ROS are known to play major roles in various plant and microorganism developmental processes, such as cell elongation (root hairs, pollen tube or appressoria growth) and during biotic interactions (Figure 1). In order to avoid ROS accumulation leading to cell death (Mittler et al. 2004) organisms have evolved enzymatic and non-enzymatic antioxidant mechanisms constantly generating and deteriorating ROS (Figure 1). In plants, ROS are unavoidable by-products of biochemical pathways, such as glycolysis and photosynthesis. As a result, plants have evolved enzymatic and non-enzymatic antioxidant mechanisms to eliminate ROS and avoid oxidative destruction (Apel and Hirt 2004). On the other hand, ROS production is necessary for cell elongation (root hairs, appresoria growth) and plant-microorganism interactions. It is therefore necessary for the plant to possess very complex and well-tuned ROS producing and scavenging systems capable of maintaining ROS homeostasis in the cells.
Enzymatic ROS scavenging mechanisms involved in plant-microorganism interactions
The large battery of ROS scavenging enzymes included in the “ROS gene network” contains catalases (Kat), superoxide dismutases (SOD), ascorbate peroxidases (APx, detected in plants) cytochrome C peroxidases (CcP, detected in fungi). They are present in several intracellular compartments as well as in the apoplast in order to regulate both intracellular and extracellular ROS accumulation (Mittler et al. 2004). Considering the extracellular region, both plant and microorganism are capable of regulating the ROS level in this area during the early steps of the interaction (Figure 2).
Figure 2. ROS regulation during development and plant-microorganism interactions. Constitutive ROS production (photosynthesis and respiration) have also been included in the cartoon. T-bars and arrows correspond to scavenging and production of ROS respectively. Pink lines stand for ROS involved in plant-microorganism interactions and blue lines for ROS involved in the development
Download figure to PowerPoint
The symbiotic Rhizobia appears to have an efficient antioxidant defense. Indeed, although ROS were present in the infection threads they weren't detected inside the bacteria progressing within the infection thread (Santos et al. 2001; Rubio et al. 2004). Indeed, S. meliloti possesses two SOD that convert O2•− to O2 and H2O2 (Santos et al. 2000; Hérouart et al. 2002) and three heme b containing catalases, which are able to scavenge H2O2 (Hérouart et al. 1996; Ardissone et al. 2004). Bacterial catalases appear to play an important role in the nodule formation process as the double katB/katC and katA/katC mutants of S. meliloti are strongly impaired in nodule formation (Jamet et al. 2003). In addition to the catalases, the S. meliloti genome contains three thiol peroxidase encoding genes: the alkyl hydroperoxide reductase ahpC-like and two organic hydroperoxide resistance ohr-like genes. Both types of enzyme display biochemically equivalent functions and catalyze the reduction of organic peroxides to the corresponding less toxic organic alcohols.
In the case of a pathogenic interaction between rice and Magnaporthe griseae, the causal agent of rice blast disease, the fungus has to overcome the plants innate immunity in order to infect it. The massive production of ROS during the early stages of interaction is part of the plants innate immunity response. To overcome this line of defense, M. griseae must be able to counter the oxidative burst by producing ROS scavenging enzymes. A novel gene related to pathogenecity has recently been isolated in M. griseae: Defense Suppressor 1 (DES1) (Chi et al. 2009) . Δdes1 deficient mutants were hypersensitive to exogenous oxidative stress and the transcription of extracellular enzymes such as peroxidases and laccases were severely reduced. In interaction with a susceptible rice cultivar, the mutants displayed an important reduction of infectious hyphal extension, leading to a decrease in pathogenicity. Interestingly, the Δdes1 deficient mutants recovered their normal infectious growth when interacting with DPI treated plant tissue. These results strongly support the possibility that ROS play a major role in the first line of plant defenses at the cell surface both as toxic molecules as well as signalling actors.
Mechanisms to generate ROS
Several different enzymes have been implicated in the generation of ROS. Among these, NADPH oxidases (NOx) correspond to one of the most studied systems that play an important role in the production of superoxide radicals during the oxidative burst to defend cells from invasion. NOx are integral membrane proteins capable of oxidizing NADPH in the cell as well as reducing molecular oxygen into superoxide radicals in the apoplast (Sumimoto 2008), which is quickly dismutated into H2O2 either spontaneously or by SOD enzymes. ROS produced by the NADPH oxidases function in defense, development and redox-dependent signalling. They share common structural features and are evolutionarily of ancient origin and thus ubiquitous in multicellular eukaryotes (Bedard and Krause 2007; Bedard et al. 2007). In plants, NADPH oxidases form a small multigenic family and are involved in diverse events including innate immunity development. Due to the fact that ROS are toxic and in many cases short-lived, the activity of these oxidases is tightly regulated both temporally and spatially.
The use of DPI, that inhibits flavoproteins such as NOx, and abolished ROS production, strongly supports the possible involvement of M. truncatula NOx homologues in ROS production. Moreover, a DPI treatment during the early stages of M. truncatula – S. meliloti interaction not only abolished ROS production but also suppressed root hair curling and infection thread formation (Peleg-Grossman et al. 2007; Cardenas et al. 2008). These results emphasize the involvement of M. truncatula NADPH oxidase homologues in the early steps of Rhizobium infection.
The involvement of plant NOx in plant-microorganism interactions have clearly been shown (see earlier reviews Apel and Hirt 2004; Torres and Dangl 2005). NOx are present in all the fungi forming fruit bodies where they seem to participate in sexual reproduction. The inactivation of Aspergillus nidulans NoxA gene resulted in a decrease of ROS production, inhibition of the formation of cleistothecia at early stages of development, stimulation of mycelium growth and suppression of asexual reproduction (Lara-Ortiz et al. 2003). In addition NOx from fungi are also important during the infection process. Moreover in the symbiotic interaction between the fungus Epistle fistulae and the plant Folium perenne, a NOXA deficient fungus mutant is unable to undergo symbiosis and induces plant death. This shows that fungus produced ROS that also play a major role in the establishment of this symbiosis (Takemoto et al. 2006; Tanaka et al. 2006).
More recently, several reports from the microbe side indicate a major role of these genes in the pathogenicity process (Egan et al. 2007; Giesbert et al. 2008) and thus, play a positive role for the pathogen. Accordingly, during plant infection, NOx from M. grisea generate ROS. This oxidative burst is associated with the development of specialised infection structures called appressoria. Pharmacological scavenging of these oxygen radicals significantly delayed the development of appressoria and affected their morphology. Using a genetic approach targeting two NOx genes, Egan et al. (2007) showed that these genes are independently required for pathogenicity of M. grisea (inability to initiate appressorium-mediated cuticle penetration for the mutants) and are involved in ROS production (Egan et al. 2007). In a similar approach, the deletion of a putative NOx from the ergot fungus of ryegrass, Claviceps purpurea, has an impact on germination of conidia and pathogenicity, although its involvement in focusing ROS production has not been shown (Giesbert et al. 2008).
Other proteins, such as class III peroxidases, regulate for ROS homeostasis. Class III peroxidases are only detected in Viridiplantae and are present as large multigenic families in all land plants (Bakalovic et al. 2006). Released from the cell surface into the apoplast, peroxidases are an important class of enzymes responsible for the stress-induced formation and degradation of ROS (Bolwell et al. 2002; Bindschedler et al. 2006; Fecht-Christoffers et al. 2006). Apart from their indirect role in H2O2 detoxification through its peroxidative activities, some apoplastic class III peroxidases can also generate O2•− or H2O2 at physiological pHs via its oxidative cycle (Minibayeva et al. 2009). The cell wall has an enormous capacity to retain proteins in normal growth conditions, most of the peroxidases for instance, which may be released following abiotic stress. The involvement of class III peroxidases during the symbiotic process has already been observed. For example, Rip1, encoding a peroxidase from Medicago is rapidly and transiently induced by Rhizobium meliloti or after NFs treatment (Cook et al. 1995). Moreover, this gene is induced by H2O2 (Ramu et al. 2002).
More recently, a class III peroxidase (Srprx1) has been shown to be crucial for the bacterial invasion of the tropical Legume, Sesbania rostrata (Den Herder et al. 2007). The expression of Srprx1 is strictly dependent on bacterial nodulation factors (NFs) and could be modulated by H2O2, a downstream signal for crack-entry invasion. Its expression was not induced after wounding or pathogen attack, indicating that the peroxidase is a symbiosis-specific isoform. More interestingly, lack of Srprx1 gene expression could cause an aberrant structure of the infection threads (Den Herder et al. 2007).
OsPrx53, encoding a peroxidase from rice, is the strongest gene induced after Glomus infection (Guimil et al. 2005). Peroxidases seem important for the initiation of symbiosis but no direct evidence has demonstrated their implication for ROS production in the early steps of interaction and the development of the infection (Guimil et al. 2005).
Furthermore, the involvement of peroxidases in H2O2 synthesis during plant-pathogen interactions has been recently highlighted (Choi et al. 2007). Thus, H2O2 production is also compromised after inoculation of Capsicum annum, silenced for a peroxidase, by avirulent Xanthomonas campestris bacteria (Choi et al. 2007). This clearly demonstrate the peroxidase involvement in ROS production (Choi et al. 2007).
Although NOx and peroxidases represent the main characterized plant ROS producing systems at the plant cell surface, one should note that several oxidative and reductive systems are present in the plant plasmalemma (Vuletic et al. 2005). More interestingly, special attention has been recently paid to the plasma membrane microdomains. Recently, a proteomic approach based on the purification of lipid-rafts in plasma membrane from M. truncatula identified several putative ROS producing systems, including peroxidase (Furt et al. 2007; Lefebvre et al. 2007). This approach does not allow the identification of NOx although their presence has been previously shown in elicitor treated tobacco cells (Mongrand et al. 2004).
Other possible sources for H2O2 in the Legume –Rhizobium symbiosis are germin-like oxalate oxidases or diamine oxidases (Wisniewski et al. 2000). Indeed, a germin-like oxidase from Pisum sativum has been characterized (PsGER1). This protein has a superoxide dismutase activity, and is associated with nodules (Gucciardo et al. 2007).