*Aims*

This study aimed to demonstrate the effectiveness of combined morphological and genetic approaches (both mitochondrial and nuclear markers) in resolving invasive bivalve identification. The research also assessed the methodology as an invasion-source tool by using *Corbicula* specimens from separate invaded sites in Ireland and Belgium.

#### **2. Materials and Methods**

#### *2.1. Specimen Collection*

Five sites in Ireland with known populations at the time this study was conducted, were sampled to collect individuals for genetic and morphological study (Figure 1; Table 1). One site in Belgium, in the Meuse River at Petit Lanaye (Table 1) was sampled on the 23/10/2013 to collect individuals solely for morphological comparison with individuals from the Irish sites.

A range of standard sampling methods were used to collect *Corbicula* specimens, depending on water depth and site location, including SCUBA diving, benthic dredge, grab sampler and kick-net. Samples in Ireland were collected from five sites between 2011 and 2013; the Shannon River at Carrick-on-Shannon, Lanesborough, Portumna and Lough Derg, as well as the River Barrow, the River Nore and the Meuse River (Table 1). The Shannon River basin sites were collected from a range of river and lake settings with the St. Mullin's and Red House sites [16] being riverine.

**Figure 1.** Irish rivers containing *Corbicula* specimens, previously described as *C. fluminea*, sampled during this study, River Nore, Barrow and Shannon. Sampling sites Carrick-on-Shannon, Lanesborough, Portumna, Lough Derg, St. Mullin's and Red House.


**Table 1.** Sampling sites and collection methods of *Corbicula* clams for DNA and morphometric analysis.

#### *2.2. Morphological Analysis*

A morphological examination of each individual from Ireland (N = 84) and Belgium (N = 51) was carried out, separating them into the morphotypes described in Europe, R, S, Rlc and also, an intermediate morphotype (Int) [19], in order to determine the form of *Corbicula* according to the morphotype descriptions in [21] and [26]. Form R, (round form) has a shell with well pronounced concentric ridges, is round and broad and generally attains a larger size than form S. Its internal colour is white but may contain pallid purple markings [26]. Form S, (saddle form) again has concentric ridges but these are less raised then in the R form and is narrower and proportionately longer. The internal shell colour is a deep purple throughout. Form Rlc (light colour R from) is superficially similar to the R form in shape, but with a lighter surface shell colour and an off-white to yellow internal colour [26]. Form Int is similar in shape to the R form but with finer less pronounced ridges on the shell [19].

All specimens selected from Ireland and the River Meuse were measured to the nearest mm for shell height (H), length (L) and width (W) using a pair of digital calipers to determine individuals of form R, S and Intermediate (Int), Rlc was not found (Table 2). The ratio between each measurement was calculated for the Irish *Corbicula*, and compared to individuals of form R, S and Int from the River Meuse. A Principal Component Analysis (PCA) was conducted in RStudio (Version 0.98.501), on the ratios between shell length, height and width, (L/H, L/W, H/W), as described in [21].


**Table 2.** Numbers of individuals (forms R, S and Intermediate (Int)) used for the morphological (Ireland and Belgium), mtCOI and microsatellites (Ireland only) analyses.

#### *2.3. DNA Extraction*

Adductor muscle tissue and foot were dissected from each of the Irish specimens for genetic analysis. The samples for the Carrick-on-Shannon, Lanesborough, River Barrow and River Nore sites were preserved in 98% ethanol and stored at ambient temperature. Samples from Lough Derg were initially preserved in methanol and transferred to ethanol after dissection.

Total genomic DNA was extracted from the adductor muscles and/or foot of 50 individual specimens (Table 2), using the DNeasy blood and tissue kit (Qiagen), according to the manufacturer's protocol. DNA extraction and microsatellite sequencing were carried out at the Laboratory of Evolutionary Genetics and Ecology, (LEGE), University of Namur, Belgium, as previously described in [11,23].

#### *2.4. Mitochondrial COI Gene Analysis*

A fragment of 710 bp of the COI gene was amplified in 25 individuals, with representatives from each Irish location. Polymerase Chain Reaction (PCR) was carried out using the universal primers LCO1490 and HCO2198 [34] following the protocol described in [23]. Amplicons were purified and sequenced with the forward universal primer HCO2198 on an automated ABI 3730XL Genetic Analyzer (Genoscreen, Lille, France). Retrieved sequences were visualized, aligned and edited using BioEdit 7.0.5.3 [35]. Corrected sequences and published sequences of the invasive lineages (FW5—Form A/R: GU721082; FW1—Form B: AF196269; FW4—Form Rlc: GU721084; FW17—Form C/S: GU721083) were added to our dataset and used to construct a haplotype median-joining network using the Network 4.6.1.2. [36].

#### *2.5. Microsatellite Marker Analysis*

Ten microsatellite markers [24] (ClA01, ClA02, ClA03, ClB03, ClB11, ClC01, ClC08, ClC12, ClE01, and ClD12) were amplified following the protocol of [24]. Microsatellite markers were read on an ABI 3130XL Genetic Analyzer with GeneScan-500 (LIZ) size standard (Applied Biosystems) and scored using GENEMAPPER (Applied Biosystems).

For each of the 50 individuals analysed, we defined a multilocus genotype (here, the unique combination of alleles for the 10 microsatellite loci). These individuals as well as 47 individuals from the European and American invasive individual's lineages previously typed (10 individuals from Form A/R, 16 from Form B, 10 from Form Rlc and 11 from Form C/S) were clustered based on their multilocus genotype using a discriminant analysis of principal components (DAPC) [37]. The DAPC analysis was performed using the package adegenet [38] implemented in R version 2.15.2 (R development core team 2008). The number of putative populations was first determined using the k-means clustering algorithm [39] for K = 1 to K = 11; 11 being the number of sampled populations added in the analysis. The appropriate number of clusters was defined using the Bayesian Information Criterion (BIC); the value at which the BIC distribution forms an elbow indicating the best clustering (Supplementary Material S1). The relationships between the BIC-defined clusters were then inferred. Six principal components and one discriminant function (98.9% of the total variance) were retained to represent the majority of the variability contained in the dataset.

## **3. Results**

## *3.1. Morphology*

From the morphological characterization, all Irish *Corbicula* clams were visually determined as the European form R as described in [19,23] with a round deep shell that may range from externally dark to golden, heavy ridges and a generally white interior (Figure 2).

**Figure 2.** Examples of *Corbicula* form R from the River Shannon at Lanesborough (juvenile specimen) and Carrick-on-Shannon and the River Barrow at St. Mullin's. Also shown is an example of Form S from the River Meuse at Maastricht.

The PCA analysis (Figure 3) of shell height vs length shows the Irish *Corbicula* grouping together with the Belgium form R and the international intermediate (Int) form. The Int form displayed an intermediate morphology between R and S, with the round body shape of R and the closely spaced shell ridges and purple interior coloring of S. The Irish individuals were different from the Belgian form S. No Irish individuals had a shell height/length ratio consistent with form S.

**Figure 3.** A principal component analysis (PCA) of the ratio between shell measurements for *Corbicula* form R from all Irish sites and forms R, S and individuals with intermediate morphotypes between form R and S (Int), from the River Meuse in Belgium.

#### *3.2. COI Sequence*

A 710 bp length fragment of the mitochondrial COI was successfully amplified for 24 of the Irish specimens (Table 2). All the COI sequences retrieved were identical to FW5, the mtCOI haplotype of form A/R distributed in Europe and Americas. All individuals sampled from the Irish populations possessed the same haplotype. No haplotypes from other invasive lineages (Figure 4) were detected from the Irish sites.

**Figure 4.** Haplotypic diversity and relationships in Invasive *Corbicula* clams inferred through median-joining Network. The European invasive haplotypes FW1 (Form B), FW4 (Form Rlc), FW17 (Form C/S), FW5 (Form A/R; Irish individuals) are plotted. Branch length was proportional to the number of mutations between the haplotypes and nodes proportional to the haplotype frequency.

#### *3.3. Microsatellite Markers*

Genetic diversity and genetic relationships in invasive *Corbicula* clams estimated through a Discriminant Analysis of Principal Components (Figure 5) based on multilocus genotypes from microsatellite marker amplifications show a common clustering between all Irish samples and the American and European invasive form A/R, suggesting that Irish *Corbicula* belongs to this same lineage.

**Figure 5.** Genetic diversity and genetic relationships in invasive *Corbicula* clams estimated through a Discriminant Analysis of Principal Components based on multilocus genotypes. Only axes 1 and 2 are represented. One dot represents one distinct multilocus genotype; individuals showing the exact same genotype are, therefore, pictured under the same dot.

## **4. Discussion**

Here, we used a multi-method approach, including mitochondrial COI sequencing, microsatellite genotyping and morphological analysis, to identify the invasive *Corbicula* lineages in Ireland. Such an integrative approach was necessary as the genus *Corbicula* is characterized by haplotypes capable of displaying highly divergent phenotypes in response to differing environmental conditions or following cross-lineages mixing.

Our results showed that only the invasive form A/R, the most widespread invasive lineage, was present from all samples collected in Irish waters, as determined using a combination of (a) morphological analysis, (b) mtDNA (COI) and (c) microsatellite markers.

No clams with either a narrow and fine-ridged shell or deep purple interior, as corresponding to form S [19,40] were observed. The PCA analysis carried out on the ratio between shell length and shell height (Figure 3) supports the correct classification of these individuals a belonging to form R. Morphologically, all clams sampled from the Irish sites conform to the invasive form A/R as described in [11,21,23,28,29], with a certain variability observed (Figure 3).

All individuals sampled from the Irish sites presented the invasive FW5 mtCOI haplotype of form A/R distributed in Europe and Americas [11,21,23,28,29]. The COI sequence data supports the results from the morphological analysis: no mismatches were observed between mtCOI and morphotype, which may be potentially caused by reproduction through androgenesis between distinct lineages [28].

All Irish sampled individuals showed a common clustering for the microsatellite data with the A/R lineage. This form has proven to be clonal and, thus, shows no genetic diversity. Indeed hundreds of individuals of this lineage, sampled at di fferent locations from across Europe, North and South America, present the exact same multilocus genotype with no genetic diversity [11]. The genetic variability detected within Irish populations is very likely due to the poor amplification of the microsatellites (Supplementary Material S2). Initial storage conditions of some samples in methanol was likely responsible for this. It is also a possible that false positives were detected from the data.

It is unlikely that other forms of *Corbicula* exist in Ireland as extensive investigations have not revealed clams displaying di ffering morphological characteristics [14,16,17] (Minchin Pers Ob.) or in subsequentially discovered populations [41,42]. As the level of genetic diversity among the invasive lineages of *Corbicula* clams (reviewed in [18]) was low, it was not possible to discriminate the origin of the Irish populations. The emerging field of massive parallel sequencing (MPS) to detect single nucleotide polymorphisms (SNPs) [43] could potentially complement the approaches used in our study. Similarly, it is not possible to draw any inferences on the number of discrete introductions that may have occurred [44].

The combined methodology of mitochondrial COI sequencing, microsatellite genotyping and morphological analysis has the potential to pick apart the exact identification of genetically complex species groups for which traditional ecological methods may have overlooked complex relationships such as in *Corbicula* [27] with its highly invasive clonal A/R form. The ability to discriminate between discreet introductions to a geographic area, and secondary spread is an invaluable tool in prioritising AIS biosecurity resources [45] and can inform horizon scanning [46].

## **5. Conclusions**

The use of a combined morphological and nuclear marker approach in resolving the identity of the invasive *Corbicula*, demonstrates that only one invasive lineage of *Corbicula* has invaded Ireland, the most prevalent and widely distributed form being A/R, a form also found across Europe and America. The extremely low genetic diversity found within this invasive lineage makes the determination of di fferences in population origins di fficult as a result, and therefore, the number of discrete introductions of *Corbicula* to Ireland remains unknown. In order to properly inform managemen<sup>t</sup> plans and limit impacts to ecosystem services, the identity of invasive *Corbicula* populations must be resolved. The use of a combined morphological, mitochondrial and nuclear marker approach in gleaning the identity of invasive *Corbicula*, as demonstrated in this study, provides a useful tool for achieving these goals, with the possibility for extending this approach to other invasive species.

**Supplementary Materials:** The following are available online at http://www.mdpi.com/2073-4441/11/8/1652/s1, supplementary S1, supplementary S2.

**Author Contributions:** Conceptualization, R.S., E.E. and F.L.; Methodology, R.S., E.E., F.L., K.V.D. and D.M.; Software, R.S. and E.E.; Validation, R.S. and E.E.; Formal analysis, R.S. and E.E.; Investigation, R.S, and E.E.; Resources, F.L. and K.V.D.; Data Curation, R.S., E.E., F.L. and K.V.D.; Writing—Original Draft Preparation, R.S. and E.E.; Writing—Review and Editing, R.S., E.E., F.L., D.M. and K.V.D.; Visualization, R.S. and E.E.; Supervision, F.L. and K.V.D.; Project Administration, R.S., E.E., F.L. and K.V.D.; Funding Acquisition, F.L. and K.V.D.

**Funding:** This research was funded by President's bursary award, Institute of Technology, Sligo. Financial support was also provided by the COST Action no. TD1209 'Alien species: Linking Information across a European Network (ALIEN Challenge)' fund. Financial support from Inland Fisheries Ireland (IFI) is also acknowledged.

**Acknowledgments:** Particular thanks to Alan Cullagh, Declan Cullagh, IFI Clonmel. Fergus Lynch and all of the IFI Drumsna staff. Joe Caffrey and Michael Millane of IFI Citywest. Julie Virgo of UNamur and Helen Moran of Geomara, Clarinbridge, Co. Galway, Ireland.

**Conflicts of Interest:** The authors declare no conflict of interest.
