INTRODUCTION
In the last decades, the invasion of non-indigenous species in the Mediterranean Sea from the Red Sea through the Suez Canal, the so-called ‘lessepsian migration’ (Por, Reference Por1971), has dramatically increased (Galil, Reference Galil2000). The lessepsian migration has always animated the interest of scientists (mainly ichthyologists), generally because of interactions between immigrants and the Mediterranean autochthonous species (see e.g. Rilov & Galil, Reference Rilov, Galil, Rilov and Crooks2009 and references therein). Conversely, less attention has been paid to the genetic structure of these colonizers (Hassan et al., Reference Hassan, Harmelin-Vivien and Bonhomme2003; Hassan & Bonhomme, Reference Hassan and Bonhomme2005; Azzurro et al., Reference Azzurro, Golani, Bucciarelli and Bernardi2006; Bernardi et al., Reference Bernardi, Golani, Azzurro, Golani and Appelbaum-Golani2010).
The bluespotted cornetfish Fistularia commersonii Rüppell, 1838 (Osteichthyes: Fistulariidae) is an Indo-Pacific species recently recorded in the Mediterranean Sea. After its first report from the Middle East coast (Golani, Reference Golani2000), the geographical distribution of this fish has rapidly extended to the eastern Mediterranean (Bilecenoglu et al., Reference Bilecenoglu, Taskavak and Kunt2002; Corsini et al., Reference Corsini, Kondilatos and Economidis2002; Karachle et al., Reference Karachle, Triantaphyllidis and Stergiou2004), to some areas near the Strait of Sicily (Azzurro et al., Reference Azzurro, Pizzicori and Andaloro2004; Ben Souissi et al., Reference Ben Souissi, Zaouali, Bradai and Quignard2004) and, more recently, to the western Mediterranean (Pais et al., Reference Pais, Merella, Follesa and Garippa2007; Hemida & Capapé, Reference Hemida and Capapé2009) and the Adriatic Sea (Dulčić et al., Reference Dulčić, Scordella and Guidetti2008). Pais et al. (Reference Pais, Merella, Follesa and Garippa2007) and Merella et al. (Reference Merella, Casu, Garippa and Pais2010) suggested that F. commersonii colonized the Mediterranean as adult, because they found several Mediterranean specimens infected by two specific Indo-Pacific digeneans that cannot be harboured by larval stages of this fish. Golani et al. (Reference Golani, Azzurro, Corsini-Foka, Falautano, Andaloro and Bernardi2007), studying a partial sequence of the mitochondrial DNA (mtDNA) control region (D-loop I, also called HVS-I), argued that a genetic bottleneck occurred in the lessepsian migration of this species, as a result of a single invasion event mediated by as few as two females. This last finding is consistent with the evidence of a decrease of genetic variability in invaders (Spencer et al., Reference Spencer, Neigel and Leberg2000; Sax et al., Reference Sax, Stachowicz, Brown, Bruno, Dawson, Gaines, Grosberg, Hastings, Holt, Mayfield, O'Connor and Rice2007). However, genetic analysis until now performed on lessepsian fish showed no significant reduction of genetic diversity (Hassan et al., Reference Hassan, Harmelin-Vivien and Bonhomme2003; Hassan & Bonhomme, Reference Hassan and Bonhomme2005). It is worth noting that some of these studies were performed using multiple molecular markers and/or different nuclear or mitochondrial regions, since this approach is advisable to get more complete and reliable results (see e.g. Pogson et al., Reference Pogson, Mesa and Boutilier1995; Neigel, Reference Neigel1997; Clements et al., Reference Clements, Gray and Choat2003).
In this context, this study is aimed to provide new insights on the number and variability of matrilineal lineages of F. commersonii in the Mediterranean Sea throughout the combined use of different mitochondrial regions. Four regions of mtDNA were chosen: the second segment of the control region, D-loop II (also called HVS-II) the 16S rDNA, 12S rDNA, and Cytochrome b (Cyt b) of the coding region.
MATERIALS AND METHODS
Twelve specimens of Fistularia commersonii from different western and eastern Mediterranean areas (sensu Spalding et al., Reference Spalding, Fox, Allen, Davidson, Ferdaña, Finlayson, Halpern, Jorge, Lombana, Lourie, Martin, McManus, Molnar, Recchia and Robertson2007) were analysed: six from Sardinia (Oristano, west coast; Torre delle Stelle, south coast; Capo Comino and Arbatax, east coast), and six from North Africa (Teboulba and Sfax, Tunisia; Tripoli, Libya) (Table 1; Figure 1).
Sites of sampling of Fistularia commersonii. (1) Oristano (ORs); (2) Torre delle Stelle (TSs); (3) Arbatax (ARs); (4) Capo Comino (CCs); (5) Teboulba (TEa); (6) Sfax (SFa); (7) Tripoli (TRa). The suffixes ‘s’ and ‘a’ in the abbreviations indicate ‘Sardinia’ and ‘North Africa’, respectively.
Data on the individuals of Fistularia commersonii examined.
DNA was extracted from small portions of the muscle (about 15 mg) by means of the QIAGEN® DNeasy Tissue kit. Mitochondrial regions were amplified using the primers HCAL2 and HCAH2 (Teske et al., Reference Teske, Cherry and Matthee2003) which amplified 396 bp of the control region D-loop II; L2510 and H3058 (Wilson et al., Reference Wilson, Vincent, Ahnesjo and Meyer2001) which amplified 564 bp of the 16S rDNA gene; L1091 and H1478 (Kocher et al., Reference Kocher, Thomas, Meyer, Edwards, Pääbo, Villablanca and Wilson1989) which amplified 350 bp of the 12S rDNA gene, L14725 (Pääbo et al., Reference Pääbo, Thomas, Whitfield, Kumazawa and Wilson1991) and H15240 (Wilson et al., Reference Wilson, Vincent, Ahnesjo and Meyer2001) which amplified 284 bp of the Cytochrome b gene. Each 25 µl PCR mixture contained about 15 ng/μl of total genomic DNA on average, 0.2 µM of each primer and 2.5 U of EuroTaq DNA Polymerase (Euroclone) in a reaction mix prepared according to the manufacturer's instructions. The MgCl2 concentration was 2.5 mM.
Polymerase chain reaction was performed in a MJ PTC-100 Thermal Cycler (MJ research), programmed for 1 cycle of 2 minutes at 94°C, 35 cycles of 1 minute at 94°C, 1 minute at 54°C and 1 minute and 30 seconds at 72°C each. At the end, a post-treatment for 5 minutes at 72°C and a final cooling at 4°C were applied. For all PCR reactions, negative controls and replicates were included. Electrophoresis runs were carried out on 2% agarose gels, made using 0.5 × TBE buffer and stained with ethidium bromide (10 mg/ml), at 4 V/cm for 20 min.
Polymerase chain reaction products were purified using ExoSAP-IT (USB Corporation), and sequenced using an external core service. Sequences were aligned using the software BioEdit 7.0.5.2 (Hall, Reference Hall1999) and deposited in GenBank (Accession Nos: 12S, GQ901988 to GQ901999; 16S, GQ902000 to GQ902011; Cyt b, GQ902012 to GQ9023; D-loop II, GQ902024 to GQ902035).
The best probabilistic model of sequence evolution was determined after evaluation by Modeltest 3.7 (Posada & Crandall, Reference Posada and Crandall1998), using the corrected Akaike information criterion (AICc). Since cluster analysis involved a dataset including four mitochondrial genes, they were tested for heterogeneity (Farris et al., Reference Farris, Kallersjo, Kluge and Bult1995) by the partition–homogeneity test, implemented in PAUP*4.0b10 (Swofford, Reference Swofford2003) in order to measure the significance of incongruence among datasets (Bull et al., Reference Bull, Huelsenbeck, Cunningham, Swofford and Waddell1993; Chippindale & Wiens, Reference Chippindale and Wiens1994; Sullivan, Reference Sullivan1996; Cunningham, Reference Cunningham1997). The 99% confidence interval was used to determine the level of significance, following Farias et al. (Reference Farias, Ortí and Meyer2000).
To characterize genetic variation among individuals, the number of polymorphic sites (S), the number of haplotypes (nH), estimates of haplotype diversity (h), and nucleotide diversity (π), were computed using the software DnaSP 4.10 (Rozas & Rozas, Reference Rozas and Rozas1999).
Genetic relationships among individuals were inferred using the program Network 4.5.1.0 (www.fluxus-engineering.com) by means of the median-joining algorithm (Bandelt et al., Reference Bandelt, Forster and Rohl1999), assigning the same weight to all the observed polymorphisms, and excluding from the analysis the insertion/deletions (indels) of presumably poor phylogenetic value (Redd & Stoneking, Reference Redd and Stoneking1999).
RESULTS AND DISCUSSION
The level of genetic variability of the lessepsian migrant Fistularia commersonii was studied in different localities of the western and eastern Mediterranean Sea never sampled before. It is worth noting that, due to the rarity and irregular findings of this species in the western basin (with some records until autumn–winter 2007–2008, but no further reports in the following period), it was not possible to collect any additional samples beyond the 12 original ones.
The best model selected by AICc on D-loop II and 16S was K80, whereas 12S and Cyt b fitted the JC model. The partition–homogeneity test did not show significant heterogeneity (P > 0.01) among the four mitochondrial fragments. Thus, the four regions analysed (D-loop II, 16S, 12S and Cyt b) were combined in a resultant total length sequence of 1594 bp (including 1 indel), with 17 polymorphic sites 15 of which were parsimony-informative. The best model selected on the pooled dataset was K80.
The genetic diversity indices of the individuals of F. commersonii analysed for each mitochondrial region are reported in Table 2. The number of haplotypes ranged from one (Cyt b) to three (16S and 12S). The highest value of haplotype diversity (h = 0.5) was found in the 16S gene and the highest value of nucleotide diversity (π = 0.009) was evidenced in the D-loop II. However, the D-loop II—notwithstanding it provided good results in terms of genetic variability in some species of Syngnathiformes (Teske et al., Reference Teske, Cherry and Matthee2003; Sanna et al., Reference Sanna, Addis, Biagi, Motzo, Carcupino and Francalacci2008)—did not give high levels of global differentiation for F. commersonii, as observed in an analogous study on mitochondrial DNA of other fish (Tang et al., Reference Tang, Liu, Maydenc and Xiong2006). Conversely, the 16S and 12S genes, although often highly conservative and generally used at higher systematic levels (Giribet et al., Reference Giribet, Edgecombe and Wheeler2001), showed to be the most variable and informative among the mtDNA regions considered. The Cyt b did not show any genetic variability, leading to consider this region of scarce importance to achieve useful information for further phylogeographic analysis on this species.
Summary of the genetic variability indices for each of the four mitochondrial regions analysed and for the pooled dataset.
The network analysis (Figure 2) evidenced the occurrence of five haplotypes, grouped into two main haplogroups, both including individuals from North Africa and Sardinia with no geographical trend. Most of the specimens exhibited one of the two main haplotypes, whilst three individuals from Sardinia were differentiated for one to three mutations.
Network plot obtained for the pooled dataset (D-loop II, 16S, 12S and Cyt b); in each spot, in white individuals from Sardinia and in black from North Africa. Numbers on network branches indicate the polymorphic sites.
The main result obtained by the analysis of the combined dataset was the finding of five mitochondrial haplotypes, pointing out the presence of more than the two single matrilineal lineages previously described by Golani et al. (Reference Golani, Azzurro, Corsini-Foka, Falautano, Andaloro and Bernardi2007) in the Mediterranean Sea, after examining a single mtDNA region. Remarkably, on the basis of parasitological evidence, Pais et al. (Reference Pais, Merella, Follesa and Garippa2007) and Merella et al. (Reference Merella, Casu, Garippa and Pais2010) suggested the possibility that F. commersonii invaded the Mediterranean with a large number of adult individuals, compatible with a not severe bottleneck event.
In this context, the present study highlights that, due to the heterogeneous levels of variability found for different mitochondrial regions of F. commersonii, the choice of a unique mtDNA segment could lead to a sub-estimation of this value even when high numbers of specimens are sampled. For this reason, the enlarged number of mitochondrial regions used in this work has allowed to detect a higher number of haplotypes in spite of the lower number of individuals analysed. Furthermore, although based on a narrow number of individuals, the present data suggest that the biogeographical barrier represented by the Strait of Sicily did not affect significantly either the dispersal or the genetic distribution of F. commersonii in the eastern and western parts of the Mediterranean Sea.
The results obtained provided new information on the levels of genetic variability for the analysed mitochondrial regions of F. commersonii. The knowledge of the polymorphism rates of D-loop II, 16S, 12S and Cyt b (Table 2) will allow to make a well-addressed choice of the most informative mtDNA regions depending on the specific purpose of the survey (such as population genetics, phylogeography and phylogeny). Indeed, different mutation and fixation rates shown by mitochondrial regions have been often used to depict multifaceted degrees of differentiation (see e.g. Saccone et al., Reference Saccone, Pesole and Sbisà1991).
In conclusion, the combined use of different mtDNA regions here proposed allowed a finer inference of the levels of genetic variability, and these results will be useful to shed some light on the debates arising from the supposed lack of reduction in genetic variability of Mediterranean populations of different lessepsian fish.
ACKNOWLEDGEMENTS
We wish to thank two anonymous referees for their helpful comments that contributed to improve this paper. This research was supported by the ‘Centro di Eccellenza’ of the Università di Sassari.
