The goal of this work was to establish which enzymes – peroxidases or NADPH oxidases – play the most important role in the resistance-related oxidative burst response of wheat to infection by brown rust (Puccinia triticina). The expression of four peroxidases and two NADPH oxidases was analysed in the susceptible wheat cv. Thatcher and isogenic lines with different Lr resistance genes after pathogen inoculation. Of the peroxidases, TaPrx118 and TaPrx112 were induced several times more strongly than TaPrx103 and TaPrx107. The induction of peroxidases was more pronounced than that of NADPH oxidases. The patterns of peroxidase expression clearly differentiated moderately resistant from highly resistant lines and corresponded to oxidative response profiles. The possible involvement of peroxidases or NADPH oxidases was verified with enzyme-specific inhibitors. The oxidative burst in the susceptible cv. Thatcher and in the lines TcLr24, TcLr25, TcLr9 was peroxidase-dependent, while the response in line TcLr26 was NADPH-oxidase-dependent. It is postulated that class III peroxidases play a leading role in the formation of reactive oxygen species molecules during the response of wheat to pathogen infection. The results suggest a high level of redundancy of some peroxidase genes induced in biotic stress. The role of both enzyme systems in wheat response/resistance to brown rust is discussed in relation to the oxidative response, the efficiency of resistance, and the presence and origin of particular Lr resistance genes.
An oxidative response is an early and complex reaction induced by biotic stress. It is already well established that reactive oxygen and nitrogen species integrate a range of complementary defence mechanisms, although the details of those still remain to be determined. The molecules play diverse roles in plant–pathogen interactions. Because of their antimicrobial properties they act directly against the invading organism. They serve as substrates for synthesis of secondary metabolites or as components enforcing the physical barrier of the cell wall. Finally, they play important roles in the signalling that coordinates plant responses to biotic stress (Floryszak-Wieczorek et al., 2007; Torres, 2010). The plant–pathogen interaction is associated with an oxidative response detectable as the localized accumulation of reactive oxygen species (ROS) molecules in attacked and/or neighbouring cells of the infection site. The detailed characteristics of the process may vary depending on the particular plant, pathogen and type of interaction. The most common feature of the process is a biphasic induction of ROS accumulation. The first phase of this response is usually observed in both compatible and incompatible interactions, while the stronger, second phase, induced during an incompatible interaction, has a crucial role in activation of plant defence (Grant et al., 2000; Orczyk et al., 2010). Accumulation of ROS in cells surrounding the infection site is the result of a sharp increase of ROS-generating enzymes, which are cross-regulated with free radical scavenging systems (Nanda et al., 2010).
Apoplast-secreted peroxidases (Daudi et al., 2012) and membrane-bound NADPH oxidases (Bolwell et al., 2002) are among the most frequently studied enzymes in host–pathogen systems because they are expected to have significant roles in the plant oxidative burst.
Based on the cellular localization of the prevailing enzyme system, the reaction might be called apoplastic or symplastic. The class III peroxidases constitute an important part of ROS homeostasis. The enzymes, which are present only in green plants, are encoded by a large multigene family (Bakalovic et al., 2006). The class III peroxidases are secreted into the apoplastic space or the plant vacuole, depending on the presence of a C-terminal extension which serves as a vacuolar retention signal (Passardi et al., 2004). Class III peroxidases contribute to ROS scavenging by their peroxidative (or catalytic) activity and they can also generate superoxide radicals (O2−) via their oxidative cycle. The oxidative cycle is involved in building up high levels of ROS during the oxidative burst (Almagro et al., 2009). It is notable that the initial peroxidase activity requires the presence of H2O2 and the final outcome of the reaction (i.e. the elimination or the accumulation of ROS) depends on the type of activity cycle (Almagro et al., 2009). The main biological function of class III peroxidases is to oxidize phenolics, lignin precursors, auxins and secondary metabolites using hydrogen peroxide (Almagro et al., 2009). It was shown that H2O2 is generated in vivo by apoplastic peroxidases in French bean (Bolwell et al., 2002). The peroxidase-generated ROS function in cell wall enforcement (Barcelo, 1995) and in biotic stress signalling (Johrde & Schweizer, 2008). There are 73 peroxidase-encoding genes in Arabidopsis thaliana and 138 in rice. It is expected that the gene family encoding peroxidases in wheat is as large. Currently, the PeroxiBase database (http://peroxibase.toulouse.inra.fr/index.php) lists 110 class III wheat peroxidases. The biological functions of most of them remain unknown.
The membrane-bound NADPH oxidases are also called the respiratory burst oxidase homologues (Rboh) because they have been identified in plants based on their homology to mammalian respiratory burst oxidase or gp91phox proteins (reviewed by Bedard & Krause, 2007). The primary products of Rboh activity are superoxide radicals (O2−), which dismutate to H2O2 in a nonenzymatic reaction or in the superoxide-dismutase-catalysed reaction. The N-terminal region contains a Ca2+-binding EF-hand motif. This implicates calcium-dependent regulation which has been experimentally confirmed for AtRbohC in calcium-regulated root-hair formation (Foreman et al., 2003). Isolation and characterization of Rboh genes strongly indicated that at least some of them are functionally involved in the pathogen-induced oxidative response. Out of 10 A. thaliana AtRboh genes, two were involved in the response to pathogen infection: AtRbohD was responsible for extracellular ROS production and AtRbohF for the hypersensitive reaction (Torres et al., 2002). The six HvRboh genes identified in barley showed conserved genomic structure and the protein domains found earlier in A. thaliana and rice orthologues. Two of them, HvRBOHF1 and HvRBOHF2, were pathogen-induced (Lightfoot et al., 2008). It has been suggested that the biological role of NADPH oxidases during plant–pathogen interactions is more complex than that of peroxidases, and might vary between different types of interaction. Silencing of NtRbohA and NtRbohB in tobacco infected with Phytophthora infestans reduced the production of ROS and also depressed the tobacco non-host resistance (Asai et al., 2008). A similar effect of reducing ROS accumulation in elicitor-induced tobacco cells was the result of NtRbohD silencing (Simon-Plas et al., 2002). During compatible and incompatible interactions of A. thaliana with Pseudomonas syringae, both AtRbohD and AtRbohF participated in the oxidative burst and in regulation of the plant response (Chaouch et al., 2012). Torres et al. (2002) and Chaouch et al. (2012) reported that AtRbohF promoted A. thaliana cell death in the process which was found to be cross-regulated by salicylic acid. In A. thaliana infected with Alternaria brassicicola the knockout mutation of AtRbohD inhibited whole-cell ROS accumulation and affected pathogen-induced cell death (Pogány et al. ,2009) [Correction added on 19 November 2012 after online publication: The analysis on Pogany et al. 2009 has been corrected]. In monocots, HvRbohA silencing reduced the generation of superoxide radicals in cells infected with Blumeria graminis and lowered the rate of fungal penetration (Trujillo et al., 2006). According to the authors, the Rboh generated superoxide along with hydroxyl radicals, the result of the Fenton reaction, causing local polymer breakdown and cell wall softening, which facilitated easier fungal penetration. The effect of higher susceptibility to B. graminis was observed after HvRbohF2 silencing (Proels et al., 2010). Interestingly, silencing of this gene did not affect accumulation of ROS in the cells of the infection site (Proels et al., 2010). The same genes, HvRbohF1 and HvRboF2, were induced in barley after infection with necrotrophic Pyrenophora teres f. sp. teres and Rhynchosporium secalis pathogens (Lightfoot et al., 2008). Although beyond the scope of this article, it is worth noting that the activity of fungal NADPH oxidases is also an important player in the infection process (Egan et al., 2007). The accumulation of ROS in the response to pathogen infection can be the result of apoplastic and/or symplastic enzyme activities (Bindschedler et al., 2006). Activation of either or both of the systems depends on the type of plant–pathogen interaction. The apoplast-secreted peroxidases were found to be the main enzymatic system responsible for ROS generation in French bean cells (Bolwell et al., 2002) and in A. thaliana plants responding to the elicitor of Fusarium oxysporum (Bindschedler et al., 2006). Conversely, the activity of symplastic NADPH oxidases prevailed in the response of rose cells to F. oxysporum (Davies et al., 2006), while in another plant–pathogen system, A. thaliana infected with Ps. syringae, both types of enzyme were jointly involved in the oxidative reaction (Grant & Loake, 2000). The role of the symplastic Rboh system in A. thaliana was further confirmed by Grant et al. (2000). Davies et al. (2006) proposed that although both systems could operate in A. thaliana, the first phase of ROS accumulation could be peroxidase-dependent. This step might further activate the symplastic NADPH oxidases and the resistance-gene-mediated hypersensitive reaction. According to this interpretation the oxidative burst and the local necrosis observed at the infection sites were the result of both types of enzyme activity. The role of each type of enzyme can be investigated using enzyme-specific inhibitors. The use of diphenylene iodonium (DPI), which is an inhibitor of NADPH oxidases, experimentally confirms the role of NADPH oxidases in the plant oxidative response. This approach was used to confirm NADPH oxidase activity in the interaction of Medicago truncatula with Sinorhizobium meliloti (Cardenas et al., 2008). The use of DPI in A. thaliana infected with Ps. syringae blocked H2O2 accumulation at the sites of hypersensitive reaction but not at the sites of papillae formation (Soylu et al., 2005). The results indicated that ROS molecules involved in these reactions were of different origin. The formation of papillae, an important element of basal resistance, relied on DPI-insensitive apoplastic peroxidases, while the ROS molecules involved in the hypersensitive reaction were the result of NADPH oxidase activity. In another experiment, inoculation of A. thaliana with an avirulent Ps. syringae pv. tomato strain and the application of DPI (an NADPH oxidase inhibitor) or NaN3 (a Prx inhibitor) significantly decreased the oxidative response, which indicated that both enzyme systems were involved (Grant et al., 2000). An earlier study proved that the oxidative response of wheat to brown rust (Puccinia triticina) infection was biphasic, with the first step common to both susceptible and resistant plants, while the second phase was induced only in the resistant lines; the entire pattern of the response was closely correlated with the efficiency of resistance conferred by a specific Lr gene (Orczyk et al., 2010).
The goal of this work was to establish which type of enzyme activity (symplastic NADPH oxidase or apoplastic peroxidase) had the prevailing role in the oxidative burst in the wheat–brown rust pathosystem. Using a defined plant–pathogen system of wheat isogenic lines with the Lr resistance gene and a single-spore brown rust isolate with a defined virulence/avirulence formula, the role of peroxidases and NADPH oxidases in the generation of ROS in response to pathogen infection was analysed. The analysis of expression was verified using enzyme-specific inhibitors. The results are discussed in relation to the spatiotemporal pattern of ROS accumulation after brown rust infection, the biphasic mode of plant oxidative response and the origin of Lr resistance genes.
Materials and methods
Plant and pathogen material
Wheat cv. Thatcher, susceptible to brown rust, and isogenic Thatcher lines TcLr9, TcLr24 and TcLr26, each carrying a single brown rust resistance Lr gene, were used (McIntosh et al., 1995). The wheat seedlings were grown in a growth chamber at 22°C, 70–95% relative humidity, with a 16-h photoperiod at an illumination intensity of 250–300 μmol m−2 s−1. A single-spore isolate (7/2006) of P. triticina with established avirulence/virulence characteristics and the following infection types was established according to Roelfs & Martens (1988): Thatcher – 4, TcLr9 – 0, TcLr19 – 0; TcLr24 – 1, TcLr25 – 2, TcLr26 – 0, TcLr29 – 1, TcLr34 – 4. The primary leaves of 7-day-old seedlings were inoculated with urediniospores as described by Orczyk et al. (2010).
RNA isolation, reverse transcription and quantitative PCR
RNA was isolated from leaves collected 0, 8, 16, 24, 32, 48, 72, 96 and 120 h post-inoculation (hpi). The RNA was extracted from leaf samples ground in liquid N2 using TRI Reagent solution (Ambion) and RNA 3-zone (Novazym) according to the manufacturers' instructions. RNA concentration and A260/A280 ratio were measured using a NanoDrop spectrophotometer (NanoDrop Technologies). A260/A280 ratio was always higher than 1·8 and agarose gels were used to further determine the quality of obtained RNA. Isolated total RNA was treated with 1 U DNase (DNase I recombinant RNase-free; Roche) and 1 U Protector RNase inhibitor (Roche) per 1 μg RNA, for 15 min at 37°C. DNase was inactivated by addition of EDTA and heating at 75°C for 10 min according to the manufacturer's instructions. Complete digestion was confirmed by PCR with primers specific for wheat 18S rRNA (Table 1) using 200 ng DNase-digested RNA as a template. No product on agarose gels was detected after 40 cycles of amplification. One microgram of RNA (nondegraded, DNase-treated with undetectable gDNA impurities) was used as a template for the reverse transcription reaction with random oligomers (RevertAid FirstStrand cDNA synthesis kit; Fermentas). The obtained cDNA was diluted fivefold and used directly as a template for quantitative PCR (qPCR). The standard qPCR reaction mix was composed of 5 μL master mix (2 × Sso-Fast EvaGreen Supermix; Bio-Rad), 0·3 μL primer F (10 μm), 0·3 μL primer R (10 μm), cDNA and water to 3·7 μL. The reaction was performed in a Rotor-Gene 6000 model 5-plex thermocycler (Corbett). Template concentrations ranging from 102 to 1011 copies of analysed amplicon per reaction were used as the standards for qPCR. The efficiencies of amplification for all primer pairs were in the range 0·8 to 1·1. The specificity of amplification was verified by melting curve analysis and agarose gel electrophoresis. The resulting data were normalized to 18S rRNA using the two standard curves method. Relative quantification was performed with rotor-gene 6000 software v. 1.7. The real-time PCR experiments were performed using cDNA from three biological replications and with three technical repetitions each. The whole procedure of RNA isolation, reverse transcription, qPCR conditions and data analysis met the MIQE criteria outlined by Bustin et al. (2009).
The primers for TaPrx103 (TAPERO) peroxidase were designed based on the coding sequence X56011 (Rebmann et al., 1991). The primers for the TaPrx107 transcript were designed based on the coding sequence AJ878510. The primers for TaPrx108 were designed based on the coding sequence EU595582 (Simonetti et al., 2009). The gene TaPrx112 is represented by four forms: TaPrx112-A (EU595569), TaPrx112-C (EU567314), TaPrx112-D (EU595568) and TaPrx112-M (EU595567). The PCR primers were designed to amplify a 320-bp fragment common to the A, C, D and M forms. Wheat cDNAs representing NADPH oxidases were identified based on their homology to the known RbohD and RbohF genes. The wheat cDNA AK335454 is a homologue (with 72·2% identity) of AtRbohD of A. thaliana, the respiratory burst oxidase homologue protein C of Brachypodium distachyon LOC100826490 (with 93·3% identity) and Zea mays NADPH oxidase LOC100136880 (with 89·8% identity). Based on the above homology, AK335454 was used to design primers of TaRbohD transcript. The wheat cDNA AY561153 (UniGene Ta.31 600, wheat LOC543151) is a homologue (with 95·1% identity) of the predicted respiratory burst oxidase homologue protein F in B. distachyon, the respiratory burst oxidase AtRbohF of A. thaliana (with 81·2% identity), Z. mays NADPH-oxidase-encoding sequence LOC100194086 (92·8% identity) and rice NADPH-oxidase-encoding sequence Os01g0734200 (92·2% identity). Based on the above homologies, the coding sequence AY561153 was used to design primers for the wheat TaRbohF gene.
Primer sequences, cDNA accession numbers and PCR conditions are summarized in Table 1. The results of quantitative RT-PCR showed relative transcript accumulation of each of the analysed genes in relation to 18S rRNA (Fig. 1). The values on the vertical axis indicate the transcript ratio of the analysed gene in relation to 18S rRNA. They were calculated using Eqn (1)
where TR is the transcript ratio, and A and 18S are the number of template molecules per quantitative PCR reaction mix (10 μL) of the analysed gene (A) and 18S rRNA, respectively. The factor of 10−6 was used to adjust the scale. The transcription rate values were comparable for all combinations of genes versus lines. The expression diagrams (Fig. 1) indicate the relative level of transcript accumulation rate between 0 and 5 days post-inoculation (dpi).
In order to compare the possible impact of the expression of each of the analysed genes on the wheat oxidative response the cumulative value of transcript rate (CVTR) was introduced. The CVTR denoted the total transcript accumulation rate during 5 days of infection and was calculated as the definite integral of the function of the accumulation rate between 0 and 5 dpi according to Eqn (2):
The result represents the area below the graph of the accumulation rate delimited by time 0 and 5 dpi for each gene. Since the CVTR value takes into account both the transcript accumulation rate and the duration of expression, it allows direct comparison of all genes in tested lines. This makes it possible to evaluate the impact of genes' expression on the analysed process. The CVTR parameter was used to compare the impact of four tested peroxidases and two NADPH oxidases on the wheat oxidative response activated after pathogen infection.
Histochemical detection of H2O2 and visualization of pathogen structures
Leaf samples were stained with DAB (3,3′-diaminobenzidine tetrahydrochloride) and with calcofluor white as described earlier (Orczyk et al., 2010). The stained samples were examined under a fluorescence microscope (Nikon Diaphot, epifluorescence optics with excitation 340–380 nm, barrier filter 420 nm and dichroic mirror 400 nm). These conditions allowed the visualization of pathogen structures and host cell necrosis as yellow-orange autofluorescence. Observations were made for at least 30 infection sites per leaf sample collected from two to five inoculated plants.
Enzyme inhibitors – identification of enzymes responsible for the oxidative burst
In order to establish which type of enzyme activity (i.e. NADPH oxidase or peroxidase) was responsible for the accumulation of H2O2 in the cells adjacent to the infection site, two enzyme-specific inhibitors were used. Diphenylene iodonium (DPI), 30 μm in 10 mm Tris pH 7·5, was used as the inhibitor of NADPH oxidases (Grant et al., 2000; Moloi & van der Westhuizenb, 2006). Sodium azide (NaN3), 100 μm in 10 mm Tris pH 7·5, was used as the inhibitor of peroxidases (Grant et al., 2000; Bindschedler et al., 2006). A mock solution (10 mm Tris–HCl, pH 7·5) or the inhibitor was applied to the lower side of the leaf. The experiment design consisted of six seedlings each of Thatcher, TcLr24 and TcLr25, and nine seedlings each of TcLr9, TcLr26 inoculated with brown rust. The inoculated leaves were infiltrated with DPI, NaN3 or mock solution 1, 2 and 3 dpi (TcLr9, TcLr26) or 4 and 5 dpi (TcLr24, TcLr25, Thatcher). Sections 2 cm long were marked on the infiltrated leaves. One hour after the infiltration the leaves were collected and double-stained with DAB and calcofluor white. There was one seedling per each treatment/time combination and the whole experiment was performed in three to five biological replications. The timings of inhibitor application and leaf sampling were chosen based on earlier results (Orczyk et al., 2010) and they corresponded to the highest accumulation of H2O2 in tested lines. The total number of infection sites and the number of DAB-stained infection sites were counted on the 2-cm-long leaf area which was subjected to the mock or inhibitor treatment. The results were used to calculate the percentage of DAB-stained infection sites (Eqn (3)) and the percentage of DAB-stained infection sites after inhibitor treatment (Eqn (4)):
The impact of the inhibitor on H2O2 accumulation in the infection sites was calculated according to Eqn (5) and was presented on a logarithmic scale (Fig. 2).
Expression patterns of peroxidases and NADPH oxidases
The first phase of expression of the analysed genes was observed between 8 and 16 hpi. It was detected for all peroxidases and one of the NADPH oxidase genes, TaRbohD. TaRbohF was the only one of the analysed genes that was not induced shortly after inoculation. Early expression of all tested genes was a very characteristic feature detected in all tested plants: the susceptible cv. Thatcher, the moderately resistant lines TcLr24 and TcLr25, and the highly resistant TcLr9 and TcLr29. This step of the expression corresponded to the first phase of the oxidative burst. It is worth noting that although the pattern of early induction was similar across the genes and the lines, the level of induction was very different. The pattern of expression between 2 and 5 dpi represented the second phase of induction. It additionally differentiated the tested combinations of genes and lines. In the susceptible cv. Thatcher, the second induction was very weak and was detected only for TaPrx107 peroxidase and for both the NADPH oxidases TaRbohD and TaRbohF (Fig. 1). In moderately resistant lines (TcLr24, TcLr25) the second induction phase was found only for TaPrx107 in both lines and TaPrx103 in TcLr24 (Fig. 1). In highly resistant lines (TcLr9, TcLr26) the second phase of induction was strong for TaPrx103 and very strong for TaPrx108 and TaPrx112. In these lines the expression of NADPH oxidases was at a relatively low (TaRbohD) or very low (TaRbohF) level (Fig. 1).
Comparison of expression by cumulative value of transcript rate
Quantitative real-time polymerase chain reaction analysis based on the standard curve method allows the most direct quantification of target cDNA. It was chosen to measure and to compare the accumulation of different transcripts. For each of the analysed genes the definite integral of the function of the accumulation rate between 0 and 5 dpi was calculated. The result, representing the area below the graph of the accumulation rate from 0 to 5 dpi, was designated the cumulative value of transcript rate (CVTR). This parameter was introduced to denote the total transcript accumulation rate during 5 days of infection.
The CVTR (Table 2) allowed direct comparison of different expression patterns of all combinations of genes. It also helped to estimate the possible impact of the expression of each gene on the wheat oxidative response. As with all PCR-based methods, this approach may be prone to systematic errors, but these can be reduced by using the standard curve method, as in this study, to quantify the target molecule, and by complying with the rigorous qPCR rules. It also offers an accessible alternative to unbiased but more expensive direct transcript quantification, such as second-generation RNA sequencing. This approach does not appear to have been described before.
Table 2. Cumulative value of transcript rate (CVTR) of tested genes in susceptible wheat cv. Thatcher and isogenic lines. CVTR is the area below the expression graphs, i.e. of each of the tested genes between 0 and 5 days post-inoculation with Puccinia triticina. qPCR analysis based on the standard curve method was chosen for a direct comparison of all combinations and also for the estimation of the possible impact of each gene's expression on the oxidative response
Cumulative value of transcript rate
Total CVTR of each analysed line
Total CVTR of each analysed gene
The highest CVTR values were obtained for the expression of the two peroxidases TaPrx112 and TaPrx108, these values being about 10 times higher than those of the other peroxidases. The total CVTR obtained for each of the tested lines clearly differentiated the lines and closely corresponded to their resistance levels. The CVTR values of the highly resistant TcLr26 and TcLr9 were more than twice as high as those of the moderately resistant lines and about four times higher than for the susceptible cv. Thatcher.
Effects of enzyme-specific inhibitor treatment
Infiltration of leaves with enzyme-specific inhibitors affected the accumulation of H2O2 in the response to rust infection. The time points selected for the inhibitor treatment corresponded to the strongest oxidative response found in Thatcher and the isogenic lines used in the study (Orczyk et al., 2010). In the susceptible Thatcher, the number of DAB-stained infection sites was reduced after sodium azide treatment to a value of 0·2 relative to that of the mock-treated control (i.e. five times lower) and inhibition was detected at both tested time points 4 and 5 dpi. DPI, an NADPH oxidase inhibitor, reduced the number of oxidative responding sites to 0·8 of the control value (Fig. 2c). In moderately resistant lines (TcLr24, TcLr25) the relative number of DAB-stained infection sites was reduced after sodium azide treatment to 0·4. In TcLr25 the oxidative response was also affected by DPI treatment (to relative values of 0·7 and 0·5 at 4 and 5 dpi, respectively), indicating involvement of the NADPH oxidase (Fig. 2c). In the highly resistant TcLr9 and TcLr26 the inhibitors were applied 1, 2 and 3 dpi. The accumulation of H2O2 was examined in stomatal guard cells 1 dpi and in mesophyll cells 1, 2 and 3 dpi. The H2O2 accumulation in TcLr9 plants treated with NaN3 was depressed in stomatal guard cells to 0·6 of the control value and in mesophyll cells to 0·4, 0·2 and 0·5 of control values at the three successive time points. The results indicated that the peroxidases played a major role in the oxidative response of these lines. The influence of DPI was low or very low (Fig. 2a,b) in all tested lines except TcLr26. The impact of DPI treatment in this particular line was stronger than the impact of sodium azide and was observed in the stomatal guard cells and in the mesophyll cells. The early response of mesophyll cells (1 and 2 dpi) was only slightly changed by either of the inhibitors, while the response 3 dpi was strongly arrested by the DPI, to a relative value of 0·2 (Fig. 2a,b).
Hydrogen peroxide accumulation in the site of pathogen infection is one of the first cytologically detectable symptoms of plant–pathogen interaction. Peroxidases and NADPH oxidases are the two main enzyme systems found to be responsible for ROS accumulation in the response to pathogen infection (for review see Apel & Hirt, 2004) and this reaction could be the result of either one or both of the types of enzyme activity (Trujillo et al., 2004; Asai et al., 2008). It was established earlier that the most characteristic features of compatible and incompatible wheat–rust interactions closely correlated with the specific patterns of the oxidative response (Orczyk et al., 2010). The present work investigated which of the enzyme systems, the apoplastic peroxidases or the membrane-bound NADPH oxidases, plays a predominant role in the generation of ROS in each of these interactions.
The expression patterns of four peroxidases and two NADPH oxidases were analysed in a genetically well-characterized system of wheat isogenic lines with the defined brown rust resistance genes. The possible involvement of either of the enzyme systems was tested with the use of enzyme-specific inhibitors. DPI was used as the inhibitor of NADPH oxidases (Grant et al., 2000; Moloi & van der Westhuizenb, 2006) and NaN3 as the peroxidase inhibitor (Grant et al., 2000; Bindschedler et al., 2006). Four genes of wheat class III peroxidases and two genes of NADPH oxidases were selected based on their reported or expected involvement in biotic stress responses. An oxidative response in different plant–pathogen systems has been found to show a characteristic biphasic pattern. It has also been detected in the interaction of wheat with brown rust (Orczyk et al., 2010). The first peak of expression (8 hpi) of all tested peroxidases (TaPrx103, TaPrx107, TaPrx108, TaPrx112 and TaRbohD) in all analysed lines correlated very well with the first symptoms of the oxidative burst in stomatal guard cells detected in both susceptible and resistant plants (Orczyk et al., 2010). Of the two NADPH oxidases, only TaRbohD was co-induced with the peroxidases shortly after inoculation. The second phase of H2O2 accumulation, observed 4 and 5 dpi in the susceptible Thatcher and moderately resistant Lr24 and Lr25 lines and 1, 2 and 3 dpi in the highly resistant Lr9 and Lr26, was localized in the mesophyll cells surrounding the infection site (Orczyk et al., 2010). In the first group (Thatcher, TcLr24, TcLr25) only one peroxidase, TaPrx107, and both NADPH oxidases were induced 3–4 dpi and remained elevated until 5 dpi. In the highly resistant lines (Lr9 and Lr26) the first phase of expression (8–16 hpi) was immediately followed by the second one (32–48 hpi), which was particularly strong for TaPrx103, TaPrx108 and TaPrx112. The elevated expression of the peroxidases returned to a low level 3–4 dpi. The same peroxidases in the moderately resistant lines did not show the second induction and remained at a low level during the entire infection cycle. Davies et al. (2006) suggested that during the first phase the peroxidase-generated ROS was a part of basal resistance and served as an activator of NADPH oxidase. The former enzyme was a major ROS generator during R-gene mediated resistance. In the wheat–rust system only the oxidative reaction of a line with Lr26 resistance gene could fit this pattern. Although in TcLr26 the peroxidases were strongly induced, the process of ROS accumulation both in stomatal guard cells and mesophyll cells was DPI-sensitive. The results imply that in this particular isogenic line (TcLr26) the activity of NADPH oxidases has a more pronounced role than peroxidases during the entire interaction with rust. This specific reaction might be related to the rye origin of the Lr26 gene (Pathan & Park, 2006) – a species with a basal type of resistance against P. triticina. The rye translocation containing Lr26 also provides other resistance genes: powdery mildew Pm8 (Hanusova et al., 1996), stem rust Sr31 and yellow rust Yr9 (Mago et al., 2005). The pattern of the oxidative response, as well as the multiple pathogen resistance, might be manifestations of a basal rye resistance translocated to wheat. The oxidative burst observed in another highly resistant line, TcLr9, contrary to that of TcLr26, was entirely peroxidase-dependent. It is worth underlining that although the gene expression patterns were similar in both lines, the accumulation of ROS was dependent on different enzyme systems.
To evaluate the possible impact of the genes' expression on the analysed process, the definite integral of the relative accumulation rates between 0 and 5 dpi was calculated and designated the CVTR. Because CVTR takes into account both the expression rate and its duration, it can be used to compare the relative impact of each enzyme on ROS accumulation. This approach showed that the highest CVTR values, obtained for TaPrx108 and TaPrx112, exceeded by several times the values found for TaPrx103 and TaPrx107, indicating the leading role of TaPrx112 and TaPrx108 in the wheat oxidative response to rust infection. This agrees with results showing that the orthologues of Prx107 in barley and rice (PeroxiBase database) the orthologues of Prx108 in Triticum monococcum and rice (Simonetti et al., 2009) and the orthologues of Prx112 in T. monococcum and in oat are pathogen-induced (Simonetti et al., 2009). In another experimental system the expression of Prx111, Prx112 and Prx113 was analysed in wheat roots infested with the nematode Heterodera avenae (Simonetti et al., 2009). In this case Prx112 expression was very similar to that in brown-rust-infected leaves. The fast increase of expression observed on the 4th day after nematode infestation corresponded to the strong induction found on the 2nd day after rust inoculation. It is noteworthy that under these two divergent biotic stresses, i.e. nematodes (Simonetti et al., 2009) and brown rust (this study), TaPrx112 was the most strongly expressed peroxidase. The sum of the cumulative values of all tested peroxidases clearly differed between susceptible, moderately resistant and highly resistant lines. The cumulative values of highly resistant lines were about twice those of the moderately resistant lines and over four times higher than those of the susceptible ones. The results, confirming the important role of the peroxidases in resistance, are consistent with conclusions from other sources. They match the results of direct functional analysis of TaPrx103 and its expected orthologue HvPrx40 (Johrde & Schweizer, 2008). The results of transient/stable expression and silencing experiments confirmed the crucial role of this peroxidase in wheat and barley resistance against B. graminis. The orthologues of TaPrx103 in T. monococcum are also induced by biotic stress (PeroxiBase database; Liu et al., 2005). The findings of the present study indicate a high level of redundancy of class III peroxidases participating in biotic stress response. All of the tested Prx genes showed a similar induction pattern, but the strength of the induction and the sum of the CVTR closely correlated with the resistance level. Consistent with this notion, the finding that silencing of HvPrx40 did not depress ROS accumulation (Johrde & Schweizer, 2008) indicates redundant activity of other peroxidases in this response. Gonzales et al. (2010) reported that 56% of barley resistance quantitative trait loci (QTL) were linked with peroxidase profiling markers. This implies co-localization of the peroxidase gene clusters with basal resistance QTLs and provides another set of data confirming the complex and redundant role of peroxidases in plant resistance to biotic stresses. This conclusion is also consistent with the high number of peroxidase genes found in different species. There are at least 110 wheat class III peroxidases reported in the PeroxiBase database (Oliva et al., 2009).
While the peroxidases have been widely reported to be an integral part of resistance, the roles of NADPH oxidases are more complex. They have been reported to have a regulatory and pleiotropic influence on pathogen-induced processes. Also, their functions might be different in dicots versus monocots, as well as in different types of interaction. The expression of TaRbohF in Thatcher and moderately resistant lines was slightly repressed at 1 dpi and induced approximately fourfold during the next 3 days of the asexual infection cycle of the pathogen. The expression patterns of both TaRbohD and TaRbohF were very similar to the results reported by Pogány et al. (2009) for A. thaliana AtRbohD and AtRbohF. The strongest induction of TaRbohF (2–4 dpi in this report) correlated precisely with the time of massive necrotic reactions in the susceptible and moderately resistant lines, at the end of the asexual infection cycle of the pathogen (Orczyk et al., 2010). This is in agreement with the suggestion that RbohF participates in pathogen-induced cell death (Torres et al., 2002; Trujillo et al., 2006; [Correction added on 19 November 2012 after online publication: The citation “Pogany et al., 2009” has been removed from the preceding sentence].). The results presented by Trujillo et al. (2004, 2006) and Christensen et al. (2004) indicated that in monocots two separate ROS sources might operate: Rboh-dependent superoxide production and peroxidase-dependent generation of H2O2. During the interaction of barley with B. graminis the first Rboh-dependent reaction is localized at the sites of fungal penetration, while peroxidase-related ROS are produced throughout the cells in close proximity to infection sites. The results of this study, based on the use of enzyme-specific inhibitors, indicated that the oxidative burst in wheat was peroxidase dependent in four lines and NADPH-oxidase-dependent in one line. This confirms the notion that the two enzyme systems have divergent roles in different plant–pathogen systems.
The research was financially supported by The National Science Centre, DEC-2011/01/B/NZ9/02 387 (granted to WO) and PBZ-MNiSW-2/3/2006 (granted to WO and AN-O).