Gyrodactylus salinae n. sp. (Platyhelminthes: Monogenea) infecting the south European toothcarp Aphanius fasciatus (Valenciennes) (Teleostei, Cyprinodontidae) from a hypersaline environment in Italy

Background Historically, non-native species of Gambusia (Poeciliidae) have been used to control larval stages of the Asian tiger mosquito, Stegomyia albopicta Reinert, Harbach et Kitching, 2004 throughout Italy. The potential utility of indigenous populations of Aphanius fasciatus (Valenciennes) (Teleostei: Cyprinodontidae) as an appropriate alternative biological control is currently being explored. A sub-sample of ten fish collected from Cervia Saline, Italy (salinity 65 ppt; 30°C) to assess their reproductive capability in captivity, harboured a moderate infection of Gyrodactylus von Nordmann, 1832 (Platyhelminthes, Monogenea). A subsequent morphological and molecular study identified this as being a new species. Results Gyrodactylus salinae n. sp. is described from the skin, fins and gills of A. fasciatus. Light and scanning electron microscopical (SEM) examination of the opisthaptoral armature and their comparison with all other recorded species suggested morphological similarities to Gyrodactylus rugiensoides Huyse et Volckaert, 2002 from Pomatoschistus minutus (Pallas). Features of the ventral bar, however, permit its discrimination from G. rugiensoides. Sequencing of the nuclear ribosomal DNA internal transcribed spacers 1 and 2 and the 5.8S rRNA gene and a comparison with all species listed in GenBank confirmed they are unique and represent a new species (most similar to Gyrodactylus anguillae Ergens, 1960, 8.3% pair-wise distance based on 5.8S+ITS2). This represents the first species of Gyrodactylus to be described from Aphanius and, to date, has the longest ITS1 (774 bp) sequenced from any Gyrodactylus. Additional sampling of Cervia Saline throughout the year, found G. salinae n. sp. to persist in conditions ranging from 35 ppt and 5°C in December to 65 ppt and 30°C in July, while in captivity a low level of infection was present, even in freshwater conditions (0 ppt). Conclusions The ability of G. salinae n. sp. to tolerate a wide range of salinities and temperatures shows its potential to readily adapt to several environmental conditions. These findings, together with the fact that A. fasciatus is a protected species and is considered as a biological control organism, necessitate further studies on the ecology and virulence of G. salinae n. sp.


Background
Toothcarp is a colloquial term used to describe members of the order Cyprinodontiformes, which comprises ten fish families, namely Anablepidae, Aplocheilidae, Cyprinodontidae, Fundulidae, Goodeidae, Nothobranchiidae, Poeciliidae, Profundulidae, Rivulidae and Valenciidae [1]. Within the family Cyprinodontidae, the south European toothcarp Aphanius fasciatus (Valenciennes) is one of the more commonly occurring neritic species within the Mediterranean and is characteristically known to be an eurythermic and euryhaline fish species, for its ability to tolerate a wide range of temperatures (5-39°C) and salinities (0-180 ppt), respectively [2]. Aphanius fasciatus has a widespread distribution along the Italian coastline, principally in brackish waters [3], although increasing anthropogenic activity has caused a general decline in numbers. For this reason, A. fasciatus is considered to be a species that is "dying out" and as such, is listed under Appendix III "Protected Fauna Species" after the Bern Convention on the conservation of European wildlife and natural habitats, and under Annex II of the Council Directive 92/43/EEC on the conservation of natural habitats and of wild fauna and flora in the European Community [4]. It has recently been used as a biological control against the larvae of the Asian tiger mosquito Stegomyia albopicta (= Aedes albopictus) Reinert, Harbach et Kitching, 2004 which are vectors for a range of human infectious diseases, including Chikungunya fever, dengue fever, West Nile fever and yellow fever [5][6][7].
During a recent research study investigating the artificial reproduction of A. fasciatus in captivity as part of a large scale restocking and mosquito control initiative in Italy [8,9], several specimens collected, under licence, on a number of occasions were found to harbour an infection of Gyrodactylus von Nordmann, 1832, principally on the skin and fins and, to a lesser degree, on the gills. Infected tissues were removed from the moribund fish and subsequently sent to the Institute of Aquaculture, University of Stirling (Scotland, UK) for identification. Given the importance of A. fasciatus as a protected species and its utility as an alternative indigeneous biological control agent to the introduced Gambusia spp.
(Poeciliidae), more information on its Gyrodactylus fauna was needed.

Specimens collection and preparation
A total of ten adult A. fasciatus (total length 4-7 cm; weight~3 g) was collected during July 2008 from isolated pools in Cervia Saline, located in the Emilia Romagna region in northern Italy, and fixed in 70% ethanol. All ten specimens collected from the drying, landlocked pools were moribund individuals, a consequence of reduced water availability and increased algal growth. The skin, fins and gills of each fish were subsequently screened for metazoan parasites using an Olympus SZ40 stereomicroscope at ×4 magnification. Specimens of Gyrodactylus were removed using mounted triangular surgical needles. All ten fish were infected but given the condition of the fish on collection, the intensity of infection can only be estimated at between 10-30 parasites per fish.
The alcohol-fixed parasites were subsequently rinsed in distilled water and representatives prepared as whole mounts using ammonium picrate glycerine following the procedure detailed by Malmberg [10]. Additional specimens had their opisthaptors removed using a scalpel, which were then individually subjected to proteolytic digestion on glass slides, as described in Paladini et al. [11]. The largely tissue-free opisthaptoral hook preparations were then mounted in ammonium picrate glycerine using an 18 × 18 mm coverslip, the edges of which were sealed with a commercial brand of nail varnish. The corresponding body of each specimen of Gyrodactylus was fixed in 90% ethanol for subsequent molecular characterisation.
For scanning electron microscopy (SEM), single specimens of Gyrodactylus were subjected to full proteolytic digestion on 11 mm round glass coverslips to obtain tissue-free attachment hooks. Each digestion took approximately 60 min, with 3 μl of digestion solution being added every 10 min, punctuated by the addition of 5 μl distilled water for five times to remove digested tissue residues and debris. For each step, the digested hook preparations were placed in a Petri dish to protect them from extraneous dust and placed in an incubator at 55°C to help digestion, followed by a final incubation at 40°C overnight to dry. The position of hooks on each coverslip was subsequently marked using tiny, adhesive, triangular labels positioned using forceps under an Olympus BX51 compound microscope at ×10 magnification. The coverslips were then attached onto aluminium stubs with bi-adhesive round labels, sputter-coated with gold using an Edwards S150B sputter coater and then examined using a JEOL JSM5200 SEM operating at an accelerating voltage of 10 kv.

Morphological analysis
For the morphological study, images of the opisthaptoral hard parts and the male copulatory organ (MCO) were captured at magnifications of ×40 and ×100 oil immersion using MRGrab 1.0.0.4 (Carl Zeiss Vision GmbH, 2001) software and a Zeiss AxioCam MRc digital camera mounted on an Olympus BX51 compound microscope, using a ×0.75 interfacing lens. Drawings of the taxonomic features were made from the captured images. Each Gyrodactylus specimen was subjected to morphometric analysis taking a total of 27 point-topoint measurements on the opisthaptoral hooks using a JVC KY-F30B 3CCD video camera mounted on an Olympus BH2 microscope using a ×2.5 interfacing lens at ×100 oil immersion and the KS300 (ver.3.0) (Carl Zeiss Vision GmbH, 1997) image analysis software, combined with the specific macro for gyrodactylids, Point-R (Bron & Shinn, University of Stirling). The point-to-point hook measurements for the specimens are given in micrometres as the mean ± 1 standard deviation followed by the range in parentheses, and follow those described in Shinn et al. [12], plus three additional measurements of the dorsal bar (total length, width and attachment point length
The obtained sequences were subjected to a BLAST search (available at http://www.ncbi.nih.gov/BLAST/) to identify similar sequences among other species of Gyrodactylus in GenBank [15]. The sequences with the highest similarity to our sequences were downloaded and the 5.8S and ITS2 fragments were aligned in Clustal W implemented in MEGA 4 [16]. Pair-wise genetic distances were computed in MEGA 4 according to the evolutionary model that was selected by jModelTest 0.1.1 [17]. The sequences were also scanned for repeat elements using the program Tandem Repeats Finder [18].

Site of infection
Skin, fins and occasionally gills.

Environmental conditions under which specimens were collected
Salinity and temperature in July 2008, 65 ppt and 30°C, respectively.

Type material
Fifteen specimens were studied for light microscopy.

General
A species profile including host and taxonomic details is provided on the on-line databases http://www.gyrodb. net [19] and http://www.monodb.org [20].

Etymology
Named after the Italian generic name for a hypersaline water body i.e. "salina" (= saline in English) and the broad salinity tolerance exhibited by this species of Gyrodactylus.

Molecular characterisation
The total fragment (1379 bp) consists of the 3' end of the 18S subunit (32 bp), the ITS1 (774 bp), the 5.8S gene (158 bp), the ITS2 (402 bp) and the 5' end of the 28S subunit (13 bp). Both specimens had identical sequences. Gyrodactylus salinae n. sp. appeared most closely related to Gyrodactylus species belonging to the G. (Paranephrotus) and G. (Neonephrotus) sub-genera (sub-genera according to Malmberg [10]) based on the pair-wise distances (Tamura-Nei gamma corrected distances [21] using the 5.8S-ITS2 fragment,    (Figure 2d-g). Given the potential overlapping distribution of these two fish hosts within the Mediterranean Sea [1], it is important to detail the subtle features in hook morphology that permit their discrimination from one another. Two paratypes of G. rugiensoides were examined for morphological comparison with G. salinae n. sp. The paratypes were re-measured and the marginal hook sickle re-drawn and overlaid with G. salinae n. sp. for a direct comparison (Table 1; Figure 2e-g). There was a good agreement between the measurements obtained in the current study and those presented in Huyse & Volckaert [22]. The present study, however, provides an additional set of measurements taken from the type material of G. rugiensoides for direct comparison with the opisthaptoral features of G. salinae n. sp. Although the hamuli roots in both species narrow after their union with the shaft, giving the anterior edge of the dorsal bar attachment point a small but distinct edge, this appears to be more prominent in G. rugiensoides than on the hamuli of G. salinae n. sp. The ventral bar attachment points also differ; those of G. rugiensoides appear flat and rectangular, while those of G. salinae n. sp. are indented. The prominent ventral bar processes and the longer, slender ventral bar membrane of G. salinae n. sp. contributes to the discrimination of the two species. The similar morphology of the marginal hook sickles of both species though, requires careful examination. The union of the marginal hook shaft with the sickle divides the width of the sickle base into 3:2 (heel:toe) in both species, however, the sickle base is deeper in G. rugiensoides than in G. salinae n. sp. with a more angular, rectangular heel and a steeper faced, more robust toe (Figure 2f-g). The sickle shaft and sickle tip of G. rugiensoides is broader, proportionately so, than that of G. salinae n. sp. giving the latter the appearance of having a more open deeper sickle aperture. A comparison of soft body features suggests that the posterior pharynx bulb of G. rugiensoides (78.4 μm in diameter) is considerably larger than that of G. salinae n. sp. which measures in 24.6 μm in diameter. Although Huyse & Volckaert [22] described the MCO of G. rugiensoides as armed with one principal spine and five small spines, a closer examination of the paratypes shows that G. rugiensoides possesses nine small spines, as that of G. salinae n. sp. The MCOs of the two species are similar in length (G. salinae n. sp. 14.4 μm vs G. rugiensoides 14.8 μm) but that of G. rugiensoides is slightly wider (12.7 vs 15.1). A molecular comparison of G. rugiensoides and G. salinae n. sp. showed that they were quite distinct. The pair-wise distance amounted to 14% (based on the 5.8S + ITS2 fragment). The ITS1 sequences were more difficult to align, due to length differences (up to 132 bp). Both species belong to the so-called marine rugiensis-group that also includes G. anguillae.
Additional sampling A. fasciatus from the Cervia Saline at several time points throughout the year, found that G. salinae n. sp. was present on their hosts in waters ranging from 35 ppt and 5°C during December to 65 ppt and 30°C during July in the wild, while under captive conditions, fish even maintained a low level of infection in freshwater (0 ppt) (pers. obs.).

Discussion
Gyrodactylus salinae n. sp. is the first species to be formally described from Aphanius fasciatus and, also the first from the genus Aphanius Nardo, 1827. Over 420 species of Gyrodactylus have been described [19,20] and only six are known to parasitise cyprinodontids, all of them are recorded from Cyprinodon spp. (see Table 3). These are G. cyprinodontis, G. nevadensis and G. saratogensis, all from Cyprinodon nevadensis nevadensis Eigenmann et Eigenmann; G. hargisi and G. mobilensis from Cyprinodon variegatus variegatus Lacepède, and G. tularosae from Cyprinodon tularosa Miller et Echelle. No supporting molecular data, however, is available for any of these species. Morphologically, the marginal hook sickles of the Gyrodactylus species described from cyprinodontid hosts are markedly different (Figure 3). The marginal hook sickles of these species were re-drawn from the paratypes and holotypes, where available, and a comparison of their morphology with G. salinae n. sp. is given in Figure 3. The marginal hook sickles of G. cyprinodontis, G. nevadensis and G. saratogensis are morphologically very similar to one another (see Figure 3a, g, i). These three species were found on the Table 2 Pair-wise genetic distances based on the 5.8S and ITS2 rDNA fragment of Gyrodactylus salinae n. sp. and the Gyrodactylus species showing highest similarity in the BLAST search on GenBank (Tamura-Nei + gamma model) same host, Cyprinodon n. nevadensis, and they differ in the size of their opisthaptor and its skeletal elements, notably the hamulus total length (G. cyprinodontis = 49 μm; G. nevadensis = 33 μm; G. saratogensis = 26 μm) which allows their ready discrimination from each other [23]. The opisthaptoral hooks of the type material of G. cyprinodontis and G. nevadensis were not completely flat and therefore reconstructed drawings of the marginal hook sickles of these two species were necessary and are presented here and compared with G. salinae n.   [26].
The fact that the current species is (both morphologically and molecularly) more closely related to Gyrodactylus species parasitising marine gobies and eels than to those infecting cyprinodontids, might suggest that the ecology of the host, rather than host phylogeny plays an important role in this host-parasite system. Malmberg [10] found G. anguillae only on migrating elvers, which are relatively small and abundantly found in estuaries, as is the common goby Pomatoschistus microps (Krøyer), host to G. micropsi. There are other examples of ecological radiations onto distant-related hosts in Gyrodactylus, e.g. the G. wageneri-group primarily infects cyprinids but they are also found on sticklebacks, percids and cottids [27].
There is nothing known yet on the effect of G. salinae n. sp. on its host, but its closest relative so far, G. anguillae, has been reported as a pest in the culture of anguillid eels [28,29].
Based on the nuclear ITS fragment, the genus Gyrodactylus can be divided in two groups, one with a 'short' (347-473 bp) and one with a 'long' (535-688 bp) ITS1 fragment [30,31]. These ranges have recently been extended both to the lower end, with Gyrodactylus cichlidarum Paperna, 1968 having the shortest ITS1 sequence determined thus far (343 bp; acc. no. DQ124228), and G. salinae n. sp. having the longest ITS1 fragment (774 bp). The sequence of G. salinae n. sp. is also unique as the 5.8S gene is 158 bp long, whereas all other Gyrodactylus species so far had an invariant length of 157 bp. The new species clusters in the marine rugiensis-group which is part of the monophyletic 'long ITS' group, and appears most closely related to G. anguillae. Despite the morphological resemblance between G. salinae n. sp. and G. rugiensoides, they are genetically quite different (especially in the ITS1 fragment). This underlines again the fact that morphological and molecular evolution is not always correlated. The close affinity with G. anguillae and Gyrodactylus cf. micropsi Huyse, Audenaert et Volckaert, 2003 as based on the genetic distances does not necessarily equal sister species relationships. The true sister species of G. salinae n. sp. has probably not been sequenced yet since the difference between Gyrodactylus species from closely related goby species was for example much smaller (2-5%) [32] than the distance between G. anguillae and G. salinae n. sp.

Conclusions
Although the specimens of A. fasciatus from Cervia Saline were moribund at the time of collection, this was not a consequence of the moderate G. salinae n. sp. infection but of the pool conditions (30°C, 65 ppt) from which they were collected which had a water depth of less than 15 cm and a heavy growth of algae. Given the ability of G. salinae n. sp. to tolerate a wide range of salinities and temperatures, it would appear that this species has the potential to readily adapt to the full spectrum of environmental conditions. What is unknown, however, is the potential of this species to survive on other neritic fish hosts when cohabiting isolated pools or interacting with other fish species in open, full strength seawater (i.e. 35 ppt). The potential utility of Aphanius fasciatus as a biological control against the larvae of the Asian tiger mosquito Stegomyia albopicta, in addition to the finding of a new parasite species, raises concerns regarding the health status of this host due to the parasitic infection. Further studies are required in order to define the potential role of G. salinae n. sp. as a pathogen, and the consequences that this finding might have on the use of indigenous populations of A. fasciatus as an alternative biological control to the use of various Gambusia species. The ecological risks of introducing Gambusia spp. to non-native environments has been stressed by many authors [33][34][35]. Cross breeding with indigenous species may lead to the extinction of the latter given that certain species have been shown to hybridise [36][37][38]. There are also the attendant risks of disease introduction linked with the movement of fish into new habitats [39]. The use of native fish, rather than introduced species, as biological controls, therefore, is highly recommended to avoid these potential ecological impacts.
In accordance with section 8.6 of the ICZN's International Code of Zoological Nomenclature, copies of this article are deposited at the following five publicly accessible libraries: Natural History Museum, London, UK; American Museum of Natural History, New York, USA; Museum National d'Histoire Naturelle, Paris, France; Russian Academy of Sciences, Moscow, Russia; Academia Sinica, Taipei, Taiwan.