SEARCH

SEARCH BY CITATION

Keywords:

  • cryptic hybridization;
  • Corbicula

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Specimen studied
  6. DNA-isolation
  7. Sequence analyses
  8. Morphometrics of RHA individuals
  9. DNA Amplification Fingerprinting from RHA individuals
  10. Nested clade analysis of RHA population
  11. PCR-RFLPs from RHA individuals
  12. Karyograms from RHA individuals
  13. Test for statistical associations between data sets from RHA population
  14. Phylogenetic analyses of worldwide Corbicula samples
  15. Results
  16. Morphometrics
  17. COI haplotypes
  18. DAF-variation
  19. ITS1 RFLPs
  20. Associations of morphotypes, multilocus genotypes, haplotype clades and RFLP pattern
  21. Karyotype analysis of population RHA
  22. Mitochondrial sequence variation and phylogenetic relations in world wide Corbicula
  23. Discussion
  24. Reproductive isolation and detection of hybridization
  25. Taxonomic, phylogenetic and phylogeographical considerations of the species status in Corbicula
  26. Conclusions
  27. References

We studied two Corbicula morphotypes in a syntopic population in the Rhine River in order to reveal their taxonomic, reproductive and phylogenetic relationship, using morphometrics, DAF-fingerprinting, mitochondrial COI and nuclear ITS1 sequence variation. Morphometric analysis showed that two statistically distinguishable morphotypes with few intermediates were present.Mitochondrial sequence analysis detected two divergent clades. DAF-fingerprinting revealed three highly distinctive multilocus genotypes. Two of the multilocus genotypes were significantly associated with different morphotypes and mitochondrial lineages. The third genotype B, however, was found in both morphotypes, intermediates and mitochondrial lineages. Conclusive evidence for hybridization came from RFLP analysis of the nuclear ITS1 locus. We interpret the hybrids as F1 hybrids between different evolutionary lineages. Integration of Corbicula sequences from all over the world into Maximum Parsimony analysis suggested a simultaneous radiation resulting in several evolutionary lineages whose species status remained doubtful. An unequivocal taxonomic assignment of the two evolutionary lineages in the Rhine population was therefore not possible.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Specimen studied
  6. DNA-isolation
  7. Sequence analyses
  8. Morphometrics of RHA individuals
  9. DNA Amplification Fingerprinting from RHA individuals
  10. Nested clade analysis of RHA population
  11. PCR-RFLPs from RHA individuals
  12. Karyograms from RHA individuals
  13. Test for statistical associations between data sets from RHA population
  14. Phylogenetic analyses of worldwide Corbicula samples
  15. Results
  16. Morphometrics
  17. COI haplotypes
  18. DAF-variation
  19. ITS1 RFLPs
  20. Associations of morphotypes, multilocus genotypes, haplotype clades and RFLP pattern
  21. Karyotype analysis of population RHA
  22. Mitochondrial sequence variation and phylogenetic relations in world wide Corbicula
  23. Discussion
  24. Reproductive isolation and detection of hybridization
  25. Taxonomic, phylogenetic and phylogeographical considerations of the species status in Corbicula
  26. Conclusions
  27. References

The worldwide anthropogenic introduction of exotic species into new environments has posed many serious ecological and economic problems. Many freshwater taxa seem to be especially successful invaders. The increasing ship traffic of cargo has led to ample opportunities for these taxa to invade foreign ecosystems with the ballast water. Besides causing ecological problems, secondary contact zones between formerly isolated but closely related taxa can lead to complex species interactions, like hybridization, character displacement or may even promote speciation (Jiggins & Mallet, 2000). Secondary contact zones thus pose questions concerning the level of reproductive isolation among the lineages, evolutionary processes and taxonomic issues (for reviews see Harrison, 1990; Arnold, 1992; Barton, 2001).

In this study, we focus on the bivalve genus Corbicula, of which several taxa were recently introduced to Europe. The genus Corbicula has attracted a considerable attention during the past decades as a result of its invasion of large parts of the world (Counts, 1986). The original distribution of the genus at the beginning of thetwentieth century was confined to Africa, Asia and Australia (Counts, 1980). Since then, the genus has attained an almost worldwide distribution (Ituarte, 1981; Mouthon, 1981; Counts, 1986).

The first record of an occurrence outside the original genus distribution dates back to 1924 from the west coast of North America (Counts, 1981). After about 40 years, the mussels reached the Atlantic coast. Two species appear to have invaded the North American continent (Siripattrawan et al., 2000), although it is not clear which taxa were involved. Further studies showed that Corbicula arrived in South America in the 1970s (Ituarte, 1994).

In Europe, Corbicula is believed to be present since the early 1980s (Mouthon, 1981). Since the first occurrence, there has been a debate about the number of species present, to which described taxon they belong and where they came from (Kinzelbach, 1991; den Hartog et al., 1992; Araujo et al., 1993). Renard et al. (2000) analysed the genetic and morphometrical variation of Corbicula in French rivers and reached the conclusion that two morphotypes were present and that they belong to the described species C. fluminea (O.F. Müller 1774) and C.fluminalis (O.F. Müller 1774). Additionally, they found another haplotype lineage, to which they attributed species status (C. spec.), although they could not assign a specific name to the taxon.

In the Rhine and other middle European rivers, at least two morphotypes, identified as C. fluminea and C. fluminalis, were described. It was, however, not clear whether they belonged to C. fluminea and C. fluminalis in the sense of Renard et al. (2000), or whether additional species were involved (Kinzelbach, 1991). In the Rhine, both morphs are simultaneous hermaphrodites. They have differing life-histories and ecological preferences (Meister, 1997), but can occur in syntopy (den Hartog et al., 1992; see below).

Surprisingly, despite the wealth of studies concerning ecology (Lee & Chung, 1980; McMahon & Williams, 1986; Stites et al., 1995; Cataldo & Boltovskoy, 1998), taxonomy (Renard et al., 2000; Siripattrawan et al., 2000) and physiology (Nomura et al., 2001; Nozawa et al., 2001; Vidal et al., 2001) of the genus Corbicula, no study has yet explored the potential reproductive interactions of different taxa in syntopy. Here, we will show that a syntopic population of two Corbicula morphs in the Rhine River consists of two divergent evolutionary lineages thatform cryptic hybrids in this secondary contact area. We have tackled this issue, using morphometrics, DAF-fingerprinting, sequencing and RFLP analyses from mitochondrial COI and nuclear ITS1 loci. In the second part of the study, we have analysed COI and 16SrDNA sequences in order to investigate the phylogenetic and evolutionary context of the hybridization. In order to achieve this goal, we have tried to discern the taxonomic relationships of the identified evolutionary lineages and their geographical origin.

Specimen studied

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Specimen studied
  6. DNA-isolation
  7. Sequence analyses
  8. Morphometrics of RHA individuals
  9. DNA Amplification Fingerprinting from RHA individuals
  10. Nested clade analysis of RHA population
  11. PCR-RFLPs from RHA individuals
  12. Karyograms from RHA individuals
  13. Test for statistical associations between data sets from RHA population
  14. Phylogenetic analyses of worldwide Corbicula samples
  15. Results
  16. Morphometrics
  17. COI haplotypes
  18. DAF-variation
  19. ITS1 RFLPs
  20. Associations of morphotypes, multilocus genotypes, haplotype clades and RFLP pattern
  21. Karyotype analysis of population RHA
  22. Mitochondrial sequence variation and phylogenetic relations in world wide Corbicula
  23. Discussion
  24. Reproductive isolation and detection of hybridization
  25. Taxonomic, phylogenetic and phylogeographical considerations of the species status in Corbicula
  26. Conclusions
  27. References

In the river Rhine at km 492, two morphotypes of C. spec. occurred syntopical. This population will be called hereafter RHA. Two hundred and fourteen individuals were collected randomly over a range of 1 km with a power shovel boat. Individuals from both morphotypes were found in the same shovel, thus confirming the finding ofMeister (1997) that they can be found in immediate proximity to each other in the sediment.

To reveal the taxonomic status of the morphotypes andtheir geographical origin, additional specimen were sampled or obtained from sampling sites in Germany, France, Israel, USA, China and Japan (Table 1a).

Table 1.  Material used in the study. (a) Specimen sampled for this study. Location of the sampling site from North to South, geographical position in decimal degrees, country, number of specimen used for COI and 16SrDNA analysis, respectively, are given. (b) Sequences available in GenBank. Taxonomic assignment, country, number of COI and 16S sequences, respectively, and GenBank accession numbers are provided.
LocationLatitude/longitudeCountryNCOIN16S
(a) Field samples
 Weser km 36053,047°NGermany 3 3
8,879°E   
 Moselle km 250,374°NGermany 1 
7,583°E   
 Moselle km 2150,320°NGermany 2 
7,469°E   
 Rhine km 56050,157°NGermany 2 2
7,691°E   
 Rhine km 50650,040°NGermany 2 2
8,230°E   
 Rhine km 492 RHA49,967°NGermany2113
8,350°E   
 Moselle km 14149,859°NGermany 3 3
6,932°E   
 Moselle km 16649,804°NGermany 2 2
6,854°E   
 Seine near48,413°NFrance 4 
 Fontainebleau2,705°E   
 Saône near46,315°NFrance 4 
 Mâcon4,823°E   
 Rhône near 45,739°NFrance 2 2
 Lyon4,834°E   
 Shennandoah 39,138°NUSA 3 3
 River Basin,77,888°W   
 Virginia Lake  near Tokio35,777°NJapan 4 4
139,418°E   
 Lake Genezareth32,812°NIsrael 5 5
35,604°E   
 Hong Kong22,253°NChina 3 3
113,917°E   
Taxonomic assignmentCountryNCOIN16SAccession no.
(b) Sequences from GenBank
 US formAUSA1 AF 196281
 US formBUSA2 AF 196279; AF 196278
C. africanaAfrika1 1AF 196275; AF152022
C. australisAustralia1 1AF 196274; AF 152023
C. sandaiJapan2 AF 196272–73;
C. japonicaJapan1 AF 196271
 Thailand1 AF 196270
 Korea1 AF 196269
C. leanaJapan1 AF 196268
C. sp.France1 AF 269095
C. fluminalisFrance1 AF 269096–98
C. flumineaFrance1 AF 269090–93
Neocorbicula limosaArgentina1 1AF 196277; AF 152025

Sequence analyses

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Specimen studied
  6. DNA-isolation
  7. Sequence analyses
  8. Morphometrics of RHA individuals
  9. DNA Amplification Fingerprinting from RHA individuals
  10. Nested clade analysis of RHA population
  11. PCR-RFLPs from RHA individuals
  12. Karyograms from RHA individuals
  13. Test for statistical associations between data sets from RHA population
  14. Phylogenetic analyses of worldwide Corbicula samples
  15. Results
  16. Morphometrics
  17. COI haplotypes
  18. DAF-variation
  19. ITS1 RFLPs
  20. Associations of morphotypes, multilocus genotypes, haplotype clades and RFLP pattern
  21. Karyotype analysis of population RHA
  22. Mitochondrial sequence variation and phylogenetic relations in world wide Corbicula
  23. Discussion
  24. Reproductive isolation and detection of hybridization
  25. Taxonomic, phylogenetic and phylogeographical considerations of the species status in Corbicula
  26. Conclusions
  27. References

Two mitochondrial loci (subunit I of the cytochrome oxidase, COI and 16S rDNA) were amplified by PCR with standard universal primers. Amplification was performed in 12.5 mL total reaction volume with standard reaction conditions. Samples were amplified for 10 cycles (92 °C for 50 s, 44 °C for 50 s and 72 °C for 40 s) and 36 cycles (92 °C for 30 s, 48 °C for 40 s and 72 °C for 40 s) after initial incubation of 90 °C for 2 min 30 s. The nuclear ITS1 locus was amplified with primers taken from Armbruster et al., (2000). Amplification was performed with 40 cycles (92 °C for 1 min, 48 °C for 1 min, and 72 °C for 1 min 30 s).

Both strands of the purified COI and 16S rDNA amplification products were directly cycle-sequenced with the Perkin Elmer Taq DyeDeoxy™ Terminator Cycle Sequencing Kit after the protocol of the supplier and read automatically on the ABI Prism 377® sequencing device. Obtained Sequences (Acc. nos. AYO97262–AYO97315) and retrieved Corbicula sequences from GeneBank (see Table 1) were aligned with Clustal W (Thompson et al., 1994) and adjusted by eye. Obtained sequences were collapsed (i.e. identical sequences from different individuals were pooled) to haplotypes with the Collapse 1.0 tool provided by David Posada (http://bioag.byu.edu/zoology/crandall_lab) and numbered consecutively (H1–H41). Sequence variation within sampling site RHA was assessed with the DNAsp 3.51 (Rozas & Rozas, 1999).

Morphometrics of RHA individuals

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Specimen studied
  6. DNA-isolation
  7. Sequence analyses
  8. Morphometrics of RHA individuals
  9. DNA Amplification Fingerprinting from RHA individuals
  10. Nested clade analysis of RHA population
  11. PCR-RFLPs from RHA individuals
  12. Karyograms from RHA individuals
  13. Test for statistical associations between data sets from RHA population
  14. Phylogenetic analyses of worldwide Corbicula samples
  15. Results
  16. Morphometrics
  17. COI haplotypes
  18. DAF-variation
  19. ITS1 RFLPs
  20. Associations of morphotypes, multilocus genotypes, haplotype clades and RFLP pattern
  21. Karyotype analysis of population RHA
  22. Mitochondrial sequence variation and phylogenetic relations in world wide Corbicula
  23. Discussion
  24. Reproductive isolation and detection of hybridization
  25. Taxonomic, phylogenetic and phylogeographical considerations of the species status in Corbicula
  26. Conclusions
  27. References

All left (anterior) shells were photographed with a digital camera from the top and side, always including a ruler onthe image. The shape of the left shell was quantified by elliptic Fourier approximation as described by Kuhl & Giardana (1982). This technique consists of decomposing aclosed contour curve in a two-dimensional plane into asum of harmonically related sequences. The program tpsDIG (Rohlf, 1999) was used to apply 60 landmarks on the outline of each view. These landmarks were used to produce a closed outline curve. Fourier decompositions are sensitive to location, size and orientation of objects. We have used the top of the whorl and the most extreme point on the opposite side of the shell as two homologueous points to obtain a normalization of orientation. The shells were then rotated along the resulting axis, centred and normalized for size. The decomposition into Fourier series was computed with EFAWin (Isaev & Denisova, 1995), using the algorithms of Ferson et al. (1985). The application of seven harmonics was sufficient to reproduce the outline with high accuracy. Additionally, measurements of maximum height, maximum breadth and maximum length for each shell were recorded. We assessed variation in shell sculpture characters as mean distance between ribs, number of ribs per centimetre and regularity of the spacing as the coefficient of variation of inter-rib distances. A Principal Component Analysis of all morphologic data was calculated with STATISTICA vers. 5.5 (StatSoft, 1995).

DNA Amplification Fingerprinting from RHA individuals

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Specimen studied
  6. DNA-isolation
  7. Sequence analyses
  8. Morphometrics of RHA individuals
  9. DNA Amplification Fingerprinting from RHA individuals
  10. Nested clade analysis of RHA population
  11. PCR-RFLPs from RHA individuals
  12. Karyograms from RHA individuals
  13. Test for statistical associations between data sets from RHA population
  14. Phylogenetic analyses of worldwide Corbicula samples
  15. Results
  16. Morphometrics
  17. COI haplotypes
  18. DAF-variation
  19. ITS1 RFLPs
  20. Associations of morphotypes, multilocus genotypes, haplotype clades and RFLP pattern
  21. Karyotype analysis of population RHA
  22. Mitochondrial sequence variation and phylogenetic relations in world wide Corbicula
  23. Discussion
  24. Reproductive isolation and detection of hybridization
  25. Taxonomic, phylogenetic and phylogeographical considerations of the species status in Corbicula
  26. Conclusions
  27. References

We used DNA Amplification Fingerprinting (DAF) to investigate the genomic variability of the population RHA. The method produces dominant marker bands from anonymous nuclear loci. The DAF technique is similar to RAPD-fingerprinting, but the bands can be reproduced with higher reliability due to the higher primer concentration (Caetano-Anollés et al., 1991). DAF analyses were carried out with 5 and 10 ng template DNA. The reaction was performed twice from the same DNA preparation in 12.5 μL volume in a thermocycler. Reaction conditions: 1.25 μL amplification buffer, 25 μmol of each dNTP, 2.5 mmol MgCl2, 20 pmol primer and 0.3 U Taq-polymerase (Quiagen). The PCR program was 95 °C/2.30 min, 40 cycles with 92 °C/20 s, 40 °C/15 s, 72 °C/30 s (ramp: 3 s/1 °C). We applied seven randomly chosen DAF primers (270-03, 460-03, 460-08, 470-03, 470-07, 470-10, C19).

Digital images of the ethidium-bromide fluorescence of DAF patterns on the 1.4% agarose gels were reversed and scored for the presence or absence of bands. The size of the amplified DAF fragments was assessed with the RFLP Analysis Application, vers. 2.01 (Scanalytics, Csp Inc.) by comparison with a length standard (100 bp ladder). Bands were scored only when they could be reproduced with two different DNA template concentrations in two amplifications. Four bands, not reproducible in all individuals were excluded from subsequent analyses. To keep different gels comparable, at least two fingerprints already scored on different previous gels were loaded on subsequent gels. For statistical analyses, the data were transformed into a binary matrix, where presence of a band was coded as 1 and absence as 0.

We subjected the presence/absence matrix to a Principal Component Analysis (PCA). This statistical technique achieves an ordination of the individuals according to the presence or absence of certain bands along principal component axes. These principal components are uncorrelated to each other and account, in decreasing order, for portions of the total DAF variation.

Nested clade analysis of RHA population

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Specimen studied
  6. DNA-isolation
  7. Sequence analyses
  8. Morphometrics of RHA individuals
  9. DNA Amplification Fingerprinting from RHA individuals
  10. Nested clade analysis of RHA population
  11. PCR-RFLPs from RHA individuals
  12. Karyograms from RHA individuals
  13. Test for statistical associations between data sets from RHA population
  14. Phylogenetic analyses of worldwide Corbicula samples
  15. Results
  16. Morphometrics
  17. COI haplotypes
  18. DAF-variation
  19. ITS1 RFLPs
  20. Associations of morphotypes, multilocus genotypes, haplotype clades and RFLP pattern
  21. Karyotype analysis of population RHA
  22. Mitochondrial sequence variation and phylogenetic relations in world wide Corbicula
  23. Discussion
  24. Reproductive isolation and detection of hybridization
  25. Taxonomic, phylogenetic and phylogeographical considerations of the species status in Corbicula
  26. Conclusions
  27. References

The phylogenetic relationships of the COI sequences from the RHA population were estimated with a Statistical Parsimony Network (SPN, Templeton et al., 1992). The 95% probability of a parsimonious linkage between haplotypes and the resulting parsimonious haplotype network was estimated with version 1.13 of the TCS program (Clement et al., 2000). The estimated haplotype network was converted into a hierarchical series of nested clades by hand, following the rules given in Templeton (1998) and Crandall (1996). Nested Clade Analysis (NCA) is a useful technique to infer hierarchical clade categories at a low taxonomic level in an objective, statistical manner. On each clade level, these clades group the closest phylogenetic relatives together. At the intraspecific level or at the border between intra- and interspecific levels, gene-genealogies may not be accurately represented by bifurcating trees, but rather by networks (Posada & Crandall, 2001). One of the reasons is that a tree representation assumes the absence of ancestral taxa from the sample. When the taxa in question are closely related, however, ancestral haplotypes are often extant.

PCR-RFLPs from RHA individuals

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Specimen studied
  6. DNA-isolation
  7. Sequence analyses
  8. Morphometrics of RHA individuals
  9. DNA Amplification Fingerprinting from RHA individuals
  10. Nested clade analysis of RHA population
  11. PCR-RFLPs from RHA individuals
  12. Karyograms from RHA individuals
  13. Test for statistical associations between data sets from RHA population
  14. Phylogenetic analyses of worldwide Corbicula samples
  15. Results
  16. Morphometrics
  17. COI haplotypes
  18. DAF-variation
  19. ITS1 RFLPs
  20. Associations of morphotypes, multilocus genotypes, haplotype clades and RFLP pattern
  21. Karyotype analysis of population RHA
  22. Mitochondrial sequence variation and phylogenetic relations in world wide Corbicula
  23. Discussion
  24. Reproductive isolation and detection of hybridization
  25. Taxonomic, phylogenetic and phylogeographical considerations of the species status in Corbicula
  26. Conclusions
  27. References

Clade membership of COI fragments not sequenced was identified with restriction enzymes. Each of the clades identified in the SPN (see Results section) had several autapomorphies, two of which were detectable with available restriction enzymes (Hall, 1999). Clade 3–1 haplotypes had a single exclusive recognition sequence for SacI, clade 3–2 haplotypes for HpaI. To assign the COI fragments to one of the haplotype clades, 8 μL of the amplification products were cut with each enzyme according to the protocol of the supplier (Amersham Pharmacia Biotech). COI fragments from individuals with known haplotype sequence were included into analysis as control. The ITS1 fragments from individuals of the RHA population were typed for the presence or absence of Bst EII recognition sites.

The PCR-RFLP products were electrophoresed on a 1.4% agarose gel containing ethidium bromide with a noncut control fragment and a 100-bp size marker ladder. Digital images of the ethidium-bromide fluorescence on the agarose gels were inverted and scored for the presence or absence of recognition sites.

Karyograms from RHA individuals

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Specimen studied
  6. DNA-isolation
  7. Sequence analyses
  8. Morphometrics of RHA individuals
  9. DNA Amplification Fingerprinting from RHA individuals
  10. Nested clade analysis of RHA population
  11. PCR-RFLPs from RHA individuals
  12. Karyograms from RHA individuals
  13. Test for statistical associations between data sets from RHA population
  14. Phylogenetic analyses of worldwide Corbicula samples
  15. Results
  16. Morphometrics
  17. COI haplotypes
  18. DAF-variation
  19. ITS1 RFLPs
  20. Associations of morphotypes, multilocus genotypes, haplotype clades and RFLP pattern
  21. Karyotype analysis of population RHA
  22. Mitochondrial sequence variation and phylogenetic relations in world wide Corbicula
  23. Discussion
  24. Reproductive isolation and detection of hybridization
  25. Taxonomic, phylogenetic and phylogeographical considerations of the species status in Corbicula
  26. Conclusions
  27. References

Karyograms were prepared from 62 individuals of the RHA population, following the protocol of Kligerman & Bloom (1977), with modifications. Living mussels were treated for 3 h with 0.05% colchizin w/v. Mantle and genital tissue was extracted and kept for 35 min in aqua dest. The tissue was transferred into Carnoy's solution (three parts 99% Ethanol, one part acetic acid) and kept there for 30 min, renewing the solution three times. The sample was then ground in 50% acetic acid and kept for 10 min shaking. The cell solution was dropped from 50 cm distance on an SDS coated microscope slide. After drying, the slides were stained for 20 min in a 2% Giemsa solution (manufacturer), rinsed with aqua dest and dried. Prior to microscopical inspection, the slides were embedded in Entelan.

Phylogenetic analyses of worldwide Corbicula samples

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Specimen studied
  6. DNA-isolation
  7. Sequence analyses
  8. Morphometrics of RHA individuals
  9. DNA Amplification Fingerprinting from RHA individuals
  10. Nested clade analysis of RHA population
  11. PCR-RFLPs from RHA individuals
  12. Karyograms from RHA individuals
  13. Test for statistical associations between data sets from RHA population
  14. Phylogenetic analyses of worldwide Corbicula samples
  15. Results
  16. Morphometrics
  17. COI haplotypes
  18. DAF-variation
  19. ITS1 RFLPs
  20. Associations of morphotypes, multilocus genotypes, haplotype clades and RFLP pattern
  21. Karyotype analysis of population RHA
  22. Mitochondrial sequence variation and phylogenetic relations in world wide Corbicula
  23. Discussion
  24. Reproductive isolation and detection of hybridization
  25. Taxonomic, phylogenetic and phylogeographical considerations of the species status in Corbicula
  26. Conclusions
  27. References

From the worldwide samples, 40 COI and 33 16S rDNA sequences were gained (see Table 1a). All 15 COI and 3 16S rDNA sequences of Corbicula species available in GeneBank (Renard et al., 2000; Siripattrawan et al., 2000; L.R. Cooley and D. O'Foighil, unpublished) were integrated into the analyses (Table 1b).

The phylogeny of all available COI sequences alone and joint COI and 16S rDNA sequences was inferred using Maximum Parsimony (MP) as optimality criterion. For the MP analyses, insertions/deletions were treated as a fifth state. Heuristic searches were conducted with 10 random sequence addition replicates. Sequences from a member of the same family, Neocorbicula limosa (Corbiculidae) were used as outgroup. Nodal support was estimated using the Bootstrap approach (Felsenstein, 1985) with 1000 replicates. The analyses were performed with paup 4.08b (Swofford, 1998).

Morphometrics

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Specimen studied
  6. DNA-isolation
  7. Sequence analyses
  8. Morphometrics of RHA individuals
  9. DNA Amplification Fingerprinting from RHA individuals
  10. Nested clade analysis of RHA population
  11. PCR-RFLPs from RHA individuals
  12. Karyograms from RHA individuals
  13. Test for statistical associations between data sets from RHA population
  14. Phylogenetic analyses of worldwide Corbicula samples
  15. Results
  16. Morphometrics
  17. COI haplotypes
  18. DAF-variation
  19. ITS1 RFLPs
  20. Associations of morphotypes, multilocus genotypes, haplotype clades and RFLP pattern
  21. Karyotype analysis of population RHA
  22. Mitochondrial sequence variation and phylogenetic relations in world wide Corbicula
  23. Discussion
  24. Reproductive isolation and detection of hybridization
  25. Taxonomic, phylogenetic and phylogeographical considerations of the species status in Corbicula
  26. Conclusions
  27. References

The scatterplot of the factor scores on the first two principal component axes of morphological variables revealed two distinct clusters with few intermediate individuals (Fig. 1, middle part). A two modal frequency distribution of the factor scores of the first axis suggested the presence of two independent normal distributions (Fig. 1, lower part). The simultaneous parameter estimation (means and standard deviations) for the mixed distributions with an EM algorithm (Hasselblad, 1966) implemented in the NOCOM program (Ott, 1979), allowed the fitting of two normal distributions and the assignment of upper and lower 95% confidence intervals for each of the detected clusters. Individuals with factor scores smaller than the upper 95% confidence interval of the left distribution had a rather round shape and were assigned the morphotype ‘Round’ (R). Individuals right of the lower 95% cf. of the right distribution exhibited a shape reminding of a bicycle saddle and were assigned to the morphotype ‘Saddle’ (S). Individuals not falling within these limits were regarded to belong to the ‘Intermediate’ (I) morphotype (Fig. 1, lower part). Only 7% of the individuals fell into this ‘I’ category, 55% were ‘R’ and 38% corresponded to the ‘S’ morphotype. The upper part of Fig. 1 shows the front and side view of representative shells of each morphotype.

image

Figure 1. Results of principal component analysis on morphologic variables of Corbicula spec. at sampling site RHA. The middle part of the figure shows the factor score scatterplot of all individuals against the first two principal component axes. Two distinct clusters with few intermediates could be distinguished, mainly separated by the first axis. The lower part presents a frequency distribution of factor scores on the first principal component axis. Two simultaneously fitted normal distributions were overimposed on the two modal distribution. The dashed lines represent the upper and lower 95% confidence interval of the left and right normal distribution, respectively. The individuals left of the upper 95% cf. of the left distribution were termed ‘Round’ morphotype (R), individuals right of the lower 95% cf. of the right distribution ‘Saddle’ morphotype (S). Individuals falling between the two confidence intervals were assigned to the ‘Intermediate’ morphotype (I). The latter were rare (7%) in the sample. The upper part shows representative individuals for the respective morphotypes.

Download figure to PowerPoint

COI haplotypes

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Specimen studied
  6. DNA-isolation
  7. Sequence analyses
  8. Morphometrics of RHA individuals
  9. DNA Amplification Fingerprinting from RHA individuals
  10. Nested clade analysis of RHA population
  11. PCR-RFLPs from RHA individuals
  12. Karyograms from RHA individuals
  13. Test for statistical associations between data sets from RHA population
  14. Phylogenetic analyses of worldwide Corbicula samples
  15. Results
  16. Morphometrics
  17. COI haplotypes
  18. DAF-variation
  19. ITS1 RFLPs
  20. Associations of morphotypes, multilocus genotypes, haplotype clades and RFLP pattern
  21. Karyotype analysis of population RHA
  22. Mitochondrial sequence variation and phylogenetic relations in world wide Corbicula
  23. Discussion
  24. Reproductive isolation and detection of hybridization
  25. Taxonomic, phylogenetic and phylogeographical considerations of the species status in Corbicula
  26. Conclusions
  27. References

The COI sequences of 21 individuals had a length of 660 bp, which is associated with a 95% confidence limit of 11 mutational steps for a parsimonious connection between haplotypes. The sequences were polymorphic at 17 sites. The average number of nucleotide differences between sequences was 6.22. Nested Clade Analysis of COI sequence variation revealed the presence of two highly divergent haplotype clades (3–1, 12 sequences; 3–2, 9 sequences) on the 3-step nesting level (Fig. 2). Haplotypes of the two major clades were separated by at least 11 mutational steps. Within clade 3–1, the average number of nucleotide differences equalled 0.50, within clade 3–2 1.15. The average number of differences among clades was 11.56.

image

Figure 2. Statistical parsimony network of COI sequence variation in population RHA with associated nested design. The nesting shows on the 3-step level two major haplotype clades (3–1, 3–2), connected by at least 11mutational steps. The size of the circles is proportional to the haplotype frequency.

Download figure to PowerPoint

The PCR-RFLP on 61 individuals gave unequivocal results in all cases: fragments that were digested with SacI were not restricted with HpaI and vice versa. We observed no case of heteroplasmy or partial digest. In total, 25 individuals belonged to haplotype clade 3–1 and 36 to haplotype clade 3–2.

Our haplotype 2 (H2) was identical to Corbiculafluminea’ sequences reported by Renard et al. (2000) (AF269091–93) and the ‘US form A’ from Siripattrawan et al. (2000) (AF196281); sequence H4 was identical to the nonidentified Corbicula species in Renard et al. (2000) (AF269095).

DAF-variation

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Specimen studied
  6. DNA-isolation
  7. Sequence analyses
  8. Morphometrics of RHA individuals
  9. DNA Amplification Fingerprinting from RHA individuals
  10. Nested clade analysis of RHA population
  11. PCR-RFLPs from RHA individuals
  12. Karyograms from RHA individuals
  13. Test for statistical associations between data sets from RHA population
  14. Phylogenetic analyses of worldwide Corbicula samples
  15. Results
  16. Morphometrics
  17. COI haplotypes
  18. DAF-variation
  19. ITS1 RFLPs
  20. Associations of morphotypes, multilocus genotypes, haplotype clades and RFLP pattern
  21. Karyotype analysis of population RHA
  22. Mitochondrial sequence variation and phylogenetic relations in world wide Corbicula
  23. Discussion
  24. Reproductive isolation and detection of hybridization
  25. Taxonomic, phylogenetic and phylogeographical considerations of the species status in Corbicula
  26. Conclusions
  27. References

Application of seven DAF primer resulted in 107 scorable bands, 95 of which were polymorphic. All individuals could be distinguished by their unique multilocus RAPD pattern. The scatterplot of factor scores on the first two principal component axes of the DAF variation ordinated the individuals into three statistically distinct clusters (Fig. 3). These clusters with no intermediates were interpreted as distinct multilocus genotypes (A, B, C). Each multilocus genotype occurred with roughly similar frequency in the 56 typed individuals (39, 32 and 29%, respectively). Figure 4 shows the band frequency of individual DAF markers for each multilocus genotype. Bands with a frequency below 10% and above 80% in the total data set were excluded from this analysis to avoid the scoring of noise or the inclusion of uninformative bands, respectively. From the remaining 70 marker bands, multilocus genotype A and C exhibited bands that did not occur in any other genotype. Genotype B had no exclusive bands; all bands were shared either with genotype A or C (Fig. 4).

image

Figure 3. Principal component analysis on DAF variation at sampling site RHA. Three multilocus genotype clusters could be distinguished (A, B, C). The 95% confidence area around each cluster is indicated by dashed lines.

Download figure to PowerPoint

image

Figure 4. Frequencies of 70 DAF markers among multilocus genotypes identified from PCA ( Fig. 3 ). The genotype B shows most of the bands that can be found in A or C, but lacks specific bands. Bands exclusive for the respective multilocus genotype can be found in the genotypes A and C.

Download figure to PowerPoint

ITS1 RFLPs

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Specimen studied
  6. DNA-isolation
  7. Sequence analyses
  8. Morphometrics of RHA individuals
  9. DNA Amplification Fingerprinting from RHA individuals
  10. Nested clade analysis of RHA population
  11. PCR-RFLPs from RHA individuals
  12. Karyograms from RHA individuals
  13. Test for statistical associations between data sets from RHA population
  14. Phylogenetic analyses of worldwide Corbicula samples
  15. Results
  16. Morphometrics
  17. COI haplotypes
  18. DAF-variation
  19. ITS1 RFLPs
  20. Associations of morphotypes, multilocus genotypes, haplotype clades and RFLP pattern
  21. Karyotype analysis of population RHA
  22. Mitochondrial sequence variation and phylogenetic relations in world wide Corbicula
  23. Discussion
  24. Reproductive isolation and detection of hybridization
  25. Taxonomic, phylogenetic and phylogeographical considerations of the species status in Corbicula
  26. Conclusions
  27. References

RFLP analysis of the ITS1 region revealed three different fragment pattern α, β and χ. Pattern β appeared to be a strictly additive combination of α and χ(Fig. 5). Twelve individuals exhibited pattern α, 11 pattern β and 12 χ. The sum of the fragment lengths was in any pattern larger than the length of the original PCR product (550 bp), suggesting the presence of at least two different ITS1 sequences in each individual.

image

Figure 5. RFLP patterns of ITS1 with Bsp EII ( α , β and γ ). Pattern β is the additive combination of α and γ . The fragment length of the original ITS1 PCR product was 550 bp. The sum of RFLP fragment lengths exceeds this value, showing that more than one ITS1 sequence was present.

Download figure to PowerPoint

Associations of morphotypes, multilocus genotypes, haplotype clades and RFLP pattern

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Specimen studied
  6. DNA-isolation
  7. Sequence analyses
  8. Morphometrics of RHA individuals
  9. DNA Amplification Fingerprinting from RHA individuals
  10. Nested clade analysis of RHA population
  11. PCR-RFLPs from RHA individuals
  12. Karyograms from RHA individuals
  13. Test for statistical associations between data sets from RHA population
  14. Phylogenetic analyses of worldwide Corbicula samples
  15. Results
  16. Morphometrics
  17. COI haplotypes
  18. DAF-variation
  19. ITS1 RFLPs
  20. Associations of morphotypes, multilocus genotypes, haplotype clades and RFLP pattern
  21. Karyotype analysis of population RHA
  22. Mitochondrial sequence variation and phylogenetic relations in world wide Corbicula
  23. Discussion
  24. Reproductive isolation and detection of hybridization
  25. Taxonomic, phylogenetic and phylogeographical considerations of the species status in Corbicula
  26. Conclusions
  27. References

In test for association among morphotypes, multilocus genotypes haplotype clades and RFLP pattern, the null hypothesis of no association could be rejected with very high probability for all tested pairwise combinations (Table 2). All possible haplotype clade/morphotype combinations could be observed, but clade 3–1 was significantly associated with morphotype R and clade 3–2 with morphotype S (Table 2a). Morphotype R co-occurred more often than expected with genotype A, genotype C was associated exclusively to morphotype S. Genotype B was the only to appear in the intermediate morphotype (I), but it occurred as well inboth R and S (Table 2b). In the test of genotype against haplotype clade, genotype A was exclusively related to clade 3–1 as well as genotype C to clade 3–2. Genotype B, however, was found with both clades (Table 2c). RFLP pattern α was strictly associated with genotype A, pattern β with genotype B and χ with C (Table 2d).

Table 2.  Frequency tables of (a) haplotype clades (row) against morphotypes (column) (b) multilocus genotypes against morphotypes (c) , multilocus genotypes against haplotype clades and (d) multilocus genotypes against ITS RFLP patterns.
 RISΣ
(a)
 3–1222325
 3–2622836
 Σ28431ΣΣ 63
inline image31.69P  < 0.000   
(b)
 A200222
 B82818
 C001616
 Σ28226ÓÓ 56
inline image43.90P  < 0.000   
 3–13–2Σ 
(c)
 A20020 
 B41418 
 C01616 
 Σ2430ΣΣ 54 
inline image55.12P  < 0.000   
 αβχΣ
(d)
 A120012
 B011011
 C001212
 Σ121112ΣΣ 35
inline image76.85P  < 0.000   

Mitochondrial sequence variation and phylogenetic relations in world wide Corbicula

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Specimen studied
  6. DNA-isolation
  7. Sequence analyses
  8. Morphometrics of RHA individuals
  9. DNA Amplification Fingerprinting from RHA individuals
  10. Nested clade analysis of RHA population
  11. PCR-RFLPs from RHA individuals
  12. Karyograms from RHA individuals
  13. Test for statistical associations between data sets from RHA population
  14. Phylogenetic analyses of worldwide Corbicula samples
  15. Results
  16. Morphometrics
  17. COI haplotypes
  18. DAF-variation
  19. ITS1 RFLPs
  20. Associations of morphotypes, multilocus genotypes, haplotype clades and RFLP pattern
  21. Karyotype analysis of population RHA
  22. Mitochondrial sequence variation and phylogenetic relations in world wide Corbicula
  23. Discussion
  24. Reproductive isolation and detection of hybridization
  25. Taxonomic, phylogenetic and phylogeographical considerations of the species status in Corbicula
  26. Conclusions
  27. References

Forty-three unique COI haplotypes were found in our sample (H1–H43). In total, 54 different COI sequences, including the Corbicula sequences available in GenBank, were analysed. From the 660 positions, 181 characters were variable (27.4%) of which 90 (13.6%) were parsimony-informative.

Heuristic search found more than 2000 equally parsimonious trees of 279 steps length on the same tree island. The bootstrap majority consensus tree is shown in Fig. 6. The trees had a consistency index of 0.778 and a retention index of 0.877. The skewness of a random tree distribution of 100 000 trees (mean length=691.5, SD=25.5, g1=−1.38, g2=2.41, outgroup excluded) provided evidence for a strong phylogenetic signal in the data (Hillis & Huelsenbeck, 1992). Adding the available 16SrDNA sequences to the respective COI sequences from the same individual yielded sequences of 1103 bp length. This procedure can be justified, because the entire mitochondrial genome can be regarded as a single gene from a phylogenetic point of view. The merging reduced the data set to 39 different mitochondrial haplotypes. Because of the sequence variability in 16S, however, some of the COI haplotypes (e.g. H2) were further differentiated. A total of 264 characters were variable (23.9%) and 113 (10.2%) parsimony-informative. The search revealed more than 2000 equally parsimonious phylogenies. The trees had a length of 375, which showed that the merged data contained a strong phylogenetic data as well, compared with a random tree distribution (mean length=51.2 SD=18.6 g1=−1.29 g2=3.86) (tree not shown).

image

Figure 6. Maximum parsimony phylogeny estimates based on COI sequences. Bootstrap values above 50% are given above the respective nodes. Sequences from GenBank are accompanied by their respective accession number (see Table 1 ). Sequences from this study are represented by their haplotype number (Hx). Haplotype clades from RHA populations ( Fig. 2 ) are indicated. Nodes separating C. japonica and C. africana from the majority of Corbicula sequences are designated by encircled numbers under the respective node.

Download figure to PowerPoint

Bootstrap values above 85% (Hillis, 1995) supported several clades in the COI phylogeny: A basal node separated a Corbicula japonica sequence (AF 196271) and other sequences from individuals sampled in Japan from all other Corbicula (Node 1 in Fig. 6). The next node suggested a split between Corbicula africana (AF 196275) and a clade containing the major part of the sequences (Node 2 in Fig. 6). Node 3 represents a clade that splits into a large polytomy (Fig. 6). This clade includes taxa to which taxonomic significance has been attributed (C.sandai, C.leana, C. australis, C. fluminea, C. fluminalis), as well as all sequences from (Renard et al., 2000; Siripattrawan et al., 2000) and this study. The clade consists of several lineages, which are more or less well supported by the bootstrap procedure (Fig. 6). Haplotype clade 3–1 is part of a lineage together with sequences from C. fluminea and US form A. Haplotype clade 3–2 is part of a separate lineage, including sequences from other sampling sites in the Rhine and from France (C. spec). The specimens from Israel, where exclusively C. fluminalis is described (Morton, 1986), form another lineage. However, it is difficult to assign unequivocal taxonomic relevance to the inferred lineages, because sequences assigned to C. leana occur in the same lineage as C.fluminea sequences. The latter can be found in different lineages, which is true for C. fluminalis assigned sequences as well (see Fig. 6).

When 16S sequences were included, the polytomy at node 3 was not resolved (tree not shown). The lineages described above were confirmed and an additional lineage from the Rhône river identified. The average sequence divergence between lineages was 0.018 ± 0.004 (mean ± SD) changes per site. The 16S sequences obtained from Japanese individuals confirmed the basal position of this lineage and suggested that they belonged to the species C. japonica. The separate position of C. africana was confirmed as well by inclusion of 16S sequence into the analysis. Phylogenies derived from maximum likelihood methods (data not shown) revealed virtually identical tree topologies.

Reproductive isolation and detection of hybridization

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Specimen studied
  6. DNA-isolation
  7. Sequence analyses
  8. Morphometrics of RHA individuals
  9. DNA Amplification Fingerprinting from RHA individuals
  10. Nested clade analysis of RHA population
  11. PCR-RFLPs from RHA individuals
  12. Karyograms from RHA individuals
  13. Test for statistical associations between data sets from RHA population
  14. Phylogenetic analyses of worldwide Corbicula samples
  15. Results
  16. Morphometrics
  17. COI haplotypes
  18. DAF-variation
  19. ITS1 RFLPs
  20. Associations of morphotypes, multilocus genotypes, haplotype clades and RFLP pattern
  21. Karyotype analysis of population RHA
  22. Mitochondrial sequence variation and phylogenetic relations in world wide Corbicula
  23. Discussion
  24. Reproductive isolation and detection of hybridization
  25. Taxonomic, phylogenetic and phylogeographical considerations of the species status in Corbicula
  26. Conclusions
  27. References

The presence of two distinctive morphotypes with few intermediate phenotypes (Fig. 1) initially suggested the co-occurence of two taxa. This corresponded well to the detection of two highly divergent haplotype clades (Fig. 2), proposing two independent evolutionary lineages. However, the clades were not exclusively associated with the respective morphotype; some individuals of the round morphotype R had mitochondrial COI haplotypes of clade 3–2 that was mostly associated with morphotype S and vice versa (Table 2a). This was suggestive of bi-directional introgression of cytoplasmic lineages via hybridization between formerly fragmented lineages and backcrossing with selection against intermediate genotypes. However, if this were true, we would have expected to find two distinctive multilocus genotypes with few intermediates.

DAF-fingerprinting, however, revealed three distinctive multilocus genotypes without intermediates (Fig. 3). Each multilocus genotype was present in approximately equal frequency (Table 2b). The absence of intermediates between multilocus genotypes (Fig. 3) suggested that backcrosses to the parental lineages were either too rare to be detected with the given sample size or even absent. All individuals could be distinguished by their unique DAF fingerprint pattern, which suggested an outcrossing breeding system. One of the genotypes (A) occurred mostly in individuals of morphotype R, whereas genotype C was exclusively found in morphotype S. Multilocus genotype B, however, occurred with equal frequency in both morphotypes and was the only one found in specimen with intermediate morphotype (I) (Table 2b). Additionally, only individuals with genotype B contained haplotypes from both haplotype clades. Genotype A was exclusively bound to haplotype clade 3–1 and genotype C to clade 3–2 (Table 2c). These findings proposed that multilocus genotype B individuals were hybrids of the parental lineages A and C. This conclusion was strengthened by the observation that genotype B individuals exhibited no exclusive bands, but appeared to be a strict combination of bands occurring in genotype A and C (Fig. 4). An a priori statistical inference of genotypic classes as proposed by Boecklen & Howard (1997) or Miller (2000) was not applicable to our data, because we did not know (i) that hybrids were present in the first place (ii) to which genotypic class the parental taxa belonged and, hence (iii) which markers were diagnostic for which parental taxon.

Conclusive evidence for genotype B individuals representing hybrids between genotype A and C arose from the strict association of restriction patterns of the ITS1 locus (Fig. 5) to the multilocus genotypes (Table 2d). The β pattern of genotype B can be interpreted as additive pattern of the α and γ fragments that were present in theparental lineage genotypes A and C, respectively. Itappeared that all genotype B individuals were heterozygous regarding this ITS1 locus. The exclusive finding of an additive ITS1-RFLP pattern in genotype B individuals indicates that they are formed as F1 hybrids only and that they are sterile among themselves. Otherwise, we would expect to find α, β and γ restriction patterns in F2 generation genotype B individuals as a result of the random distribution of alleles to the gametes during meiosis. However, additional experiments with other codominant markers are necessary to confirm this inference.

The observation that all individuals karyotyped appeared to have the same diploid chromosome number (n=36) regardless of multilocus genotype suggested a regular formation of the hybrid zygotes without pre or postzygotic polyploidization.

These findings allowed the inference that two morphologically and genetically distinct evolutionary lineages co-occurred. The lineage with the rounded morphotype (R) was associated with haplotype clade 3–1, the multilocus genotype A and the ITS1 restriction site pattern α. Wewill call it hereafter lineage 1. The lineage with the saddle-like appearance (S) was linked with clade 3–2, genotype B and RFLP pattern γ (lineage 2). However, the two lineages formed F1 hybrids (genotype B, RFLP pattern β), resembling phenotypically either lineage 1 or lineage 2. We have termed this ‘cryptic’ hybridization, because it is morphologically not apparent, as the hybrids mostly resemble either one or the other parent species. The detection of both haplotype clades in the hybrids indicates that both parental lineages could act as female.

Hybrids found were apparently F1 generation only. This suggests a sterility of the hybrids. However, given the relative low number of hybrids identified, we cannot exclude the possibility of rare backcrosses. The finding of two individuals with genotype A, but saddle-shaped shells (S) may indicate the introgression of lineage 2 genes into lineage 1 (Table 2b).

The inference of cryptic hybridization in a syntopic population of Corbicula morphotypes is the necessary basis for further studies in progress. One issue currently addressed is the lack of intermediate phenotypes. Considering only genotype B individuals in morphological analysis, it becomes apparent that they occupy more or less continuously the entire morphospace, except for the most extreme regions of the parent taxa. The intermediates are, however, relatively rare in comparison with individuals that fall into the range of the morphological variation of the parental taxa. This suggests that cross mating of the parent taxa can produce almost the entire range of phenotypes in the hybrids, but intermediates do not or only rarely reach the adult stage. We hypothesize that the two morphotypes present adaptive peaks and the virtual absence of the intermediate phenotype is due to selection against it and not against the hybrid genotype per se. The study of cytonuclear disequilibrium (Arnold, 1993; Scribner et al., 2001) may reveal the evolutionary forces and dynamics of this phenomenon.

Taxonomic, phylogenetic and phylogeographical considerations of the species status in Corbicula

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Specimen studied
  6. DNA-isolation
  7. Sequence analyses
  8. Morphometrics of RHA individuals
  9. DNA Amplification Fingerprinting from RHA individuals
  10. Nested clade analysis of RHA population
  11. PCR-RFLPs from RHA individuals
  12. Karyograms from RHA individuals
  13. Test for statistical associations between data sets from RHA population
  14. Phylogenetic analyses of worldwide Corbicula samples
  15. Results
  16. Morphometrics
  17. COI haplotypes
  18. DAF-variation
  19. ITS1 RFLPs
  20. Associations of morphotypes, multilocus genotypes, haplotype clades and RFLP pattern
  21. Karyotype analysis of population RHA
  22. Mitochondrial sequence variation and phylogenetic relations in world wide Corbicula
  23. Discussion
  24. Reproductive isolation and detection of hybridization
  25. Taxonomic, phylogenetic and phylogeographical considerations of the species status in Corbicula
  26. Conclusions
  27. References

The inference of a hybridization raises several questions: (i) at which taxonomic level occurred the hybridization (ii) how large is the evolutionary divergence of the lineages involved and (iii) – in case of hybridization in a secondary contact area – where is the geographical origin of the parental lineages? All this was not known in the case of Corbicula. We have therefore used samples from potential source areas all over the world to infer the phylogenetic relation of the lineages. We have integrated previous work on the subject (Renard et al., 2000; Siripattrawan et al., 2000) to obtain a taxonomic assignment of the two lineages found in the Rhine River.

However, this venture proved to be difficult. Comparison with available sequences in GenBank revealed that haplotypes from clade 3–1 corresponded to C. fluminea as defined by Renard et al. (2000), respectively, US form A (Siripattrawan et al., 2000). It seems therefore likely thatthis lineage invaded the European rivers with cargo from North America, as proposed by several authors (Kinzelbach, 1991; Araujo et al., 1993). Clade 3–2 haplotypes correspond to Renard et al. (2000) C. spec. and not to C. fluminalis as previously thought. Neither the geographical origin nor the potential invasion routes of this evolutionary lineage could be revealed by this study. Perhaps the integration of more material from the Near East and the Caucasian region (Illies, 1978) in future studies can help to identify the source of these Corbicula lineages.

The polytomy at node three (Fig. 6) suggested a simultaneous divergence of the lineages. Power analysis (Walsh et al., 1999; Braun & Kimball, 2001; Walsh & Friesen, 2001) indicated that the combined COI and 16S data set of 1103 bases should have been sufficient to resolve sequential branchings that occurred during an interval of at least 7.2% (±2.2%) of the total divergence time between the lineages. This indicates that the radiation of the Corbicula lineages indeed occurred during a relatively short time interval, if not simultaneously, and is not due to a poor resolution caused by a lack of data. Applying the molecular clock rates estimated for the mitochondrial 16S rDNA by Canapa et al. (1996) for several Veneridae (Bivalvia) of 0.14–0.36% sequence divergence per million years, we could place the divergence among the lineages, with all caution, into the early Pleistocene. During the Pleistocene, however, the genus Corbicula was widespread in Europe as evidenced by fossils (Meijer & Preece, 2000), although exact species identification is difficult as the present taxonomic problems illustrate. Even if East Asian lineages are involved and the introduction is mainly because of anthropogenic dispersal, the presence of Corbicula in European stream systems should perhaps less be seen as an invasion of an exotic species, but rather as a postglacial recolonization of a formerly occupied range (Preece, 1997).

The taxonomy of the genus is the focus of a continuing debate (Britton & Morton, 1986; Woodruff et al., 1993). Several dozens of species have been described for Asia and Africa, until a revision of (Morton, 1986) reduced the worldwide number of species to two. However, not all authors follow this interpretation, claiming that more species exist (Park et al., 2000; Siripattrawan et al., 2000). The latter conclusion is supported by the present study, suggesting that at least C. japonica and C. africana constitute separate evolutionary entities.

Related to the taxonomic confusion is the question of the breeding system. Some authors report for C.fluminea (O.F. Müller 1774) and/or C. fluviatilis (O.F. Müller 1774) that the studied populations consisted more or less completely of hermaphroditic individuals (Britton & Morton, 1986; Morton, 1986; Meister, 1997) whereas inother populations separate sexes seem to prevail (Kraemer & Galloway, 1986; Morton, 1987). Widely varying outcrossing rates, from complete selfing to obligate outcrossing have been described (Britton & Morton, 1986; Kraemer & Galloway, 1986). Park et al. (2000) claim that triploid populations (n=54) exist. However, it has to be kept in mind that the species identification is difficult and equivocal, leaving the possibility that different studies have in fact worked with different evolutionary entities. The recent spread of several lineages and the potential to form hybrids in secondary contact zones may have contributed to thetaxonomic uncertainties that have hampered the attempts to fit local populations into the scheme of existing taxonomic names. Considering their potentially incomplete reproductive isolation we think that the Corbicula lineages radiating from node 3 in a short time frame (Fig. 6) should perhaps rather be considered as an incipient species group than several well-defined species. The taxonomic and biogeographical problems still posed by these Corbicula lineages demand for the integration of more material from the whole genus range in future studies.

Conclusions

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Specimen studied
  6. DNA-isolation
  7. Sequence analyses
  8. Morphometrics of RHA individuals
  9. DNA Amplification Fingerprinting from RHA individuals
  10. Nested clade analysis of RHA population
  11. PCR-RFLPs from RHA individuals
  12. Karyograms from RHA individuals
  13. Test for statistical associations between data sets from RHA population
  14. Phylogenetic analyses of worldwide Corbicula samples
  15. Results
  16. Morphometrics
  17. COI haplotypes
  18. DAF-variation
  19. ITS1 RFLPs
  20. Associations of morphotypes, multilocus genotypes, haplotype clades and RFLP pattern
  21. Karyotype analysis of population RHA
  22. Mitochondrial sequence variation and phylogenetic relations in world wide Corbicula
  23. Discussion
  24. Reproductive isolation and detection of hybridization
  25. Taxonomic, phylogenetic and phylogeographical considerations of the species status in Corbicula
  26. Conclusions
  27. References

The integrated approach using morphometric and molecular genetic methods revealed the coexistence of two evolutionary distinct lineages in a syntopic population in the River Rhine. These lineages formed cryptic hybrids that exhibited the phenotype of the parental lineages; intermediate phenotypes, however, were rare.

The initiation of hybrid studies with or without molecular markers often has been based on meristic or morphological criteria solely. As the presented example shows, such an approach may lead to a serious underestimation of the extent of natural hybridization, due to a lack of intermediate phenotypes.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Materials and methods
  5. Specimen studied
  6. DNA-isolation
  7. Sequence analyses
  8. Morphometrics of RHA individuals
  9. DNA Amplification Fingerprinting from RHA individuals
  10. Nested clade analysis of RHA population
  11. PCR-RFLPs from RHA individuals
  12. Karyograms from RHA individuals
  13. Test for statistical associations between data sets from RHA population
  14. Phylogenetic analyses of worldwide Corbicula samples
  15. Results
  16. Morphometrics
  17. COI haplotypes
  18. DAF-variation
  19. ITS1 RFLPs
  20. Associations of morphotypes, multilocus genotypes, haplotype clades and RFLP pattern
  21. Karyotype analysis of population RHA
  22. Mitochondrial sequence variation and phylogenetic relations in world wide Corbicula
  23. Discussion
  24. Reproductive isolation and detection of hybridization
  25. Taxonomic, phylogenetic and phylogeographical considerations of the species status in Corbicula
  26. Conclusions
  27. References
  • Araujo, R., Moreno, D. & Ramos, M.A. 1993. The Asiatic Clam Corbicula fluminea (Müller 1774) (Bivalvia: Corbiculidae) in Europe. Am. Malacol. Bull. 10: 3949.
  • Armbruster, G.F.J., Van Moorsel, C.H.M. & Gittenberger, E. 2000. Conserved sequence patterns in the non-coding ribosomal ITS-1 of distantly related snail taxa. J. Molluscan Studies 66: 570573.
  • Arnold, M.L. 1992. Natural hybridization as an evolutionary process. Annu. Rev. Ecol. Syst 23: 237261.
  • Arnold, J. 1993. Cytonuclear disequilibria in hybrid zones. Annu. Rev. Ecol. Syst 24: 521554.
  • Bahl, A. & Pfenninger, M. 1996. A rapid method of DNA isolation using laundry detergent. Nucl. Acids Res. 24: 15871588.
  • Barton, N.H. 2001. The role of hybridization in evolution. Molec. Ecol. 10: 551568.
  • Boecklen, W.J. & Howard, D.J. 1997. Genetic analysis of hybrid zones: numbers of markers and power of resolution. Ecology 78: 26112616.
  • Braun, E.L., Kimball, R.T. 2001. Polytomies, the power of phylogenetic inference, and the stochastic nature of molecular evolution: a comment on Walsh et al. (1999). Evolution 55: 12611263.
  • Britton, J.C. & Morton, B. 1986. Polymorphism in Corbicula fluminea (Bivalvia: Corbiculoidea) from North America. Malacol. Rev. 19: 143.
  • Caetano-Anollés, G., Bassam, B.J. & Gresshoff, P. 1991. DNA Amplification fingerprinting: a strategy for genome analysis. Plant Mol. Biol. Reporter 9: 294307.
  • Canapa, A., Marota, I., Rollo, F. & Olmo, E. 1996. Phylogenetic analysis of Veneridae (Bivalvia): comparison of molecular and palaeontological data. J. Mol. Evol. 43: 517522.
  • Cataldo, D. & Boltovskoy, D. 1998. Population dynamics of Corbicula fluminea (Bivalvia) in the Parana River Delta (Argentina). Hydrobiologia 380: 153163.
  • Clement, M., Posada, D. & Crandall, K.A. 2000. TCS: a computerprogram to estimate gene genealogies. Mol. Ecol. 9: 16571659.DOI: 10.1046/j.1365-294X.2000.01020.x
  • Counts, C.L.I. 1980. The genus Corbicula (Bivalvia: Corbiculidae) in Africa and South America: zoogeographic and taxonomic problems. Bull. Am. Malacol. Union 46: 7172.
  • Counts, C.L.I. 1981. Corbicula fluminea (Bivalvia: Corbiculidae) in British Colombia. Nautilus 95: 1213.
  • Counts, C.L.I. 1986. The zoogeography and history of the invasion of the United States by Corbicula fluminea (Bivalvia: Corbiculidae). Am. Malacol. Bull. Special Edition No. 2: 739.
  • Crandall, K.A. 1996. Multiple interspecies transmissions of human and simian T-cell leukemia/lymphoma virus type I sequences. Mol. Biol. Evol 13: 115131.
  • Felsenstein, J. 1985. Confidence limits on phylogenies using the bootstrap. Evolution 39: 783791.
  • Ferson, S.F., Rohlf, F.J. & Koehn, R.K. 1985. Measuring shape variation of two-dimensional outlines. Syst. Zool. 34: 5968.
  • Hall, T.A. 1999. BioEdit: a user-friendly biological sequence alignment editor and analysis program for Windows 95/98/NT. Nucl. Acids. Symp. Ser. 41: 9598.
  • Harrison, R.G. 1990. Hybrid zones: windows on evolutionary process. Oxf. Surv. Evol. Biol. 7: 69128.
  • Den Hartog, C., Van Den Brink, F.W.B. & Van Der Velde, G. 1992. Why was the invasion of the river Rhine by Corophium curvispinum and Corbicula species so successful? J. Natural History 26: 11211129.
  • Hasselblad, V. 1966. Estimation of parameters for a mixture of normal distributions. Technometrics 8: 431444.
  • Hillis, D.M. 1995. Approaches for assessing phylogenetic accuracy. Syst. Biol. 44: 316.
  • Hillis, D.M. & Huelsenbeck, J.P. 1992. Signal, noise, and reliability in molecular phylogenetic analyses. J. Heredity 83: 189195.
  • Illies, J.H. 1978. Limnofauna Europaea. Gustav Fischer Verlag Stuttgart 532S.
  • Isaev, M.A. & Denisova, L.N. 1995. The Computer Programs for Shape Analysis of Plant Leaves. Proceedings of the ‘Mathematics. Computer. Education’ International Conference, Pushchino.
  • Ituarte, C.F. 1981. Primera noticia acerca de la introduccion de pelecipods asiaticos en al area Rioplatense. Neotropica 27: 7982.
  • Ituarte, C.F. 1994. Corbicula and Neocorbicula (Bivalvia, Corbiculidae) in the Parana, Uruguay, and Rio-De-La-Plata Basins. Nautilus 107: 129135.
  • Jiggins, C.D. & Mallet, J. 2000. Bimodal hybrid zones and speciation. Trends Ecol. Evol. 15: 250255.
  • Kinzelbach, R. 1991. Die Körbchenmuscheln Corbicula fluminalis, Corbicula fluminea und Corbicula fluviatilis in Europa (Bivalvia: Corbiculidae). Mainzer Naturwissenschaftliches Arch. 29: 215228.
  • Kligerman, A.D. & Bloom, S.E. 1977. Rapid chromosome preparation from solid tissue of fishes. J. Fish. Res. Board Canada 34: 266269.
  • Kraemer, L.R. & Galloway, M.L. 1986. Larval development of Corbicula fluminea (Müller) (Bivalvia: Corbiculacea): an appraisal of its heterchrony. Am. Malacol. Bull. 4: 6179.
  • Kuhl, F.P. & Giardana, G.R. 1982. Elliptic Fourier features of a closed contour. Comp. Graph. Image Processing 18: 236258.
  • Lee, T.Y. & Chung, E.Y. 1980. Reproductive cycle of the marsh clam Corbicula fluminea (Müller). Publications Inst. Mar. Sci. University Busan 12: 4754.
  • McMahon, R.F. & Williams, C.J. 1986. A reassessment of growth rate, life span, life cycles and population dymamics in a natural population and field caged individuals of Corbicula fluminea (Müller) (Bivalvia: Corbiculacea). Am. Malacol. Bull. Special Edition No. 2: 151166.
  • Meijer, T. & Preece, R.C. 2000. A review of the occurrence of Corbicula. the Pleistocene of North-West Europe. Geologie En Mijnbouw-Netherlands J. Geosciences 79: 241255.
  • Meister, A. 1997. Lebenszyklus, Autökologie und Populationsökologie der Körbchenmuscheln Corbicula fluminea und Corbicula fluminalis (Bivalvia, Corbiculidae) im Inselrhein. Schriftenreihe des HLfU 238.
  • Miller, L.M. 2000. Classifying genealogical origins in hybrid populations using dominant markers. J. Heredity 91: 4649.
  • Morton, B. 1986. Corbicula in Asia – an updated synthesis. Am. Malacol. Bull. Special Edition No. 2: 113124.
  • Morton, B. 1987. Polymorphism in Corbicula fluminea (Bivalvia: Corbiculoidea) from Hong Kong. Malacol. Rev. 20: 105127.
  • Mouthon, J. 1981. Sur la presence in France et en Portugal de Corbicula (Bivalvia, Corbiculidae) orginaire de Asie. Basteria 45: 109116.
  • Nomura, T., Yamamoto, I., Morishita, F., Furukawa, Y. & Matsushima, O. 2001. Purification and some properties of alanine racemase from a bivalve mollusc Corbicula japonica. J.Exp Zool. 289: 19.
  • Nozawa, H., Mori, T. & Seki, N. 2001. Different effects of NaCl on activities of transglutaminases from marine and freshwater shellfish muscles. Fish Sci. 67: 383385.
  • Ott, J. 1979. Maximum likelihood estimation by counting methods under polygenic and mixed models in human pedigrees. Am. J. Hum Genet 42: 255257.
  • Park, G.M., Yong, T.S., Im, K.I. & Chung, E.Y. 2000. Karyotypes of three species of Corbicula (Bivalvia: Veneroida) in Korea. J. Shellfish Res. 19: 979982.
  • Posada, D. & Crandall, K.A. 2001. Intraspecific gene genealogies: trees grafting into networks. Trends Ecol. Evol. 16: 3745.
  • Preece, R.C. 1997. The spatial response of non-marine molluscs to past climate changes. NATO ASI Series I Global Environ. Change 47: 163177.
  • Renard, E., Bachmann, V., Cariou, M.L. & Moreteau, J.C. 2000. Morphological and molecular differentiation of invasive freshwater species of the genus Corbicula (Bivalvia, Corbiculidea) suggest the presence of three taxa in French rivers. Mol Ecol. 9: 20092016.DOI: 10.1046/j.1365-294X.2000.01104.x
  • Rohlf, F.J. 1999. Shape statistics: procrustes superimpositions and tangent spaces. J. Classification 16: 197223.
  • Rozas, J. & Rozas, R. 1999. DnaSp, Version 3: an integrated program for molecular population genetics and molecular evolution analysis. Bioinformatics 15: 174175.
  • Sambrook, J., Fritsch, E.F. & Maniatis, T. 1989. Molecular Cloning: A Laboratory Manual. Cold Spring Harbour Laboratory, Cold Spring Harbour.
  • Scribner, K.T., Page, K.S. & Bartron, M.L. 2001. Hybridization infreshwater fishes: a review of case studies and cytonuclearmethods of biological inference. Rev. Fish Biol. Fish 10: 293323.
  • Siripattrawan, S., Park, J.K. & Foighil, D. 2000. Two lineages of the introduced Asian freshwater clam Corbicula occur in North America. J. Molluscan Studies 66: 423429.
  • Sokal, R.R. & Rohlf, F.J. 1981. Biometry, 2nd edn. Freeman, New York.
  • Stites, D.L., Benke, A.C. & Gillespie, D.M. 1995. Population-dynamics, growth, and production of the Asiatic clam, Corbicula fluminea, in a Blackwater River. Can J. Fish Aquat Sci. 52: 425437.
  • Swofford, D.L. 1998. PAUP*. Phylogenetic Analysis Using Parsimony (*and Other Methods). Sinauer, Sunderland.
  • Templeton, A.R. 1998. Nested clade analyses of phylogeographic data: testing hypotheses about gene flow and population history. Mol Ecol. 7: 381397.DOI: 10.1046/j.1365-294x.1998.00308.x
  • Templeton, A.R., Crandall, K.A. & Sing, C.F. 1992. A cladisticanalysis of phenotypic associations with haplotypes inferred from restriction endonuclease mapping and DNA sequence data. III. Cladogram estimation. Genetics 132: 619633.
  • Thompson, J.D., Higgins, D.G. & Gibson, T.J. 1994. Clustal-W – improving the sensitivity of progressive multiple sequence alignment through sequence weighting, position-specific gappenalties and weight matrix choice. Nucl. Acids Res. 22: 46734680.
  • Vidal, M.L., Basseres, A. & Narbonne, J.F. 2001. Potential biomarkers of trichloroethylene and toluene exposure in Corbicula fluminea. Environ. Toxicol. Pharmacol. 9: 8797.
  • Walsh, H.E., Friesen, V.L. 2001. Power and stochasticity in the resolution of soft polytomies: a reply to Braun et al. Evolution 55: 12641266.
  • Walsh, H.E., Kidd, M.G., Moum, T. & Friesen, V.L. 1999. Polytomies and the power of phylogenetic inference. Evolution 53: 932937.
  • Woodruff, D.S., Kijviriya, V. & Upatham, E.S. 1993. Genetic relationships among Asian Corbicula: Thai Clams are referable to topotypic Chinese Corbicula fluminea. Am. Malacol. Bull. 10: 5153.