In total, 114 Silver pomfret individuals were collected from the northern waters in Kuwait, Sonmiani Bay, Ormara, Pasni, Xiamen, and Taiwan between 2010 and 2014 (Figure 1). All individuals were identified according to their morphological characteristics (Li et al. 2013) to ensure the accuracy of the species identification. Then, the back-muscle tissues were excised and preserved in 95% alcohol for the subsequent experiments.

Figure 1. 

Sampling locations of P. argenteus.

Genomic DNA was isolated from muscle tissue by proteinase K digestion and extracted with Qiagen DNeasy kit. The extracted DNA was assessed by 1.5% agarose gel electrophoresis and stored at –20 °C for PCR amplification. The mtDNA CR was amplified with the primers F-gao: 5'-GAAGTTAAAATCTTCCCTTTTGC-3' (forward), and R-gao: 5'-GGCCCTGAAGTAGGAACCAAA-3' (reverse). Each PCR was performed in a 25 μL reaction mixture containing 17.5 μL of ultrapure water, 2.5 μL of 10× PCR buffer, 2 μL of dNTPs, 1 μL of each primer (5 μM), 0.15 μL of Taq polymerase, and 1 μL of DNA template. PCR amplification was performed in a Biometra thermal cycler under the following conditions: 5 min of initial denaturation at 95 °C; 30 cycles of 45 s at 94 °C for denaturation, 45 s at 50 °C for annealing, and 45 s at 72 °C for extension; and a final extension at 72 °C for 10 min. PCR products were purified, and both strands were sequenced. The newly isolated nucleotide sequences were deposited in GenBank under accession numbers MF402948–MF402998. Two CR sequences of Pampus chinensis (Euphrasen, 1788) were used as the out-group.

CR sequences were edited and aligned using DNASTAR software. Polymorphic sites, haplotype number, and molecular diversity indices for each population were calculated using ARLEQUIN version 3.5 (Excoffier et al. 2005). Genetic relationships among haplotypes were reconstructed using the neighbor-joining (NJ) method implemented with 1000 replicates in MEGA 5.0 (Tamura et al. 2011). Analysis of molecular variation (AMOVA) was performed using ARLEQUIN to investigate the partition of genetic variation among the populations. An unrooted minimum spanning tree (MST) was constructed via the MINSPNET algorithm as implemented in ARLEQUIN to show the relationship among haplotypes and subsequently drawn by hand (Excoffier et al. 2005). Historical demography/spatial expansions were inferred by neutrality testing and mismatch distribution analysis, as implemented in ARLEQUIN. Deviations from neutrality were evaluated using Fu’s F_S and Tajima’s D. Nucleotide mismatch distributions were applied to assess population growth and spatial range expansion. A molecular clock-based time estimate provided an approximate timeframe for evaluating phylogeographical hypotheses. Historical demographic expansions were also investigated by the examination of frequency distributions of pair-wise differences between sequences (mismatch distribution), based on three parameters: θ₀, θ₁ (θ before and after population growth) and τ (time since expansion, expressed in units of mutational time) (Rogers and Harpending 1992). The values of τ were transformed to estimates of real time since expansion with the equation τ=2×μ×t where μ is the mutation rate for the whole sequence under study and t is the time since expansion. In the present study, a sequence divergence rate of 0.5–1× 10⁻⁷/site/year was applied to the CR sequences of P. argenteus (Bowen et al. 2001). Bayesian skyline plots were created with BEAST v.8 (Drummond and Rambaut 2007). In the present study, a sequence divergence rate of 5%–10% /Myr (Bowen et al. 2001) was applied to the CR sequences of P. argenteus.

After a manual correction, the CR fragment sequences were 450~453 bp in length, including a 70-bp partial fragment of the tRNA^pro, and no variable site was detected in the tRNA fragment. After deleting the 70-bp fragment, the obtained target fragment was 380~383 bp in length, which corresponded to the 15,699–16,080 bp region of the complete mitogenome of P. argenteus (KJ569773). Thirty-eight variable sites and 23 parsimony-informative sites were assessed within the target fragment. There were 23 transitions, five transversions, and five insertions/deletions. The ratio of transitions to transversions was 4.6, indicating that the mutations in the CR sequence of P. argenteus had not reached saturation. The A+T content (70.33%) was significantly higher than the G+C content, indicating a significant AT preference.

All variable sites defined 51 haplotypes among the 114 individuals. No haplotype was shared between the six populations, and forty-two specific haplotypes were detected among all individuals, accounting for 82.4% of the total haplotypes (Table 1). The population from Ormara exhibited the most specific haplotypes (10), whereas the population from Taiwan exhibited the fewest specific haplotypes (4). Hap_27 was the dominant haplotype and was shared by 22 individuals (Table 2). High levels of haplotype diversity (h) were detected within each population, demonstrating a high level of genetic diversity in this species. In contrast, low levels of nucleotide diversity (π) were observed. Overall, the average values of the h and π were 0.932 ± 0.013 and 0.018 ± 0.010, respectively (Table 1).

A NJ tree was constructed based on the 51 CR haplotypes using P. chinensis as out-group, and two deeply divergent lineages were identified in the six populations that were not geographically concordant (Figure 2). Lineage A comprised 30 haplotypes (74 individuals), whereas lineage B comprised 21 haplotypes (40 individuals). No significant differences were observed in the haplotype distribution of the two lineages, and both lineages were found in all populations (Table 2, Figure 2). Lineage A obviously dominated the population from Kuwait (81.82%), followed by the population from Pakistan (62.07%). However, lineage B dominated the Chinese populations (41.18%) compared with its presence in the Arabian Sea populations (32.5%). The frequency of lineage A in each population was reduced from 81.82% (Kuwait) to 50% (Taiwan). Two shallow lineages were also detected in MST among the 51 CR haplotypes (Figure 3). There were no significant differences in the distribution of the haplotypes in both lineages other than the specific ones mentioned above (Table 2).

Figure 2. 

NJ tree of CR haplotypes of P. argenteus. Pampus chinensis was used as the out-group. Bootstrap supports >50 in 1,000 replicates are shown.

Figure 3. 

Unrooted minimum spanning tree depicting the genetic relationship among the CR haplotypes of P. argenteus. Circle sizes are proportional to the haplotype frequency. Perpendicular tick marks on the lines joining the haplotypes represent the number of nucleotide substitutions.

Based on the best model, i.e., TrN+G, the net genetic distance between lineage A and lineage B was 0.0058. Based on the 5–10%/MY (million years) divergence rate, the time since the population divergence occurred was estimated to be 0.06–0.12 million years ago, dating back to the late Pleistocene.

The F_ST values between six populations were low (from 0.012 to 0.062) and statistically non-significant, except for those from the Xiamen and Kuwait populations (Table 3). All results showed that genetic differentiation was not significant between different populations. The negative F_ST values suggested that the genetic differentiation among individuals was higher than that within populations. AMOVA revealed that the variability among the samples yielded an F_ST= 0.3328 (P<0.001) as one gene pool, and the divergence was attributable to 66.72% of the genetic variation among the populations. To further investigate the possible effects, the populations were partitioned into two and three gene pools, and both pools yielded a significant divergence within the populations (P < 0.001) (Table 4). Therefore, the divergence among the populations was very weak with no statistical significance, whereas the genetic divergence was mainly derived from within the populations with statistical significance. Thus, in all cases, no significant genetic structure was identified across the entire geographical sampling range of the Silver pomfret.

The observed mismatch distributions of Silver pomfret were established for the two lineages (Figure 4). The mismatch distribution of lineage B exhibited a unimodal pattern that was consistent with the expected distribution in the population expansion model. In contrast, the mismatch distribution of lineage A was bimodal with two peaks. The F_S test was significantly negative for both lineages (P < 0.05), whereas the D test was not significantly negative (P > 0.05). However, both the SSD and HRI tests were not statistically significant (P > 0.05), suggesting that the two lineages were consistent with the null hypothesis of the sudden expansion model. Thus, the studied Silver pomfret populations experienced a recent expansion event.

Figure 4. 

Mismatch distributions of control region haplotypes of P. argenteus.

The peak τ of the nucleotide mismatch distribution provides information that can be utilized to estimate the approximate time of the population expansion. In this study, the τ values of lineages A and B were 14.889 and 5.426, respectively (Table 5). Therefore, the time since the expansion of lineages A and B was estimated to be 3.9 × 10⁵–7.8 × 10⁵ and 1.4 × 10⁵–2.8 × 10⁵ years ago based on the divergence rate of the mitochondrial CR (5–10%/MY) and τ values, respectively. The ratio values (θ₁/θ₀), which estimated the effective female population sizes after and before the expansion, were 2000 for lineage A and infinite for lineage B (Table 5).

The Bayesian skyline plots revealed a detailed demographic history of population size changes, from which we could see that both lineages A and B had undergone population expansion in the late Pleistocene. The effective population size of lineage A increased slowly after the last glacial maximum (LGM) approximately 3.2 × 10⁵ years before the present, and the effective population size of lineage B increased sharply from 1.7 × 10⁵ years ago (Figure 5).

Figure 5. 

Bayesian skyline plots showing N_efT (N_ef = effective population size; T = generation time) changes over time for P. argenteus based on CR sequences. The upper and lower limits of the blue line represent the 95% confidence intervals of highest posterior densities (HPD) analysis. The black line represents median estimates of N_efT.

The genetic diversity in species is a result of the long-term evolution of organisms, and the level of genetic diversity is closely related to the survival and evolutionary potential of the species, of which h and π are two important indicators. In this study, high h and low π were detected in six P. argenteus populations, and the results supported the second population rapid growth hypothesis of marine fishes as interpreted by Grant and Bowen (1998).

Currently, the high diversity in this species may be related to the following aspects. First, this species has an extensive distribution area and varying habits. Silver pomfret are found from the Taiwan Strait to the Indian Ocean. This long coastline has created diverse marine ecological environments in which this species is successfully adaptive to local habitat conditions. Second, this species has numerous effective populations. Despite a declining trend in the amount of pomfret resources, numerous recruitment populations are available to ensure an effective population, which was evaluated by acoustic fishery resources (Jia et al. 2004). Third, this species has advantages due to its geographical distribution. Since the Quaternary, substantial increases and decreases in global temperature have resulted in a decreased genetic diversity, and the most seriously affected species are located at the southern and northern edges of the distribution area (Hewitt 1996). The distribution area of P. argenteus occurs in a relatively central location that is less affected by global glacial climate fluctuations; therefore, a relatively high genetic diversity may more easily occur.

Two lineages were tested using a NJ tree and MST based on all haplotypes. The spatial variation in the haplotype frequencies between the two lineages was absent, indicating a high degree of genetic homogeneity among the six populations. The genetic structure may be a result of both historical and contemporary processes. All population structure analyses were concordant with the null hypothesis of panmixia despite the well-defined phylogeographical structures of the mtDNA haplotypes. Numerous studies have confirmed that the phylogeographical patterns and population genetic structures of marine species are related to specific geological events or environmental factors, and the isolation due to Pleistocene glaciation is likely the main reason for the genetic differentiation of species (Liu et al. 2007; Shen et al. 2011; Han et al. 2012; Qiu et al. 2016). Therefore, we hypothesized that the phylogeographical structure of the Silver pomfret was connected to a second admixture in which its populations historically experienced exotic divergence and mixed again with differentiated populations (Avise 2000). The biomass of marine organisms decreases sharply with periodic climate fluctuations. In addition, few surviving individuals remained in the limited shelters during the LGM. Due to the rising sea levels that occurred after the last glacial period, the Silver pomfret potentially underwent a re-colonization event, and the effective maternal population size grew rapidly. The derivative populations that experienced population isolation were connected to the second admixture, which eliminated the partial genetic divergence that previously accumulated. In this case, the genetic differentiation that was detected across the distribution range of the Silver pomfret was compatible with the hypothesis of recent range expansion and insufficient time to attain migration-drift equilibrium (Slatkin 1993), which was considered the most important factor.

In addition to historical events, contemporary factors, including oceanic currents and life history characteristics of the species, are important factors that affect the genetic structure of species in marine environments. Similar to numerous marine pelagic fishes, the Silver pomfret exhibits a highly migratory behavior and large population size and dispersal potential during the early life stage, which could lead to frequent gene flow among different populations (Ward et al. 1994). In general, ocean currents play an important role in transporting the larvae of marine organisms, which could allow substantial dispersal and high connectivity among different populations (Liu et al. 2007). Thus, the Silver pomfret was hypothesized to travel a long range, given its larval stage and the current velocity near the coasts of China and Pakistan, which could sufficiently explain the low level of genetic divergence in this species.

Unfortunately, we only collected six geographical populations of the Silver pomfret, which is not enough for an even sampling throughout its entire distribution in the Indo–Pacific Ocean. The population genetics of this species may be one-sided in this study and remain to be discussed further. Therefore, the Silver pomfret samples of an intermediate distribution need to be collected and will further verify our results.

It is necessary to assess the genetic population diversity and genetic structure of marine fish for fisheries management and conservation. The contemporary genetic structure of the Silver pomfret revealed in this study can preliminarily improve genetic knowledge and provide a firm basis of fishery stocks in the Indo-Pacific Oceans. Although the Silver pomfret currently exhibits a relatively high genetic diversity, it is likely to experience a disaster similar to that experienced by traditional economic fish, e.g., Larimichthys polyactis (Bleeker, 1877), L. crocea (Richardson, 1846) and Trichiurus haumela (Forsskål, 1775), if attention is not paid to the conservation of resources. In fact, a decline in the Silver pomfret resources has been reported in some waters due to over-fishing and the devastation of marine ecology. Therefore, fishery management measures regarding the Silver pomfret must be implemented in a timely manner.