Integrative taxonomy reveals the occurrence of the Asian freshwater snail Sinotaia cf. quadrata in inland waters of SW Europe

A multisource approach to assessing freshwater biodiversity reveals the first occurrence of the Asian freshwater snail Sinotaia cf. quadrata (Gastropoda, Viviparidae) in inland waters of the northern Iberian Peninsula (Spain) and southwestern France. We present a detailed characterization, combining morphological traits and molecular tools for a genetic identification, and illustration of the species and its ultrastructure, as well as key information regarding its reproductive biology and ecology at the new locations. We discuss the most plausible introduction pathways and vectors, its potential invasiveness and the subsequent impacts that it may generate in the receiving localities. Our preliminary results raise concerns about the threat of Sinotaia cf. quadrata to human health and ecosystem dynamics, since this species acts as an intermediate host for several parasite species. Finally, this study confirms once again the usefulness of an integrative taxonomical approach for shedding light on hidden biodiversity and the invasion of freshwater ecosystems.


Introduction
Non-indigenous species (NIS) that become invasive can cause strong ecological impacts in receiving environments worldwide, including irreversible habitat alterations and damage to ecosystem functioning (Carlton and Geller 1993;Justine et al. 2014;Tricarico et al. 2016). Invasive species are an important cause of biodiversity loss and ecosystem homogenization through competition (Cowie 2005;Frederico et al. 2019) and alteration of physical or chemical properties of habitats (Katsanevakis et al. 2014a, b). Those consequences lead to loss of ecosystem services, impacts on food provision, tourism and lifecycle maintenance, often affecting human health and economic activities (Katsanevakis et al. 2014a, b;Borrell et al. 2017;Tricarico et al. 2017). During recent decades there has been an alarming increase in the number of non-indigenous species establishing populations worldwide (Simberloff 2014). Therefore, the early detection of these non-native species is crucial for proper management and for preventing or mitigating potential threats (Arias and Torralba-Burrial 2014;Pochon et al. 2015;Tricarico et al. 2017). In this regard, molecular tools have been proven useful for species identifications, early detection and effective ecosystem monitoring when facing invaders (Ardura et al. 2017).
Freshwater ecosystems are particularly vulnerable to non-native species introductions, as they are usually closely related to human activities, and most aquatic organisms present a high dispersal rate . European freshwater environments host 756 non-native species (Nunes et al. 2015), of which 30 are molluscs (Cianfanelli et al. 2016). Aquaculture, sport fishing and aquarium and pet trade are among the most remarkable introduction pathways for non-native species in freshwater habitats (Mazza et al. 2014). To date, 18 introduced mollusc species have been reported in the Iberian freshwater environments and five of them, belonging to the Gastropoda, have developed an invasive behavior in the receiving ecosystems (i.e. Physella acuta (Draparnaud, 1805), Melanoides tuberculata (Müller, 1774), Potamopyrgus antipodarum (Gray, 1843), Pomacea canaliculata (Lamarck, 1822) and P. insularum (d'Orbigni, 1835)) (García-Berthou et al. 2007).
Members of the family Viviparidae are operculate freshwater snails with high intraspecific variability, occurring naturally in temperate and tropical regions worldwide (Ovando and Cuezzo 2012;Hirano et al. 2019;Stelbrink et al. 2020). Two species of viviparids have been reported as invasive in North America: Heterogen japonica (Martens, 1861) and Cipangopaludina chinensis (Gray in Griffith and Pidgeon, 1833), with devastating effects on the distribution and abundance of native snails (Johnson et al. 2009;Solomon et al. 2010). Furthermore, C. chinensis has been also reported in Europe, from the Netherlands (Soes et al. 2011). Another species of the family, Sinotaia cf. quadrata (Benson, 1842), has been recorded as invasive in Argentina, South America (Ovando and Cuezzo 2012;Ferreira et al. 2017) and more recently, from Central Italy (Cianfanelli et al. 2017).
During the course of a series of eco-monitoring programs aimed at investigating the freshwater fauna of rivers and ponds from Asturias (northern Spain) and Cazaux (southwest France), several specimens conchologically consistent with the diagnosis of the viviparid genus Sinotaia were found. These constitute the first record of this genus from the Iberian Peninsula. In this paper we present a detailed characterization, combining morphological traits and molecular tools for a genetic identification, and illustration of the species, as well as some notes on its reproductive biology and ecology at the new location. Moreover, we discuss the most plausible introduction pathways and vectors, its current status (established or invasive) and the potential impacts that it may generate in the receiving localities, including associated biosanitary risks.

Specimen collection, morphological analysis and taxonomic procedures
Studied specimens (6 females, 1 male and 2 juveniles) were collected during May 2017 and June 2018 in the Nora River (43°22′N; 5°47′W, Colloto, Asturias, northern Spain). Another specimen was collected during August 2017 in the Cazaux Lake (44°29′N; 1°9′W, Gironde/Landes, Nouvelle Aquitaine, southwestern France). All collected specimens were prepared for preservation in situ and subsequently fixed and stored in 70% ethanol. The specimens were deposited at the Zoological Collection of the Department of Organisms and System (BOS) of the University of Oviedo (https://bos.uniovi.es/). Specimens were examined under a dissecting stereomicroscope. Temporary glycerol slides of radulae and other internal structures were examined under a compound light microscope. Selected specimens were dissected to observe the internal anatomy and, in females, the developing eggs and embryos. Photographs were taken with a Canon EOS 1200D Digital SLR Camera with EF-S 18-55 mm f/3.5-5.6 III Lens; photomicrographs were taken with a Nikon Digital Sight DS-L1 camera mounted on a Nikon SMZ-U stereomicroscope. Selected juvenile specimens were prepared for scanning electron microscopy (SEM) for the study of their general morphology. Specimens were dehydrated in an ascending series of ethanol, critical point dried using acetone as the transition liquid, mounted on aluminum stubs and sputter coated with gold. Samples were then imaged using a JEOL 6610 LV scanning electron microscope. Line drawings were made with the aid of a camera lucida and digital photography.

DNA extraction, PCR amplification and sequencing
DNA was extracted from 20-50 mg of ethanol-preserved tissues from the foot of each of the eight studied specimens, using E.Z.N.A Mollusc DNA Kit (Omega Bio-tek, Omega Bio-Tek, Norcross, GA, USA) and following the manufacturer's protocol. The success of the extraction was checked trough a horizontal electrophoresis (1% agarose gel), and DNA samples were stored at −20 °C. The mitochondrial cytochrome c oxidase subunit I (COI) gene was amplified by means of polymerase chain reaction (PCR) in a total volume of 25 μl, using the universal primers LCO1490 and HCO2198 (Folmer et al. 1994). The reaction mixture contained 2.5 μL template DNA, 2.5 μM MgCl 2 , 1.25 μM deoxyribonucleotides triphosphate, 0.5 μM of each primer, 0.2 U Taq polymerase and the appropriate buffer at 1x final concentration. PCR conditions used were an initial denaturation step of 94 °C for 4 min, then 45 cycles of 94 °C for 30 s, 48 °C for 1 min, 72 °C for 2 min, and finally an extension of 72 °C for 7 min and 20 °C for 1 min.

Genetic analysis
The consensus sequences obtained were edited and aligned using ClustalW in the freeware BIOEDIT (Hall 1999). Although all the studied individuals were sequenced, sequences obtained for two of the eight studied individuals were shorter than the others, and therefore they were excluded for the analysis. After alignment and corrections, preliminary genetic species identification was attempted using nBlast to search in BOLD (http://boldsystems.org/index.php/IDS_OpenIdEngine) and GenBank (https://blast.ncbi.nlm.nih.gov/Blast.cgi) databases. Phylogenetic analysis was conducted using 61 COI sequences downloaded from Genbank from a wide range of Viviparidae. The taxon range includes the genus Sinotaia Haas, 1939 (Bellamyinae clade A from Stelbrink et al. (2020)) (Supplementary material Table S1). Other sequences already included in JCR published papers were included in the analyses (Schultheiß et al. 2014;Hirano et al. 2015Hirano et al. , 2019Gu et al. 2015a, b;Cianfanelli et al. 2017 andStelbrink et al. 2020) (Table S1). The sequences were analyzed using the MEGA 7 software ). The Model Test software included in Mega 7 was used to predict the nucleotide substitution model showing the best BIC scores (Bayesian information criteria) (Nei and Kumar 2000). A Maximum Likelihood tree was done using the Tamura 3 parameter model (T92+G+I) of molecular evolution and 2000 bootstrap replicates. Moreover, the Network 5 program using the median-joining model (Bandelt et al. 1999; Fluxus Technology Ltd. 2020) was used for obtaining a haplotype network from representative species in the genus Sinotaia and including Torotaia as an external group.  Material examined: 7 specimens from Nora River (Asturias, northern Spain), May 2017; 2 specimens from Nora River (Oviedo, Asturias, northern Spain), June 2018; 1 specimen from Cazaux Lake (Arcachon, Aquitania, southwestern France), August 2017.

Systematics and taxonomy
Description: Thick, elongated and conical shell, 18-27 mm width and 25-36 mm height, yellowish-green or brown-olive, outer margin of aperture blackish ( Figure 1A-I). Aperture pear-circular ( Figure 1A, D, G), blackish axial growth lines usually present on first whorls, low spiral carina. Protoconch with 2-4 spiral raised lines with thin and triangular hairs in juveniles ( Figure 2); often ground down in adults. Six to seven whorls slightly convex, separated by deep sutures, with a thin brownish band below suture on spire. Carinated whorls with 2-3 spiral keels, sometimes with spiral lines between keels that may carry triangular lamellae often absent in adult stages ( Figure 2). Juveniles present a protoconch with 2-4 raised lines with long periostracal thin and triangular hairs (Figure 2A, B). The first part of the foot present dark spots in juveniles. Circular, horny operculum, smaller than the aperture, with concentric growth lines, occupying dorsal foot surface ( Figure 1K, J). Taenioglossate radula, narrow and no longer than 2 mm ( Figure 3A-G). Rachidian tooth with rectangular central cusp followed by 4-6 triangular cusps on each side ( Figure 3A-G). Two lateral teeth per row, curved and multicuspid, similar to the central teeth. Body with width cylindrical snout with an aperture in the anterior margin ( Figure 4A, B). Two large tentacles (double length of the snout) with a short ommatophore in the outer surface ( Figure 2A). Males with modified right tentacle that serves as a copulatory organ ( Figure 4B). Males and female specimens differing in the right margin of the mantle cavity, which is filled by the oviduct in females ( Figure 4C). This oviduct works as a brood pouch filled by capsules, covered by a thin membrane, with different stages of developing oocytes, embryos and juvenile snails ( Figure 4C).
Remarks: Studied S. cf. quadrata specimens displayed high radular variability, confirming previous observations by Ferreira et al. (2017) from Studied specimens were 25.9 ± 2.5 mm height and 16.2 ± 1.9 mm width (mean ± SD). Six out of 7 individuals were females (from Nora Rver), all of them carrying developing embryos and juveniles. Males were smaller than females (Height males = 22.45 ± 1.5 mm; Height females = 27 ± 1.3 mm), and sutures between whorls were deeper in males. The two specimens found in 2018 were juveniles of 9-10 mm height and 7-8 mm width.

Biology and ecology
Sinotaia cf. quadrata is an ovoviviparous species (commonly referred only as "viviparous"), in which the eggs develop within the female body and are birthed live. Studied specimens presented gonadal asynchrony with a continuous development of germ cells. Fertilized and unfertilized eggs and embryos in different stages of development were found along the female gonad ( Figure 4A-C). Mature oocytes were subspherical to slightly trapezoidal in shape ( Figure 4C) with a mean diameter of 2.07 mm (N = 15; SD = 0.11). Brooded juveniles ranged in shell height from 2 to 5 mm, with a mean of 3.4 mm (N = 20; SD = 0.86).
When the snails were prepared for the morphological study and placed in small Petri dishes (containing clean water) under illuminated light, two trematode cercariae were observed leaving the snails. The cercariae presented the following combination of morphological features: non-forked tail longer than its body and without finger-like processes; body entirely anterior to the tail and oral and ventral suckers present. Furthermore, colonies of cocci-and bacilli-shaped bacteria were found on the shell surface and the external body parts of the snails, as well as on the radula ( Figure 3C-H).

Genetic results
Fragments of 574-bp COI were successfully obtained from six studied specimens, and a total of two different haplotypes (Sinotaia AstuH1 and Sinotaia AstuH2) were obtained and deposited in GenBank under accession numbers MN737101 and MN737102 respectively. The NCBI Blast procedures conducted for genetic identifications revealed that more than 95 (Sinotaia AstuH1) and more than 100 (Sinotaia AstuH2) deposited sequences in Genbank coming from more than 8 putatively different species (B. aeruginosa, B. purificata, B. quadrata, B. dispiralis, B. angularis, B. lapillorum, S. quadrata histrica, B. turritus) showed all 100% query coverage, E value = 0 and more than 99% of identity with the Asturian haplotypes.
Phylogenetic relationships using COI sequences from a wide range of Viviparidae were assessed through Maximum Likelihood estimations ( Figure 5). The analysis revealed a similar topology and clades previously reported by Stelbrink et al. (2020). Our specimens showed two different entities (haplotypes) and were located within the Bellamyinae Clade A species complex, closely related to S. quadrata, S. quadrata histrica, S. purificata and S. aeruginosa, all of them with Asian origin ( Figure 5). In this clade we found low bootstrapping values (i.e. 67%) ( Figure 5). The African species complex (Bellamyinae Clade B) showed higher support as evidenced by boostrapping values (i.e. 90%).
The haplotypes analyses using Network revealed two different origins for the new reported haplotypes Sinotaia AstuH1 and Sinotaia AstuH2 haplotypes ( Figure 6). The Sinotaia AstuH1 (MN737101) haplotype found in France and Asturias was connected to one named S. purificata haplotype (KF535431) found in China by Gu et al. (2015a, b) (Figure 6). The second Asturias haplotype (MN737102) coincided with a S. quadrata haplotype (MN998013) found in Korea (Stelbrink et al. 2020) and were connected to a haplotype named S. quadrata histrica in Japan (LCO28489) (Hirano et al. 2015) (Figure 6). Few mutational changes connected most haplotypes from the reported species S. quadrata, S. quadrata histrica, S. purificata and S. aeruginosa. Despite this, haplotypes representing those species and also Sinotaia sp. specimens reported by Stelbrink et al. (2020) are shown slightly further apart (in terms of a high number of mutational changes) from the  cluster of haplotypes/species where the Asturian haplotypes were localized ( Figure 6). These results support the morphological identification of our specimens, which undoubtedly belong to the Asian genus Sinotaia. Moreover, they seem to be part of a highly genetically diverse single species (probably S. quadrata) as suggested by morphology. Both statements justified the use of "cf." (short for the Latin: confer/conferatur), evidencing the uncertainty or the provisional taxonomic status of S. quadrata (Ferreira et al. 2017).

Discussion
The herein reported occurrence of S. quadrata from northern Spain represents the second record of the family Viviparidae in the Iberian Peninsula. The only previously reported species of the family is V. viviparus, which was recorded from three localities of Portugal and northern Spain (Morelet 1845). However, the current occurrence of V. viviparus in Iberia is considered unlikely, since the original record was only based on a few empty shells collected from Portugal (Morelet 1845) and new records of this species were absent in more recent malacological studies from Portugal and northern Spain (Castillejo 1982;Holyoak et al. 2019). Sinotaia cf. quadrata is an Asian viviparid snail, native to China, Korea and Taiwan (Lee 2009), which has subsequently been introduced to Japan, Thailand and Philippines (Ovando and Cuezzo 2012).
The 648 bp region of the COI gene forms the primary barcode sequence for members of the animal kingdom and almost all results show that more than 95% of species in test assemblages of varied animal groups have been shown to possess distinctive COI sequences Hebert 2007, 2013). Comparison of 5' COI sequences from 13,000 pairs of congeneric species has showed a mean divergence of 11.3% (Hebert et al. 2003a). Ninety-eight percent of the species pairs under study by Hebert et al. (2003a, b) exhibited greater than 2% COI divergence (i.e., 10 substitutions per 500 bp). Cases of incomplete resolution involve species that are closely allied or taxonomically understudied groups Hebert 2007, 2013). It seems clear that the subfamily Bellamyinae and the genus Sinotaia in particular can be one of these cases mentioned above. Recently, Kagawa et al. (2019) reported that shell morphologies of Sinotaia spp. in Japan easily change under the influence of the environment and advised that high flexibility in shell morphology should be carefully considered when managing or studying these freshwater molluscs. Gu et al. (2019) reported that the Bellamya monophyletic Chinese group showed very little genetic differentiation among species due to reversal of the river courses and the influence of intermingling of different lineages in China (Gu et al. 2019). In river snails, due to the lack of comprehensive taxonomic and genetic data sets, the phylogenetic relationships and their global diversification dynamics are still poorly understood (Stelbrink et al. 2020). These authors claimed that the genetic dataset from recent phylogenetic studies (i.e. Gu et al. 2019;Hirano et al. 2019) is taxonomically very incomplete, including misplaced species and/or using inadequate mutation rates (Stelbrink et al. 2020).
The genetic study conducted here demonstrates the inefficacy of blast assignments for species identifications in the genus Sinotaia due to the huge amount of misidentified sequences that are currently available in the Bold and Genbank databases. Moreover, even using only published sequences with voucher IDs for phylogenetic and network haplotypes analyses, our results suggest the occurrence of a single species (i.e. S. cf. quadrata) with a confused nomenclature probably due to a high phenotypic plasticity both in adults and juvenile stages. Studied specimens of S. cf. quadrata (Spain and France), S. aeruginosa from China (Gu et al. 2015a), S. purificata from China (Gu et al. 2015b), S. quadrata histrica from Japan (Hirano et al. 2015), S. quadrata from China from Gu et al. (2015a, b) and Stelbrink et al. (2020), and a recent haplotype found in Italy (Cianfanelli et al. 2017), showed all to have very low levels of COI genetic divergence among what are supposedly considered as different species. It is probable that the use of more barcodes from other ribosomal or nuclear genes could help elucidate the discernment of Sinotaia spp. nomenclature. Stelbrink et al. (2020), using analyses based on the genes 28S rRNA, COI, and H3, found very high levels of support for their phylogenetic trees. It is obvious that the taxonomy of the genus Sinotaia requires revision, but this is outside of the scope of the present study. Even so, it is clear from the morphological observations, and the genetic results obtained here, that our studied specimens belong to the genus Sinotaia, and fit correctly with the S. quadrata species complex. Furthermore, the haplotypes network analysis may suggest multiple introduction events and may be from different origins. Multiple introduction episodes and adaptation of introduced populations constitute a serious risk and conditions for invasive outbreaks in all receiving ecosystems (Frankham 2005).
The population found at Nora River can be considered as established based on the finding of adult and reproductive-active females bearing embryos in 2017, and the presence of juveniles in 2018. Different studies have stated that S. quadrata is tolerant of a wide range of environmental variables, including pH and conductivity (Ovando and Cuezzo 2012;Ferreira et al. 2017), which could be an advantage for successful establishment. Further, the viviparous strategy ensures fecundity and short lifespan, favoring successful reproduction (Ferreira et al. 2017). Another feature that may enhance their establishment is the presence of a strong and thick shell that may protect them from potential predators (Ovando and Cuezzo 2012). Even so, future samplings are required in the area of study to confirm the persistence of this species in Nora River. A common introduction pathway of non-native species in aquatic environments is the aquarium trade; many gastropod species, including Sinotaia spp., have been introduced unintentionally in association with ornamental plants or fishes (Arias and Torralba-Burrial 2014;Ng et al. 2016;Patoka et al. 2017). The common water hyacinth (Eichhornia crassipes (Mart.) Solms, 1883) was found in the study area during our samplings, although this was removed later (authors pers. obs.). Sinotaia cf. quadrata may have been introduced in different juvenile stages with this ornamental plant. Another possible pathway could be an intentional introduction for human consumption, since these snails constitute a common food resource in Asia (Qian et al. 2014). This mechanism has been proposed as the most plausible introduction method to some regions of Italy, with presence of resident Asian communities (Cianfanelli et al. 2017). Another possible way of introduction was discussed by Ferreira et al. (2017) from Argentine populations. These authors consider that, since Sinotaia spp. are commonly used in aquaculture as a food source of carp and allied species, S. cf. quadrata may be accidentally introduced with fishes to artificial ponds in Argentina. Subsequently, the occasional floods could have connected the artificial ponds with natural water bodies, favoring the dispersion of the snail in the wild (Ferreira et al. 2017).
Sinotaia cf. quadrata may compete with native species for space or resources, or even change physicochemical parameters, turbidity or organic matter concentration of the water. Further, they may alter plant and algae biomass, thereby affecting the native communities (Ovando and Cuezzo 2012). Sinotaia chinensis, a species similar to S. quadrata, was introduced to North America and competed with other freshwater snails (e.g. Physa sp. and Lymnaea sp.), decreasing their abundances (Johnson et al. 2009). Moreover, S. quadrata also feeds on eggs and larval stages of some freshwater fishes such as the bluegill (Lepomis macrochirus Rafinesque, 1819), even producing a significant decrease in their abundance (Nakao et al. 2006). In the Nora Rriver, the demersal eggs and embryos of two common fish species, minnows (Phoxinus bigerri Kottelat, 2007) and brown trout (Salmo trutta Linnaeus, 1758), are similar to those of bluegills and unlike the bluegill, these species do not provide parental care to their offspring, making their predation by S. cf. quadrata even easier.
The trematode cercariae found leaving the body of one of the S. cf. quadrata studied specimens, are morphologically consistent with the diagnosis of the echinostome-like cercaria (Schell 1970). Several members of the genus Echinostoma Rudolphi, 1809 and close-related genera can infect humans as well as other mammals. In humans, these trematodes cause the symptomatology called Echinostomiasis, an intestinal infection commonly evidenced by a minor affliction, but in certain situations, the echinostomes can cause severe infections with ulceration of the intestinal mucosa and subsequent abdominal pain, diarrhea, anemia, and/or edema (Schell 1970). Previous studies revealed that S. quadrata populations from Taiwan and Japan might host metacercariae of echinostomid trematodes (Graczick and Fried 1998). Echinostomid metacercaria may affect humans by eating undercooked infected snails (Graczick and Fried 1998). Furthermore, populations of S. quadrata from Japan may also host the parasitic nematode Angiostrongylus cantonensis (Chen, 1935) (Lu et al. 2018). This nematode can cause Angiostrongyliasis in humans, damaging the brain and even the lungs, and causing meningitis (Lu et al. 2018). Consequently, the introduction of this species in the Nora River, northern Spain, implies a biosanitary risk, as they may infect the human population nearby and spread these diseases. Besides, our individuals presented bacteria on the radula and shell surface ( Figure 3C, D, H) in densities high enough to make difficult the observation of the radular structures. Specimens of S. quadrata in some areas of Italy (where this species had been released for food consumption) presented a rate of bacterial content 234 times higher than the rate suitable for human consumption and thus, their harvestmen were forbidden by the local government in order to avoid a serious sanitary risk (Empoli 2019).
Finally, our results confirm once again, the usefulness of an integrative taxonomical approach to shedding light on hidden biodiversity and the invasion process of freshwater ecosystems.