Plant Cell, Tissue and Organ Culture (PCTOC)

, Volume 109, Issue 2, pp 201–211

Evaluation of key factors influencing Agrobacterium-mediated transformation of somatic embryos of avocado (Persea americana Mill.)


  • Elena Palomo-Ríos
    • Departamento de Biología Vegetal, Instituto de Hortofruticultura Subtropical y Mediterránea “La Mayora” (IHSM-UMA-CSIC)Universidad de Málaga
  • Araceli Barceló-Muñoz
    • IFAPA Centro de Churriana
    • Departamento de Biología Vegetal, Instituto de Hortofruticultura Subtropical y Mediterránea “La Mayora” (IHSM-UMA-CSIC)Universidad de Málaga
  • Fernando Pliego-Alfaro
    • Departamento de Biología Vegetal, Instituto de Hortofruticultura Subtropical y Mediterránea “La Mayora” (IHSM-UMA-CSIC)Universidad de Málaga
Original Paper

DOI: 10.1007/s11240-011-0086-5

Cite this article as:
Palomo-Ríos, E., Barceló-Muñoz, A., Mercado, J.A. et al. Plant Cell Tiss Organ Cult (2012) 109: 201. doi:10.1007/s11240-011-0086-5


Key factors influencing the efficiency of transformation of embryogenic cultures, induced from immature zygotic embryos, of avocado cv. ‘Duke 7’ were evaluated. Initially, the sensitivity of somatic embryos to the antibiotics kanamycin, used for selection, carbenicillin, cefotaxime and timentin, all used for elimination of Agrobacterium cells, were evaluated. Isolated globular somatic embryos were more sensitive to kanamycin than embryogenic masses, and 25 mg l−1 kanamycin completely restricted callus proliferation. Cefotaxime at 500 mg l−1 partially inhibited proliferation of embryogenic cultures, while both carbenicillin and timentin did not affect callus growth. For genetic transformation, somatic embryos were infected with A. tumefaciens containing the pBINUbiGUSint plasmid. After 2 days, the embryos were transferred to selection medium supplemented with 50 mg l−1 kanamycin and 250 mg l−1 timentin for 2 months. Then, kanamycin level was increased to 100 mg l−1 for two additional months. The A. tumefaciens strain AGL1 yielded higher transformation rates, 6%, than EHA105 or LBA4404, 1.2%. The percentage of kanamycin resistant calli obtained was significantly influenced by the embryogenic line used as source of explants. Genetic transformation was confirmed by PCR and Southern blot analysis. A significant improvement in the germination rate was obtained when transgenic embryos were cultured in liquid MS medium with 4.44 μM BA and 2.89 μM GA3 for 3 days in a roller drum and later transferred to the same medium gelled with 7 g l−1 agar. Plants from five independent transgenic lines were acclimated and grown in the greenhouse, being phenotipically similar to control plants.


Genetic transformationSomatic embryogenesisTransgenic fruit trees





Gibberellic acid


Murashige and Skoog medium


Embryogenic avocado culture medium


Somatic embryo


Avocado (Persea americana Mill.) is a tropical tree cultivated for its fruits, which are a well-balanced source of nutrients and vitamins, but also a rich source of oil (Chanderbali et al. 2008). World avocado production reached 3.5 million tons in 2008 (FAOSTAT 2010), being Mexico, Chile and Indonesia the leading producing countries. The international trade in fresh avocado fruit is very important. In 2007, the estimated value of avocado exports was 1,460 million US$ (FAOSTAT 2010). This species is highly heterozygous, with a long juvenile phase and a high rate of flower abscission and immature fruit drop (Litz et al. 2005). Due to these limitations breeding programs have been relatively unsuccessful, and most important cultivars, e.g. ‘Hass’ and ‘Fuerte’, have been derived from open pollination and dooryard tree selection (Litz et al. 2007).

A major objective of avocado breeding programs is to develop improved rootstocks with enhanced resistance to fungal soil-borne pathogens. Phytophthora root rot, caused by the pathogen Phytophthora cinnamomi Rands., is one of the most important avocado diseases (Litz et al. 2007). There are several rootstocks, e.g. ‘Duke 7’, ‘Merensky I’, ‘Merensky II’ and ‘Thomas’, which are tolerant to this pathogen (Newett et al. 2002). White root rot caused by the ascomycete fungus Rosellinia necatrix is also an important pathogen in the Mediterranean area, especially Israel and Spain (Pliego et al. 2009). Genetic transformation could be a useful tool to enhance disease tolerance in this species. However, this technique is hampered by the recalcitrant nature of avocado explants to in vitro regeneration. Most protocols for avocado regeneration are based on embryogenic cultures derived from immature zygotic embryos (Sánchez-Romero et al. 2005; Litz et al. 2007). Although these cultures can be easily obtained, the conversion of somatic embryos occurs at very low rates.

Few attempts on avocado transformation have been reported so far. Cruz-Hernández et al. (1998) obtained transgenic embryogenic cultures via inoculation with Agrobacterium tumefaciens carrying a co-integrate vector containing the gus and nptII genes. After continuous selection in liquid medium, some transgenic calli were obtained, but transgenic plants could not be recovered. Using a similar protocol, Raharjo et al. (2008) obtained transgenic avocado plants, presumably from a single transformation event, expressing a plant defensin gene under the control of the CaMV35S promoter. Despite these successful reports, the characterization of some critical transformation parameters, such as explant type, the level of selective agent to kill non transformed cells, or Agrobacterium strain, are still deficient. In this investigation we have evaluated some of these parameters, and we describe a novel protocol that can be successfully used for obtaining transgenic avocado plants.

Materials and methods

Plant material and culture conditions

Embryogenic avocado (P. americana Mill.) cultures were established from immature zygotic embryos, cv. ‘Duke 7’, according to Pliego-Alfaro and Murashige (1988) on Murashige and Skoog (MS) medium (Murashige and Skoog 1962) supplemented with 0.41 μM picloram (MSP medium) and solidified with 6 g l−1 agar (Sigma A-1296). Cultures were incubated in the dark at 25 ± 1°C. Embryogenic cultures were maintained in this MSP medium and subcultured at monthly intervals. Three embryogenic lines, D2, D2.3 and D6, originated from different zygotic embryos, were used.

Transformation vector and Agrobacterium strains

The pBINUbiGUSInt (Humara et al. 1999) binary vector was used in transformation experiments. This plasmid harbours the neomycin phosphotransferase II (nptII) and the β-glucuronidase (uidA) genes under the control of the nopaline synthase promoter and the promoter of the ubi1 gene of maize polyubiquitin, respectively. The uidA gene contained the PIV2 intron of the ST-L1 gene from Solanumtuberosum within its coding sequence, preventing its expression in Agrobacterium.

Three A.tumefaciens strains were used in the transformation experiments: LBA4404 (Hoekema et al. 1983), EHA105 (Hood et al. 1986) and AGL1 (Lazo et al. 1991). The binary vector was introduced into the disarmed bacterial strains by the freeze–thaw method (Höfgen and Willmitzer 1988). For avocado transformation, Agrobacterium cultures were grown for 24 h at 28°C in LB medium at 250 rpm. Then, bacterial suspension was centrifuged at 4,000×g, the pellet washed with 10 mM MgSO4 and finally diluted in liquid MSP medium at 0.5 OD600nm.

Kanamycin selection experiments

Tolerance of avocado somatic embryos (SE) to the aminoglycoside antibiotic kanamycin was tested in different experiments. Firstly, embryogenic masses, 0.5 g per plate, were cultured on Petri dishes containing solid MSP medium supplemented with kanamycin at 0, 50, 100, 150 and 200 mg l−1. These embryogenic masses consisted on pieces of embryogenic callus with somatic embryos at different developmental stages. Embryogenic calli were recultured onto fresh medium every 2 weeks, and the callus weight recorded after 4 weeks of culture. Five plates per treatment were employed. Secondly, the effect of kanamycin on the proliferation of isolated SE, instead of embryogenic masses, was tested. Groups of three globular SE, 1–2 mm size, were cultured in MSP solid medium supplemented with 0, 6, 12.5, 25 and 50 mg l−1 kanamycin. Embryogenic masses derived from the proliferation of isolated SE were recultured every month onto fresh medium, and the callus weight as well as the percentage of dead explants, were recorded after two subcultures. Six Petri dishes with ten groups of three SE each one per treatment were used.

Effect of bactericidal antibiotics on the growth of embryogenic cultures

The effect of the antibiotics carbenicillin, cefotaxime and timentin, used to eliminate Agrobacterium, on the growth of isolated SE was evaluated. Globular SE, 1–2 mm size, were cultured in groups of three embryos for 4 weeks on solid embryogenic medium supplemented with carbenicillin (500 mg l−1), cefotaxime (500 mg l−1) or timentin (250 mg l−1). These concentrations were found optimal for restricting bacterial growth on Agrobacterium-inoculated SE in previous experiments. The embryos were recultured every month onto fresh medium and the callus fresh weight was recorded after two subcultures. Seven dishes with ten groups of three SE each per treatment were used. In all experiments, antibiotics were filter-sterilised and added to the cooled media after autoclaving.

Transformation experiments

Based on the results of the experiments indicated above, as well as on the protocol previously developed for transformation of olive embryogenic cells (Torreblanca et al. 2010), an initial procedure, for avocado transformation, was established as follows. Small SE at the globular stage, 1–2 mm size, were selected by hand from embryogenic culture stocks at 3–4 weeks after subculture. The embryos were immersed for 20 min in the diluted Agrobacterium suspension, strain AGL1, under mild agitation. Then, explants were blotted dry with sterile filter paper and cultured in solid MSP medium for 48 h. Afterwards, explants were washed for 2 h with liquid MSP medium supplemented with 300 mg l−1 timentin, blotted dry and transferred in groups of three SE to selection medium, i.e. solid MSP medium supplemented with 250 mg l−1 timentin and 50 mg l−1 kanamycin. Explants were recultured onto fresh selection medium bi-weekly during the first month and every month thereafter. After 2–3 months of culture in the selection medium, some SE showed secondary embryogenesis and formed a yellow callus. Each kanamycin proliferating mass was considered as an independent putative transformed line. These calli were grown for at least two additional months onto selection medium supplemented with 100 mg l−1 kanamycin, before transferring to maturation medium without antibiotics for SE development.

Several transformation parameters were evaluated using the above described procedure. In the first experiment, the transformation efficiency of three different Agrobacterium strains, LBA4404, EHA105 or AGL1, harboring pBINUbiGUSInt plasmid, was evaluated, using SE from the D2 embryogenic line. Then, the effects of the genotype and kanamycin selection level were tested. In this experiment, SE from the embryogenic lines D2, D2.3 and D6 were inoculated with A. tumefaciens AGL1 strain and cultured in the selection medium supplemented with 25 or 50 mg l−1 kanamycin for 2 months. Afterwards, kanamycin resistant lines were transferred to selection medium with 100 mg l−1 kanamycin. In all experiments, the transformation efficiency was estimated as the number of calli actively growing in the presence of 100 mg l−1 kanamycin divided by the number of groups of three SE inoculated. All transformation experiments were made by duplicate with ten Petri dishes, containing ten groups of three SE, per treatment.

GUS activity was assessed in putative transgenic SE at different developmental stages by the histological GUS assay (Jefferson 1987). Additionally, as a proof of transformation, some kanamycin resistant calli were challenged to grow in liquid MSP medium supplemented with high kanamycin concentrations. To this purpose, 0.4 g of control and transgenic calli (TC) were grown on 250-ml culture flasks containing 40 ml of liquid MSP medium supplemented with kanamycin at 50, 100, 200 and 400 mg l−1. Cultures were incubated in the dark in an orbital shaker at 120 rpm, and the callus fresh weight was recorded after 1 month of culture.

Transgenic plant recovery

Transformed embryogenic masses from seven D2.3 and three D2 independent transgenic lines were cultured for 5 weeks onto maturation medium, MS medium with macronutrients of Gamborg et al. (1968) and solidified with 10 g l−1 agar (Márquez-Martín et al. 2011), for embryo development. Resulting embryos were recultured on the same medium for an additional 5 weeks period. Afterwards, mature, white-opaque SE larger than 0.4 cm were selected and cultured in MS medium supplemented with 4.44 μM 6-benzyladenine (BA) and 2.89 μM gibberellic acid (GA3) for embryo germination (Witjaksono and Litz 1999a). To improve the germination rate, the effect of a pre-treatment of SE in liquid germination medium was evaluated. In this experiment, mature SE were transferred to test tubes containing 5 ml liquid germination medium for 3 days, and cultured in a roller drum, under 40 μmol m−2 s−1 irradiance level. Afterwards, SE were transferred onto solid germination medium for a month. This pre-treatment in liquid medium and the subsequent transfer to solid medium was repeated 3 times.

Some of the germinated SE gave rise to shoots that reached 5–10 mm in size. These shoots could be excised and multiplied in Gamborg solid medium supplemented with 0.3 mg l−1 BA (Barceló-Muñoz et al. 1999). Fifteen mm long shoots from the proliferation medium could be rooted following a 3-days exposure to liquid MS medium with macroelements at 0.3× and supplemented with 1 mg l−1 indolebutyric acid (Barceló-Muñoz et al. 1999). Shoots from sprouted embryos which failed to elongate had to be recovered by micrografting onto in vitro grown Topa-Topa seedlings, following the procedure of Pliego-Alfaro and Murashige (1987). After 6–8 weeks, shoots excised from sprouted grafts, could be rooted following the same procedure described above (Barceló-Muñoz et al. 1999). Rooted plantlets were transplanted to trays containing a mix of peat, coconut fiber and perlite and acclimated to ex vitro conditions following the procedure of Barceló-Muñoz et al. (1999).

Molecular analysis of transgenic material

The transgenic nature of avocado plants was confirmed by both polymerase chain reaction (PCR) amplification and Southern blot analysis. Genomic DNA was isolated from embryogenic calli and leaves of control and putative transgenic lines using QUIAGEN DNeasy® Plant Mini Kit. Aliquots of these DNA extracts were used to amplify by PCR a 220 bp fragment belonging to the nptII gene (Youssef et al. 2009). For Southern blot analysis, 5–10 μg of DNA isolated from embryogenic calli were digested overnight with BglII, fractionated in a 0.8% agarose gel and transferred to a Hybond N+ membrane. The filter was hybridized at 64°C with a 700 bp digoxigenin labelled probe obtained by PCR amplification of the nptII gene from pBINUbiGUSInt plasmid (Álvarez et al. 2004). Primers used in PCR were 5′-CGCAGGTTCTCCGGCCGCTTGGGTG-3′ and 5′-AGCAGCCAGTCCTTCCGCTTCAG-3′ for the amplification of 220 bp nptII fragment, and 5′-GAGGCTATTCGGCTATGACTG-3′ and 5′-ATCGGGAGCGGCGATACCGTA-3′ for 700 bp nptII fragment.

Statistical analysis

Data were subjected to analysis of variance using SPSS software. Tests for normality and homogeneity of variance were performed prior to ANOVA and the Tukey test was used for mean separation. Frequency analyses were performed with the G-test of independence, using BIOMstat software (Sokal and Rohlf 1995).


Effect of antibiotics in the growth of avocado embryogenic cultures

In the first series of experiments, the effect of the addition of the antibiotic kanamycin for selection of transgenic material on the growth of avocado SE, line D2, was tested. Explants were cultured as masses of embryogenic callus or as groups of three isolated SE, 1–2 mm diameter, at the globular stage. When using embryogenic masses, a 62% reduction in the callus fresh weight was observed at the 50 mg l−1 kanamycin treatment when compared with the control after 4 weeks of culture (Fig. 1a). At higher kanamycin concentrations, the increments on callus fresh weight were negligible (Fig. 1a). Preliminary experiments showed that isolated embryos were more susceptible to kanamycin than embryogenic masses. Thus, in this case, the range of kanamycin concentrations tested was lower than the one used with embryogenic masses. Kanamycin at 6 mg l−1 reduced the growth of SE by 71% (Fig. 1b). At higher antibiotic concentrations, callus growth was totally impaired. Kanamycin notably reduced embryo viability at 6 mg l−1, and the percentage of necrotic explants was close to 100% at the 25 mg l−1 kanamycin concentration. Figure 2a shows the aspect of isolated SE from line D2 after a month of culture in the presence of the different kanamycin concentrations. These experiments were also performed with other avocado embryogenic lines, derived from different zygotic embryos, obtaining similar results (data not shown).
Fig. 1

Effect of kanamycin on the growth of avocado embryogenic callus cultured as embryogenic masses (a) or groups of three isolated globular embryos (b). Data correspond to mean ± SD of six replicates, after 1 month of culture
Fig. 2

a Proliferation of avocado somatic embryos after a month of culture in solid basal medium with increased kanamycin concentrations (0–50 mg l−1). Initial explants were groups of three isolated globular somatic embryos. b Aspect of Agrobacterium-inoculated avocado cultures after 2 months of selection in 50 mg l−1 kanamycin. Some explants showed a good proliferation rate in this selection medium while the remainder explants appeared necrotic. c Histological GUS assay in control, non-inoculated, (left) and kanamycin resistant (right) avocado calli. Bar corresponds to 1 mm

Besides the selective agent, selection media for Agrobacterium inoculated explants must be supplemented with a bactericidal antibiotic at high concentrations for eliminating Agrobacterium cells. When avocado SE were inoculated with A. tumefaciens, strains AGL1, EHA105 and LBA4404, timentin at 250 mg l−1 and cefotaxime at 500 mg l−1 were the most effective agents controlling overgrowth of bacterial cells, whereas carbenicillin at 500 mg l−1 did not totally restrict bacterial growth (results not shown). The effect of these bactericidal antibiotics on the growth of avocado somatic embryos was tested. In this experiment, isolated SE were chosen as initial explant, since kanamycin experiments had showed that this kind of explant was more suitable for selection experiments than embryogenic masses. As it can be observed in Fig. 3, fresh weight of embryogenic callus derived from the growth of isolated SE was significantly decreased when the medium was supplemented with cefotaxime at 500 mg l−1. Carbenicillin at 500 mg l−1 also exerted a slight inhibitory effect during the first month of culture, but callus growth recovered during the second month. Finally, timentin at 250 mg l−1 did not affect the growth of avocado calli. Thus, timentin at 250 mg l−1 was chosen to control bacterial growth on further transformation experiments.
Fig. 3

Effect of bactericidal antibiotics on the growth of avocado somatic embryos. Globular somatic embryos were cultured as group of three embryos in media supplemented with carbenicillin or cefotaxime at 500 mg l−1 or timentin at 250 mg l−1, and the fresh weight was recorded after 1 and 2 months of culture. Mean separation at one (lower letters) and 2 month samplings (capital letters) were performed by Tukey test at P = 0.05

Development of a transformation protocol

In previous experiments, we were unable to obtain any transgenic avocado callus using the procedure described by Cruz-Hernández et al. (1998). Thus, we performed a series of experiments to develop a novel transformation protocol. Initially, the use of isolated globular embryos of small size, 1–2 mm, instead of proembryogenic masses, as explants for Agrobacterium inoculation was evaluated. Twelve out of 200 inoculated explants showed a good proliferation rate in the selection medium after 2 months of culture while the remainder explants appeared to be necrotic (Fig. 2b). All these putative transgenic calli proliferated well when transferred to selection medium supplemented with 100 mg l−1 kanamycin. Samples of these kanamycin resistant calli were subjected to the histochemical GUS assay, showing all of them a strong GUS signal (Fig. 2c). The final transformation rate based on the number of independent calli actively growing in the presence of kanamycin at 100 mg l−1 was 6%. As an evidence of transformation, two independent transgenic calli, TC4.1 and TC5.7, were cultured in liquid medium at high kanamycin concentrations during a month. Kanamycin exerted a strong inhibitory effect on the growth of control, non-transformed embryogenic callus cultured in liquid medium, and the growth of this callus was totally inhibited at the lowest kanamycin concentration, 50 mg l−1 (Fig. 4). By contrast, both transgenic lines only showed a partial reduction in their growth rate at kanamycin concentrations in the range of 50–200 mg l−1. Transgenic line TC5.7 showed the highest tolerance to the antibiotic, being able to grow at 400 mg l−1 kanamycin.
Fig. 4

Fresh weight increments in control and transgenic avocado embryogenic calli after 1 month of culture in liquid medium supplemented with different kanamycin concentrations. Two independent transgenic calli lines, TC 4.1 and TC 5.7, were analyzed. Data represent means ± SD of five replicates

In a different experiment, the effect of the Agrobacterium strain on the transformation efficiency was evaluated, following the transformation protocol above described. AGL1 yielded the highest transformation rate, 6.0 ± 0.02%, mean ± SE, n = 3, whereas only a 1.2 ± 0.01% of putative transgenic calli were recovered with the use of EHA105 and LBA4404 strains. Finally, to test the effects of the genotype and kanamycin selection level in the transformation efficiency, three independent embryogenic lines (D2, D2.3 and D6), derived from different zygotic embryos, were inoculated with the AGL1 A. tumefaciens strain and cultured in selection media supplemented with kanamycin at 25 or 50 mg l−1. In general, transformation rates were slightly lower when explants were initially subjected to 50 mg l−1 kanamycin, especially in the case of D2.3 and D6 lines. However, the differences between kanamycin selection levels were not statistically significant. By contrast, the transformation rate was strongly influenced by the embryogenic line used as source of explants. Embryogenic line D2 yielded a mean transformation rate significantly higher than those obtained with the other two lines (Fig. 5). Major steps of the developed transformation protocol are shown in Fig. 6.
Fig. 5

Effect of the embryogenic line used as source of explants on the transformation efficiency. Somatic embryos from three independent embryogenic lines, derived from three zygotic embryos, were inoculated with A. tumefaciens AGL1 strain and the percentages of kanamycin resistant calli were recorded after 4 months of culture on selection medium. Data represent means ± SE. Mean separation was performed by G-test of independence at P = 0.05
Fig. 6

Schematic diagram showing the major steps of the protocol used to obtain avocado transgenic plants

Genomic DNA was extracted from control and putative transgenic calli, and the transgenic nature of this material was confirmed by both PCR and Southern blot analysis. Figure 7a shows the PCR amplification of a 220 bp fragment from the nptII gene. No amplification was detected in control DNA, whereas all transgenic lines analyzed showed a positive signal. Using primers from the VirG gene from Agrobacteriumrhizogenes it was also confirmed that transgenic DNA was not contaminated with bacterial DNA (results not shown). For Southern blot analysis, genomic DNA was digested with BglII which recognizes one site inside the T-DNA of the pBINUbiGUSInt plasmid. The membrane was hybridized with a 700 bp fragment from the nptII gene. No hybridization signal was detected in control DNA, whereas the 5 independent lines analyzed showed several hybridization bands, ranging the number of transgene copies inserted between 1 and 4 (Fig. 7b).
Fig. 7

a PCR amplification of nptII gene in DNA isolated from control (c) and transgenic embryogenic calli (TL). Each lane corresponds to an independent transgenic line. P plasmid pBINUbiGUSInt. MW molecular weight marker Lambda DNA/HindIII. Arrows correspond to molecular weight fragments of 6.5, 4.3, 2.3, 2.0 and 0.5 kb, from top to bottom. b Southern blot analysis of DNA from control (c) and several independent embryogenic lines (T4–T16). DNA was digested with BglII, and the filter was hybridized with a 700 bp probe obtained by PCR amplification of the nptII gene from pBINUbiGUSInt plasmid. MW molecular weight marker Lambda DNA/HindIII

Transgenic plantlet recovery

Somatic embryos from three D2 and seven D2.3 independent transgenic lines were matured following the protocol previously described by Márquez-Martín et al. (2011). White-opaque mature SE were obtained after 10 weeks of culture with a mean value of 5.2 ± 0.6 mature SE per 100 mg of embryogenic calli. Transgenic mature SE showed a strong blue coloration when assayed with the histological GUS assay (Fig. 8a). Mature SE larger than 0.4 cm length were germinated in solid medium as described by Witjaksono and Litz (1999a). However, the percentage of SE that developed shoots was extremely low, c. 0.01%, in one of the D2.3 transgenic lines, and no germination was observed in the other D2.3 and D2 lines. To improve the germination rate, a pre-treatment consisting of culturing white opaque SE in liquid germination medium for 3 days and subsequent transferring of the SE to solid medium for a month was tested. This treatment was repeated three times. Using this improved germination protocol, a germination rate in the range 0.5–2% was observed in five of the D2.3 transgenic lines, while no germination was observed in embryos of the two other D2.3 and the three D2 transgenic lines.
Fig. 8

a Histological GUS assay in control (left) and transgenic (right) mature somatic embryos. Bar corresponds to 5 mm. b Transgenic plant growing in Gamborg medium. c PCR amplification of nptII gene in DNA isolated from control (c) and transgenic leaves (TP). Each lane represent an independent transgenic plant. P corresponds with plasmid pBINUbiGUSInt. d Histological GUS assay in control (upper) and transgenic (lower) leaves. e Transgenic avocado plant growing in the greenhouse. Pictures correspond to material derived from different D2.3 independent avocado transgenic lines

Shoots over 5 mm in length from germinated embryos could be multiplied and rooted according to the procedure of Barceló-Muñoz et al. (1999). Sprouted shoots from germinated embryos which failed to elongate, ca. ≤ 5 mm, could be recovered through micrografting using the procedure of Pliego-Alfaro and Murashige (1987). After 6–8 weeks, shoots from sprouted grafts could be excised and rooted using the same procedure described for ungrafted shoots (Barceló-Muñoz et al. 1999) Using both protocols, we could recover plants from the five independent transgenic D2.3 lines where shoot sprouting had been observed (Fig. 8b). Genomic DNA was isolated from leaves and the transgenic nature of these plants was verified by PCR amplification of nptII gene (Fig. 8c). Leaves from these plants showed GUS activity when assayed with the histochemical GUS assay (Fig. 8d). Rooted plants which survived the acclimation phase, to ex vitro conditions, are being currently grown in the greenhouse (Fig. 8e).


Effects of antibiotics on the growth of avocado somatic embryos

In vitro regeneration of avocado plants is extremely difficult, and this recalcitrant character has hindered the development of efficient protocols for its genetic transformation. As an initial step to develop a reliable transformation protocol for this species, the effect of antibiotics, commonly used in the Agrobacterium-mediated transformation system, in the growth of embryogenic cultures, i.e. kanamycin used as selective agent, and β-lactams used to eliminate Agrobacterium, were evaluated. In vitro tolerance of fruit trees explants to antibiotics used as selective agents varies widely (Srinivasan and Scorza 1999). In avocado, kanamycin tolerance seems to depend on the genotype and the type of explant used, e.g., in cv. ‘Hass’, the growth of SE was only restricted to 50% at 100 mg l−1 kanamycin when cultured in semi-solid medium (Cruz-Hernández et al. 1998) while the growth of ‘Duke 7’ SE was almost totally restricted at this antibiotic concentration. The level of kanamycin tolerance of avocado embryogenic cultures was drastically reduced when isolated globular SE were cultured instead of embryogenic masses, and 12.5 mg l−1 kanamycin was enough to inhibit growth. A similar differential sensitivity of isolated embryos versus embryogenic masses to antibiotics has been observed in olive (Pérez-Barranco et al. 2009). Zhang et al. (2001) found that kanamycin sensitivity in cotton somatic embryos varied during embryo development, being mature embryos the most tolerant. In our case, both kind of explants, embryogenic masses and isolated SE, were at similar developmental stage. Then, the superior behavior of kanamycin restricting the growth of isolated globular embryos should be likely due to a better diffusion of the antibiotic into the explant cells.

Carbenicillin, cefotaxime and timentin, in the range 250–500 mg l−1, are the most frequently used antibiotics to eliminate Agrobacterium cells after explant infection. These β-lactams inhibit bacterial cell wall synthesis, without being toxic to plant cell. However, there are numerous reports showing an effect, positive or negative, of these bactericidal antibiotics in the regeneration process (Padilla and Burgos 2010). It has been suggested that these antibiotics or part of them have growth regulator-like activity (Holford and Newbury 1992; ur Rahman et al. 2004; Shehata et al. 2010). In this investigation, we found that cefotaxime at 500 mg l−1, the concentration needed to avoid Agrobacterium contamination in avocado inoculated SE, reduced the growth of embryogenic cultures, whereas 500 mg l−1 carbenicillin and 250 mg l−1 timentin did not affect embryogenesis. It has also been shown that cefotaxime at 500 mg l−1 reduced shoot regeneration in tobacco (Nauerby et al. 1997) and strawberry (Husaini 2010), and, at higher concentrations, the production of SE in walnut (Tang et al. 2000). Cefotaxime also partially inhibited the development of mature SE in Sitka spruce (Sarma et al. 1995). As observed in walnut SE (Tang et al. 2000), timentin at 250 mg l−1 was the most effective suppressing A. tumefaciens proliferation in infected avocado SE. Taking into account these results, this antibiotic was selected for transformation experiments.

Development of a transformation protocol

An early report by Cruz-Hernández et al. (1998) described the obtainment of transgenic avocado embryogenic calli via inoculation with A. tumefaciens. Although transgenic calli showed a high tolerance to kanamycin, used as selective agent, transgenic plants could not be recovered. In the protocol developed by Cruz-Hernández et al. (1998), proembryonic masses were wounded with a camel hair brush, inoculated with a diluted A. tumefaciens culture in liquid medium and co-cultured for 3 days. Then, proembryonic masses were continuously cultured in liquid medium in the presence of kanamycin for transgenic selection. More recently, Raharjo et al. (2008) used the protocol described by Cruz-Hernández et al. (1998) with some modifications to obtain transgenic plants containing a defensin gene for fungal resistance. Basically, these modifications consisted in the use of phosphinotricin to select transgenic material instead of kanamycin and the application of a 2 weeks period without selection after co-culturing embryogenic masses with Agrobacterium. Despite this successful report, the use of this transformation protocol entails some concerns. It is well known that prolonged culturing in liquid medium induces the degeneration of embryogenic cultures and increases the risk of somaclonal variations (Etienne and Bertrand 2003; Von Arnold 2008). Along this line, Witjaksono and Litz (1999b) found that the time required for disorganization of avocado proembryonic masses or globular embryos cultured as suspension, with the consequent loss of embryogenic potential, could be as short as 3 months for some genotypes. Additionally, the selection in liquid phase makes it difficult the identification of independent transformation events within a given flask and it does not allow the calculation of a transformation rate which could be useful for treatments comparison in experiments devoted to improve the transformation efficiency. In fact, neither Cruz-Hernández et al. (1998) nor Raharjo et al. (2008) reported the efficiencies of their transformation experiments. Moreover, all the transgenic plants recovered by Raharjo et al. (2008) showed the same pattern of bands when DNA was subjected to Southern blot analysis, leading these authors to suggest that all of them belonged to the same transformation event. In our case, the protocol described by Cruz-Hernández et al. (1998) was unsuccessful. This was probably due to the sensitivity of embryogenic cells of the ‘Duke 7’ genotype to the extreme conditions, indicated in this protocol, for selection of transgenic cells, ca. a high kanamycin concentration (100 mg l−1) and a liquid selection medium.

In this research, a novel avocado transformation protocol that avoids the use of liquid medium and allows the easy isolation of independent transgenic lines has been developed. This protocol is based on those described by Álvarez et al. (2004) and Torreblanca et al. (2010) for the transformation of cork oak and olive SE, respectively. Main features of this protocol are the use of globular SE as explants for Agrobacterium inoculation, the selection of transgenic cells in solid medium and the employment of a hypervirulent A. tumefaciens strain, AGL1. This protocol yields transformation rates in the range of 1–6%. Agrobacterium strain and the embryogenic line used as source of explants exerted a great influence in the transformation efficiency. The number of kanamycin resistant calli was significantly higher when using the hypervirulent strain AGL1, despite AGL1 shares the same Agrobacterium chromosome, C58, and Ti plasmid, pTiBo542ΔT-DNA, than EHA105 (Hellens et al. 2000). The AGL1 strain has also been proven to be more efficient than other A. tumefaciens strains in the transformation of other species (Weir et al. 2001; Álvarez et al. 2004; Bartlett et al. 2008; Li et al. 2010; Torreblanca et al. 2010; Zhao et al. 2011). Due to the hypervirulent character of this Agrobacterium strain, a two-day coculturing period was chosen to avoid bacterial overgrowth, despite some authors found optimal an extended co-culture phase (Yang et al. 2010; Parimalan et al. 2011). Regarding the source of explants for transformation, significant differences were found in the transformation rate among the three embryogenic lines tested, although transgenic material could be successfully recovered from all the lines. This result suggests that the genetic background of the explants has a critical influence in the transformation process. Following the protocol above described, transgenic callus lines corresponding to different transformation events have been recovered, as it was shown by the different pattern of nptII hybridization bands obtained in the Southern blot analysis of genomic DNA.

Somatic embryos from several transgenic lines were matured according to Márquez-Martín et al. (2011) and germinated following the procedure of Witjaksono and Litz (1999a). Germination of mature avocado SE is generally low and very dependent on the genotype (Sánchez-Romero et al. 2005; Litz et al. 2007). According to Pliego-Alfaro and Murashige (1988), the failure to develop shoots is likely due to the disorganization of the apical meristem. Additionally, embryo conversion in transgenic material could be more difficult as result of the transformation procedure, since Raharjo et al. (2008) reported a 0.0016% shoot recovery from transgenic avocado SE. In ‘Duke 7’ transgenic SE, the development of shoots was also sporadic when using the procedure of Witjaksono and Litz (1999a). However, shoot emergence was significantly improved when SE were pretreated in liquid germination medium. As cytokinins are important for the organization of the apical meristem (Gordon et al. 2007), it is possible that the liquid treatment enhances the effect of the BA present in the germination medium. Cytokinins have also been shown to improve embryo conversion in other species (Ceasar and Ignacimuthu 2010; Chen et al. 2010). Despite this improvement, plants from two D2.3 and three D2 transgenic lines could not be recovered, although embryos showed a normal aspect. Lack of response on the D2 lines could be an age-genotype effect, e.g. conversion rate of control, non transgenic, D2 somatic embryos has been observed to drastically decrease with time in culture. Shoots larger than 5 mm, from germinated embryos, could be successfully micropropagated using the protocol of Barceló-Muñoz et al. (1999). However, transgenic shoots which failed to elongate could be recovered by using the micrografting technique of Pliego-Alfaro and Murashige (1987). Raharjo et al. (2008) also used a micrografting procedure for rescuing transgenic avocado plants. Finally, several plants from 5 independent transformation events of embryogenic line D2.3 could be recovered. These plants showed GUS activity in the leaves and were phenotipically similar to control plants.

In conclusion, an efficient protocol for the transformation of avocado plants using embryogenic cultures has been developed. Transformation efficiencies in the range of 1–6%, based on the number of kanamycin resistant calli, are obtained. Embryo conversion has been improved following a liquid medium pretreatment of the mature white opaque transgenic embryos. This protocol is currently being used to obtain transgenic plants expressing antifungal proteins to enhance tolerance to R. necatrix.


This research was funded by Ministerio de Ciencia e Innovación of Spain and Feder European Union Funds (Grant No. AGL2008-05453-C02-01/AGR). The authors thank Dr. Ricardo J. Ordás, Universidad de Oviedo, Spain, for providing the AGL1 strain with the pBINUbiGUSint plasmid, and Dr. Clara Pliego for her valuable support in the molecular analysis.

Copyright information

© Springer Science+Business Media B.V. 2011