Fish sampling

samples were collected from fishermen at three sites herein called three populations, including Thanh Binh district, Dong Thap province (DT), Chau Thanh district, An Giang province (AG), and Ke Sach district, Soc Trang province (ST). These sampling sites (Figure 1) are located at the two main river branches, Tien (DT) and Hau branch (AG and ST), of the Mekong River. The fish were identified based on morphological keys11, 12. This pangasiid species is characterized by a small body size (18-20 cm), large eyes, and long barbels reaching the pectoral-fin base. Samples from the three populations (n=23 for DT, n=20 for AG, and n=29 for ST) were chosen for genetic analysis. Fin clips of these samples were preserved in 95% ethanol until DNA analysis.

Genetic diversity analysis methods

DNA was extracted from fin clips by using the ammonium acetate protocol13. Approximately 25 mg of fin clip from each fish was first cut and then placed in an Eppendorf tube containing 650 µl of lysis buffer solution and 5 µl of Proteinase-K. After being incubated overnight at 55°C, all samples were added to 240 µl ammonium acetate and then incubated at 4°C for 30 minutes to precipitate the protein. Next, the tube was centrifuged at 10,000 rpm for 15 minutes two times, and 650 µl of the above solution was pipetted into a new tube. Then, 100% ethanol was added to the tube, incubated, and centrifuged to obtain DNA, which remained at the bottom of the tubes. Finally, the tube was air-dried, and 100 µl TE was added to obtain DNA.

Two samples from each population were selected randomly for primer screening. A total of 15 primers from ISSR libraries were screened. Six primers that had high polymorphisms and were visible on gels were chosen (Table 1). PCR amplifications were conducted in a 10 μL mixture containing 5 μL Promega PCR Master Mix (including Taq DNA polymerase supplied in a proprietary reaction buffer (pH 8.5), 400 μM dNTPs, and 3 mM MgCl), 0.4 μL primer (10 μM), 1.0 μL DNA, and 3.6 μL nuclease-free water. Thermal conditions for PCRs included one denaturing cycle at 95°C for 5 minutes, exceeded by 38 repeated cycles of 95°C for 30 seconds, annealing temperature (Ta) from 44°C to 50°C (Table 1) for 40 seconds, extension at 72°C for 1 minute; and one final extension has one cycle at 72°C for 5 minutes.

PCR results after amplification were checked with 1.2% agarose electrophoresis. Each gel contained a 1 kb DNA ladder (ABM, Canada) and individuals representative of the three populations. The electrophoresis process was carried out at 50 V for 80 minutes. Then, the gels were dipped in ethidium bromide (5 μg/mL) solution for staining for 15 minutes. The bands were visualized by scanning in a UV transilluminator, and band sizes were estimated based on the DNA ladder.

ISSR bands of each gel were scored by using a binary data matrix, ‘1’ for the presence of the band and ‘0’ for the absence of the band. To minimize errors in band scoring, only clear and intensified bands were scored (low-intensity bands were not scored18, and scored bands were compared between gels of the same primers). Furthermore, scoring was performed independently by two lab technicians, and the results were then discussed to finalize the data.

Genetic diversity parameters, including the effective number of alleles (Ne), percentage of polymorphism, private alleles, unbiased expected heterozygosity (uHe), and Shannon index (I), were estimated for each population by using GenAlEx 6.5 software19. Genetic differences among the three populations were evaluated based on the overall genetic difference (G, Nei’s genetic distance, and principal coordinates analysis (PCoA). In addition, a dendrogram was constructed based on the unweighted pair-group method with arithmetic average (UPGMA) method using Popgene version 1.3.120 and MEGA 6.021.

ISSR variability

There were 72 samples from three populations using six primers (Table 1) for ISSR analysis. A total of 67 bands were generated ranging from 390 bp (17898A) to 3000 bp (Chiu-SSR1, Chiu-SSR2, and Micro11) (Figure 2). The number of bands of each primer fluctuated from 7 (ISSR15) to 15 (Micro11). Band numbers from each population ranged from 65 bands (AG) to 66 bands (DT and ST), in which one private band was found in the ST population.

Genetic diversity parameters of the three populations

The genetic diversity parameters of are presented in Table 2. The percentage of polymorphic loci varied from 80.6% to 95.03%. The number of effective alleles (Ne) fluctuated from 1.381 to 1.454; the Shannon index (I) ranged from 0.369 to 0.435, and the unbiased expected heterozygosity (uHe) ranged from 0.243 to 0.286. The ST population had the highest genetic diversity values (Ne=1.454±0.038, I=0.435±0.024, and uHe=0.286±0.019), and the lowest was the AG population (Ne=1.381±0.040, I= 0.369±0.029, and uHe=0.243±0.021).

Genetic differences among the three populations

The overall genetic difference among populations (G) was 0.053, and the number of migrants per generation (Nm) was 8.90. The values of Nei's unbiased genetic distance and genetic identity among the three populations ranged from 0.015 to 0.035 and from 0.966 to 0.985, respectively (Table 3). The genetic distance between the ST and AG populations was the largest, and the AG - DT pair was the lowest. In addition, the genetic relationship among populations of this fish was shown by the UPGMA dendrogram using Nei's unbiased genetic distance (Figure 3). The AG and DT populations are more genetically similar than the ST population.

The molecular analysis of variance (AMOVA) showed that the genetic variation within populations (97%) was larger than that among populations (3%). Principal coordinates analysis (PCoA) indicated the genetic difference among the three populations in a two-axis (coordinate) plot, with 8.54% and 6.51% of the total genetic variation explained by coordinates 1 and 2, respectively (Figure 4). The AG and the majority of individuals from DT populations were clustered together in the same region, whereas most individuals from the ST population were distributed further apart from the plot. The ST population, followed by the DT population, was more widely scattered on the coordinate plot than the AG population, indicating that these two (ST and DT) populations had larger levels of genetic diversity.

Levels of genetic diversity of populations

The study found that all populations of had high levels of genetic diversity (with Ne from 1.381 to 1.454, I from 0.369 to 0.435, and uHe from 0.243 to 0.286) in comparison with other fish species based on ISSR markers. A study on showed lower values of genetic diversity parameters, with Ne from 1.352 to 1.376, I from 0.299 to 0.320, and uHe from 0.208 to 0.22222. In yellow catfish ()23, the results of Ne ranged from 1.225 to 1.284, lower compared to the present study. On the other hand, has a higher and equivalent level of genetic diversity than some scale fish species. Kissing gourami () had less genetic diversity than (Ne from 1.295 to 1.387, I from 0.269 to 0.386)24. In addition, genetic parameters based on eight ISSR primers of three populations in the Mekong Delta were comparable to with Ne ranging from 1.422 to 1.458, I from 0.364 to 0.395, and uHe from 0.250 to 0.271 25. Although the AG population of had lower genetic parameters than the other two populations, the magnitudes of the differences were small, ranging from 3% to 5% for Ne and from 9% to 15% for I or uHe.

The genetic diversity of a population depends on internal factors of the population, such as population size and random genetic drift, and external factors of the gene exchange process due to the migration of individuals among populations4. The level of genetic diversity has a positive correlation with population size (or the number of individuals)26. Compared to other species, can be found in abundance, especially during the flooding season (from July to November) in tributaries of the Tien and Hau Rivers6, 27. Based on the catch data, 78.86 tons/year in Tien Giang Province, it can be predicted that has a large population size in the Mekong Delta, resulting in the high level of genetic diversity of this fish. In small populations, random genetic variation is powerful, causing some genes to be completely lost 28. In kissing gourami (), a low level of genetic diversity was observed in a closed population resulting from a small population size24. Moreover, the ability to exchange genes during migration is also a factor that can increase the genetic diversity of the population4. in Brazil was detected to have low genetic diversity, which might be related to the sedentary habits of the species, reducing gene flow between populations of different locations29.

Genetic differentiation among populations

The genetic differentiation of the populations was low, as indicated by the low values of G (0.053) and Nei’s unbiased genetic distances and the high value of the number of migrants per generation (Nm=8.9). These parameters are indicative of genetic exchange among populations4. Nei’s unbiased genetic distances among populations (from 0.015 to 0.035) were lower than those of other freshwater species in the Mekong Delta. The Nei’s values ranged from 0.023 to 0.102 for kissing gourami 24 and from 0.022 to 0.057 for (with G = 0.09)25. Baird et al. (2001) reviewed different hypotheses about the migration patterns of . The fish can migrate upstream from the Mekong River floodplain for spawning and from the main river to smaller streams and tributaries. Because of this migratory behavior, genetic flow occurs strongly. Consequently, gene flow reduces genetic differences between populations and increases the variation within populations4. An example is Malayan leaf fish (), which migrate during the flooding season from the Mekong River into wetland areas and then migrate back to the river in the dry season and have low genetic differences among populations30. In addition, another study found that the low genetic distance and weak genetic structure of were affected by their migration behavior22.

The genetic differences among the three populations of may be affected not only by their migratory behavior but also by the geographical distance and the ecological conditions of the rivers. The geographical distance of the ST population and the other two populations was furthest, corresponding to larger genetic differences of ST-AG and ST-DT. Meanwhile, sampling sites in AG and DT are located in the Hau and Tien Rivers, but they are connected by the Vam Nao River. This hydrological connectivity allows gene flow between the AG and DT populations, resulting in their low genetic difference. Similarly, in wild populations in the Mekong Delta, a population with the greatest hydrological distance had the largest genetic distance from the other populations 25. On a geographically regional scale, Duong et al. (2019) reported that genetic differences among wild striped snakehead () populations in Vietnam and Cambodia were positively correlated with hydrological distances 31. Moreover, the distribution of river systems and ecological conditions can affect the migration of fish and ultimately the population structure32. The differences in river systems and associated ecological factors were determinants of the genetic differentiation of striped snakehead populations33 and climbing perch strains in Thailand34.

The findings of the present study have implications for the domestication and wild population management of In the Mekong Delta, domestication of has been in progress with preliminary results of artificial reproduction (unpublished information). Because of their high genetic diversity, the three populations, of which the ST population is relatively better, could be good resources for domestication. Furthermore, combining these groups would boost genetic variety in a base population, allowing for more effective domestication programs35. In terms of wild fish management, low genetic differentiation among the three populations (G=0.053) or large gene flow (Nm=8.9) implies that they can be considered a “panmictic population”, in which free genetic exchange occurs among individuals from different locations36. In this case, it is suggested that they can be managed as one population37.