External Characteristics
The body length, fork length, body weight (total and without internal organs), sex ratio, and age composition of two masu salmon groups are summarized in Table 2. The proportion of females is 11% in Mijiang salmon (hereafter MJ salmon) and 75% in farmed cherry salmon (hereafter JP salmon). In the two masu salmon groups, 2-year-old fishes are the main components, accounting for 40% and 50% for MJ and JP salmon, respectively. The farmed salmon had a body length of 16.46 ± 0.52 cm, fork length of 15.19 ± 0.52 cm, head length of 3.56 ± 0.12 cm, total body weight of 49.51 ± 7.25 g and weight of 41.47 ± 5.68 g excluding internal organs. The C.D values of the 11 proportional traits are shown in Table 3. Only in Bd / Sl and CPl / Sl, the C.D values were 1.658 and 2.902, were greater than 1.28. The C.D values between the other proportional traits were all less than 1.28.
COI variation in masu salmon
COI gene consists of 23.5% A, 31.8% T, 17.9% G, and 26.7% C in MJ salmon. The average content of A+T is higher than G+C. The content of G is the lowest, showing an obvious anti-G bias, which is consistent with the characteristics of vertebrate mitochondrial DNA. The DNA sequencing of COI detected 15 variable sites, including 13 synonymous and 2 non-synonymous substitutions (Table 4). MJ salmon detected five variable sites (4 synonymous and 1 non-synonymous substitution), which differ in the 170 and 299 nucleotide positions. The non-synonymous substitution in the 562 nucleotide position was performed in all salmon subspecies, while the mutation in the 325 nucleotide position only existed in biwa salmon. In addition, the unique variable site of satsukimasu salmon and Formosa landlocked salmon are in 185 and 599 positions, respectively. In the population of MJ salmon, two special variable sites were observed in 674 and 299 nucleotide positions.
Twelve haplotypes were defined as “a” to “l” from the 64 individual sequences, including five haplotypes from MJ salmon, three haplotypes from cherry salmon, two haplotypes each from satsukimasu salmon and biwa salmon, and one haplotype from formosa landlocked salmon. Especially, MJ and cherry salmon had an identical haplotype “b”, and 11 other haplotypes were obtained from a single population. The distribution of most haplotypes has obvious geographic characteristics. Haplotypes “a” and “c” were only observed in cherry salmon from Japan, while haplotypes “f” and “g” were observed in biwa salmon from Biwa Lake. Haplotype “e” is the sole haplotype in formosa landlocked salmon, and haplotypes “i” to “l” from MJ salmon is unique. In a network (Fig. 1), haplotype “b” located in the central. Only the haplotypes of biwa salmon separate from “b” clearly, while those of the remaining masu salmon subspecies intermingle with each other.
Haplotype diversity (Hd) and nucleotide diversity (π) are important indicators for evaluating population genetic diversity; the higher the value, the richer the population genetic diversity. In MJ salmon, Hd was remarkably high with a value of 0.817±0.025. We used the JP salmon for comparison, yielding a Hd value if 0.268±0.113. The π value of MJ salmon was (0.00243± 0.00022) was at a low level (Table 5).
Genetic distances among the 12 haplotypes were obtained with their standard deviations (Table 6) by using the Kimura’s two parameter method 28. The genetic distances of haplotype “b” between MJ and JP salmon was 0.0000. The inter-group genetic distance ranged from 0.0016 to 0.0095 between MJ and cherry salmon and from 0.0000 to 0.0063 between MJ and JP salmon. In addition, we found that the intra-subspecies variations (from 0.0016 to 0.0063) in MJ salmon was nearly equal to the inter-group genetic distance. The genetic distances of only biwa salmon were significantly isolated, with values 0.0063–0.0128 in MJ salmon, 0.0048–0.0127 in cherry salmon, and 0.0063–0.0095 in JP salmon.
The base saturation analysis of COI sequence is shown in Figure 2. A linear relationship was observed between Ts, Tv, and F84 distance, indicating that the base transformation and transversion of COI sequence did not reach saturation during genetic evolution, and this information could be used for phylogenetic analysis. To draw a neighbor-joining tree from COI gene, we included the sequence for O. mykiss as an outgroup for the O. masou complex (Fig. 3). Biwa salmon, satsukimasu salmon, and formosa landlocked salmon were separated, while MJ, JP, and cherry salmon were shown in the same branch after 1,000 bootstrap replicates.
Microsatellite variation in masu salmon
All the 14 loci were highly polymorphic in the Mijiang population examined without linkage disequilibrium (LD) between loci. Table 7 summarizes the number of alleles, allelic richness (Ar), and the observed (ho) and expected heterozygosity (he) per population of 14 microsatellite loci. Overall, 186 alleles were found in the population. The mean allelic richness among all loci within populations was 9.38 in the MJ salmon populations. The number of alleles in MJ salmon was 11.0714. The observed heterozygosity was 0.6061, while the expected heterozygosity was 0.7863. Overall, the mean expected heterozygosity was slightly higher than the observed heterozygosity. Nine loci in population MJ showed a significant deviation from Hardy–Weinberg equilibrium (HWE), as determined using GENEPOP. In addition, tests for LD suggested no significant linkages in the Mijiang population. The mode shift indicator showed a normal L-shaped distribution, thus explaining the genetic bottlenecks in Mijiang population.
AMOVA with both markers (Table 8) revealed the population structure in masu salmon that the variation within populations (COI, 39.0%, P < 0.001; microsatellites, 61.0%, P < 0.001) was significantly higher than that among populations (COI, 2.5%, p=0.000; microsatellites, 97.5%, P < 0.001), indicating the congruence of both mitochondrial and microsatellite markers in estimating the genetic structure of masu salmon.
The FIS, FST, and Nm values in masu salmon based on microsatellites are summarized in Table 9. The FST value is 0.0540 in the intra-group of MJ salmon and −0.0232 in the intra-group of JP salmon. The value of FIS was similar in two populations (0.1995 and 0.2294). Otherwise, the pairwise FST were estimated as 0.375 (P < 0.01) and 0.024 (P < 0.01) between MJ and JP salmon for COI and microsatellite markers in Table 10, respectively (although COI was larger than for microsatellites). Nm values are responsible for gene flow with a value of 10.595 in MJ salmon. Nm and FST values are inversely proportional.