Introduction

Maize is one of the most important crops around the world because of its importance as food and feed in the past and present; thus, breeding technology in this crop has been the subject of intense efforts resulting in several biotechnology approaches applied mainly in order to incorporate desiderable traits on several maize lines (reviewed by Armstrong 1998). Among the biotechnology tools, those related to transferring DNA have received special attention, resulting in several strategies such as biolistic (Fromm et al. 1990; Gordon-Kamm et al. 1990) or Agrobacterium tumefaciens (Ishida et al. 1996; Frame et al. 2002; Huang and Wei 2005). A lot of advantages have been associated with the A. tumefaciens-mediated transformation, as is the possibility of transferring large DNA segments into recipient cells, the generation of a high number of stable, low-copy number, correctly expressing transgenic events, while producing less transgene rearrangements (Frame et al. 2002; Shou et al. 2004). Additionally, this represents a simple technology with low cost. In light of the above mentioned advantages, several studies related to the Agrobacterium-mediated transformation of maize have been focused in diverse factors associated with this system, such as the plant genotype, types and developmental stages of the explants, strains and cell bacterial density, the addition of phenolic compounds, co-cultivation period, vectors, pH, temperature as well as the composition of the medium (Armstrong 1998; Amoah et al. 2001; Huang and Wei 2005; Frame et al. 2006). Some promising developed protocols have been based on the use of super-binary vectors (Ishida et al. 1996; Negrotto et al. 2000) or the addition of antioxidants through the Agro-infection (Frame et al. 2002; Lupotto et al. 2004).

To our knowledge, almost all the previously reported protocols were developed using the maize model genotype A188 and their hybrid Hi-II (A188 × B73), which are characterized by high frequency of embryogenic callus induction and plant regeneration (D’Halluin et al. 1992); these genotypes have poor agronomical value, therefore it would be a time-consuming and costly procedure to introduce transgenes into local maize varieties by backcrossing (O’Kennedy et al. 2001; Lupotto et al. 2004). Additionally, the recovery of progeny with both the transgenic trait and suitable agronomic traits is often difficult due to incompatible heterotic groups and poor combination ability (O’Kennedy et al. 2001). Thus, some crosses between A188 and other agronomically important inbred lines have been used to generate transgenic maize plants (Lupotto et al. 2004), but this only partially resolves the above mentioned problems. Additionally, all the “laboratory genotypes” from which protocols have been developed, are adapted to temperate zones and little or not attention has been focused on the transformation of lines adapted to tropical regions, where crop productivity is often low due to stresses such as virus, insects, poor and metal-contaminated soils, and adverse climatic conditions (Bohorova et al. 1995; Rascón-Cruz et al. 2004). That is why improvements of this transformation technology are needed to introduced successfully useful genes directly into the genome of several agronomically important and/or specific land-adapted inbred lines (Carvalho et al. 1997; Rascón-Cruz et al. 2004).

A successful genetic transformation requires an efficient gene delivery coupled with a plant tissue culture system whose whole transgenic plants may be regenerated from transformed cells (Ishida et al. 1996; Negrotto et al. 2000; Amoah et al. 2001; Frame et al. 2006). In this sense the maize embryo-scutellum having a lot of competent cells for somatic embryogenesis, is a proved primary explant from which fertile plants can be regenerated at high frequency (Carvalho et al. 1997; Brettschneider et al. 1997; Rascón-Cruz et al. 2004), and has been selected as a target tissue for transformation by electroporation (D’Halluin et al. 1992) and biolistic (Fromm et al. 1990; Gordon-Kamm et al. 1990; Brettschneider et al. 1997). The fact that in maize embryos infected with A. tumefaciens transformation mainly occurs in host cells belonging to the embryo axis side and the apical end (Frame et al. 2002), and non-embryogenic calli emerged from these, this is an obstacle for reaching efficient rates of transformed embryogenic-calli generation (Welter et al. 1995).

In this report, we describe a transformation system of a tropical-adapted maize line by way we have achieved a direct production (i.e., without a previously cross with any other line) of fertile transgenic plants. Diverse components of the A. tumefaciens-transformation process were taking into account: (i) the effect of silver nitrate enhancing the production of regenerable embryogenic-calli beginning from the embryo-scutellum, and (ii) transformation rates and the distribution of the T-DNA toward the embryo-scutellum by applying physical microwounds and vacuum infiltration. Additionally, molecular evidence of transgene integration into fertile maize plants is presented.

Materials and methods

Plant material

Preliminary studies in which several tropical maize lines were pre-screened to identify genotypes susceptible to Agro-infection, and capable of producing regenerative type-II calli has been conduced in our lab (unpublished data); according to these, at the present work we used the maize (Zea mays) tropical line LPC13 supplied by the Instituto Nacional de Investigaciones Forestales, Agrícolas y Pecuarias de México. Immature zygotic embryos (1.5–2.0 mm) were aseptically dissected from greenhouse-grown ears harvested 12 to 15 days after pollinization (DAP).

Induction and maintenance of calli and plant regeneration

Induction and maintenance of embryogenic-callus from immature embryos were evaluated based on the previously described N6C1 medium (Bohorova et al. 1995), which consisted of N6 basal medium (Chu et al. 1975) supplemented with 200 mg l−1 casein hydrolysate, 2.302 mg l−1 l-proline, 3% sucrose and 2 mg l−1 Dicamba; addition of three doses of AgNO3 at 10 (N6C1-10), 15 (N6C1-15) and 20 mg l−1 (N6C1-20) into the base N6C1 medium, were evaluated for its ability to induce embryogenic type-II callus from the embryo-scutellum as has already been demonstrated on several tropical maize lines (Vain et al. 1989; Welter et al. 1995; Carvalho et al. 1997). For callus initiation and maintenance, immature embryos were cultured oriented with the embryo-axis side in contact with the medium (scutellum side up) and incubated in the darkness at 27°C; embryogenic tissue was subcultured to fresh medium every 20 days.

Regeneration capacity of the different induced calli was evaluated for three different regeneration media, all of them based on basal MS medium (Murashige and Skoog 1962). MS1, hormone-free and 20 g l−1 sucrose (Lupotto et al. 2004); MS2, hormone-free, 60 g l−1 sucrose and 100 mg l−1 myo-inositol (Frame et al. 2002); and MS3, 20 g l−1 sucrose, 0.5 mg l−1 indol−3-acetic acid (IAA) and 1 mg l−1 6-bencylamino purine (6-BAP) (Bohorova et al. 1995). The pH of all media were adjusted with NaOH to 5.8 and solidified with 7.5 g l−1 purified agar. For regeneration, calli were incubated in a growth chamber at 26°C with a 16:8 light:dark photoperiod.

Agrobacterium tumefaciens strains and transformation vector

Strains LBA4404, EHA101, and EHA105 (Hood et al. 1993) of A. tumefaciens, all containing the standard binary vector pCAMBIA3301 (CAMBIA, Canberra, Australia), were tested in their ability to infect immature embryos of maize. The T-DNA region of this plasmid (Fig. 1) contains the reporter gene gus and the selectable marker gene bar, both under the control of CaMV 35S promoter. Moreover, the gus gene contained a portable intron (gus-int) to prevent GUS activity coming from A. tumefaciens cells.

Fig. 1
figure 1

Schematic representation of the T-DNA region of the standard binary transformation vector pCAMBIA3301. LB, left border; RB, right border; bar, phosphinothricin acetyltransferase gene; gus-int, β-glucuronidase gene containing an intron; P35S, CaMV 35S promoter; T35S, CaMV 35S terminator; TN, nopaline synthase terminator. For Southern analysis, a 700 bp probe comprising a region between P35S and the bar gene, and labeled by PCR with digoxigenin is indicated by the filled box. B, BamHI

Full size image

Tissue culture media and transformation strategy

Infection, cocultivation and resting media used in the present work were those reported by Frame et al. (2002). All of these contained N6 salts and vitamins (Chu et al. 1975), 1.5 mg l−1 2,4-dichlorophenoxyacetic acid, and 0.7 g l−1 l-proline and addition of the following components: infection medium contained 68.4 g l−1 sucrose, 36 g l−1 glucose (pH 5.2), and supplemented with 100 μM AS (Sigma, St. Louis) before use; cocultivation medium contained 30 g l−1 sucrose, 350 mg l−1 l-Cysteine, 0.85 mg l−1 silver nitrate, 100 μM AS, and 3 g l−1 gelrite (pH 5.8); resting medium contained 30 g l−1 sucrose, 0.5 g l−1 MES, 0.85 mg l−1 silver nitrate, 250 mg l−1 cefotaxime, and 7.5 g l−1 purified agar (pH 5.8). Selection was applied using callus-induction medium supplemented with bialaphos at 1.5 mg l−1 (selection I) and 3 mg l−1 (selection II).

For the infection, A. tumefaciens cultures were grown for 2 to 4 days at 20°C on YEP medium (5 g l−1 yeast extract, 10 g l−1 peptone, 5 g l−1 NaCl, 15 g l−1 agar, pH 6.8), amended with 50 mg l−1 kanamycin (for transformation vector) and 50 mg l−1 gentamycin (for strain genomic background). Prior to infection, one full loop (3 mm) of bacterial culture was scraped and suspended in infection medium (68.4 g l−1 sucrose, 36 g l−1 glucose, pH 5.2) and supplemented with 100 μM AS as described by Frame et al. (2002). Bacterial suspensions of 5 ml were measured and adjusted to three different optical densities (OD550 nm): 0.1–0.5, >0.5–1.0, and >1.0–1.5. The dissection of immature embryos from the ears and their infection with the pre-induced Agrobacterium cells were done as described by Negrotto et al. (2000), except that some immature embryos were subjected to physical microwound treatments and/or vacuum infiltration before and during the infection, respectively.

Physical microwound treatments

Particle gun wounding treatments (bombardment) were done as described by Brettschneider et al. (1997). The embryos were arranged with the scutellum exposed on the surface of the infection medium (bacterial-free) amended with 15 g l−1 agar, around the perimeter of a 1-cm circle in 60 mm × 20 mm Petri plates. Particle-bombardments were performed using a Particle Inflow Gun-PIG applying 28 mmHg vacuum as described by Droste et al. (2000). Helium pressures of 60, 80, and 120 lb/in2 were evaluated, and all the shots were done at a distance of 19 cm from the stopping screen using 1.8 μm average diameter tungsten particles (GE, Cleveland). Uncharged particles were prepared according to Brettschneider et al. (1997). The embryos were incubated with the bacteria soon after particle-bombardment. Ultrasound treatment (sonication) was performed according to Santarém et al. (1998). Immature embryos were imbibed in 1 ml of infection medium bacterial-free and treated with ultrasound for 5, 10, and 30 s at 50 Hz using an ultrasonic water bath (Branson, Ultrasonic, Corporation, EUA); after sonication, medium was replaced for the same volume of bacterial suspension.

For each physical microwound treatment, 30 wounded-embryos were submerged into the bacterial suspension for 15 min; as a control, 30 non-damaged embryos were infected as well. Additionally, in order to allow the bacterial cells to penetrate into the subperipherical host cells, 28 mmHg of vacuum infiltration were applied, during the infection step, to 30 wounded-embryos generated by each physical microwound treatment, and to 30 non-damaged embryos used as a control. After infection, embryos were transferred to the surface (scutellum side up) of cocultivation medium and incubated here for 3–5 days at 20°C in the dark; after which, embryos were transferred to resting medium for 7 days (27°C, dark).

Histochemical GUS assays

Transient and stable GUS activity studies were carried out on immature embryos 3 days after cocultivation according with Jefferson (1987). Immature embryos were placed in a GUS assays mix; after 5 min of vacuum infiltration, the reactions were incubated overnight at 37°C. GUS activity was determined by placing the embryos on a grid and estimating both the blue spots and the percentage of the embryos surface that showed blue sectors. Histochemical GUS assays were also used to assess stable expression of the gus gene both in PPT-resistant calli samples and transformed leaf tissue which was visualized by removing the GUS assay mix, and rinsed twice with 70% ethanol.

Selection of transformed embryos/calli and recovery of transgenic plants

After resting, embryos were transferred to the first selection medium (1.5 mg l−1 bialaphos) for 3 to 4 weeks before they were transferred to selection II medium until the first putative transformed calli began to emerge. Putatively transformed plants were regenerated from bialaphos-resistant calli by transferring tissues to Magenta boxes containing regeneration medium, supplemented with 3 mg l−1 bialaphos, and placed them in a growth chamber at 27°C with a 16:8 light:dark photoperiod.

Molecular analysis and progeny segregation analysis

Total genomic DNA was isolated from approximately 1 g of fresh leaf tissue from putative transformed and control non-transformed plants as described by Bohorova et al. (1995). Approximately 30 μg of genomic DNA from each sample were digested with a threefold excess of the BamHI restriction enzyme, separated on 0.8% agarose gels and transferred to nylon membranes. A 700-bp PCR digoxigenin-labeled probe comprising a region between CaMV 35S promoter and the bar gene was used for Southern blot hybridizations (Southern 1975).

To establish segregation ratios for expression of the bar gene in progeny, a Basta™ test was applied (Rascón-Cruz et al. 2004). Transgenic T0 plants were self-fertilized and progeny (T1) grown in the greenhouse were examined for resistance to PPT and GUS expression. Herbicide resistance was scored by painting the fifth or sixth leaf near the tip of the youngest fully extended leaf with a 2% (v/v) Basta™ solution containing 0.1% Tween 20. One week after the first Basta™ treatment, the plants were painted again with a second Basta™ application and assessed for damage 1 week after this.

Results and discussion

Callus induction

Approximately 80% of the immature embryos grown on medium containing AgNO3 were able to generate calli, compared to 58.2% of those induced in absence of AgNO3. The total rate of embryos producing callus increased 20–25% by the addition of silver nitrate to the induction medium, while the fraction of induced type II-callus increased threefold to fourfold at all the concentrations of AgNO3 with respect to the control medium; 15 mg l−1 seemed to be the optimal level for the induction of friable embryogenic type-II callus because a high concentration (20 mg l−1) changed this tendency (Table 1). As it is shown at Fig. 2a, soft, granular and non-embryogenic calli were induced at the apical end and on the embryo axis side (exs) from the immature embryos grown in absence of AgNO3. Besides inducing higher rates of calli, the addition of silver nitrate into the induction medium, promoted the emergence of the somatic embryogenesis from embryo-scutellum (Fig. 2b); moreover, a high fraction from calli coming from this tissue was embryogenic/organogenic-like which generated several organized somatic embryos arisen from an undifferentiated base (Fig. 2c).

Table 1 Effect of AgNO3 on frequencies of type-II callus induction from immature embryos and regeneration capacity of the induced calli
Full size table
Fig. 2
figure 2

Callus induction from immature zygotic embryos and plant regeneration. (a) Soft, granular non-embryogenic callus emerged from embryo axis side (exs) after 10 days of culture on N6C1 medium, where little or null callus induction was observed from embryo-scutellum (es). (b) Development of compact, nodular embryogenic callus on embryo-scutellum (es) of immature embryos cultured for 15 days on N6C1-15 medium. Somatic embryos are indicated by arrows. (c) Late embryogenic culture of maize with organized somatic embryos arising from an undifferentiated base. (d) Plantlets developing from embryogenic callus plated on regeneration medium. (e) Well-developed plantlets with established root system. Bar = 1.0 mm

Full size image

The ion Ag+ promotes somatic embryogenesis via an increase in the endogenous ABA levels; additionally it also competes for the binding site of the ethylene produced during the embryo extraction without interrupting the ethylene biosynthesis, important to the growth and development of plants (Vain et al. 1989; Songstad et al. 1991). Thus, the enhanced somatic embryogenesis showed by the addition of AgNO3, agreed with previous observations made on some maize genotypes. Vain et al. (1989) demonstrated that 5–10 mg l−1 AgNO3 enhanced the induction of type-II callus from immature embryos of the maize line A188. A positive response was observed within a range of 1.8–18 mg l−1 for inbred line B73 and its derivatives (Songstad et al. 1991; Welter et al. 1995). Even though a few tropical maize genotypes have shown the capacity to produce type-II callus, this was improved by the addition of 15 mg l−1 AgNO3 (Bohorova et al. 1995; Carvalho et al. 1997). According to Huang and Wei (2004), the addition of AgNO3 (10 mg l−1) supported an increased frequency of embryogenesis using mature embryos of maize elite inbred lines. Additionally, increased somatic embryogenesis by the addition of AgNO3 has been reported in different cultures such as wheat (Wu et al. 2006), coffee (Fuentes et al. 2000) and L. barbarum (Li et al. 2001). In the present work, rates of somatic embryogenesis were increased while the doses of AgNO3 were also increased until a concentration of 15 mg l−1; when applying doses of 20 mg l−1 the induction diminished (Table 1). This tendency has been reported by Huang and Wei (2004), when low concentrations of AgNO3 significantly increased the production of embryogenic callus in subculturing medium, but concentration up to 20 mg l−1 inhibited it strongly. Li et al. (2001) also reported that when excess of AgNO3 is used, Ag+ turn on toxic and somatic embryogenesis is remarkably inhibited. Thus, determining the minimal amount of Ag+ necessary for inducing this embryogenic response is essential.

Plant regeneration

In spite that some calli can be classified as embryogenic, not necessarily implies that they possess regenerability; thus, type-II calli induced from all the tested media were evaluated for their regenerative capability. Although different regeneration media were initially evaluated, best results were found with the MS3 medium (Bohorova et al. 1995); thus, it was selected for posterior plant regeneration. Significant increases in regeneration rates were found when calli were induced under presence of AgNO3; thus, whereas only 4.4 plants per g FW of calli induced from control medium were regenerated, up to 12.3 and 19.1 plants per g fresh weight (FW) of calli induced on the presence of 10 mg l−1 (corresponding to 2.8-fold) and 15 mg l−1AgNO3 (corresponding to 4.4-fold) (Table 1) were regenerated. It is well known that AgNO3 influences cell division, cytodifferentiation, and regeneration capacity (Songstad et al. 1991). At present work, the capacity to form embryogenic calli was correlated with the efficiency to regenerate plantlets (Table 1). Similar results were reported by Carvalho et al. (1997) when testing the regeneration ability of embryogenic calli induced in several maize inbred lines; the highest number of regenerated plants/g calli in each inbred were observed in medium containing 15 mg l−1 AgNO3. In addition, improved somatic embryogenesis and regenerative capability by using AgNO3 have been reported in other cultures as wheat (Wu et al. 2006) and coffee (Fuentes et al. 2000). Since ethylene is considered to suppress morphogenesis in vitro, AgNO3 and other ethylene inhibitors as aminoethoxynylglycin (AVG) have been used to improve the regenerative capability of maize and other crops (Songstad et al. 1991; Fuentes et al. 2000; Wu et al. 2006).

Effect of the bacterial strain and Agrobacterium cell density on the efficiency of embryo transformation

The transformation efficiency of each tested strain on the tropical maize line LPC13 was evaluated by the transient GUS expression showed by infected-embryos. Results indicated that the EHA101 and EHA105 strains were able to generate at least 30% of GUS-positive embryos in all the tested Agrobacterium cell densities, while those cultivated with LBA4404 strain had little or null GUS staining (Data not shown). In the analyses presented here, an embryo having at least one blue spot was scored as a GUS-positive. Agrobacterium cell density had a marked effect on transient expression, having the best results at OD550 nm = >0.5–1.0, showing rates of GUS-positive embryos up to 67% (EHA105) and 59% (EHA101) (Data not shown). According with the mentioned results, the following experiments were performed using the strain EHA105 at OD550 nm = >0.5–1.0. The A. tumefaciens hypervirulent strain EHA105 has shown to be highly efficient for transforming derivative hybrids of the laboratory maize line A188 (Lupotto et al. 2004), and other important cultures as soybean (Santarém et al. 1998) and rice (Cheng et al. 1998). The previously reported genotype-dependency between the bacterial and vegetal cells was confirmed; thus, searching for the most efficient bacterial strain should be carried out for each variety of interest.

Effect of physical wounds and vacuum infiltration on transient GUS expression

The effect of the applied physical treatments was evaluated taking into account the rate of GUS-positive embryos (A) and the percentage area of transient GUS expression showed by them (B). When vacuum infiltration was applied, improved frequencies up to 17.6 and 28.9% of GUS-positive embryos were obtained in control non-damaged and particle-wounded embryos (Table 2), respectively. Enhanced percentages of transient blue area of 30.2% and 33.6% were observed in particle-wounded embryos bombarded at 80 and 120 lb/in2, respectively; while higher increases up to 160% were obtained in the control non-damaged and particle-wounded embryos bombarded at 60 lb/in2 (Fig. 3a–d). Amoah et al. (2001) reported similar results when infiltration was applied for 1 h in Agro-infected wheat inflorescences.

Table 2 Effect of wound treatments applied on maize immature embryos in transient gus gene expression
Full size table
Fig. 3
figure 3

Effect of physical wounds on rate and distribution of transient GUS expression of infected embryos. (a) In non-infiltrated control embryos, blue area was visible on the apical end and on the embryo axis side. (b) Non-wounded but infiltrated embryos. (c) Non-infiltrated but particle-wounded embryos at 60 lb/in2. (d) Particle-wounded and infiltrated embryos at 60 lb/in2. (e) Wounds applied by sonication (10 s) on non-infiltrated embryos allow to extend the blue area toward all the scutellar side in a homogeneous way. (f) Infiltration applied on wounded embryos by 10 s of sonication decreases the transient GUS expression. Bar = 0.5 mm

Full size image

Regarding the sonication treatment, we found a contrary effect when vacuum infiltration was applied, since a dramatic reduction was observed in the percentage of both GUS-positive embryos and the blue staining area at all the sonication times assayed (Fig. 3e and f); thus, we observed a reduction of 31.1% in the percentage of GUS-positive embryos sonicated for 30 s, and a decrease of 61.1% in the transient blue area showed by the GUS-positive embryos sonicated for 10 s (Table 2). This may indicate a synergistic effect between sonication and infiltration damages, which tends to reduce the transient GUS expression perhaps due to a decrease in cell viability.

Due to the particular nature of each microwound treatment (i.e., bombardment and sonication) and the potential synergistic damage over the immature embryos when they are combined with vacuum infiltration, we compared the effect of each microwound treatment alone or combined with vacuum infiltration, as a particular treatment. Not significant differences were observed in the total transient GUS expression (A × B) rates between particle-wounded embryos bombarded at 60–80 lb/in2 and the control non-damaged ones (Table 2). However, differences in the location of blue staining were found in these treatments; while in non-damaged embryos the transient GUS expression was visible at the embryo axis (Fig. 3a and b), in the particle-wounded embryos the blue staining was located mainly in the embryo-scutellum (Fig. 3d and e). This could impact the generation of transformed embryogenic callus induced. Using maize elite inbred lines, Lupotto et al. (2004) reported that in all the cases in which pre-bombarded embryos were used, infectivity of the Agrobacterium increased without altering the capability of the infected tissue to further develop an embryogenic callus. Increases in the transformation frequencies of tobacco leaf explants and sunflower meristems by using a transformation protocol based on particle-wounded/Agrobacterium-treated tissues, were also reported by Bidney et al. (1992). Similarly, Droste et al. (2000) reported up to 270 blue foci per infected clumb when using particle-wounded soybean embryogenic clumbs prior to Agro-infection, while transient GUS expression was not detected in non-damaged ones. At the present work, decreases of transient GUS expression were observed in particle-wounded embryos by bombardment at 120 lb/in2, which could be due to the magnitude of the damage generated suggesting that the particles should penetrate cells preserving the integrity of them.

Best rates of total transient GUS expression (A × B) for 31% and 47% were obtained using wounded-embryos by sonication for 5 and 10 s, representing increases of 40% and 115%, respectively, compared to the control (+V) treatment from which only 21.86% of total transient GUS expression was obtained (Table 2). Nevertheless, when ultrasound treatment was increased from 10 to 30 s, a dramatic reduction of 47.11 to 7.91% (respectively) was observed, which could be due to a decrease in cell viability caused by long times of exposures to ultrasounds (Santarém et al. 1998; Liu et al. 2006). A greater percentage of transient GUS expression on sonicated/infected-embryos was visible on the embryo-scutellum (Fig. 3e); this represents an important improvement toward the potential regeneration of whole transgenic plants, since the regenerable embryogenic calli are mainly induced from that tissue. To our knowledge this is the first report in which sonication is evaluated for its ability to promoting increases in transformation rates of maize immature embryos by A. tumefaciens; however, several protocols usually termed sonication-assisted Agrobacterium-mediated transformation (SAAT) have been developed in other cultures. Santarém et al. (1998) reported increases of 20% in the rates of GUS expression using SAAT-treated immature soybean cotyledons, instead of Agro-infected non-damaged ones. Improvements on the proportion of GUS-positive wheat inflorescences of 35% were obtained by applying ultrasounds for 6 s prior the Agro-infection; however, sonication reduced the number of spots produced per explants (Amoah et al. 2001). Better rates of transient GUS expression were obtained in ultrasound-wounded embryos of loblolly pine infected with wild A. tumefaciens strains than in those non-damaged embryos infected with A. tumefaciens strains carrying extra copies of the virB, virC and virG genes (Tang 2003). Transient GUS and GFP expression were also enhanced when sonication was applied in sunflower shoot-tip explants before the Agro-infection; however, this did not increase the stable expression in regenerated shoots (Weber et al. 2003).

Although the addition of acetosyringone (AS) to induce the vir-genes is routinely used in several Agrobacterium-transformation protocols, physical wounds lead to increase in the number and distribution of cells having transient GUS expression, producing not only attachment sites for bacterial cells, but also inducing synthesis of endogenous elicitors of the bacterial vir-genes (Bidney et al. 1992; Santarém et al. 1998; Weber et al. 2003); additionally, the wounded-induced plant cell elicitors could lead to a higher expression of the vir-genes, initially induced by AS.

Optimization of transformation conditions and generation of bialaphos-resistant plants

After the statistic analysis of the treatments shown at the Table 2 (last column), we took into account the most successful combinations which are expected to produce a big number of regenerated transformant plants; thus, five transformation procedures were designed: (1) control, infecting non-damaged embryos; (2) B60V and (3) B80V, infecting particle-wounded embryos bombarded at 60 and 80 lb/in2; these three were accompanied by vacuum infiltration during the infection step; (4) S5 and (5) S10, infecting ultrasound-wounded embryos for 5 and 10 s. All the procedures were carried out using the A. tumefaciens strain EHA105::pC3301 at an OD550 nm = 0.5–1.0.

Although bialaphos-resistant calli were induced from all the transformation procedures, the frequencies were different among them (Table 3). Even though the transient GUS expression in the infected non-damaged embryos was high at the embryo axis (Fig. 3b), a low frequency of bialaphos-resistant calli (9.5%) was obtained from the control procedure; this could be due to generation of embryogenic calli from the scutellar non-transformed cells (Table 3). Higher frequencies of bialaphos-resistant calli (16.7–33.5%) were obtained from all the other procedures, which presented the transient GUS expression mainly on the scutellum cells. These results confirmed that a successful genetic transformation requires not only efficient gene delivery, but also having tissue culture conditions allowing a good proliferation of transformed cells (Ishida et al. 1996; Negrotto et al. 2000; Amoah et al. 2001).

Table 3 Efficiency of designed A. tumefaciens-mediated maize transformation procedures
Full size table

Analysis of stable events and regeneration of bialaphos-resistant plants

Regarding the stable GUS expression, some bialaphos-resistant calli were randomly assayed from the total generated ones (Fig. 4a); all of them resulted positive (Fig. 4b). Since the gus gene in this construct contains an intron, blue staining was indicative of plants rather than A. tumefaciens expression of the transgene. The rest of the bialaphos-resistant calli were placed in regeneration medium amended with bialaphos under the conditions described in “Materials and methods” section; almost all of them developed green tissue containing a few leaves; some of them were assayed for stable GUS expression resulting positive as well (Fig. 4c). Frequencies of bialaphos-resistant calli in procedures using ultrasound-wounded embryos were higher than those using particle-wounded embryos (Table 3); nevertheless, the regenerative capability in B8V was higher than those using sonication. In the procedures using ultrasound-wounded embryos, a detrimental effect on the regenerative capability was observed when the sonication period was increased from 5 to 10 s (Table 3); this may be due to an increase in the magnitude of the damage in the target tissue, affecting proliferation of the transformed cells and consequently their regenerative capability. Thus, it is important to minimize the applied wound (Bidney et al. 1992; Brettschneider et al. 1997; Droste et al. 2000).

Fig. 4
figure 4

Fertile transgenic tropical maize plants generated by the A. tumefaciens-mediated transformation of the immature embryo-scutellum. (a) Putative transformed callus emerging from infected embryos plated on selection medium with 3 mg l−1 bialaphos. (b) Stable gus expression in bialaphos-resistant callus (left and upper); non-transformed callus without GUS expression (lower). (c) Stable GUS expression in excised leaf of a regenerated transformed plant (right); GUS staining was not detected in leaves of non-transformed plant (left). (d–e) Whole transgenic plants growing in the greenhouse at flowering stage. (f) Normal fertile ear emerged from transgenic tropical maize (left), which seems similar as the non-transformed control (right). (g) Test of the putative transformant for resistance to Basta™. Young leaves of non-transformed (upper, left) and putative transformed (lower, right) plants 2 weeks after a 2% Basta™ solution was applied over these

Full size image

A total of 96 independent bialaphos-resistant plants were regenerated from all five conditions; the majority (>90%) of them survived the transfer to soil, grew into normal and fertile plants, and formed a normal tassel and ear (Fig. 4d and e). In addition, the majority (about 80%) of them produced as many seeds as seed-derived control plants by self-pollination (Fig. 4f).

Molecular and progeny analyses of transgenic plants

Samples of genomic DNAs from soil-adapted plants, showing moderate to high resistance to Basta™ herbicide, were assayed to assess stable integration of the bar gene (Fig. 4g). Southern blot analysis confirmed that 16 out the 48 tested plants were transgenic, yielding various hybridizing bands (longer than 1.8 kbp). Digestions of genomic DNAs were carried out by using BamHI, which possessed a unique restriction site in the T-DNA (Fig. 1); thus, each hybridization band was estimated as one transgene insertion copy in the plant genome, which varied from approximately one to six for the evaluated clones (Fig. 5). The low number of integration loci showed here corresponds with the well-known insertion pattern arose by this natural transformation system (Ishida et al. 1996; Frame et al. 2002; Shou et al. 2004). No hybridization signal could be detected for the non-transformed control plant.

Fig. 5
figure 5

Southern-blot analysis of genomic DNAs from independent T 0 events of PPT-resistant maize plants. Extracted genomic DNAs were digested with BamHI and allowed to hybridize with a 700 bp PCR-digoxigenin labeled probe comprising a region from p35S to bar gene. (a) Transgenic plants generated by infected embryos particle-wounded at 60 lb/in2 (B60V−5, -9, and -10), or 80 lb/in2 (B80V-5, -9, -11, and -16); ctrl-1 and ctrl-2, putative transformed plants generated by the control treatment. (b) Transgenic plants generated by infected embryos wounded by applying sonication during 5 s (S5-2, -3, and -5), or 10 s (S10-5, -6, and -7). M, molecular weight marker; Wt, non-transformed control

Full size image

On the basis of the average stable transformation efficiency (number of molecular-confirmed plants recovered per total number of embryos infected), the best results were obtained in the procedures S5 (6.82%) and B80V (5.41%), while lower efficiencies were obtained in the procedures B60V (3.13%) and S10 (2.63%). None confirmed plants were raised when non-damaged embryos were used (Table 3). Thus, on the basis of the results showed here, transformation of the tropical maize line LPC13 by the A. tumefaciens-mediated transformation was only possible when wounded immature embryos were used as target tissue. Transformation efficiencies were similar to those previously reported with A. tumefaciens-mediated transformation using immature embryos of the laboratory maize line Hi-II (5.5% by Frame et al. 2002), or inbred temperate lines (2.4–5.3% by Huang and Wei 2005); similarly, evaluating some tissue culture media regimes in post-infection embryogenic callus induction and transformation, Frame et al. (2006) reported 2.8–8.0% of transformation efficiencies using temperate maize inbred lines. Additionally, efficiencies (2.0–4.0%) were reported by Brettschneider et al. (1997) when using a transformation protocol based in biolistic.

Progeny from some independent events generated from all the transformation procedures were analyzed by GUS histochemical or in planta Basta™ assays to determine goodness-of-fit to the expected 3:1 Mendelian ratio for self pollinations. It was demonstrated that the introduced genes were inherited as a single Mendelian locus in all events (Table 4).

Table 4 Progeny analysis of independent T 0 transformantsa
Full size table

Concluding remarks

A successfully transformation procedure was developed, in which transgenic tropical maize plants could be directly generated without preliminary crosses with laboratory and/or temperate lines. This was based in: (i) induction of somatic embryogenesis in the embryo-scutellum (target tissue), and (ii) optimization of the A. tumefaciens-mediated gene delivery toward this tissue by applying physical wounds prior the Agro-infection. Because the genotype-dependence is still an important issue involved in the Agrobacterium-maize transformation, the procedure reported within this manuscript could not be necessary successful in other tropical maize genotypes; nevertheless, some of the strategies reported here could be useful and get us an additional parameter to evaluate.

To our knowledge, this is the first report in which the Agrobacterium-mediated transformation technology is being extended toward the tropical-adapted maize; it opens the way to generate transgenic plants improved in various agronomical, nutritional, and nutraceutical traits in a simple and short-time way.