Exploring variation in pea protein composition by natural selection and genetic transformation

Pea (Pisumsativum L.) seeds are a rich and valuable source of proteins, which can have potential for food industrial applications. Pea storage proteins are classified into two major classes: the salt-soluble globulins, and the water-soluble albumins. The globulins are subdivided into two major groups based on their sedimentation coefficient: the 11S fraction (comprising the class ofleguminwith variousisoforms) and the 7S fraction (comprising the classes ofvicilinandconvicilin, each with variousisoforms).Pea cultivars with extreme variation in globulin composition (i.e. lacking a particular class of proteins) might become important for the food industry, because they could provide new raw materials for specific applications, like the production of Novel Protein Foods (NPFs), which receive attention as possible meatreplacers.This thesis aimed at (i) to determine the existing natural variation in pea's globulin content and composition, in order to identify suitable cultivars for the production ofNPFs, (ii) to develop a more efficient protocol for genetic modification of pea, and (iii) to modulate pea protein composition, based ondsRNAdirected silencing.An inventory of protein content and composition of pea was performed to characterize the genetic variation for these traits (Chapter 2). To include a wide range of natural genetic variation, cultivars from a wide geographic distribution, with differences in leaf and seed characteristics, were selected and characterized. Large variation was observed between the various lines. Results on protein content showed a variation from 16.3% to 36.6% of dry matter (DM), with an overall average content of 26.6%. Globulins content varied between 49.2% and 81.8% of the proteins of the total pea protein extract (TPPE). On individual globulins level,legumincontent varied between 5.9% and 24.5%.Vicilinwas the most abundant protein of pea, and its content varied between 26.3% and 52.0% of the TPPE. The processedvicilinwas the predominant of the two, with values between 17.8% and 40.8%, whereas the non-processed ones constituted between 3.1% and 13.5% of the TPPE.Convicilinwas the least abundant globulin having an average content of 6.1%. Its content ranged from 3.9% to 8.3%. Finally, the globulin-related proteins were present in amounts ranging from 2.8% to 17.3% of the TPPE. The globulins showed the largest relative variation of the four globulin classes.It is known that a lowvicilin/convicilinratio can result in poorgelation. Based on our data (Chapter 2) and the literature, it is concluded that pea isolates have a morefavourableprotein composition for gelling applications as compared to those from soybean. Moreover, the genetic variation for this trait appears to be larger in pea than in soybean, which might offer opportunities to reduce theconvicilincontent further.Our inventory did not show cultivars lacking a specific globulin. Such cultivars might be important, because they could have morefavourablephysical properties for the production ofNPFs. To produce such lines genetic modification approaches were employed. To carry out genetic modification a reliable protocol is needed. At the time this study started, protocols for the production of genetically modified peas were available, but particularly the regeneration of plants from transgenic cells was very inefficient. Most of the plants obtained were either escapes orchimeric(not all cells of a plant are genetically modified). Therefore, our study focused on obtaining a novel regeneration protocol, which in combination with the transformation procedure would result in an improved method for obtaining transgenic pea lines.The novel regeneration protocol started with subculture of stem tissue with one node (Chapter 3) or whole seeds (Chapter 4) on TDZ supplemented medium. Repeated subculture of stem tissue with one node resulted in a greenhyperhydrictissue in the swollen bases of the multiple shoots, which is fully covered with small buds [bud-containing tissue (BCT)] was formed. BCT fragments were isolated andsubculturedin the same medium and, as a result, they were able to reproduce themselves in a cyclic fashion. Subculture of BCT on medium supplemented with a combination of GA 3 ,cytokininsandauxinsresulted in the production of rooted shoots. In-vitro plants were transferred to the greenhouse foracclimatisationand further development. All tested pea cultivars ('Espace', 'Classic', 'Solara', and 'Puget') responded in the same way.Culture of seeds for a relatively long time (Chapter 4) on TDZ supplemented medium resulted in the production of very high numbers of shoots together with BCT which was identical to the BCT described in Chapter 3. This protocol resulted in the faster production of BCT as compared to the protocol described in Chapter 3.The regeneration protocol from Chapter 3 was combined with genetic modification (Chapter 5). Transgenic pea plants were obtained after co-cultivation of bud containing tissue (BCT) andshootyBCT withAgrobacteriumtumefaciens strain AGL0(pG49A). The binary vector pG49A contained an interrupted inverted repeat of aleguminA gene, flanked by the promoter of thetrypsin/chymotrypsininhibitor gene, together with theluciferasegene for selection of transgenic tissue.Luciferasepositive tissue was identified, isolated, andsubculturedon TDZ-supplemented medium. On this medium, BCT can be multiplied, and theshootyBCT will become pure BCT again. Theluciferasebased selection procedure was repeated until (almost) completeluciferasepositive BCT cultures were obtained. Plants (S 0 ) of 23 transgenic lines were grown in the greenhouse. The S 0 plants were smaller in size and produced less seeds than the control plants. All lines producedluciferasepositive seeds. The transgenic nature of 5 S 0 plants was further confirmed using Southern blot analysis. Protein analysis with SDS PAGE electrophoresis of the seeds of 8 lines indicated differences in protein composition, although our data were not conclusive on whether the amount ofleguminA or otherleguminswas affected. Further experiments should show whether the protein compositional changes resulted from silencing of theleg Agene, or other factors such as genetic or epigenetic changes in the genetically modified plants, caused by the tissue culture procedures.Seeds of 6 lines were grown to produce S 1 . The S 1 plants were comparable in height to control plants. However, the number of seeds per plant was significantly lower.The developed transformation protocol is highly repeatable. Each experiment resulted in genetic modified plants, in contrast to other systems, which have low repeatability. However, our system is more time consuming than those developed by others. Therefore, the regeneration system, which produces BCT directly from the seeds without the need for production of in-vitro plants, should be combined with genetic modification in the future. Furthermore, the selection of transgenic tissue should be optimised using selectable marker genes such asnpt IIor pat .