Our initial aim was to develop a robust protocol for the stable genetic transformation of winter triticale. The approach taken rested heavily on transformation protocols established for its parental species, wheat and rye. The amenability of immature embryos to co-cultivation with A. tumefaciens in liquid culture was trialed, as this provides an efficient means of processing immature embryos in barley , and the same approach is effective in rye . Unfortunately, however, it does not seem to work well in wheat (unpublished data), where desiccation of the immature embryos appears to improve the transformation efficiency . In practice, triticale behaved like wheat in this respect, as transformation was only obtained when the immature embryos were co-cultivated on filter paper soaked with co-culture medium (Table 1). A similar study focusing on the spring triticale cultivar ‘Wanad’ compared the effectiveness of the three selectable marker genes BAR, HPT and NPTII driven by one of maize UBI-1, cauliflower mosaic virus 35S or A. tumefaciens NOS promoter, respectively , and concluded that the best combination was NOS::NPTII, even though NOS performs poorly in a monocotyledonous host . In the present study, the HPT selectable marker gene was preferred, a gene which has also proven useful e.g. in barley [18, 19], wheat and maize . The gfp reporter gene was an efficient tool for monitoring transgenesis and the subsequent expression of the transgene (Figure 1F-J) [13, 20].
Two of the three multiple T-DNA insertion events involved independent integration sites. In barley, >50% of multiple transgenic events induced by agro-infection involved only one integration site . A more comprehensive analysis of transformation outcomes has been made in Arabidopsis thaliana, where the number of T-DNA copies integrated at a single site appears to be dependent not only on the identity of the A. tumefaciens strain and the explant, but also on the transformation methodology as well as the origin of replication of the vector providing the T-DNA . The occurrence of transformation events with multiple T-DNA copies being integrated in independent genomic loci of triticale opens up the opportunity to generate transgenic segregants with reduced copy number. Moreover, co-introduction of effector and selectable marker gene using two different T-DNAs may give rise to selectable marker-free transgenics after independent segregation of the loci in the T1. While a similar case unintendedly occurred in the present study (TG5E01 T1 plant 6), a directed approach using barley has recently been presented by Kapusi et al. . Although Agrobacterium-mediated transformation generally results in the less frequent integration of truncated transgenes than biolistic transfer, as many as 44% of primary wheat transgenics have been shown to carry incomplete T-DNAs  with many involving truncations at the left T-DNA border . In barley, meanwhile, only 3% (of 260 primary transgenics analysed) retained the full T-DNA . Truncation of the T-DNA can be expected to result in a loss of transgene function, as was indeed the case in the present experiments that revealed truncations in 37.5% of the integrated T-DNAs analysed.
The non-Mendelian segregation of transgenes among T1 progeny is a commonplace observation, and several hypotheses have been promoted to explain this phenomenon, such as T0 chimerism, multiple independently assorting insertion loci and transgene silencing induced by multiple transgene copies or DNA rearrangements [24, 27–29]. In some cases, false positives can arise due to the expression of non-incorporated transgene cassettes including those carried by persisting Agrobacterium. Non-Mendelian transgene segregation has been noted in triticale , but since this observation was based on a histochemical reporter gene assay and did not include any DNA analysis, its basis could not be ascertained. In the present study, fewer transgenic progeny was obtained than expected in the case of TG5E03, which suggests this plant to be chimeric with regard to transgenicity. This interpretation is corroborated by DNA gel blot, PCR and leaf assay, which indicated that all functional elements be present in at least one of the two coupled T-DNA copies (Figure 4A-F). Nonetheless, one of the copies may have produced aberrant mRNA causing post-transcriptional gene silencing. However, the non-Mendelian segregation observed in the T2 families derived from TG5E03 is anticipated to be caused solely by transgene silencing in some siblings, because chimerism can generally be ruled out in generations later than T0.
As monitored using confocal laser scanning microscopy, gfp expression was widely distributed, but concentrated in the cytosol (Figure 5). This localization mirrors what has been observed in transgenic barley and wheat , where the level of reporter gene expression in the aleurone and the endosperm was comparable to that driven by either the barley bi-functional α-AMYLASE/SUBTILISIN INHIBITOR (ISA) or the wheat EARLY-MATURING (EM) promoter [31, 32].