Genetic diversity in oysters

• Main Author: English, Louise,(Louise Jasmins)
• Format: Thesis
• Language: English
• Published: 2001
• Subjects: Oysters; Crassostrea; Sydney rock oyster; Dun; Pacific; Crassostrea gigas
• Online Access: https://eprints.utas.edu.au/19736/; https://eprints.utas.edu.au/19736/1/whole_EnglishLouiseJasmins2001_thesis.pdf

This project examined the effects on genetic diversity of oysters by hatchery techniques and selective breeding. The edible oyster industry in Australia comprises of two main species: Crassostrea gigas and Saccostrea glomerata, which are produced by hatcheries or natural spatfall, respectively. This study examined the levels of genetic diversity as a result of unintentional selection in C. gigas, and as a result of intentional selection (a selective breeding program to increase meat weight) in S. glomerata. This study has implications for aquaculture species worldwide as it examines the levels of genetic diversity in introduced (eg C. gigas) and a previously declining, native (S. glomerata) species, which comprise two of the main groups of species utilised for aquaculture. The oyster aquaculture industry in Tasmania is based on Pacific oysters, C. gigas, and derived from imports of Japanese oysters made some 40 years ago. Fears were held that introduction and subsequent domestication had eroded the genetic diversity. As the industry wished to embark on selective breeding programs, a genetic audit of the hatchery stocks was required. Four Australian established (ie non-hatchery produced) populations, and two of Japanese populations that originally imported, were analysed to determine the amount of genetic diversity present. Three different genetic techniques were employed - allozyme electrophoresis, microsatellite DNA and mitochondrial DNA RFLP (restriction fragment length polymorphism) analyses. Using 17 allozyme loci, three hatchery and four naturalised populations of Crassostrea gigas (Thunberg) in Australia were compared with one another and with two endemic Japanese populations. All populations showed a high degree of genetic variability. The percent of polymorphic loci ranged from an average of 70.6% (hatcheries) through 73.5% (naturalised and Japan). Mean observed heterozygosities ranged from 0.267 (naturalised) through 0.285 (hatcheries) to 0.291 (Japan). Mean numbers of alleles per locus ranged from 3.0 (hatcheries) through 3.3 (naturalised) to 3.5 (Japan). Most loci and populations showed good fits to Hardy-Weinberg expectations; the few significant exceptions were heterozygote deficiencies. Tests of allele frequency differentiation among the nine populations revealed that 11 of the 16 variable loci showed significant (a reduced to 0.05/16 = 0.0031) inter-population heterogeneity after both x 2 . and GsT analysis (Table 2.5). Four loci — GDA, 6PGDH, PEPS-I and PROT-2 — were non-significant for both analyses, and EST-D was significant for x2 analysis (P < 0.001), but not after Bonferroni correction for GsT analysis (P = 0.004). Five populations (BEA, BRI, DUN, SMI, SEN) conformed to Hardy-Weinberg equilibrium for all loci. A few populations and loci did not conform (after Bonferroni correction, using a = 0.0031): HIR (AK, DIA, PGI), NSW (PEPS-2), PIT (DIA), and SWA (DIA) (data not shown). All the non-conforming samples showed heterozygote deficiencies, which were significant in two cases: HR (DIA: x2 = 35.67, P < 0.001; D — 0.314, P < 0.001); and SWA (DIA: X2 79.04, P = 0.001, D = - 0.397, P < 0.001). Allele-frequency differences among populations were minor, although sometimes statistically significant: only about 1% of the allele frequency variation could be attributed to among-population differences. The levels of homozygous excess observed in this study were lower than that previously reported for this species, but may be due to the same reasons such as gel scoring errors, null alleles, selection, inbreeding, population admixture or sampling error. Four microsatellite loci (consisting of two dinucleotide, one tetranucleotide and one tetranucleotide/dinucleotide motifs) were used to analyse the populations. Primers were designed on clone sequence containing at least five repeats of a microsatellite motif. Ten sets of primers were trialled but only four sets had variability levels useful for population genetics analyses, as the others had between 23-34 alleles or were monomorphic. High levels of genetic variability were observed (mean polymorphism = 0.889, mean heterozygosity = 0.188). Tests of allele frequency differentiation among the nine populations revealed that three of the four variable loci showed significant (a reduced to 0.05/4 = 0.0125) inter-population heterogeneity after x2 analysis and Gsr analysis — cmrCgl 7, BV59 and cmrCg61. The amount of differentiation among the populations was, however, small. Across all loci, only 4% of the genetic variation could be attributed to differences among populations. No population conformed to Hardy-Weinberg equilibrium for all loci. Where populations did not conform to Hardy-Weinberg equilibrium, a significant excess of homozygotes was observed. Although null alleles have been previously reported for the loci used in this study null alleles do not appear to explain the heterozygote deficiencies observed in this study, based on analysis by the NullTest program (W. Amos, pers. comm.) as the frequency of the proposed null alleles seems too high. Unbiased (Nei, 1978) genetic distances over the four loci were estimated between all pair-wise combinations of populations. All pair-wise population distances are very small (Nei D<0.03). However, the standard errors of the distances are large (ranging to 0.0295±0.0867 between the SWA/BEA/BRI/HIR/PIT/NSW/SMI/SEN and DUN clusters), indicating that the groupings, based on only four loci, are not statistically meaningful. There were no significant changes of genetic diversity between the populations. Overall, the use of microsatellites confirmed the results of the allozyme study of these populations. - the introduction of oysters from Japan to Tasmania, and their subsequent domestication, have not substantially eroded the genetic basis of the Tasmanian stock. Very little variation was found using mitochondrial DNA (mtDNA) RFLP analysis using 12 restriction enzymes in one hatchery and one endemic population. The mtDNA fragment used was found to contain two conserved genes explaining the lack of variability observed. The technical limitations and lack of knowledge of the oyster mtDNA genome made this approach inappropriate for population genetics analysis. However, the identification of the proximity of the 16srRNA, and COIII genes observed in this study gives further insight into the mtDNA gene order of C. gigas. Together with previous findings of the mtDNA RE site map (Oohara and Mori, 1989), the location of the cytochrome b gene . (Li and Hedgecock, 1998), this study demonstrates a block of COIII, 16srRNA and cytochrome b mtDNA genes —a combination unique to bivalves. Two generations of a selected breeding line for increased whole weight (using mass selection) and unselected group of the Sydney rock oyster Saccostrea glomerata were examined using allozyme electrophoresis (a total of 14 loci analysed). Genetic variability levels were determined for each group - all were high and similar to one another (mean percentage polymorphic, 66.7, mean observed heterozygosity, 0.222; mean number of alleles, 2.5). Genotype frequencies at all loci in all groups conformed to Hardy-Weinberg equilibrium, except for the ESTD-2 locus in the second generation (P<0.001), which had a large and significant excess of homozygotes (Selander index D = -0.451, P<0.001). Thus significant allele frequency differences were observed at seven loci between control and second generation, at eight loci between second and third generation, and at only one locus for control and third generation. This suggests that the second generation sample is responsible for most of the heterogeneity observed. Despite the difference between actual and estimated broodstock numbers, the expected numbers of alleles of the second and third generations of the selected breeding line were very close to the observed numbers in all cases, suggesting that random genetic drift (sampling variation) alone was the cause of allelic variation between the groups. The results of this allozyme survey indicate that the selective breeding for increased whole weight has not substantially eroded levels of genetic variation. There are high levels of genetic variation present in the control group and in the two generations of the selected breeding line. The limited, but statistically significant, heterogeneity between the second generation and other samples appears to be due to a sampling artifact; it is likely to be biologically unimportant. The type of selective breeding program used (mass selection) aided the retention of genetic diversity across the generations sampled. By using oysters from 4 different areas and using large numbers of oysters for spawning helped to achieve the levels of genetic diversity observed. In summary, this study has not only substantially broadened the base of knowledge in the two oysters species investigated, but also shown that intentional (in the form of selective breeding) and unintentional (in the form of introduction and subsequent domestication) selection of aquaculture species need not result in major allele loss. Without major allele loss an aquaculture species then has the potential for successful selective breeding for desired traits; as the genetic diversity and hence likelihood of finding particular gene(s) of interest have not been diminished significantly, and has every chance of success. That the two commercially important edible oyster species (introduced and native) in this study have not shown a large magnitude of genetic diversity loss despite either introduction, naturalisation and domestication, or selective breeding serves as a encouraging example to the oyster industries and researchers involved as well as those in other countries and indeed other aquaculture species.