Transformation of mature plant material

Improvement of woody species through genetic engineering will have limited appli- cations unless tissue from mature plants of elite varieties can be readily transformed. Transformation and regeneration of these plants are usually restricted to juvenile tis- sues (Cervera et al., 1998b), which exhibit greater regenerative potential and greater competence for transformation and suscep- tibility to Agrobacterium than mature tis- sues. Regenerated plants from juvenile tissues will have juvenile characters, and several years will be needed before horti- cultural and commercial traits of the trans- genic plants can be evaluated. Development of transformation procedures that could by- pass the juvenile stage would greatly reduce the time involved in improving woody trees by genetic engineering.

Citrus species show juvenile phases ranging between 5 and 20 years in subtrop- ical areas, which are marked by vigorous and upright growth habit, a characteristic

leaf shape, development of long thorns in the leaf axils and lack of fl ower initiation. We have chosen Pineapple sweet orange as a model to develop a transformation proce- dure for mature tissue of a selected cultivar (Fig. 15.1).

Previous experiments in our laboratory had confirmed the limited regenerative potential of explants from aged mature citrus plants. Thus, to increase the regener- ation ability of mature explants, buds from adult trees were grafted on vigorous seedlings. Regeneration from stem seg- ments from the first, second and third

fl ushes of newly grafted invigorated mature sweet orange plants was evaluated in com- parison with regeneration from stem seg- ments from juvenile plants. The results indicated that explants from the fi rst and second fl ushes produced similar regenera- tion frequencies, signifi cantly higher than that of the explants from the third fl ush. The fi rst fl ush of the adult plants was selected as the source of tissue for genetic transformation experiments.

Buds collected from Pineapple sweet orange maintained in a screenhouse of the pathogen-free Germplasm Bank Collection

Fig. 15. 1. Procedure for the genetic transformation of adult sweet orange plants. (A) After co-cultivation of internodal stem segments with Agrobacterium tumefaciens, the explants are transferred to a regeneration–selection medium which promotes the regeneration of transgenic shoots. (B) Apical portions of shoots are grafted in vitro on decapitated citrange seedlings. (C) Several weeks after grafting, developed scions are grafted again on vigorous rootstocks in the greenhouse. (D) After one year, transgenic plants started to fl ower and set fruit, confi rming their mature nature. They showed a normal phenotype and development. (E) Release of genetically modifi ed citrus plants under controlled fi eld conditions at the IVIA.

of the IVIA are grafted on seedlings of C. volkameriana Ten & Pasq. in the green- house (18–27˚ C). Then, newly elongated mature shoots are allowed to develop the fi rst fl ush. Preparation of explants, transfor- mation and regeneration conditions are the same as described above for transformation of juvenile material. Internodal stem seg- ments (1 cm long) in a semi-hardened stage are cut transversely from the stem pieces, inoculated and co-cultivated with A. tume- faciens, and transferred to SRM, as described above. Shoot regeneration is observed after 2–5 months in selective medium (Fig. 1A). Whole transgenic plants are obtained from these shoots by in vitro grafting (Fig. 1B) and a posterior grafting in the greenhouse (Fig. 1C).

The putative mature transgenic sweet orange plants show the morphology and growth habits of an adult plant, as com- pared with control mature plants. In fact, whereas juvenile plants show a pro- nounced thorniness, transgenic mature plants are almost thornless, similar to the mature plants from which the explants are taken for transformation. After 14–18 months in the greenhouse, the trans- genic and control plants usually start to fl ower, confi rming their mature nature (Fig. 1D).

These results confi rm the maintenance of the ontogenic mature stage of the invig- orated mature plants as well as the trans- genic plants. Interestingly, transgenic events keep their epigenetic mature state even after a process of dedifferentiation, callus induction and redifferentiation, nec- essary to shift the cells to a competent stage for transformation. Therefore, we are able to transform and regenerate mature tissues of citrus directly, by-passing the juvenile stage. This process greatly shortens the period of time until flowering and bearing fruit, and decreases the time to achieve horticulturally acceptable charac- teristics by years. It is the fi rst time that a procedure for genetic transformation and regeneration of mature tissues of woody fruit plants has been developed (Cervera et al., 1998b).

Improvement in the effi ciency of generating citrus transgenic plants

Since some citrus genotypes are particu- larly recalcitrant to transformation, we have investigated the use of proper marker genes and Agrobacterium vectors with the aim of increasing the efficiency of generating transgenic citrus plants.

The selection of genetic transformants requires the use of marker genes that func- tion as reporters of gene expression and so permit the recovery of transgenic plants. Since the fi rst demonstration of the gfp gene as a vital marker gene in both bacteria and Caenorhabditis elegans (Chalfie et al., 1994), it has attracted increasing interest and is considered to have several advan- tages over other visual marker genes. The fl uorescence emission of GFP only requires the excitation of living cells by UV or blue light.

Sour orange is a species among citrus genotypes particularly recalcitrant to trans- formation. The production of transgenic plants is rare and circumstantial. In many recalcitrant plants, cells competent for transformation are restricted to specifi c tis- sues. To develop transformation procedures in these cases, it is important to localize the sites of transgene expression in order to favour the regeneration of whole plants from such competent cells. When uidA is used as a reporter gene, the diffusion of the GUS indigo dye reaction, the possibility of the presence of bacteria expressing the uidA gene, the possible contact between different transformation events and the destructive character of GUS assays all con- tribute to the imprecise localization of cells or tissues competent for transformation. GFP provides the possibility to perform in vivo monitoring of Agrobacterium -inocu- lated plant tissues and thus allowed local- ization of competent cells for transformation in dedifferentiated callus from the cambium tissue of citrus explants (Ghorbel et al., 1999). Therefore, treatments favouring the development of such callus tissue are necessary to increase transforma- tion frequencies and to enhance a more effi -

cient recovery of citrus transgenic plants (Ghorbel et al., 2000).

It has been important in many plants to use reporter genes, mainly uidA, to test the infl uence of factors affecting both transfor- mation and the regeneration of transfor- mants and escapes. However, since assays to test reporter gene activity are destructive, such factors cannot be evaluated continu- ously and simultaneously. Moreover, the destructive character of GUS assays pre- cludes the recovery of the GUS + putative transgenic regenerated shoots. To select transgenic shoots from systems in which escapes and chimeras regenerate at high frequencies, such as in citrus, all the regen- erated shoots have to be analysed, which is a very laborious process. We have shown that GFP expression permits a rapid and easy discrimination of transgenic and escape shoots in citrus (Fig. 15.2C). Competition for growth between trans- formed and non-transformed shoots could be avoided by eliminating the escapes soon after their origin (Ghorbel et al., 1999).

Selectable markers, such as NPTII, have been widely used to produce trans-

genic plants. However, the presence of these markers may restrict the release of transgenic products due to their adverse infl uence on consumer perceptions. In fact, the presence and expression of selectable markers coding for antibiotic resistance has been a topic of much discussion for the commercialization of certain transgenic crops in Europe. Recently, a European Union Directive forbid the commercializa- tion of transgenic foods with antibiotic transgenes, effective from 2008. We have used GFP for the early detection and rescue of transgenic buds, circumventing the use of antibiotic marker genes to produce trans- genic plants. Green fl uorescent shoots can be regenerated from explants inoculated with Agrobacterium but cultured in a medium without kanamycin (Ghorbel et al., 1999) (Fig. 15.2C). This opens up the possi- bility of producing transgenic plants with- out using selective agents.

Moreover, we have investigated the use of appropriate Agrobacterium strains to transform economically important citrus genotypes effi ciently. Citrus is not naturally infected by A. tumefaciens. In fact, crown

Fig. 15.2. Use of uidA and gfp as reporter genes in the recovery of transgenic citrus plants. (A) GUS expression of a transgenic sweet orange fruit (left) and a non-transgenic control fruit (right). (B) GUS expression of cultured embryos recovered from seeds of an adult transgenic plant. (C) GFP-positive shoot grafted on a decapitated citrange seedling illuminated with blue light. (D) GFP expression of juice vesicles from a transgenic (left) and a non-transgenic (right) lime fruit.

gall is not a disease affecting citrus. However, we have demonstrated that A. tumefaciens strain A281 is not only able to transform citrus effi ciently but that it is even super-virulent in many citrus geno- types (Cervera et al., 1998c; R. Ghorbel et al., unpublished results). In addition, we have found a correlation between super-vir- ulence and the super-transformation ability of its disarmed derivative EHA105. The super-virulence of A281 in many plants has been attributed to higher expression of vir genes after induction. In fact, we have demonstrated that the vir region of its Ti plasmid pTiBo542 provides this ability in citrus (Ghorbel et al., 2001a).

Furthermore, introduction of extra copies of virG from pTiBo542 into EHA105 increases stable transformation frequency in citrus. When additional copies of virG from pTiBo542 are introduced into C58 (pMP90), stable transformation frequency is drastically increased, approximately 80- fold in lime, suggesting that not only the quantity but also the quality of virG from pTiBo542 is a determinant in the super- transformation ability of Agrobacterium in citrus. In lemon, introduction of additional copies of virG from pTiBo542 in EHA 105 permits the transformation frequency to be doubled. Thus, additional copies of virG from pTiBo542 can be used as a tool to increase transformation frequency in those citrus genotypes more recalcitrant to Agrobacterium -mediated transformation (Ghorbel et al., 2001a).

Greenhouse and fi eld performance of transgenic plants

In vegetatively propagated and long-lived perennial fruit crops, an important requi- site for evaluating the validity of genetic transformation technology in improvement programmes is the stability of the modifi ed genome and transgene integration and expression over long periods of time. We have kept in a screenhouse for a period of 4–5 years a collection of 70 transgenic cit- range plants in order to investigate: (i) the

origin of morphological variants in the transgenic population; (ii) how factors related to T-DNA integration, the regenera- tion process and the expression of trans- genes may be infl uenced by transformation conditions; (iii) whether transgenes are stably integrated and expressed over long time periods in citrus plants grown under natural environmental conditions; and (iv) whether correlations between integration patterns and transgene expression could be established in these plants.

A detailed survey of the plants allowed detection of four phenotypic off-type plants and a large variation of transgene integration and expression patterns among the popula- tion. Off-type plants were analysed and characterized as nucellar tetraploids, proba- bly originating from tetraploid starting nucellar tissues rather than from somaclonal variation events. Transgene integration and expression analyses revealed that: (i) a sig- nificant negative correlation was found between copy number and transgene expression; (ii) rearrangements of the T- DNA inserts did not imply low expression levels; and (iii) stability of integration and expression of transgenes was confi rmed for all the transformants grown under natural environmental conditions. These combined features validate transformation as a tool for genetic improvement of citrus (Cervera et al., 2000a).

A release of genetically modifi ed citrus plants under controlled fi eld conditions has also been ongoing since 1997 (Fig. 15.1E). The release site is located at the IVIA, in an experimental fi eld with an extension of 1638 m2. There are 130 trees, including 16 transgenic plants of Pineapple sweet orange, 16 transgenic plants of lime and 16 transgenic plants of Carrizo citrange (two plants from eight independent transgenic lines for each case). In addition, there are eight non-transgenic control plants from each of the species and an external border of 58 non-transgenic trees of Clemenules clementine. The purpose of the release is to investigate morphological and phenological characteristics of the transgenic trees, expression of the transgenes in leaves, fl ow-

ers and fruit, stability of the transgenes, transmission of the transgenes to the prog- eny, and the possibility of transgene disper- sal through the pollen to non-transgenic monoembryonic citrus trees (Clemenules clementine). The trial has been approved by the Spanish Ministry of Environment (permit No. B/ES/96/15) and is in accor- dance with Article 9 of Directive 90/220 of the European Union. This is the fi rst release of transgenic citrus plants to the fi eld. To date, the transgenic plants are morphologi- cally and phenologically normal, as are the non-transgenic controls, and transgenes are stably expressed over different seasons and in different plant tissues and organs (Fig. 15.2A, B and D).