Multiple population structure of the giant eel Anguilla marmorata in Thua Thien Hue, Vietnam base on COI sequences

The giant mottled eel Anguilla marmorata is a species of great economic, ecological, and conservative values in the Southeast Asian region. The research aims to conservation and evaluation of the genetic diversity of A. marmorata populations living in Thua Thien Hue, Vietnam. The sequencing of the barcode region of the cytochrome c oxidase subunit 1 gene was carried out for 48 individuals of A. marmorata , which were collected from ve different ecosystem regions. The sequences were analyzed using various genetic, phylogenetic, and population analyses to assess their variability. A total of 20 polymorphic sites and 17 haplotypes were identied. Very low xation (Fst = -0.073 – 0.003; < 0.05) was found for A. marmorata populations in the region. The populations showed signs of recent populations’ expansion, besides negative Tajima’s D test, Fu’s Fs, Fu and Li’s D* and F* test values. The genetic evolution of eel occurs in a randomized pattern over a large population, with the likelihood of rare alleles appearing in the population. In Thua Thien Hue (Viet Nam) territories, we indicated the ve areas for conservation units with moderate eels’ populations’ diversity. To assess the genetic diversity of the A. marmorata population distributed in Thua Thien Hue (TTHue), Vietnam regarding the distribution environment and examining the evolution and development of populations, identifying conservation units for A. marmorata in TTHue, we designed this study basing on the amplication and sequencing of polymerase chain reaction (PCR) for the COI subunit of A. marmorata collected in 5 different ecological regions. and 16S rRNA genes (Gagnaire et al. 2009). COI gene barcode region exhibited a credible population-based diversity that exceeded that obtained with sequencing the mitochondrial Cytb gene, showing a percentage of haplotypes from the total sequences equaling 45% for COI, versus 15% for Cytb (El-Nabi et al. 2017). For the rst time in Viet Nam, we could identify the level of genetic diversity of A. marmorata based on COI markers basic. Results of the COI gene showed high levels of haplotype diversity (h = 0.7556–0.9000) and low levels of nucleotide diversity (π = 0.00161–0.00242). 17 haplotypes were identied in the 48 samples analyzed (35.42%). It is a similar trend with some other researchers on Anguilla species. In A. marmorata, Famil et al were found 44 haplotypes, h = 0.937 ± 0.013 and π = 0.861 ± 0.002 (%) in Indonesia base on analyzed the Cyt b gene (Fahmi et al. 2015) and 14 haplotypes, nucleotide diversity (p = 2.156%) and haplotype diversity (h = 0.780) from mtCR sequences in Pohnpei and Kosrae (Donovan et al. 2012). On another hand, 129 haplotypes were also identied with h = 0.92–1.00 and π = 0.13–1.06% for 6 other species of tropical eels (A. interior, A. nebulosi nebulosi, A. bicolor pacic, A. bicolor bicolor, A. celebesensis, A. borneensis) in Indonesia determined by Famil et al. (2015). El-Nabi et al (2017) found 44 segregating nucleotide sites, 33 haplotypes, haplotype diversity (h = 0.94), and nucleotide diversity (π = 0.008) when analyzed the COI of A. anguilla with 525 nucleotides in length (El-Nabi et al. 2017).

A. marmorata has a catadromous life-history strategy (Arai 2016) with distances from several hundred to thousands of kilometers (Arai 2014). During migration between oceans and freshwater during special stages of the life cycle, strong environmental changes have shaped not only their physiological characteristics (Wang et al. 2014), (Le et al. 2009) but also the genetic structure of eels (Ishikawa et al. 2004); (Fahmi et al. 2015); (Li et al. 2015); (Pavey et al. 2015); (Laporte et al. 2016), (El-Nabi et al. 2017). Besides, upheaval conditions of river management, water containment, over shing, and pollution may have in uenced the movement of upstream eels (Wasserman et al. 2011), (Le et al. 2009) and downstream (Linh et al. 2010), (Huyen et al. 2012), (Huyen and Phu 2015) led to the increase of population decline risk of the giant mottled eel in the wild. Although in the world A. marmorata is ranked at LC level (least concern) (Pike et al. 2019), but it has been listed in the Viet Nam Red Data Book as VU (Vulnerable) lever since 2007 (Ministry of Science Technology and Environment 2007).
DNA barcoding provides an e cient method for species-level identi cations (Hubert et al. 2008), to address questions relating to the ecology and evolution of natural systems (Kress et al. 2015) based on short, standardized gene regions (Hebert and Barett 2005), (Hebert et al. 2003). For animals, markers located on the mitochondrial genome, such as cytochrom b (cytb) gene, cytochrom c oxyase 1 (COI) gene, and D-loop region…. often have been recommended using by scientists (Kress et al. 2015). In which, a short DNA sequence in the mitochondrial gene encoding COI with 600 base pairs (bp) (Hebert and Barett 2005), (Hebert et al. 2003) has been accepted as a practical, standardized, species-level DNA barcode for many groups of animals (Hebert et al. 2003), (Kress et al. 2015). To assess the genetic diversity of the A. marmorata population distributed in Thua Thien Hue (TTHue), Vietnam regarding the distribution environment and examining the evolution and development of populations, identifying conservation units for A. marmorata in TTHue, we designed this study basing on the ampli cation and sequencing of polymerase chain reaction (PCR) for the COI subunit of A. marmorata collected in 5 different ecological regions.

Ethics statement
All animal protocols were approved by the Committee on the Ethics of Animal Experiments of Hue University, Vietnam (permit No. DHH2019-02-113), and were performed strictly with the Guide for shing capture and animals of Institute of Biotechnology. Animals were shing capture by shermen and allowed the Provincial Department of Fisheries.

Collection of specimens
Wild adult specimens of A. marmorata were collected from 5 localities in Thua Thien Hue, Viet Nam, including Phong Dien (PD), Thao Long dam (DTL), Truoi dam (DTR), Nam Dong (ND), Bu Lu and Lang Co (PL) from October 2017 to October 2018 (Fig. 1). A. marmorata species were identi ed using the keys developed by Watanabe (2004). Tissues from the adductor muscle were dissected from fresh specimens, preserved in 95% ethanol, and frozen at -80 °C until DNA extraction.

Sequence alignment and molecular phylogenetic analysis
For all sequence analyses, COI genetic similarities were evaluated using the Basic Local Alignment Search Tool (BLAST) (http://www.ncbi.nlm.nih.gov/BLAST) to identify A. marmorata sequences. The raw DNA sequences were edited using BioEdit (Hall 1999) and the pairwise, as well as multiple alignments of sequences, which was performed using ClustalW (Larkin et al. 2007) alignment editor. Multiple sequence alignment was also checked manually, and the consensus sequences were obtained. Molecular phylogenetic analyses were performed in MEGA. X software .
Sequence data were subsequently analyzed for Neighbour Joining methods with bootstraps of 1000 replicates (Felsenstein 1985). Nucleotide composition analysis was carried out using BioEdit. Transition and transversions were equally weighted. The nal consensus sequences were submitted to the National Center for Biotechnology Information (NCBI) database with accession number -MN067923 to MN067970. Standard genetic diversity indices, such as the number of haplotypes, polymorphic sites (S), haplotype numbers (h), haplotype diversity (Hd), nucleotide diversity (π), population mutation rates based on the number of segregation sites (θω) and mean the number of pairwise differences (θπ) was calculated with DnaSP 5.0 (Librado and Rozas 2009). Statistical analysis to distinguish DNA sequences evolving randomly (neutrality) with those evolving under a non-random process was done using Tajima's D (Tajima 1989); Fu's Fs tests; Fu & Li's Fs test and Fu & Li's D test (Tamura and Nei 1993). The Network ver. 5.0 software (http://www. uxus-engineering.com/) was implemented to estimate phylogenetic relationships among the unique haplotypes. Genetic differentiation, genetic distance, and migration rate among the populations were estimated by calculating the F statistic (Fst) between the populations and testing their signi cance with 1000 permutations.

Genetic variation
Sequences of the 843 bp COI gene were determined in 48 specimens, 20 polymorphic sites, and 17 haplotypes were detected. 12 haplotypes were found in only one population, 2 (H1 and H9) were found equally between two populations, only 1 (H7) was found in three populations, and 2 (H4 and H5) were found in all ve populations (Table 1 and Fig. 2). Haplotype positions were showed in Table 1.  39  75  171  174  189  210  237  270  342  387  420  582  585  611  657  690  747  804  808   H1  MN067923,  MN067925,  (Tajima 1989), (Tamura and Nei 1993). The results in Table 2 show that all exposed exhibited signi cant negative values for all neutrality tests with p < 0.05. These indicators support standing 'expansion in these areas. The population of eel in TTHue evolved by a random selection expanded population, and rare alleles appear in the population with high frequency. Also, the value of Fu and Li's D * = -2.03402 (p < 0.05) indicated that in the study population, many individuals were showing a big difference compared to other individuals in the population. Besides, when considering the values of Tajima's D test, Fu's Fs, Fu and Li's D * and F * test in 5 individual populations according to the sampling area, the trend was similar but the value p > 0.05 except for Fu's Fs results for the COI gene in the PL. This shows that the difference in the individual eel is only signi cant when selected with large populations (14 individuals or more).

Haplotype network analysis
The median-joining network (Figure 2) illustrates the polymorphic sites, including the number and frequency of the haplotypes for COI sequences. The COI network was radial-like with a high number of unique haplotypes closely related to 1 central haplotype (H4). Dominant haplotype H4 accounted for 27.87% (20/48) of all 48 specimens. This also indicated that H4 is an ancestral haplotype. H5 also showed the presence of all sub-populations in the study area but with little correlation with other haplotypes, which indicated the speci city of it. The haplotypes tended to be dispersed in comparison to the central haplotype, revealing a tendency to generate discrepancies among independent individuals. The giant mottled eel A. marmorata populations in TTHue are strongly correlated with each other, and there seems no separation of an individual population.
The evolutionary history was inferred using the Unweighted pair group method with arithmetic mean (UPGMA) (Sneath and Sokal 1973) in gure 3 with 1000 replicates of the bootstrap test (Felsenstein 1985). The sum of the branch length is 0.07028378. The tree is drawn to scale, with branch lengths in the same units as those of the evolutionary distances used to infer the phylogenetic tree. The evolutionary distances were computed using the Nei-Gojobori method (Nei and Gojobori 1986) in MEGA X . The UPGMA tree showed that most of the haplotypes were weakly associated (less than 50% bootstrap support) or unresolved, the differences among which were possibly due to low nucleotide. Haplotypes of COI of A. marmorata in TTHue were clustered into 2 obvious branches. H8 and H14 stand in the same independent branches with the remaining 14 haplotypes with bootstrap support of 56%. H5 and H15 together with H12 and H13 were grouped into 2 subgroups in branch 2 with the highest correlation (bootstrap> 60%). H4, H7, and H17 also form a group but have an extremely low bootstrap support rate of only 17%. The remaining haplotypes have a discrete distribution in the branch.

Discussion
A. marmorata with long life and long migrate distance could have a strong population genetic structure. Genetic diversity within and between populations provides a potential genetic resource for future adaptation and can be vital for the tness of a population (Xu et al. 2012). It is mainly explained by several historical and contemporary processes, such as genetic drift, effective migration, natural selection, fragmentation, and range expansion (Slatkin 1985). Earlier works have analyzed eel population genetics-based mainly on allozymes (Pantelouris et al. 1970(Pantelouris et al. , 1971 and demonstrated genetic differentiation between geographical locations. The genetic structure of A. marmorata was later reassessed based on the region of the Cyt b (Fahmi et al. 2015), mitochondrial (mt) DNA (Ishikawa et al. 2004), mitochondria control region (mtCR) (Donovan et al. 2012) and 16S rRNA genes (Gagnaire et al. 2009). COI gene barcode region exhibited a credible population-based diversity that exceeded that obtained with sequencing the mitochondrial Cytb gene, showing a percentage of haplotypes from the total sequences equaling 45% for COI, versus 15% for Cytb (El-Nabi et al. 2017). For the rst time in Viet Nam, we could identify the level of genetic diversity of A. marmorata based on COI markers basic. Results of the COI gene showed high levels of haplotype diversity (h = 0.7556-0.9000) and low levels of nucleotide diversity (π = 0.00161-0.00242). 17 haplotypes were identi ed in the 48 samples analyzed ( Researchers often use Fst to assess gene ow, a higher Fst value indicates a lower level of gene ow (Nm) and higher genetic differentiation among populations (Hedrick 2005). Fst re ects the level of inbreeding within populations (Wright 1984) or the extent to which populations are differentiated (Hartl and Clark 2007). The presence of genetic structure is an outcome of limited gene ow and a high level of genetic drift within each reproductively isolated group. Fst values below 0.05 indicate negligible genetic differentiation, whereas values greater than 0.25 indicate high genetic differentiation within the analyzed population (Weir 1996). Fst values of A. marmorata in TTHue, Viet Nam were signi cant but weak (Fst = -0.073-0.003; p < 0.05) lower than that was found for A. anguilla populations in Egypt and worldwide (Fst = 0-0.04) (El-Nabi et al. 2017) and high Nm values were detected among the populations. This may be because primitive and highly conservative of the COI gene in this species. However, DTR and ND showed high and signi cant Fst values (Fst = 0.073, p > 0.05); this indicated a high degree of genetic differentiation among 2 populations, which endorsed a signal of isolation due to distance. Populations may be divided by major geographic barriers such as land barriers and oceanographic patterns.
The negative Tajima's d test and Fu's Fs neutrality test values were obtained for all tested populations, thus rejecting the null hypothesis of neutral evolution of the COI marker. This indicated that most populations of A. marmorata in TTHue, Vietnam have been in expansion. These results also showed a similar trend of variation when compared with Famil et al. (2015) and El-Nabi et al. (2017) studies.
Results of the median-joining network of the COI gene showed several ambiguous connections among the 5 populations. The demographic analyses using UPGMA for all haplotypes show 2 similar branches one branch contains two types of haplotypes (COI) from two populations ND and PD and the other branch contains most of haplotypes from the remaining 5 populations. The rami cation of the haplotypes has a low Boostrasp's support value. This can be explained by several reasons. The rst, Anguilid have a catadromous life-history strategy, spawning in remote tropical seas with larvae that are transported back by currents to their nursery grounds in freshwater or coastal areas (Arai 2016). The golden eel growth stage may be as short as two to three years in warm productive habitats, but about six to 20 years or more in more northerly latitudes, e.g. in the Pearl River, China (Williamson and Boëtius 1993). Intense migration and long-lived habitats in the inland ecosystems have driven the formation of haplotypes independently of individual species. Secondly, A. marmorata, a catadromous eel, migrates upstream on nights, following the lunar cycle. The dramatic environmental changes between ocean and freshwater during particular phases of their life cycle shape their physiological features, e.g. visual sensitivity, olfactory ability, and salinity tolerance (Wang et al. 2014). This can also be proposed as a cause of the occurrence of rare genes and single haplotypes. This induces then isolation by time (IBT) of spawning groups.
IBT causes a restriction in gene ow, taking place between early and late spawners (Hendry and Day 2005). And the last, based on analyzed three different phylogenetic trees of A. marmorata population in Thua Thien Hue, Vietnam, Huyen and Linh (2020) showed that there was the high genetic similarity of individuals in eel populations in Thua Thien Hue and it was divided into two separate groups that are guided by the migration process and speci c ecological (Huyen and Linh 2020).

Declarations
Availability of data and materials Median-joining network for COI haplotypes of A. marmorata in Thua Thien Hue, Viet Nam. On the connecting lines, red numbers present the variable sites between each haplotype pair. Different colors represent the different populations in the network (Yellor, Green, Blue, and Pink colors represented for DTL -Thao Long Dam, DTR -Truoi Dam, ND -Nam Dong and PD -Phong Dien respectively. PL -Phu Loc, include Lang Co and Bu Lu, was indicated by red color).