New North American paratenic hosts of Anguillicola crassus and molecularly-inferred source of invasion

Brown bullhead (Ameiurus nebulosus), pumpkinseed (Lepomis gibbosus) and bluegill (Lepomis macrochirus) collected from the Paskamansett River, Massachusetts, were infected with larval Anguillicola crassus. These are new records of paratenic hosts of A. crassus in North America, although infected brown bullhead and pumpkinseed have been found in Europe. Prevalence was 64–100% and mean abundance 5.50–9.32. Morphological identification of L3 was confirmed by molecular sequence analysis of mitochondrial cytochrome c oxidase subunit I (COI, barcode) and the nuclear large subunit nLSU (28S) ribosomal RNA genes. Analysis of COI sequences from larval and adult worms from the Paskamansett River and nLSU (sequences from those L3 and adults from the Mira River, Cape Breton, Nova Scotia) showed that the COI gene better discriminated among species of Anguillicola than did the nLSU gene. At least 9% difference in sequence similarity was found between A. crassus and other species with COI, whereas there was overlap in the nLSU sequence similarity between A. crassus and other Anguillicola species. Comparative sequence analysis suggests that North American A. crassus originated from Japan, whereas European A. crassus originated from Taiwan, in agreement with previous studies. Two single nucleotide polymorphism (SNP) markers, SNP390 and SNP91 in COI and nLSU, respectively, are informative for differentiating between North American and European sources. Analyses with both genes also supports the monophyly of the Anguillicolidae.


Introduction
The nematode Anguillicola crassus Kuwahara, Niimi and Itagaki, 1974 infects the swimbladder of eels (Anguilla spp.).Considered native to Japanese eels Anguilla japonica Temminck and Schlegel, 1846, the parasite was first observed in Europe in the European eel Anguilla anguilla (Linnaeus, 1758) in 1982, subsequently spreading throughout most of that eel's range in less than 20 years (Taraschewski 2006;Lefebvre et al. 2012a, b).This invasive parasite is a concern because it causes lesions, inflammation, haemorrhaging and fibrosis in the European eel's swimbladder, possibly affecting its ability to migrate to spawning grounds in the Sargasso Sea (Lefebvre et al. 2012a).The parasite also has been reported in American eels Anguilla rostrata (Lesueur, 1817) from North American waters in Mexico, Texas, South Carolina, North Carolina, Rhode Island, Massachusetts, Maine, Chesapeake Bay, Hudson River drainage, New Brunswick and Cape Breton-Nova Scotia (Aieta and Oliveira 2009;Lefebvre et al. 2012a;Li et al. 2015).The parasite was transferred to the genus Anguillicoloides following a systematic revision based on morphological characteristics (Moravec 2006).However, a phylogenetic analysis of three gene sequences of the Anguillicolidae suggest that the genera Anguillicola and Anguillicoloides are mono-Table 1. Prevalence (P, %), mean intensity (I, + S.E.) and mean abundance (A, + S.E.) of L 3 of Anguillicola crassus in paratenic fish hosts from the Paskamansett River, Massachusetts, collected in August, 2012.N, number of fish examined; L, mean standard length (mm) and M, mean mass (g) of fish.
Species (common  phyletic, and thus the original name has been retained (Laetsch et al. 2012).
In this study, we report new paratenic hosts of A. crassus from North American waters.Specifically, paratenic hosts are recorded from the Paskamansett River, Massachusetts, where prevalence of A. crassus in the eel has fluctuated between 85 and 95% (Oliveira, unpublished).We also confirm the identification of L3 of A. crassus from paratenic hosts and adult parasites using the barcode cytochrome c oxidase subunit I (COI) gene, and for the first time provide the analysis of nuclear 28S large subunit (nLSU) ribosomal RNA gene of A. crassus from North American worms.

Fish collections and parasite analyses
Forage fish belonging to nine species (see Table 1) were collected opportunistically by electrofishing from the Paskamansett River (41.67778ºN; 70.97695ºW) in August 2012.Fish were placed on ice and frozen upon return to the laboratory.Upon thawing, fish were identified, measured for length and mass, and the viscera removed.The swim bladder, stomach and intestine were placed in a Petri dish.The mesenteric tissues were separated and examined using a stereomicroscope.After opening the swimbladder, stomach and intestine and discarding their contents, these organs were squashed between glass plates and stereomicroscopically examined.Nematodes found were photographed and fixed in 70% ethanol.Larval identification was based on the morphological description in Moravec (2013).American eels were collected while migrating from the Paskamansett River (41.57090ºN; 71.00160ºW) using fyke nets with 6 mm mesh in October 2011, and the Mira River (46.04723ºN; 60.01927ºW) using minnow traps or beach seines in Cape Breton, Nova Scotia from August to December 2012.Eels from the Paskamansett River were brought back to the laboratory live and examined for adult parasites in the swimbladder.Those from the Mira River were kept on ice, frozen upon return to the laboratory, thawed, the swimbladders removed and examined as above.Specimens of A. crassus were fixed in 95% ethanol and refrigerated.
Abundance refers to the number of A. crassus in an individual fish, including uninfected fish.Intensity is the number of A. crassus in an infected fish.Prevalence is defined as the proportion of fish infected with A. crassus, expressed as a percentage (Bush et al. 1997).Fulton's condition factor was measured as K = weight (g) / (total length (cm) 3 × 1000 (Ricker 1975).Correlations between the abundance of A. crassus and the length, weight and condition of brown bullheads were tested using Spearman's correlations.

DNA purification and polymerase chain reaction (PCR)
We used GeneJET ™ Genomic DNA Purification Kit (Thermo Scientific, cat #K0722) to purify the genomic DNA of adult and L3 A. crassus.We processed 0.5 cm of the adult worm's anterior portion tissue to avoid any host blood in the digestive tract and DNA was eluted with 60 l Elution Buffer.The same procedure was used to purify single L3 DNA using one-third of the kit's reagent volumes and DNA was purified in 20 l Elution Buffer.
A universal primer set was designed to amplify nuclear large subunit nLSU (also 28S) rRNA gene of the conserved regions at the ends of five Anguillicola species.M13 sequences (underlined) were tagged to the 5'ends of M13Ang28SF 5'-TGTAAAACGA CGGCCAGTATGAAGCGGATAGAGTTAACG-3' (forward) and M13Ang28SR 5'-CAGGAAACA GCTATGACATTTGCACGTCAGAACCGCTTC-3' (reverse) primers.The nLSU gene (687 bp) was amplified in a volume of 25 l containing 0.3 M of each primer (Eurofins Genomics), 0.625 Units of DreamTaq ™ DNA Polymerase (Thermo), 1X DreamTaq PCR buffer containing 2.0 mM MgCl 2 (Thermo), 0.2 mM deoxynucleotide dNTP solution mix (New England BioLabs, cat# N0447L) and 2-3 l of adult or 5-6 l of larval total DNA.The mixture was denatured at 95 C for 2 min followed by 5 cycles at 95 C for 30 sec, 60.2 C for 30 sec and 72 C for 1 min, followed by 40 cycles at 95 C for 30 sec, 65.1 C for 30 sec and 72 C for 1 min and the last extension at 72 C for 10 min.
For the barcode gene COI, we used M13-tagged degenerate primers to amplify a 524 bp sequence within the conserved Folmer region (Folmer et al. 1994).M13AngCOIF 5'-TGTAAAACGACGGCCA GTRTCWTTTTTRATTCGTTTTG-3' (forward) and M13AngCOIR 5'-CAGGAAACAGCTATGACTW CGRTCYATYAAYARYATAG-3' (reverse) primers were used, in which W = A or T, R = A or G, and Y = C or T. COI amplification was executed in a similar volume, PCR ingredients and protocol as the nLSU gene amplification except for using different primers and annealing temperatures.In the first 6 cycles, the annealing temperature was 51.8 C for 30 sec then 58.1 C for 30 sec in the last 39 cycles and the last extension was 72 C for 10 min.PCR products of both genes (5 l) were verified on 1% agarose gel for the right size and absence of nonspecific byproduct prior to purification.For sequencing, 10 l of the successful amplicon was purified with 4 l of ExoSAP-IT (affymetrix, USB Corporation, USA) and was incubated at 37 C for 18 min for unused nucleotides and primer degradation, followed by a second incubation at 80 C for 15 min for enzyme inactivation.The purified DNA was halved for bidirectional sequencing with the universal M13 forward and reverse primers.COI and nLSU genes were sequenced at Genome Quebec (McGill University, Canada).Sequences were edited using Geneious software version 6.1.8(http://www.geneious.com,Kearse et al. 2012).The accession numbers of the deposited COI (n = 60) in Genbank are MF458488-MF458547 and PopSet # 1229247113.The accession numbers of nLSU sequences (n = 67) are MF449305-MF449371 and PopSet # 1214998041.Morphological specimens (n = 25) were deposited at the Royal Ontario Museum (ROM, Toronto, ON), with accession numbers ROMIZ F625-ROMIZ F649.

COI and nLSU sequence manipulation
A comprehensive COI gene sequence alignment analysis (n = 596) was performed using the A. crassus L3 sequences from in this study (n = 31) isolated from paratenic hosts and from adult worms (n = 29) from eels.In addition, Anguillicola species COI sequences from four different PopSets available in Genbank were used in analysis (1) PopSet # 338970719, n = 95 (Laetsch et al. 2012) (Wielgoss et al. 2008).Furthermore, four non-Anguillicola COI sequences were included in the sequence alignment to serve as outgroups for tree rooting.We extracted only the regions that would align to Anguillicola COI from the entire mitochondrial genome of Toxocara canis (accession # EU730761, region 67-730), Ascaris suum (accession # HQ704901, region 73-729), Dracunculus medinensis (accession # HQ216219, region 305-961) and the entire published COI sequence of Strongylida sp.AM-2008 (accession # FJ172978) was used.The 592 aligned Anguillicola COI sequences belong to Anguillicola (Anguillicola) globiceps Yamaguti, 1935, n = 1; Anguillicola papernai Moravec and Tcharaschewsky, 1988, n = 38; Anguillicola novaezelandiae Moravec and Tcharaschewsky, 1988, n = 11;Anguillicola australiensis Johnston and Mawson, 1940, n = 15;new A. crassus addressed in this study (n = 60) and other A. crassus (n = 467).ClustalW multiple sequence alignment software (Larkin et al. 2007) was run from inside Geneious molecular software with optimized settings such as IUB matrix, gap open cost 15 and gap extension cost 6.66.After initial multiple sequence alignment and trimming, another ClustalW sequence alignment was performed to deduce the pairwise sequence similarities.Finally, a Bayesian phylogenetic inference tree (MrBayes 3: Ronquist and Huelsenbeck 2003) of Anguillicola COI sequences was constructed.MrBayes plugin was run using Geneious software.All Bayesian inference default settings were unchanged except the variation rate was set to equal, COI sequence for T. canis was selected to be the main root outgroup of the tree and JC69 was chosen as the base substitution model (Jukes and Cantor 1969).Initially, we constructed a tree using all 596 COI sequences, however, for clarity we used a concise COI tree consisting of only 94 A. crassus, 10 sequences of other Anguillicola species and the four outgroup sequences.To prepare the concise phylogenetic COI tree (n = 108), ClustalW alignment settings as mentioned above were used and the trimmed sequences were realigned to infer pairwise comparison.Finally, a COI Bayesian tree was reconstructed and genetic distances were determined.

Intra-and Interspecific genetic variation
Four aligned non-Anguillicola plus 583 Anguillicola COI sequences were processed with TaxonDNA Species identifier (Meier et al. 2006) to explore intra-and interspecific genetic distances, match sequences and cluster sequences based on pairwise distances.For more details on the setup and procedure of the manipulated COI and nLSU sequences, see Supplementary material Figures S1 and S2 captions.

Allelic frequency analyses
Anguillicola crassus COI sequences (n = 527) from three continents were analyzed for their single nucleotide polymorphism (SNP) on the locus 390 (SNP 390 ).Numbering of this SNP is based on our deposited A. crassus COI sequences in Genbank (PopSet 1229247113).One European sequence (accession # EU376851) from Lake Neagh, Northern Ireland was excluded from the analysis because its locus is occupied by base-A.The remaining 526 A. crassus COI gene sequences possess either the base "T" or "G" on SNP 390 (see Results).For the nuclear gene, the allelic frequency distribution of the SNP 91 of 146 A. crassus nLSU sequences was analyzed.The expected values in Contingency tables (Supplementary material Tables S1 and S2) are equal to (row total × column total)/n, where n is the total number of observations.GraphPad Prism version 5.00 was used for Chi-square and Fisher exact test analyses and Allelic frequency graph production.
Molecular sequence and phylogenetic tree analyses COI sequences (Table 2) belonging to 31 L3 and 29 adult worms from the Paskamansett River and nLSU sequences (Table 3) from 35 L3 from the Paskamansett River and 32 adults from either the Mira River (n = 6) or the Paskamansett River (n = 26) were compared to the entire Genbank sequence database.The best matching "hits" obtained from the Blast analyses with percent similarity 99.8%-100% revealed that these sequences belong to Anguillicola crassus (Tables 2, 3).
Pairwise nLSU or COI sequence analyses obtained from either Geneious or Genbank differed slightly in their identity (ID) ± 0.9% (Tables 2 and 3), indicating that our sequence alignments were properly done and can be used in constructing the phylogenetic COI and nLSU trees.The slight differences in ID % are simply because Genbank Blast reports the identity % without fractions while Geneious accurately describes pairwise matchings with decimals.The North American A. crassus COI sequences herein have the best matches to several conspecifics isolated from Mikawa Bay-Aoki, Japan (MIK) or to those from St. Jones River (StJ) in Delaware, USA (Wielgoss et al. 2008).Moreover, two sequences collected from China (accession # JF805656 and JF805655) and an isolate originated from Taiwan (accession # JF805712) are identical to most of our COI sequences (Table 2, Figure 1).American A. crassus nLSU sequences from the present study (Table 3) are matched with 100% identity to other 14 nLSU sequences collected from Taiwan with accession # FJ748532-FJ748544 and FJ748547 (Heitlinger et al. 2009).The second best matches to our nLSU sequences are only 97.3% with A. crassus nLSU sequences from Europe (see Clade B in Table 3 and Figure 2).
The COI gene appears better than nLSU in delineating A. crassus from the other four Anguillicola species.For instance, delineation of COI between conspecific and congeneric sequences is > 9% (Table 2), while there is an overlap in nLSU sequence similarity between American and Asian A. crassus conspecifics and congeneric sequences of A. crassus and either A. australiensis or A. novaezelandiae (Table 3).COI has a large barcoding gap > 7% that easily discriminates A. crassus from the other four species (Supplementary material Figure S1).In contrast, nLSU's barcoding gap is not clear because some Japanese A. crassus sequences overlap with A. australiensis and A. novaezelandiae in the first interspecific peak (Supplementary material Figure S2). A. crassus has two interspecific peaks in both COI and nLSU genes in which the first peak represents the genetic distance between A. crassus and either A. australiensis or A. novaezelandiae while the second peak is mainly associated with A. papernai, A. globiceps and to a lesser extent A. australiensis or A. novaezelandiae (Supplementary material Figures S1 and S2).
Bayesian phylogenetic inference of Anguillicola COI tree (n = 108) suggests that the five Anguillicola species originated from the same common Anguillicolidae ancestor (Figure 1, the open circle).Three A. crassus clades can be separated based on their COI sequences (Figure 1).Clade A consists mainly of North American A. crassus including all new sequences herein (n = 60), plus 28 (out of 32) StJ isolates (Wielgoss et al. 2008) and some Asian sequences, mainly from Mikawa Bay, Japan.Clade B contains all sequences isolated from Europe and some sequences originating from China, Taiwan (Kao River), Japan (Yamaguchi and Mikawa Bay) plus four North American sequences from the StJ.Clade C includes several Asian COI sequences clusters from China, Taiwan and the two bays in Japan, none of which occur in Clades A or B. A single European isolate from Turkey with accession # JF805721 is present in Clade C (Table 2, Figures 1 and 3).

Specific Single nucleotide polymorphism (SNP) and insertion/deletion (INDEL) genetic markers:
No indels were identified within the sequenced region of A. crassus COI gene sequences (n = 527) or in the other four Anguillicola species COI (n = 65).However, we found several SNPs, one of which, at locus # 390, that clearly can distinguish between North American and European isolates based on its allelic distribution.The G-allele is present in 321 of 323 European COI sequences while the T-allele occurs only in one COI sequence (Table 4).The latter COI was from an A. crassus isolate from France (accession # EU376745).In contrast, the frequency of these two alleles G/T is reversed in American A. crassus where T-allele predominates in 88 out of 92 sequences (Table 4).The Asian nematodes show more heterozygosity with respect to SNP 390 although the G allele is still predominant (Table 4 and Figure S3a).
The frequency distribution of T/G alleles in COI SNP 390 within A. crassus from the three continents is depicted in Supplementary material Figure S3a and their location is illustrated in a multiple COI gene sequence alignment in Figure 3.These T and G alleles do not affect COI protein functionality as they occur on the third nucleotide of the ACT or ACG codons, both of which code for the same amino acid, threonine (T), in the expressed mitochondrial COI protein (Supplementary material Figure S4).Table 2.A comparative ClustalW pairwise sequence alignments of the mitochondrial barcoding COI gene in Anguillicola crassus.Data obtained from either online service tool (Genbank /BLAST) or Geneious molecular software.Two A. crassus COI gene sequences from the current study (i.e. the queries) were blasted against other 22 hit sequences of Genbank's databases.The latter consist of 18 conspecific and 4 congeneric sequences.Extracted Genbank analysis data such as hit's name, accession #, the alignment length (in percent) and the degree of similarity (i.e.ID %) with the query are described.Sequence similarities generated by Geneious were compared to Blast analysis to validate Geneious alignment optimization prior to constructing COI phylogenetic tree.Selective genetic distances used in phylogenetic COI tree (see Figure 1) are also included on the table.Base substitution model such as Jukes-Cantor (JC) was used for genetic distance correction.Data were arranged according to genetic distance (ascending) and sequence ID % (descending) orders.Conspecific A. crassus sequences were categorized to three main clades A, B and C based on their distance range against the queries.Full description of country and region abbreviations are as mentioned in Figures 1 and 2.

Genbank analysis (BLAST)
Geneious  (Supplementary material Table S1) and SNP 91 of the nLSU gene (Supplementary material Table S2) in North American and European A. crassus sources deviated significantly (chi-square and Fisher exact tests, p <0.0001) from Hardy-Weinberg equilibrium.

Discussion
Worm L3 morphology and measurements correspond to the description in Moravec (2013).Bluegill is a new host record for this parasite.The North American native brown bullhead and pumpkinseed from Kolenhaven in the Albertcanal, Belgium were reportedly infected (Thomas and Ollevier 1992).
All larval nematodes appeared viable with no evidence of strong host reactions, suggesting the three species from the Paskamansett River are good paratenic hosts.Experimental transmission of larval A. crassus to eels was successful in cases where paratenic hosts exhibited weak host reactions (Székely 1996).Few dead worms were observed in pumpkinseed from Lake Balaton and this species was considered a good paratenic host (Székely 1995).
Anguillicola crassus has been found in American eels in Mexico, Texas, South Carolina, North Carolina, Delaware, New York, Maryland, Chesapeake Bay, Hudson River drainage, New Brunswick and Cape Breton (Barse et al. 2001;Moser et al. 2001;Aieta and Oliveira 2009;Fenske et al. 2010;Lefebvre et al. 2012a;Denny et al. 2013).However, molecular confirmation was limited to a single study that used COI based identification of A. crassus from the StJ in Delaware, USA (Wielgoss et al. 2008).Here, we confirm the molecular based identification of L3 of A. crassus from paratenic hosts using the barcode COI gene and provide for the first time nLSU gene sequences of A. crassus in North America.A previous study found the barcoding gap sequence divergence between conspecific and congeneric sequences within Anguillicola COI was 5.6% (Laetsch et al. 2012), higher than the recommended threshold of 2% for Metazoa (Hebert et al. 2003).In re-evaluating the barcoding gap of this gene with a more comprehensive analysis (n = 592), which included more American isolates (n = 92), the gap divergence was 7% (Supplementary material Figure S1).Both current and past studies support the use of the barcode gene as the gene of choice in the molecular-based identification of A. crassus.In contrast, the nLSU barcoding gap was not clear in Laetsch et al. (2012) or herein.Wielgoss et al. (2008) compared COI sequences in 16 different A. crassus populations from Europe (11), America (1) and Asia (4).The latter came from River Kao-Ping (KAO) in Taiwan or from two rivers in Japan, River Fushino, Prefecture of Yamaguchi (YAM) and Mikawa Bay-Aoki (MIK).Because of a single eel that was infected with 17 A. crassus, the 29 MIk COI sequences [see PopSet # 170297347 (accession # EU376652-EU376680)] were subdivided into MIK-1 and MIK-2.Japanese MIK-1 sequences were similar to KAO sequences and regardless of their geographic distance, they both share sequence similarity with European A. crassus populations (Wielgoss et al. 2008).In contrast, MIK-2, also from Mikawa Bay-Aoki, is a very unique infrapopulation of A. crassus; both MIK-2 and YAM sequences are dissimilar to the MIK-1/KAO samples, and their COI sequences are highly differentiated from European locations (Wielgoss et al. 2008).The single American population included in that study was from StJ and twenty-nine of 32 of these sequences share identity with MIK-2 (Wielgoss et al. 2008).In an reassessment of StJ sequences along with 60 new ones from the Paskamansett River, American nematodes show highest similarity with MIK-2, in agreement with Wielgoss et al. (2008) as well as one sequence from Taiwan (accession # JF805712) and two from China (accession # JF805656 and JF805655).The SNP 390 marker was very useful in differentiation between North American and European COI sources; it is a synonymous SNP as it does not change the COI protein sequence in either European or American A. crassus (Supplementary material Figure S4).
Both mitochondrial COI and nuclear LSU genetic distances (Tables 2 and 3) reveal that Clades A and B are more related to each other than to Clade C. The highest genetic diversity of A. crassus is found in the native area (Japan, China and Taiwan); the parasite is more differentiated in these countries than in any European or North American countries (Wielgoss et al. 2008;Laetsch et al. 2012;Lefebvre et al. 2012a).We propose the presence of different A. crassus clusters in East Asia, some of which are more related to each other than to others.Two of these related groups are the likely ancestors of European and American worms.Those related Asian groups are present in Clade A and B while the least related Asian isolates are clustered in Clade C.
The invasion of A. crassus into Europe apparently came from Taiwan in the 1980s via the live Japanese eel trade and to North America from Japan in the 1990s (Neumann 1985;Fries et al. 1996;Barse et al. 2001;Wielgoss et al. 2008;Laetsch et al. 2012).Based on their COI gene sequence, the specimens of A. crassus collected from Paskamansett River and Mira River are closely related to Japanese MIK-2 isolates (Wielgoss et al. 2008).However, according to the nLSU sequence analysis, North American isolates are identical to worms collected from Taiwan (Heitlinger et al. 2009).This is the first study combining COI gene and nLSU gene sequence analyses together from the same North American individual worms.It is possible that Taiwanese conspecifics that share identical nLSU sequences with our worms were collected from Taiwan but their native origin was Japan: the country of A. crassus collection does not necessarily reflect the native origin (Lefebvre et al. 2012b).COI gene sequence analyses indicate our worms from Clade A are similar to Mikawa Bay parasites as well as a few other sequences collected from China and Taiwan (n = 3).Notably, 14 other samples from China and 54 from Taiwan did not match North American A. crassus COI sequences (see Table 4).However, Wielgoss et al. (2008) found A. crassus belonging to two separate genotypes from Mikawa Bay, and our COI cluster shows these worms belong to either Clade A (with North American parasites) or B (with European parasites).Thus, we believe that the nematodes from Massachusetts and Cape Breton originated from a Japanese source, in agreement with previous studies (Barse et al. 2001;Wielgoss et al. 2008;Laetsch et al. 2012).While 4 specimens from StJ also aligned with Clade B, this does not rule out a Japanese origin.Regardless, the separation of the majority of North American from European A. crassus according to both gene analyses clearly supports distinct colonization histories.
Anguillicola crassus introduction may have occurred through a single invasion to either Europe or America (Wielgoss et al. 2007(Wielgoss et al. , 2008;;Laetsch et al. 2012), but a secondary mode of A. crassus introduction is also possible (Aieta and Oliveira 2009;Denny et al. 2013).The discontinuous distribution of A. crassus along the northeastern coast of North America supports the possibility of secondary introductions (Aieta and Oliveira 2009).The presence of a few American StJ worms (n = 4) associated with European A. crassus within Clade B (Figures 1 and 3) suggests that these American isolates could have resulted from a secondary A. crassus introduction to North America from elsewhere, perhaps directly from Taiwan or indirectly via Europe.Whether A. crassus invasion occurred in single or multiple incidences is uncertain but one scenario, supported by the absence of a significant correlation between genetic and geographic distances, is that multiple introductions of A. crassus to Europe occurred (Wielgoss et al. 2008).
The five species of anguillicolids were divided into two genera according to taxonomic standards: Anguillocola (containing a single species, Anguillicola globiceps) and Anguillicoloides, containing the remaining species (Moravec 2006).Our current molecular analysis of nLSU and COI genes and that of a previous study (Laetsch et al. 2012) found that Anguillicola globiceps and the other four species are monophyletic and support the reassignment of all the five species to the genus Anguillicola.
In this study, we used two genes for the molecular identification of American A. crassus collected from new North American paratenic hosts and eels from two localities along the eastern seaboard.We recommend including the nLSU gene along with the COI barcode for future molecular analyses of North American A. crassus to help confirm this invasive parasite's colonization history and spread across its distributional range.

Figure 4 .
Figure 4.A full-length alignment of thirteen A. crassus nLSU gene sequences (672 bp) shows the sequence differences between the three clades A-C: North American A. crassus (n = 67) and their Taiwanese ancestors (n = 14) form Clade A possess eight insertions which are absent in other clades (stars); an SNP on locus 91 occupied by "A" base in all North American A. crassus nLSU and their Taiwanese ancestors (black circle); nLSU of Clade C is has three extra SNPS (open circles) and a single insertion (the square) that are not present in other clades.

Table 3 .
A comparative ClustalW pairwise sequence alignments of the nuclear ribosomal large subunit (nLSU = 28S) rRNA gene of Anguillicola crassus.Data obtained from either online service tool (Genbank / BLAST) or Geneious molecular software.Three A. crassus nLSU gene sequences from the current study (i.e. the queries) were blasted against other 15 hit sequences of Genbank's databases.The latter consist of 11 conspecific and 4 congeneric sequences.Extracted Genbank analysis data such as hit's name, accession #, the alignment length (in percent) and the degree of similarity (i.e.ID %) with the query are described.Sequence similarities generated by Geneious multiple sequence alignment software and from the BLAST report were compared to validate Geneious alignment optimization prior to drawing nLSU gene phylogenetic tree.Selective genetic distances used in phylogenetic nLSU tree (see Figure2) are also added on the table, in which Jukes-Cantor (JC) model was used for genetic distance correction.Data were arranged to the genetic distance (ascending) and sequence ID % (descending) orders.Conspecific A. crassus sequences were categorized to three main clades A-C based on their distance range against the queries.Full description of country and region abbreviations are as mentioned in Figures 1 and 2 except of CB, CAN (Cape Breton-Nova Scotia, Canada).