MATERIALS AND METHODS

Sample collection:A total of 66 specimens of C. striata with eight to ten individuals as representatives from each of seven locations (Table 1) were collected and used in this study. Samples were caught by simple fishing methods such as nets. Collection was made from November, 2013 to June, 2014. Species level was identified at the time of collection. The GPS location of collection sites was recorded (Table 1) and locations of sampling sites were showed in Fig. 1. Fin tissues (2 x 2 mm) of each were collected and stored individually in 95% ethanol. The voucher specimens were collected in Thepsatri Rajabhat University.

Mitochondrial DNA analysis:Total genomic DNA was extracted from fin samples using an Aquagenomic Kit (Bio-Syntech, Salt Lake City, Utah, USA), according to the manufacturer’s instructions. DNA was amplifiedusing Universal Primers for mtDNA COI genes as follows:

The PCR was conducted in a Peltier thermal cycler (MJ Research Waltham, MA, USA), with the following profile: pre-denaturation at 95°C for 5 min, 30 cycles of denaturation at 94°C, annealing at 50°C and extension at 70°C for 1 min each, followed by final extension at 72°C for 5 min. The PCR products were purified using Wizard^®SV Gel and a PCR Clean-Up System by Promega (Promega Madison, WI).

Table 1:	Sample location and size (n) of Channa striata analyzed in this study

Fig. 1:	Locations of sampling sites indicated by letters in black circle. Letters code correspond to locations listed in Table 1

The purified PCR product was sequenced on an ABI3730XL genetic Analyzer (Applied Biosystems, Foster City, CA, USA).

Data analysis:Sequences were edited using MEGA 6.0v6.01 (Tamura et al., 2013) and aligned with Clustal W 1.6, implemented in the same software. Molecular diversity indices [number of haplotypes, transitions, transversions and haplotype diversity (h) and nucleotide diversities (π), besides a population comparison by pairwise F-statistics, were calculated by using Arlequin version 3.1 software (Schneider et al., 2000), in order to reveal the level of genetic variation and population structure. The program network (http://www.fluxus-technology.com) was used to construct a median-joining network.

RESULTS

A total of 396 bp of COI sequencealignments, with 66 (16.7%) variable nucleotide positionsand 59 (14.9%) parsimoniously informative sites, were obtained. Sixty of the base substitutions were transitions, whereas only six was a transversion. All the variable sites occurred at the third codon position. Altogether, 33 haplotypes were identified from all the 66 individuals sampled.

Table 2:	Population name, number of examined individuals, haplotype and nucleotide diversities and number of haplotypes in each population

Table 3:	Pairwise Fst between populations of Channa striata, based on the COI gene
*Significant Fst (p< 0.05) based on 10000 permutations of haplotype frequencies among samples

Fig. 2:	Median-joining network for the 33 haplotypes of mtDNA COI (396 bp) with circle diameter proportional to sample sizes

The mean base composition (%) was 30.5A:25.5G:23.3T:20.7C. Haplotypes in samples from Lopburi were common but not exclusive to the LB population. Fourteen haplotypes were shared by multiple populations and 19 haplotypes have been found to be population specific. Median-joining network demonstrated that haplotype clades did not correspond to the geographic distribution.

High genetic variations in each population were revealed from their haplotype and nucleotide diversity values (Table 2). Haplotype nucleotide diversities ranged from 0.750-1.000 and 0.00433-0.02672, respectively. As a whole, high mean population haplotype (0.913) and nucleotide (0.01707) diversities were observed. Genetic diversity was slightly higherin AY, AT, SB, SP and PJ regions (h: 0.857-1.000 and π: 0.00433-0.02672), than in KP populations (h: 0.750 and π: 0.02441). Population differentiation (F_ST), overlapping haplotype distribution, haplotype sharing and low nucleotide mutation among all the populations were apparent from the data. Nevertheless, genetic differentiation was observed among LB, SP, KP and PJ populations (F_ST ranging from 0.27891-0.40627) (Table 3). Figure 2 shows a Median-joining network for the 33 haplotypes of mtDNA COI (396 bp) as visualization of haplotypes relationships. The network does not indicate a differential geographic distribution of haplotypes among populations.

DISCUSSION

The haplotype diversity values inferred from mtDNA COI gene sequencerevealed high genetic diversity of C. striata popuations investigated in this study. This was supported by previous studies exhibiting relatively high population haplotype diversity of C. striata populations mainly from the Mekong and Chao Phraya rivers (0.97) inferred from mtDNA cytochrome b gene sequence (Adamson, 2010), C. striata populations in Malaysia gene sequence and (0.90) inferred from mtDNA ND5 gene sequence (Tan et al., 2013) and C. striata populations in Malaysia gene sequence and (0.80) inferred from mtDNA COI gene sequence (Song et al., 2013). Similar high genetic diversities have been observed in other freshwater fish populations; the cyprinid, Acrossocheilus paradoxus (1.00) inferred from multiple mtDNA segments (Wang et al., 2000), four Hawaiian freshwater fishes, Lentipes concolor, Stenogobius hawaiiensis, Sicyopterus stimpsoni and Awaous guamensis (0.47-0.98) inferred from both coding and non-coding regions of mtDNA (Chubb et al., 1998). This relatively high haplotype diversity observed in most C. striata populations was often noted in freshwater fishes inhabiting non-glaciated regions (during the past glaciations era) or temperate regions (Bernatchez and Wilson, 1998; Roos, 2004). Based on F_ST analyses, there was no genetic differentiation observed among AY, AT, SB, LB and SP populations. This might result from high gene flow. However, genetic differentiation was observed among LB, SP, KP and PJ populations. These populations were incompletely divergent, evident by haplotype distribution, haplotype sharing and low nucleotide mutation. Similar haplotype sharing have been found in C. striata population between the East Peninsular Malaysia and Sumatra populations and this result revealed extraordinary migration ability of C. striata (>500 km) through ancient connectivity (Tan et al., 2013).

The low level of genetic divergence within geographical regions which is contradictory to the low migratory behavior of C. striata (Halls et al., 1998; Amilhat and Lorenzon, 2005), implies the existence of factors other than free gene flow. The possible alternative factors are either recent population expansion with insufficient time for coalescence (Wang et al., 2000) or interconnection of the areas studied. Furthermore, this region is also liable to high level of flooding, thereby facilitating lateral dispersal to flood plains, watersheds or between adjacent populations (Hurwood and Hughes, 1998; Wang et al., 2000).

Due to their high economic value, the conservation of C. striata species is of great concern. Based onFST analyses, our analysis demonstrated the suitability of partial COI genes for determining the genetic diversity in C. striata population. The newly obtained sequences provided relevant information for identification of Channa species in the future. Nevertheless, based on the limited genetic information, all inferences should be reconfirmed using further, faster evolving markers, such as microsatellites.

Based on its high genetic variability fish populations in the river basins of cental Thailand are the mostappropriate wild species candidate population to receive priority as a baseline stock for selective breeding. Information on stock structure is required for conservation management of native populations of C. striata. The information of stock structure is vital for planning management and conservation of natural resources, in addition to being useful in genetic improvement programs (Abdul Muneer et al., 2009).

Furthermore, habitat fragmentation potentially cause further losses in genetic diversity or an increase in inbreeding depression. Therefore, habitat protection from agricultural and industrial activities, including the development of surrounding areas which prevent population connectivity, must be carefully regulated. As an alternative, the introduction and improvement of C. striata aquaculture is a valuable option for ensuring the maintenance of this species. Nevertheless, hatchery facilities and operation, as well as brood-stock selection, must be systematically controlled, in order to prevent fugitive genetic contamination into the wild and other adverse genetic risks (Mahidol et al., 2007).

CONCLUSION

Haplotype and nucleotide diversity values of COI gene sequences in each populations demonstrated high genetic variation of striped snakehead fish (Channa striata) in river basins of central Thailand. The haplotype clades were not corresponding to the geographic distribution of this species. The high genetic variations of striped snakehead fish in this area are probably influenced by high gene flow between regional population due to fish dispersal by flood and aquaculture. Based on their high genetic variability and positive contribution towards genetic variation in this region, AY, AT, SB, LB and SP populations of wild species are the most appropriate candidates to receive priority as a baseline stock for selective breeding. Overall, this study provided highly relevant genetic information for planning management, conservation and ranching guidelines for C. striata in river basins of cental Thailand.

ACKNOWLEDGMENT

This work was supported by the Higher Education Research Promotion and National Research University project of Thailand, Office of the Higher Education Commission (grant number: 2557A13162001). Spacial thanks to Miss Pongpun Prasankok for data analysis.