Deep-water parasite diversity in Lake Tanganyika: description of two new monogenean species from benthopelagic cichlid fishes

Lake Tanganyika is the world’s second deepest lake. Its diverse cichlid assemblage offers a unique opportunity for studying a deep-water host-parasite model in freshwater. Low host specificity and a broad host range including representatives of the Bathybatini tribe in the only monogenean parasite described from this habitat, Cichlidogyrus casuarinus Pariselle, Muterezi Bukinga & Vanhove, 2015 suggest a link between lower specificity and lower host density. Conversely, high host specificity and species richness are reported for monogeneans of the lake’s littoral cichlids. We further investigated whether the deep-water environment in Lake Tanganyika is really monogenean species-depauperate by investigating the monogenean fauna of Trematocara unimaculatum (a representative of the tribe Trematocarini, the sister lineage of the Bathybatini) and Benthochromis horii, a member of the tribe Benthochromini, found in the same deep-water habitat as the already known hosts of C. casuarinus. Sclerotised structures of the collected monogenean individuals were characterised morphologically using light microscopy and morphometrics. Both examined cichlid species are infected by a single monogenean species each, which are new to science. They are described as Cichlidogyrus brunnensis n. sp., infecting T. unimaculatum, and Cichlidogyrus attenboroughi n. sp., parasitising on B. horii. Diagnostic characteristics include the distal bifurcation of the accessory piece in C. brunnensis n. sp. and the combination of long auricles and no heel in C. attenboroughi n. sp. In addition C. brunnensis n. sp. does not resemble C. casuarinus, the only species of Cichlidogyrus thus far reported from the Bathybatini. Also Cichlidogyrus attenboroughi n. sp. does not resemble any of the monogenean species documented from the pelagic zone of the lake and is among the few described species of Cichlidogyrus without heel. As two new and non-resembling Cichlidogyrus species are described from T. unimaculatum and B. horii, colonisation of the deep-water habitat by more than one morphotype of Cichlidogyrus is evident. Based on morphological comparisons with previously described monogenean species, parasite transfers with the littoral zone are possible. Therefore, parasites of pelagic cichlids in the lake do not seem to only mirror host phylogeny and the evolutionary history of this host-parasite system merits further attention.

Lake Tanganyika in the African Rift Valley is the second deepest lake in the world and figures as a natural experiment displaying the greatest diversity of speciation mechanisms in cichlids compared to the other major African lakes [28,31]. Lake Tanganyika is inhabited by more than 200 cichlid species [31], classified into 15 different tribes [32][33][34]. Six monogenean genera have been reported to infect African cichlid fishes. Cichlidogyrus Paperna, 1960 is the most species-rich one [15,24,35]. To date, 22 species of Cichlidogyrus have been described in Lake Tanganyika from 18 cichlid hosts representing seven different tribes [12][13][14][15][16][17]. Most records originated from the littoral zone where these parasites were shown to display quite strong host specificity [27]. Cichlid species richness in Lake Tanganyika decreases with water depth [31], with most of the diversity found in shallow littoral habitats. This situation is caused by three main factors: reduction of niche diversity in the deep-water habitat, the fact that the short-wave length blue light spectrum in the depths does not promote diversification mechanisms, and the absence of strong geographic barriers [36][37][38][39]. The same pattern of lower species richness, along with lower host specificity, was documented in monogeneans and suggested to be a consequence of lower host availability. For example, the generalist C. casuarinus Pariselle, Muterezi Bukinga & Vanhove, 2015 has been reported from six bathybatine cichlid species belonging to the genera Bathybates and Hemibates [13,Kmentová et al.,unpublished observation].
To further our knowledge of the monogenean diversity in deep-water Tanganyika cichlids, we examined Trematocara unimaculatum Boulenger, 1901, a representative of the tribe Trematocarini, the sister group of the Bathybatini [34,40], and Benthochromis horii Takahashi, 2008, a cichlid species belonging to another deep-water tribe, Benthochromini, which is only distantly related to the Bathybatini [31,34], but found in the same habitat as the previously reported hosts of C. casuarinus.

Methods
Fishes were bought on fish markets in the capital city of Burundi, Bujumbura (3°23'S, 29°22'E) in September 2013 and identified in situ. Eight specimens of Trematocara unimaculatum and ten of Benthochromis horii were dissected according to the standard protocol of Ergens & Lom [41]. Gills were preserved in ethanol until subsequent inspection for monogeneans under an Olympus SYX7 stereomicroscope. Parasites were mounted on slides under coverslips using Hoyer's medium, enabling visualisation of sclerotised structures [42]. Measurements of sclerotised structures were taken at a magnification of 1000× (objective × 100 immersion, ocular × 10) using an Olympus BX51 microscope with incorporated phase contrast and the software MicroImage version 4.0. In total, 26 different metrical features were measured and are presented in micrometres. The terminology follows [17,43] while "straight length" and "straight width" mean the linear length of the measured structure. Drawings were made using an Olympus BX51 microscope equipped with a drawing tube and OLYMPUS KL 1500 LED illumination and edited with a graphics tablet compatible with Adobe Illustrator 16.

Results
Trematocara unimaculatum and B. horii were each infected by a single different monogenean species belonging to Cichlidogyrus. Following Paperna [44] and Pariselle et al. [45], the genus is characterised by a haptor consisting of two pairs of medium-sized anchors, seven pairs of marginal hooks, two transversal bars (the dorsal one with two auricles; the ventral one curved and articulated), a male copulatory organ (MCO) with copulatory tube and most often a heel and (see [15]) an accessory piece; and a vagina which is not always sclerotised. Both collected Cichlidogyrus spp. are new to science and their descriptions are presented below. Since the species description in dactylogyrid monogenean taxonomy is more than anything else based on the morphology of their sclerotised structures [19,46], the depiction of soft parts and internal organs is omitted and a differential diagnosis focused on details of the parasites' hard parts is provided.  Table 1.] Dorsal anchor with poorly incised roots and well-developed, regularly curved short point. Ventral anchors larger than dorsal anchors, with longer distance between base and point of hook, poorly incised roots, short point. Dorsal bar large, wide, straight, with relatively short, wide auricles. Ventral bar thick, short, branches straight with constant width. Hooks 7 pairs, pairs 1-4, 6 and 7 relatively short (sensu [35]) compared to pair 5, considering ontogenetic development as pair 5 retains its larval size); pair 7 largest. MCO small, with narrow, thin-walled tubular copulatory tube; accessory piece of same length as copulatory tube with distal bifurcation starting in distal quarter, and short heel. Sclerotised vagina not observed.

Differential diagnosis
The anchors of this species resemble those of its non- Tanganyika , 1916), respectively, in their broad base and almost non-incised roots of the anchors. However, the exceptional shape of its accessory piece, with a forked ending, as well as the large ventral anchor in comparison to the dorsal one, make C. brunnensis n. sp. clearly distinguishable among all species of Cichlidogyrus described so far. Moreover, the shape of the anchors, specifically their poorly incised roots and the proportionally large hook, is unique among all other known congeners in Lake Tanganyika casuarinus. The latter species infects bathybatine cichlids, the sister group of the Trematocarini, which among others also comprises T. unimaculatum, the type-host of C. brunnensis n. sp., but C. casuarinus is morphologically very different; it is easily distinguished by the spirallycoiled thickening in the wall of the copulatory tube. Although the forked shape of the accessory piece of C. brunnensis n. sp. is similar to an undescribed species collected from Limnochromis auritus [50], a member of Lake Tanganyika's benthic deep water cichlid tribe Limnochomini, differences in haptoral sclerotised structures, namely the length of the dorsal bar auricles (on average 11.1 μm in C. brunnensis vs c.45 μm in the undescribed species) and the shape of the anchors allows clear distinction between them.
Cichlidogyrus attenboroughi n. sp.  Table 1.] Dorsal anchors with different root size and regularly curved points. Ventral anchors larger in total size and more similar root size than dorsal anchors. Dorsal bar thin, with relatively long narrow auricles. Ventral bar thin, long, with constant width. Hooks 7 pairs, pair 4 relatively long; pairs 1 and 5 of equal length. MCO with slender copulatory tube with relatively thick wall; accessory piece broader than copulatory tube. No heel or sclerotised vagina observed.

Differential diagnosis
There are congeners in Lake Tanganyika that share with C. attenboroughi n. sp. the small size of the first pair of hooks and the presence of long dorsal bar auricles, namely C. makasai and C. vandekerkhovei recorded from Ophthalmotilapia ventralis (Boulenger, 1898),O. nasuta (Poll & Matthes, 1962) and O. boops (Boulenger, 1901). Cichlidogyrus attenboroughi n. sp. differs from these species by the shorter length of the auricles (18.4 μm in C. attenboroughi vs 20 μm in C. makasai and 30 μm in C. vandekerkhovei) and in possessing a MCO with a simple accessory piece and without a heel. Because of the shape of the ventral anchor and the equal size among the first and the second pairs of hooks this species could be mistaken with C. gistelincki infecting Ctenochromis horei (Günther, 1894), C. irenae described from Gnathochromis pfefferi (Boulenger, 1898) and C. steenbergei parasitising Limnotilapia dardennii (Boulenger, 1899) all of which are also present in Lake Tanganyika. However, in contrast to these species, there is no developed heel in the MCO of C. attenboroughi n. sp. Three other species have been described lacking a heel: C. haplochromii Paperna & Thurston, [48] described from Pharyngochromis darlingi (Boulenger, 1911); C. tilapiae Paperna, 1960 recorded from, among others, Oreochromis leucostictus (Trewavas, 1933) and Sarotherodon galilaeus (Linnaeus, 1758) and C. arfii Pariselle & Euzet, 1995 described from Pelmatochromis buettikoferi (Steindachner, 1894). However, the haptoral region of the latter species cannot be confused with C. attenboroughi n. sp. because of the different edge of the dorsal anchor roots, the relative length of the first pairs of hooks (the smallest pair in C. attenboroughi n. sp. and the biggest pair in C. arfii) or the length of the auricles (18.4 μm in C. attenboroughi n. sp. vs 9 μm in C. arfii) [51]. The most evident difference between C. attenboroughi n. sp. and C. tilapiae is the size of the dorsal anchor as well as the maximal straight width of the dorsal bar and the length   [45]. Differences with C. haplochromii are visible in the dorsal bar, which has shorter and less slender auricles in comparison to C. attenboroughi n. sp. [48].

Discussion
The presence of a single monogenean species, phenotypically substantially different from C. casuarinus, on T. unimaculatum, indicates that closely related deep-water cichlid lineages have been colonised by several Cichlidogyrus lineages. Moreover, comparison with other species reported from Lake Tanganyika so far points to multiple origins of the deep-water representatives of Cichlidogyrus. Indeed, this is suggested by the phenotypic similarity of C. brunnensis n. sp. to an undescribed species collected from the benthic cichlid Limnochromis auritus (Boulenger, 1901) [50] and its quite different morphology of sclerotised structures as compared to C. attenboroughi n. sp., which infects another deep-water cichlid species, Benthochromis horii. Furthermore, C. attenboroughi n. sp. shares morphological characteristics of its haptoral region with two species of Cichlidogyrus (C. vandekerkhovei and C. makasai) recorded from three species of Ophthalmotilapia as well as with species described from tropheine cichlids [14,15]. Interestingly, C. casuarinus infecting Bathybatini and C. nshomboi infecting Boulengerochromis microlepis (Boulenger, 1899), a cichlid species phylogenetically closely related to the Trematocarini and Bathybatini [52][53][54], are similar to C. centesimus. The latter infects three species of the genus Ophthalmotilapia, which belongs to the tribe Ectodini. Cichlidogyrus casuarinus, C. nshomboi and C. centesimus share the unique spirally coiled thickening of the wall of the copulatory tube [12,13,15]. Both ectodine and tropheine cichlids occur in shallow water and are only distantly related to Bathybatini, Trematocarini, Benthochromini and Boulengerochromis. Therefore, other scenarios such as host habitat preferences influencing the chance of transmission [55] or shared morphological characters of the host affecting monogenean phenotypes might have played a role in the evolutionary history of this monogenean assemblage [56,57].

Conclusions
The inventory of monogeneans from Lake Tanganyika has been supplemented by the description of C. brunnensis n. sp. and C. attenboroughi n. sp. collected from T. unimaculatum and B. horii, respectively. These are the first monogeneans reported from the respective cichlid species and tribes. The known host range of C. casuarinus still remains limited to the genera Bathybates and Hemibates, although further investigations are needed to confirm this observation. Our result is consistent with taxonomic hypotheses that include the Trematocarini as a separate tribe [34,58], which, together with the Bathybatini, constitute the sister group of the Boulengerochromini [53]. Cichlidogyrus casuarinus, parasitising bathybatine cichlids, morphologically resembles C. nshomboi (collected from B. microlepis, see [12,13]) more than C. brunnensis n. sp. which infects T. unimaculatum, a representative of the Bathybatini's sister group Trematocarini. Hence, probably other speciation mechanisms rather than co-speciation have occurred in the evolutionary history of this deep-water parasite-host system. An exhaustive list of Cichlidogyrus species occurring on deep-water cichlid species in Lake Tanganyika, together with genetic analyses and a cophylogenetic approach, are needed to verify these alternative scenarios. The reported lower monogenean host specificity is probably correlated with small diversity and population densities of hosts [58][59][60] influenced by lower temperature and reduction of light as communication of cichlids is mainly based on visual signals [61]. Decline of parasite diversity is probably also related to the distance from the shore as well as to specific host behavioural characteristics [60]. Although the deepwater monogenean fauna in Lake Tanganyika seems to be less species-rich than the littoral one [27] it is premature to exactly quantify differences in monogenean species richness per host, or to conclude whether the deepwater monogenean fauna is indeed depauperate.