Maize transformation with Agrobacterium tumefaciens
--Lupotto, E, Reali, A, Passera, S, Chan, M-T*
*Collaboration with: Institute of Molecular Biology, Academia Sinica, Nangkang Taipei, Taiwan

The application of Agrobacterium-mediated transformation to monocotyledonous species, such as rice and maize, has been recently reported. The main characteristics of the Agrobacterium system in these species, as for dicotyledonous species, are: i) high frequency of transformation; ii) proper integration of the foreign gene into the host genome; iii) low copy number of the gene inserted, resulting in most cases in a correct expression of the transgene itself. A limitation of the system is represented by the strict interaction between the genotype of the plant and the Agrobacterium strain, and the need to identify and to supplement specific signal molecules for the vir genes induction during the co-cultivation period (acetosiryngone or sinapinic acid).

Although not being natural hosts for Agrobacterium tumefaciens, monocotyledonous species seem to be in some instances susceptible to the infection. Studies on Agrobacterium infection of maize were first reported by Grimsley et al. (Nature 325:177-179,1987) and Gould et al. (Plant Physiol. 95:426-434,1991), but the first evidences of the possibility of application of Agrobacterium-mediated transformation of cereal species comes from the work of Chan (Chan et al., Plant Cell Physiol. 33:577-583,1992) and Hiei (Hiei et al., Plant J. 6:271-282, 1994), who first obtained transgenic rice plants by means of transformation of immature embryos with Agrobacterium tumefaciens. Most recently, the technique has been successfully applied to maize, and transgenic maize plants obtained at high frequency (Ishida et al., Nature Biotechnol. 14:745-750,1996). Ishida and co-workers report on the efficient transformation of maize inbred A188, and of some crosses between A188 and other inbreds.

Due to the ease of handling the Agrobacterium-mediated transformation, it would be advantageous to identify key parameters which allow the methodology to be applied to agronomically relevant genotypes, such as important inbred lines utilized in programmes of maize breeding. The present work was developed with the aim to extend the methodology of transformation with Agrobacterium tumefaciens to maize genotypes of relevant agronomical importance and, within these genotypes, to identify key parameters needed for transformation, and subsequent tissue culture and regeneration of transgenic plants.

To this purpose, eleven Lo inbred lines of maize, bred by our Institute, have been considered for Agrobacterium-mediated transformation. Among these, four were previously tested in vitro for tissue culture response and plant regeneration: Lo1054, Lo1056, Lo951 and Lo881 (Locatelli et al., MNL.66:17, 1992). Results obtained were compared with two reference genotypes: A188 and B73.

Five Agrobacterium strains, characterized by different chromosomal backgrounds and Ti-plasmid content, were chosen, all carrying in a binary system the chimeric construct -INTGUS- driven by the CaMV 35S promoter, for early detection of transformation events with the histochemical staining of the tissues for B-glucuronidase expression. The Agrobacterium strains, indicated as: C58c1(pGUSINT), Agt121(pGUSINT), EHA101(pMTCA23GUSINT), EHA105(PGUSINT), EHA105(pMT1) and LBA4404(pTOK233) were previously described in detail (Chan et al., Plant Cell Physiol. 33:577, 1992; Smith, RH and Hood, EE, Crop Sci. 35:301, 1995; Hiei et al.,Plant J. 6: 271-282, 1994).

Maize plants were grown in the field during summer seasons 1996 and 1997, hand pollinated, and immature embryos dissected at 12 DAP, explanted on N6I medium (Lupotto and Lusardi, Maydica 33:163, 1988). For transformation, immature embryos and primary embryogenic calli were co-cultivated with Agrobacterium. The procedure adopted followed the protocol described by Hiei et al. (1994). For the infection, bacteria were grown in AB medium supplemented with each strain's proper antibiotic selection, for three days. The bacteria were collected and diluted in low pH (pH 5.2) infection medium (LSinf) at high density (O.D. 1-1.2) in the presence of 100 uM AS. Explanted embryos were immersed in LSinf, vortexed at maximal speed 20 seconds, and incubated 10 minutes. Blotted dry embryos were subsequently co-cultivated for three days on LSD1.5, 100uM AS medium, transferred on LSD1.5 plus 250 mg/l cefotaxime (or 200 mg/l timentin) for 2 days, and tested for infection with histochemical GUS assay, or cultured for embryogenic callus induction.

For the experiments of timing, in the case of infection of A188 with Agrobacterium strains 2 and 6, embryos were explanted at 11, 12, 13, 14, and 15 DAP. Callus induction frequency was measured as % embryos giving an embryogenic callus. The extent of infection was rated from 0 (not infected) to 5 (heavily infected). The frequency of transformation was recorded as % embryos exhibiting dark blue areas.

Preliminary experiments showed that a microbombardment performed with the particle gun device, improved the experiments in creating microwounding in the scutellum, thus enhancing the extent of infected tissues. In the case of pre-bombardment, 25 embryos were arranged with the scutellum exposed on the surface of culture medium and bombarded at 900 PSI, at a distance of 5 cm from the stopping screen, in 26 in Hg vacuum, with 1 um diameter gold particles, according to the customer instructions for PDS 1000/He device.

The efficiency of transformation was evaluated for each genotype, considering the interaction maize X Agrobacterium strain, the medium used for infection and co-cultivation, the use of elicitors (sinapinic acid and acetosyringone), and the age of the immature embryos. Each inbred line considered was also evaluated for tissue culture response, in order to choose the best combination Agrobacterium strain X responding inbred for routine transformation procedures.

The results obtained showed that Agrobacterium strains C58c1(pGUSINT) and EHA101(pMTCA23GUSINT) were most effective in the attachment to the maize genotypes A188, Lo 1095, Lo1061, Lo951 and Lo1056 (rated 4-5); effective, though at a lesser extent, were genotypes Lo1054, Lo881 and B73 (rated 3). No infection was detected in the cases of genotypes Lo1074, which was in any case not attached by any of the tested strains, as well as Lo1023. The Agrobacterium strain LBA4404(pTOK233) was effective on A188, B73 and Lo1061. No attachment in any of the genotypes tested was detected for strains EHA105(pGUSINT) and EHA105(pMT1). The supplement of 100 uM acetosyringone in the LSinf and LSD1.5 co-cultivation medium was beneficial for Agrobacterium attachment.

Considering the inbred A188 as reference genotype, and Agrobacterium strains C58c1(pGUSINT) and EHA101(pMTCA23GUSINT), an experiment of timing for infection was performed with the aim to identify the proper stage of embryo development for maximal attachment. Embryos were explanted every 24 hours from day 11 to day 15 after pollination, days 11 and 12 being the optimal stage for tissue culture and regeneration (87 and 94% respectively for day 11 and 12). Interestingly, this stage also resulted in the highest rate of infection (76 and 90% respectively for day 11 and 12).

A first set of experiments were performed according to the main protocol of Ishida et al. (1996), where dissected embryos are vortexed in the presence of the bacteria. In this case, some wounding occurs in the scutellar region, where the Agrobacterium attaches. However, in this case, bacteria preferentially attached to the ventral part of the embryo, in the region of the embryo axis. In order to facilitate Agrobacterium infection of the embryogenic-competent scutellar cells, we adopted the strategy of microwounding by means of microbombardment of the scutella with gold particles. In this case, the areas of infection were significantly enlarged. It was subsequently shown that the joint technologies (Agrobacterium plus microwounding) did not alter the subsequent capability of the infected tissue in further development.

The data obtained from the work allow drawing the following considerations. In maize, immature embryos and embryogenic primary calli derived from them can effectively be infected with an array of Agrobacterium tumefaciens strains, by means of a co-cultivation procedure which includes preculture of the Agrobacterium in a minimal medium, and the addition of acetosyringone 100 uM in the co-cultivation medium. Regarding the phase of infection and transformation, histologically detected with GUS assay, several genotypes have been revealed to be susceptible, and among these are several agronomically interesting inbred lines. This fact validates the approach and extends the results published by Ishida et al. (1996), who suggested that only A188 and A188-derived crosses could be infected. There was in any case clearly a strong genotypic interaction between the maize genotype and the Agrobacterium strain. Unwounded immature embryos offer to Agrobacterium, as preferential attachment site, the embryo axis, which is not useful for further callus induction and plant regeneration. In this case, the infection occurs in non-competent tissues for embryogenesis. This problem can be overcome by means of microwounding of the scutellar surface with particle gun mediated- bombardment with gold particles (PDS 1000/He Bio Rad device). Primary embryogenic calli also offer an opportune target tissue and can be infected at high rate with the same procedure. Utilizing embryogenic calli as target, we had to face problems of tissue necrosis and difficulty in callus growth recovery. The problem is on the way to being solved with the use of antioxidants in the culture medium, and the use of timentin instead of cefotaxime as antibiotic for Agrobacterium elimination. In our experiments, particle bombardment previous to infection increased the percentage of the Agrobacterium-infected embryos and increased the extent of the infected areas at the surface of the scutellum. Finally, the combination of acetosyringone as elicitor of virulence and the preculture of the bacteria in minimal medium, allowed the effective infection of a broad range of genotypes of agronomical interest.

Please Note: Notes submitted to the Maize Genetics Cooperation Newsletter may be cited only with consent of the authors

Return to the MNL 72 On-Line Index
Return to the Maize Newsletter Index
Return to the Maize Genome Database Page