Weak Population Structure of Perna Viridis in Indo-Pacic Region based on Nuclear and Mitochondrial DNA Marker

Perna viridis is a mussel commonly distributed along the Asian Indo-Pacic coasts. It is one of the main cultured species of that region. Previous studies focused mostly on the native populations within single countries; with the present study we analyzed the genetic diversity of P. viridis in a large study area, spanning from Oman to southern China. Three molecular markers were used, namely portions of the nuclear ITS region, and the mitochondrial COI gene and D-Loop region. The nuclear marker showed moderate levels of genetic diversity (haplotype diversity h = 0.543 to 0.897) and nucleotide diversity π = 0.0022 to 0.0064); whereas mitochondrial markers exhibited higher levels of genetic variability (h = 0.858 to 0.964 and π = 0.0012 to 0.0079). The estimates of inter-sample genetic divergence (F ST ) and the analysis of molecular variance highlighted that the Thai population is genetically divergent from the others. Our results showed the genetic variation of P. viridis at the rim of South China Sea and obtained the genetic basic information of P. viridis.


Introduction
The Asian green-lipped mussel, Perna viridis (Linnaeus, 1758) (Bivalvia, Mytilidae), is distributed along the Asian coasts of the Indo-Paci c region [1]. Furthermore, P. viridis is appreciated seafood rich in EPA, DHA and microelement in its soft part in its distributed areas [2]. With shortly breeding culture and comparatively high tolerance to most environmental conditions [3], P. viridis become a major aquaculture species in many Asian countries [4,5]. The life cycle of P. viridis sustained one year from the larva to the sexual mature individual. The spawning period is from April to November and the pelagic larvae stage lasts about 17 to 25 days [6]. In this period, the ocean currents in the South China Sea are active potentially promoting P. viridis larval dispersal.
In a previous genetic study using 19 microsatellite loci, carried out in samples collected from ten locations of peninsular Malaysia, Ong et al. [7] found a general picture of genetic homogeneity within the study area. Similar results were found in the Gulf of Thailand by Prakoon et al. [5] using ve microsatellite loci. Even comparing three natural samples with two cultured populations of P. viridis by using nine microsatellite loci, no signi cant genetic divergence was observed [1]. In China, Ye et al. [8] and Ye et al. [9] also found that the genetic differentiation of the populations of P. viridis was non-signi cant and low based on the mitochondrial DNA and microsatellite markers. It was di cult to investigate whether the local samples were affected by other countries or regions in the South China Sea and showed a non-signi cant divergence. However, most previous reports of the genetic characterization of P. viridis focused on the relationship between location samples in a restricted study area. None of report studied the individuals of P. viridis between different sea areas at the rim of South China Sea. The genetic relationship and population structure of different P. viridis populations are ambiguity among the natural habitat in the vast South China Sea.
In this study, portions of one nuclear DNA marker (internal transcribed spacers, ITS region) and two mitochondrial DNA markers (Cytochrome C oxidase subunit I, COI; mitochondrion control region, D-Loop region) were employed to investigate the genetic diversity of P. viridis in an area spanning from Arabian Sea to South China Sea. With high variable region and easily sequenced, these markers were reliable and common tools for studying genetic diversity and evolution in most species and widely used in mollusk species to estimate interand intra-populations genetic variability [10][11][12].. Thus, it was bene cial to use different markers for revealing the genetic diversity and genetic structure between different location samples. We explored whether the genetic differentiation of P. viridis were similar at the rim of South China Sea. As a reference, several individuals from Oman also were tested to compare with the results of P. viridis in the East Asia and Southeast Asia. We expected the study could reveal the genetic relationship of wild samples of P. viridis at the rim of South China Sea and provide materials of P. viridis for the sheries management.

Material And Methods
Sample collections and DNA extraction Six samples of Perna viridis were collected in Zhangzhou (ZZ), Zhuhai (ZH), Zhanjiang (ZJ), Beihai (BH), in the South China Sea; Suratthani (ST) in the Gulf of Thailand; and Muscat (MC) in Arabian Sea (Table 1 and Fig. 1).
The adductor muscle was removed from each individual, xed in absolute ethanol and stored at -20°C until DNA extraction. The improved salting-out method was employed to extract genomic DNA [13]. After extraction, DNA quality was assessed using 2% agarose gels electrophoresis with Super Stain (Cwbiotech Co. Ltd., Peking, China). With the NanoDrop 2000 Spectrophotometer (Thermo Scienti c), DNA concentration was evaluated by absorption at 260/280 nm. DNA was diluted to a nal concentration of 50-60ng/µl and stored at -20°C for further analysis.

Sequencing
The sequences of ITS region were obtained by using primers ITS5 (5'-GCATCGATGAAGAACGCAGC-3') and ITS28 (5'-GGAATTCTCCTCCGCTTATTGATATGC-3') [14,15] All the PCR products were tested in the electrophoresis on 1.5% agarose gels. Then all of them were sent to the sequencing company to bi-directional sequence (BGI Tech Solutions Co., Ltd., shanghai, China).

Data analysis
The software ClustalX v2.0 [16] was used to align the DNA fragments. Each fragment was compared with complete genome of P. viridis online (https://www.ncbi.nlm.nih.gov/) to verify the species and the fragment position.
The sequences of ITS, COI and D-Loop were performed the tests as followed (COI gene and D-Loop region were subsequently concatenated and investigated to consider). Haplotypes were distinguished using the program DnaSP v6.0 and this program was also used to group the specimens and estimate the genetic parameters [17].
To estimate the optimal nucleotide substitution model, the software jModelTest v2 was used to perform the model test [18]. was utilized to calculate the Nm (Effective number of migrants per generation) [20]. The program MrBayes was used to construct the Bayesian tree of the haplotypes [21]. The program Phylip was used to plot the ML tree of the haplotypes [22]. using R packages abe4. A median-joining network of haplotypes was built using the program Network v. 4.6.1.0 [25]. To analyse the occurrence of genetic structuring, we used the software BAPS v6.0 [26][27][28]. This program was able to detect panmictic clusters of genetically similar individuals. BAPS was set with six replicate runs for each value of k (the maximum number of genetic cluster) up to k = 20. In addition, we set several reference individuals n = 200 and repeated the admixture analysis 20 times for individual.
The software DnaSP v. 6.0 was used to obtain mismatch distribution of mitochondrial DNA genes and constructed the plot of mismatch distribution by R packages ggplot2. The parameter of population expansion (τ) was estimated by the general non-linear least-square method and the con dence interval was calculated by the parametric bootstrap approach method [29]. The formula t = τ / 2µk (k: the length of the sequence; µ: the mutation rate of the gene) was used to estimate the actual expansion time (t). The generation time was one year for the P. viridis. Due to lack of speci c record in green-lipped mussel fossil and the molecular clock of different makers was different, in this study, 1%~2%/Ma was chosen as evolution rate to calculate a conservative time range of expansion [15].

Genetic diversity analysis
There are 132 sequences were aligned and classi ed. Hap1 was the main haplotype accounting for 41.67% with 55 individuals (Fig. 2). Most individuals were gathered in Hap1, Hap2 and Hap25. Other haplotypes appeared in few or single individuals. Using the optimal model (GTR + I model), 12.80% of variation was attributed among populations and 3.20% of variation was among populations within groups. Apart from that, the genetic difference within populations explained 84.00% of the total variation (Table 2). In Φ -statistics, Φ CT = 0.128 (P > 0.05), Φ SC = 0.037 (P < 0.05), Φ ST = 0.160 (P < 0.05). This result suggested lacking adequate genetic exchange within populations would cause the divergence between three groups and the in uence of genetic divergence was resulted by inter-populations no inter-groups.  Muscat were obviously divergence with others. In the plot of haplotypes network, all haplotypes were clustered into two parts and there was a difference between samples from Muscat and Suratthani (Fig. 2). Haplotypes structure of MC was different from that of the samples of China and Thailand, which means that population of MC lacked genetic exchange with others. Based on the Bayesian clustering method to simulate and evaluate genetic structure, when K = 2 (Fig. 4), MC was the rst site isolated from other sites. It supported that except MC, other populations had widely genetics exchanges among them and population of MC limited genetics exchange with other populations. Therefore, the best K value supported two clusters were the most reliable clustering in the population structure.

Evolution And Dynamics Analysis
Due to the non-coding region of ITS region and D-Loop region, sequence of COI was selected to analysis the expansion. The detected D and F values were negative in ST population and the SSD value was non-signi cant. The plot of mismatch distribution con rmed the model of population expansion (Fig. 5). Thus, it revealed that only the ST population had a recent population expansion (Table 3). Based on the formula t = τ / 2µk, the expansion time was estimated between 0.2 Ma to 0.24 Ma [15]. in other ve populations were less than 10 and most of individuals were detected the Hap1. In the plots of network and Bayesian and ML tree of haplotypes of ITS region (Figs. 2 and 6), obvious divergence was not found. There was no evident difference in haplotypes between three countries. There was no discernible divergence between endemic haplotypes. It is evident from the evolutionary tree that haplotypes of six populations were mixed and clustered.
For COI & D-Loop region among six populations, all populations were estimated high haplotypes diversity (h > 0.85). All Chinese populations were found a higher the nucleotide diversity (π > 0.007) than the populations in Thailand and Oman (π < 0.003). Overall, the genetic diversity of

Evolution Dynamics
There was no indication for possible positive selection acting on the mitochondrial COI gene through analyses of the Ka/Ks test. Generally, marine animals were subjected to purifying selection and the mitochondrial genes were commonly subjected to the purifying selection to pursue thermoregulation and e cient energy metabolism [30][31][32]. This result revealed that in the China South Sea, the populations of P. viridis possibly lived under pressure of rising ocean temperatures, frequency human activities and others environmental factors.
For the neutrality test, except the lacked data of MC, only the population of ST was negative with no-signi cant SSD value ( Table 3). The mismatch distribution showed that population of ST had a recent population expansion event. However, other populations revealed no recent population expansion events. These results were possibly caused by the reasons as followed. Firstly, for Pleistocene in China South Sea, the relatively rich current system and wide ocean area could ensure the genetic exchange within populations and protect the habitats of P. viridis from destruction. Secondly, in the gulf of Thailand, due to a relatively closed environment, the population of P. viridis was limited without enough genetic exchange. In Pleistocene, the gulf was isolated with the external environment. Weak population structure could not withstand the environmental change, especially for the sedentary mussels. Last, in Thailand, most aquaculture of P. viridis derived from the same place. It was possible that some data of populations in Thailand was not collected. Recently, the globally aquaculture was in fast development. Most countries chose to develop Mollusk for its importance and economic-value species. In this tactic, more improved varieties were introduced and bred. Thus, the mantel tests did not show a signi cant relation between genetic distance and geographic distance (Fig. 7). These results suggested that the study of genetic diversity and structure of P. viridis were possibly developed into two groups and the populations of China were obviously different from the outside populations. The results showed that different between mtDNA and nucleotide DNA was due to genetically different ways of the two DNA that generated slightly different results.

Conclusion
It was important to analysis the population genetic and structure to obtain the genetic information of P. viridis. In this study, the populations of ST and MC were divergence with the populations of China. The difference among populations of China was limited. Therefore, it was good way to introduce the good breed of P. viridis in different countries to improve the resilience and yield of the native species.

Con ict of interest
The authors declare that the research was conducted in the absence of any commercial or nancial relationships that could be construed as a potential con ict of interest.