Genetic characteristics of yellow seabream Acanthopagrus latus (Houttuyn, 1782) (Teleostei: Sparidae) after stock enhancement in southeastern China coastal waters

14 Yellowfin seabream is an important economic fish that is widely distributed in the East and South 15 China seas. Many attempts to enhance stocks of yellowfin seabream have occurred in China, but a 16 lack of genetic information for this species after stock release represents an obstacle to its 17 management and conservation. To provide scientific guidance for sustainable germplasm resource 18 development, we sequence the mitochondrial DNA (mtDNA) control region (CR) of 123 yellowfin 19 seabream from 6 sample populations (Xiamen, Dongshan I, Dongshan II, Yangjiang, 20 Fangchenggang, and Beibu Gulf). Populations of both wild and cultured yellowfin seabream have 21 high genetic diversity, which we relate to their breeding habits and growth rate. A neighbor-joining 22 tree of CR haplotypes reveals no specific phylogenetic structure corresponding to location of fish 23 capture. Both neutral test and nucleotide mismatch distribution analyses suggest that yellowfin 24 seabream have experienced population expansion events. Pleistocene glacial periods and recent 25 stock releases have played important roles in the formation of present-day phylogeographical 26 patterns. Our study provides baseline information which will assist future research on genetic 27 structure, genetic diversity, and historical demography of yellowfin seabream after stock release in 28 southeast China coastal waters. The use of exotic seeds should be avoided in stock breeding and 29 release, and relevant follow-up surveys and genetic monitoring should be undertaken to clarify the 30 genetic impact of exotic seed use on wild populations. 31

, and also provides reference opinions for the rational utilization and sustainable development 40 of species germplasm resources (Pimm et al. 2014;Ward 2000). The most common method of 41 studying genetic diversity, genetic structure, and historical population demography involves the use 42 of molecular markers (Gao et al. 2007). Because mitochondrial DNA (mtDNA) is only inherited 43 maternally, reorganization of genetic material does not occur during this process, and the speed of 44 its evolution exceeds that of nuclear genes (Xu et al. 2014); mtDNA is often used to appraise both 45 species diversity and in phylogenetic research (Li et al. 2018(Li et al. , 2019. The mtDNA control region 46 (CR) is a non-coding sequence region which is subject to less natural selection pressure and has a 47 faster mutation and evolution rate than other mtDNA coding sequences (Xiao et al. 2000); it has 48 been widely used in marine fish genetic diversity research (Zhou et

Data analysis 94
The original yellowfin seabream mtDNA CR sequence was edited and corrected manually using 95 SeqMan in the DNASTAR software package, combined with PCR amplification of the same primers 96 DL-S and DL-R. Genetic diversity indices (number of haplotypes, mutation sites, transitions and 97 indels, haplotype diversity (h), nucleotide diversity (π) (Nei 1987), and mean number of pairwise 98 differences (k) (Tajima 1983)) were calculated using ARLEQUIN version 3.0 (Excoffier et al. 2005). 99 Blackhead seabream Acanthopagrus schlegelii schlegelii (Bleeker, 1854) was used as an outgroup 100 in the construction of a neighbor-joining tree (NJ) (Saitou and Nei 1987) in MEGA 5.0, using yellowfin seabream haplotypes and the Kimura two-parameter (K2P) model. 1000 nonparametric 102 self-expanding analyses were used for repeated tests, and the confidence of each branch of the 103 phylogenetic tree was calculated (Tamura et al. 2011). An unrooted minimum spanning tree (MST) 104 was constructed using the MINSPNET algorithm in ARLEQUIN version 3.0 (Excoffier et al. 2005) 105 to determine relationships among haplotypes; the MST topological structure was drawn manually.

107
Neutrality tests and mismatch distribution analyses were used to identify historical demographic 108 patterns (Fu 1997;Tajima 1989); the D test of Tajima (1989) and Fs test of Fu (1997) were used to 109 test for neutrality. Both mismatch analyses and neutrality tests were performed in ARLEQUIN 3.0 110 (Excoffier et al. 1992). Pairwise genetic divergences between sample populations were tested using 111 the conventional population index FS (Excoffier et al. 1992 were transformed to estimates of 'real time since expansion' using the formula: τ = 2 μkt (Rogers 118 and Harpending 1992), where t is an expected date when changes occurred, μ is the substitution rate 119 of CR, and k is fragment length. 120

121
Genetic diversity and Hap_29) with BB; no haplotype was shared by all six populations (Table 2).

138
The six combined populations had high haplotype (0.9959 ± 0.0018) and low nucleotide (0.0207 139 ± 0.0105) diversity. Among sample populations, wild individuals from DS I had the highest 140 diversity (0.9957 ± 0.0153), and wild individuals from YJ had the lowest (0.9778 ± 0.0540) ( Table  141 1). 142 143 r 144 The genetic distance between BB and DS sample populations was the largest (0.026), while that 148 between YJ and DS sample populations was the lowest (0.013) (Table 3). Based on the 100 mtDNA 149 CR haplotypes from the 6 sample populations, an NJ tree depicting 2 large lineages (with low 150 confidence) is apparent; lineage 1 contains 73 haplotypes (92 individuals) and lineage 2 contains 27 151 haplotypes (31 individuals). There is no apparent pedigree structure corresponding to capture 152 location (Fig. 2). The genetic distance between the two lineages is 0.032; of the 73 haplotypes in 153 lineage 1, all 20 from the XM sample population are included, as are 16 from DS I, 12 from DS II, 154 8 from YJ, 10 from FC, and 17 from BB sample populations. Lineage 2 comprised remainding 155 haplotypes, with no specific internal topological structure. 156 157 An unrooted MST was constructed based on the 100 mtDNA CR haplotypes (Fig. 3). All sequences 158 exhibited multiple primary haplotypes, with other haplotypes radially distributed around them, 159 without obvious phylogenetic structure corresponding to different sample populations.

161
Pairwise FS values estimated from mtDNA CR sequences ranged −0.024-0.456 (Table 4)   Raggedness index tests were not significant (P > 0.05). As such there is no significant deviation 210 from the expected distribution under the lineage expansion model, which we can then use to analyze 211 historical lineage dynamics ( Table 6). The unpaired nucleotide distribution of haplotype lineage 1 212 is bimodal (Fig. 4), indicating that branches within the lineage exist, with one peak corresponding 213 to the difference between all individuals, and the other to the small branch difference between them, 214 as is the case for lineage 2. Neutral tests on each haplotype lineage (Table 6) produced significantly 215 population expansion events. The D test on both lineages is also negative, and both are significant 217 (P < 0.05).

219
Observed τ values provide an estimated time for population expansion, with values for haplotype 220 lineage 1 being 6.227 (95% CI: 2.582-20.242) and for lineage 2, 3.748 (95% CI: 2.598-4.826). 221 According to the divergence rate and τ value of 5%-10%/MY, the timing of expansion of lineage 1 222 occurred between 95,500 and 47,800 YBP (years before present), and that for lineage 2 between 223 57,500 and 28,700 YBP. Both expansions occurred in the late Pleistocene.  and wild populations will increase (Bert 2007). In such instances the genetic diversity of the released 248 population will be lower than that of the wild population.

250
Because fitness is closely related to genetic diversity, the fitness of released fish will be strongly 251 influenced by the genetic diversity of parental stock (Bert 2007). Reduced genetic diversity can lead 252 to increased expression of harmful recessive genes and a decline in certain traits, leading to changes 253 in population fitness, reduced survival rates, fecundity, and growth, and lowered adaptive capacity 254 (Asahida et al. 2004). If fish with low fitness are released into natural waters, crossbreeding between 255 them and wild populations might greatly increase the frequency and expression of harmful recessive 256 genes in wild populations, and even lead to long-term degradation of a species. To avoid releasing 257 fish with low genetic diversity and poor fitness, parents of comparable or greater genetic diversity 258 than wild fish should be selected for breeding, and the number of parents involved in breeding 259 should be increased to as many as possible.

261
We report 100 haplotypes from 6 yellowfin seabream sample populations, with 89 of them exclusive 262 to a single sample, and with no dominant haplotype shared by all samples. The short-distances this 263 species migrates might explain the lack of shared and exclusive haplotypes and the high genetic 264 diversity in different geographical locations, which is comparable in wild and cultured populations. 265 The genetic diversity of wild fish from Xiamen in 2019 was slightly lower than it was in 2008 and  Note: h, haplotype diversity; π, nucleotide diversity; k, mean number of pairwise differences.