Gyrodactylus magadiensis n. sp. (Monogenea, Gyrodactylidae) parasitising the gills of Alcolapia grahami (Perciformes, Cichlidae), a fish inhabiting the extreme environment of Lake Magadi, Kenya

A new species of Gyrodactylus von Nordmann, 1832 is described from the gills of Alcolapia grahami, a tilapian fish endemic to Lake Magadi. This alkaline soda lake in the Rift Valley in Kenya is an extreme environment with pH as high as 11, temperatures up to 42 °C, and diurnal fluctuation between hyperoxia and virtual anoxia. Nevertheless, gyrodactylid monogeneans able to survive these hostile conditions were detected from the gills the Magadi tilapia. The worms were studied using light microscopy, isolated sclerites observed using scanning electron microscopy, and molecular techniques used to genetically characterize the specimens. The gyrodactylid was described as Gyrodactylus magadiensis n. sp. and could be distinguished from other Gyrodactylus species infecting African cichlid fish based on the comparatively long and narrow hamuli, a ventral bar with small rounded anterolateral processes and a tongue-shaped posterior membrane, and marginal hooks with slender sickles which are angled forward, a trapezoid to square toe, rounded heel, a long bridge prior to reaching marginal sickle shaft, and a long lateral edge of the toe. The species is also distinct from all other Gyrodactylus taxa based on the ITS region of rDNA (ITS1–5.8s–ITS2), strongly supporting the designation of a new species. These findings represent the second record of Gyrodactylus from Kenya, with the description of G. magadiensis bringing the total number of Gyrodactylus species described from African cichlids to 18.


Introduction
Fishes of the genus Alcolapia Thys van den Audenaerde, 1969 are African cichlids that occur in two of the most severe environments in the eastern African Rift Valley, namely the Soda Lakes Magadi and Natron, located in Kenya and Tanzania, respectively [53,67]. These two lakes were once a continuous water body, separated~10,000 years ago [7,14,22,52]. Fish subsist in scattered lagoons and sites at the periphery of Lake Magadi where particularly extreme conditions

Materials and methods Collection
Fish were collected (permit number NCST/RRI/12/1/ MAS/99) from the Fish Spring Lagoon of Lake Magadi ( Fig. 1B and C) in July 2013 and June 2018. Fish were specifically collected from the most peripheral pool (1 in Fig. 1C). Fish were euthanised on site by severing the spinal cord and the whole fish preserved in either formalin (July 2013) or absolute ethanol (June 2018). At the University of Johannesburg, gills were removed from the preserved fish and studied for the presence of parasites, with attached worms removed from the gills using fine dissecting needles and detached worms picked carefully from the fixative.

Morphometric analyses
Formalin fixed worms were washed in water, dehydrated in a series of ethanol (30%, 50%, and 70% ethanol), and subsequently mounted and cleared in glycerine ammonium picrate (GAP) [40] for examination of the haptoral sclerites and male copulatory organ (MCO). Some specimens were also stained with Horen's trichrome [41] and cleared and mounted in lactophenol. Light microscopy, using both phase and differential interference contrast approaches, were used to study the shape and dimensions of sclerotized structures using a Zeiss Axioplan 2 imaging light microscope with Axiovision 4.7.2 software (Carl Zeiss, Jena, Switzerland). Micrographs were used to draw taxonomically important structures. The measurements of hamuli and other haptoral sclerites (28 point-to-point measurements) were taken according to Shinn et al. [60] and García-Vásquez et al. [15], and the measurements of the parasite's whole body were taken according to Christison et al. [8]. All measurements are in micrometres unless stated otherwise, and presented as a mean and range in parentheses. Measurements of the haptoral sclerites are presented in full alongside the only other Gyrodactylus species described from Kenya (G. malalai) in Table 2. After analyses, worms were removed from the slides, dehydrated through a graded ethanol series, and mounted in Canada balsam for permanent storage [11].
Due to the fixation of samples in formalin and high concentration ethanol, SEM of whole worms did not produce viable results. To study the sclerites at higher magnification, worms were digested on a concavity slide using the digestion buffer from a DNA extraction kit, digested tissue removed, rinsed, and prepared for SEM [35,42,[58][59][60]. Slides were gold sputter coated using an Emscope SC500 Sputter Coater (Quorum Technologies, Newhaven, UK) and studied with using a Vega 3 LMH scanning electron microscope (Tescan, Brno, Czech Republic) at 3.4 kV.
Etymology: The species is named after Lake Magadi from which the specimens were collected.   All 10 specimens produced identical sequence data for the ITS fragment analysed. The amplified region was 882 bp in length, with the size of the 18S, ITS1, 5.8S, ITS2, and 28S rDNA fragments 24, 385, 158, 291, and 24 bp long, respectively. Alignment of the sequence to other data retrieved from GenBank produced a 1088 bp alignment, of which 636 positions were conserved, 446 variable, and 399 parsimony informative. Gyrodactylus magadiensis n. sp. was only distantly related to most other Gyrodactylus species (Table 4), most closely to Gyrodactylus branchicus Malmberg, 1964 (23.2%) and most distantly to Gyrodactylus katamba García-Vásquez, Guzmán-Valdivieso, Razo-Mendivil, and Rubio-Godoy 2018 (25.4%). Distances of 0. 45-25.4% were observed between species included in these analyses, while intraspecific distances of 0.00-1.14% were seen. The latter would suggest that taxa with more than 1.14% sequence divergence are distinct species, indicating that sequences for distinct species with less than that (in this case G. katamba and Gyrodactylus lamothei   , 1910) need to be revised to produce a robust criteria to identify species based on ITS rDNA. Topologies of phylogenetic analyses based on ML and BI methods produced similar results, thus a single combined tree is shown in Figure 4. In all cases, G. magadiensis n. sp. formed a distinct, well supported linage from its congeners. Gyrodactylus magadiensis n. sp. appeared to be most closely related to a clade of G. katamba and G. lamothei in all cases.

Differential diagnosis
In comparison to other Gyrodactylus species described from African cichlid fishes, the marginal hooks of G. magadiensis n. sp. are most similar to those of Gyrodactylus cichlidarum Paperna, 1968, Gyrodactylus yacatli García-Vásquez, Hansen, Christison, Bron and Shinn, 2011 and Gyrodactylus ulinganisus García-Vásquez, Hansen, Christison, Bron and Shinn, 2011 in that the sickle is smoothly curved (except G. yacatli) and the toe is almost square. The marginal hooks of the new species can be distinguished from these species in that the toe is more pronounced than in G. cichlidarum and G. ulinganisus; the indentation on inferior edge between the toe tip and shaft attachment point is more pronounced than in G. cichlidarum and G. ulinganisus; the bridge prior to reaching the marginal sickle shaft is longer than in G. cichlidarum and G. ulinganisus; the lateral edge of the toe is longer than in G. yacatli; and the sickle is angled forward (similar only to G. yacatli). The heel of G. magadiensis n. sp. is also notably rounded, only slightly similar to that of Gyrodactylus thysi Vanhove, Snoeks, Volckaert and Huyse, 2011, Gyrodactylus thlapi Christison, Shinn and van As, 2005 and the illustration of Gyrodactylus niloticus Cone, Arthur and Bondad-Reantaso, 1995 by Cone et al. [9] (junior synonym of G. cichlidarum [13]).
In terms of the ventral bar, G. magadiensis n. sp. can be differentiated from other Gyrodactylus species infecting African cichlids based on the distinct tongue shape of the membrane, medial notch in posterior of membrane, rounded lateral ends of the bar itself, and anterolateral processes rounded and slightly curved outward. These features are most similar to the ventral bar of G. cichlidarum, but specifically the shape of the membrane can easily distinguish these species. The long and narrow nature of the hamuli are reminiscent of those of G. malalai, Gyrodactylus ergensi Přikrylová, Matějusová, Musilová and Gelnar, 2009 and Gyrodactylus nyanzae Paperna, 1973. However, the hamuli of G. magadiensis n. sp. can be distinguished by the lack of an indentation of the root above the attachment of the dorsal bar as in G. malalai and G. ergensi, and the more robust root in comparison to G. nyanzae. The MCO of G. magadiensis n. sp. has six spinelets, whereas most of the other species for African cichlids have 4, 5 or 7 (with the exception of Gyrodactylus shariffi Cone, Arthur and Bondad-Reantaso, 1995 and G. cichlidarum which can have six).

Discussion
Gyrodactylus magadiensis n. sp. is the second record of a Gyrodactylus species infecting a cichlid from Kenya. However, this is the first gyrodactylid to be described from the extreme conditions of Lake Magadi and from Alcolapia grahami. The ability of G. magadiensis n. sp. to persist and thrive in the extreme condition of Lake Magadi is truly impressive. Although G. salinae survives in the extreme salinity and temperature of Cervia Saline in Italy [46], this species is genetically very distant from G. magadiensis n. sp. and the morphologies of the species do not share many similarities. It is thus unlikely that these species, which are both able to survive extreme conditions, originated from the same lineage, indicating that this adaptation has occurred at least twice convergently.
Based on both morphology and genetic data, the material studied here is markedly distinct from all other information available for African gyrodactylid monogeneans. Regarding the genetic identity of G. magadiensis n. sp., a gyrodactylid infecting a marine Gasterosteiformes host (Gasterosteidae) collected in Finland and Belgium is the most closely related (23.2%) based on uncorrected p-distances, while none of the sequences for Gyrodactylus species from African cichlids were identified as close relatives using BLASTn, even though most of these African species have representative sequence data. In fact, it would appear that most of the species identified as close relatives to G. magadiensis n. sp. are marine species, which is puzzling. However, as can be seen from the topology of the phylogenetic analyses shown in Figure 4, G. magadiensis n. sp. groups with a clade of G. lamothei and G. katamba, both species from freshwater systems in Mexico. This association is only weakly supported by BI analyses (0.48), but more strongly by ML approaches (82%).

Conclusion
Gyrodactylus magadiensis n. sp. is described here on the basis of its morphology and genetic identity. The species can be distinguished from congeners parasitising other African cichlids based on the comparatively long and narrow hamuli, a ventral bar with small rounded anterolateral processes and a tongue-shaped posterior membrane, and marginal hooks with slender sickles which are angled forward, a trapezoid to square toe, rounded heel, a long bridge prior to reaching the marginal sickle shaft, and a long lateral edge of the toe. Genetically, this species is distinct from all other monogenean species, with more than 23.2% pairwise divergence between it and its closest relatives.