species of Pseudorhabdosynochus and one species of Echinoplectanum.

Introduction Diplectanid monogeneans are ectoparasites on the gills of fish, generally smaller than 1 mm, but often present in high numbers, even on wild fish. Some diplectanid species can pullulate on fish in confinement and thus become a major health problem for fish in aquaculture [1][2][3][4][5]. On groupers especially, several diplectanid species have been reported as a concern for mariculture [6][7][8][9][10][11]. It is therefore important to provide taxonomic identification keys to parasite species, which are currently lacking for diplectanids from groupers in the Mediterranean Sea.
On the basis of type-material and specimens newly collected from off Tunisia and Libya (North Africa), we found that P. riouxi, P. bouaini, and P. enitsuji are morphologically close. The three species exhibit a sclerotized vagina with a conspicuous spherical secondary chamber. Hence, the 'P. riouxi group' is proposed here to accommodate them. In contrast, P. dolicocolpos and P. sinediscus are separated from all Mediterranean congeners by their unique vaginal structure. A key for diplectanid species of Mycteroperca spp. in the Mediterranean Sea is provided.
For fish identifications by barcoding, the Cytochrome c Oxidase subunit I (COI) sequences from our sampled specimens were compared with other COI sequences from different geographical regions published in GenBank; sequences and detailed comparisons were previously published [22,31,32].

Fish sampling and identification
Fish were purchased at the fish markets in Sfax, Tunisia and Tripoli, Libya. These were previously caught by fishermen in the nearby coastal waters of the Mediterranean Sea. In all cases, the fish were dead when available for parasitological studies. Fish were identified morphologically according to keys [33,34], and these identifications were challenged by the COI sequences of individual fish (Table 1). Fish nomenclature follows [35] and [36].
as 'b'); and c) dehydrated in an ethanol series, stained with carmine and permanently mounted in Canada balsam (designated as 'c') [39]. Specimens were drawn using an Olympus BH2 microscope equipped with drawing apparatus and DIC optics. The terminology for the sclerotized parts, i.e. the male quadriloculate organ and the vagina follows Justine (2007) [14]. Measurements, in micrometers, were taken with the help of a custom-made transparent rule and are expressed as the mean followed in parentheses by the range, the standard deviation when n!30, and (n) the number of observations; measurements were taken as in Fig 1. The measurements of the right-hand haptoral hard-parts and left-hand equivalents were pooled. The measurements of the holotype are separated and indicated by 'h'. Drawings were scanned and redrawn on a computer using Adobe Illustrator. The museum abbreviation used is as follows: MNHN, Muséum National d'Histoire Naturelle, Paris; NHMUK, Natural History Museum, London.   Fig 2I).
Sclerotized vagina comprises trumpet, primary canal, primary chamber, secondary canal and secondary chamber. Trumpet funnel-shaped, sometimes with "ring-like" distal limit, followed by long primary canal looped in its proximal or medium portion, with thick wall. Primary chamber, elongate with wide lumen, much smaller than secondary chamber that is robust and spherical. Primary and secondary chambers surrounded by very thick and rigid wall. Secondary canal linking the two chambers not always visible (Fig 2A-2H  Comments. Pseudorhabdosynochus riouxi was originally described by Oliver (1986) as Cycloplectanum riouxi from Mycteroperca marginata in the Mediterranean Sea off France [30]. It was then transferred to Pseudorhabdosynochus by Kritsky & Beverley-Burton (1986) [44].
Although all slides were prepared by the same method, "mounted in carmine" (i.e. mounted in Canada balsam after carmine staining), Oliver (1986) mentioned that the specimens of P. riouxi found in the Mediterranean Sea differed greatly in the measurements from those of the Atlantic Ocean. Yang et al. (2005) subsequently provided a new illustration of P. riouxi on the basis of eight paratypes (1982.12.14.1-3) deposited in the NHMUK collections by Guy Oliver; these paratypes were collected from the same host, M. marginata, in the Atlantic Ocean but in a different location, off Ireland's coast [30,45].    After examination of the holotype, paratype, and additional voucher specimens deposited in the MNHN collections by Guy Oliver after his retirement, we consider that our newly collected specimens from the same host and from the same region (i.e. the Mediterranean Sea) are conspecific with P. riouxi (Table 2, Figs 2 and 3). However, they are different from Yang et al.'s drawing by the proportions of vaginal parts and by the total number of concentric rows of rodlets in the squamodiscs (14-17 in the Mediterranean Sea vs. 23-24 in the Atlantic Ocean, as indicated by Oliver, 1986).
It is likely that the species of Pseudorhabdosynochus on M. marginata in the northeastern Atlantic is different from P. riouxi, and its status would require additional study and collection of new specimens from the same locality. It would also be interesting to know whether the specimens collected off the Canary Islands, mentioned but not described [43], are closer to the Mediterranean species or to the North Atlantic species; the Canary Islands are in the Atlantic but south of the Mediterranean Sea. However, the status of P. riouxi itself, based on its holotype, which is from the Mediterranean Sea, is unchanged, and specimens from the Mediterranean Sea with the same vaginal structure can safely be attributed to P. riouxi.
Differential diagnosis. Pseudorhabdosynochus riouxi most closely resembles P. bouaini Neifar & Euzet, 2007 and P. enitsuji Neifar & Euzet, 2007 from the goldblotch grouper Mycteroperca costae in the general morphology of the sclerotized vagina. In the three congeners, the sclerotized vagina has a conspicuous spherical secondary chamber. Pseudorhabdosynochus riouxi is readily distinguished from P. bouaini by the length of its primary canal (very long in P. riouxi vs. very short in P. bouaini), and by the shape of its primary chamber (more elongate in P. riouxi). It differs from P. enitsuji by the size of the vagina (carmine: 37-53 in P. riouxi vs. 27-34 in P. enitsuji), by the shape of the primary canal (looped in P. riouxi vs. roughly straight or curved) and its length (longer in P. riouxi), and by the external diameter of the secondary chamber (carmine: 17-26 in P. riouxi vs. [15][16][17][18].
A series of species from off the American Atlantic coast show vaginal morphologies that are similar to P. riouxi, but none is identical. The comparison with these species is given below in the paragraph concerning the' P. riouxi group'.
Molecular identification of fish via DNA barcoding: The COI sequences from four specimens from Tunisia were already published and identified as M. costae ( Table 1).

Redescription (Figs 4 and 5;
Sclerotized vagina comprises anterior trumpet followed by primary canal; primary canal heavily sclerotized, very short, roughly straight, its lumen wide; connection between primary canal and primary chamber anterior; primary chamber very similar in wall thickness to primary canal, heavily sclerotized, ovate with wide lumen; secondary canal inserted at anterior limit of secondary chamber, well visible, its lumen thin; secondary chamber robust, spherical, much larger than primary chamber (Fig 4B-4I). Total length of sclerotized vagina h 40, b 47     [46], the type-locality of P. bouaini should be corrected from "off Sfax" to "off Zarzis".
In the slides containing the paratypes MNHN HEL2-Th75 and MNHN voucher specimens labelled as P. bouaini, additional specimens of P. enitsuji were found. The two species co-occur on the gills of M. costae in the Mediterranean Sea and eastern Atlantic Ocean.
Differential diagnosis. Just like P. bouaini, P. enitsuji and P. riouxi possess a sclerotized vagina with a conspicuous spherical secondary chamber. Among species from the Mediterranean Sea, Pseudorhabdosynochus bouaini is closest morphologically to P. enitsuji. The two species can be differentiated by the size of the vagina, i.e. larger vagina in P. bouaini (picrate: 40 vs. 33; Berlese: 48 vs. 39), the length of the primary canal (shorter in P. bouaini), and the external diameter of the secondary chamber (picrate: 28 in P. bouaini vs. 19; Berlese: 28 in P. bouaini vs. 23). Pseudorhabdosynochus bouaini is readily distinguished from P. riouxi by its very short primary canal (vs. very long and looped in P. riouxi) and by its very small primary chamber.
A series of species from off the American Atlantic coast show vaginal morphologies that are similar to P. bouaini, but none is identical. The comparison with these species is given below in the paragraph concerning the 'P. riouxi group'.
Molecular identification of fish via DNA barcoding: The COI sequences from four specimens from Tunisia were already published and identified as M. costae ( Table 1).
Site of infection: Gill lamellae. Type-locality: Off Sfax, Tunisia [24]. Other localities: Off Zarzis (Tunisia) [24]; off Dakar (Senegal) [24]; off Tripoli (fish market), Libya (present study).   7F). Male copulatory organ a quadriloculate organ (Fig 6A), first (anterior) chamber as      Sclerotized vagina comprises trumpet funnel-shaped, primary canal heavily sclerotized, roughly straight; followed by curved primary chamber; secondary canal opening into distal limit of secondary chamber, secondary chamber robust and spherical, with thick wall, much larger than primary chamber (Figs 6B-6J, 8 Nomenclatural comments. The type-slide of Pseudorhabdosynochus enitsuji deposited by Neifar & Euzet in the MNHN collections contains more than one specimen of P. enitsuji (3 specimens) among which the holotype was not distinctly marked. Neifar & Euzet (2007) provided a composite drawing of this species but the holotype was not represented. We consider that this is a case, although atypical, in which the holotype is "lost or destroyed" since it is impossible to recognize it among several specimens. According to Article 75.3 of the Code [46], the specimen labelled B in our Fig 8 is designated here as the neotype of P. enitsuji. Differential diagnosis. Pseudorhabdosynochus enitsuji is close to P. bouaini and P. riouxi in terms of general structure of sclerotized parts, i.e. male copulatory organ, squamodiscs, haptoral parts and mainly the sclerotized vagina. This species is distinguished from its two congeners by having the smaller size of vagina and secondary chamber (more details above in P. riouxi and P. bouaini).
A series of species from the American Atlantic coast show vaginal morphologies that are similar to P. enitsuji, but none is identical. The comparison with these species is given below in the paragraph concerning the' P. riouxi group'.
Among the Mediterranean species of Pseudorhabdosynochus infesting the gills of members of Mycteroperca, only P. riouxi, P. bouaini, and P. enitsuji share the common characteristic of a sclerotized vagina with a conspicuous and heavily sclerotized spherical secondary chamber (Fig 9). In order to facilitate their distinction within Pseudorhabdosynochus, which currently includes more than 80 valid species described on groupers throughout the world, the 'P. riouxi group' is proposed here to accommodate them.
Species of the 'P. riouxi group' are well differentiated from their congeners in the Mediterranean Sea, but the fact that several species of Pseudorhabdosynochus may occur on both sides of the Atlantic [21,25] requires a comparison with species from the western side of the Atlantic.
Six species, found on Mycteroperca spp. off the American Atlantic coast, share a single heavily sclerotized spherical chamber with the 'P.  [25,47,48]. Indeed, Kritsky et al. ([25] p. 13) remarked that groupers assigned to Mycteroperca are parasitized by "a complex of similar species", which have in common a sclerotized vagina "with a single subspherical to ovate chamber and a distal tube that is strongly recurved near its articulation with the vaginal vestibule". Unfortunately, the descriptions or redescriptions of these species did not include a detailed interpretation of the canals and chambers of the vagina and it is not ascertained whether the "single subspherical to ovate chamber" is really the secondary chamber, as in the Mediterranean species of the 'P. riouxi group'. These species also have in common with 'the P. riouxi group' the characteristic of squamodiscs with closed internal rows [25,48]. We consider it likely that they are close to the 'P. riouxi group'.
None of the above six species have vaginal morphology identical to the three species of the "P. riouxi group" in the Mediterranean Sea. Differentiation on other characters is as follows: P. capurroi has characteristic "twisted" lateral bars [25,48]. P. contubernalis has an male copulatory organ (MCO) with a very thin-walled first chamber and tegumental scales on the peduncle [25].
P. kritskyi has an MCO with a very short cone and dorsal bars with a large median end [25]. P. vascellum has an MCO with a delicate, almost non-existent cone [25]. The "P. riouxi group" could be extended to these six species and thus could include nine species; we prefer to keep only the Mediterranean species within this group and wait for additional molecular information to ascertain relationships of the "P. riouxi group" to species off the American Atlantic coast.
Molecular identification of fish via DNA barcoding: The COI sequences from four specimens from Tunisia were already published and identified as M. costae ( Table 1).
Differential diagnosis. No species of Pseudorhabdosynochus in the Mediterranean Sea shows a sclerotized vagina similar to that of P. dolicocolpos. Pseudorhabdosynochus dolicocolpos is readily distinguished from the species of the 'P. riouxi group' and all other Mediterranean species of Pseudorhabdosynochus by its very long primary canal and the general structure of its sclerotized vagina. This species has a ring-shaped trumpet, a character also found in the 'beverleyburtonae group' [22]; however, P. dolicocolpos can be differentiated from the species of this group by the general structure of the rest of the vagina.
Of the described species not from the Mediterranean Sea, a single species has vaginal morphology resembling that of P. dolicocolpos: P. variabilis Justine, 2008 from M. morrhua off New Caledonia [27]. Both species also share similar shapes of haptoral parts. However, P. variabilis has an MCO with a thin-walled first chamber (vs. thick in P. dolicocolpos), a relatively longer secondary canal and a shorter sclerotized vagina (c. 50 vs. 75-120 μm), and squamodiscs with 2-3 circular rows (vs. 1-2 V-shaped rows). It is, however, striking that the two species of Pseudorhabdosynochus with this vaginal morphology are both parasites of species of Mycteroperca, although widely separated geographically (Mediterranean Sea vs. South Pacific). A common origin is likely.
Molecular identification of fish via DNA barcoding: The COI sequences from four specimens from Tunisia were already published and identified as M. costae ( Table 1).
Sclerotized vagina comprises anterior trumpet, primary canal, primary chamber, and secondary canal (secondary chamber absent). Anterior trumpet, heavily sclerotized, funnelshaped; primary canal long, S-shaped, its wall slightly thinner distally, opens into posterior The 'Pseudorhabdosynochus riouxi group' (Monogenea, Diplectanidae) part of primary chamber; primary chamber small, heavily sclerotized; very short secondary canal, with thin lumen, emerges from anterior part of primary chamber; secondary chamber not seen, probably very small and embedded in wall of primary chamber (Fig 12A-12D). Total length of sclerotized vagina b 29 (25-34, n = 7).
Comments. The inventory number assigned to the holotype of P. sinediscus as MNHN HEL10-Th83 on the label of the type-slide does not match the one published with the original description of this species (MNHN HEL11-Th84).
Differential diagnosis. Pseudorhabdosynochus sinediscus is characterized, among its Mediterranean congeners, by the general structure of its vagina in which the secondary chamber is not visible. However, in terms of vaginal morphology, P. sinediscus is close to P. regius Chaabane, Neifar & Justine, 2015 from Mycteroperca rubra and the 'beverleyburtonae group' [22] from M. marginata, M. costae and the mottled grouper M. rubra (Bloch). Pseudorhabdosynochus sinediscus and P. regius are readily distinguished by the shape of their primary canal (Sshaped in P. sinediscus vs. straight in P. regius). This species shares the same shape of the vagina with the 'beverleyburtonae group' (S-shaped) but differs from it by the small size of its vagina.
In addition, P. sinediscus is the only species within Pseudorhabdosynochus known to have a haptor without squamodiscs. All other currently known species of Pseudorhabdosynochus have two squamodiscs. Pseudorhabdosynochus monosquamodiscusi Balasuriya & Leong, 1995, from Lates calcarifer (Bloch), has a single squamodisc [49], but this species is now considered a member of Laticola and a junior synonym of L. latesi [50].    (Table 7). Pharynx not visible. Sclerotized male copulatory organ, funnel-shaped, with muscular reservoir at its anterior extremity with four muscular layers, and spiny cirrus in its posterior part; length b 112 (100-135, n = 5). Sclerotized vagina not seen.
Comments. This species was first described by Euzet & Oliver as Diplectanum echinophallus, from the dusky grouper in the Mediterranean Sea [51], and later included by Oliver in Cycloplectanum Oliver, 1968 due to the presence of two innermost rows forming circles in its squamodiscs [42].  redescribed the species and considered it a member of their new genus Echinoplectanum [16]. Among the seven species assigned to Table 7. Taxonomic key to Diplectanids on Mycteroperca spp. in the Mediterranean Sea (Fig 14). Shading of cells applied only for readability.

Taxonomic key
A taxonomic key to the species of diplectanids on Mycteroperca spp. in the Mediterranean Sea is proposed in Table 7. The main morphological features used in the taxonomic key are illustrated in Fig 14.