Genetic And Epigenetic Changes After Allo-Polyploidization And Effect Of Ploidy Increase On Transgene Expression In Citrus

The somatic hybridization technique enabling recombination of nuclear and cytoplasmic genetic information from both parents, is especially useful in circumventing the natural polyembryony and pollen/ovule sterility of perennial citrus crops. To date, numerous citrus interspecific and intergeneric somatic hybrids, aiming at combining beneficial agronomic traits from both parents, have been produced and evaluated for cultivar improvement. In addition to practical breeding, somatic hybrids also provide unique materials for studying the mechanism underlying the adaptation of two parental genomes within a common cytoplasm. This system has some advantages in elucidating the immediate changes that occur in newly synthesized allotetraploid over the important paradigm of synthetic allotetraploid arisen from sexual crossing and/or chromosome doubling. We focus on the potentially high dynamic TE component and random loci of the genome to unravel the early genetic and epigenetic changes associated with somatic hybridization. Then, to evaluate the effects of ploidy increase on the expression of GFP gene in citrus somatic hybrids at the plant level, EGFP expression in cells and protoplasts from the mature leaves of the diploid and allotetraploid plants was compared and quantitatively detected. Furthermore, the Mt-GFP version was for the first time transformed into citrus via the embryogenesis pathway for regeneration of transgenic plants from differernt ploidy callus. The performance of the three different subcellularly targeted GFPs was compared in living citrus leaf cells, and ploidy-dependent reduction of transgene expression was discussed. The main results are as follows:1. Characteristics of genetic and epigenetic changes in the three sets of synthetic allotetraploid somatic hybrids.1) For the three sets of the somatic hybrid and their diploid parents, an average of 2.68-4.89 polymorphic bands per primer set was yielded by AFLP marker. The genetic similarity coefficient between somatic hybrid and callus genotypes was higher than somatic hybrid and leaf genotypes. The genetic similarity between the set of plants was further investigated by constructing a dendrogram using AFLP data, MV and V were grouped together while M was placed in a single cluster. These results suggest the proportions of callus-genome-specific genetic and epigenetic alterations were significantly greater than that of leaf-genome-specific alterations.2) SSAP was performed using six newly found retrotransposon families. Overall polymorphism levels are variable depending on the retrotransposon population analyzed, but when the six data sets are combined, polymorphism levels are higher than those for AFLP, with an average of 4.11-6.16 polymorphic bands per primer set. Likewise, the similarity coefficient based on SSAP-Combined data ranged from 0.14 (callus and leaf genotypes) to 0.54 (somatic hybrid and leaf genotypes). MV and V were clustered, but M was placed far apart. The results based on AFLP or SSAP markers are very similar. These results suggest a generalized behavior of dominance for the leaf parent over the ones of all the other parental species combined by somatic hybridization and that the parental genomes differ in their ability to be modified in epigenetics in the polyploidy background. We also show that different retrotransposon populations reorganize differently, and more drastic changes were observed for the RLC1 population with more bands expected from callus-parent-origin. Southern blot analysis using a probe specifically detecting the conserved region of RLC1 family revealed the alteration in some bands in callus parents. 3) The analysis of MSAP was conducted to reveal the number of random loci presenting each possible methylation status in allotetraploid somatic hybrid and its diploid parents. The inner methylation of double-stranded DNA level was the highest, while the external cytosine methylation was the lowest. The total methylation levels of random loci in allotetraploid somatic hybrid and its diploid parents ranged from 29.57% to 48.70%. For MSAP markers, the methylation changes of random loci in callus parent were significantly higher than leaf parent.4) The methylation levels of the six combined transposable elements have also been observed in the same genotypes. Similarly, the inner methylation of double-stranded DNA level was the highest, while the external cytosine methylation was the lowest. The total methylation levels of regions flanking TE insertions ranged from 44.31% to 53.70%. For MSTD-combined markers, we also observed that the proportion of callus-genome-specific methylation alterations was significantly higher than leaf-genome-specific methylation alterations.2. Characteristics of genetic and epigenetic changes in the three sets of autotetraploid. 1) AFLP marker techniques were used to detect random genomic loci polymorphisms in autotetraploid. The number of AFLP polymorphism bands ranged from 3 to 17. The similarity coefficient based on AFLP data ranged from 0.96 to 0.99. These results suggest the rapid elimination of DNA sequences from autopolyploid genomes seems less genome restructuring than allopolyploids.2) SSAP was performed using six newly found retrotransposon families. The number of SSAP-Combined polymorphism bands ranged from 4 to 39. Likewise, the similarity coefficient based on SSAP-Combined data ranged from 0.95 to 0.99. We also observed that the rapid elimination of DNA sequences from autopolyploid genomes seems less genome restructuring than allopolyploids.3) The methylation levels of the random genomic loci have also been observed in the same genotypes. The total methylation levels of random loci in autotetraploid ranged from 30.13% to 34.92%. The methylation changes of random loci ranged from 2.18% to 7.42%. 4) The analysis of MSTD was subsequently conducted to reveal the number of CpG loci flanking TE insertions presenting each possible methylation status in allotetraploid. Similarly, the total methylation levels of regions flanking TE insertions ranged form 33.19% to 44.16%. Methylation changes of CpG sequences flanking the six transposable elements ranged from 3.28% to 10.19%. These results show DNA methylation changes resulting from chromosome doubling was significantly less than resulting from interspecific hybridisation.3. Characteristics of genetic and epigenetic changes in the two sets of cybrids.1) AFLP was performed to detect random genomic loci polymorphisms in cybrid. The polymorphic bands per primer set ranged from 0.36 to 1.23. The similarity coefficient based on AFLP data ranged from 0.92 to 0.97. These results disclose the cytoplasmic role on nuclear—cytoplasmic interactions in cybrids seems less genome restructuring than allopolyploids.2) SSAP marker techniques were used to detect TE-specific loci polymorphisms. The polymorphic bands per primer set ranged from 0.50 to 1.04. The similarity coefficient based on AFLP data ranged from 0.94 to 0.95. The lacking of mobilization of transposon family was revealed by SSAP.3) The methylation levels of the random genomic loci have also been observed in the same genotypes. The total methylation levels of random loci in autotetraploid ranged from 30.09% to 35.32%. The methylation changes of random loci ranged from 7.43% to 11.58%.4) The analysis of MSTD was subsequently conducted to reveal the number of CpG loci flanking TE insertions presenting each possible methylation status in allotetraploid. Similarly, the total methylation levels of regions flanking TE insertions ranged form 40.04% to 41.00%. Methylation changes of CpG sequences flanking the six transposable elements ranged from 7.75% to 8.24%. These results show DNA methylation changes resulting from cytoplasmic genomes was significantly less than resulting from interspecific hybridisation.4. Effect of ploidy increase on transgene expression.1) By using epifluorescent microscope and confocal laser scanning microscope, EGFP expression was compared in cells and protoplasts from mature leaves of these diploid cybrids and allotetraploid somatic hybrids. Results showed that the intensity of GFP fluorescence per cell or protoplast in diploid was generally brighter than in allotetraploid.2) To verify if the different GFP fluorescence is due to the level of GFP protein, Western blotting was employed to investigate the GFP protein level in both fusion parents and the regenerated plants. GFP protein level of the diploid cybrid was revealed significantly higher than that of the allotetraploid somatic hybrid.3) To confirm stable gene integration and the copy number, Southern blot analysis was performed on regenerated plants and parents. The plants showed one similar hybridization signal.4) To accurately determine whether the alteration of the GFP fluorescence lead to the underlying changes in GFP mRNA abundance, each line was subsequently analyzed by real time PCR. The results indicated that the GFP transcript level of diploid cybrid was significantly higher than allotetraploid somatic hybrid and slightly lower than that of leaf parent, either actin gene or eIF-1a gene was used as endogenous control.5. Transformation of the Mt-GFP gene to produce different ploidy transgenic Valencia orange lines.1) The plasmid pBI-mgfp4-coxIV encoding the Mt-GFP gene was successfully transferred into different ploidy embryogenic callus of Valencia sweet orange via Agrobacterium tumefaciens-mediated transformation. Putative transgenic plantlets expressing GFP were analyzed by PCR and Southern blot analysis to verify the presence of the Mt-GFP gene in their genome. PCR and Southern blot analysis revealed that 100% of the callus with green fluorescence resulted from positive transgenic events. This demonstrates that though some silenced transgenic callus may have been lost, it is still a good marker gene to identify the transgenic events at an early stage in Agrobacterium-mediated transformation and is a great assistance for citrus breeding and gene function analysis.2) To verify that the fluorescent particles observable in the guard cells are indeed mitochondria, we co-localize the Mt-GFP fusion protein with a mitochondrial-specific dye in citrus protoplasts.3) The performance of the three different subcellularly targeted GFPs was compared in living citrus leaf cells. The results demonstrate that the subcellular distribution of the three subcellular targeted GFP is very distinct in citrus leaf cells and the cell lines containing Mt-GFP gene can be further used in citrus basic cell fusion research.