Introduction

Alien species cause many ecological and social problems, such as extinction of native species, degradation of agricultural land, and threats to people during outdoor activities (Pyšek et al., 2020; Cuthbert et al., 2022). In particular, freshwater environments, which are under severe threat from human activities (Vörösmarty et al., 2010), are highly affected by invasive species (Dextrase & Mandrak, 2006; Dudgeon et al., 2006; Reid et al., 2019). For example, among the 100 of the World's Worst Invasive Alien Species, about 15% are species associated with freshwater, such as Cyprinus carpio Linnaeus, 1758 and Dreissena polymorpha (Pallas, 1771) (ISSG, 2022). Considering the potential impact of alien species, it is therefore necessary to clarify the process of introduction and spread of alien species in freshwaters in terms of conservation.

Here we focused on the invasion of Gyraulus chinensis (Dunker, 1848) in Europe, which is a freshwater snail belonging to the family Planorbidae (Gastropoda: Hygrophila). This planorbid snail is native to East-Southeast Asia, although its exact range is unclear due to taxonomic inadequacies (Meier-Brook, 1983; Saito et al., 2018b, 2020). It differs from the other European representatives of the genus by having distinct black dots on a distant part of the mantle (Brown et al. 1998; Glöer, 2019; Rowson et al., 2021). The earliest documented record in Europe is from Camargue, France, in 1973 (Brown et al., 1998). Subsequently, this snail is reported to have spread over a large area in Europe, and there are also records from Africa, Australia, Middle East (Israel), and Caribbean (Martinique) as an alien species (Brown et al., 1998; Pointier, 2008; Appleton & Miranda, 2015; Mienis et al., 2015; Ponder et al., 2022). On the other hand, despite its global distribution, little is known about the nature of G. chinensis as an invasive species. At least in its native range, the snail serves as an intermediate host for parasites that are harmful to human health (Lo, 1995; Toledo et al., 1998). Although G. chinensis was a part of the community that has replaced native fauna in the Arno River in Italy (Haubrock et al., 2021), its direct impact has not been considered. Given its widespread invasion in Europe, G. chinensis may pose an existing and/or potential threat to snails of the same genus and family, which include endangered species (IUCN, 2022). In other representatives of Hygrophila, negative impacts on other native species of the same family (Lymnaeidae; Pseudosuccinea columella (Say, 1817); Vignoles et al., 2018) and the potential for introgressive hybridisation with closely related species have been reported (e.g. Ferrissia californica (Rowell, 1863); Saito et al., 2018a).

In this study, we address the clarification of the introduction and spread of G. chinensis in Europe using literature information and molecular approaches. The synthesis of literature information will trace the process of invasion in Europe and provide a better understanding of the spread process. In addition, the molecular approach will help validate the invasions of members of the genus Gyraulus, which are sometimes difficult to identify from shell morphology alone and cause taxonomic confusion (Meier-Brook, 1983; Lorencová et al., 2021), and provide insights into the origin of the introduction. By combining these approaches, this study aims to reconstruct the history of the invasion of G. chinensis. This reconstruction will not only address the further spread and potential impacts of this species, but will also expand our knowledge of the spread of alien freshwater invertebrates by comparing the history with the invasion processes of other species.

Methods

Literature survey

Some of the records of Gyraulus chinensis in Europe were documented in local malacological journals and illustrated books, which are not collected in online databases, such as Web of Science and Google Scholar. Therefore, in order to obtain as many records as possible in this study, we manually searched the scientific literature and integrated it into a list after validation (Online Resource 1). Water bodies that exist outdoors and have no artificial temperature control were treated as outdoor environments, while other water bodies, such as aquaria and heated ponds, were treated indoors. However, it cannot be completely ruled out that the sites where G. chinensis was detected were under the influence of hot water runoff, as water temperature was not documented in most surveys. In addition, some records that were not adequately documented to be outdoors were treated as possibly outdoor habitat (labelled “possibly outdoors” in Online Resource 1). The observed locations were narrowed down as much as possible using all available information. We then mapped the records using QGIS 3.16 (QGIS Development Team, 2020). In cases where it was not possible to determine the year of record, the latest possible year was used for mapping for simplicity.

Taxon sampling and DNA methods

We assembled all potential specimens of Gyraulus chinensis from our samples collected in the course of earlier investigations in outdoor freshwater environments in Albania and Italy, and indoor habitat in the Czech Republic (Fig. 1; Tables 1; Online Resource 2). To select these specimens, we followed morphological characters of G. chinensis reported in the literature (Brown et al. 1998; Glöer, 2019; Rowson et al., 2021). Some of the Italian specimens were collected in the 1990s and early 2000s. In addition, the Albanian populations were the first records. For comparison, we also collected and analysed G. chinensis specimens from native Asian regions, including Hong Kong, near the type locality of G. chinensis, and the other six Gyraulus species (Tables 1; Online Resource 2). Some of these samples were used in two previous studies, and fixation and extraction were performed using the methods described in the respective study (Saito et al., 2018a, b; Lorencová et al., 2021). Other newly collected specimens were fixed in 96% ethanol or dried at ambient temperature and humidity. Total DNA was extracted using the E.Z.N.A. Mollusc DNA Kit (Omega BioTek, GA, USA) according to the manufacturer's instructions.

Fig. 1
figure 1

Map of Gyraulus chinensis distribution in Europe, based on literature review and our samples (Online Resource 1). Symbols and country colours are used to indicate outdoor vs indoor records. In addition, the first year of outdoor record is noted

Full size image
Table 1 Gyraulus samples used for mitochondrial and nuclear phylogenies
Full size table

We then amplified two mitochondrial regions, cytochrome b (CytB) and cytochrome c oxidase subunit I (CO1), and two nuclear genes, internal transcribed spacers 1 (ITS1) and 2 (ITS2) of the rRNA gene cassette. The primers and protocols for amplification were the same as in Lorencová et al. (2021). Each amplified product was purified with ExoSAP (Affymetrix, CA, USA) and sequenced at the OMICS Core facility, BIOCEV (Czech Republic) using the BigDye® Terminator v3.1 Cycle Sequencing Kit (Applied Biosystems, CA, USA) or SEQme s.r.o. (Czech Republic) using the Gerbera Sequencing Kit v3.1 (SEQme, Czech Republic). After removing the primer sequences, all sequences were deposited in GenBank (Accession Nos OQ121085–OQ121110, OQ127675–OQ127702, OQ133368–OQ133396, and OQ133368–OQ133396).

Molecular phylogenetics and population genetics

Phylogenetic analyses for molecular identification of the introduced Gyraulus chinensis, which taxonomy is insufficient, were performed using two mitochondrial (CytB and CO1) and two nuclear regions (ITS1 and ITS2). Sequences of six Gyraulus species from GenBank, validated in Lorencová et al. (2021), were used for comparison. The resulting CO1 + CytB dataset included 39 individuals with 78 sequences (25 of which were from GenBank), 33 for ITS1 (10 from GenBank) and 39 for ITS2 (10 from GenBank) (Table 1). Phylogenetic analyses had to be conducted separately for ITS1 and ITS2 regions because we were not able to amplify both markers for some specimens. In addition, with reference to Lorencová et al. (2021), Planorbis planorbis (Linnaeus, 1758) was used as an obvious outgroup. Each dataset was aligned using MUSCLE (Edgar, 2004) implemented in MEGA 7 (Kumar et al., 2016) with default settings. To reduce uncertainty for phylogenetic analyses, gaps in the alignments of CO1, ITS1 and ITS2 were eliminated using trimAl 1.2 (Capella-Gutiérrez et al., 2009). Finally, identical haplotypes were stacked with FaBox 1.61 (Villesen, 2007).

Maximum likelihood analyses (ML) were performed using W- IQ-TREE (Nguyen et al., 2015; Trifinopoulos et al., 2016) with the ‘-spp’ option (same set of branch lengths in all partitions, allowing each partition to evolve at its own rate). Evolutionary models and partitions for ML analyses were selected using PartitionFinder 2 (Lanfear et al., 2016) based on AICc (Table 2). The ML tree topologies were evaluated using ultrafast bootstrapping (UFB; Hoang et al., 2018) with 10,000 replicates, Shimodaira-Hasegawa approximate likelihood ratio test (SH-aLRT; Kishino & Hasegawa, 1989; Shimodaira & Hasegawa, 1999) with 10,000 replicates, and a Bayes-like transformation of aLRT (aBayes; Anisimova et al., 2011). Bayesian inference analyses (BI) were performed with MrBayes 3.2.7a (Ronquist & Huelsenbeck, 2003) using four simultaneous chains and sampling trees every 1000 generations. The ‘usebeagle’ option was turned off to avoid unexpected errors. Four chains were run for 1 million generations in all datasets. After using Tracer v. 1.7 (Rambaut et al., 2018) to confirm that the effective sample size (ESS) of all parameters in each dataset was greater than 200, we discarded the first 101 trees as burn-in after checking for convergence. The remaining trees were then used to estimate each tree with maximum clade reliability and to evaluate each tree topology with Bayesian posterior probability (BPP).

Table 2 Appropriate evolutionary model and partitioning estimated for each phylogeny in this study
Full size table

To reveal the origin(s) and invasion history of G. chinensis in Europe, we explored the genetic structure of the G. chinensis populations using the CO1 region, which has the largest number of relevant sequences in GenBank. However, the taxonomy of the genus Gyraulus is not fully clarified, and it is difficult to distinguish G. chinensis a priori based on the GenBank registry name. Therefore, after obtaining all Gyraulus sequences from GenBank, we performed preliminary phylogenetic analyses and selected sequences that fell into the monophyletic clade that was sister to G. parvus (Say, 1817). This clade also contained sequences previously considered to represent G. chinensis and also those from near to the type locality. Consequently, 164 haplotypes from 284 sequences, including sequences from Hong Kong, close to the type locality of G. chinensis, were used for the subsequent analysis (Online Resource 2). In addition, two sequences of G. parvus were selected as outgroups. Phylogenetic analysis was then performed using this CO1 dataset. After alignment, the sequences were stacked by haplotype, and then the ML and BI analyses were performed in the same flow as the previous analyses (the evolutionary models in these analyses are listed in Table 2). However, the number of generations in the BI analysis was 3 million and the burn-in was the first 301 trees. Finally, we presented a haplotype network using the monophyly in the CO1 phylogeny for G. chinensis and included all haplotypes detected in Europe. The haplotype network was constructed with the TCS algorithm (Clement et al., 2000) using PopART (Leigh & Bryant, 2015).

Results

Distribution of Gyraulus chinensis

Based on the collected 59 literature records and our survey, Gyraulus chinensis has been reported outdoors from Albania, Austria, France, Germany, Greece, Italy, Montenegro, the Netherlands, Portugal and Spain, and indoors from Austria, the Czech Republic, Ireland, Germany, the Netherlands, Norway, Sweden and the United Kingdom (Fig. 1; Online Resource 1). In Belgium, the occurrence was highly speculated (Van Goethem, 1992; Sablon & Vercauteren, 2011), but no evidence was found in the literature during our study. The south of Wonseradeel in the Netherlands was the northern limit of occurrence of G. chinensis in the outdoor environments.

After the first outdoor record from France (1973), the species was also reported outdoors from Italy, Portugal and Spain by the 1980s. Findings in other European regions are more recent. In south-eastern Europe, the species was first reported outdoors from Greece in 2003, from Montenegro in 2013 and from Albania in 2018. In Central Europe, the oldest outdoor records are from the Netherlands (1976). In Austria and Germany, possible occurrences in the outdoor environments were reported already in 1993 and 1998, but reliable outdoor records come from 2001 and 2009. The oldest records from indoor habitats are from Ireland and the United Kingdom (1974), and from Austria and the Netherlands as early as the 1990s. In other countries, it has been recorded indoors since the 2000s. In France, Spain, Italy, and south-eastern Europe, there are no documented records in an explicitly indoor habitat.

Molecular identification

Our three (CO1 + CytB, ITS1 and ITS2) phylogenies showed that most Gyraulus species formed monophylies that were well supported (> 90 and > 0.95) by at least two evaluation values (Fig. 2; Online Resource 3). Gyraulus chinensis in Europe has nine haplotypes in CO1 + CytB, eleven haplotypes in ITS1 and five haplotypes in ITS2, although some European samples were not sequenced. Samples from Albania, Italy, and Japan shared one haplotype in ITS2. In three phylogenies, the European G. chinensis haplotypes were included in the well-supported (CO1 + CytB: UFB = 99/ SH -aLRT = 95/aBayes = 1.00/BPP = 1.00; ITS1: 100/100/1.00/1.00; ITS2: 91/96/1.00/1.00) monophyly of G. chinensis with samples from Japan, Taiwan, and Hong Kong (Fig. 2; Online Resource 3). This is an important outcome as Hong Kong sequences represent data from close to the type locality.

Fig. 2
figure 2

Maximum likelihood phylogenies of the genus Gyraulus inferred from nuclear internal transcribed spacer 2 (ITS2), and mitochondrial cytochrome b (Cytb) and cytochrome c oxidase subunit 1 (CO1) DNAs. The numbers at each node indicate support values by four different methods. The labels on each haplotype represent the IDs of the included samples (Table 1). The bars next to the labels indicate the species based on a previous molecular phylogeny (Lorencová et al. 2021). The colours of each ID represent the countries or regions collected

Full size image

Population genetic structure

Our CO1 phylogeny revealed the phylogenetic relationship of Gyraulus chinensis and closely related lineages (Fig. 3). Many basal nodes were not supported, while most terminal nodes were well supported and often consisted of geographically related haplotypes. All European G. chinensis were included in one relatively well-supported monophyly (91/97/1.00/-), and in addition, the remaining haplotypes from Europe, with the exception of one haplotype from Italy, were included in another well-supported monophyly (100/100/1.00/-).

Fig. 3
figure 3

Maximum likelihood phylogeny of Gyraulus chinensis and closely related lineages inferred from mitochondrial cytochrome c oxidase subunit 1 (CO1) DNA. The numbers at each node indicate the support values by four different methods (they are not shown at the terminal nodes due to visibility). The label on each operational unit represents the haplotype numbers (Online Resource 2), and for European samples the country where the haplotype was collected. The colours of each label represent the countries or geographical areas, which are shown on the map at the top left. The geographical areas have been defined with reference to biogeographical boundaries for convenience. This map was created from the Global Self-consistent, Hierarchical, High-resolution Geography Database, version 2.3.5 (Wessel & Smith 1996, 2016), using QGIS version 3.16 (QGIS Development Team, 2020). The sites were identified based on the publications and GenBank registry information of each sequence (Online Resource 2)

Full size image

The haplotype network, illustrated by the monophyly with all nine European haplotypes, showed the geographical relationship between the individual haplotypes (Fig. 4; Table 2). Haplotype No. 1 included eight sequences from Albania and one sequence from the Ryukyu Islands, Japan, and Taiwan. All other European sequences were included in eight haplotypes that did not include sequences from outside Europe. Each haplotype, with the exception of haplotype No. 7 from Italy, had the most closely related haplotypes containing sequence(s) from the southern part of East Asia, namely Taiwan, Hong Kong, and the Ryukyu Islands of Japan. In addition, haplotype No. 7 consisted of two Italian sequences from Arenzano and Venaria, and the most closely related haplotype was from Kyushu Island in Japan.

Fig. 4
figure 4

The TCS network of mitochondrial CO1 haplotypes for the monophyly from the phylogeny in Fig. 3. The numbers with circles represent the nominal haplotype numbers (Online Resource 2), and each circle size represents the haplotype frequency. The bar on the branch indicates a mutation. Each colour represents countries or geographical areas shown on the map in the upper right part. This map was created from the Global Self-consistent, Hierarchical, High-resolution Geography Database, version 2.3.5 (Wessel & Smith, 1996, 2016), using QGIS version 3.16 (QGIS Development Team, 2020)

Full size image

Discussion

Origins of Gyraulus chinensis invasion

This is the first study to demonstrate the introduction of Gyraulus chinensis into Europe based on extensive molecular phylogenies, as previously this was only supported based on morphological identification (Figs. 2; Online Resource 3). Although there have been some previous sequences from Montenegro (von Oheimb et al., 2013; Saito et al., 2022) and British aquaria (Rowson et al. 2021) assumed to represent G. chinensis, they were identified as Gyraulus sp. Therefore, our study provides the first indisputable evidence for G. chinensis occurrence in Europe based on sequences from several European countries. Moreover, our CO1 phylogeny and haplotype network revealed a strong genetic relationship between G. chinensis from Europe and from Hong Kong (New Territories), close to the type locality of G. chinensis (Hong Kong Island; Dunker, 1848; Figs. 3, 4). Therefore, despite the confusion in the taxonomy of G. chinensis, the European G. chinensis collected in this study should strictly be G. chinensis. However, because our sampling may not have captured the entire European diversity of G. chinensis populations, it cannot be ruled out that there are more lineages of G. chinensis sensu lato introduced into Europe than those already detected. Therefore, further sampling and taxonomic studies of G. chinensis and related species are needed to clarify this issue.

Based on our molecular results, it is difficult to fully determine the origin of introduced G. chinensis in Europe, because the GenBank sequences of G. chinensis and related lineages, although relatively extensive, could not cover all native distributions, and CO1 sequences can be affected by introgressive hybridisation and ancestral polymorphisms (Ballard, 2000; Ballard & Whitlock, 2004). Nevertheless, the European populations may have originated in the south-eastern part of the Eurasian continent, particularly on Kyushu Island, the Ryukyu Islands, Taiwan, and Hong Kong. At a minimum, we can conclude that European haplotypes have little relationship with haplotypes from South Asia, Central Asia, Northeast Asia and North Asia.

The haplotypes of G. chinensis in Europe were diverse, with haplotype No. 7 from Italy being genetically quite distant. Moreover, the localities of the first outdoor records are geographically distant and dated over a long period of time (Fig. 1; Online Resource 1). These findings suggest the possibility that introduction occurred from more than one source or, less likely, from a source that exhibited high haplotype diversity. Most records of G. chinensis in the outdoor environments are from southern Europe, and date from the early stages of the invasion, while records from the north are relatively recent and rare, as suggested by our field and literature surveys (Fig. 1; Online Resource 1). This suggests that temperature differences between the native range and introduced areas were important in maintaining populations during the early stages of the invasion. In general, temperature can be a major factor determining the success or failure of an alien species introduction and establishment (Hayes & Barry, 2008; Petitpierre et al., 2012; Guisan et al., 2014), and it has also been suggested that niche conservatism for temperature may have influenced the formation of the distribution of Hygrophila snails (Saito et al., 2021).

Mechanisms of introduction

Although the specific mechanisms of introduction remain unclear, the most likely hypothesis is invasion via the aquarium trade. Invasion of freshwater organisms via the aquarium trade is often considered an influential mechanism (e.g. Papavlasopoulou et al., 2014; Nunes et al., 2015; Ng et al., 2016; Vinarski, 2017). In particular, small invertebrates, such as Hygrophila snails, are often imported and introduced as unintended hitchhikers of target species in the trade (Duggan, 2010; Patoka et al., 2016, 2017; Duggan & Pullan, 2017). In our case, Gyraulus chinensis is commonly found in indoor aquaria in botanical gardens (Fig. 1; Online Resource 1), and the Czech haplotypes from indoor aquaria are genetically close to the Italian and Albanian haplotypes in the outdoor environments. In addition, CO1 sequences from UK aquaria (but not published) have been reported to be identical to the sequences from Taiwan (Rowson et al., 2021). These records suggest that aquaria are important sources of G. chinensis invasion in Europe. This hypothesis is also supported by the presumed origin(s) of European G. chinensis on the south-eastern Eurasian continent, as European countries mainly import freshwater organisms from the tropics, such as Singapore and Hong Kong (Madsen & Frandsen, 1989; Brunel, 2009). In addition, the genetic diversity of G. chinensis in Europe could be explained by the fact that the introduced populations in indoor aquaria across Europe might represent a source of potentially high genetic diversity. On the other hand, there is no evidence to rule out other mechanisms of introduction and spread. In particular, due to the wide and dense distribution of G. chinensis, it is difficult to explain all dispersal in Europe by the aquarium trade alone. Other mechanisms such as release of contaminants by fisheries, transport by humans, external and internal transport by birds, and natural dispersal via water currents should be examined in further studies. Based on the literature records (Fig. 1; Online Resource 1), natural dispersal in particular could be the major contributor to dispersal, as many records in Italy and Spain have been documented within a short period of time.

Distribution in Europe

Currently, the records of introduced Gyraulus chinensis populations are concentrated in southern Europe (Fig. 1). There are also records in the northern regions, i.e. the Netherlands, Austria, and Germany, but findings in the outdoor environments seem to be much rarer than in southern Europe (Neckheim, 2017). Moreover, most of these northern records are from urban areas and even in the outdoor environments, it cannot be excluded that they inhabit water bodies affected by hot water discharge. Indeed, in the Lower Rhine, G. chinensis only occurs in water bodies with hot water discharge (van der Velde et al., 2002). As mentioned above, this limitation of invasion might be caused by climatic differences between the native range and Europe, except for the southern parts. However, climate change and adaptation of alien species may allow them to spread to new environments (Sax et al., 2007; Walther et al., 2009; Chapman et al., 2017; Hulme, 2017). Furthermore, as G. chinensis is observed in many indoor aquaria, there is a large invasion debt (i.e. the time lag of introduction; Rouget et al., 2016) in many areas of Europe. Therefore, its spread may continue in the future under the climate warming scenario. On the other hand, the consequences of G. chinensis invasion have hardly been assessed and documented so far. The impacts of non-native species vary depending on the regions and ecosystems (Uchida et al., 2016) and may even change due to their adaptive evolution (Prentis et al., 2008; Whitney & Gabler, 2008). Therefore, it will be necessary to further assess and carefully monitor the impacts.