Abstract: This study examines eight mugilid species: Mugil cephalus, Chelon labrosus, Oedalachelis labeo, Liza abu, Liza aurata, Liza saliens and Liza ramada from the Mediterranean Sea and Mugil soiuy Black Sea on the basis 16S rDNA gene of mitochondrial DNA. The 16S rDNA dataset contained 121 variable and parsimony informative sites and the mean nucleotide diversity (Pi) was found to be 0.05. Haplotype diversity was found to be 0.88 and 7 different haplotypes were observed. Species specific haplotypes were detected and only C. labrosus and L. ramada shared the same Haplotype (H₃). Sequencing analysis revealed that M. cephalus was clearly separated from the other species. For inter-generic comparisons, there was no genetic difference between C. labrosus and L. ramada and C. labrosus and O. labeo should be considered within the genus Liza. Moreover M. soiuy and L. abu should be considered under the genus Liza, or new genus name should be given for these two species.

INTRODUCTION

Mugilid species are distributed worldwide and inhabit marine, estuarine and freshwater environments in all tropical and temperate regions. Various Mugilid species play an important role in the fisheries and aquaculture of many regions of the world (Nash and Shehadeh, 1980), especially in aquacultural practices based on natural food webs.

The family Mugilidae includes 17 genera and >75 species (Nelson, 2006) in the world. Mugilid species are commonly found along the Mediterranean and Black Seas and represented with four genera and nine species. Mugil cephalus, Liza aurata, Liza ramada, Liza saliens, Chelon labrosus and Oedalechilus labeo have Atlanto Mediterranean distribution. L. carinata has recently spread to the South-Eastern Mediterranean from the Red Sea through the Suez Canal (Thomson, 1997). Mugil soiuy was initially introduced into the Azof Sea and now has been reported in the Black Sea and more recently in the Aegean Sea (Kaya et al., 1998). Liza abu is a freshwater species and found in the Orontes River, Turkey connected to Mediterranean Sea (Turan et al., 2004).

Due to the very conservative morphology displayed by mugilids many investigations based on various morphological characters did not elucidate their system atic problem (Schultz, 1946; Trewavas and Ingham, 1972; Stiassny, 1993), while the use of the pharyngobranchial organ as a key character to address the identification and taxonomy of mugilids also provided poor results (Harrison and Howes, 1991). Thomson (1997) revised Mugilidae family and reported 14 genera and a total of 64 valid species, most of these species are representatives of Liza and Mugil genera. Despite the fact that the Mugilid species have been revised many times, genetic identification and the taxonomic status of some species and genera within the family is still confused (Rossi et al., 1998; Turan et al., 2005).

Nowadays, mitochondrial DNA analysis is a very useful tool for molecular genetic studies because of its special features (Meyer et al., 1990; Billington and Hebert, 1991; Normark et al., 1991; Meyer, 1992). In most species mitochondrial DNA (mtDNA) is highly variable and is therefore a good marker for detecting possible genetic differentiation. Therefore, mtDNA variation has been widely investigated among several fish species including mugilidae (Papasotiropoulos et al., 2002; Caldara et al., 1996; Turan, 2008; Turan et al., 2009).

Although, the systematic relationships among these species have been investigated with the use of nonmorphological characters, such as biochemical and cytogenetic markers (Delgado et al., 1992; Rossi et al., 1996, 1997, 2000; Gornung et al., 2001, 2004; Nirchio et al., 2003), allozyme electrophoresis (Autem and Bonhomme, 1980; Papasotiropoulos et al., 2001; Rossi et al., 2004; Turan et al., 2005) and mtDNA analyse (Caldara et al., 1996; Murgia et al., 2002; Papasotiropoulos et al., 2002, 2007; Rossi et al., 2004; Fraga et al., 2007) they did not result in a clear taxonomic perspective (Cataudella et al., 1974; Delgado et al., 1992; Crosetti et al., 1993; Rossi et al., 1997, 2000).

In this study, taxonomic description and status of 8 mugilid species (Mugil cephalus, M. soiuy, Liza ramada, L. aurata, L. abu, L. saliens, Chelon labrosus, Oedalechilus labeo), living in the Mediterranean Sea, were investigated with mitochondrial 16S rDNA gene sequence data.

MATERIALS AND METHODS

Samples from eight mugilid species (Mugil cephalus, Liza ramada, Liza aurata, Liza abu, Liza saliens, Chelon labrosus, Oedalechilus labeo) were collected from Iskenderun Bay in North-eastern Mediterranean Sea and only one species Mugil soiuy was sampled from Trabzon in the Black Sea.

The number and location of the samples used in the sequence analysis are given in Table 1. The samples were placed on ice and kept frozen at -40°C in the laboratory. In the laboratory fin clips and muscle tissue were collected and preserved in 95% ethanol for DNA extraction.

Total genomic DNA was extracted from a peace of fin tissue (approximately 2 mm²) using AGOWA mag Midi DNA isolation Kits (AGOWA, Berlin, Germany). The amplification of the mitochondrial 16S rDNA gene was performed using PCR with a profile of 94°C for 4 min, followed by 35 cycles of 94°C/30 sec strand denaturation, 52°C/20 sec annealing and 72°C/1 min 30 sec primer extension and a final 7 min elongation at 72°C. The 16S rDNA amplification conditions were: 1.5 μL 10x polymerase buffer, 0.5 μL dNTP (10 mM), 0.3 μL Teg DNA polymerase (3 U μL^-1) equivalent to Taq DNA polymerase, 0.05 μL 16Fi1 40 primer (100 μM) (5'-CG(CT)AAGGGAA(ACT)GCTGAAA-3'), 0.05 μL 16 Fi1 524 primer (100 μM) (5'-CCGGTCTGAACTCAGATCACGT AG-3'), 3-5 μL DNA from AGOWA purification and water for a total reaction volume of 15 μL.

Amplified DNA was purified with Exo/Sap enzymes (Cleveland, Ohio, USA) following the manufacturer’s instructions. Finally, all the samples were sequenced in both directions using 16Fiseq1463 (5'-TGCACCAT TAGGATGTCCRGATCC AAC-3') and 16sarL (5'-CGCCTG TTTAACAAAAACAT-3') primers. The sequencing products were loaded onto an ABI3730 (Applied Biosystems) automated sequencer.

Sequences were aligned and ambiguous bases resolved by eye using Sequencer v.4.5 (Gene Codes Corp) and final alignment was done manually with BioEdit (Hall, 1999). MtDNA sequence data were analysed to assess levels of pairwise nucleotide variation and to determine nucleotide composition for each taxon using MEGA 3.1 (Kumar et al., 2004). The molecular phylogenetic tree was constructed using the two distinct phylogenetic approaches: a distance-based method using Neighbor Joining (NJ) (Saitou and Nei, 1987) and a cladistic approach using the Maximum Parsimony (MP). Bootstrap resampling was applied to assess the relative stability of NJ trees produced with different substitution models.

Nucleotide diversity and DNA total divergence (Dxy; were estimated using MEGA v.4 (Tamura et al., 2007). In all models, phylogenetic trees were rooted using out group species Carangoides armatus, which belongs to the family Carangidae and its sequence is published in GenBank under accession number NC_004405. The sequences have been deposited in the GenBank with accession numbers given in (Table 1).

Table 1:	Sampling coordinates and GenBank accession numbers for the 16S rDNA segment sequenced in this study
n-S: Number of specimens used in the sequencing

RESULTS AND DISCUSSION

After alignment, the partial 16S rDNA gene sequences consisted of 788 bp. Examination of the gene reveals a moderate amount of guanine (G; 21.4%) and abundance of adenine (A; 31%). The 16S rDNA dataset contained 121 variable and parsimony informative sites and the mean nucleotide diversity (Pi) was found to be 0.05. Haplotype diversity was found to be 0.88 and 7 different haplotypes were observed. Variable nucleotide positions and frequencies of haplotypes are given in Fig. 1. Species specific haplotypes were detected and only C. labrosus and L. ramada shared the same Haplotype (H₃) (Table 2).

Pairwise genetic distances between the species were given in (Table 3). For inter-generic comparisons, there was no genetic difference between C. labrosus and L. ramada and highest genetic divergences were observed between M. cephalus and O. labeo.

The three different phylogenetic approaches resulted in same tree topologies and the clades are well supported (Fig. 2). C. labrosus and L. ramada clustered together on the same branch and were sister group to L. saliens, O. labeo clustered with L. aurata and M. soiuy branched with L. abu. On the other hand, M. cephalus highly divergently clustered outside of this group.

The present study based on 16S rDNA gene sequencing data of four genera and eight species of the family Mugilidae revealed different systematic classification from the current classification and question existence of Chelon and Oedalechilus genera. Mugilid species are generally morphologically uniform; the genus Oedalechilus by the different pattern of the jaw structure, that features a very broad upper lip, with the maxillare extending vertically backwards under the preorbitale. In other Mediterranean mugilid species (L. abu, L. saliens, L. ramada, L. aurata and M. cephalus), the upper lip is not broad and the mouth is rather horizontal, with the same extending maxillare. However, M. cephalus shows at least some features that make it diagnosable compared to Liza, Chelon and Oadalechilus and all the genetic and molecular studies (Caldara et al., 1996; Cataudella et al., 1974; Autem and Bonhomme, 1980; Rossi et al., 1998; Papasotiropoulos et al., 2001, 2002; Turan et al., 2005; Fraga et al., 2007; Semina et al., 2007) reported that M. cephalus shows the highest degree of genetic divergence among the other three genera. The present study is not congruent with the present monophyletic status of Liza and Mugil genera. There was no enough genetic differentiation of Chelon and Oadalechilus genera from Liza genus and C. labrosus and O. labeo seems to be belongs to the genus Liza.

The close relationship between L. saliens C. labrosus and O. labeo suggest that the separation of Chelon, Oadelechilus and Liza in three separate genera might be unnatural, making the monophyletic origin of the genus Liza questionable. In the previous studies, similar controversy was also reported. Papasotiropoulos et al. (2002) investigated the phylogenetic relationship of five mugilid species by means of mtDNA PCR-RFLP and found that L. saliens and C. labrosus were the most closely relates species.

Table 2:	Distribution and frequency of 16S rDNA haplotypes of mugilid species
H: Haplotypes

Fig. 1:	Variable nucleotide positions and frequencies of 16S rDNA haplotypes in Mugilid species. For all haplotypes variable nucleotides are indicated, while identity is shown by dashes

Table 3:	Total genetic distance between the species

Fig. 2:	Neighbour joining phylogenetic tree for 16S rDNA. Bootstrap values are shown on the tree. C. armatus was used as an outgroup

Turan et al. (2005) investigated phylogenetic relationships of the nine species of the family Mugilidae based on allozyme data and suggested that C. labrosus and O. labeo should be placed in the genus Liza. Rossi et al. (2004) and Papasotiropoulos et al. (2007) investigated phylogenetic relationships of six mugilids (M. cephalus, C. labrosus, O. labeo, L. aurata, L. ramada and L. saliens) and five mugilid species (M. cephalus, C. labrosus, L. aurata, L. ramada and L. saliens) and found low level of sequence divergence between Chelon labrosus and Liza aurata using allozyme data and three segments of mtDNA. Investigations based on nuclear and mtDNA data provide good support for the hypothesis that C. labrosus and O. labeo should be placed in the genus Liza. The existence of such disorder in phylogenetic studies of Mugilidae in the study is not uncommon (Cataudella et al., 1974; Menezes et al., 1992; Caldara et al., 1996; Papasotiropoulos et al., 2002).

The present study also revealed close relationship between M. soiuy and L. abu. The molecular phylogenetic position of M. soiuy and L. abu was considered in this study first time. L. abu is a freshwater species and M. soiuy is in fact a freshwater fish originating from Amu Darya River Basin in Far East Asia (Berg, 1965). This species was later introduced to the Sabolat (Hacibey) Lagoon, 60 km from Odessa in the Northeastern Black Sea M. soiuy found a suitable environment on the eastern Black Sea coast of Turkey, after leaving the Sea of Azov and following the Northeastern coast of the Black Sea. In time, the species migrated to the west, reaching the Sea of Marmara and Aegean Sea via the Istanbul Bogazi (Kaya et al., 1998). The present study gives new hypothesis that M. soiuy and L. abu should be considered in the genus Liza, or new genus name should be given for these two species.

16S rDNA gene sequencing data revealed a high level of genetic variability among the species examined. Only C. labrosus and L. ramada shared the same Haplotype (H₃) and the taxonomic tree support structuring of haplotypes by species designation. Any species-specific haplotypes could serve as an strict diagnostic marker for taxonomic and aquaculture practice purposes, which is especially, important at the larval and fingerling stages since the morphological and physiological characters do not show significant differences. The identification of several fish species by use of molecular markers proves a reliable tool facilitating discrimination of morphologically similar species. For example, McDowell and Graves (2002) used different mitochondrial and nuclear markers and concluded that the mitochondrial ND 4 and the nuclear locus unambiguously identified marlins, spearfishes, sailfishes and swordfishes; xiphiid billfishes that represent an important commercial and recreational fisheries resource.

CONCLUSION

The two genera Chelon and Oedalechilus should be considered together within the genus Liza. The present data also first time report the taxonomic revision the species of M. soiuy and L. abu, which should be considered together under the same genus.