PLOIDY DEGREE OF PONTOSCOLEX CORETHRURUS COMPLEX SPECIMENS 

Get Complete Project Material File(s) Now! »

Reproductive isolation, polyploidy and speciation

In order to know the mode and process of speciation in nature, it is necessary to determine how reduction of gene flow in the ‘grey zone’ occurs between species, sufficiently to allow each of them to become irrevocably committed to different evolutionary paths (Bush, 1994). Reproductive isolation is essential for the gene flow to be stopped, which is a consequence of a gradual process happening in the ‘grey zone’ and has two different types; pre- and postzygotic barriers (Widmer et al., 2009). Pre-zygotic barriers could be related to differences of habitat, behavioral changes, temporal differences or gametic incompatibilities between species. For instance, the recent evolution of Senecio eboracensis from Senecio vulgaris, two plant species which are interfertile and are both visited by generalist pollinators, is due to selfing of these species by which species boundaries are maintained (review in Widmer et al., 2009). Prezygotic barriers in animals are controlled by few major genes and are important during early stages of speciation, whereas postzygotic barriers are controlled by numerous genes of minor effect and accumulate more gradually (review in Widmer et al., 2009). Therefore, prezygotic barriers evolve faster than postzygotic ones in animals. Postzygotic isolation is related to hybrid sterility, hybrid inviability, and embryo mortality, and among them hybrid sterility most likely evolves faster than hybrid inviability (review in Widmer et al., 2009). For example, in a study based on data available in literature on 368 species, Price and Bouvier (2002) showed that complete loss of F1 hybrid fertility of birds takes on the order of millions of years, while loss of F1 hybrid viability occurs over longer timescales. The evolution of complete reproductive isolation among species may take hundreds to millions of generations to occur, and during this long history populations change in size and spatial distributions, and the processes that enhance or erode barriers to gene flow including hybridization may occur (Abbott et al., 2013). Hybridization in a spatial context mostly occurs in ‘hybrid zones’; which is at abrupt parapatric boundaries. In this context, the exchange of genes occurs between locally adapted populations. In a temporal context, hybridizations occur with ‘secondary contacts’; which is after a period of independent evolution (Abbott et al., 2013).
In some cases, speciation is much faster i.e., polyploid speciation (the presence of three or more complete chromosome sets in an organism, in this case the organism may be immediately isolated reproductively from the other individuals of the same species). This type of speciation has been highly reported for plant species. For instance, Wood et al. (2009) showed that 15% of angiosperm and 31% of fern speciation events are accompanied by ploidy increase.

Polyploidy in earthworms

Parthenogenetic reproduction is common in earthworms, which could be related to the lack of sex determining chromosomes in them (Lynch, 1984). Parthenogenetic reproduction in earthworms could result in or maintains odd or even ploidy degrees, while sexually reproducing polyploid species have even numbers of ploidy (Terhivuo and Saura, 2006). For instance, for the athecal individuals (without spermathecae) of Amynthas catenus the ploidy degree are; 2n, 3n, and 4n, while the sexthecals (with spermathecae) are diploid (2n) (Shen et al., 2011).
Sometimes different ploidy degrees within one morphospecies could be observed by morphological differences and ecological preferences. For instance, polyploid species; Dendrobaena subrubicunda (4n) and Octolasion cyaneum (10n), are larger specimens than their diploid forms, because of clitellum displacement in some segments (Muldal, 1952). Difference in ecological preferences has also been observed for different ploidy degrees within a species. Viktorov (1997) demonstrated niche partition for diploid and polyploid individuals in sympatry. For instance, in south Siberia diploid Eisenia nordenskioldi f. pallida are endogeics while octoploid E. nordenskioldi f. typica are anecics (Viktorov, 1997). As mentioned above parthenogenetic species are widespread in a wider range than sexual ones in a pattern called “geographic parthenogenesis” which is also observed for the earthworms. For instance, sexual forms of Achaea trapezoids earthworm species are found in the circum-Mediterranean areas, while parthenogenetic ones are found in a wider range in the rest of the world (De Sosa et al., 2017b).
Despite the fact that more and more species complexes are found in earthworms, no studies have investigated the possible ploidy degree differences between cryptic species within a complex. This hypothesis could show the possibility of reproductive isolation among cryptic species which has occurred by polyploidization which is an abrupt speciation mechanism.

DNA material and COI sequences

A total of 299 specimens belonging to the Pontoscolex corethrurus complex were collected from six countries in tropical and subtropical zones (Table C 3 in Supplementary data). These countries included: French Guiana which is situated in the Guayana shield (i.e., the putative native zone of the morphospecies), Brazil, Mexico, Gabon, Taiwan, and Thailand. In each country, 1-8 sites were sampled. After collection, specimens were washed in distilled water, and conserved individually in 100% ethanol. A piece of 20 mg of dorsal tegument was cut between the tail and the clitellum. DNA extraction was performed using the Nucleospin tissue kit (Macherey Nagel, France). A total of 662 sequences of a fragment of the cytochrome oxidase I gene (COI) were obtained from a previous study (Taheri et al., in revision).

COI sequences analyses

A phylogenetic tree based on 662 COI sequences was built (Figure C 6 in Supplementary data) using the model HKY+G which was the best fitting evolutionary model selected with MrModelTest (Nylander, 2016) using the Akaike Information Criterion. Two specimens of Pontoscolex spiralis from Guadeloupe and Puerto Rico were chosen as outgroups (Genbank accession numbers XXX15). Maximum likelihood (ML) analysis was performed with PhyMl online analysis (http://www.atgc-montpellier.fr/phyml/, Guindon et al., 2010). Clade support was assessed using bootstrap with 1000 pseudoreplicates. Trees were then edited using FigTree v. 1.4.2. (Rambaut, 2014).
A total of 479 COI sequences from 49 locations were assigned to P. corethrurus L1. Among them, 269 sequences from 12 populations were used for a population genetics analysis. For these samples, number of polymorphic sites, number of haplotypes, haplotype (gene) diversity and nucleotide diversity (Pi; Nei, 1987) were calculated using DnaSP v. 5 software (Librado & Rozas, 2009). A haplotype network was constructed based on derived haplotypes in Network v. 5.0.0.1, using median joining calculations (Bandelt et al., 1999).

AFLP procedure

AFLP analysis was carried out according to Vos et al. (1995) with few modifications: approximately 50 ng/μl of purified genomic DNA of each specimen was digested with two digestive enzymes. The first digestion was done in 10μl with Taq1 (20U, New England BioLabs, Ipswich, MA, USA (NEB)), Buffer Taq1 (10X, NEB), BSA (1mg.ml-1, NEB) and incubated for 1h30 at 65°C by thermal cycler (T100TM, BIO-RAD Laboratories Inc., Foster city, CA, USA). The second digestion of the DNA from the last step was done in 15μl with a solution of: EcoRI (20U, NEB), Buffer EcoRI (10X, NEB), BSA (1mg.ml-1, NEB) and incubated for 1h30 at 37°C. We verified DNA digestion quality by gel electrophoresis and negative controls were run at each time. Adapters were ligated using 50pmoles μl-1 of double- stranded Taq1 adapter (Taq top 5’-GACGATGAGTCCTGAC and Taq bottom 5’-CGGTCAGGACTCAT, Eurofins Genomics, Germany) and 5pmoles μl-1 of double-stranded EcoRI adapter (Eco top 5’-CTCGTAGACTGCGTACC-3’ and Eco bottom 5’-AATTGGTACGCAGTCTAC-3’, Eurofins Genomics, Germany), T4 DNA ligase buffer (10X, Promega, Madison, WI, USA), ATP (10mM, NEB), BSA (1mg.ml-1, NEB), T4 DNA ligase (3U, Promega). Samples with a total volume of 50μl (prepared solution + digested DNA from previous step + adjusted water) was then incubated for 3h at 37°C. Digestion-ligation production of 20μl was diluted with 40μl of AE buffer (QIAGEN Sciences, Maryland, USA). Then each dilution was divided in 3 separate samples of 15μl each. Pre-selective PCRs contained 5pmoles of E01 primer (GACTGCGTACCAATTCA) and 5pmoles of T01 primer (GATGAGTCCTGACCGAA), MgCl2 (25mM, Promega), dNTPs (10mM, InvitrigenTM, Life technologies, Carlsbad, CA, USA), GoTaq buffer (5X, Promega), DNA Taq polymerase (5U, Promega), adjusted with water to have a total volume of 50μl in each sample (added to diluted ligation product). The PCR pre-selective was done by an initial denaturation step at 94°C for 2min, followed by annealing setup of 30 cycles containing; 94°C for 30s, 56°C for 1 min and 72°C for 60s, and finally elongation setup of 72°C for 10 minutes. After this setup, the three sub samples were reassembled in one, and then 1μl of each sample was diluted in 19μl of AE buffer (QIAGEN). Selective PCR reactions contained 5μl of pre-amplified DNA with 15μl of a solution containing MgCl2 (25mM, Promega), dNTPs (10mM, Invitrigen), GoTaq buffer (5X, Promega), 5pmoles.μl-1 T32 primer (5’-GATGAGTCCTGACCGAAAC-3’) or 5pmoles.μl-1 T38 primer (5’-GATGAGTCCTGACCGAAACT-3’), DNA Taq polymerase (5U, Promega) and one primer combination E32-FAM was used (5’-GACTGCGTACCAATTCAA-3’). A touchdown thermal cycling (PTC-100) started with denaturing setup of 94°C for 2 min, following by 9 cycles; 94°C for 30s, 65°C for 30s with 1°C diminution per cycle, 72°C for 60s, following by 26 cycles containing 94°C for 30s, 56°C and 72°C during 1min and finally 72°C for 10 minutes. After each setup DNA solutions were centrifuged by ROTANTA 460R (Hettich Lab Technology, Tuttlingen, Germany). Amplified products were mixed with formamide (Hi_DiTM, Applied Biosystems, Foster city, CA, USA (AB)) and a GenescanTM-500LIZTM (AB) size standard (9.5μl of formamide and 0.5 μl of GenescanTM-500LIZTM for 2μl of amplified product). Fragments were separated on a ABI PRISMTM 3130 Genetic Analyzer (platform INSERM, Henri Mondor hospital, Créteil). Raw data were visualized and the fragments manually scored using Genemapper V5 (Applied Biosystem) software. Processed data were exported as presence/absence matrix.

Reproducibility of AFLP profiles

Genetic markers, including AFLPs, can be prone to genotyping errors with various potential sources (Bonin et al., 2004). Therefore, we estimated the genotyping error rate of our dataset by re-genotyping and blind scoring of 54 randomly chosen individuals (25% of the samples) within P. corethrurus L1 from 15 locations (i.e., Bahia cacao, Bahia pasture, Belem, Caxiuana, Joinville, Orléans, Sao Paolo, Cayenne, Mancha, Tlalcotlen, Mitaraka, PararéA, PararéB, Taiwan National University, and Chachoengsao). The error rate was calculated by mismatch error rate based on the formula proposed by Bonin et al. (2004), multiplied by the total number of markers (i.e., 394).

AFLP analyses of cryptic lineages

The Structure software (Pritchard et al., 2000) was used to assign the individuals for which COI sequences was not available to the L1, L3 and L4 cryptic species of the P. corethrurus complex. For that, a ‘reference’ data set was built, composed of 113 AFLP profiles from L1, L3 and L4 individuals of known COI sequences. We first checked if AFLP data allowed to distinguish the lineages within the complex using the Structure software by modelling cluster assignments for K = 1–5 clusters using non-admixture model. We made 5 independent runs for each K to confirm consistency across runs in all simulations. In a second time, we used the ‘reference’ data to assign 54 AFLP profiles from individuals for which COI sequences were not available, to the cryptic species of the P. corethrurus complex. The used model was non-admixture, and K was set between 1 to 5, with 5 independent runs per K.
In three populations (MNH, TLC and NTU) coexistence of cryptic species was found, Structure was run to find potential hybrids among them. The model selected was admixture, and K was set between 1 to 5 with 15 independent runs per K. For each analysis, we performed a burn-in period of 10000 iterations and 100000 Markov chain Monte Carlo iterations.
After each analysis, to determine the most likely value of K, we used the ΔK method of Evanno et al. (2005) implemented in Structure Harvester (Earl & vonHoldt, 2012). Results from different replicate runs were combined into one output using Clumpp software v. 1.1.2 (Jakobsson & Rosenberg, 2009). Results visualization were done by Distruct (Rosenberg, 2007).
We also analyzed possible hybridization occurrence within MNH, TLC and NTU populations with the program NewHybrids (Anderson & Thompson, 2002; Anderson, 2008) based on statistical model-based Bayesian methods. The software considers six genotype categories: pure species A, pure species B, F1 hybrid, F2 hybrid, and the F1 backcross to pure species A or pure species B, with the results of estimated posterior probability to assign each individual to one of the six genotypic classes. We assigned each individual to the most likely (probability>0.50) NewHybrids genotypic class. Given the large number of polymorphic loci (308) a burn-in period of 75,000 repetitions was defined, and 100,000 MCMC iterations, with no previous population information, and ‘Jeffery’s like priors’ were considered.

Population genetics of P. corethrurus L1

A total of 226 AFLP profiles corresponding to 12 populations from 5 countries (Brazil, French Guiana, Mexico, Gabon and Thailand) were used to investigate the genetic variation within and between populations of P. corethrurus L1. Genetic diversity statistics, including number of different haplotypes by considering the ‘error rate’ calculated from replicates, genotype diversity, gene diversity, proportion of variable markers and the frequency down-weighed marker value (DW) (Schönswetter & Tribsch, 2005) were calculated using AFLPdat program in RStudio (RStudioTeam, 2015). Genotype and gene diversity were calculated based on Nei’s formula (Nei, 1987). Similar haplotypes within locations were removed and further analyses on P. corethrurus L1 were carried out on a dataset without clones.
In order to evaluate the evidence for recombination, measures of multilocus gametic disequilibrium were calculated and tested for significance with 500 randomizations in Multilocus software (http://www.bio.ic.ac.uk/evolve/software/multilocus/), using 3 different methods. Since in a pair of binary character data, the presence of all four possible combinations of characters (0/0, 1/0, 0/1, 1/1) indicate incompatibility which is more parsimoniously explained by sexual recombination than by three mutation events, the proportion of compatible pairs of loci was computed to probe the predominant mating system in the populations. Moreover, the index of association (IA) and an alternative measure of index of association that is less sensitive to the number of loci (𝑟̅d, Agapow & Burt, 2001) were computed. We further tested for gametic disequilibrium based on the distribution of allelic mismatches between pairs of genotypes over all loci using an exact test implemented in Arlequin. To adjust for multiple comparisons, the SGoF method (Carvajal-Rodríguez et al., 2009) as implemented in the software Myriad (http://myriads.webs.uvigo.es/MyriadsReadme.htm) was applied.
We identified loci with greatly increased differentiation between populations using a genomic scan approach. These Fst outlier may be interpreted as signature of (i) divergent selection (ii) intrinsic (i.e. environment independent) pre- or post-zygotic genetic incompatibilities (Bierne et al., 2011) or (iii) differential introgression from a sister species (Gosset & Bierne, 2013). Tests for Fst outliers were carried out using Bayescan 2.01 (Foll et al., 2010), running with 100,000 iterations (sample size 5000 * thinning interval 10 + 50000 burn-in), after 20 pilot runs (5000 iterations each), and selecting outliers at a threshold of prior odds (PO) >10, as suggested by the manual for datasets with hundreds of loci. Further analyses of genetic structure were carried out on two datasets: the original dataset and the dataset without outlier loci.
We quantified the amount of genetic differentiation of population groups using a hierarchical Analysis of Molecular Variance (AMOVA) implemented in Arlequin V3.5 (Excoffier & Lischer, 2010). In this analysis, the AFLP data set was partitioned at three levels: groups of native versus introduced populations, among-populations within groups and among all populations. One thousand random permutations were used to infer the significance of the variance components. In addition, an unbiased estimate of differentiation among populations, θ(II) was obtained using the Bayesian method proposed by Holsinger et al. (2002) and implemented in the software Hickory v1.1. The data were run with the default parameters using the f-free model. To illustrate the relationships among populations, split networks were constructed using the software Splitstree version 4.1.4.6 (Huson & Bryant, 2006) on AFLP profiles. We used the distance-based Neighbor-Net (N-net) method for construction of networks. The N-net provides good visualization of the data when it presents complex evolutionary steps or reticulate relationship among genotypes (Huson & Bryant, 2006). The networks were constructed based on Nei’s distance (GD) matrix between populations calculated with a Bayesian method using AFLP-Surv version 1.0 (Vekemans et al., 2002) with non-uniform distribution by assuming deviation from Hardy-Weinberg equilibrium; Fis values were estimated by Hickory software using the full model (Holsinger et al., 2002). Analyses were done with 1000 permutations and 1000 bootstrap values.

Principal stages of the analyses

A workflow of principal stages of the analyses is presented in Figure C 1: first, assignment of AFLP sequences in P. corethrurus complex to already known COI lineages (L1, L3 and L4); second, potential hybridizations of species in P. corethrurus complex (L1, L3, and L4) for three populations in sympatry based on AFLP profiles; finally, the phylogeography and population genetics study of P. corethrurus L1 populations based on 479 COI sequences and 226 AFLP profiles.

Table of contents :

I. INTRODUCTION
I.1. BIOLOGICAL INVASION
I.2. SPECIATION AND CRYPTIC SPECIES
I.3. REPRODUCTIVE ISOLATION, POLYPLOIDY AND SPECIATION
I.4. MODEL OF STUDY: PONTOSCOLEX CORETHRURUS
I.5. THESIS OBJECTIVES AND STRUCTURE OF THE DOCUMENT
II. CHAPTER 1: WHAT DO ‘POSITIVE’ AND ‘NEGATIVE’ IMPACTS OF INVASIVE SPECIES DEPEND ON?
II.1. CHAPTER’S FOREWORD
II.2. ARTICLE
III. CHAPTER 2: HOW MANY INVASIVE LINEAGES WITHIN THE COMPLEX OF SPECIES PONTOSCOLEX CORETHRURUS?
III.1. CHAPTER’S FOREWORD
III.2. ABSTRACT
III.3. INTRODUCTION
III.4. MATERIAL AND METHODS
III.5. RESULTS
III.6. DISCUSSION
III.7. CONCLUSION
III.8. ACKNOWLEDGEMENT
III.9. REFERENCES
III.10. SUPPLEMENTARY DATA
IV. CHAPTER 3: PHYLOGEOGRAPHY AND POPULATION GENETICS OF AN INVASIVE PEREGRINE EARTHWORM SPECIES
IV.1. CHAPTER’S FOREWORD
IV.2. ABSTRACT
IV.3. INTRODUCTION
IV.4. METHODS
IV.5. RESULTS
IV.6. DISCUSSION
IV.7. CONCLUSION
IV.8. REFERENCES
IV.9. SUPPLEMENTARY DATA
V. CHAPTER 4: PLOIDY DEGREE OF PONTOSCOLEX CORETHRURUS COMPLEX SPECIMENS 
V.1. CHAPTER’S FOREWORD
V.2. INTRODUCTION
V.3. MATERIALS AND METHODS
V.4. RESULTS AND DISCUSSION
V.5. CONCLUSION AND PERSPECTIVES
V.6. REFERENCES
VI. GENERAL CONCLUSIONS AND PERSPECTIVES
VI 1. SPECIATION, HYBRIDIZATION AND GENETIC VARIATION WITHIN PONTOSCOLEX CORETHRURUS COMPLEX
VI.2. P. CORETHRURUS L1: THE INVASIVE SPECIES
VI.3. MIXED REPRODUCTIVE STRATEGY WITHIN A SPECIES COMPLEX AND/OR WITHIN A SPECIES
VII. THESIS APPENDIX
VII.1. APPENDIX 1
VIII. REFERENCES (OUT OF CHAPTERS)
IX. LIST OF FIGURES AND TABLES
X. THESE EN FRANÇAIS : PROCESSUS MACRO- ET MICRO-EVOLUTIFS AU SEIN D’UN COMPLEXE D’ESPECES, CAS D’ETUDE DE L’ESPECE TROPICALE ET INVASIVE DE VERS DE TERRE ; PONTOSCOLEX CORETHRURUS

GET THE COMPLETE PROJECT