Primary and cyclic somatic embryogenesis in cassava (Manihot esculenta Crantz)

Cassava is one of the major food crops in the tropics. Several of the major problems in cassava can probably only be solved by breeding with cellular and molecular techniques, e.g., the introduction of specific genes (virus resistance, protein content, quality aspects and so on). These genes can be directly applied in existing varieties of vegetatively propagated crops like cassava. Genetic modification requires efficient, genotype-independent regeneration methods. Plant regeneration can be accomplished by two different pathways: organogenesis and somatic embryogenesis. In both organogenesis and embryogenesis, the regenerated structures either originate directly from the explant or indirectly from callus induced from the explant. In most species transformed plants are obtained by indirect regeneration, either by organogenesis or somatic embryogenesis. The callus phase is used to select and multiply transformed cells. Because organogenesis for cassava appeared to be not repeatable, somatic embryogenesis was further investigated. Somatic embryogenesis is defined as the process in which a bipolar embryo is formed which has no vascular connection with the parental tissue. It has been described in more than 200 species [Chapter 11. It was shown by others that in cassava (primary) embryos originated directly from young leaves or zygotic embryos. Direct embryogenesis has been used successfully in a some species for plant transformation. In these species primary somatic embryos themselves were an excellent source to start a new cycle of (secondary) embryogenesis. Repeated subculture of somatic embryos allowed the development of continuously proliferating embryogenic cultures (cyclic embryogenesis). The phase of embryo proliferation was used to select and multiply transformed cells. An overview of culture regimes which allows continuous proliferation of somatic embryos is given in Chapter 2.In initial experiments first cycle or primary embryos were formed from young cassava leaf explants derived from greenhouse grown plants. After 10 days of culture nodular or globular embryos were visible. Globular embryos developed into torpedo shaped embryos which germinated after tranfer to the a medium without auxins. Germinated embryos (GE) are defined as structures with a distinct hypocotyl and large green cotyledons. Five of the six tested South American and Indonesian clones formed germinated embryos. The number of germinated embryos produced, was strongly influenced by the genotype and by hardly controlled growing conditions of the donor plants in the greenhouse. The production of the Colombian clone M.Col22 varied between 0 and 22.1 GE per initial leaf explant (GE/IE). The other clones were considerably lower in their response [Chapter 3]. Therefore, M.Col22 was chosen as a model plant.To create uniform growing conditions in vitro grown donor plants were used as source for leaf explants. Using the same culture conditions as applied for greenhouse derived leaf explants, this approach gave less variation in germinated embryos but also a much lower production (< 1 GE/IE). Doubling of the 2,4-D concentration in the embryo induction medium increased the production to a maximum of 3.5 GE/IE. The embryogenic capacity of M.Col 22 could be further increased to 6.6 GE/IE by growing donor plants at reduced irradiance. The highest production (9.9 GE/IE) was obtained by a pretreatment of donor plants with 2,4-D, a few days before the isolation of leaf explants.Another advantage of the 2,4-D pretreatment of donor plants was studied in Nigerian clones. Only 5 out of 11 invitro grown clones formed globular embryos and only in 2 some of the globular embryos developed into germinated embryos. After 2,4-D pretreatment of the donor plants, 10 out of the 11 clones formed globular embryos and in 8 of them germinated embryos were formed [Chapter 6].Only torpedo shaped and germinated embryos initiate a new cycle of embryogenesis after reculture on induction medium. Germinated embryos were the best starting material to initiate cyclic cultures [Chapter 4]. Independent of the genotype, germinated embryos formed new germinated embryos at a high rate and the embryogenicity seemed not to be changed after one year of culture [Chapters 4 and 6]. The production of cyclic germinated embryos for M.Col22 varied between 6.8 and 9.9 GE/IE. The production of germinated embryos in liquid medium was significantly higher than on solid medium. Also fragmentation of the initial germinated embryos, before starting a new cycle of embryogenesis, enhanced the production. With both improvements, the production of M.Co122 increased to about 30 GE/IE [Chapter 5].Culture of torpedo shaped and germinated embryos on BA supplemented medium allowed their development into shoots.-As for the induction of new embryos, germinated embryos were also the best material to be cultured for shoot development [Chapter 4]. The frequency of shoot development appeared to be genotype dependent [Chapters 3 and 6]. In the clone M.Co122 more than 50 percent and in the clone Tjurug only 10 percent of the germinated embryos developed into shoots. All shoots, independent of the genotype, formed roots on growth regulator- free medium [Chapters 3, 4 and 5].Cyclic embryos originated directly from the cotyledons of the somatic embryo by a budding process. The origin appeared to be multicellular. The first embryogenic divisions started with cells in or near the vascular strands. These initial divisions led either directly to a somatic embryo or to meristematic tissue, of which later embryos were formed [Chapter 7].Almost 500 regenerants of up to the seventeenth cycle embryos were evaluated in vitro for somaclonal variation. Only one regenerant had a visible deviation from control plants (variegated leaves) which was assumed to be of genetic origin [Chapter 8]. About 110 regenerants were transferred to the greenhouse and evaluated for more than 1 year. Plants of the regenerants showed fewer virus-like symptoms than control plants. The root tubers of control plants were more uniform than that of regenerated plants. Some plants of the regenerants had irregularly shaped roots which were not observed in control plants. Not all plants of a particular regenerant had a abnormal root tuber phenotype and this is a clear indication that the cause of this variation is most probably epigenetic (physiological) and, therefore, is expected to disappear with prolonged multiplications. Cassava has proven to be amenable for Agrobacterium -mediated transformation and the transformed cells are able to divide. Unfortunately, the majority of them developed into callus cells and only a few into embryogenic competent cells. Culture procedures which increase the recovery of embryogenic competent cells from transformed cells together with an efficient non-destructive selection procedure should allow the development of an efficient, genotype-independent transformation procedure. This is of importance for breeding of this vegetatively propagated crop, cassava.