Fusarium wilt, caused by Fusarium oxysporum f. sp. cubense (Foc), is one of the most devastating diseases of banana (Musa spp.). Apart from resistant cultivars, there are no effective control measures for the disease. We investigated whether the transgenic expression of apoptosis-inhibition-related genes in banana could be used to confer disease resistance. Embryogenic cell suspensions of the banana cultivar, ‘Lady Finger’, were stably transformed with animal genes that negatively regulate apoptosis, namely Bcl-xL, Ced-9 and Bcl-2 3′ UTR, and independently transformed plant lines were regenerated for testing. Following a 12-week exposure to Foc race 1 in small-plant glasshouse bioassays, seven transgenic lines (2 × Bcl-xL, 3 × Ced-9 and 2 × Bcl-2 3′ UTR) showed significantly less internal and external disease symptoms than the wild-type susceptible ‘Lady Finger’ banana plants used as positive controls. Of these, one Bcl-2 3′ UTR line showed resistance that was equivalent to that of wild-type Cavendish bananas that were included as resistant negative controls. Further, the resistance of this line continued for 23-week postinoculation at which time the experiment was terminated. Using TUNEL assays, Foc race 1 was shown to induce apoptosis-like features in the roots of wild-type ‘Lady Finger’ plants consistent with a necrotrophic phase in the life cycle of this pathogen. This was further supported by the observed reduction in these effects in the roots of the resistant Bcl-2 3′ UTR-transgenic line. This is the first report on the generation of transgenic banana plants with resistance to Fusarium wilt.
If you can't find a tool you're looking for, please click the link at the top of the page to "Go to old article view". Alternatively, view our Knowledge Base articles for additional help. Your feedback is important to us, so please let us know if you have comments or ideas for improvement.
Fusarium wilt, also known as Panama disease, has been and continues to be a major constraint and serious threat to banana (Musa spp.) production worldwide (Ploetz and Pegg, 2000). The disease is caused by the soil-borne fungal pathogen Fusarium oxysporum forma specialis (f. sp.) cubense (Foc), of which four physiologically distinct ‘races’ (referred to as races 1–4) have been identified based on their variation in virulence to specific host cultivars. Races 2 and 3 are not considered economically important, as they do not infect commercially relevant banana cultivars. Foc race 1, however, infects commercially important cultivars such as Gros Michel (Musa spp. AAA group) and ‘Lady Finger’ (AAB). In the mid-1950s, Foc race 1 decimated the major export cultivar (Gros Michel) in South and Central America, essentially eliminating its use and leading to the adoption of the race 1-resistant Cavendish subgroup of cultivars (AAA) as the dominant export commodity. Foc race 4 infects all race 1-susceptible cultivars as well as the Cavendish cultivars and, until relatively recently, only affected bananas in subtropical climates [and was therefore designated subtropical race 4 (SR4)]. More recently, a newly discovered Foc variant called tropical race 4 (TR4) has been identified, which affects Cavendish cultivars, and other locally important types such as the plantains, growing in tropical regions. This variant is apparently spreading and has been responsible for significant plantation losses is South-east Asia, particularly Malaysia, China, Philippines and Indonesia as well as northern Australia (Ploetz and Pegg, 2000; Ploetz, 2006).
There are currently no effective methods available for controlling Foc. Chemical control is ineffective, and Fusarium chlamydospores remain viable in the soil for several decades rendering infested ground unsuitable for growing susceptible banana cultivars. It is generally accepted that the only option for controlling the disease is through the use of genetically resistant cultivars generated by conventional breeding or genetic modification. Although sources of resistance to Foc have been identified in wild bananas (Ploetz, 2006), the exploitation of these ‘resistance genes’ by conventional breeding has been hampered by the extremely low fertility of commercial banana cultivars, which essentially do no set seeds, thus precluding traditional breeding strategies. As such, genetic modification is generally regarded as the most viable strategy for developing bananas with enhanced agronomic traits, particularly with the recent availability of efficient and reliable banana transformation protocols (Becker et al., 2000; Khanna et al., 2004).
Different forms of programmed cell death (PCD), including apoptosis and autophagy, occur in plants during normal growth and development as well as in response to environmental stresses and pathogen attack (Greenberg, 1996; Lenz et al., 2011). We have previously shown that the extensive cell death observed in banana cell suspensions exposed to Agrobacterium tumefaciens was accompanied by several features characteristic of apoptosis, including DNA laddering and fragmentation and the formation of apoptotic-like bodies (Khanna et al., 2007). Further, these cellular responses were shown to be inhibited in cells expressing the antiapoptosis genes Bcl-xL, Bcl-2 3′ UTR and Ced-9. The transgenic modification of pathways controlling PCD in plants has indicated that engineering resistance to biotic stresses in plants is a valid strategy (Dickman et al., 2001). Dickman et al. (2001) showed that the constitutive expression of various antiapoptotic Bcl-2 gene family members (including Bcl-xL, Ced-9 and Bcl-2 3′ UTR) in transgenic tobacco plants resulted in high levels of resistance to a broad range of necrotrophic fungi. As Fusarium oxysporum is a necrotrophic fungus (Vajna, 1985; Trusov et al., 2006) that kills host cells prior to infection through the predicted deployment of toxins and enzymes that induce cell death, we hypothesized that transgenic expression of antiapoptosis genes would confer resistance to Foc. In this paper, we provide proof-of-principle that antiapoptosis transgenes do confer resistance against Foc race 1 in banana.
Generation and characterization of transgenic plants
Binary vectors were generated containing the antiapoptosis genes Bcl-xL, Ced-9 and Bcl-2 3′UTR, all under the control of the maize polyubiquitin (Ubi-1) constitutive promoter. Embryogenic cell suspensions (ECS) of the banana cultivar ‘Lady Finger’ were transformed with each of the constructs using an Agrobacterium-mediated transformation protocol from Khanna et al. (2004), and transgenic lines were regenerated. Thirty-one independently transformed lines (7 × Bcl-xL, 11 × Ced-9, 13 × Bcl-2 3′ UTR) were identified based on PCR and Southern analysis, and these were selected for further study.
To determine whether transgene expression resulted in any deleterious phenotypic effects, each of the transgenic lines was multiplied and ten plants from each line were acclimatized and grown in the glasshouse for 8 weeks. Plants from three Bcl-xL lines, two Ced-9 lines and five Bcl-2 3′ UTR lines displayed a range of mild phenotypic abnormalities that included stunting as well as altered leaf morphology and phyllotaxy. In most cases, the abnormal phenotype was present in every replicate of a given line. None of the cell control lines displayed abnormalities. Phenotypic abnormalities have previously been reported in plants expressing high levels of antiapoptosis genes (Dickman et al., 2001); however, abnormal phenotypes, or ‘off-types’, also occur naturally in nontransgenic tissue-cultured banana caused by somaclonal variations (Côte et al., 1993; Israeli et al., 1996; Reuveni et al., 1996). It is not known whether the abnormalities observed in some transgenic plants generated in this study were the result of naturally occurring somaclonal variations or transgene expression or both. However, the observation that (i) all but one of these abnormalities were off-types commonly observed associated with somaclonal variation in tissue-cultured, nontransgenic banana plants (Israeli et al., 1991) and (ii) none of the abnormalities were typical of those previously associated with antiapoptosis transgene expression (Dickman et al., 2001), suggested that they were unlikely to be transgene-related.
Two independent ‘small-plant’ bioassay glasshouse trials were conducted to assess the transgenic banana lines for resistance to Foc race 1. The first trial (T1) included 15 independently transformed banana lines (6 × Bcl-xL, 5 × Ced-9 and 4 × Bcl-2 3′ UTR), while the second trial (T2) included the most promising line/s from trial 1 plus an additional 16 independently transformed banana lines (1 × Bcl-xL, 6 × Ced-9 and 9 × Bcl-2 3′ UTR). For each of the trials, ten plants of each transgenic line were inoculated, while ten plants of both wild-type susceptible ‘Lady Finger’ and resistant Cavendish cv ‘Grand Naine’ plants were included as controls. Two plants from each transgenic and wild-type line were also included as noninoculated controls. Twelve-week postinoculation, the external (yellowing, wilting, stem-splitting) and internal symptoms (corm discoloration) of Fusarium wilt infection were assessed. In the second trial, a subset of plants was evaluated at 23-week postinoculation to assess the potential for long-term resistance to Fusarium wilt in the glasshouse.
In both trials, typical external disease symptoms first appeared on the susceptible ‘Lady Finger’ wild-type positive control plants within 3–5 weeks. At 12-week postinoculation (Figure 1a), these plants showed mean yellowing and wilting scores of 2 and 2.2, respectively (on a scale of 1–5), and mean stem-splitting scores of 2.8 (on a scale of 1–3) in T1 and scores of 2.6, 2.9 and 2.6, respectively, in T2 (Table 1). When the internal symptoms were assessed, an average of 44.9% and 50.6% corm discoloration was observed in trials 1 and 2, respectively (Table 1 and Figure 2a). As expected, plants from the resistant ‘Grand Naine’ wild-type line treated with Foc race 1 displayed no symptoms of wilting and stem-splitting, with corm discoloration at 2.8% and 0.1% in trials 1 and 2, respectively (Table 1 and Figure 2a). For the ‘Grand Naine’ controls, the average wilting and stem-splitting ratings were significantly less than the ‘Lady Finger’ controls; however, the average yellowing ratings were not, suggesting this symptom was not a reliable disease indicator in small-plant assays. None of the noninoculated wild-type plants of either cultivar showed external or internal signs of infection (Figures 1 and 2b).
Table 1. Assessment of external symptoms of Foc race 1 infection on Bcl-xL, Ced-9 and Bcl-2 3′ UTR-transgenic lines
Results are presented as score means based on ten replicates.
†Based on 1–5 scale.
‡Based on 1–3 scale.
*Significantly different from their respective susceptible LF control lines in either trial 1 (T1) or trial 2 (T2) with P <0.05 based on LSD post hoc test.
Bcl-2 3′ UTR-6 (T1)
Bcl-2 3′ UTR-6 (T2)
Bcl-2 3′ UTR-17 (T2)
Seven transgenic lines [2 × Bcl-xL (15 and 30), 3 × Ced-9 (1, 4 and 6) and 2 × Bcl-2 3′ UTR (6 and 17)] identified from both trials had significantly lower disease ratings than the wild-type susceptible ‘Lady Finger’ control banana plants. For line Bcl-xL-15 and line Ced-9-6, only the stem-splitting rating and percentage corm discoloration were significantly lower than the susceptible controls, while for line Bcl-xL-30, the wilting and stem-splitting ratings and percentage corm discoloration were significantly lower (Table 1 and Figure 2). For two Ced-9 lines (Ced-9-1 and Ced-9-4) and two Bcl-2 3′ UTR lines (Bcl-2 3′ UTR-6 and Bcl-2 3′ UTR-17), all external symptom ratings and the amount of vascular discoloration were significantly lower than the susceptible controls (Table 1, Figures 1b,c and 2).
Because all of the external symptom ratings for Ced-9-4, Bcl-2 3′ UTR-6 and 17 in trial 2 were significantly lower than the susceptible controls, one replicate for each line was not sacrificed for evaluation of internal symptoms at week 12, but was instead kept in the glasshouse for an additional 11 weeks. Noninoculated wild-type ‘Lady Finger’ plants were also kept as controls. At 23-week postinoculation, the external and internal symptoms were assessed. The Bcl-2 3′ UTR-17, Ced-9-4 and Bcl-2 3′ UTR-17 plants had increased external symptom ratings over that additional 11-week period along with an increase in the amount of corm discoloration. In contrast, none of the external symptom ratings of the Bcl-2 3′ UTR-6 plant increased over that period nor did the amount of vascular discoloration.
Further characterization of the most resistant plant lines
To determine whether there was any correlation between the expression of the antiapoptosis transgenes and the level of disease resistance, the seven most promising transgenic lines were analysed by semi-quantitative RT-PCR and/or Western analysis. As a result of a lack of suitable antibodies against Ced-9, Western analysis could only be carried out on the 2 × Bcl-xL (15 and 30) lines. For the 3 × Ced-9 (1, 4 and 6) and 2 × Bcl-2 3′ UTR (6 and 17) lines, RT-PCR was used to assess transcript levels. Both RT-PCR and Western analyses revealed that higher transcript/protein levels were correlated with the resistance phenotype. For example, the amounts of Ced-9 and Bcl-2 3′ UTR-specific transcripts were found to be highest in the most resistant lines, Ced-9-1 and Bcl-2 3′UTR-6 (Figure 3a), respectively. Similarly, Western analysis revealed that higher amounts of Bcl-xL were present in line Bcl-xL-30 than in line Bcl-xL-15 (Figure 3b), with the former line also displaying a higher level of disease resistance. An unexpected observation was the cross-reactivity of the anti-Bcl-xL antibodies with a lower molecular weight protein present in all wild-type and transgenic lines tested.
Apoptotic response of banana roots to Foc
Foc is considered to be a necrotrophic pathogen. As necrotrophic pathogen appear to induce PCD during the course of infection (Dickman et al., 2001), we examined whether Foc could induce PCD-like features in banana roots and, if so, whether these effects could be prevented/reduced in the resistant transgenic banana line Bcl-2 3′ UTR-6. The terminal deoxynucleotidyl transferase-mediated dUTP nick-end labelling (TUNEL) assay is commonly used to detect two characteristic features of PCD associated with apoptosis, namely DNA fragmentation and the formation of apoptotic bodies, which occur from coalescence of specifically cleaved DNA. The assay relies on the presence of specific nicks in the DNA, which can be identified by terminal deoxynucleotidyl transferase, an enzyme that catalyses the addition of dUTPs that are secondarily labelled with a marker (fluorescein).
Roots from three wild-type susceptible ‘Lady Finger’ plants and three wild-type resistant ‘Grand Naine’ plants, as well as three clones of the transgenic banana line Bcl-2 3′ UTR-6, were initially incubated separately in a solution of sugar-free liquid rooting medium without Foc as controls. Root tissue samples taken at 48 h were analysed by TUNEL assays, and all samples contained largely intact nuclei with very few, if any, green (TUNEL-positive) apoptotic cells indicating that experimental conditions were not causing any background apoptotic effects (Figure 4a-i, b-i, c-i). In contrast, DNase-treated roots included as positive controls displayed a relatively large number of green, apoptotic cells (Figure 4a-iii, b-iii, c-iii).
When roots from wild-type ‘Lady Finger’ banana plants were incubated with an Foc spore suspension, TUNEL-positive nuclei were observed as early as 6-h postinoculation (Figure 4a-v) with the incidence of TUNEL-positive cells increasing to a maximum at 48 h (Figure 4a-vii, ix, xi). At 48-h postinoculation, a slight necrosis/browning type cell death was observed in the root tissue, at which time distorted nuclei and apoptotic-like bodies containing fragmented DNA were also observed. In comparison, roots from the Bcl-2 3′ UTR-transgenic plants exposed to Foc contained very few TUNEL-positive nuclei, like the nonexposed controls, and these were only detected after 12- and 24-h exposure to Foc (Figure 4b-v, vii). In addition, the tissue browning phenotype previously noted in Foc-exposed wild-type plants was not observed.
As a further control, the roots from the wild-type ‘Grand Naine’ plants, which are known to show field resistance to Foc race 1, were exposed to an Foc race 1 spore suspension. TUNEL assays (Figure 4c-vii) revealed that only a small number of positive cells, similar to nonexposed controls, were present in roots of the ‘Grand Naine’ cultivar primarily after 24-h exposure. These results indicate that (i) DNA fragmentation and cell browning observed in the wild-type ‘Lady Finger’ banana roots following exposure to Foc is a consequence of PCD which, in the case of the Bcl-2 3′ UTR-transgenic plants, could be reduced via expression of antiapoptosis genes and (ii) the lack of PCD in banana root cells following exposure to Foc is correlated with a resistance phenotype.
The transgenic modification of pathways regulating PCD in plants is emerging as a promising strategy for engineering broad-spectrum resistance to both biotic and abiotic stresses in plants (Dickman et al., 2001; Lincoln et al., 2002; Li and Dickman, 2004). Such a strategy is particularly suited for generating disease-resistant bananas as there are limited options for genetic improvement of this crop. In this study, we have provided proof-of-principle that antiapoptosis genes can be used in banana to confer resistance against Foc race 1. Of the 31 transgenic ‘Lady Finger’ banana lines challenged with Foc race 1 in small-plant glasshouse bioassays, two Bcl-xL, three Ced-9 and two Bcl-2 3′ UTR-transgenic lines exhibited significantly less external and internal disease symptoms than wild-type susceptible ‘Lady Finger’ control plants after a 12-week exposure to the pathogen. Of these lines, the transgenic line Bcl-2 3′ UTR-6 showed a level of resistance similar to the ‘Grand Naine’ cultivar. Importantly, further monitoring of the line Bcl-2 3′ UTR-6 transgenic showed that the resistance continued for at least 5 months after inoculation.
Despite limited published evidence, Foc is considered to be a necrotrophic pathogen. Such pathogens, by definition, require dead cells for nutrition, growth and development. A characteristic feature associated with many necrotrophs is the appearance of host PCD during the course of infection. Dickman et al. (2001) reported apoptotic responses in plants infected with several different necrotrophic fungi, including Sclerotinia sclerotiorum, Botrytis cinerea and Cercospora nicotianae, and concluded that apoptosis-like PCD occurs in compatible plant–necrotrophic pathogen interactions. In this study, we used TUNEL assays to demonstrate that an apoptotic-like cell death is triggered in susceptible wild-type banana root cells as early as 6-hour postexposure to a suspension of Foc spores. Further evidence suggesting that the observed DNA fragmentation was apoptotic in nature was provided from TUNEL assays on roots from a Foc race 1-resistant Bcl-2 3′ UTR-transgenic line in which these effects were significantly attenuated or absent. These results provide further evidence that Foc race 1 is a necrotrophic pathogen or at least has a necrotrophic stage in its life cycle.
Little is known regarding the molecular mechanisms underlying Foc infection and PCD induction in banana. Toxins produced by many necrotrophic fungi, including oxalic acid from S. sclerotiorum, AAL from Alternaria alternata, Fumonisin B1 from Fusarium verticillioides and victorin from Cochliobolus victoriae, induce characteristic hallmarks of apoptosis in plants (Abbas et al., 1995; Navarre and Wolpert, 1999; Kim et al., 2008). Further, in Fusarium species such as F. verticillioides and F. moniliforme, several secondary metabolites, including fumonisins, are produced during infections and act as powerful elicitors of PCD (Wang et al., 1996; Stone et al., 2000). Based on what is known about other F. oxysporum f. sp, Foc is predicted to produce mycotoxins that may be involved in stimulating cell death in the host and facilitating fungal growth. However, whether this and/or other mycotoxins are responsible for the PCD observed in Foc-infected ‘Lady Finger’ banana cells is unknown.
The mechanism(s) by which antiapoptosis genes confer protection against necrotrophic fungi is still unclear. Cytological studies on tobacco transformed with Bcl-xL and Ced-9 revealed that these proteins localized to mitochondria and other organelles including the vacuole and chloroplast (Chen and Dickman, 2004). Based on these observations, it was suggested that the proteins might improve the overall function of organelles by assisting in the generation of ATP in the mitochondrion or photorespiration to prevent ROS production in the chloroplast under stress conditions (Chen et al., 2003; Chen and Dickman, 2004; Li and Dickman, 2004). By analogy, Bcl-xL and/or Ced-9 may prevent cell death and enhance plant resistance characteristics by contributing to the maintenance of organelle homeostasis (Qiao et al., 2002). The mechanism/s by which Bcl-2 3′ UTR confers resistance to plant cells is also unknown. However, the Bcl-2 3′ UTR transcript is thought to interact with proapoptotic proteins at the RNA level (Dickman et al., 2001) to mediate their degradation via the ubiquitin/proteasome pathway (Martin Dickman, unpublished data).
The generation of ‘Lady Finger’ banana plants with resistance to Foc race 1 is significant, considering the serious economic and social impact caused by Fusarium wilt and the fact that no sustainable control strategy currently exists for this disease. The glasshouse tested resistant lines described here are very promising; however, testing of the transgenic plants under field conditions will ultimately be required, preferably through at least two crop cycles as symptoms of Foc infection can sometimes take up to 12 months to appear. Although the concentration of Foc in naturally infested soils is typically unknown, it is likely that the Foc inoculum pressure in the field sites would be dramatically lower than that of the very high, confined inoculum load used in our small-plant glasshouse assays. As such, the number of transgenic plants resisting infection to Foc race 1 in the field might be considerably higher than that observed in glasshouse trials.
A further important next step will be to test the transgenic banana lines generated in this study for resistance to Foc race 4, and particularly Foc TR4 that is emerging as a major threat to the global banana industry (Ploetz, 2006; Buddenhagen, 2009; Dita et al., 2010). As the apoptosis-related genes used in this study have been previously shown to generate broad resistance to a diversity of necrotrophic pathogens in other plants, it is possible that the transgenic banana generated in this study will also show resistance to the necrotrophic fungal disease black Sigatoka, caused by Mycosphaerella fijiensis, which also threatens the international viability of the crop. Assessment of plants for developmental and agronomic qualities under field conditions is also required, as well as evaluation of any increased susceptibility to biotrophic banana pathogens such as viruses.
Binary vectors pPTN254, pPTN261, pPTN396 and pPTN290 containing the maize polyubiquitin-1 (Ubi-1) promoter controlling the expression of antiapoptosis genes Bcl-xL, Ced-9, Bcl-2 3′ UTR and the UidA reporter gene encoding β-glucuronidase (GUS), respectively, were as described previously (Khanna et al., 2007). All genes were fully sequenced in their original vector prior to transformation to verify the presence and integrity of the coding sequences and the promoter/gene and gene/terminator borders.
Transformation of banana and molecular characterization of transgenic plants
Embryogenic cell suspensions (ECSs) of the banana cultivar ‘Lady Finger’ (Musa spp. AAB group) were initiated, maintained and transformed as described previously (Khanna et al., 2004). The molecular analysis of transgenic plants was essentially done as previously described by Khanna et al. (2007). RNA was extracted from 50 mg of fresh leaf tissue using an RNeasy Plant Mini kit (Qiagen, Valencia, CA) following the manufacturer’s instructions and was treated with RNase-free DNase (Promega, Madison, WI). RT-PCR was performed using a Titan One Tube RT-PCR Kit (Roche Diagnostics Australia Pty Ltd, Castle Hill, NSW, Australia), with each reaction containing 10 ng DNase-treated total RNA and the appropriate gene-specific primers. For Western analysis, crude plant protein extract (20 μg) from 100 mg of leaf tissue was separated on a 10% denaturing SDS polyacrylamide gel and transferred to nitrocellulose membrane. The membrane was incubated with rabbit monoclonal anti-Bcl-xL (BL804) (Bethyl Laboratories, Inc., Montgomery, TX) primary antibody (diluted 1/1000 in blocking solution) for 16 h at 4 °C. Following a subsequent incubation in HRP-conjugated secondary antibody in blocking solution for 1 h at 4 °C, the SuperSignal West Femto (Pierce Biotechnology, Inc., Rockford, IL) substrate was used for signal detection at 4 °C until the desired intensity was achieved.
Fungal cultures and inoculum preparation
Fusarium oxysporum f. sp. cubense race 1 (VCG 0124/5) isolate was obtained from the Department of Employment, Economic Development and Innovation (DEEDI) herbarium, Indooroopilly Research Centre, Australia. The fungus was grown on 1/4 strength potato dextrose agar (PDA) supplemented with 50 mg/L streptomycin and incubated for 3–5 days at 25 °C. The inoculum used for small-plant bioassays was in the form of Foc-colonized Japanese millet (Echinochloa esculenta) grain (Smith et al., 2008). For in vitro root assays, spore suspensions were used as inoculum. These were prepared by inoculating 100 mL of Fusarium liquid growth medium-A (without yeast extract and supplemented with 20.6 g/L tri-sodium citrate) (Ahamed and Vermette, 2009) with two 7-mm-diameter plugs of PDA-grown Foc. After shaking at 120 r.p.m. for 3 days at 25 °C, the medium was filtered through three layers of cheesecloth and centrifuged at 3850 g for 5 min, and the pelleted conidia (a mixture of macroconidia and microconidia) were resuspended in 5 mL distilled water. Spore concentration was calculated using a haemocytometer before dilution to 105 spores/mL in sugar-free liquid rooting medium M5 (Côte et al., 1996).
Tissue-cultured transgenic plants of the ‘Lady Finger’ cultivar and wild-type banana plants of ‘Lady Finger’ and Cavendish cv ‘Grand Naine’ were acclimatized and transplanted into 100-mm-diameter pots in soil (Searles Premium Potting Mix, Kilcoy, Qld, Australia) for 8 weeks in a glasshouse at 27 °C under natural light conditions.
Selected transgenic banana lines and wild-type control plants were inoculated with Foc race 1 in two independent glasshouse trials. For logistical reasons, each independent transgenic line was assayed once in trial 1 or trial 2 with the exception for Bcl-2 3′ UTR line 6, which was assayed in both trials. For each line tested in both trials, ten replicate plants were inoculated and two plants were noninoculated. Plants were inoculated with Foc race 1 using a modified version of the Foc small-plant bioassay protocol described by Smith et al. (2008). Briefly, 200-mm pots were half-filled with potting mix (7-mm gravel, sand, perlite and vermiculite in a ratio of 2 : 2 : 1 : 1), and 20 mL of Foc-colonized millet grain (corresponding to an average of 2 × 108 Foc spores) was placed on the surface. Acclimatized banana plants were placed directly on the millet grain, and the pots were filled with potting mixture. A tablespoon of Osmocote Plus slow-release fertilizer (Scotts Australia Pty Ltd, Baulkham Hills, NSW, Australia) was added to each pot. Plants were maintained in a greenhouse with an average temperature between 23 and 25 °C during the summer and autumn months.
Assessment of disease symptoms
Twelve weeks after inoculation, external and internal symptoms of Fusarium wilt infection were assessed using a modified version of the method described by Smith et al. (2008). External symptoms were assessed by scoring each plant for the intensity of the three main disease symptoms. Yellowing and wilting were assessed using a 5 point scoring scale where 1 = healthy, no sign of symptoms (except natural degradation of lower leaves), 2 = slight symptoms, mainly on lower leaves, 3 = advanced symptoms (∼50%), 4 = extensive symptoms (∼90%) and 5 = entire plant affected (dead plant). Splitting of the stem was assessed using a 3-point scoring scale where 1 = no sign of splitting, 2 = slight splitting at the base of the plant and 3 = extensive splitting. Following assessment of external symptoms, plants were removed from their pots, the stem was cut longitudinally, and digital images of the rhizome were taken using a Canon Ixus 75 digital camera (Canon, Tokyo, Japan). The percentage of discoloration (browning) of the stellar region of the corm was subsequently assessed from the digital images using a MATLAB®-based program to ensure accuracy and reproducibility.
In vitro root assays
Tissue-cultured wild-type and transgenic ‘Lady Finger’ banana plants (5–7 cm tall) growing on M5 medium were transferred into liquid M5 medium in a 50-mL tube, and roots were allowed to develop for 2 weeks. The plants were then transferred into separate fresh 50-mL tubes containing 7 mL Foc spore suspension (105 spores/mL) and incubated at 25 °C with shaking (120 r.p.m.) for 48 h. Negative control samples were incubated in sugar-free M5 medium only. Root tip fragments (10 mm) for terminal deoxynucleotidyl transferase-mediated dUTP nick-end labelling (TUNEL) assays were taken from each plant at 6-, 12-, 24- and 48-h postinoculation, washed in 10 mm phosphate-buffered saline (PBS), fixed for 24 h in 4% (v/v) paraformaldehyde in 10 mm PBS (pH 7.4) at 4 °C and stored at 4 °C in 70% ethanol.
Root tip fragments were washed three times in fresh 10 mm PBS (pH 7.4) before being assayed by TUNEL essentially as described by Khanna et al. (2007). Positive control samples were made by subjecting root tissue to 1 unit of RNase-free DNase (Promega) for 10 min at room temperature. Stained root fragments were squash-mounted onto slides and examined under a BX41 microscope (Olympus Imaging Australia Pty Ltd, Macquarie Park, NSW, Australia) equipped with U-MWIBA3 and U-MWIY2 filters (Olympus), a DP71 microscope digital camera (Olympus) and the DP Manager software (Olympus).
Correlations between data were established using a correlation matrix of all external and internal symptoms. Owing to the ordinal scale of measurements, a Spearman’s Rho correlation coefficient was calculated to determine the most appropriate choice for further statistical analysis. Based on a high correlation coefficient, a multivariate analysis of variant (MANOVA) model was fitted using a general linear model (GLM) approach. This model included all four symptoms (yellowing, wilting, stem-splitting and vascular discoloration) concurrently. The LSD post hoc test was used to compare each plant line against the wild-type ‘Lady Finger’ control line with statistical significance reported at a level of P <0.05, using a two-tailed test. Statistical analysis was made using SPSS 15.0 (SPSS Inc., Chicago, IL).
This work was financially supported by the Australian Research Council, while J-YP was supported by a QUT International Postgraduate Research Scholarship. The authors wish to thank Dr Linda Smith for providing advice on the small-plant bioassays and Fiona Banks for technical assistance.