Despite significant improvement for insect pest resistance that has been made possible through conventional breeding programmes, insect pests continue to cause yield losses of 5-30% in the rice crop. Advances in biotechnology such as the introduction and expression of novel genes such as Bt from Bacillus thuringiensis in different rice cultivars have complemented conventional breeding efforts for insect pest resistance (Datta et al. 1998; Tu et al. 2000). To improve transformation efficiency, it is necessary to use selectable markers as an integral part of the transformation process irrespective of the transformation methods employed. However, the production of marker-free transgenic plants is an integral step in crop improvement through transformation, as it addresses public concerns regarding antibiotic selectable markers, thereby making it easier to commercialize genetically modified crops and incorporate them into conventional breeding programmes. Several strategies, such as sitespecific recombination, homologous recombination, transposition and co-transformation, and the use of non-antibiotic selectable markers, have been developed to eliminate marker genes Puchta 2003).

Here, we report the production of marker-free transgenic indica rice cv. Azucena through the use of a co-transformation strategy. The transformation procedure was performed using particle bombardment (Biolistic^TM) that involved two plasmid vectors. The vectors used were pCIB4421, which contains a truncated cry1A(b) gene driven by the tissue-specific PEPC promoter and having a PEPC intron and 35S terminator sequence, and a separate vector, pGL2, that contains the selectable marker hygromycin phosphotranferase (hph) gene in between the 35S promoter and polyA tail (Datta et al. 1998). In the event of co-transformation, there are chances that the selectable marker gene and the gene of interest may integrate at distant genetic loci, so that, during subsequent segregating generations, meiosis will ensure the separation of these loci among the progenies.

Southern blot analysis of the primary transgenic showed a complex multi-tier banding pattern apart from the expected 1.8-kb cry1A(b) band that is not uncommon in the biolistic method of transformation (figure not shown).

Screening of 350 T₂ progenies of two independently transformed primary transgenic plants through Southern blot analysis for the cry1A(b) gene revealed two types of banding patterns among the progenies (Fig. 1). A majority of the progenies showed the same banding pattern as that of the T₀, while some of them showed a simple integration of the 1.8 kb expected size. From this result, we inferred that, during transformation, there were at least two separate integration events of cry1A (b) at different loci in the genome: one having an intact copy (or copies) and the other with rearranged copies of the gene. It is evident that these two loci segregated from each other in the T₂ generation. However, the progenies with both integration patterns showed 100% mortality of yellow stem borer (bioassay results not shown).

When the cry1A(b) genes screened by Southern blot were re-probed with the radiolabelled 1.1 kb hph probe, no hybridization was detected for the progenies having the simple

integration pattern. To confirm this result, we performed Southern analysis for both cry1A(b) and hph separately: on a selected set of progenies showing a simple integration pattern and on two T₂ plants (lanes 2 and 19) having the parental integration pattern (Figs. 2a and 2b). The Southern blot confirmed that, while the plants showing a complex integration pattern for cry1A(b) were positive for the hph gene, the plants having a simple integration of cry1A(b) did not have the hph gene. Therefore, we can conclude that the hph gene was tightly linked with the transgenic locus having rearranged copies of cry1A(b), and that it co-segregated together with this locus during meiosis to give hph-free plants with a simple integration pattern. This result correlates with the earlier reported segregation of the marker gene integrated in different individual loci (Tu et al. 2003). This study shows the possibility of obtaining transgenic plants free from selectable markers through delivery of a marker gene by a co-transformation strategy and screening for the segregation of the marker gene among the progenies in subsequent generations. The method could be improved in the future by optimizing various parameters such as the use of minimal and optimal vectors, the ratio of selectable markers and genes of interest, etc. However, the production of a large number of plants and detailed insights into molecular analysis of unlinked loci in subsequent generations may help to achieve success in the development of marker-free transgenic plants that can be used in crop improvement programmes.

Acknowledgement

Thanks are due to BMZ/GTZ, Germany, and the Rockefeller Foundation, USA, for supporting the project and Bill Hardy for editorial help.

References

Datta, K., A. Vasquez, J. Tu, L. Torrizo, M.F. Alam, N. Oliva, E. Abrigo, G.S. Khush and S.K. Datta, 1998. Constitutive and tissue-specific differential expression of cryIA (b) gene in transgenic rice plants conferring resistance to rice insect pest. Theor. Appl. Genet 97: 20-30.

Puchta, H., 2003. Marker-free transgenic plants. Plant Cell, Tissue and Organ Culture 74: 123-134.

Tu, J., K. Datta, N. Oliva, G. Zhang, C. Xu, G.S. Khush, Q. Zhang and S.K. Datta, 2003. Site-independent integrated transgenes in the elite restorer rice line Minghui 63 allow removal of a selectable marker from the gene of interest by self-segregation. Plant Biotechnology Journal, 1: 155.l.

Tu, J., G. Zhang, K. Datta, C. Xu, Y. He, Q. Zhang, G.S. Khush and S.K. Datta, 2000. Field performance of transgenic elite commercial hybrid rice expressing Bacillus thuringiensis part-endotoxin. Nature Biotechnology 18: 1101-1104.