A short-length single chimeric transgene induces simultaneous silencing of Agrobacterium tumefaciens oncogenes and resistance to crown gall


E-mail: nalbur@cebas.csic.es


Transformation with self-complementary oncogene sequences was used to silence the Agrobacterium tumefaciens oncogenes ipt and iaaM. The silencing response was triggered by using a very short chimeric sequence where conserved fragments from both oncogenes were fused in one unique transgene. Most T0 transgenic tobacco lines and T1 seedlings evaluated in vitro had intermediate or very low susceptibility to A. tumefaciens as compared with the wildtype plants. A greenhouse evaluation of whole plants confirmed the lines that were resistant. Low levels of transgene hairpin RNA (hpRNA) coupled with small interfering RNA (siRNA) accumulation correlated with oncogene silencing and, therefore, resistance to crown gall. After infection with the oncogenic strain, much lower levels of the oncogenes’ mRNA were found in resistant lines than in wildtype plants. The frequency of resistant lines, with few or no symptoms, produced with the chimeric construct was similar to the highest reported efficiencies obtained by using sense and antisense whole oncogene sequences.


Agrobacterium tumefaciens is a plant pathogen able to infect plants through wounds. Bacterial infection produces tumours in the plant known as crown galls, a disease that affects many perennial fruit, nut and ornamental crops, causing large annual losses to growers and nurseries, all around the world (De Cleene & De Ley, 1976; Alekseeva et al., 2008). Tumours in the root collar and even in aerial parts of the plant may reduce plant productivity and increase susceptibility to pathogens and environmental stresses.

Traditional breeding strategies to produce resistant plants have been successful only in those species where resistant germplasm exists (Moriya et al., 2010) and, even then, many decades are necessary to obtain resistant rootstocks.

Different methods of crown gall control have been tested, including biological control using the non-pathogenic strain K84 of Agrobacterium rhizogenes (Rhouma et al., 2008), Agrobacterium vitis VAR03-1 (Kawaguchi et al., 2008), rhizobacteria BSCH14 and O33 (Rhouma et al., 2008), or Bacillus subtilis (Gupta & Khosla, 2007). Additionally, solarization has been used in combination with biological antagonists (Gupta & Khosla, 2007) or acetyl salicylic acid treatments (Anand et al., 2008) to reduce tumour formation. The disadvantage of all these treatments is that, even in the best case, complete resistance was not conferred.

Infection of plant cells by A. tumefaciens involves transfer of T-DNA from the bacterium to the plant cell nucleus and its integration into the plant genome (Gelvin, 2000; Pacurar et al., 2011). Within the T-DNA, two types of genes are found: those responsible for the synthesis of opines and oncogenes. Opines are amino-acid-derived molecules that bacteria can use as a unique source of carbon and nitrogen. Expression of oncogenes is responsible for the overproduction of cytokinins and auxins that results in the initiation of uncontrolled cell division and growth and the formation of a tumour (Zhu et al., 2000). These oncogenes include ipt, whose product catalyses the condensation of AMP and isopentenyl pyrophosphate to form zeatin, and iaaM, which encodes a tryptophan monooxygenase that converts tryptophan into the auxin precursor indole-3-acetamide. This is then converted to indole-3-acetic acid by the third oncogene (iaaH) product, indole-3-acetamide hydrolase (Binns & Costantino, 1998). Inactivation of the expression of ipt and of either of the other two genes involved in auxin biosynthesis prevents tumour formation (Zhu et al., 2000). However, it has been shown that the 6b T-DNA is an oncogene and can cause tumours by itself, so even if ipt and iaaM expression is abolished, some strains may still induce tumours, although this is dependent on Agrobacterium strain and plant species (Hooykaas et al., 1988).

Biotechnological approaches, based on genetically engineering plants with different genes, have been used recently to control crown gall disease. Transformation with complete or truncated A. tumefaciens proteins involved in T-DNA transfer has been shown to be a possible strategy for conferring crown gall resistance. Arabidopsis thaliana plants transformed with the VirD2 gene were less susceptible to subsequent A. tumefaciens infection than untransformed ones. However, constitutive over-expression of VirD2 was toxic to Arabidopsis plants and they were able to survive only if they had a very low (undetectable) expression level or if a wound-inducible promoter was used (Hwang et al., 2006).

Recently, transgenic grape plants transformed with a truncated VirE2 gene showed increased resistance to crown gall without any deleterious effect of transgene over-expression (Krastanova et al., 2010).

Other approaches for the production of plants resistant to crown gall have focused on the silencing of bacterial genes involved in tumour production by using sense-strand RNAs containing premature stop codons, provided that translation can be initiated from a start codon located downstream of the premature stop (Lee et al., 2003). Transgenic apples resistant to crown gall have been obtained using biotechnological strategies that produce sense and antisense strands of part of the oncogenes’ sequence and place these sequences between opposing, strong, constitutive promoters (Viss et al., 2003).

However, probably the most efficient strategies are those that use sense and antisense sequences of ipt and iaaM, split by an intron, that produce hairpin RNA (hpRNA) transcripts. Several pieces of work have demonstrated a high efficiency in the production of virus-resistant plants through the use of self-complementary viral sequences that are transcribed as hpRNAs (Waterhouse et al., 1998; Smith et al., 2000; Wesley et al., 2001). RNA-mediated silencing is a process based on sequence homology that is triggered by double-strand RNA (dsRNA) and leads to the suppression of gene expression. The enzyme Dicer processes dsRNA into small fragments of 21–25 nucleotides. These small interfering RNAs (siRNAs) are then incorporated into an ‘RNA induced silencing complex’ (RISC), which is then directed to degrade specific single-strand RNA sequences. Specificity is provided by the small RNAs (Mansoor et al., 2006). RNA silencing is an area of intense research that is producing new discoveries in the control of gene expression, plant development and defence against pathogens. One application of this research has been the genetic modification of plants by specifically silencing target genes or their promoters. This is accomplished through cloning sense and antisense sequences of the target gene, with an intron separating the two sequences, which are under the control of the same promoter. After transcription, these sequences produce hpRNA that triggers the silencing of the gene (Smith et al., 2000). The silencing efficiencies of hpRNA and antisense RNA have been compared in several plant species: the hp strategy increases gene silencing to 90–100% of transgenic plants (Wesley et al., 2001). A high percentage of plants transformed with these constructs, where genes are silenced, are immune to a late viral infection (Smith et al., 2000). The induction of hpRNA-mediated, post-transcriptional silencing is, currently, the system most widely used to silence genes in plants (Mansoor et al., 2006).

Transformation with self-complementary oncogene sequences has been used in Arabidopsis and tomato (Escobar et al., 2001), as well as in embryogenic lines from seed cotyledons of walnut cv. Chandler (Escobar et al., 2002), to silence the A. tumefaciens oncogenes ipt and iaaM. Plants transformed with sense and antisense sequences from ipt and iaaM, split by one intron, produced hpRNA and were highly resistant to crown gall. The high level of cross homology of oncogenes among different A. tumefaciens strains (around 90% identity in all sequenced genes) suggests that an ipt and iaaM silencing strategy may confer a wide spectrum of resistance (Escobar et al., 2001).

Escobar et al. (2001, 2002) introduced a large-size T-DNA, because most of the complete sequences from both genes were cloned in sense and antisense, split by an intron, each one with its promoter and terminator sequences. The objective of the present work, however, was to demonstrate the possibility of using a short-length chimeric transgene to silence both A. tumefaciens oncogenes simultaneously, by using conserved gene sequences. The strategy to design the transgene followed previously published methodology (Bucher et al., 2006), which produced tomato plants resistant to four viruses by silencing a different gene in each of the four tomato viruses with a single sequence.

Materials and methods

Plasmid construction and plant transformation

Sequences from different Agrobacterium tumefaciens strains (representing octopine-, nopaline- and succinamopine-producing types) in the GenBank database for ipt (AF242881, U83987, U83986, NC_002377, X77327, X14410, AB032122, AB025109, X17428) and iaaM (AF242881, U83987, NC_003065, NC_002377, AF126446, AB025110) oncogenes, were aligned and compared. Conserved sequences were identified approximately 125 bp after the start codon in the ipt gene and 1220–1270 bp for the iaaM gene, depending on the A. tumefaciens strain. Primers were designed to amplify a 250-bp segment (from each gene), overlapping where the segments of the cassette should be fused (Fig. 1a; Table S1). It was necessary to avoid restriction sites for the enzymes used during later cloning as well as take care that potential start codons in the sequences amplified from each gene were out of frame. First, a PCR with the appropriate primers to amplify the segments separately was run (Bucher et al., 2006) using A. tumefaciens strain C58 DNA as template. Then, the PCR products were used in another round of PCR without primers, resulting in the fusion of the two segments and the generation of a 500-bp chimeric transgene (Shevchuk et al., 2004). The product of this PCR was used in an additional PCR with primers complementary to the fragment ends (Fig. 1a), to produce a sufficient amount of the transgene. The chimeric cassette flanked by XhoI and EcoRI (sense) or XbaI and HindIII (antisense) was cloned into inverted repeats, around the intron, into pHANNIBAL (Wesley et al., 2001). Then, this plasmid was digested with NotI and the complete expression unit of the hairpin RNA was cloned into the pART27 binary plasmid (Gleave, 1992) containing the nptII gene for antibiotic resistance; the new pSilenceCG plasmid generated was transformed into A. tumefaciens strain EHA105. The new construct (Fig. 1b) was checked by restriction digestion and sequencing.

Figure 1.

 Construction of the chimeric transgene. (a) Sequences of primers used to amplify and join highly conserved fragments of the ipt and iaaM oncogenes. The fusions of the fragments were in sense (upper sequence) and antisense (lower sequence). (b) Sense and antisense fragments were then cloned at both sides of an intron to produce hairpin RNA. The figure shows the schematic representation of the transcription unit, with the NotI sites at both ends to clone it within the T-DNA of the pART27 plasmid.

Bacteria were cultured overnight in Luria broth medium (LB) with the corresponding antibiotics. Then, the culture was diluted to an OD600 of 0·3 and used to infect Nicotiana tabacum cv. Xanthi leaf discs, using previously published procedures (Clemente, 2006).

DNA extraction, PCR and Southern blot analysis

The DNA was extracted from 100 mg leaf tissue of putative transformed shoots using the DNeasy Plant Mini Kit (QIAGEN GmbH). The PCR evaluation was carried out with primers specific for the introduced transgenes nptII (Miguel & Oliveira, 1999) and ipt-iaaM (primers A1 and A2 in Fig. 1), after discarding A. tumefaciens contamination by using primers specific for virG genes (Petri et al., 2004). Amplified DNA was detected under UV light after electrophoresis of the amplification reaction mixture for each sample on 1% (w/v) agarose gels and staining with ethidium bromide. For Southern analysis, 20-μg BamHI-digested DNA samples were separated on 1% (w/v) agarose gels and transferred to positively charged nylon membranes (Roche GmbH) by capillary blotting. The 696-bp PCR nptII fragment was labelled with digoxigenin (DIG) using the PCR DIG Probe Synthesis Kit (Roche GmbH). Prehybridization and hybridization of blots to the labelled probe were performed at 42°C. The blots were then washed twice at 23°C in 2× SSC (0·3 m NaCl, 0·03 m sodium citrate) plus 0·1% (w/v) sodium dodecyl sulfate (SDS) for 15 min, and twice at 65°C in 0·5× SSC, 0·1% SDS for 15 min. Hybridizing bands were visualized with anti-DIG antibody-alkaline phosphatase and CDP-Star (Roche GmbH) on X-ray films.

Greenhouse evaluation of transgenic lines

Non-transformed control tobacco plants, maintained in vitro, were acclimatized in the greenhouse. They were then infected with the disarmed EHA105 strain and with three oncogenic A. tumefaciens strains following previously described procedures (Cervera et al., 1998; Petri et al., 2004). Oncogenic strains belong to different opine groups, as defined by their Ti plasmids: C58 is a nopaline strain, Ach5 is an octopine strain and A281 is an L,L-succinamopine strain (Cervera et al., 1998). Briefly, the bacterial strains were cultured in LB until the stationary phase and then diluted to an OD600 of 0·3 in SIM (simplified induction medium) (Alt-Mörbe et al., 1989) supplemented with 500 μm acetosyringone and left for at least 5 h to induce virulence. Stems from plants around 40 cm in height were T-wounded with a scalpel and inoculated with 5 μL of the bacterial culture diluted in water to final densities of 106, 107, 108, and 109 CFU mL−1. The wounds were covered with Parafilm for 2 weeks to avoid desiccation. Crown gall symptoms in the wounds (degree of tumour formation) were assigned a value from 0 to 3 (Fig. 2b).

Figure 2.

 Greenhouse evaluation of transgenic tobacco plants for crown gall resistance. (a) Observed (circles) and predicted (dashed line) stem diameter increases, and observed (squares) and predicted (solid line) symptoms, after infection with increasing concentrations of the oncogenic Agrobacterium tumefaciens strain C58. (b) Symptom evaluation scale from 0 to 3, used during the greenhouse tests. (c) Stem diameter increases and symptom evaluation in wildtype plants infected with the oncogenic C58 or disarmed EHA105 A. tumefaciens strain, as well as in transgenic lines infected with strain C58.

T0 transgenic plants were acclimatized in the greenhouse and challenged with the strains and bacterial concentrations selected. After 8 weeks, tumour symptoms and increase in stem diameter at the inoculation site were recorded.

Northern blots

For siRNA analysis, total RNA was extracted with the Tri® Reagent (Sigma-Aldrich Co.), following the manufacturer’s directions, from leaves of transgenic shoots as well as from wildtype tobacco leaves used as a negative control. Samples equivalent to 20 μg total RNA, according to NanoDrop® 1000 analysis, were re-dissolved in 50% formamide and, after heat treatment, were loaded onto a polyacrylamide gel (20%). After electrophoresis, the nucleic acids were electroblotted onto a positively charged nylon membrane (Roche GmbH). Hybridization was performed in Dig Easy Hyb buffer (Roche GmbH) using an in vitro-transcribed, DIG-labelled RNA probe corresponding to the full length of the chimeric gene described above. Detection of the siRNA was performed with anti-DIG antibody–alkaline phosphatase and CDP-Star (Roche GmbH) on X-ray films.

Gene expression analysis in transgenic plants and detection of the ipt and iaaM oncogenes by qRT-PCR

The RNA samples were digested with DNase I using the DNA-free Kit (Ambion), and quantified using a spectrophotometer (NanoDrop 1000).

cDNA was synthesized using the RETROscript cDNA Synthesis Kit (Ambion). Reverse transcription was carried out at 42°C for 1 h in a 20-μL reaction mixture containing 1 μg total RNA, reverse transcription buffer, dNTPs at 500 μm each, 5 μm oligo-dT primers, 0·5 units RNase inhibitor and 5 units Moloney Murine Leukemia Virus (MMLV)-Reverse Transcriptase.

The expression levels of the transgene, oncogenes and a β-actin gene used as a normalizer were determined by real-time qPCR using the GeneAmp 7500 sequence detection system (Applied Biosystems), in different lines. The cDNA was first synthesized as described above, from three independent RNA extractions, and PCR was carried out on the cDNA in triplicate, in 96-well plates, using the SYBR Green Master Kit (Applied Biosystems).

Primers were designed by following the instructions given with the sequence-detection-system software. The primers used were as follows: for β-actin (forward 5′-CTGGCATTGCAGATCGTATGA-3′ and reverse 5′-GCGCCACCACCTTGATCTT-3′), for the transgene (S1 and S2 in Fig. 1 and Table S1), for the ipt oncogene (forward 5′-TGGGCTTCCAGTCCTTCG-3′ and reverse 5′-TCCGCCTTCCGGTTGACA-3′), and for the iaaM oncogene (forward 5′-CCTGTTTTTGCTGAGCGGTAA-3′ and reverse 5′-ACCCGGTAACGCATTTCATC-3′). Each set of primers was mixed, at a final concentration of 300 nm, with 2 μL cDNA and SYBR Green. After denaturation at 95°C for 10 min, amplification occurred in a two-step procedure: 15 s denaturation and 1 min of annealing and extension at 60°C for 40 cycles. These conditions were used for both target and reference genes and the absence of primer-dimers was checked in controls lacking templates. Transcript levels were calculated using relative standard curves for both target and normalizing genes, which were made after running serial dilutions, covering a 1- to 1000-pg range of purified cDNA of the genes, as described in ABI PRISM 7700 User Bulletin #2 (Applied Biosystems). The RNAs were normalized between samples to the β-actin level.


Evaluation of transgenic tobacco plants for crown gall resistance

Tobacco transgenic lines were produced following standard procedures. Plants surviving antibiotic selection were checked by PCR and Southern blots as shown for some of the lines in Fig. 3a,b. The number of T-DNA integrations ranged from only one (lines 809-4 and 809-7) to complex integrations, as for line 809-8 (Fig. 3b).

Figure 3.

 Molecular evaluation of transgenic tobacco plants. (a) PCR amplification of a 700-bp fragment from the nptII transgene and 507 bp from the chimeric transgene ipt-iaaM in transgenic tobacco plants. Lane M: 100-bp DNA ladder, P: positive control (pSilenceCG plasmid), C: negative control. (b) Southern blot analysis with an nptII probe of 11 putative tobacco lines transformed with the pSilenceCG plasmid. Lane C: negative control (non-transformed tobacco DNA), P: positive control (pSilenceCG plasmid), M: Lambda HindIII Dig-labelled molecular marker.

A preliminary in vitro evaluation (see Data S1) with explants from wildtype plants demonstrated that the three oncogenic strains produced callus symptoms and weight increases. Furthermore, there were no differences among the three oncogenic strains (Fig. S1a). In vitro evaluation of the T0 transgenic tobacco plants suggested a degree of resistance in most lines (Fig. S1b). However, the reactivity of stem explants to the PGR-free medium did not allow clear differentiation between wildtype plants infected with the disarmed or oncogenic strain (Fig. S1b).

The kanamycin-resistant T1 plants were also subjected to preliminary in vitro evaluation by infecting different explants (hypocotyls, first true leaves, and roots), and the results depended on the type of explant used (Data S2, Table S2, Fig. S2).

An in planta evaluation was also carried out, following previously described procedures (Cervera et al., 1998). In this and other papers, the sensitivity of different plant species to A. tumefaciens was tested at high concentrations of inoculum. Different inoculum densities were tested on wildtype plants to determine the minimum concentration that produced clear tumour symptoms on the plants (Fig. 2a). An A. tumefaciens concentration of 108 colony forming units per millilitre (CFU mL−1) was enough to produce clear symptoms (scored from 0 to 3, as indicated in Fig. 2b) of crown gall disease. In fact, there was a linear increase in symptoms and in stem diameter with the increase in the A. tumefaciens concentration in the aqueous solution (Fig. 2a). Additionally, it was observed that the three oncogenic A. tumefaciens strains produced similar symptoms and increases in stem diameter and that all of them produced significantly greater increases in symptoms and stem diameter (P < 0·001 for both parameters) than the disarmed strain EHA105 (data not shown).

When transgenic lines were challenged in the greenhouse with the oncogenic strain C58 at a concentration of 109 CFU mL−1, 8 out of the 10 lines tested showed a significantly smaller diameter increase and seven out of 10 fewer symptoms than wildtype plants infected with the same strain. Line 937-6 (resistant) showed similar behavior to wildtype plants infected with the disarmed strain EHA105, whereas the six remaining lines (tolerant) exhibited intermediate behaviour, with symptoms and diameter increases that differed from those of wildtype plants infected with the disarmed strain (Fig. 2c).

It is important to note that, because the inoculum pressure used in these evaluations was very high (the probability of plants being challenged with this bacterial concentration in natural conditions is low), line 937-6 has a high degree of resistance (Fig. 4e).

Figure 4.

 Determination of transgene and oncogene expression and fate in tobacco plants. (a) Northern blot analysis of small interfering RNA (siRNA) accumulation. (b) Quantitative real-time PCR analysis of the accumulation of the transgene hpRNA in transgenic lines. (c) Similar analysis for the accumulation of ipt and iaaM mRNAs. Leaf discs from wildtype plants were infected with oncogenic strain C58 or disarmed strain EHA105, whereas leaf discs from different transgenic lines were infected with oncogenic strain C58. (d) Inoculated stems of wildtype whole tobacco plants infected with the oncogenic strain C58 developed a large callus at wounded sites, whereas plants from line 937-6 (e) did not show any symptoms; bar = 1 cm.

Accumulation of transgene-specific siRNAs and transgene hpRNA

RNA was isolated from six lines, representing different degrees of resistance, and assayed for the presence of siRNAs of approximately 21–25 nucleotides, homologous to the iaaM and ipt conserved fragments introduced into the transformed plants. As expected, the non-transformed control and the sensitive line 809-8 did not show any accumulation of siRNA. A signal was found, however, for the rest of the lines, including the resistant line 937-6 and the tolerant lines 809-7, 809-3 and 809-5. Surprisingly, a strong signal was found also for the sensitive line 809-4 (Fig. 4a).

Transgene RNA levels were also quantified by real-time reverse transcriptase-PCR (RT-PCR). All resistant plants accumulated very low amounts of transgene hpRNA (0·8 ± 0·3 to 17·0 ± 6·7 pg transgene hpRNA per pg β-actin mRNA) (Fig. 4b). The highest accumulation was found for the non-resistant line 809-4 (59·1 ± 14·2 pg transgene hpRNA per pg β-actin mRNA), that also accumulated siRNAs (Fig. 4a,b). In line 809-8 only 18·7 ± 11·1 pg transgene hpRNA per pg β-actin mRNA were accumulated, while siRNAs were not detected in its northern blot (Fig. 4a,b).

Leaf explants from the resistant line 937-6 did not accumulate mRNA from ipt or iaaM oncogenes after infection with the oncogenic A. tumefaciens strain (Fig. 4c). Accumulation of oncogenic mRNAs was similar to that in wildtype plants infected with the disarmed EHA105 strain (Fig. 4c). Explants from sensitive lines and wildtype plants infected with the oncogenic C58 strain showed much higher levels of iaaM and ipt mRNAs.


In this work, transformation with a chimeric transgene reduced the appearance of symptoms or induced complete resistance to A. tumefaciens in tobacco transgenic plants.

Active post-transcriptional gene silencing (PTGS) has been consistently associated with the accumulation of siRNAs, which may act as ‘guides’ for the degradation of homologous mRNAs in the cytosol. The bands of 21–25 nucleotides were expected in the clearly resistant lines, such as 937-6, and in those with reduced symptoms, but a strong signal was also present in the sensitive line 809-4.

From the results, it seems that accumulation of transgene-derived siRNAs is required for PTGS-mediated resistance against crown gall, but that intron-hairpin lines accumulating high levels of transgene-derived siRNAs may also be susceptible to crown gall. These results agree with those obtained by López et al. (2010) in Mexican lime transformed with sense, antisense and intron-hairpin cDNAs from viral sequences: all resistant lines accumulated transgene-derived siRNAs, but this was not necessarily associated with resistance to Citrus tristeza virus and only a fraction of the transgene-derived siRNAs were competent for RNA silencing, the rest being quickly degraded.

PTGS would be expected to reduce the level of transgene hpRNA as a consequence of degradation by the Dicer enzyme. Transgene hpRNA accumulation and resistance were correlated because all resistant plants accumulated very low amounts of transgene hpRNA, which suggests its degradation. These results agree with those of López et al. (2010) where Citrus tristeza virus resistance was correlated with low accumulation of the transgene-derived transcript rather than with high accumulation of transgene-derived siRNAs. However, the non-resistant line 809-4 showed a high accumulation of transgene hpRNA and also of siRNAs; this points to inefficient degradation of the transgene hpRNA. Another situation is represented by line 809-8, which accumulated low amounts of transgene hpRNA, with no signal for siRNAs detected. Similar results were found in gall-sensitive, transgenic apple plants that accumulated very low levels of transgene dsRNA and failed to silence the oncogenes, although the authors did not evaluate accumulation of siRNA in these lines (Viss et al., 2003). A number of factors may be responsible for the low accumulation of dsRNA or hpRNA transgenes in the absence of PTGS; for instance, weak transcription as a result of chromosomal location or, alternatively, gene rearrangements which may have removed promoters or transgene sequences from the integrated T-DNAs or may have led to transgene silencing. The former is consistent with the large number of T-DNA insertions in line 809-8 (Fig. 3b), with a complex pattern of integrations, which may have triggered the silencing of transgene expression and which would explain both the low accumulation of its hpRNA and the absence of siRNAs. As expected, line 937-6 had the lowest hpRNA accumulation, which correlates with the lack of disease symptoms. Furthermore, the expression of oncogenes in leaf explants from the resistant line 937-6 was silenced after infection by A. tumefaciens strain C58.

Regarding the efficiency of producing plants resistant to crown gall with each of the transformation strategies tested to date, Arabidopsis plants transformed with the VirD2 gene produced tumours after infection with oncogenic A. tumefaciens A208, at a rate four- to 10-fold lower than in wildtype plants. The VirD2 protein is involved in the protection and integration efficiency of T-DNA in the plant genome and competition probably occurs between the protein expressed in the plant and that introduced after infection, linked to the T-DNA, for limiting plant factors involved in T-DNA integration in the genome (Hwang et al., 2006).

However, as indicated before, survival of transformed VirD2 plants was very low when expression of the full-length protein was under the control of a constitutive promoter. When a wound-inducible promoter was used, the rate of tumour reduction was two- to fourfold that of wildtype plants. Unfortunately, VirD2 transgenic plants were re-transformed with very low efficiency, which would limit the possibility of stacking desired characteristics in those plants, mediated by A. tumefaciens (Hwang et al., 2006).

Around 15% of Vitis plants transformed with the virE2 gene were considered resistant to crown gall because they did not show tumours at all or half, or less, of the inoculation sites had small tumours (Krastanova et al., 2010). The molecular mechanism underlying the resistance mediated by a chimeric virE2 is unclear, but it is hypothesized that, following infection, the truncated VirE2 protein, devoid of a single-stranded DNA-binding domain, competes with the wildtype VirE2 of A. tumefaciens for VirE2-interacting host factors essential for T-DNA transformation (Krastanova et al., 2010).

When constructs encoding sense-strand RNAs from ipt and iaaM genes with premature stop codons were used to produce crown-gall-resistant tobacco plants, the oncogenes were simultaneously silenced in 3% of the plants and these were resistant to the disease (Lee et al., 2003). Of the apple plants produced by Viss et al. (2003) using the same strategy but with constructs encoding sense and antisense RNA, 12% were crown-gall-resistant. In fact, self-complementary RNA constructs split by an intron have been the most efficient strategy with regard to producing crown-gall-resistant plants. Sixty percent of tested tomato transformants and 50% of tested Arabidopsis transformants were completely resistant in the study of Escobar et al. (2001). These plants were functionally resistant because there was an inhibition of tumour production; however, they could be transformed again and there were no significant differences in the number of roots able to form herbicide-resistant calluses between double-transgenic lines and plants that had not been previously transformed. This means that the expression of hpRNA transcripts induces silencing of A. tumefaciens oncogenes but does not affect the ability to integrate T-DNA in the plant cell. These authors also found that around 54% of transgenic, self-complementary RNA walnut lines were resistant to crown gall, versus only 20% of resistant lines transformed with sense and antisense constructs and none of those transformed with antisense constructs (Escobar et al., 2002). Additionally, the strategy of using the self-complementary RNA designed to initiate PTGS of iaaM and ipt gave a broad-spectrum crown gall resistance, to several different A. tumefaciens strains (Escobar et al., 2001).

In this work the size of the T-DNA construct was greatly reduced. As the ultimate objective of this research is to transform recalcitrant species, such as Prunus, it is paramount to improve transformation efficiency as much as possible, and a shorter T-DNA length has been reported to improve transformation frequency (Park et al., 2000). Additionally, a shorter length reduces the chances of DNA recombinations or rearrangements during the T-DNA transfer. With the construct described in this manuscript, around 70% of the lines showed a significant reduction in the development of the disease and 10% of the lines were absolutely resistant after infection with the oncogenic C58 strain, without significant differences from wildtype plants infected with the disarmed strain. Additionally, it is expected that the resistance will have a broad spectrum, given the approximately 90% level of cross-strain DNA sequence conservation in the iaaM and ipt oncogenes (Escobar et al., 2001) and the selection of highly conserved sequences within each gene when the chimeric transgene in the present study was designed.

The resistant line 937-6 was challenged with the A. tumefaciens strains Ach5 and A281, belonging to two other opine groups. As expected, plants from this line showed complete resistance to the Ach5 strain, but surprisingly, only showed a slight reduction in symptoms, compared with the wildtype plants, when infected with strain A281. Fragment sequences in the chimeric transgene showed very high homology to all iaaM sequences found in GenBank. The homology for the ipt fragment used in the transgene in this study was over 95% for most of the strains found, with the exception of the ipt gene from A281 with 90% homology, as indicated in previous studies of the ipt gene in the Ti plasmid Bo542 from strain A281 (Strabala et al., 1989). A detailed re-examination of the construct in the present study showed that mismatches between the chosen ipt fragment and the A281 ipt gene sequence are regularly spaced, in such a way that it can be estimated that they will be in the central and/or 5′ region of the generated siRNAs (Waterhouse et al., 2001). Similar nucleotide mismatches have been related to a drastic decrease in the silencing effectiveness of siRNAs (Pusch et al., 2003), but in the present case, the problem could be solved easily by adding an additional fragment, specifically designed to have a high homology with the ipt from A281, to the chimeric transgene.

Biochemical pathways for endogenous plant hormones may involve reactions analogous to those catalysed by iaaM and ipt, although no sequences in plant genes similar to those of oncogenes have apparently been published. This means that oncogenes may be silenced in plant cells infected by A. tumefaciens without changing the plant hormonal balance. This would be in agreement with the fact that no abnormal phenotypes were observed in the transgenic tobacco plants in this study, nor have they been reported by other authors (Lee et al., 2003). The exception is line 809-8 in the present study, which has a large number of insertions and had difficulty in proliferating in vitro, although its T1 did germinate and grow normally, as did the T0 plants in the greenhouse. Mean stem diameter before infection was 8·2 mm for control plants and ranged from 7·1 to 8·5 in the transgenic plants, depending of the specific line. As far as is known, abnormal phenotypes have not been reported in plant species such as tomato or Arabidopsis (Escobar et al., 2001), walnut (Escobar et al., 2002) or apple (Viss et al., 2003), in which oncogenes were also silenced.

Although some anti-silencing activity has been described in tumours produced by A. tumefaciens, it seems to be a consequence of the uncontrolled cell proliferation and, in contrast with viruses, where silencing-suppressive proteins have been found frequently, these types of protein do not seem to occur in A. tumefaciens (Dunoyer et al., 2006). Therefore, resistance induced by silencing of oncogenes may be very durable, because its specificity is mediated by RNA hybridization rather than protein interactions (Escobar et al., 2001).


The authors thank Dr Leandro Peña for providing the A. tumefaciens strains and Dr David J. Walker for critical reading of the manuscript. This work was partially supported by projects AGL2010-20270 (co-financed by FEDER funds) and BioCarm BIO-AGR07/04-0011.