Introduction
Freshwater ecosystems have been especially affected by the unintentional introduction of exotic species (Sala et al. Reference Sala, Chapin, Armesto, Berlow, Bloomfield, Dirzo, Huber-Sanwald, Huenneke, Jackson, Kinzig, Leemans, Lodge, Mooney, Oesterheld, Poff, Sykes, Walker, Walker and Wall2000; Gherardi, Reference Gherardi and Gherardi2007; Havel et al. Reference Havel, Kovalenko, Thomaz, Amalfitano and Kats2015). The analysis of the mechanisms underlying successful invasions contributes to the protection of aquatic ecosystems against the impacts of future invaders (Nunes et al. Reference Nunes, Tricarico, Panov, Cardoso and Katsanevakis2015). Among these mechanisms, parasitic interactions could play a critical role in the invasion success of an exotic host, through their impact on both exotic and native species (Dunn and Hatcher, Reference Dunn and Hatcher2015). Since the review of Torchin et al. (Reference Torchin, Lafferty, Dobson, McKenzie and Kuris2003) and subsequent studies (Torchin and Mitchell, Reference Torchin and Mitchell2004; Gendron et al. Reference Gendron, Marcogliese and Thomas2012; Paterson et al. Reference Paterson, Townsend, Tompkins and Poulin2012), it is suggested that invasive species are released from their enemies as they occupy new areas – a phenomenon named the ‘enemy release hypothesis’ (Williamson, Reference Williamson1996); but some exotic parasites may also be introduced with the exotic host. In this case, they can colonize native communities and decrease their performance, giving a competitive advantage to their initial exotic host (Gendron et al. Reference Gendron, Marcogliese and Thomas2012). On the other hand, introduced species can be then colonized by non-specific native parasites from their new environment (Poulin and Mouillot, Reference Poulin and Mouillot2003; Gendron et al. Reference Gendron, Marcogliese and Thomas2012). These spillover–spillback phenomena, i.e. the transfer of parasites from invasive hosts to native ones (spillover), and from native hosts to invasive ones (spillback), can lead to the partial substitution of the parasite community in an invasive host along its invasive pathway (Kelly et al. Reference Kelly, Paterson, Townsend, Poulin and Tompkins2009).
The round goby, Neogobius melanostomus (Pallas, 1814), is a Ponto-Caspian species that spread only recently in Western Europe (Roche et al. Reference Roche, Janac and Jurajda2013). Until the 1980s, its uppermost distribution range was limited to the northeast of Bulgaria along the Danube River (Francová et al. Reference Francová, Ondračková, Polačik and Jurajda2011). The Danube and the Rhine rivers have been connected after the construction of the Rhine–Main–Danube corridor in 1992 (Leuven et al. Reference Leuven, van der Velde, Baijens, Snijders, van der Zwart, Lenders and bij de Vaate2009). The round goby was recorded downstream of the Vienna hydropower dam (Austrian stretch of the Danube) in 2000 and soon after in the Dutch Rhine Delta in 2004 (Van Beek, Reference Van Beek2006). First records in the Upper Rhine date back to 2011 and 2012 in the Gambsheim fishway and several places along the French–German border, i.e. the unchannelled part of the Upper Rhine (Manne et al. Reference Manne, Poulet and Dembski2013). The round goby is now widespread along the Rhine River and its abundance during an electrofishing can reach 84% of the total catch (Manne et al. Reference Manne, Poulet and Dembski2013), with an estimated density ranging from 2 to 8 individuals m−2. The geographic expansion of N. melanostomus is still in progress, since it was observed for the first time in the Seine Basin in 2015, within the Rouen harbour (Agence Française de la Biodiversité, personal communication). Its flexible trophic diet, with a majority of crustaceans and insects (Brandner et al. Reference Brandner, Auerswald, Cerwenka, Schliewen and Geist2013), and its aggressive behaviour are considered as the main reasons for its invasion success in introduced aquatic communities (Steinhart et al. Reference Steinhart, Marschall and Stein2004; Borza et al. Reference Borza, Eros and Oertel2009; Stevove and Kovac, Reference Stevove and Kovac2013).
The aim of our study is to document the parasite assemblage of the round goby N. melanostomus in populations recently established in the Upper Rhine (French–German border), and compare it to the parasite assemblages reported from populations closest from its native area. We hypothesized that the round goby has a poor macroparasite assemblage in the newly invaded area compared with the native one. Depending on the phenomenon driving the changes in its parasite community, the parasite assemblage of the round goby from the Old Rhine should be more similar to the Danube ones or to local ones (spillback phenomenon; Kelly et al. Reference Kelly, Paterson, Townsend, Poulin and Tompkins2009). To test these hypotheses, we first documented the potential changes in the parasite assemblage reported from the round goby throughout its European invasive pathway along the Rhine–Main–Danube corridor through a literature review. We also collected a large number of N. melanostomus from the Upper Rhine (and some native fishes) to establish parasite assemblage in the newly invaded area. We expected the acanthocephalan Pomphorhynchus laevis to be one of the most abundant parasite of the round goby (Francová et al. Reference Francová, Ondračková, Polačik and Jurajda2011), and we therefore focused on this species to better address the changes in parasite diversity driven by invasion. With a wide geographical distribution across Europe, P. laevis is one of the most common acanthocephalan parasites of freshwater fishes (Perrot-Minnot et al. Reference Perrot-Minnot, Špakulová, Wattier, Kotlík, Düşen, Aydoğdu and Tougardin press, and references therein). This intestinal parasite of freshwater fish uses several amphipod species as intermediate hosts. It uses a broad range of freshwater and brackish-water fish species as final hosts, mainly Cyprinidae, the largest family of freshwater fish, but also Salmonidae (Médoc et al. Reference Médoc, Rigaud, Motreuil, Perrot-Minnot and Bollache2011; Perrot-Minnot et al. Reference Perrot-Minnot, Špakulová, Wattier, Kotlík, Düşen, Aydoğdu and Tougardin press). Pomphorhynchus laevis has occasionally and locally integrated in its life cycle an additional fish species as paratenic host, i.e. a facultative host used for the completion of the life cycle but in which no development occurs (Médoc et al. Reference Médoc, Rigaud, Motreuil, Perrot-Minnot and Bollache2011).
The biogeographic history of P. laevis has been recently reconstructed, and reveals the existence of two lineages genetically and geographically distinct in the Danubian system, one in the Danube, Volga and Vistula rivers, and the other in its tributaries (Perrot-Minnot et al. Reference Perrot-Minnot, Špakulová, Wattier, Kotlík, Düşen, Aydoğdu and Tougardin press). We therefore analysed the genetic diversity of P. laevis in our samples based on the mitochondrial DNA c oxidase subunit I (COI) gene, and used this sequence information to identify the geographic origin of the Upper Rhine lineages of P. laevis and propose a scenario for its introduction.
Material and methods
Literature review
A literature review was performed on Web of Science (up to July 2017) to make a census of papers focusing on round goby parasites along the Rhine–Main–Danube corridor. The main key-words used were ‘Neogobius melanostomus’, ‘parasite’ and ‘European freshwater’. The references in each paper have also been checked to decrease the chance of missed studies. A total of 13 papers were found, covering 20 locations along the invasion corridor (Fig. 1). The presence and the prevalence of each parasite species were recorded for each location (Table 1).
Fig. 1. Location of the scientific studies published on Neogobius melanostomus and its parasites along the Rhine–Main–Danube corridor. (a, b) Kvach and Skóra (Reference Kvach and Skóra2007); (c, s) Kvach et al. (Reference Kvach, Kornyychuk, Mierzejewska, Rubtsova, Yurakhno, Grabowska and Ovcharenko2014); (d, g, i) Francová et al. (Reference Francová, Ondračková, Polačik and Jurajda2011); (e, i) Ondračková et al. (Reference Ondračková, Francová, Dávidová, Polačik and Jurajda2010); (f) Košuthová et al. (Reference Košuthová, Košco, Letková, Košuth and Manko2009); (h) Ondračková et al. (Reference Ondračková, Dávidová, Pečinková, Blažek, Gelnar, Valová, Černý and Jurajda2005); (j, k) Mühlegger (Reference Mühlegger2008); (l, n) Emde et al. (Reference Emde, Kochmann, Kuhn, Plath and Klimpel2014); (m) sampling location for the present study (vicinity of Ottmarsheim, Upper Rhine); (o) Emde et al. (Reference Emde, Rueckert, Palm and Klimpel2012); (p) Ondračková et al. (Reference Ondračková, Valová, Hudcová, Michálková, Šimková, Borcherding and Jurajda2015); (q) Kvach and Winkler (Reference Kvach and Winkler2011); (r) Kvach and Skóra (Reference Kvach and Skóra2007); (t) Rolbiecki (Reference Rolbiecki2006); (u) Herlevi et al. (Reference Herlevi, Puntila, Kuosa and Fagerholm2017).
Table 1. Data from a literature review of parasite assemblages in Neogobius melanostomus along the Rhine–Main–Danube corridor, restricted to the three mean macroparasites reported
Richness: total number of parasite species found in the study; prevalence: the percentage of fishes parasitized; mean abundance: mean number of parasites found in all individual fishes sampled; native area: lower part of the Danube River (Reference Francová, Ondračková, Polačik and JurajdaFrancová et al. 201Reference Bollache, Devin, Wattier, Chovet, Beisel, Moreteau and Rigaud1); non-native area: area where the round goby is considered as an invasive fish.
(a, b) Kvach and Skóra (Reference Kvach and Skóra2007); (c, s) Kvach et al. (Reference Kvach, Kornyychuk, Mierzejewska, Rubtsova, Yurakhno, Grabowska and Ovcharenko2014); (d, g, i) Francová et al. (Reference Francová, Ondračková, Polačik and Jurajda2011); (e, i) Ondračková et al. (Reference Ondračková, Francová, Dávidová, Polačik and Jurajda2010); (f) Košuthová et al. (Reference Košuthová, Košco, Letková, Košuth and Manko2009); (h) Ondračková et al. (Reference Ondračková, Dávidová, Pečinková, Blažek, Gelnar, Valová, Černý and Jurajda2005); (j, k) Mühlegger (Reference Mühlegger2008); (l, n) Emde et al. (Reference Emde, Kochmann, Kuhn, Plath and Klimpel2014); (m) sampling location for the present study (vicinity of Ottmarsheim, Upper Rhine); (o) Emde et al. (Reference Emde, Rueckert, Palm and Klimpel2012); (p) Ondračková et al. (Reference Ondračková, Valová, Hudcová, Michálková, Šimková, Borcherding and Jurajda2015); (q) Kvach and Winkler (Reference Kvach and Winkler2011); (r) Kvach and Skóra (Reference Kvach and Skóra2007); (t) Rolbiecki (Reference Rolbiecki2006); (u) Herlevi et al. (Reference Herlevi, Puntila, Kuosa and Fagerholm2017).
Host sampling
Fish were collected in three sites of the Upper Rhine River located near Ottmarsheim (Grand Est, France), 20 km downstream of Basel, along the left bank. Two samplings were collected, one in February and one in May 2016, e.g. before the spring flood of the Rhine River. The sites are located on a relic of the uppermost stretch of the Upper Rhine, called the Old Rhine, a 50 km long by-passed single-bed channel located downstream of the Kembs dam. Fish were sampled from the three sites (A, B, C hereafter) that are quite close from each other. These locations belong to a morphodynamic restoration programme of controlled bank erosion conducted by Electricité de France (EDF) (Garnier and Barillier, Reference Garnier and Barillier2015). Site A is the upstream site (47°44′51·87″N, 7°32′38·72″E) and can be considered as a positive control where geomorphic units and microhabitats are varied. Site B (47°44′43·95″N, 7°32′38·72″E) is 400 m downstream and has been the subject of an ecological restoration, with controlled bank erosion and artificial groynes implementation. The project was initiated by EDF. These actions are aimed at using the natural erosion capacity of floods to supply the Old Rhine River with aggregates, and to diversify the river mosaic thus allowing a potential gain in the biodiversity of alluvial environments. Site C (47°46′03·15″N, 7°31′54·84″E) is 2 km downstream of site B and can be considered as a negative control with a low diversity of habitats and a bank mostly composed of big rocks and a concrete area.
Electrofishing was used to collect the dominant fish species at each site. On completion of sampling, a total of 179 N. melanostomus were collected, 63 in site A, 57 in site B and 59 in site C. We also collected 18 barbels, Barbus barbus (L., 1758), and 11 chubs, Leuciscus cephalus (L., 1758) at the same time. To comply with animal welfare rules, fish were anaesthetized with precise doses of clove oil to cause death before transportation. The length (±1 mm) and the weight (±0·01 g) of each fish were determined before the dissection with aseptic precautions. During dissections, fish were sexed based on gonadal structure. For each individual, the eight gill arches were dissected off, and observed under a Leica ×40 binocular microscope. Macroparasites were also collected in the body cavity, and within the gut dissected under a binocular microscope. All parasites collected were stored separately in 99% ethanol.
DNA extraction, sequencing and phylogenetic analyses
Most of the parasites collected were surrounded by a membranous layer, which could represent a major host–parasite interface containing host haemocytes (Dezfuli et al. Reference Dezfuli, Simoni, Duclos and Rossetti2008). In order to limit contamination with host DNA during the extraction of parasite DNA, this membranous layer was systematically removed with sterile material. Each parasite was then placed in 99% ethanol. The DNA extraction was made on a selected number of P. laevis of each N. melanostomus, between one and three parasites in each sampled organ of each fish.
Individual parasite samples were incubated during 90 min at 55 °C in 700 µL of proteinase K 1 mg/mL (Euromedex, Souffelweyersheim, France) in 1% SDS, 500 mm NaCl, 10 mm TrisHCl, 50 mm EDTA. Then, the samples were mixed with 700 µL of phenol : chloroform : isoamyl alcohol (25 : 24 : 1; Euromedex). After centrifugation, the supernatant was collected and mixed with an equal volume of NaAc:ethanol and placed at −20 °C overnight for DNA precipitation. After centrifugation and washing with 70% EtOH, the dried pellets were suspended in 50 µL of Tris-EDTA with RNase solution (6 µL of RNAse, 594 µL of TE) at 55 °C for 60 min.
The first sequencing trials of cytochrome COI using universal primers (Folmer et al. Reference Folmer, Black, Hoeh, Lutz and Vrijenhoek1994) systematically led to the amplification of COI from the host N. melanostomus. After having tested several alternatives including blocking primers, we chose to design new primers with limited hybridization to N. melanostomus: the forward primer 5′-TGTATGTTTTGGTTGGTGTGTGAGG and the reverse primer 5′-GGTGCTGATACAAAATAGGTGAACC (synthetized by Eurofins Scientific, Luxembourg). The condition of PCR amplification followed Perrot-Minnot (Reference Perrot-Minnot2004): 1× reaction buffer, 1·5 mm of MgCl, 0·1 µ m of each primer, 0·1 mm of dNTP, 0·5 unit of GoTaqG2 polymerase (Promega, Madison, Wisconsin, USA) and 5 µL of DNA diluted at 50 ng/μL. The thermal cycling was programmed as 94 °C for 2 min, followed by 40 cycles with 20 s at 94 °C, 20 s at 50 °C and 50 s at 65 °C, with a final elongation of 5 min at 65 °C. PCR products were checked on 1·5% agarose gel and sequenced by Eurofins.
The DNA sequences and electropherograms were visualized, checked and aligned using MEGA6 (Tamura et al. Reference Tamura, Stecher, Peterson, Filipski and Kumar2013) and the haplotype reduction of the whole alignment was led on the FaBox web interface (http://users-birc.au.dk/biopv/php/fabox/). The haplotype network was generated using pegas R-package (Paradis, Reference Paradis2010) in R v.3.2 (R Core Team, 2017). In order to identify the lineages occurring in the Upper Rhine, we compared these sequences to recently released sequences of P. laevis from the Western Europe (Perrot-Minnot et al. Reference Perrot-Minnot, Špakulová, Wattier, Kotlík, Düşen, Aydoğdu and Tougardin press) (Genbank accessions MF563495–MF563527). For this purpose, the neighbour-joining phonetic tree was built in MEGA.
Results
Variation of the parasite assemblage of N. melanostomus along its invasion corridor
The bibliographical study along the corridor Rhine–Main–Danube revealed the presence of 37 different macroparasite species. Their presence and their prevalence vary depending on the site and on the river considered (Fig. 2). Three locations have been investigated in the Black Sea (Kvach and Skóra, Reference Kvach and Skóra2007; Kvach et al. Reference Kvach, Kornyychuk, Mierzejewska, Rubtsova, Yurakhno, Grabowska and Ovcharenko2014) revealing a total of 20 different macroparasites, five of them being found in the three locations. In the Danube, studies showed a decrease of parasite diversity compared with the Black Sea (Ondračková et al. Reference Ondračková, Dávidová, Pečinková, Blažek, Gelnar, Valová, Černý and Jurajda2005, Reference Ondračková, Francová, Dávidová, Polačik and Jurajda2010; Mühlegger, Reference Mühlegger2008; Košuthová et al. Reference Košuthová, Košco, Letková, Košuth and Manko2009; Francová et al. Reference Francová, Ondračková, Polačik and Jurajda2011). Eighteen parasite species were reported in eight locations, with a mean diversity of 4·8 species per location. Eight parasite species found in the Danube samples have already been reported from the Black Sea. Only one study focuses on the parasites of N. melanostomus in the Main River (Emde et al. Reference Emde, Kochmann, Kuhn, Plath and Klimpel2014). The two main parasites observed in the round goby were the nematode Raphidascaris acus (Bloch, 1779), already reported in the Black Sea, and P. laevis, already reported in the downstream stretch of the Danube but not in the Black Sea. In the Baltic Sea, five studies conducted on the parasite community of the round goby reported a total of 20 species, including five species that had never been reported in the other hydrosystems (Rolbiecki, Reference Rolbiecki2006; Kvach and Skóra, Reference Kvach and Skóra2007; Kvach and Winkler, Reference Kvach and Winkler2011; Kvach et al. Reference Kvach, Kornyychuk, Mierzejewska, Rubtsova, Yurakhno, Grabowska and Ovcharenko2014; Herlevi et al. Reference Herlevi, Puntila, Kuosa and Fagerholm2017). In the Rhine, three studies focus on three locations, but only in the lower (and middle) part of the river, downstream of the confluence of the Main–Danube canal (Emde et al. Reference Emde, Rueckert, Palm and Klimpel2012, Reference Emde, Kochmann, Kuhn, Plath and Klimpel2014; Ondračková et al. Reference Ondračková, Valová, Hudcová, Michálková, Šimková, Borcherding and Jurajda2015). One location has been surveyed in two periods of the year, autumn and spring (Ondračková et al. Reference Ondračková, Valová, Hudcová, Michálková, Šimková, Borcherding and Jurajda2015). Only six species have been reported, including one which had never been observed before in the Main, the Danube or the Black Sea, the nematode Paracuaria adunca (Emde et al. Reference Emde, Rueckert, Palm and Klimpel2012).
Fig. 2. Parasitic diversity along the Rhine–Main–Danube corridor. Each column represents one site except for the Rhine River, where the first two columns correspond to the same site surveyed at two different periods of the year (references in Fig. 1).
The present overview of parasite diversity along the Rhine–Main–Danube corridor therefore revealed three dominant parasite species in most sites along the corridor, with a high prevalence (Table 1): the acanthocephalan P. laevis, the nematode R. acus and the digenean trematode Diplostomum spp. In this study, we only observed P. laevis parasites in the 179 round goby collected in the Upper Rhine River.
Lack of parasite diversity in the Upper Rhine River
Out of the 179 round gobies sampled, 109 were parasitized, which represents a prevalence of 60·9%. We collected 1666 parasites in the body cavity, the gut, the liver or the gonads, all of which belonging to Pomphorhynchus species based on visual inspection. The parasites were enveloped in a membranous layer, at the larval stage called cystacanth. Once this layer was removed, the parasites presented diverse morphologies: invaginated proboscis or not, large and short body or small and thin body, proboscis with spines or not. Their identification according to their morphology was therefore difficult. The intensity, i.e. the mean number of parasite per infected fish was 15·3, with a minimum of one parasite (mean length of the fishes: 7·1 cm) and a maximum of 120 parasites (length of the fish: 15 cm). Fifty per cent of the goby harboured between one and five parasites, and only 8·3% carried more than 40 parasites. Out of the 18 barbels sampled, four were parasitized. The 21 parasites from barbels were all collected inside the gut. They were all at the adult stage, with their proboscis attached to the inner side of the intestine wall. Out of the 11 chub collected, only one was parasitized, with one parasite. This parasite was also at the adult stage, inside the gut of the fish. A total of 242 Pomphorhynchus parasite samples from 109 round gobies were identified using molecular method, plus 10 parasites from four barbels and one parasite from one chub. The molecular barcoding approach on the mDNA COI assigned all the 253 partial sequences to P. laevis.
Genetic diversity of the P. laevis in the Upper Rhine River
Using the specific primers designed for this study, we obtained partial COI sequences of 557 bp length from the 253 parasites. This dataset included 33 haplotypes (Genbank accessions: MF563495–MF563527) representing a haplotype diversity of 0·65 ± 0·032, despite a weak nucleotide diversity of 0·0018 ± 0·00 015. Eight of the 33 haplotypes represented the majority (87·4%) of the samples (Table 2), the remaining haplotypes being represented by only one or two samples. The haplotypes A and B represented 145 and 35 samples, respectively. The median-joining network showed a radial structure centred on the haplotype A (Fig. 3). This radial unimodal shape is accentuated by the poor representation of the other haplotypes. The highest divergence between two haplotypes is brought by seven mutations, and the distance between two neighbour haplotypes does not exceed two mutations. There was no evidence for structuring in the haplotype network driven by spatial location, nor by fish hosts, nor by location within the host (viscera or body cavity).
Fig. 3. Median-joining network comprising the 33 COI haplotypes of P omphorhynchus laevis from the Upper Rhine River. Each circle represents a haplotype and its size is proportional to the haplotype frequency. Haplotype A gathers 145 individuals, haplotype B gathers 35 individuals, other median circle dots gather from 12 to 2 individuals. Small white circles were found in only one individuals. Line lengths in the network corresponds to the number of mutational changes between haplotypes, and grey lines represent other equivalent lops between close haplotypes. Black dots represent haplotypes missing in the study sampling.
Table 2. Distribution of the eight most common haplotypes of P omphorhynchus laevis recorded in the Upper Rhine River
Phylogenetic position of the Upper Rhine populations in the phylogeography of P. laevis
In order to establish the native or exotic status of the lineage of P. laevis found in the Upper Rhine, we placed these haplotypes in a phylogenetic tree comprising published sequences of P. laevis from Europe (Fig. 4). All samples from the Upper Rhine are clustering with haplotypes from the Danube and the Vistula rivers. They are therefore gathered in a lineage distinct from the Western Europe lineage comprising samples from France (Rhone and Loire drainages, and Meuse River; Fig. 4).
Fig. 4. Neighbour-joining phenetic tree of the lineage of P omphorhynchus laevis in Europe, based on 45 European haplotypes from a previous study, and the 33 haplotypes identified in the present study. The numbers correspond to bootstrap values supported by each node. The phylogenetic tree has been built using two acanthocephalans species close to P. laevis as external groups, Pomphorhynchus tereticollis (n = 5) and Echinorhynchus truttae.
Discussion
In the Upper Rhine, the invasive N. melanostomus is parasitized only by the acanthocephalan P. laevis. However, N. melanostomus is infected by a diversity of macroparasites along the Rhine–Main–Danube corridor (Molnar, Reference Molnar2006; Emde et al. Reference Emde, Rueckert, Palm and Klimpel2012). Some macroparasites are specific to an area [e.g. the nematode Cosmocephalus obvelatus (Creplin, Reference Creplin1825), in the Baltic Sea (Kvach and Winkler, Reference Kvach and Winkler2011; Kvach et al. Reference Kvach, Kornyychuk, Mierzejewska, Rubtsova, Yurakhno, Grabowska and Ovcharenko2014), while others are present at several locations along the corridor (e.g. the nematode R. acus)]. Torchin et al. (Reference Torchin, Lafferty, Dobson, McKenzie and Kuris2003) proposed that at the beginning of invasion process, the newly introduced host loses a part of its parasite community. The differences in the parasite community between the native and the invasive populations would vanish after several years, once the invasive population is well settled (Gendron et al. Reference Gendron, Marcogliese and Thomas2012). If the parasite community of N. melanostomus shows a large diversity along its invasive pathway, it is yet reduced to one species in the Upper Rhine River. Indeed, we did not find in the Upper Rhine the parasites associated to N. melanostomus in its native range (Kvach and Skóra, Reference Kvach and Skóra2007; Kvach et al. Reference Kvach, Kornyychuk, Mierzejewska, Rubtsova, Yurakhno, Grabowska and Ovcharenko2014). This loss of parasites is in agreement with the ‘enemy release hypothesis’ (Williamson, Reference Williamson1996), which states that exotic species arrive almost without any parasite. In addition, the mDNA COI sequences of P. laevis from the Upper Rhine clearly fit within a phylogenetic lineage described in the Danube and the Vistula rivers (Perrot-Minnot et al. Reference Perrot-Minnot, Špakulová, Wattier, Kotlík, Düşen, Aydoğdu and Tougardin press). This lineage could have been transported by N. melanostomus from the Danube to the Rhine, as a ‘hitchhiker’ parasite. Associated with the ‘enemy release’ phenomenon (i.e. the loss of initial parasites), the Danubian P. laevis found in N. melonostomus could testify a previous spillback event during the passing of the round goby along the Danube River. Our additional fishing of two well-known definitive hosts of P. laevis, the barbel, B. barbus and the chub, L. cephalus (Sures and Siddall, Reference Sures and Siddall1999; Thielen et al. Reference Thielen, Zimmermann, Baska, Taraschewski and Sures2004) demonstrate that P. laevis from the Danube uses these native fish to complete its life cycle in the Upper Rhine. This is in agreement with the fish hosts from which this lineage was recorded in the Danube (barbels and gobies, Perrot-Minnot et al. Reference Perrot-Minnot, Špakulová, Wattier, Kotlík, Düşen, Aydoğdu and Tougardin press). This result could testify to a spillover event with the transfer of P. laevis from an exotic host (N. melanostomus) to a native one. Concomitantly with our study, the introduction of P. laevis in the Rhine River (Germany, Switzerland) from the Ponto-Caspian region has been recently reported, although the precise genetic lineage of P. laevis was not identified (Hohenadler et al. Reference Hohenadler, Nachev, Thielen, Taraschewski, Grabner and Sures2017). Interestingly, the introduction of Ponto-Caspian P. laevis was accompanied by the displacement of Pomphorhynchus tereticollis within about a decade, as evidenced based on historical sampling of Pomphorhynchus from eels (Hohenadler et al. Reference Hohenadler, Nachev, Thielen, Taraschewski, Grabner and Sures2017). From our data, we cannot conclude that N. melanostomus is the only dispersal vector of P. laevis, and other intermediate hosts could have played a role in this dispersal. For instance, a Ponto-Caspian amphipod, Dikerogammarus villosus (Sowinsky, 1894), is an intermediate host for P. laevis (Rewicz et al. Reference Rewicz, Grabowski, MacNeil and Bacela-Spychalska2014), and it has become an invasive species westward, including the Upper Rhine (Bollache et al. Reference Bollache, Devin, Wattier, Chovet, Beisel, Moreteau and Rigaud2004). It is therefore a likely candidate for P. laevis introduction (Hohenadler et al. Reference Hohenadler, Nachev, Thielen, Taraschewski, Grabner and Sures2017), together with other species from the Ponto-Caspian region such as the gobies. In addition, other transfers of parasites, led by spillover and spillback phenomena, have probably occurred all along the corridor, resulting in a large diversity of parasite communities of the round goby.
With this study, we revealed some gaps in the knowledge of the life cycle of P. laevis in the Upper Rhine River. More specifically, most parasites found in the round goby were non-mature and found outside the intestinal tract (where sexual reproduction takes place), which indicates that the round goby is probably not a definitive host for P. laevis. The significance of this host as a paratenic host or as a dead-end for P. laevis still needs to be established. In the former case, predatory fish of the goby could act as definitive hosts and allow the parasite to resume its cycle. For instance, some large barbels could feed on N. melanostomus (Emde et al. Reference Emde, Rueckert, Palm and Klimpel2012). This hypothesis awaits further investigations, for instance by testing the viability of the cystacanths found in the body cavity of the round goby (Médoc et al. Reference Médoc, Rigaud, Motreuil, Perrot-Minnot and Bollache2011). According to Kennedy (Reference Kennedy2006), N. melanostomus could be a paratenic host for P. laevis only if the cystacanths can resume their development once they are transferred to a definitive host. In this case, parateny could have a positive effect on the life cycle of P. laevis: the concentration of parasites in the paratenic host allows a delayed and massive contamination of the definitive host (Kennedy, Reference Kennedy2006). If the round goby is a dead-end host, the P. laevis in the round goby cannot resume their cycle at all, which leads to a decrease of the parasite population. To test this hypothesis, the prevalence and intensity of the parasite in several hosts over time should be studied, in order to detect a dilution effect in the P. laevis community (Emde et al. Reference Emde, Rueckert, Palm and Klimpel2012). The ecological consequences in the Upper Rhine River could be multiple, for instance a disappearance of the parasite and a lack of regulation of the round goby population, this latter already constituting a worrying percentage of fish abundance in the Upper Rhine.
Acknowledgements
This study was undertaken under the auspices of Electricité de France (EDF). We thank Yann Eckenschwiller, Sophie Louis, Victorien Tallet, Coralie Tarrene and Camille Berolo from the ‘Fédération de pêche 68’ for their contributions to data acquisition (electrofishing).
Financial Support
This research received no specific grant from any funding agency, commercial or not-for-profit sectors.
