• Open Access

Transgenic rice with inducible ethylene production exhibits broad-spectrum disease resistance to the fungal pathogens Magnaporthe oryzae and Rhizoctonia solani


Correspondence (Tel +(814) 867 0324; fax +(814) 863 7217; email yuy3@psu.edu)


Rice blast (Magnaporthe oryzae) and sheath blight (Rhizoctonia solani) are the two most devastating diseases of rice (Oryza sativa), and have severe impacts on crop yield and grain quality. Recent evidence suggests that ethylene (ET) may play a more prominent role than salicylic acid and jasmonic acid in mediating rice disease resistance. In this study, we attempt to genetically manipulate endogenous ET levels in rice for enhancing resistance to rice blast and sheath blight diseases. Transgenic lines with inducible production of ET were generated by expressing the rice ACS2 (1-aminocyclopropane-1-carboxylic acid synthase, a key enzyme of ET biosynthesis) transgene under control of a strong pathogen-inducible promoter. In comparison with the wild-type plant, the OsACS2-overexpression lines showed significantly increased levels of the OsACS2 transcripts, endogenous ET and defence gene expression, especially in response to pathogen infection. More importantly, the transgenic lines exhibited increased resistance to a field isolate of R. solani, as well as different races of M. oryzae. Assessment of the growth rate, generational time and seed production revealed little or no differences between wild type and transgenic lines. These results suggest that pathogen-inducible production of ET in transgenic rice can enhance resistance to necrotrophic and hemibiotrophic fungal pathogens without negatively impacting crop productivity.


Rice (Oryza sativa) is of utmost importance to the human population, as more than half of the global population is dependent on the crop for the majority of its food requirements. As the human population is expected to rise to about 9 billion by the year 2050, rice crop yields will need to at least double by that time (Skamnioti and Gurr, 2009). One of the major limiting factors to yield is the occurrence of diseases caused by various fungal, bacterial and viral pathogens. Rice blast, caused by the ascomycete hemibiotrophic fungus Magnaporthe oryzae, is the most devastating rice disease in the world and often results in yield loss as high as 30% (Skamnioti and Gurr, 2009). Besides rice, M. oryzae can infect other grass species, such as perennial ryegrass (Lolium perenne) causing grey leaf spot, and wheat (Triticum spp.) causing wheat blast. Host resistance to M. oryzae is conferred by both race-specific resistance (R) genes, as well as by nonrace specific resistance quantitative trait loci (QTLs). R gene-mediated resistance frequently leads to a rapid and complete inhibition of the pathogen colonisation; however, this resistance is narrow spectrum, meaning that each R gene only recognises pathogen races that carry the corresponding avirulence (Avr) gene. As a result, R gene-mediated resistance is prone to breakdown due to point mutations, deletion and/or recombination of Avr genes in the pathogen, which leads to disease susceptibility (Bonman, 1992; Dai et al., 2010). Besides R gene-mediated resistance, another type of genetic resistance worth elucidating in plants is known as broad-spectrum resistance. Broad-spectrum resistance is defined as resistance that is effective against two or more pathogen species, and/or many different races within one pathogen species (Kou and Wang, 2010; Wisser et al., 2005).

Sheath blight, caused by the fungus Rhizoctonia solani anastomosis group 1-IA, is the second-most devastating disease of rice, with yield loss between 10% and 25% (Banniza and Holderness, 2001). Unlike the case of M. oryzae, there are no known major R genes corresponding to R. solani, and resistance is conferred solely by the additive effect of nonrace-specific resistance QTL (Lee and Rush, 1983; Li et al., 1995; Liu et al., 2009; Pinson et al., 2005). Resistance QTL is thought to confer a variety of traits, including components of basal resistance, developmental or morphological phenotypes that are not conducive to infection, production of antimicrobial compounds by the plant (phytoalexins and phytoanticipins), hormonal production and signalling, or other types of defence mechanisms (Poland et al., 2008). Due to the specific and sometimes transient nature of R-gene-mediated resistance, along with the lack of known R genes for pathogens such as R. solani, it is important to study the more broad spectrum and quantitative types of host resistance.

The phytohormone ethylene (ET) is a small, gaseous molecule that plays numerous roles in plant growth, development and response to environmental stresses. Depending on the plant–pathogen combination and specific environmental conditions, ET may act as a positive or negative modulator of disease resistance (Broekaert et al., 2006; Geraats et al., 2003; Hoffman et al., 1999; van Loon et al., 2006). It is known that partial submergence of rice plants results in the biosynthesis and physical entrapment of ET within the hollow aerenchyma tissue (Steffens and Sauter, 2009). In paddy fields, enhanced resistance to M. oryzae infection was observed in rice plants grown under flood or anaerobic conditions (Lai et al., 1999). Singh et al. (2004) proposed that the flood or hypoxia-induced ET biosynthesis in rice is critical for mediating horizontal resistance to blast infection. They demonstrated that application of AVG (aminoethoxyvinylglycine hydrochloride, an ET biosynthesis inhibitor) increased blast disease severity and negated the flood-induced resistance in rice plants. By contrast, application of ethephon (2-chloroethylphosphonic acid, an ET generator) significantly enhanced rice blast resistance in disease susceptible cultivars (Singh et al., 2004). Exogenous ET is also known to induce pathogenesis-related (PR) genes such as PR1, PR5 and PR10 in rice plants (Agrawal et al., 2003). Besides M. oryzae, ET has also been implicated in Medicago resistance to R. solani. Overexpression of an ET response factor (MtERF1-1) in the roots of Medicago truncatula resulted in enhanced resistance to Rsolani AG8, whereas an ethylene-insensitive ein2 (sickle) mutant was highly susceptible to the fungus (Anderson and Singh, 2011; Anderson et al., 2010; Penmetsa et al., 2008).

In the ET biosynthetic pathway, the conversion of S-adenosyl-l-methionine (AdoMet) to 1-aminocyclopropane-1-carboxylic acid (ACC) is the first committed step, which is catalysed by ACC synthase (ACS; Chae and Kieber, 2005). In higher plants, ACS genes are encoded by a multigene family, with nine members in Arabidopsis and six members in rice. Differential expression of individual ACS genes has been demonstrated in various plants to control specific aspects of plant development, maintenance and response to environmental cues. For example, Arabidopsis ACS2 and ACS6 are expressed after wounding and Pseudomonas syringae inoculation (Liang et al., 1996). Tomato ACS2, ACS4 and ACS6 play differential roles in fruit development and ripening (Alexander and Grierson, 2002), and ACS2 and ACS6 are additionally induced by ozone stress (Moeder et al., 2002). Maize ACS6 is responsible for leaf senescence under normal and drought conditions (Young et al., 2004). During the rice–M. oryzae interaction, endogenous ET levels increased within 48 h after inoculation with either avirulent or virulent isolates, with a significantly higher production of ET in the incompatible Pii R gene-mediated interaction (Iwai et al., 2006). Whereas OsACS3 and OsACS4 were expressed constitutively, OsACS1 and OsACS2 were significantly induced upon M. oryzae infection, along with the induction of an ACC oxidase (ACO) gene, OsACO7. In a follow-up study, silencing of OsACS2 and OsACO7 by RNA interference (RNAi) resulted in increased susceptibility to rice blast (Seo et al., 2010), suggesting that OsACS2 and ET production play a positive role in rice resistance to M. oryzae infection.

In this study, we generated transgenic rice with inducible overproduction of ET by placing the ET biosynthetic gene OsACS2 under the control of a strong, pathogen-inducible PBZ1 promoter. Molecular, physiological and pathological analyses reveal that inducible overexpression of OsACS2 in transgenic rice lines results in an induction of ET production and PR gene expression as well as enhanced resistance to both rice blast and sheath blight pathogens.


Generation and verification of transgenic OsACS2 overexpression lines

Agrobacterium-mediated transformation of the PBZ1::OsACS2 construct yielded 22 independent T0 lines, with 55 plants total. Five-week-old T0 plants were analysed for basal levels of OsACS2 expression through quantitative real-time PCR (qRT-PCR), using a pair of OsACS2-specific primers. The resulting values were variable due to the nature of the stress-inducible PBZ1 promoter. However, at least five independent lines showed a higher than threefold induction of OsACS2 mRNA as compared with nontransformed cv. Kitaake lines (Figure 1b). The basal ET levels of these five transgenic lines were also higher than that of the wild-type plant (Figure 1c), which was consistent with the qRT-PCR data. Three independent lines (OX-7, OX-8 and OX-20) were chosen to advance to the T1 and T2 generations for further analysis based on the high-level inducibility of the OsACS2 transgene and ET production.

Figure 1.

Generation and verification of OsACS2-overexpression lines. (a) The PBZ1 promoter::OsACS2 construct. (b) Relative quantity of OsACS2 mRNA in leaves of wild type (WT) and T0 overexpression (OX) lines. (c) Ethylene production in foliar tissues of WT and T0 transgenic lines. The data were averaged from three leaf replicates with standard error. Experiments were conducted twice with similar results.

Transgenic lines show inducible overexpression of OsACS2 and increased production of ethylene

Upon comparison of the dH2O- and benzothiodiazole (BTH)-treated plants, the dH2O-treated OsACS2-OX lines OX-7 and OX-20 showed no significant change in the expression of OsACS2 (Figure 2a), which can be explained by the lack of induction of PBZ1 promoter under optimal growth conditions in absence of environmental stress. However, OX-8 showed a slight, but not significant increase of OsACS2. In contrast, the OsACS2-OX plants treated with BTH showed a drastic increase of OsACS2 mRNA over the wild-type Kitaake (F = 19.66; P = 0.018). A second set of water- and BTH-treated lines were harvested and used to measure the ET levels (Figure 2b). Interestingly, all water-treated OsACS2-OX lines showed significantly higher production of ET as compared with wild type (WT) Kitaake, despite lack of induction of OsACS2 (F = 103.99; P < 0.0001). Basal ET levels of OX-8 were the highest among the OX and WT lines, which mirrored the slightly elevated levels of OsACS2 mRNA of that same line. This pattern was amplified in the BTH-treated lines, as all OX lines showed a significant increase in ET production, with OX-8 containing the highest level of ET. The increase in both OsACS2 expression and ET production in BTH-treated T2 lines shows a successful introduction of a functional and strongly inducible OsACS2 transgene that is stably maintained through generations.

Figure 2.

Induction of the OsACS2 transgene and ethylene production in T2 homozygous lines. (a) Relative quantity of OsACS2 mRNA in wild type (WT) and OsACS2-OX lines at 24 h post-treatment with either water or 0.25 mm benzothiodiazole (BTH). (b) Ethylene production in WT and OsACS2-OX lines at 24 h post-treatment with either water or 0.25 mm BTH. (c) Relative quantity of OsACS2 mRNA in WT and OsACS2-OX lines at 72 h postinoculation with either water or Magnaporthe oryzae isolate IC17-18/1. (d) Ethylene production at 0, 24, 48 and 72 h postinoculation with M. oryzae isolate IC17-18/1. Experiments were conducted three times, and the data were averaged from three independent replicates with standard error.

To examine the patterns of OsACS2 expression and ET production in OsACS2-OX lines after pathogen challenge, WT cv. Kitaake and OsACS2-OX lines were inoculated with M. oryzae isolate IC17-18/1. OsACS2 transcript levels were measured using qRT-PCR (Figure 2c). Similar to that of the BTH-treatment experiment, OsACS2 was not significantly expressed in the OsACS2-OX lines as compared with the WT cv. Kitaake under basal conditions. At 72 h postinoculation, levels of OsACS2 mRNA in the OsACS2-OX lines were significantly higher than that of the WT cv. Kitaake (F = 14.12; P = 0.014). This result indicates that the OsACS2 transgene is induced to a higher degree than the endogenous OsACS2 transcripts after M. oryzae infection. The production of ET in OsACS2-OX and WT cv. Kitaake lines was measured at 0, 24, 48 and 72 h postinoculation by M. oryzae isolate IC17-18/1 (Figure 2d). The OX-8 and OX-20 lines showed a marked increase in ET production over WT cv. Kitaake at all four time points; however, the kinetics of ET production were slightly different. The ET production of OX-20 peaked earlier (24 h postinoculation) than that of OX-8 (48 h postinoculation). The ET production of OX-7 was similar to that of WT cv. Kitaake, with the exception of 48 h postinoculation, in which OX-7 line produced a higher amount of ET than both cv. Kitaake and OX-20. Despite the differences in the kinetics of ET production, the OsACS2-OX lines produced significantly more ET than that of WT cv. Kitaake lines (F = 10.79; P < 0.0001) after inoculation with M. oryzae.

Increased expression of PR genes in OsACS2 overexpression lines

Quantitative real-time PCR was used to measure the basal levels of OsPR1b and OsPR5 transcripts in both 5-week-old T0 and 2-week-old T2 transgenic plants. OsPR1b was shown to be significantly induced in all three OsACS2-OX lines as compared with the nontransformed control, ranging from 10-fold to about 60-fold increase in both generations (Figure 3a). Likewise, all three OsACS2-OX lines also showed significantly higher expression of OsPR5 in both generations, however, to a smaller degree than that of OsPR1b with fold increases from 2.0 to 7.9 in the T0 generation, and 2.3–4.2 in the T2 generation (Figure 3b). These results show that OsACS2-OX lines display higher basal expression of PR genes, which may imply higher levels of host resistance.

Figure 3.

Enhanced expression of PR genes in OsACS2-overexpression lines. (a) Relative expression of OsPR1b in T0 (top) and T2 (bottom) lines. (b) Relative expression of OsPR5 in T0 (top) and T2 (bottom) lines. Experiments were conducted once in T0 lines and three times in T2 lines. The T0 data are the average of three technical replicates; T2 data are the average of three independent biological replicates with standard error.

OsACS2 overexpression lines show increased resistance to different races of Magnaporthe oryzae

To evaluate the disease resistance of transgenic rice lines, two isolates of M. oryzae with differing degrees of virulence were spray-inoculated onto OsACS2-OX lines and wild-type Kitaake. Disease severity was assessed through counting the number of lesions per leaf (an indicator of disease incidence) as well as measuring the lesion size (an indicator of disease severity). The first isolate tested was M. oryzae IC17-18/1, which is moderately virulent on cv. Kitaake. The OsACS2-OX lines all showed a significant reduction in lesion number (F = 25.7; P < 0.0001), (Figure 4a,c); however, it was noted that the overexpression lines OX-8 and OX-20 showed a more significant reduction in lesion number than OX-7. All three overexpression lines also showed a significant reduction in lesion size (F = 21.42; P < 0.0001) (Figure 4b,c). A second M. oryzae isolate, IE1K-FN9, is highly virulent on cv. Kitaake and has the ability to overcome the Pi-ta resistance gene widely deployed in the US rice cultivars. Similar to the results from the previous IC17-18/1 inoculation, the OsACS2-OX lines showed significant reductions in both lesion number (F = 23.9; P < 0.0001) and lesion size (F = 29.63; P < 0.0001) (Figure 4d–f). It was noted that OX-8 showed the most significant reduction in lesion number as compared with OX-7 and OX-20; however, all three overexpression lines were significantly more resistant as compared with cv. Kitaake. These results suggest that increased resistance conferred by overexpression of OsACS2 is not limited to single races of M. oryzae.

Figure 4.

Increased resistance of OsACS2-overexpression lines to moderately virulent (IC17) and highly virulent (IE1K) isolates of Magnaporthe oryzae. (a) Lesion number per leaf after inoculation with Moryzae IC17-18/1. (b) Lesion size after inoculation with M. oryzae IC17-18/1. (c) Rice blast symptoms on wild type (WT) and OsACS2-overexpression lines inoculated with M. oryzae IC17-18/1. (d) Lesion number per leaf after inoculation with M. oryzae IE1K-FN9. (e) Lesion size after inoculation with M. oryzae IE1K-FN9. (f) Rice blast symptoms on WT and OsACS2-overexpression lines inoculated with M. oryzae IE1K-FN9. About 20 plants per line were evaluated in each experiment, and the data were averaged from three independent replicates. Letters represent each significance groups, determined through one-way analysis of variance.

OsACS2-overexpression lines show increased resistance to Rhizoctonia solani

After establishing that overexpression of OsACS2 confers resistance to M. oryzae in a nonrace-specific manner, the question arose if this resistance could protect against a completely different species of fungus. The OsACS2-OX lines and wild-type control were inoculated with R. solani field isolate RR0140 (Wamishe et al., 2007) using the mycelia ball method (Park et al., 2008). The OsACS2-OX lines showed a 35%–45% reduction in lesion size compared with nontransformed cv. Kitaake (F = 19.8; P < 0.0001) (Figure 5a,b). These results demonstrate that OsACS2-overexpression lines are not only more resistant to the hemibiotrophic rice blast fungus, but also exhibit enhanced resistance to the necrotrophic sheath blight pathogen.

Figure 5.

Increased resistance of OsACS2-overexpression lines to Rhizoctonia solani isolate RR0140. (a) Lesion length on the wild-type plant and OsACS2-OX lines. (b) Comparison of sheath blight symptoms between wild type and OsACS2-OX lines. At least eight plants per line were evaluated in each experiment, and the data represent the average of three independent replicates with standard error. Letters represent different significance groups according to one-way analysis of variance.

Inducible overexpression of OsACS2 does not negatively affect seed production and agronomic traits

Regardless of the strategies and genes used to genetically improve crop cultivars, one of the most important aspects to check is the potential effect of transgene on plant growth and grain production. To gauge the effect of inducible overexpression of OsACS2 on these agronomically important traits, twelve plants from each transgenic lines were grown in glasshouse conditions over two different seasons, and the growth rate, time until maturity and seed production were measured. There were no differences between WT cv. Kitaake and OsACS2-OX lines until 5 weeks postgermination, which was close to the end of the vegetative stage (approximately 25–30 cm). After 6 weeks, the OX-7 and OX-20 lines grew slightly slower than WT cv. Kitaake and were about 7–10 cm shorter at maturity (66 and 63 cm, respectively, compared with 74 cm for Kitaake). Interestingly, OX-8 line grew significantly slower than nontransformed and other transformed lines between 5 and 8 weeks postgermination, but underwent rapid growth after week 8 and reached a similar height to OX-7 and OX-20 at maturity (Figure 6a).

Figure 6.

Evaluation of plant growth and development as well as grain production of OsACS2-overexpression lines. (a) Plant height during rice growth and maturation. (b) Time required to reach each growth and reproductive stage. (c) The number of panicles of each plant at maturity. (d) The number of seeds per panicle. (e) The weight per 100 seeds. At least 12 plants per line were evaluated in each experiment, and the data represent the average of two independent replicates with standard error.

To assess the maturation rate of the OsACS2-OX lines, the time elapsed until each growth stage was recorded for each individual plant. All plants reached the booting stage at weeks 6–7, heading at weeks 7–8 and flowering between weeks 7 and 9. OX-7 and OX-20 plants reached each reproductive stage in the shortest time, flowering about 0.5 weeks earlier than the wild-type plant. Similar to the growth rate data, OX-8 reached each stage latest, 1 week behind OX-7 and OX-20 and 0.5 weeks later than Kitaake. However, in the ripening phase, OX-8 underwent rapid grain-filling and ripening as compared with all other lines, and reached maturity between 12 and 13 weeks, similar to OX-7 and cv. Kitaake. OX-20 matured in the shortest time, by 12 weeks postgermination (Figure 6b).

To analyse yield-related components, the number of panicles per plant, the number of seeds per panicle and the weight of 100 seeds were measured (Figure 6c–e). There were no significant differences for each of these parameters between WT and OsACS2-OX lines; however, OX-7 had a slightly higher number of panicles (average of 3.6 per plant, compared with 3 per plant for Kitaake, OX-8 and OX-20) as well as number of seeds per panicle (56 seeds, as compared to approximately 44 seeds in Kitaake, OX-8 and OX-20). For all lines, the weight per 100 seeds was about 2.7 g, with little variation. From this data, we can conclude that inducible overexpression of OsACS2 has no negative effect on these agronomically important traits, and may even promote slightly more rapid maturation and yield in a glasshouse setting.


The development of crop cultivars through transgenic methods is becoming increasingly common, due to advances in gene discovery and overall improvement in the efficiency and efficacy of stable transformation methods. The use of transgenic approaches to improve host resistance is beneficial, as it mitigates the need for pesticides and is more efficient than conventional breeding methods (Gust et al., 2010). Breeding strategies for host genetic resistance have traditionally focused on using major R genes to ensure a complete, efficient form of resistance. The drawbacks to this method are that the majority of R genes only protect against a narrow spectrum of pathogens, and are prone to breakdown due to introductions of different isolates, or mutations within pathogen populations. For example, breakdown of rice blast R gene Pi-ta occurs due to the unstable telomeric location of the corresponding Avr Pita gene in M. oryzae (Dai et al., 2010; Orbach et al., 2000; Zhou et al., 2007). For this reason, it is important to discover other, less specialised sources of host genetic resistance. To date, transgenic modification or altered expression of many rice genes has been shown to increase host resistance to different fungal and bacterial pathogens. These include receptors OsWAK1 (Li et al., 2009), OsSERK1 (Hu et al., 2005) and OsBRR1 (Peng et al., 2009), salicylic acid (SA) signalling component AtNPR1 and its rice homologue OsNH1 (Chern et al., 2001, 2005), jasmonic acid (JA) biosynthetic gene OsAOS2 (Mei et al., 2006), transcription factors OsWRKY13 (Qiu et al., 2007), OsWRKY45 (Shimono et al., 2007) and OsWRKY71 (Liu et al., 2007), and PR genes such as PR5 (Datta et al., 1999), among many others.

Plant hormones such as SA, JA and ET play diverse roles in mediating defence signalling and disease resistance responses (Grant and Jones, 2009; Lopez et al., 2008; Robert-Seilaniantz et al., 2007). These hormones often invoke the activation of defence-related kinases, transcription factors and PR genes, many of which have been found to improve host resistance in transgenic rice (reviewed in Delteil et al., 2010). SA is a phenolic compound that has been shown to be involved in the activation of systemic acquired resistance in many dicotyledonous plants (Durrant and Dong, 2004; Park et al., 2007). Because rice plants contain a high basal level of SA, which is not significantly induced by pathogen infection, SA may not serve as an effective defence signal in rice (Silverman et al., 1995; Yang et al., 2004). However, endogenous SA was shown to protect rice plants from oxidative damage caused by biotic and abiotic stresses and to mediate host resistance in adult plants (Iwai et al., 2007; Yang et al., 2004). Furthermore, SA signalling components appear to be functional in rice because overexpression of the rice NPR1 homologue, OsNH1, resulted in increased PR gene expression and enhanced resistance to X. oryzae pv. oryzae (Chern et al., 2005). Possibly due to high basal level of SA in rice, which may inhibit jasmonate biosynthesis, JA levels increase only slightly after M. oryzae infection (Mei et al., 2006). However, JA was shown to be capable of mediating defence signalling and disease resistance in rice. Application of exogenous JA or methyl JA (MeJA) activates the expression of several PR genes, along with an increased level of phytoalexins in rice (Tamogami et al., 1997). In addition, overexpression of the jasmonate biosynthetic gene allene oxide synthase 2 (OsAOS2) resulted in increased JA production and PR gene expression as well as enhanced resistance to M. oryzae (Mei et al., 2006). Recently, increasing evidence from our lab and other groups suggests that ET may play a more prominent role than SA and JA in rice disease resistance (Bailey et al., 2009; Seo et al., 2010; Singh et al., 2004). Applications of ethephon decreased the incidence of rice blast in field conditions (Singh et al., 2004) and triggered PR gene expression in rice cell cultures (Agrawal et al., 2003). More importantly, inoculation of rice by M. oryzae triggers induction of ET biosynthetic genes (Iwai et al., 2006), and knockdown of these genes results in increased susceptibility to M. oryzae (Seo et al., 2010). Therefore, it is imperative to explore the transgenic approach for improving rice disease resistance based on the genetic modification for enhanced ET biosynthesis.

In this study, we created transgenic rice exhibiting inducible overproduction of ET, by placing the ET biosynthetic gene OsACS2 under control of the PBZ1 promoter. This pathogen-inducible promoter was chosen for the purpose of fine-tuning the host response to a pathogen, as opposed to a constitutive promoter, as continuous overexpression of a transgene may potentially have a negative effect on other aspects of plant health, yield or tolerance to other environmental stresses (Brown, 2002; Kim et al., 2009; Potenza et al., 2004). OsACS2 was chosen to overexpress because this gene is significantly induced by M. oryzae and closely associated with ET-mediated rice defence response (Iwai et al., 2006). There is some variability in induction of OsACS2 in the OsACS2-OX lines, particularly in the elevated basal expression of OsACS2 in OX-8 (as seen in Figures 2a,b). This might be caused by position effects where nearby cis-elements at the transgene insertion site in the OX-8 line resulted in elevated basal expression of OsACS2. Despite this, both OsACS2 transcripts and ET levels were greatly increased in T0 and T2 generations after BTH- and pathogen-activation of the PBZ1 promoter as compared with basal conditions. Therefore, we can conclude that the OsACS2 transgene is inducible, functional and stable through multiple generations.

Inoculation of OsACS2-overexpression lines with a moderately virulent isolate of M. oryzae (IC17-18/1) showed about a 50% reduction in both lesion number (incidence) and lesion size (severity), along with a significant increase in ET production at 48 and 72 h postinoculation. It is important to note that disease symptoms do appear on OsACS2-overexpression lines; however, blast lesions occur later and are smaller, and the overall disease severity is lower than that of nontransformed lines, classifying this as more of a partial resistance. We also inoculated the OsACS2-overexpression lines with a highly virulent isolate, IE1K-FN9, in order to gauge the efficacy of ET-mediated resistance against a more pathogenic isolate. Infection by IE1K-FN9 resulted in many lesions and extensive damage on leaves of nontransformed cv. Kitaake, but much fewer lesions on OsACS2-overexpression lines. Interestingly, the use of an unrelated, necrotrophic fungal pathogen, R. solani, showed the same trend of a significant reduction in lesion size, but not a complete inhibition of the pathogen growth. This is consistent with our recent observation that transgenic suppression of ET biosynthesis gene via RNAi or treatment of ET biosynthetic inhibitor aminooxyacetic acid (AOA) resulted in increased susceptibility to Rsolani (Yang lab, unpubl. data). Taken together, our results suggest that ET-mediated disease resistance may be broad spectrum in nature, and is potentially more durable than R gene-mediated resistance.

The data derived from this study show that increased ET production enhances host resistance in rice against M. oryzae and R. solani. However, the mechanism of ET-mediated resistance is still not clear. A simple explanation is that it can be attributed to the enhanced expression of PR genes at the end of the ET signalling pathway. It has been shown that application of exogenous ET results in the activation of rice PR genes with GCC-box-containing promoters, including OsPR1b and OsPR5 (Agrawal et al., 2003). Therefore, inducible overproduction of endogenous ET may lead to increased expression of PR genes in the transgenic lines. Indeed, the OsACS2-overexpression lines were found to have increased expression of both OsPR1b and OsPR5. The higher expression of OsPR5 in the T0 generation could tentatively be explained by the increased plant age. Aged tissues often show many hallmarks of stress-related responses, including elevated SA levels, production of reactive oxygen species and PR gene expression (Kus et al., 2002; Wyatt et al., 1991). It is likely that in the older T0 plants, the PBZ1 promoter is active at a low level, which increases expression of the OsACS2 transgene.

Another possible explanation for ethylene-mediated resistance is that the increased ET production is affecting other hormonal pathways such as the abscisic acid (ABA) level and/or signalling. Under submerged conditions, ET was shown to induce ABA 8′-hydroxylase, which oxidises ABA to its inactive form, phaseic acid (Saika et al., 2007). Such a reduction of active ABA may lead to enhanced rice blast resistance because ABA appears to positively regulate rice susceptibility to M. oryzae infection (Bailey et al., 2009; Jiang et al., 2010; Koga et al., 2004). By contrast, suppression of an ABA-inducible MAP kinase down-regulated the ABA pathway, but up-regulated the ET pathway, leading to enhanced rice blast resistance (Bailey et al., 2009; Xiong and Yang, 2003). Therefore, it is also possible that ET indirectly promotes host resistance to rice blast by suppressing the ABA level and/or signalling.

An alternative explanation is that the increased resistance is mediated not by ET, but by cyanide. Cyanide is a by-product of the ET biosynthetic pathway and is produced in equal amounts to ET (Peiser et al., 1984). Seo et al. (2010) hypothesised that Pii-mediated resistance was due to increased cyanide production after demonstrating that exogenous application of either cyanide or the ET precursor ACC complemented ET-deficient rice lines, but ethephon failed to. However, this is contradictory to the results published by Singh et al. (2004), which demonstrated a marked reduction in blast disease incidence in several cultivars after ethephon application. In addition, results from our lab indicated that ET-insensitive rice, which is not deficient in ET or cyanide levels, has increased susceptibility to M. oryzae (Bailey et al., 2009; Yang lab, unpubl. data). Due to these contradictory results, additional studies are necessary to clarify the roles of ET biosynthesis, cyanide production, ET signalling and other factors involved in ET-mediated rice disease resistance.

In summary, we have developed transgenic rice lines with pathogen-inducible overexpression of the ET biosynthetic gene OsACS2. These lines show inducible overproduction of ET and enhanced resistance to the fungal diseases rice blast and sheath blight, without negatively impacting agronomically important traits such as yield in glasshouse conditions. Future work will focus on the field applications of OsACS2-OX lines, in terms of agronomic traits along with the efficacy against a greater diversity of pathogens. In addition, the resistance of OsACS2-overexpression lines to multiple fungal pathogens suggests that enhancement of ET-mediated resistance may be applicable in other monocot species for controlling devastating fungal diseases such as wheat blast and turfgrass grey leaf spot.

Materials and methods

Gene construct and rice transformation

The full length cDNA sequence was amplified by PCR using rice OsACS2 cDNA (AK064250) as a template with a pair of specific primers containing KpnI or SalI restriction sites (Forward primer: 5′ ACG GTA CCA TGG CGT ACC AGG GCA TCG AC 3′; Reverse primer: 5′ ACC GTC GAG TCT GCT GGC TTA ATC AGC TG 3′). After restriction digestion, the PCR fragment was inserted into the modified vector pCAMBIA1300P, which contains a pathogen-inducible rice PBZ1 promoter (Lee, 2002). The resulting PBZ1::OsACS2 construct was then transformed into Agrobacterium tumefaciens strain EHA105 through electroporation. The Agrobacterium-mediated rice transformation was carried out using calli derived from mature seeds of cultivar Kitaake according to a previously described protocol (Hiei et al., 1994; Mei et al., 2006).

Plant materials and growth conditions

Seeds of both nontransformed cultivar Kitaake and self-pollinated T1 OsACS2-OX lines were germinated in water at 37 °C for 2 days before planting in MetroMix 360 soil (Sun Gro, Bellevue, WA). Rice plants were maintained in glasshouse with 12 h of light at 28 °C day and 24 °C night, respectively. Seven days postgermination, seedlings were fertilised with 0.5% ammonia sulphate and 0.1% Sprint iron solution. To select hygromycin-resistant transgenic seedlings from segregating progeny, leaf segments were placed in a 25 μg/mL hygromycin solution, and checked 2 days later for hygromycin resistance. Homozygous T2 OsACS2-OX lines were selected based on the transgene segregation, and further propagated for subsequent experimental studies.

Chemical treatments

To test the stability and inducibility of the transgene, 2-week-old rice seedlings of WT cv. Kitaake and the homozygous T2 OsACS2-OX lines were either sprayed with dH2O (basal) or 0.25 mm BTH. BTH has been previously shown to effectively induce the PBZ1 promoter (Lee et al., 2001). Twenty-four hrs after the treatment, rice leaves were collected for the ET measurement as well as the RNA extraction and qRT-PCR assays.

Fungal isolates and disease assays

The two field isolates of M. oryzae used in this study were IC17-18/1 and IE1K-FN9 (Zhou et al., 2007). IC17-18/1 is moderately virulent on cv. Kitaake and its infection results in necrotic lesions without causing full leaf senescence. IEK1-FN9 is highly virulent on cv. Kitaake, by producing a higher number of lesions and causing extensive leaf damage and sometimes full leaf senescence. Both M. oryzae isolates were maintained on oatmeal agar for 7–10 days before inoculation. Two-week-old plants at a two to three leaf stage were spray-inoculated with 2.5 × 105 conidia/mL suspension plus 0.1% Tween-20 until runoff occurred. Inoculated plants were then maintained under moist conditions at room temperature for 24 h, before moving to a growth chamber (28 °C day/24 °C night, 12 h light). Disease severity was scored 7 days postinoculation by counting the number of lesions, as well as measuring the length of the three largest lesions per leaf. In contrast to small reddish lesions associated with hypersensitive response, the susceptible interaction typically results in expanding lesions with grey and necrotic centres.

Sheath blight inoculation was performed using a field isolate (RR0140) of R. solani according to the method described by Park et al. (2008). Rhizoctonia solani mycelia (maintained on potato dextrose agar at 28 °C) were inoculated into 250 mL of potato dextrose broth and incubated on a 28 °C shaker for 7 days. Excess broth was strained out, and mycelia were separated into 5-mm-diameter balls. Each mycelial ball was secured against the sheath of 6-week-old plants by aluminium foil, which was removed once disease symptom initiated at 2 days postinoculation. Disease severity was evaluated 7 days postinoculation by measuring the length of each lesion.

RNA extraction and cDNA synthesis

Rice leaf tissues were snap-frozen in liquid nitrogen, and total RNA was extracted from 100 mg of frozen tissues using the TRIzol reagent (Invitrogen, Carlsbad, CA). RNA pellets were washed with 70% ethanol and resuspended in sterile deionized water pretreated with diethylpyrocarbonate (DEPC). Each RNA sample was treated with DNase I (New England Biolabs, Ipswich, MA) to remove traces of genomic DNA. cDNA synthesis was carried out using the High Capacity cDNA Reverse Transcription Kit (Applied Biosystems, Foster City, CA) according to the manufacturer's protocol.

Quantitative real-time polymerase chain reaction

All qRT-PCR was performed using a Step One Plus Real-Time PCR system (Applied Biosystems) with DyNAmo SYBR green PCR kit (New England Biolabs) according to the manufacturer's instructions. PCR cycling conditions included a DNA denaturing stage of 94 °C for 15 min, followed by 40 cycles of 94 °C for 30 s, 60 °C for 45 s and 72 °C for 45 s. The specific primer were synthesised for detecting transcripts of OsACS2 (forward primer: 5′ TGC GCC TTA CTA CGT CGA CTA CAT 3′; reverse primer: 5′ ACG CAC TAA CGC ACG TCT CTA CAA 3′; Accession number AK06425), OsPR1b (forward primer 5′ATT TAT TCG AGC GCC ACA TGA CGG 3′; reverse primer: 5′ GAC GAG TGG TCA AAC ATT GCA AGC 3′; Accession number U89895), and OsPR5 (forward primer: 5′ TAC AAC GTC GCC ATG AGC TTC T 3′; reverse primer: 5′ TGG GCA GAA GAC GAC TTG GTA GTT 3′; Accession number X68197). Relative expression data were normalised using the rice ubiquitin 1 gene (forward primer: 5′ TGG TCA GTA ATC AGC CAG TTT G 3′; reverse primer: 5′CAA ATA CTT GAC GAA CAG AGG C 3′).

Ethylene measurement

Ethylene production in rice leaves was quantified using a gas chromatograph (Hewlett-Packard 6890, Hewlett Packard, Palo Alto, CA). Rice leaves were excised from treated or untreated plants, weighted and incubated in a sealed 4-mL glass vial for 2–3 h. Three 1-mL air samples were injected per plant sample at specific time points. ET amounts were measured in parts per million (ppm) based on an ET standard and converted to nL/g/h.

Assessment of agronomic traits

The growth rate, time until maturation, and yield were assessed using twelve plants from each OsACS2-OX line and nontransformed cv. Kitaake in glasshouse conditions (28 °C, 12 h light). Two independent biological replicates were performed in two different seasons. The growth rate was calculated by measuring each plant once per week from soil surface to meristem, and the rice growth stage was noted according to the standard published by the Rice Knowledge Bank (www.knowledgebank.irri.org). Yield was assessed by counting the number of panicles per plant as well as the number of seeds per panicle, and weighing seeds in three separate groups of 100 for each line.


This work was supported by the grants from USDA/NRI (2008-35301-19028) and NSF Plant Genome Research Programme (DBI-0922747). We would like to thank Dr. Kathleen Brown and Ms. Amy Monko for help with the gas chromatography, Dr. Germán Sandoya for help with statistical analyses, and the Rice Genome Resource Centre in Japan for providing the OsACS2 cDNA clone.